From ceae29a41166ae524a4fc177b17f3922f3e16a8b Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Tue, 4 Nov 2008 13:44:07 +0000
Subject: [PATCH 001/532] Avoid assigning a const char * to a char *.

MFC after:	3 weeks
---
 lib/libutil/login_class.c | 11 +++++------
 1 file changed, 5 insertions(+), 6 deletions(-)

diff --git a/lib/libutil/login_class.c b/lib/libutil/login_class.c
index 94cb3aa14e6..507451958d1 100644
--- a/lib/libutil/login_class.c
+++ b/lib/libutil/login_class.c
@@ -142,14 +142,13 @@ substvar(const char * var, const struct passwd * pwd, int hlen, int pch, int nle
 	int	tildes = 0;
 	int	dollas = 0;
 	char	*p;
+	const char *q;
 
 	if (pwd != NULL) {
-	    /* Count the number of ~'s in var to substitute */
-	    for (p = (char *)var; (p = strchr(p, '~')) != NULL; p++)
-		++tildes;
-	    /* Count the number of $'s in var to substitute */
-	    for (p = (char *)var; (p = strchr(p, '$')) != NULL; p++)
-		++dollas;
+	    for (q = var; *q != '\0'; ++q) {
+		tildes += (*q == '~');
+		dollas += (*q == '$');
+	    }
 	}
 
 	np = malloc(strlen(var) + (dollas * nlen)

From e37e92fe4cd3149b76061e7b880f462c8c028e74 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Tue, 4 Nov 2008 13:44:43 +0000
Subject: [PATCH 002/532] Avoid assigning a const char * to a char *.

MFC after:	3 weeks
---
 lib/libutil/login_cap.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/lib/libutil/login_cap.c b/lib/libutil/login_cap.c
index 529ce822732..8fee76050b0 100644
--- a/lib/libutil/login_cap.c
+++ b/lib/libutil/login_cap.c
@@ -61,6 +61,8 @@ static char * internal_string = NULL;
 static size_t internal_arraysz = 0;
 static const char ** internal_array = NULL;
 
+static char path_login_conf[] = _PATH_LOGIN_CONF;
+
 static char *
 allocstr(const char *str)
 {
@@ -215,15 +217,14 @@ login_getclassbyname(char const *name, const struct passwd *pwd)
 
 	if (dir && snprintf(userpath, MAXPATHLEN, "%s/%s", dir,
 			    _FILE_LOGIN_CONF) < MAXPATHLEN) {
-	    login_dbarray[i] = userpath;
 	    if (_secure_path(userpath, pwd->pw_uid, pwd->pw_gid) != -1)
-		i++;		/* only use 'secure' data */
+		login_dbarray[i++] = userpath;
 	}
 	/*
 	 * XXX: Why to add the system database if the class is `me'?
 	 */
-	if (_secure_path(_PATH_LOGIN_CONF, 0, 0) != -1)
-	    login_dbarray[i++] = _PATH_LOGIN_CONF;
+	if (_secure_path(path_login_conf, 0, 0) != -1)
+	    login_dbarray[i++] = path_login_conf;
 	login_dbarray[i] = NULL;
 
 	memset(lc, 0, sizeof(login_cap_t));

From 06927900e3ce8e03f2d01e69c5f047739eb421af Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Tue, 4 Nov 2008 13:49:18 +0000
Subject: [PATCH 003/532] Committed to wrong branch.

---
 lib/libutil/login_cap.c   |  9 ++++-----
 lib/libutil/login_class.c | 11 ++++++-----
 2 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/lib/libutil/login_cap.c b/lib/libutil/login_cap.c
index 8fee76050b0..529ce822732 100644
--- a/lib/libutil/login_cap.c
+++ b/lib/libutil/login_cap.c
@@ -61,8 +61,6 @@ static char * internal_string = NULL;
 static size_t internal_arraysz = 0;
 static const char ** internal_array = NULL;
 
-static char path_login_conf[] = _PATH_LOGIN_CONF;
-
 static char *
 allocstr(const char *str)
 {
@@ -217,14 +215,15 @@ login_getclassbyname(char const *name, const struct passwd *pwd)
 
 	if (dir && snprintf(userpath, MAXPATHLEN, "%s/%s", dir,
 			    _FILE_LOGIN_CONF) < MAXPATHLEN) {
+	    login_dbarray[i] = userpath;
 	    if (_secure_path(userpath, pwd->pw_uid, pwd->pw_gid) != -1)
-		login_dbarray[i++] = userpath;
+		i++;		/* only use 'secure' data */
 	}
 	/*
 	 * XXX: Why to add the system database if the class is `me'?
 	 */
-	if (_secure_path(path_login_conf, 0, 0) != -1)
-	    login_dbarray[i++] = path_login_conf;
+	if (_secure_path(_PATH_LOGIN_CONF, 0, 0) != -1)
+	    login_dbarray[i++] = _PATH_LOGIN_CONF;
 	login_dbarray[i] = NULL;
 
 	memset(lc, 0, sizeof(login_cap_t));
diff --git a/lib/libutil/login_class.c b/lib/libutil/login_class.c
index 507451958d1..94cb3aa14e6 100644
--- a/lib/libutil/login_class.c
+++ b/lib/libutil/login_class.c
@@ -142,13 +142,14 @@ substvar(const char * var, const struct passwd * pwd, int hlen, int pch, int nle
 	int	tildes = 0;
 	int	dollas = 0;
 	char	*p;
-	const char *q;
 
 	if (pwd != NULL) {
-	    for (q = var; *q != '\0'; ++q) {
-		tildes += (*q == '~');
-		dollas += (*q == '$');
-	    }
+	    /* Count the number of ~'s in var to substitute */
+	    for (p = (char *)var; (p = strchr(p, '~')) != NULL; p++)
+		++tildes;
+	    /* Count the number of $'s in var to substitute */
+	    for (p = (char *)var; (p = strchr(p, '$')) != NULL; p++)
+		++dollas;
 	}
 
 	np = malloc(strlen(var) + (dollas * nlen)

From ea234e656896e0185f838fa5be93dec0a241c98f Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Tue, 4 Nov 2008 13:54:07 +0000
Subject: [PATCH 004/532] MFH (r184633, r184635): build at WARNS level 6

---
 lib/libutil/Makefile      |  2 ++
 lib/libutil/login_cap.c   |  9 +++++----
 lib/libutil/login_class.c | 11 +++++------
 3 files changed, 12 insertions(+), 10 deletions(-)

diff --git a/lib/libutil/Makefile b/lib/libutil/Makefile
index ce544c320a8..777f95e9a1e 100644
--- a/lib/libutil/Makefile
+++ b/lib/libutil/Makefile
@@ -15,6 +15,8 @@ SRCS=	_secure_path.c auth.c gr_util.c expand_number.c flopen.c fparseln.c \
 	stub.c trimdomain.c uucplock.c
 INCS=	libutil.h login_cap.h
 
+WARNS?=	6
+
 CFLAGS+= -DLIBC_SCCS
 
 .if ${MK_INET6_SUPPORT} != "no"
diff --git a/lib/libutil/login_cap.c b/lib/libutil/login_cap.c
index 529ce822732..8fee76050b0 100644
--- a/lib/libutil/login_cap.c
+++ b/lib/libutil/login_cap.c
@@ -61,6 +61,8 @@ static char * internal_string = NULL;
 static size_t internal_arraysz = 0;
 static const char ** internal_array = NULL;
 
+static char path_login_conf[] = _PATH_LOGIN_CONF;
+
 static char *
 allocstr(const char *str)
 {
@@ -215,15 +217,14 @@ login_getclassbyname(char const *name, const struct passwd *pwd)
 
 	if (dir && snprintf(userpath, MAXPATHLEN, "%s/%s", dir,
 			    _FILE_LOGIN_CONF) < MAXPATHLEN) {
-	    login_dbarray[i] = userpath;
 	    if (_secure_path(userpath, pwd->pw_uid, pwd->pw_gid) != -1)
-		i++;		/* only use 'secure' data */
+		login_dbarray[i++] = userpath;
 	}
 	/*
 	 * XXX: Why to add the system database if the class is `me'?
 	 */
-	if (_secure_path(_PATH_LOGIN_CONF, 0, 0) != -1)
-	    login_dbarray[i++] = _PATH_LOGIN_CONF;
+	if (_secure_path(path_login_conf, 0, 0) != -1)
+	    login_dbarray[i++] = path_login_conf;
 	login_dbarray[i] = NULL;
 
 	memset(lc, 0, sizeof(login_cap_t));
diff --git a/lib/libutil/login_class.c b/lib/libutil/login_class.c
index 94cb3aa14e6..507451958d1 100644
--- a/lib/libutil/login_class.c
+++ b/lib/libutil/login_class.c
@@ -142,14 +142,13 @@ substvar(const char * var, const struct passwd * pwd, int hlen, int pch, int nle
 	int	tildes = 0;
 	int	dollas = 0;
 	char	*p;
+	const char *q;
 
 	if (pwd != NULL) {
-	    /* Count the number of ~'s in var to substitute */
-	    for (p = (char *)var; (p = strchr(p, '~')) != NULL; p++)
-		++tildes;
-	    /* Count the number of $'s in var to substitute */
-	    for (p = (char *)var; (p = strchr(p, '$')) != NULL; p++)
-		++dollas;
+	    for (q = var; *q != '\0'; ++q) {
+		tildes += (*q == '~');
+		dollas += (*q == '$');
+	    }
 	}
 
 	np = malloc(strlen(var) + (dollas * nlen)

From 1b3515f39bcd2de8409fb54b2e4355553d016a81 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Fri, 30 Jan 2009 13:54:03 +0000
Subject: [PATCH 005/532] WIP

---
 lib/libutil/Makefile          |  10 +-
 lib/libutil/libutil.h         |   9 +
 lib/libutil/quotafile.3       |  69 +++++++
 lib/libutil/quotafile.c       | 272 ++++++++++++++++++++++++++
 libexec/rpc.rquotad/Makefile  |   4 +-
 libexec/rpc.rquotad/rquotad.c |  28 +--
 sys/ufs/ufs/quota.h           |  42 ++++-
 sys/ufs/ufs/ufs_quota.c       | 346 +++++++++++++++++++++++++++++++---
 sys/ufs/ufs/ufs_vfsops.c      |  12 ++
 sys/ufs/ufs/ufsmount.h        |   1 +
 usr.sbin/edquota/Makefile     |   3 +
 usr.sbin/edquota/edquota.c    |  99 +++-------
 12 files changed, 760 insertions(+), 135 deletions(-)
 create mode 100644 lib/libutil/quotafile.3
 create mode 100644 lib/libutil/quotafile.c

diff --git a/lib/libutil/Makefile b/lib/libutil/Makefile
index 1ea4c73767d..178948d3a16 100644
--- a/lib/libutil/Makefile
+++ b/lib/libutil/Makefile
@@ -12,8 +12,8 @@ SRCS=	_secure_path.c auth.c expand_number.c flopen.c fparseln.c gr_util.c \
 	hexdump.c humanize_number.c kinfo_getfile.c kinfo_getvmmap.c kld.c \
 	login.c login_auth.c login_cap.c \
 	login_class.c login_crypt.c login_ok.c login_times.c login_tty.c \
-	logout.c logwtmp.c pidfile.c property.c pty.c pw_util.c realhostname.c \
-	stub.c trimdomain.c uucplock.c
+	logout.c logwtmp.c pidfile.c property.c pty.c pw_util.c quotafile.c \
+	realhostname.c stub.c trimdomain.c uucplock.c
 INCS=	libutil.h login_cap.h
 
 WARNS?=	6
@@ -30,7 +30,7 @@ MAN+=	kld.3 login.3 login_auth.3 login_tty.3 logout.3 logwtmp.3 pty.3 \
 	login_cap.3 login_class.3 login_times.3 login_ok.3 \
 	_secure_path.3 uucplock.3 property.3 auth.3 realhostname.3 \
 	realhostname_sa.3 trimdomain.3 fparseln.3 humanize_number.3 \
-	pidfile.3 flopen.3 expand_number.3
+	pidfile.3 flopen.3 expand_number.3 quotafile.3
 MAN+=	login.conf.5 auth.conf.5
 MLINKS+= kld.3 kld_isloaded.3 kld.3 kld_load.3
 MLINKS+= property.3 properties_read.3  property.3 properties_free.3
@@ -58,5 +58,9 @@ MLINKS+=pidfile.3 pidfile_open.3 \
 	pidfile.3 pidfile_write.3 \
 	pidfile.3 pidfile_close.3 \
 	pidfile.3 pidfile_remove.3
+MLINKS+=quotafile.3 quota_open.3 \
+	quotafile.3 quota_read.3 \
+	quotafile.3 quota_write.3 \
+	quotafile.3 quota_close.3
 
 .include <bsd.lib.mk>
diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index 3187fb37349..48ab2188c35 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -140,6 +140,15 @@ int pidfile_close(struct pidfh *pfh);
 int pidfile_remove(struct pidfh *pfh);
 #endif
 
+#ifdef _UFS_UFS_QUOTA_H_
+struct quotafile;
+struct quotafile *quota_open(const char *);
+struct quotafile *quota_create(const char *);
+void quota_close(struct quotafile *);
+int quota_read(struct quotafile *, struct dqblk *, int);
+int quota_write(struct quotafile *, const struct dqblk *, int);
+#endif
+
 __END_DECLS
 
 #define UU_LOCK_INUSE (1)
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
new file mode 100644
index 00000000000..a414a99a027
--- /dev/null
+++ b/lib/libutil/quotafile.3
@@ -0,0 +1,69 @@
+.\"-
+.\" Copyright (c) 2008 Dag-Erling Coïdan Smørgrav
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in the
+.\"    documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd November 4, 2008
+.Dt QUOTAFILE 3
+.Os
+.Sh NAME
+.Nm quota_open
+.Nm quota_create
+.Nm quota_read
+.Nm quota_write
+.Nm quota_close
+.Nd "Manipulate quota files"
+.Sh LIBRARY
+.Lb libutil
+.Sh SYNOPSIS
+.In ufs/ufs/quota.h
+.In libutil.h
+.Ft "struct quotafile *"
+.Fn quota_open "const char *path"
+.Ft "struct quotafile *"
+.Fn quota_create "const char *path"
+.Ft int
+.Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int type"
+.Ft int
+.Fn quota_write "struct quotafile *qf" "const struct dqblk *dqb" "int type"
+.Ft int
+.Fn quota_close "struct quotafile *qf"
+.Sh DESCRIPTION
+.Sh RETURN VALUES
+.Sh SEE ALSO
+.Xr quotactl 2 ,
+.Xr quota.user 5 ,
+.Xr quota.group 5
+.Sh HISTORY
+The
+.Nm
+functions first appeared in
+.Fx 8.0 .
+.Sh AUTHORS
+.An -nosplit
+The
+.Nm
+functions and this manual page were written by
+.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org .
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
new file mode 100644
index 00000000000..a4258b1ef30
--- /dev/null
+++ b/lib/libutil/quotafile.c
@@ -0,0 +1,272 @@
+/*-
+ * Copyright (c) 2008 Dag-Erling Coïdan Smørgrav
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer
+ *    in this position and unchanged.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#include <sys/types.h>
+#include <sys/endian.h>
+#include <sys/stat.h>
+
+#include <ufs/ufs/quota.h>
+
+#include <errno.h>
+#include <fcntl.h>
+#include <grp.h>
+#include <pwd.h>
+#include <libutil.h>
+#include <stdint.h>
+#include <stdlib.h>
+#include <string.h>
+#include <unistd.h>
+
+struct quotafile {
+	int fd;
+	int type; /* 32 or 64 */
+};
+
+struct quotafile *
+quota_open(const char *fn)
+{
+	struct quotafile *qf;
+	struct dqhdr64 dqh;
+	int serrno;
+
+	if ((qf = calloc(1, sizeof(*qf))) == NULL)
+		return (NULL);
+	if ((qf->fd = open(fn, O_RDWR)) < 0) {
+		serrno = errno;
+		free(qf);
+		errno = serrno;
+		return (NULL);
+	}
+	qf->type = 32;
+	switch (read(qf->fd, &dqh, sizeof(dqh))) {
+	case -1:
+		serrno = errno;
+		close(qf->fd);
+		free(qf);
+		errno = serrno;
+		return (NULL);
+	case sizeof(dqh):
+		if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) != 0) {
+			/* no magic, assume 32 bits */
+			qf->type = 32;
+			return (qf);
+		}
+		if (be32toh(dqh.dqh_version) != Q_DQHDR64_VERSION ||
+		    be32toh(dqh.dqh_hdrlen) != sizeof(struct dqhdr64) ||
+		    be32toh(dqh.dqh_reclen) != sizeof(struct dqblk64)) {
+			/* correct magic, wrong version / lengths */
+			close(qf->fd);
+			free(qf);
+			errno = EINVAL;
+			return (NULL);
+		}
+		qf->type = 64;
+		return (qf);
+	default:
+		qf->type = 32;
+		return (qf);
+	}
+	/* not reached */
+}
+
+struct quotafile *
+quota_create(const char *fn)
+{
+	struct quotafile *qf;
+	struct dqhdr64 dqh;
+	struct group *grp;
+	int serrno;
+
+	if ((qf = calloc(1, sizeof(*qf))) == NULL)
+		return (NULL);
+	if ((qf->fd = open(fn, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0) {
+		serrno = errno;
+		free(qf);
+		errno = serrno;
+		return (NULL);
+	}
+	qf->type = 64;
+	memset(&dqh, 0, sizeof(dqh));
+	memcpy(dqh.dqh_magic, Q_DQHDR64_MAGIC, sizeof(dqh.dqh_magic));
+	dqh.dqh_version = htobe32(Q_DQHDR64_VERSION);
+	dqh.dqh_hdrlen = htobe32(sizeof(struct dqhdr64));
+	dqh.dqh_reclen = htobe32(sizeof(struct dqblk64));
+	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh)) {
+		serrno = errno;
+		unlink(fn);
+		close(qf->fd);
+		free(qf);
+		errno = serrno;
+		return (NULL);
+	}
+	grp = getgrnam(QUOTAGROUP);
+	fchown(qf->fd, 0, grp ? grp->gr_gid : 0);
+	fchmod(qf->fd, 0640);
+	return (qf);
+}
+
+void
+quota_close(struct quotafile *qf)
+{
+
+	close(qf->fd);
+	free(qf);
+}
+
+static int
+quota_read32(struct quotafile *qf, struct dqblk *dqb, int id)
+{
+	struct dqblk32 dqb32;
+	off_t off;
+
+	off = id * sizeof(struct dqblk32);
+	if (lseek(qf->fd, off, SEEK_SET) == -1)
+		return (-1);
+	switch (read(qf->fd, &dqb32, sizeof(dqb32))) {
+	case 0:
+		memset(&dqb, 0, sizeof(*dqb));
+		return (0);
+	case sizeof(dqb32):
+		dqb->dqb_bhardlimit = dqb32.dqb_bhardlimit;
+		dqb->dqb_bsoftlimit = dqb32.dqb_bsoftlimit;
+		dqb->dqb_curblocks = dqb32.dqb_curblocks;
+		dqb->dqb_ihardlimit = dqb32.dqb_ihardlimit;
+		dqb->dqb_isoftlimit = dqb32.dqb_isoftlimit;
+		dqb->dqb_curinodes = dqb32.dqb_curinodes;
+		dqb->dqb_btime = dqb32.dqb_btime;
+		dqb->dqb_itime = dqb32.dqb_itime;
+		return (0);
+	default:
+		return (-1);
+	}
+}
+
+static int
+quota_read64(struct quotafile *qf, struct dqblk *dqb, int id)
+{
+	struct dqblk64 dqb64;
+	off_t off;
+
+	off = sizeof(struct dqhdr64) + id * sizeof(struct dqblk64);
+	if (lseek(qf->fd, off, SEEK_SET) == -1)
+		return (-1);
+	switch (read(qf->fd, &dqb64, sizeof(dqb64))) {
+	case 0:
+		memset(&dqb, 0, sizeof(*dqb));
+		return (0);
+	case sizeof(dqb64):
+		dqb->dqb_bhardlimit = be64toh(dqb64.dqb_bhardlimit);
+		dqb->dqb_bsoftlimit = be64toh(dqb64.dqb_bsoftlimit);
+		dqb->dqb_curblocks = be64toh(dqb64.dqb_curblocks);
+		dqb->dqb_ihardlimit = be64toh(dqb64.dqb_ihardlimit);
+		dqb->dqb_isoftlimit = be64toh(dqb64.dqb_isoftlimit);
+		dqb->dqb_curinodes = be64toh(dqb64.dqb_curinodes);
+		dqb->dqb_btime = be64toh(dqb64.dqb_btime);
+		dqb->dqb_itime = be64toh(dqb64.dqb_itime);
+		return (0);
+	default:
+		return (-1);
+	}
+}
+
+int
+quota_read(struct quotafile *qf, struct dqblk *dqb, int id)
+{
+
+	switch (qf->type) {
+	case 32:
+		return quota_read32(qf, dqb, id);
+	case 64:
+		return quota_read64(qf, dqb, id);
+	default:
+		errno = EINVAL;
+		return (-1);
+	}
+	/* not reached */
+}
+
+#define CLIP32(u64) ((u64) > UINT32_MAX ? UINT32_MAX : (uint32_t)(u64))
+
+static int
+quota_write32(struct quotafile *qf, const struct dqblk *dqb, int id)
+{
+	struct dqblk32 dqb32;
+	off_t off;
+
+	dqb32.dqb_bhardlimit = CLIP32(dqb->dqb_bhardlimit);
+	dqb32.dqb_bsoftlimit = CLIP32(dqb->dqb_bsoftlimit);
+	dqb32.dqb_curblocks = CLIP32(dqb->dqb_curblocks);
+	dqb32.dqb_ihardlimit = CLIP32(dqb->dqb_ihardlimit);
+	dqb32.dqb_isoftlimit = CLIP32(dqb->dqb_isoftlimit);
+	dqb32.dqb_curinodes = CLIP32(dqb->dqb_curinodes);
+	dqb32.dqb_btime = CLIP32(dqb->dqb_btime);
+	dqb32.dqb_itime = CLIP32(dqb->dqb_itime);
+
+	off = id * sizeof(struct dqblk32);
+	if (lseek(qf->fd, off, SEEK_SET) == -1)
+		return (-1);
+	return (write(qf->fd, &dqb32, sizeof(dqb32)));
+}
+
+static int
+quota_write64(struct quotafile *qf, const struct dqblk *dqb, int id)
+{
+	struct dqblk64 dqb64;
+	off_t off;
+
+	dqb64.dqb_bhardlimit = htobe64(dqb->dqb_bhardlimit);
+	dqb64.dqb_bsoftlimit = htobe64(dqb->dqb_bsoftlimit);
+	dqb64.dqb_curblocks = htobe64(dqb->dqb_curblocks);
+	dqb64.dqb_ihardlimit = htobe64(dqb->dqb_ihardlimit);
+	dqb64.dqb_isoftlimit = htobe64(dqb->dqb_isoftlimit);
+	dqb64.dqb_curinodes = htobe64(dqb->dqb_curinodes);
+	dqb64.dqb_btime = htobe64(dqb->dqb_btime);
+	dqb64.dqb_itime = htobe64(dqb->dqb_itime);
+
+	off = sizeof(struct dqhdr64) + id * sizeof(struct dqblk64);
+	if (lseek(qf->fd, off, SEEK_SET) == -1)
+		return (-1);
+	return (write(qf->fd, &dqb64, sizeof(dqb64)));
+}
+
+int
+quota_write(struct quotafile *qf, const struct dqblk *dqb, int id)
+{
+
+	switch (qf->type) {
+	case 32:
+		return quota_write32(qf, dqb, id);
+	case 64:
+		return quota_write64(qf, dqb, id);
+	default:
+		errno = EINVAL;
+		return (-1);
+	}
+	/* not reached */
+}
diff --git a/libexec/rpc.rquotad/Makefile b/libexec/rpc.rquotad/Makefile
index feacce57ef4..fb3f38bc90c 100644
--- a/libexec/rpc.rquotad/Makefile
+++ b/libexec/rpc.rquotad/Makefile
@@ -6,7 +6,7 @@ MAN  =	rpc.rquotad.8
 
 WARNS ?= 6
 
-DPADD=	${LIBRPCSVC}
-LDADD=	-lrpcsvc
+DPADD=	${LIBRPCSVC} ${LIBUTIL}
+LDADD=	-lrpcsvc -lutil
 
 .include <bsd.prog.mk>
diff --git a/libexec/rpc.rquotad/rquotad.c b/libexec/rpc.rquotad/rquotad.c
index 559a57a1f33..63bb0c76be6 100644
--- a/libexec/rpc.rquotad/rquotad.c
+++ b/libexec/rpc.rquotad/rquotad.c
@@ -23,6 +23,7 @@ __FBSDID("$FreeBSD$");
 #include <errno.h>
 #include <fstab.h>
 #include <grp.h>
+#include <libutil.h>
 #include <pwd.h>
 #include <signal.h>
 #include <stdio.h>
@@ -247,9 +248,10 @@ initfs(void)
 int
 getfsquota(long id, char *path, struct dqblk *dqblk)
 {
+	struct quotafile *qf;
 	struct stat st_path;
 	struct fs_stat *fs;
-	int	qcmd, fd, ret = 0;
+	int qcmd, ret = 0;
 
 	if (stat(path, &st_path) < 0)
 		return (0);
@@ -265,32 +267,16 @@ getfsquota(long id, char *path, struct dqblk *dqblk)
 		if (quotactl(fs->fs_file, qcmd, id, dqblk) == 0)
 			return (1);
 
-		if ((fd = open(fs->qfpathname, O_RDONLY)) < 0) {
+		if ((qf = quota_open(fs->qfpathname)) == NULL) {
 			syslog(LOG_ERR, "open error: %s: %m", fs->qfpathname);
 			return (0);
 		}
-		if (lseek(fd, (off_t)(id * sizeof(struct dqblk)), L_SET) == (off_t)-1) {
-			close(fd);
-			return (1);
-		}
-		switch (read(fd, dqblk, sizeof(struct dqblk))) {
-		case 0:
-			/*
-                         * Convert implicit 0 quota (EOF)
-                         * into an explicit one (zero'ed dqblk)
-                         */
-			bzero(dqblk, sizeof(struct dqblk));
-			ret = 1;
-			break;
-		case sizeof(struct dqblk):	/* OK */
-			ret = 1;
-			break;
-		default:	/* ERROR */
+		if (quota_read(qf, dqblk, id) != 0) {
 			syslog(LOG_ERR, "read error: %s: %m", fs->qfpathname);
-			close(fd);
+			quota_close(qf);
 			return (0);
 		}
-		close(fd);
+		quota_close(qf);
 	}
 	return (ret);
 }
diff --git a/sys/ufs/ufs/quota.h b/sys/ufs/ufs/quota.h
index 28f4e92d8fc..82b0170b036 100644
--- a/sys/ufs/ufs/quota.h
+++ b/sys/ufs/ufs/quota.h
@@ -81,10 +81,13 @@
 
 #define	Q_QUOTAON	0x0100	/* enable quotas */
 #define	Q_QUOTAOFF	0x0200	/* disable quotas */
-#define	Q_GETQUOTA	0x0300	/* get limits and usage */
-#define	Q_SETQUOTA	0x0400	/* set limits and usage */
-#define	Q_SETUSE	0x0500	/* set usage */
+#define	Q_GETQUOTA32	0x0300	/* get limits and usage (32-bit version) */
+#define	Q_SETQUOTA32	0x0400	/* set limits and usage (32-bit version) */
+#define	Q_SETUSE32	0x0500	/* set usage (32-bit version) */
 #define	Q_SYNC		0x0600	/* sync disk copy of a filesystems quotas */
+#define	Q_GETQUOTA	0x0700	/* get limits and usage (64-bit version) */
+#define	Q_SETQUOTA	0x0800	/* set limits and usage (64-bit version) */
+#define	Q_SETUSE	0x0900	/* set usage (64-bit version) */
 
 /*
  * The following structure defines the format of the disk quota file
@@ -93,7 +96,7 @@
  * the vnode for each quota file (a pointer is retained in the ufsmount
  * structure).
  */
-struct dqblk {
+struct dqblk32 {
 	u_int32_t dqb_bhardlimit;	/* absolute limit on disk blks alloc */
 	u_int32_t dqb_bsoftlimit;	/* preferred limit on disk blks */
 	u_int32_t dqb_curblocks;	/* current block count */
@@ -104,6 +107,30 @@ struct dqblk {
 	int32_t   dqb_itime;		/* time limit for excessive files */
 };
 
+struct dqblk64 {
+	u_int64_t dqb_bhardlimit;	/* absolute limit on disk blks alloc */
+	u_int64_t dqb_bsoftlimit;	/* preferred limit on disk blks */
+	u_int64_t dqb_curblocks;	/* current block count */
+	u_int64_t dqb_ihardlimit;	/* maximum # allocated inodes + 1 */
+	u_int64_t dqb_isoftlimit;	/* preferred inode limit */
+	u_int64_t dqb_curinodes;	/* current # allocated inodes */
+	int64_t   dqb_btime;		/* time limit for excessive disk use */
+	int64_t   dqb_itime;		/* time limit for excessive files */
+};
+
+#define dqblk dqblk64
+
+#define Q_DQHDR64_MAGIC "QUOTA64"
+#define Q_DQHDR64_VERSION 0x20081104
+
+struct dqhdr64 {
+	char	  dqh_magic[8];		/* Q_DQHDR64_MAGIC */
+	uint32_t  dqh_version;		/* Q_DQHDR64_VERSION */
+	uint32_t  dqh_hdrlen;		/* header length */
+	uint32_t  dqh_reclen;		/* record length */
+	char	  dqh_unused[44];	/* reserved for future extension */
+};
+
 #ifdef _KERNEL
 
 #include <sys/queue.h>
@@ -125,7 +152,7 @@ struct dquot {
 	u_int32_t dq_id;		/* identifier this applies to */
 	struct ufsmount *dq_ump;	/* (h) filesystem that this is
 					   taken from */
-	struct dqblk dq_dqb;		/* actual usage & quotas */
+	struct dqblk64 dq_dqb;		/* actual usage & quotas */
 };
 /*
  * Flag values.
@@ -199,10 +226,13 @@ void	dqinit(void);
 void	dqrele(struct vnode *, struct dquot *);
 void	dquninit(void);
 int	getinoquota(struct inode *);
-int	getquota(struct thread *, struct mount *, u_long, int, void *);
 int	qsync(struct mount *mp);
 int	quotaoff(struct thread *td, struct mount *, int);
 int	quotaon(struct thread *td, struct mount *, int, void *);
+int	getquota32(struct thread *, struct mount *, u_long, int, void *);
+int	setquota32(struct thread *, struct mount *, u_long, int, void *);
+int	setuse32(struct thread *, struct mount *, u_long, int, void *);
+int	getquota(struct thread *, struct mount *, u_long, int, void *);
 int	setquota(struct thread *, struct mount *, u_long, int, void *);
 int	setuse(struct thread *, struct mount *, u_long, int, void *);
 vfs_quotactl_t ufs_quotactl;
diff --git a/sys/ufs/ufs/ufs_quota.c b/sys/ufs/ufs/ufs_quota.c
index e5cb9562500..2adf7ef8fd7 100644
--- a/sys/ufs/ufs/ufs_quota.c
+++ b/sys/ufs/ufs/ufs_quota.c
@@ -39,6 +39,7 @@ __FBSDID("$FreeBSD$");
 
 #include <sys/param.h>
 #include <sys/systm.h>
+#include <sys/endian.h>
 #include <sys/fcntl.h>
 #include <sys/kernel.h>
 #include <sys/lock.h>
@@ -73,6 +74,7 @@ static char *quotatypes[] = INITQFNAMES;
 
 static int chkdqchg(struct inode *, ufs2_daddr_t, struct ucred *, int, int *);
 static int chkiqchg(struct inode *, int, struct ucred *, int, int *);
+static int dqopen(struct vnode *, struct ufsmount *, int);
 static int dqget(struct vnode *,
 	u_long, struct ufsmount *, int, struct dquot **);
 static int dqsync(struct vnode *, struct dquot *);
@@ -80,6 +82,14 @@ static void dqflush(struct vnode *);
 static int quotaoff1(struct thread *td, struct mount *mp, int type);
 static int quotaoff_inchange(struct thread *td, struct mount *mp, int type);
 
+/* conversion functions - from_to() */
+static void dqb32_dq(const struct dqblk32 *, struct dquot *);
+static void dqb64_dq(const struct dqblk64 *, struct dquot *);
+static void dq_dqb32(const struct dquot *, struct dqblk32 *);
+static void dq_dqb64(const struct dquot *, struct dqblk64 *);
+static void dqb32_dqb64(const struct dqblk32 *, struct dqblk64 *);
+static void dqb64_dqb32(const struct dqblk64 *, struct dqblk32 *);
+
 #ifdef DIAGNOSTIC
 static void dqref(struct dquot *);
 static void chkdquot(struct inode *);
@@ -90,7 +100,7 @@ static void chkdquot(struct inode *);
  *
  * This routine completely defines the semantics of quotas.
  * If other criterion want to be used to establish quotas, the
- * MAXQUOTAS value in quotas.h should be increased, and the
+ * MAXQUOTAS value in quota.h should be increased, and the
  * additional dquots set up here.
  */
 int
@@ -522,6 +532,13 @@ quotaon(struct thread *td, struct mount *mp, int type, void *fname)
 		return (EALREADY);
 	}
 	ump->um_qflags[type] |= QTF_OPENING|QTF_CLOSING;
+	if ((error = dqopen(vp, ump, type)) != 0) {
+		ump->um_qflags[type] &= ~(QTF_OPENING|QTF_CLOSING);
+		UFS_UNLOCK(ump);
+		(void) vn_close(vp, FREAD|FWRITE, td->td_ucred, td);
+		VFS_UNLOCK_GIANT(vfslocked);
+		return (error);
+	}
 	MNT_ILOCK(mp);
 	mp->mnt_flag |= MNT_QUOTA;
 	MNT_IUNLOCK(mp);
@@ -734,8 +751,9 @@ quotaoff(struct thread *td, struct mount *mp, int type)
 /*
  * Q_GETQUOTA - return current values in a dqblk structure.
  */
-int
-getquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+static int
+_getquota(struct thread *td, struct mount *mp, u_long id, int type,
+    struct dqblk64 *dqb)
 {
 	struct dquot *dq;
 	int error;
@@ -766,7 +784,7 @@ getquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
 	error = dqget(NULLVP, id, VFSTOUFS(mp), type, &dq);
 	if (error)
 		return (error);
-	error = copyout(&dq->dq_dqb, addr, sizeof (struct dqblk));
+	*dqb = dq->dq_dqb;
 	dqrele(NULLVP, dq);
 	return (error);
 }
@@ -774,23 +792,21 @@ getquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
 /*
  * Q_SETQUOTA - assign an entire dqblk structure.
  */
-int
-setquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+static int
+_setquota(struct thread *td, struct mount *mp, u_long id, int type,
+    struct dqblk64 *dqb)
 {
 	struct dquot *dq;
 	struct dquot *ndq;
 	struct ufsmount *ump;
-	struct dqblk newlim;
+	struct dqblk64 newlim;
 	int error;
 
 	error = priv_check(td, PRIV_VFS_SETQUOTA);
 	if (error)
 		return (error);
 
-	ump = VFSTOUFS(mp);
-	error = copyin(addr, &newlim, sizeof (struct dqblk));
-	if (error)
-		return (error);
+	newlim = *dqb;
 
 	ndq = NODQUOT;
 	ump = VFSTOUFS(mp);
@@ -839,23 +855,21 @@ setquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
 /*
  * Q_SETUSE - set current inode and block usage.
  */
-int
-setuse(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+static int
+_setuse(struct thread *td, struct mount *mp, u_long id, int type,
+    struct dqblk64 *dqb)
 {
 	struct dquot *dq;
 	struct ufsmount *ump;
 	struct dquot *ndq;
-	struct dqblk usage;
+	struct dqblk64 usage;
 	int error;
 
 	error = priv_check(td, PRIV_UFS_SETUSE);
 	if (error)
 		return (error);
 
-	ump = VFSTOUFS(mp);
-	error = copyin(addr, &usage, sizeof (struct dqblk));
-	if (error)
-		return (error);
+	usage = *dqb;
 
 	ump = VFSTOUFS(mp);
 	ndq = NODQUOT;
@@ -888,6 +902,90 @@ setuse(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
 	return (0);
 }
 
+int
+getquota32(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk32 dqb32;
+	struct dqblk64 dqb64;
+	int error;
+
+	error = _getquota(td, mp, id, type, &dqb64);
+	if (error)
+		return (error);
+	dqb64_dqb32(&dqb64, &dqb32);
+	error = copyout(&dqb32, addr, sizeof(dqb32));
+	return (error);
+}
+
+int
+setquota32(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk32 dqb32;
+	struct dqblk64 dqb64;
+	int error;
+
+	error = copyin(addr, &dqb32, sizeof(dqb32));
+	if (error)
+		return (error);
+	dqb32_dqb64(&dqb32, &dqb64);
+	error = _setquota(td, mp, id, type, &dqb64);
+	return (error);
+}
+
+int
+setuse32(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk32 dqb32;
+	struct dqblk64 dqb64;
+	int error;
+
+	error = copyin(addr, &dqb32, sizeof(dqb32));
+	if (error)
+		return (error);
+	dqb32_dqb64(&dqb32, &dqb64);
+	error = _setuse(td, mp, id, type, &dqb64);
+	return (error);
+}
+
+int
+getquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk64 dqb64;
+	int error;
+
+	error = _getquota(td, mp, id, type, &dqb64);
+	if (error)
+		return (error);
+	error = copyout(&dqb64, addr, sizeof(dqb64));
+	return (error);
+}
+
+int
+setquota(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk64 dqb64;
+	int error;
+
+	error = copyin(addr, &dqb64, sizeof(dqb64));
+	if (error)
+		return (error);
+	error = _setquota(td, mp, id, type, &dqb64);
+	return (error);
+}
+
+int
+setuse(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
+{
+	struct dqblk64 dqb64;
+	int error;
+
+	error = copyin(addr, &dqb64, sizeof(dqb64));
+	if (error)
+		return (error);
+	error = _setuse(td, mp, id, type, &dqb64);
+	return (error);
+}
+
 /*
  * Q_SYNC - sync quota files to disk.
  */
@@ -1024,6 +1122,47 @@ dqhashfind(struct dqhash *dqh, u_long id, struct vnode *dqvp)
 	return (NODQUOT);
 }
 
+/*
+ * Determine the quota file type.
+ */
+static int
+dqopen(struct vnode *vp, struct ufsmount *ump, int type)
+{
+	struct dqhdr64 dqh;
+	struct iovec aiov;
+	struct uio auio;
+	int error, vfslocked;
+
+	auio.uio_iov = &aiov;
+	auio.uio_iovcnt = 1;
+	aiov.iov_base = &dqh;
+	aiov.iov_len = sizeof(dqh);
+	auio.uio_resid = sizeof(dqh);
+	auio.uio_offset = 0;
+	auio.uio_segflg = UIO_SYSSPACE;
+	auio.uio_rw = UIO_READ;
+	auio.uio_td = (struct thread *)0;
+	vfslocked = VFS_LOCK_GIANT(vp->v_mount);
+	error = VOP_READ(vp, &auio, 0, ump->um_cred[type]);
+	VFS_UNLOCK_GIANT(vfslocked);
+
+	if (error != 0)
+		return (error);
+	if (auio.uio_resid > 0) {
+		/* assume 32 bits */
+		return (0);
+	}
+
+	if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) == 0 &&
+	    be32toh(dqh.dqh_version) == Q_DQHDR64_VERSION &&
+	    be32toh(dqh.dqh_hdrlen) == (uint32_t)sizeof(struct dqhdr64) &&
+	    be32toh(dqh.dqh_reclen) == (uint32_t)sizeof(struct dqblk64))
+		ump->um_qflags[type] |= QTF_64BIT;
+	/* XXX: what if the magic matches, but the sizes are wrong? */
+
+	return (0);
+}
+
 /*
  * Obtain a dquot structure for the specified identifier and quota file
  * reading the information from the file if necessary.
@@ -1032,6 +1171,8 @@ static int
 dqget(struct vnode *vp, u_long id, struct ufsmount *ump, int type,
     struct dquot **dqp)
 {
+	uint8_t buf[sizeof(struct dqblk64)];
+	off_t base, recsize;
 	struct dquot *dq, *dq1;
 	struct dqhash *dqh;
 	struct vnode *dqvp;
@@ -1121,8 +1262,7 @@ hfound:		DQI_LOCK(dq);
 	if (numdquot < desireddquot) {
 		numdquot++;
 		DQH_UNLOCK();
-		dq1 = (struct dquot *)malloc(sizeof *dq, M_DQUOT,
-		    M_WAITOK | M_ZERO);
+		dq1 = malloc(sizeof *dq1, M_DQUOT, M_WAITOK | M_ZERO);
 		mtx_init(&dq1->dq_lock, "dqlock", NULL, MTX_DEF);
 		DQH_LOCK();
 		/*
@@ -1169,20 +1309,52 @@ hfound:		DQI_LOCK(dq);
 	DQREF(dq);
 	DQH_UNLOCK();
 
+	/*
+	 * Read the requested quota record from the quota file, performing
+	 * any necessary conversions.
+	 *
+	 * The record's offset within the file depends on the size of the
+	 * record, which means we need to know whether it's a 32-bit file
+	 * or a 64-bit file.
+	 *
+	 * Luckily, root's record is always at offset 0, and most of it is
+	 * unused, so we can use it to store a magic number indicating the
+	 * file format.  Due to an acute lack of imagination, this magic
+	 * number, stored in the first byte of root's record and hence the
+	 * first byte of the file, is 64.
+	 *
+	 * Another lucky break is that quotaon() always loads root's
+	 * record, to get the default values for dq_btime and dq_itime, so
+	 * we will always have a chance to check the file format before
+	 * being asked for a "real" record.
+	 */
+	if (id == 0 || (ump->um_qflags[type] & QTF_64BIT)) {
+		recsize = sizeof(struct dqblk64);
+		base = sizeof(struct dqhdr64);
+	} else {
+		recsize = sizeof(struct dqblk32);
+		base = 0;
+	}
 	auio.uio_iov = &aiov;
 	auio.uio_iovcnt = 1;
-	aiov.iov_base = &dq->dq_dqb;
-	aiov.iov_len = sizeof (struct dqblk);
-	auio.uio_resid = sizeof (struct dqblk);
-	auio.uio_offset = (off_t)id * sizeof (struct dqblk);
+	aiov.iov_base = buf;
+	aiov.iov_len = recsize;
+	auio.uio_resid = recsize;
+	auio.uio_offset = base + id * recsize;
 	auio.uio_segflg = UIO_SYSSPACE;
 	auio.uio_rw = UIO_READ;
 	auio.uio_td = (struct thread *)0;
 
 	vfslocked = VFS_LOCK_GIANT(dqvp->v_mount);
 	error = VOP_READ(dqvp, &auio, 0, ump->um_cred[type]);
-	if (auio.uio_resid == sizeof(struct dqblk) && error == 0)
-		bzero(&dq->dq_dqb, sizeof(struct dqblk));
+	if (auio.uio_resid == recsize && error == 0) {
+		bzero(&dq->dq_dqb, sizeof(dq->dq_dqb));
+	} else {
+		if (ump->um_qflags[type] & QTF_64BIT)
+			dqb64_dq((struct dqblk64 *)buf, dq);
+		else
+			dqb32_dq((struct dqblk32 *)buf, dq);
+	}
 	if (dqvplocked)
 		vput(dqvp);
 	else
@@ -1281,6 +1453,8 @@ dqrele(struct vnode *vp, struct dquot *dq)
 static int
 dqsync(struct vnode *vp, struct dquot *dq)
 {
+	uint8_t buf[sizeof(struct dqblk64)];
+	off_t base, recsize;
 	struct vnode *dqvp;
 	struct iovec aiov;
 	struct uio auio;
@@ -1327,12 +1501,26 @@ dqsync(struct vnode *vp, struct dquot *dq)
 	dq->dq_flags |= DQ_LOCK;
 	DQI_UNLOCK(dq);
 
+	/*
+	 * Write the quota record to the quota file, performing any
+	 * necessary conversions.  See dqget() for additional details.
+	 */
+	if (ump->um_qflags[dq->dq_type] & QTF_64BIT) {
+		dq_dqb64(dq, (struct dqblk64 *)buf);
+		recsize = sizeof(struct dqblk64);
+		base = sizeof(struct dqhdr64);
+	} else {
+		dq_dqb32(dq, (struct dqblk32 *)buf);
+		recsize = sizeof(struct dqblk32);
+		base = 0;
+	}
+
 	auio.uio_iov = &aiov;
 	auio.uio_iovcnt = 1;
-	aiov.iov_base = &dq->dq_dqb;
-	aiov.iov_len = sizeof (struct dqblk);
-	auio.uio_resid = sizeof (struct dqblk);
-	auio.uio_offset = (off_t)dq->dq_id * sizeof (struct dqblk);
+	aiov.iov_base = buf;
+	aiov.iov_len = recsize;
+	auio.uio_resid = recsize;
+	auio.uio_offset = base + dq->dq_id * recsize;
 	auio.uio_segflg = UIO_SYSSPACE;
 	auio.uio_rw = UIO_WRITE;
 	auio.uio_td = (struct thread *)0;
@@ -1345,7 +1533,8 @@ dqsync(struct vnode *vp, struct dquot *dq)
 	DQI_LOCK(dq);
 	DQI_WAKEUP(dq);
 	dq->dq_flags &= ~DQ_MOD;
-out:	DQI_UNLOCK(dq);
+out:
+	DQI_UNLOCK(dq);
 	vfslocked = VFS_LOCK_GIANT(dqvp->v_mount);
 	if (vp != dqvp)
 		vput(dqvp);
@@ -1384,3 +1573,98 @@ dqflush(struct vnode *vp)
 	}
 	DQH_UNLOCK();
 }
+
+/*
+ * 32-bit / 64-bit conversion functions.
+ *
+ * 32-bit quota records are stored in native byte order.  Attention must
+ * be paid to overflow issues.
+ *
+ * 64-bit quota records are stored in network byte order.
+ */
+
+#define CLIP32(u64) (u64 > UINT32_MAX ? UINT32_MAX : (uint32_t)u64)
+
+static void
+dqb32_dq(const struct dqblk32 *dqb32, struct dquot *dq)
+{
+
+	dq->dq_bhardlimit = dqb32->dqb_bhardlimit;
+	dq->dq_bsoftlimit = dqb32->dqb_bsoftlimit;
+	dq->dq_curblocks = dqb32->dqb_curblocks;
+	dq->dq_ihardlimit = dqb32->dqb_ihardlimit;
+	dq->dq_isoftlimit = dqb32->dqb_isoftlimit;
+	dq->dq_curinodes = dqb32->dqb_curinodes;
+	dq->dq_btime = dqb32->dqb_btime;
+	dq->dq_itime = dqb32->dqb_itime;
+}
+
+static void
+dqb64_dq(const struct dqblk64 *dqb64, struct dquot *dq)
+{
+
+	dq->dq_bhardlimit = be64toh(dqb64->dqb_bhardlimit);
+	dq->dq_bsoftlimit = be64toh(dqb64->dqb_bsoftlimit);
+	dq->dq_curblocks = be64toh(dqb64->dqb_curblocks);
+	dq->dq_ihardlimit = be64toh(dqb64->dqb_ihardlimit);
+	dq->dq_isoftlimit = be64toh(dqb64->dqb_isoftlimit);
+	dq->dq_curinodes = be64toh(dqb64->dqb_curinodes);
+	dq->dq_btime = be64toh(dqb64->dqb_btime);
+	dq->dq_itime = be64toh(dqb64->dqb_itime);
+}
+
+static void
+dq_dqb32(const struct dquot *dq, struct dqblk32 *dqb32)
+{
+
+	dqb32->dqb_bhardlimit = CLIP32(dq->dq_bhardlimit);
+	dqb32->dqb_bsoftlimit = CLIP32(dq->dq_bsoftlimit);
+	dqb32->dqb_curblocks = CLIP32(dq->dq_curblocks);
+	dqb32->dqb_ihardlimit = CLIP32(dq->dq_ihardlimit);
+	dqb32->dqb_isoftlimit = CLIP32(dq->dq_isoftlimit);
+	dqb32->dqb_curinodes = CLIP32(dq->dq_curinodes);
+	dqb32->dqb_btime = CLIP32(dq->dq_btime);
+	dqb32->dqb_itime = CLIP32(dq->dq_itime);
+}
+
+static void
+dq_dqb64(const struct dquot *dq, struct dqblk64 *dqb64)
+{
+
+	dqb64->dqb_bhardlimit = htobe64(dq->dq_bhardlimit);
+	dqb64->dqb_bsoftlimit = htobe64(dq->dq_bsoftlimit);
+	dqb64->dqb_curblocks = htobe64(dq->dq_curblocks);
+	dqb64->dqb_ihardlimit = htobe64(dq->dq_ihardlimit);
+	dqb64->dqb_isoftlimit = htobe64(dq->dq_isoftlimit);
+	dqb64->dqb_curinodes = htobe64(dq->dq_curinodes);
+	dqb64->dqb_btime = htobe64(dq->dq_btime);
+	dqb64->dqb_itime = htobe64(dq->dq_itime);
+}
+
+static void
+dqb64_dqb32(const struct dqblk64 *dqb64, struct dqblk32 *dqb32)
+{
+
+	dqb32->dqb_bhardlimit = CLIP32(dqb64->dqb_bhardlimit);
+	dqb32->dqb_bsoftlimit = CLIP32(dqb64->dqb_bsoftlimit);
+	dqb32->dqb_curblocks = CLIP32(dqb64->dqb_curblocks);
+	dqb32->dqb_ihardlimit = CLIP32(dqb64->dqb_ihardlimit);
+	dqb32->dqb_isoftlimit = CLIP32(dqb64->dqb_isoftlimit);
+	dqb32->dqb_curinodes = CLIP32(dqb64->dqb_curinodes);
+	dqb32->dqb_btime = CLIP32(dqb64->dqb_btime);
+	dqb32->dqb_itime = CLIP32(dqb64->dqb_itime);
+}
+
+static void
+dqb32_dqb64(const struct dqblk32 *dqb32, struct dqblk64 *dqb64)
+{
+
+	dqb64->dqb_bhardlimit = dqb32->dqb_bhardlimit;
+	dqb64->dqb_bsoftlimit = dqb32->dqb_bsoftlimit;
+	dqb64->dqb_curblocks = dqb32->dqb_curblocks;
+	dqb64->dqb_ihardlimit = dqb32->dqb_ihardlimit;
+	dqb64->dqb_isoftlimit = dqb32->dqb_isoftlimit;
+	dqb64->dqb_curinodes = dqb32->dqb_curinodes;
+	dqb64->dqb_btime = dqb32->dqb_btime;
+	dqb64->dqb_itime = dqb32->dqb_itime;
+}
diff --git a/sys/ufs/ufs/ufs_vfsops.c b/sys/ufs/ufs/ufs_vfsops.c
index 14ce4c6b033..2ddde018f1a 100644
--- a/sys/ufs/ufs/ufs_vfsops.c
+++ b/sys/ufs/ufs/ufs_vfsops.c
@@ -130,6 +130,18 @@ ufs_quotactl(mp, cmds, id, arg, td)
 		error = quotaoff(td, mp, type);
 		break;
 
+	case Q_SETQUOTA32:
+		error = setquota32(td, mp, id, type, arg);
+		break;
+
+	case Q_SETUSE32:
+		error = setuse32(td, mp, id, type, arg);
+		break;
+
+	case Q_GETQUOTA32:
+		error = getquota32(td, mp, id, type, arg);
+		break;
+
 	case Q_SETQUOTA:
 		error = setquota(td, mp, id, type, arg);
 		break;
diff --git a/sys/ufs/ufs/ufsmount.h b/sys/ufs/ufs/ufsmount.h
index 126867baa2e..83f9af06b59 100644
--- a/sys/ufs/ufs/ufsmount.h
+++ b/sys/ufs/ufs/ufsmount.h
@@ -120,6 +120,7 @@ struct ufsmount {
  */
 #define	QTF_OPENING	0x01			/* Q_QUOTAON in progress */
 #define	QTF_CLOSING	0x02			/* Q_QUOTAOFF in progress */
+#define QTF_64BIT	0x04			/* 64-bit quota file */
 
 /* Convert mount ptr to ufsmount ptr. */
 #define VFSTOUFS(mp)	((struct ufsmount *)((mp)->mnt_data))
diff --git a/usr.sbin/edquota/Makefile b/usr.sbin/edquota/Makefile
index e5a3b12559a..82794925397 100644
--- a/usr.sbin/edquota/Makefile
+++ b/usr.sbin/edquota/Makefile
@@ -6,4 +6,7 @@ MAN=	edquota.8
 
 WARNS?=	4
 
+DPADD=	${LIBUTIL}
+LDADD=	-lutil
+
 .include <bsd.prog.mk>
diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index f38fa77b39e..76baa0417fa 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -55,17 +55,21 @@ __FBSDID("$FreeBSD$");
 #include <sys/mount.h>
 #include <sys/wait.h>
 #include <ufs/ufs/quota.h>
+
 #include <ctype.h>
 #include <err.h>
 #include <errno.h>
 #include <fstab.h>
 #include <grp.h>
+#include <inttypes.h>
+#include <libutil.h>
 #include <pwd.h>
 #include <signal.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
+
 #include "pathnames.h"
 
 const char *qfname = QUOTAFILENAME;
@@ -102,11 +106,11 @@ main(int argc, char *argv[])
 {
 	struct quotause *qup, *protoprivs, *curprivs;
 	long id, protoid;
-	long long lim;
+	uintmax_t lim;
 	int i, quotatype, range, tmpfd;
 	uid_t startuid, enduid;
-	u_int32_t *limp;
-	char *protoname, *cp, *oldoptarg;
+	u_int64_t *limp;
+	char *protoname, *cp, *oldoptarg, *end;
 	int eflag = 0, tflag = 0, pflag = 0, ch;
 	char *fspath = NULL;
 	char buf[MAXLOGNAME];
@@ -172,11 +176,10 @@ main(int argc, char *argv[])
 					break; /* XXX: report an error */
 				}
 				if (limp != NULL) {
-					lim = strtoll(cp, NULL, 10);
-					if (lim < 0 || lim > UINT_MAX)
-						errx(1, "invalid limit value: "
-						    "%lld", lim);
-					*limp = (u_int32_t)lim;
+					lim = strtoumax(cp, &end, 10);
+					if (end == cp || *end != '\0' || lim > UINT64_MAX)
+						errx(1, "invalid limit: %s", cp);
+					*limp = (u_int64_t)lim;
 				}
 			}
 			qup->dqblk.dqb_bsoftlimit =
@@ -331,10 +334,11 @@ getentry(const char *name, int quotatype)
 struct quotause *
 getprivs(long id, int quotatype, char *fspath)
 {
+	struct quotafile *qf;
 	struct fstab *fs;
 	struct quotause *qup, *quptail;
 	struct quotause *quphead;
-	int qcmd, qupsize, fd;
+	int qcmd, qupsize;
 	char *qfpathname;
 	static int warned = 0;
 
@@ -358,46 +362,19 @@ getprivs(long id, int quotatype, char *fspath)
 		warnx("warning: quotas are not compiled into this kernel");
 				sleep(3);
 			}
-			if ((fd = open(qfpathname, O_RDONLY)) < 0) {
-				fd = open(qfpathname, O_RDWR|O_CREAT, 0640);
-				if (fd < 0 && errno != ENOENT) {
-					warn("%s", qfpathname);
-					free(qup);
-					continue;
-				}
-				warnx("creating quota file %s", qfpathname);
-				sleep(3);
-				(void) fchown(fd, getuid(),
-				    getentry(quotagroup, GRPQUOTA));
-				(void) fchmod(fd, 0640);
-			}
-			if (lseek(fd, (off_t)id * sizeof(struct dqblk),
-			    L_SET) < 0) {
-				warn("seek error on %s", qfpathname);
-				close(fd);
+			if ((qf = quota_open(qfpathname)) == NULL &&
+			    (qf = quota_create(qfpathname)) == NULL) {
+				warn("%s", qfpathname);
 				free(qup);
 				continue;
 			}
-			switch (read(fd, &qup->dqblk, sizeof(struct dqblk))) {
-			case 0:			/* EOF */
-				/*
-				 * Convert implicit 0 quota (EOF)
-				 * into an explicit one (zero'ed dqblk)
-				 */
-				bzero((caddr_t)&qup->dqblk,
-				    sizeof(struct dqblk));
-				break;
-
-			case sizeof(struct dqblk):	/* OK */
-				break;
-
-			default:		/* ERROR */
+			if (quota_read(qf, &qup->dqblk, id) != 0) {
 				warn("read error in %s", qfpathname);
-				close(fd);
+				quota_close(qf);
 				free(qup);
 				continue;
 			}
-			close(fd);
+			quota_close(qf);
 		}
 		strcpy(qup->qfname, qfpathname);
 		strcpy(qup->fsname, fs->fs_file);
@@ -418,38 +395,22 @@ getprivs(long id, int quotatype, char *fspath)
 void
 putprivs(long id, int quotatype, struct quotause *quplist)
 {
+	struct quotafile *qf;
 	struct quotause *qup;
-	int qcmd, fd;
+	int qcmd;
 	struct dqblk dqbuf;
 
 	qcmd = QCMD(Q_SETQUOTA, quotatype);
 	for (qup = quplist; qup; qup = qup->next) {
 		if (quotactl(qup->fsname, qcmd, id, &qup->dqblk) == 0)
 			continue;
-		if ((fd = open(qup->qfname, O_RDWR)) < 0) {
+		if ((qf = quota_open(qup->qfname)) == NULL) {
 			warn("%s", qup->qfname);
 			continue;
 		}
-		if (lseek(fd, (off_t)id * sizeof(struct dqblk), L_SET) < 0) {
-			warn("seek error on %s", qup->qfname);
-			close(fd);
-			continue;
-		}
-		switch (read(fd, &dqbuf, sizeof(struct dqblk))) {
-		case 0:			/* EOF */
-			/*
-			 * Convert implicit 0 quota (EOF)
-			 * into an explicit one (zero'ed dqblk)
-			 */
-			bzero(&dqbuf, sizeof(struct dqblk));
-			break;
-
-		case sizeof(struct dqblk):	/* OK */
-			break;
-
-		default:		/* ERROR */
+		if (quota_read(qf, &dqbuf, id) != 0) {
 			warn("read error in %s", qup->qfname);
-			close(fd);
+			quota_close(qf);
 			continue;
 		}
 		/*
@@ -474,16 +435,10 @@ putprivs(long id, int quotatype, struct quotause *quplist)
 			qup->dqblk.dqb_itime = 0;
 		qup->dqblk.dqb_curinodes = dqbuf.dqb_curinodes;
 		qup->dqblk.dqb_curblocks = dqbuf.dqb_curblocks;
-		if (lseek(fd, (off_t)id * sizeof(struct dqblk), L_SET) < 0) {
-			warn("seek error on %s", qup->qfname);
-			close(fd);
-			continue;
-		}
-		if (write(fd, &qup->dqblk, sizeof (struct dqblk)) !=
-		    sizeof (struct dqblk)) {
+		if (quota_write(qf, &qup->dqblk, id) != 0) {
 			warn("%s", qup->qfname);
-			}
-		close(fd);
+		}
+		quota_close(qf);
 	}
 }
 

From 08dc3a5e5c3a8be3076365832986e7a69d0ae92d Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Wed, 4 Feb 2009 00:21:36 +0000
Subject: [PATCH 006/532] Finish conversion of edquota to work with 64-bit
 quotas. Add -h option to request "humanized" values such as 1M, 1G, 1T, etc.

Discussed with: Dag-Erling Smorgrav
Sponsored by: Rsync.net
---
 usr.sbin/edquota/Makefile  |   1 +
 usr.sbin/edquota/edquota.8 |  13 +++
 usr.sbin/edquota/edquota.c | 214 +++++++++++++++++++++++++++----------
 3 files changed, 170 insertions(+), 58 deletions(-)

diff --git a/usr.sbin/edquota/Makefile b/usr.sbin/edquota/Makefile
index 82794925397..b62221257f5 100644
--- a/usr.sbin/edquota/Makefile
+++ b/usr.sbin/edquota/Makefile
@@ -4,6 +4,7 @@
 PROG=	edquota
 MAN=	edquota.8
 
+CSTD=	c99
 WARNS?=	4
 
 DPADD=	${LIBUTIL}
diff --git a/usr.sbin/edquota/edquota.8 b/usr.sbin/edquota/edquota.8
index 7cf72dd1118..9c90de468b6 100644
--- a/usr.sbin/edquota/edquota.8
+++ b/usr.sbin/edquota/edquota.8
@@ -42,6 +42,7 @@
 .Op Fl u
 .Op Fl f Ar fspath
 .Op Fl p Ar proto-username
+.Op Fl h
 .Ar username ...
 .Nm
 .Op Fl u
@@ -55,6 +56,7 @@
 .Fl g
 .Op Fl f Ar fspath
 .Op Fl p Ar proto-groupname
+.Op Fl h
 .Ar groupname ...
 .Nm
 .Fl g
@@ -97,6 +99,14 @@ unless the environment variable
 specifies otherwise.
 .Pp
 The quotas may then be modified, new quotas added, etc.
+Block quotas can be specified in bytes (B), kilobytes (K),
+megabytes (M), terabytes (T), pedabytes (P), or exabytes (E).
+If no units are specified, kilobytes are assumed.
+If the
+.Fl h
+flag is specified, the editor will always display the 
+block usage and limits in a more human readable format
+rather than displaying them in the historic kilobyte format.
 Setting a quota to zero indicates that no quota should be imposed.
 Setting a hard limit to one indicates that no allocations should
 be permitted.
@@ -159,6 +169,9 @@ and
 .Ar ihlim
 values is omitted, it is assumed to be zero, therefore
 indicating that no particular quota should be imposed.
+Block quotas can be specified in bytes (B), kilobytes (K),
+megabytes (M), terabytes (T), pedabytes (P), or exabytes (E).
+If no units are specified, kilobytes are assumed.
 .Pp
 If invoked with the
 .Fl f
diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index 76baa0417fa..2798d8a7742 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -72,10 +72,22 @@ __FBSDID("$FreeBSD$");
 
 #include "pathnames.h"
 
+/* Let's be paranoid about block size */
+#if 10 > DEV_BSHIFT
+#define dbtokb(db) \
+	((off_t)(db) >> (10-DEV_BSHIFT))
+#elif 10 < DEV_BSHIFT
+#define dbtokb(db) \
+	((off_t)(db) << (DEV_BSHIFT-10))
+#else
+#define dbtokb(db)	(db)
+#endif
+
 const char *qfname = QUOTAFILENAME;
 const char *qfextension[] = INITQFNAMES;
 const char *quotagroup = QUOTAGROUP;
 char tmpfil[] = _PATH_TMP;
+int hflag;
 
 struct quotause {
 	struct	quotause *next;
@@ -87,9 +99,11 @@ struct quotause {
 #define	FOUND	0x01
 
 int alldigits(const char *s);
-int cvtatos(time_t, char *, time_t *);
-char *cvtstoa(time_t);
+int cvtatos(u_int64_t, char *, u_int64_t *);
+char *cvtstoa(u_int64_t);
+u_int64_t cvtval(u_int64_t, char);
 int editit(char *);
+char *fmthumanval(int64_t);
 void freeprivs(struct quotause *);
 int getentry(const char *, int);
 struct quotause *getprivs(long, int, char *);
@@ -106,11 +120,10 @@ main(int argc, char *argv[])
 {
 	struct quotause *qup, *protoprivs, *curprivs;
 	long id, protoid;
-	uintmax_t lim;
 	int i, quotatype, range, tmpfd;
 	uid_t startuid, enduid;
-	u_int64_t *limp;
-	char *protoname, *cp, *oldoptarg, *end;
+	u_int64_t lim;
+	char *protoname, *cp, *endpt, *oldoptarg;
 	int eflag = 0, tflag = 0, pflag = 0, ch;
 	char *fspath = NULL;
 	char buf[MAXLOGNAME];
@@ -123,7 +136,7 @@ main(int argc, char *argv[])
 	protoprivs = NULL;
 	curprivs = NULL;
 	protoname = NULL;
-	while ((ch = getopt(argc, argv, "ugtf:p:e:")) != -1) {
+	while ((ch = getopt(argc, argv, "ughtf:p:e:")) != -1) {
 		switch(ch) {
 		case 'f':
 			fspath = optarg;
@@ -135,6 +148,9 @@ main(int argc, char *argv[])
 		case 'g':
 			quotatype = GRPQUOTA;
 			break;
+		case 'h':
+			hflag++;
+			break;
 		case 'u':
 			quotatype = USRQUOTA;
 			break;
@@ -142,45 +158,44 @@ main(int argc, char *argv[])
 			tflag++;
 			break;
 		case 'e':
-			if ((qup = malloc(sizeof(*qup))) == NULL)
+			if ((qup = malloc(sizeof(*qup) + BUFSIZ)) == NULL)
 				errx(2, "out of memory");
-			bzero(qup, sizeof(*qup));
+			bzero(qup, sizeof(*qup) + BUFSIZ);
 			i = 0;
 			oldoptarg = optarg;
 			for (cp = optarg; (cp = strsep(&optarg, ":")) != NULL;
 			    i++) {
 				if (cp != oldoptarg)
 					*(cp - 1) = ':';
-				limp = NULL;
 				switch (i) {
 				case 0:
 					strlcpy(qup->fsname, cp,
 					    sizeof(qup->fsname));
 					break;
 				case 1:
-					limp = &qup->dqblk.dqb_bsoftlimit;
-					break;
+					lim = strtoll(cp, &endpt, 10);
+					qup->dqblk.dqb_bsoftlimit =
+						cvtval(lim, *endpt);
+					continue;
 				case 2:
-					limp = &qup->dqblk.dqb_bhardlimit;
-					break;
+					lim = strtoll(cp, &endpt, 10);
+					qup->dqblk.dqb_bhardlimit =
+						cvtval(lim, *endpt);
+					continue;
 				case 3:
-					limp = &qup->dqblk.dqb_isoftlimit;
-					break;
+					qup->dqblk.dqb_isoftlimit =
+						strtoll(cp, NULL, 10);
+					continue;
 				case 4:
-					limp = &qup->dqblk.dqb_ihardlimit;
-					break;
+					qup->dqblk.dqb_ihardlimit =
+						strtoll(cp, NULL, 10);
+					continue;
 				default:
 					warnx("incorrect quota specification: "
 					    "%s", oldoptarg);
 					usage();
 					break; /* XXX: report an error */
 				}
-				if (limp != NULL) {
-					lim = strtoumax(cp, &end, 10);
-					if (end == cp || *end != '\0' || lim > UINT64_MAX)
-						errx(1, "invalid limit: %s", cp);
-					*limp = (u_int64_t)lim;
-				}
 			}
 			qup->dqblk.dqb_bsoftlimit =
 			    btodb((off_t)qup->dqblk.dqb_bsoftlimit * 1024);
@@ -285,10 +300,10 @@ static void
 usage(void)
 {
 	fprintf(stderr, "%s\n%s\n%s\n%s\n%s\n%s\n%s\n%s\n",
-		"usage: edquota [-u] [-f fspath] [-p username] username ...",
+		"usage: edquota [-u] [-f fspath] [-p username] [-h] username ...",
 		"       edquota [-u] -e fspath[:bslim[:bhlim[:islim[:ihlim]]]] [-e ...]",
 		"               username ...",
-		"       edquota -g [-f fspath] [-p groupname] groupname ...",
+		"       edquota -g [-f fspath] [-p groupname] [-h] groupname ...",
 		"       edquota -g -e fspath[:bslim[:bhlim[:islim[:ihlim]]]] [-e ...]",
 		"               groupname ...",
 		"       edquota [-u] -t [-f fspath]",
@@ -314,14 +329,17 @@ getentry(const char *name, int quotatype)
 		if ((pw = getpwnam(name)))
 			return (pw->pw_uid);
 		warnx("%s: no such user", name);
+		sleep(3);
 		break;
 	case GRPQUOTA:
 		if ((gr = getgrnam(name)))
 			return (gr->gr_gid);
 		warnx("%s: no such group", name);
+		sleep(3);
 		break;
 	default:
 		warnx("%d: unknown quota type", quotatype);
+		sleep(3);
 		break;
 	}
 	sleep(1);
@@ -499,21 +517,36 @@ writeprivs(struct quotause *quplist, int outfd, char *name, int quotatype)
 		err(1, "%s", tmpfil);
 	fprintf(fd, "Quotas for %s %s:\n", qfextension[quotatype], name);
 	for (qup = quplist; qup; qup = qup->next) {
-		fprintf(fd, "%s: %s %lu, limits (soft = %lu, hard = %lu)\n",
-		    qup->fsname, "kbytes in use:",
-		    (unsigned long)(dbtob(qup->dqblk.dqb_curblocks) / 1024),
-		    (unsigned long)(dbtob(qup->dqblk.dqb_bsoftlimit) / 1024),
-		    (unsigned long)(dbtob(qup->dqblk.dqb_bhardlimit) / 1024));
-		fprintf(fd, "%s %lu, limits (soft = %lu, hard = %lu)\n",
+		fprintf(fd, "%s: in use: %s, ", qup->fsname,
+		    fmthumanval(qup->dqblk.dqb_curblocks));
+		fprintf(fd, "limits (soft = %s, ",
+		    fmthumanval(qup->dqblk.dqb_bsoftlimit));
+		fprintf(fd, "hard = %s)\n",
+		    fmthumanval(qup->dqblk.dqb_bhardlimit));
+		fprintf(fd, "%s %llu, limits (soft = %llu, hard = %llu)\n",
 		    "\tinodes in use:",
-		    (unsigned long)qup->dqblk.dqb_curinodes,
-		    (unsigned long)qup->dqblk.dqb_isoftlimit,
-		    (unsigned long)qup->dqblk.dqb_ihardlimit);
+		    qup->dqblk.dqb_curinodes,
+		    qup->dqblk.dqb_isoftlimit,
+		    qup->dqblk.dqb_ihardlimit);
 	}
 	fclose(fd);
 	return (1);
 }
 
+char *
+fmthumanval(int64_t blocks)
+{
+	static char buf[7], numbuf[20];
+
+	if (hflag) {
+		humanize_number(buf, sizeof(buf) - (blocks < 0 ? 0 : 1),
+		    dbtob(blocks), "", HN_AUTOSCALE, HN_NOSPACE);
+		return (buf);
+	}
+	snprintf(numbuf, 20, "%lluK", dbtokb(blocks));
+	return(numbuf);
+}
+
 /*
  * Merge changes to an ASCII file into a quotause list.
  */
@@ -522,8 +555,9 @@ readprivs(struct quotause *quplist, char *inname)
 {
 	struct quotause *qup;
 	FILE *fd;
-	unsigned long bhardlimit, bsoftlimit, curblocks;
-	unsigned long ihardlimit, isoftlimit, curinodes;
+	u_int64_t ihardlimit, isoftlimit, curinodes;
+	u_int64_t bhardlimit, bsoftlimit, curblocks;
+	char bhardunits, bsoftunits, curblockunits;
 	int cnt;
 	char *cp;
 	struct dqblk dqblk;
@@ -549,21 +583,40 @@ readprivs(struct quotause *quplist, char *inname)
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    " kbytes in use: %lu, limits (soft = %lu, hard = %lu)",
-		    &curblocks, &bsoftlimit, &bhardlimit);
-		if (cnt != 3) {
+		    " in use: %llu%c, limits (soft = %llu%c, hard = %llu%c)",
+		    &curblocks, &curblockunits, &bsoftlimit, &bsoftunits,
+		    &bhardlimit, &bhardunits);
+		/*
+		 * The next three check for old-style input formats.
+		 */
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			 " in use: %llu%c, limits (soft = %llu%c hard = %llu%c",
+			    &curblocks, &curblockunits, &bsoftlimit,
+			    &bsoftunits, &bhardlimit, &bhardunits);
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			" in use: %llu%c, limits (soft = %llu%c hard = %llu%c)",
+			    &curblocks, &curblockunits, &bsoftlimit,
+			    &bsoftunits, &bhardlimit, &bhardunits);
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			" in use: %llu%c, limits (soft = %llu%c, hard = %llu%c",
+			    &curblocks, &curblockunits, &bsoftlimit,
+			    &bsoftunits, &bhardlimit, &bhardunits);
+		if (cnt != 6) {
 			warnx("%s:%s: bad format", fsp, cp);
 			return (0);
 		}
-		dqblk.dqb_curblocks = btodb((off_t)curblocks * 1024);
-		dqblk.dqb_bsoftlimit = btodb((off_t)bsoftlimit * 1024);
-		dqblk.dqb_bhardlimit = btodb((off_t)bhardlimit * 1024);
+		dqblk.dqb_curblocks = cvtval(curblocks, curblockunits);
+		dqblk.dqb_bsoftlimit = cvtval(bsoftlimit, bsoftunits);
+		dqblk.dqb_bhardlimit = cvtval(bhardlimit, bhardunits);
 		if ((cp = strtok(line2, "\n")) == NULL) {
 			warnx("%s: %s: bad format", fsp, line2);
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    "\tinodes in use: %lu, limits (soft = %lu, hard = %lu)",
+		    "\tinodes in use: %llu, limits (soft = %llu, hard = %llu)",
 		    &curinodes, &isoftlimit, &ihardlimit);
 		if (cnt != 3) {
 			warnx("%s: %s: bad format", fsp, line2);
@@ -598,8 +651,8 @@ readprivs(struct quotause *quplist, char *inname)
 			qup->dqblk.dqb_isoftlimit = dqblk.dqb_isoftlimit;
 			qup->dqblk.dqb_ihardlimit = dqblk.dqb_ihardlimit;
 			qup->flags |= FOUND;
-			if (dqblk.dqb_curblocks == qup->dqblk.dqb_curblocks &&
-			    dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes)
+			/* No easy way to check change in curblocks */
+			if (dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes)
 				break;
 			warnx("%s: cannot change current allocation", fsp);
 			break;
@@ -658,8 +711,7 @@ readtimes(struct quotause *quplist, char *inname)
 	FILE *fd;
 	int cnt;
 	char *cp;
-	time_t itime, btime, iseconds, bseconds;
-	long l_itime, l_btime;
+	u_int64_t itime, btime, iseconds, bseconds;
 	char *fsp, bunits[10], iunits[10], line1[BUFSIZ];
 
 	fd = fopen(inname, "r");
@@ -682,14 +734,12 @@ readtimes(struct quotause *quplist, char *inname)
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    " block grace period: %ld %s file grace period: %ld %s",
-		    &l_btime, bunits, &l_itime, iunits);
+		    " block grace period: %llu %s file grace period: %llu %s",
+		    &btime, bunits, &itime, iunits);
 		if (cnt != 4) {
 			warnx("%s:%s: bad format", fsp, cp);
 			return (0);
 		}
-		btime = l_btime;
-		itime = l_itime;
 		if (cvtatos(btime, bunits, &bseconds) == 0)
 			return (0);
 		if (cvtatos(itime, iunits, &iseconds) == 0)
@@ -723,21 +773,21 @@ readtimes(struct quotause *quplist, char *inname)
  * Convert seconds to ASCII times.
  */
 char *
-cvtstoa(time_t secs)
+cvtstoa(u_int64_t secs)
 {
 	static char buf[20];
 
 	if (secs % (24 * 60 * 60) == 0) {
 		secs /= 24 * 60 * 60;
-		sprintf(buf, "%ld day%s", (long)secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%llu day%s", secs, secs == 1 ? "" : "s");
 	} else if (secs % (60 * 60) == 0) {
 		secs /= 60 * 60;
-		sprintf(buf, "%ld hour%s", (long)secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%llu hour%s", secs, secs == 1 ? "" : "s");
 	} else if (secs % 60 == 0) {
 		secs /= 60;
-		sprintf(buf, "%ld minute%s", (long)secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%llu minute%s", secs, secs == 1 ? "" : "s");
 	} else
-		sprintf(buf, "%ld second%s", (long)secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%llu second%s", secs, secs == 1 ? "" : "s");
 	return (buf);
 }
 
@@ -745,7 +795,7 @@ cvtstoa(time_t secs)
  * Convert ASCII input times to seconds.
  */
 int
-cvtatos(time_t period, char *units, time_t *seconds)
+cvtatos(u_int64_t period, char *units, u_int64_t *seconds)
 {
 
 	if (bcmp(units, "second", 6) == 0)
@@ -757,13 +807,60 @@ cvtatos(time_t period, char *units, time_t *seconds)
 	else if (bcmp(units, "day", 3) == 0)
 		*seconds = period * 24 * 60 * 60;
 	else {
-		printf("%s: bad units, specify %s\n", units,
+		warnx("%s: bad units, specify %s\n", units,
 		    "days, hours, minutes, or seconds");
 		return (0);
 	}
 	return (1);
 }
 
+/*
+ * Convert a limit to number of disk blocks.
+ */
+u_int64_t
+cvtval(u_int64_t limit, char units)
+{
+
+	switch(units) {
+	case 'B':
+	case 'b':
+		limit = btodb(limit);
+		break;
+	case '\0':	/* historic behavior */
+	case ',':	/* historic behavior */
+	case ')':	/* historic behavior */
+	case 'K':
+	case 'k':
+		limit *= btodb(1024);
+		break;
+	case 'M':
+	case 'm':
+		limit *= btodb(1048576);
+		break;
+	case 'G':
+	case 'g':
+		limit *= btodb(1073741824);
+		break;
+	case 'T':
+	case 't':
+		limit *= btodb(1099511627776);
+		break;
+	case 'P':
+	case 'p':
+		limit *= btodb(1125899906842624);
+		break;
+	case 'E':
+	case 'e':
+		limit *= btodb(1152921504606846976);
+		break;
+	default:
+		warnx("%llu%c: unknown units, specify K, M, G, T, P, or E\n",
+		    limit, units);
+		break;
+	}
+	return (limit);
+}
+
 /*
  * Free a list of quotause structures.
  */
@@ -833,6 +930,7 @@ hasquota(struct fstab *fs, int type, char **qfnamep)
 	}
 	if (statfs(fs->fs_file, &sfb) != 0) {
 		warn("cannot statfs mount point %s", fs->fs_file);
+		sleep(3);
 		return (0);
 	}
 	if (strcmp(fs->fs_file, sfb.f_mntonname)) {

From b83b5d3b796b885ad19cce99cd7a3ab81e41a2e1 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Thu, 5 Feb 2009 14:06:43 +0000
Subject: [PATCH 007/532] Updates to edquota based on feedback from Dag-Erling
 Smorgrav.

---
 usr.sbin/edquota/Makefile  |  2 +-
 usr.sbin/edquota/edquota.8 | 15 ++++---
 usr.sbin/edquota/edquota.c | 91 +++++++++++++++++++++++---------------
 3 files changed, 67 insertions(+), 41 deletions(-)

diff --git a/usr.sbin/edquota/Makefile b/usr.sbin/edquota/Makefile
index b62221257f5..1196e472141 100644
--- a/usr.sbin/edquota/Makefile
+++ b/usr.sbin/edquota/Makefile
@@ -4,7 +4,7 @@
 PROG=	edquota
 MAN=	edquota.8
 
-CSTD=	c99
+CSTD=	gnu99
 WARNS?=	4
 
 DPADD=	${LIBUTIL}
diff --git a/usr.sbin/edquota/edquota.8 b/usr.sbin/edquota/edquota.8
index 9c90de468b6..326c8373c25 100644
--- a/usr.sbin/edquota/edquota.8
+++ b/usr.sbin/edquota/edquota.8
@@ -39,10 +39,9 @@
 .Nd edit user quotas
 .Sh SYNOPSIS
 .Nm
-.Op Fl u
+.Op Fl uh
 .Op Fl f Ar fspath
 .Op Fl p Ar proto-username
-.Op Fl h
 .Ar username ...
 .Nm
 .Op Fl u
@@ -54,9 +53,9 @@
 .Ar username ...
 .Nm
 .Fl g
+.Op Fl h
 .Op Fl f Ar fspath
 .Op Fl p Ar proto-groupname
-.Op Fl h
 .Ar groupname ...
 .Nm
 .Fl g
@@ -100,8 +99,11 @@ specifies otherwise.
 .Pp
 The quotas may then be modified, new quotas added, etc.
 Block quotas can be specified in bytes (B), kilobytes (K),
-megabytes (M), terabytes (T), pedabytes (P), or exabytes (E).
+megabytes (M), terabytes (T), petabytes (P), or exabytes (E).
 If no units are specified, kilobytes are assumed.
+Inode quotas can be specified in kiloinodes (K),
+megainodes (M), terainodes (T), petainodes (P), or exainodes (E).
+If no units are specified, the number of inodes specified are used.
 If the
 .Fl h
 flag is specified, the editor will always display the 
@@ -170,8 +172,11 @@ and
 values is omitted, it is assumed to be zero, therefore
 indicating that no particular quota should be imposed.
 Block quotas can be specified in bytes (B), kilobytes (K),
-megabytes (M), terabytes (T), pedabytes (P), or exabytes (E).
+megabytes (M), terabytes (T), petabytes (P), or exabytes (E).
 If no units are specified, kilobytes are assumed.
+Inode quotas can be specified in kiloinodes (K),
+megainodes (M), terainodes (T), petainodes (P), or exainodes (E).
+If no units are specified, the number of inodes specified are used.
 .Pp
 If invoked with the
 .Fl f
diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index 2798d8a7742..3331548c86c 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -158,13 +158,11 @@ main(int argc, char *argv[])
 			tflag++;
 			break;
 		case 'e':
-			if ((qup = malloc(sizeof(*qup) + BUFSIZ)) == NULL)
+			if ((qup = calloc(1, sizeof(*qup) + BUFSIZ)) == NULL)
 				errx(2, "out of memory");
-			bzero(qup, sizeof(*qup) + BUFSIZ);
-			i = 0;
 			oldoptarg = optarg;
-			for (cp = optarg; (cp = strsep(&optarg, ":")) != NULL;
-			    i++) {
+			for (i = 0, cp = optarg;
+			     (cp = strsep(&optarg, ":")) != NULL; i++) {
 				if (cp != oldoptarg)
 					*(cp - 1) = ':';
 				switch (i) {
@@ -183,12 +181,14 @@ main(int argc, char *argv[])
 						cvtval(lim, *endpt);
 					continue;
 				case 3:
+					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_isoftlimit =
-						strtoll(cp, NULL, 10);
+						cvtval(lim, *endpt);
 					continue;
 				case 4:
+					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_ihardlimit =
-						strtoll(cp, NULL, 10);
+						cvtval(lim, *endpt);
 					continue;
 				default:
 					warnx("incorrect quota specification: "
@@ -300,10 +300,10 @@ static void
 usage(void)
 {
 	fprintf(stderr, "%s\n%s\n%s\n%s\n%s\n%s\n%s\n%s\n",
-		"usage: edquota [-u] [-f fspath] [-p username] [-h] username ...",
+		"usage: edquota [-uh] [-f fspath] [-p username] username ...",
 		"       edquota [-u] -e fspath[:bslim[:bhlim[:islim[:ihlim]]]] [-e ...]",
 		"               username ...",
-		"       edquota -g [-f fspath] [-p groupname] [-h] groupname ...",
+		"       edquota -g [-h] [-f fspath] [-p groupname] groupname ...",
 		"       edquota -g -e fspath[:bslim[:bhlim[:islim[:ihlim]]]] [-e ...]",
 		"               groupname ...",
 		"       edquota [-u] -t [-f fspath]",
@@ -536,14 +536,14 @@ writeprivs(struct quotause *quplist, int outfd, char *name, int quotatype)
 char *
 fmthumanval(int64_t blocks)
 {
-	static char buf[7], numbuf[20];
+	static char numbuf[20];
 
 	if (hflag) {
-		humanize_number(buf, sizeof(buf) - (blocks < 0 ? 0 : 1),
+		humanize_number(numbuf, sizeof(numbuf) - (blocks < 0 ? 0 : 1),
 		    dbtob(blocks), "", HN_AUTOSCALE, HN_NOSPACE);
-		return (buf);
+		return (numbuf);
 	}
-	snprintf(numbuf, 20, "%lluK", dbtokb(blocks));
+	snprintf(numbuf, sizeof(numbuf), "%lluk", dbtokb(blocks));
 	return(numbuf);
 }
 
@@ -555,9 +555,8 @@ readprivs(struct quotause *quplist, char *inname)
 {
 	struct quotause *qup;
 	FILE *fd;
-	u_int64_t ihardlimit, isoftlimit, curinodes;
-	u_int64_t bhardlimit, bsoftlimit, curblocks;
-	char bhardunits, bsoftunits, curblockunits;
+	u_int64_t hardlimit, softlimit, curitems;
+	char hardunits, softunits, curitemunits;
 	int cnt;
 	char *cp;
 	struct dqblk dqblk;
@@ -584,47 +583,66 @@ readprivs(struct quotause *quplist, char *inname)
 		}
 		cnt = sscanf(cp,
 		    " in use: %llu%c, limits (soft = %llu%c, hard = %llu%c)",
-		    &curblocks, &curblockunits, &bsoftlimit, &bsoftunits,
-		    &bhardlimit, &bhardunits);
+		    &curitems, &curitemunits, &softlimit, &softunits,
+		    &hardlimit, &hardunits);
 		/*
 		 * The next three check for old-style input formats.
 		 */
 		if (cnt != 6)
 			cnt = sscanf(cp,
 			 " in use: %llu%c, limits (soft = %llu%c hard = %llu%c",
-			    &curblocks, &curblockunits, &bsoftlimit,
-			    &bsoftunits, &bhardlimit, &bhardunits);
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(cp,
 			" in use: %llu%c, limits (soft = %llu%c hard = %llu%c)",
-			    &curblocks, &curblockunits, &bsoftlimit,
-			    &bsoftunits, &bhardlimit, &bhardunits);
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(cp,
 			" in use: %llu%c, limits (soft = %llu%c, hard = %llu%c",
-			    &curblocks, &curblockunits, &bsoftlimit,
-			    &bsoftunits, &bhardlimit, &bhardunits);
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6) {
 			warnx("%s:%s: bad format", fsp, cp);
 			return (0);
 		}
-		dqblk.dqb_curblocks = cvtval(curblocks, curblockunits);
-		dqblk.dqb_bsoftlimit = cvtval(bsoftlimit, bsoftunits);
-		dqblk.dqb_bhardlimit = cvtval(bhardlimit, bhardunits);
+		dqblk.dqb_curblocks = cvtval(curitems, curitemunits);
+		dqblk.dqb_bsoftlimit = cvtval(softlimit, softunits);
+		dqblk.dqb_bhardlimit = cvtval(hardlimit, hardunits);
 		if ((cp = strtok(line2, "\n")) == NULL) {
 			warnx("%s: %s: bad format", fsp, line2);
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    "\tinodes in use: %llu, limits (soft = %llu, hard = %llu)",
-		    &curinodes, &isoftlimit, &ihardlimit);
+		    " in use: %llu%c, limits (soft = %llu%c, hard = %llu%c)",
+		    &curitems, &curitemunits, &softlimit,
+		    &softunits, &hardlimit, &hardunits);
+		/*
+		 * The next three check for old-style input formats.
+		 */
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			 " in use: %llu%c, limits (soft = %llu%c hard = %llu%c",
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			" in use: %llu%c, limits (soft = %llu%c hard = %llu%c)",
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
+		if (cnt != 6)
+			cnt = sscanf(cp,
+			" in use: %llu%c, limits (soft = %llu%c, hard = %llu%c",
+			    &curitems, &curitemunits, &softlimit,
+			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 3) {
 			warnx("%s: %s: bad format", fsp, line2);
 			return (0);
 		}
-		dqblk.dqb_curinodes = curinodes;
-		dqblk.dqb_isoftlimit = isoftlimit;
-		dqblk.dqb_ihardlimit = ihardlimit;
+		dqblk.dqb_curinodes = cvtval(curitems, curitemunits);
+		dqblk.dqb_isoftlimit = cvtval(softlimit, softunits);
+		dqblk.dqb_ihardlimit = cvtval(hardlimit, hardunits);
 		for (qup = quplist; qup; qup = qup->next) {
 			if (strcmp(fsp, qup->fsname))
 				continue;
@@ -651,8 +669,8 @@ readprivs(struct quotause *quplist, char *inname)
 			qup->dqblk.dqb_isoftlimit = dqblk.dqb_isoftlimit;
 			qup->dqblk.dqb_ihardlimit = dqblk.dqb_ihardlimit;
 			qup->flags |= FOUND;
-			/* No easy way to check change in curblocks */
-			if (dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes)
+			if (dqblk.dqb_curblocks == qup->dqblk.dqb_curblocks &&
+			    dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes)
 				break;
 			warnx("%s: cannot change current allocation", fsp);
 			break;
@@ -853,8 +871,11 @@ cvtval(u_int64_t limit, char units)
 	case 'e':
 		limit *= btodb(1152921504606846976);
 		break;
+	case ' ':
+		errx(2, "No space permitted between value and units\n");
+		break;
 	default:
-		warnx("%llu%c: unknown units, specify K, M, G, T, P, or E\n",
+		errx(2, "%llu%c: unknown units, specify K, M, G, T, P, or E\n",
 		    limit, units);
 		break;
 	}

From cdab8b664d3e54562c025e13b2a69b0abbab991e Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 10 Feb 2009 08:09:03 +0000
Subject: [PATCH 008/532] More updates to edquota based on feedback from
 Dag-Erling Smorgrav.

---
 usr.sbin/edquota/edquota.c | 169 +++++++++++++++++++++++++++----------
 1 file changed, 126 insertions(+), 43 deletions(-)

diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index 3331548c86c..94dff1ef0f7 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -101,9 +101,11 @@ struct quotause {
 int alldigits(const char *s);
 int cvtatos(u_int64_t, char *, u_int64_t *);
 char *cvtstoa(u_int64_t);
-u_int64_t cvtval(u_int64_t, char);
+u_int64_t cvtblkval(u_int64_t, char, const char *);
+u_int64_t cvtinoval(u_int64_t, char, const char *);
 int editit(char *);
-char *fmthumanval(int64_t);
+char *fmthumanvalblks(int64_t);
+char *fmthumanvalinos(int64_t);
 void freeprivs(struct quotause *);
 int getentry(const char *, int);
 struct quotause *getprivs(long, int, char *);
@@ -165,6 +167,12 @@ main(int argc, char *argv[])
 			     (cp = strsep(&optarg, ":")) != NULL; i++) {
 				if (cp != oldoptarg)
 					*(cp - 1) = ':';
+				if (i > 0 && !isdigit(*cp)) {
+					warnx("incorrect quota specification: "
+					    "%s", oldoptarg);
+					usage();
+					/* Not Reached */
+				}
 				switch (i) {
 				case 0:
 					strlcpy(qup->fsname, cp,
@@ -173,34 +181,34 @@ main(int argc, char *argv[])
 				case 1:
 					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_bsoftlimit =
-						cvtval(lim, *endpt);
+						cvtblkval(lim, *endpt,
+						    "block soft limit");
 					continue;
 				case 2:
 					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_bhardlimit =
-						cvtval(lim, *endpt);
+						cvtblkval(lim, *endpt,
+						    "block hard limit");
 					continue;
 				case 3:
 					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_isoftlimit =
-						cvtval(lim, *endpt);
+						cvtinoval(lim, *endpt,
+						    "inode soft limit");
 					continue;
 				case 4:
 					lim = strtoll(cp, &endpt, 10);
 					qup->dqblk.dqb_ihardlimit =
-						cvtval(lim, *endpt);
+						cvtinoval(lim, *endpt,
+						    "inode hard limit");
 					continue;
 				default:
 					warnx("incorrect quota specification: "
 					    "%s", oldoptarg);
 					usage();
-					break; /* XXX: report an error */
+					/* Not Reached */
 				}
 			}
-			qup->dqblk.dqb_bsoftlimit =
-			    btodb((off_t)qup->dqblk.dqb_bsoftlimit * 1024);
-			qup->dqblk.dqb_bhardlimit =
-			    btodb((off_t)qup->dqblk.dqb_bhardlimit * 1024);
 			if (protoprivs == NULL) {
 				protoprivs = curprivs = qup;
 			} else {
@@ -212,6 +220,7 @@ main(int argc, char *argv[])
 			break;
 		default:
 			usage();
+			/* Not Reached */
 		}
 	}
 	argc -= optind;
@@ -518,28 +527,29 @@ writeprivs(struct quotause *quplist, int outfd, char *name, int quotatype)
 	fprintf(fd, "Quotas for %s %s:\n", qfextension[quotatype], name);
 	for (qup = quplist; qup; qup = qup->next) {
 		fprintf(fd, "%s: in use: %s, ", qup->fsname,
-		    fmthumanval(qup->dqblk.dqb_curblocks));
+		    fmthumanvalblks(qup->dqblk.dqb_curblocks));
 		fprintf(fd, "limits (soft = %s, ",
-		    fmthumanval(qup->dqblk.dqb_bsoftlimit));
+		    fmthumanvalblks(qup->dqblk.dqb_bsoftlimit));
 		fprintf(fd, "hard = %s)\n",
-		    fmthumanval(qup->dqblk.dqb_bhardlimit));
-		fprintf(fd, "%s %llu, limits (soft = %llu, hard = %llu)\n",
-		    "\tinodes in use:",
-		    qup->dqblk.dqb_curinodes,
-		    qup->dqblk.dqb_isoftlimit,
-		    qup->dqblk.dqb_ihardlimit);
+		    fmthumanvalblks(qup->dqblk.dqb_bhardlimit));
+		fprintf(fd, "\tinodes in use: %s, ",
+		    fmthumanvalinos(qup->dqblk.dqb_curinodes));
+		fprintf(fd, "limits (soft = %s, ",
+		    fmthumanvalinos(qup->dqblk.dqb_isoftlimit));
+		fprintf(fd, "hard = %s)\n",
+		    fmthumanvalinos(qup->dqblk.dqb_ihardlimit));
 	}
 	fclose(fd);
 	return (1);
 }
 
 char *
-fmthumanval(int64_t blocks)
+fmthumanvalblks(int64_t blocks)
 {
 	static char numbuf[20];
 
 	if (hflag) {
-		humanize_number(numbuf, sizeof(numbuf) - (blocks < 0 ? 0 : 1),
+		humanize_number(numbuf, blocks < 0 ? 7 : 6,
 		    dbtob(blocks), "", HN_AUTOSCALE, HN_NOSPACE);
 		return (numbuf);
 	}
@@ -547,6 +557,20 @@ fmthumanval(int64_t blocks)
 	return(numbuf);
 }
 
+char *
+fmthumanvalinos(int64_t inos)
+{
+	static char numbuf[20];
+
+	if (hflag) {
+		humanize_number(numbuf, inos < 0 ? 7 : 6,
+		    inos, "", HN_AUTOSCALE, HN_NOSPACE | HN_DIVISOR_1000);
+		return (numbuf);
+	}
+	snprintf(numbuf, sizeof(numbuf), "%llu", inos);
+	return(numbuf);
+}
+
 /*
  * Merge changes to an ASCII file into a quotause list.
  */
@@ -607,42 +631,48 @@ readprivs(struct quotause *quplist, char *inname)
 			warnx("%s:%s: bad format", fsp, cp);
 			return (0);
 		}
-		dqblk.dqb_curblocks = cvtval(curitems, curitemunits);
-		dqblk.dqb_bsoftlimit = cvtval(softlimit, softunits);
-		dqblk.dqb_bhardlimit = cvtval(hardlimit, hardunits);
+		dqblk.dqb_curblocks = cvtblkval(curitems, curitemunits,
+		    "current block count");
+		dqblk.dqb_bsoftlimit = cvtblkval(softlimit, softunits,
+		    "block soft limit");
+		dqblk.dqb_bhardlimit = cvtblkval(hardlimit, hardunits,
+		    "block hard limit");
 		if ((cp = strtok(line2, "\n")) == NULL) {
 			warnx("%s: %s: bad format", fsp, line2);
 			return (0);
 		}
-		cnt = sscanf(cp,
-		    " in use: %llu%c, limits (soft = %llu%c, hard = %llu%c)",
+		cnt = sscanf(&cp[7],
+		    " in use: %llu%c limits (soft = %llu%c, hard = %llu%c)",
 		    &curitems, &curitemunits, &softlimit,
 		    &softunits, &hardlimit, &hardunits);
 		/*
 		 * The next three check for old-style input formats.
 		 */
 		if (cnt != 6)
-			cnt = sscanf(cp,
-			 " in use: %llu%c, limits (soft = %llu%c hard = %llu%c",
+			cnt = sscanf(&cp[7],
+			 " in use: %llu%c limits (soft = %llu%c hard = %llu%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
-			cnt = sscanf(cp,
-			" in use: %llu%c, limits (soft = %llu%c hard = %llu%c)",
+			cnt = sscanf(&cp[7],
+			" in use: %llu%c limits (soft = %llu%c hard = %llu%c)",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
-			cnt = sscanf(cp,
-			" in use: %llu%c, limits (soft = %llu%c, hard = %llu%c",
+			cnt = sscanf(&cp[7],
+			" in use: %llu%c limits (soft = %llu%c, hard = %llu%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
-		if (cnt != 3) {
-			warnx("%s: %s: bad format", fsp, line2);
+		if (cnt != 6) {
+			warnx("%s: %s: bad format cnt %d", fsp, &cp[7], cnt);
 			return (0);
 		}
-		dqblk.dqb_curinodes = cvtval(curitems, curitemunits);
-		dqblk.dqb_isoftlimit = cvtval(softlimit, softunits);
-		dqblk.dqb_ihardlimit = cvtval(hardlimit, hardunits);
+		dqblk.dqb_curinodes = cvtinoval(curitems, curitemunits,
+		    "current inode count");
+		dqblk.dqb_isoftlimit = cvtinoval(softlimit, softunits,
+		    "inode soft limit");
+		dqblk.dqb_ihardlimit = cvtinoval(hardlimit, hardunits,
+		    "inode hard limit");
 		for (qup = quplist; qup; qup = qup->next) {
 			if (strcmp(fsp, qup->fsname))
 				continue;
@@ -669,8 +699,10 @@ readprivs(struct quotause *quplist, char *inname)
 			qup->dqblk.dqb_isoftlimit = dqblk.dqb_isoftlimit;
 			qup->dqblk.dqb_ihardlimit = dqblk.dqb_ihardlimit;
 			qup->flags |= FOUND;
-			if (dqblk.dqb_curblocks == qup->dqblk.dqb_curblocks &&
-			    dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes)
+			/* Humanized input returns only approximate counts */
+			if (hflag ||
+			    (dqblk.dqb_curblocks == qup->dqblk.dqb_curblocks &&
+			     dqblk.dqb_curinodes == qup->dqblk.dqb_curinodes))
 				break;
 			warnx("%s: cannot change current allocation", fsp);
 			break;
@@ -836,7 +868,7 @@ cvtatos(u_int64_t period, char *units, u_int64_t *seconds)
  * Convert a limit to number of disk blocks.
  */
 u_int64_t
-cvtval(u_int64_t limit, char units)
+cvtblkval(u_int64_t limit, char units, const char *itemname)
 {
 
 	switch(units) {
@@ -872,11 +904,62 @@ cvtval(u_int64_t limit, char units)
 		limit *= btodb(1152921504606846976);
 		break;
 	case ' ':
-		errx(2, "No space permitted between value and units\n");
+		errx(2, "No space permitted between value and units for %s\n",
+		    itemname);
 		break;
 	default:
-		errx(2, "%llu%c: unknown units, specify K, M, G, T, P, or E\n",
-		    limit, units);
+		errx(2, "%llu%c: unknown units for %s, specify none, K, M, G, T, P, or E\n",
+		    limit, units, itemname);
+		break;
+	}
+	return (limit);
+}
+
+/*
+ * Convert a limit to number of inodes.
+ */
+u_int64_t
+cvtinoval(u_int64_t limit, char units, const char *itemname)
+{
+
+	switch(units) {
+	case 'B':
+	case 'b':
+	case '\0':	/* historic behavior */
+	case ',':	/* historic behavior */
+	case ')':	/* historic behavior */
+		break;
+	case 'K':
+	case 'k':
+		limit *= 1000;
+		break;
+	case 'M':
+	case 'm':
+		limit *= 1000000;
+		break;
+	case 'G':
+	case 'g':
+		limit *= 1000000000;
+		break;
+	case 'T':
+	case 't':
+		limit *= 1000000000000;
+		break;
+	case 'P':
+	case 'p':
+		limit *= 1000000000000000;
+		break;
+	case 'E':
+	case 'e':
+		limit *= 1000000000000000000;
+		break;
+	case ' ':
+		errx(2, "No space permitted between value and units for %s\n",
+		    itemname);
+		break;
+	default:
+		errx(2, "%llu%c: unknown units for %s, specify none, K, M, G, T, P, or E\n",
+		    limit, units, itemname);
 		break;
 	}
 	return (limit);

From 581fce8e7df72f43c9aa9eb4b45815991911b0fb Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 10 Feb 2009 08:11:44 +0000
Subject: [PATCH 009/532] Bug fixes found from using these functions in
 edquota.

---
 lib/libutil/quotafile.3 | 4 ++--
 lib/libutil/quotafile.c | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index a414a99a027..d819f619905 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -45,9 +45,9 @@
 .Ft "struct quotafile *"
 .Fn quota_create "const char *path"
 .Ft int
-.Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int type"
+.Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
 .Ft int
-.Fn quota_write "struct quotafile *qf" "const struct dqblk *dqb" "int type"
+.Fn quota_write "struct quotafile *qf" "const struct dqblk *dqb" "int id"
 .Ft int
 .Fn quota_close "struct quotafile *qf"
 .Sh DESCRIPTION
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index a4258b1ef30..c175df0b21c 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -231,7 +231,7 @@ quota_write32(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = id * sizeof(struct dqblk32);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb32, sizeof(dqb32)));
+	return (write(qf->fd, &dqb32, sizeof(dqb32)) == -1);
 }
 
 static int
@@ -252,7 +252,7 @@ quota_write64(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = sizeof(struct dqhdr64) + id * sizeof(struct dqblk64);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb64, sizeof(dqb64)));
+	return (write(qf->fd, &dqb64, sizeof(dqb64)) == -1);
 }
 
 int

From 916e406eb5992cdb49a4c90f634e3aa398aaad6a Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Fri, 13 Feb 2009 06:12:15 +0000
Subject: [PATCH 010/532] Move hasquota() function to libutil.

---
 lib/libutil/libutil.h      |   2 +
 lib/libutil/quotafile.c    |  56 +++++++++++++++++++-
 usr.sbin/edquota/edquota.c | 101 ++++++++++---------------------------
 3 files changed, 83 insertions(+), 76 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index 48ab2188c35..7796de0ac78 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -142,11 +142,13 @@ int pidfile_remove(struct pidfh *pfh);
 
 #ifdef _UFS_UFS_QUOTA_H_
 struct quotafile;
+struct fstab;
 struct quotafile *quota_open(const char *);
 struct quotafile *quota_create(const char *);
 void quota_close(struct quotafile *);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write(struct quotafile *, const struct dqblk *, int);
+int hasquota(struct fstab *, int, char **);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index c175df0b21c..2520e9fcc64 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -29,16 +29,19 @@
 
 #include <sys/types.h>
 #include <sys/endian.h>
+#include <sys/mount.h>
 #include <sys/stat.h>
 
 #include <ufs/ufs/quota.h>
 
 #include <errno.h>
 #include <fcntl.h>
+#include <fstab.h>
 #include <grp.h>
 #include <pwd.h>
 #include <libutil.h>
 #include <stdint.h>
+#include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
@@ -48,6 +51,8 @@ struct quotafile {
 	int type; /* 32 or 64 */
 };
 
+static const char *qfextension[] = INITQFNAMES;
+
 struct quotafile *
 quota_open(const char *fn)
 {
@@ -231,7 +236,7 @@ quota_write32(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = id * sizeof(struct dqblk32);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb32, sizeof(dqb32)) == -1);
+	return (write(qf->fd, &dqb32, sizeof(dqb32)) == sizeof(dqb32));
 }
 
 static int
@@ -252,7 +257,7 @@ quota_write64(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = sizeof(struct dqhdr64) + id * sizeof(struct dqblk64);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb64, sizeof(dqb64)) == -1);
+	return (write(qf->fd, &dqb64, sizeof(dqb64)) == sizeof(dqb64));
 }
 
 int
@@ -270,3 +275,50 @@ quota_write(struct quotafile *qf, const struct dqblk *dqb, int id)
 	}
 	/* not reached */
 }
+
+/*
+ * Check to see if a particular quota is to be enabled.
+ */
+int
+hasquota(struct fstab *fs, int type, char **qfnamep)
+{
+	char *opt;
+	char *cp;
+	struct statfs sfb;
+	static char initname, usrname[100], grpname[100];
+	static char buf[BUFSIZ];
+
+	if (!initname) {
+		(void)snprintf(usrname, sizeof(usrname), "%s%s",
+		    qfextension[USRQUOTA], QUOTAFILENAME);
+		(void)snprintf(grpname, sizeof(grpname), "%s%s",
+		    qfextension[GRPQUOTA], QUOTAFILENAME);
+		initname = 1;
+	}
+	strcpy(buf, fs->fs_mntops);
+	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
+		if ((cp = index(opt, '=')))
+			*cp++ = '\0';
+		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
+			break;
+		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
+			break;
+	}
+	if (!opt)
+		return (0);
+	if (cp)
+		*qfnamep = cp;
+	else {
+		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
+		    QUOTAFILENAME, qfextension[type]);
+		*qfnamep = buf;
+	}
+	/*
+	 * Ensure that the filesystem is mounted.
+	 */
+	if (statfs(fs->fs_file, &sfb) != 0 ||
+	    strcmp(fs->fs_file, sfb.f_mntonname)) {
+		return (0);
+	}
+	return (1);
+}
diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index 94dff1ef0f7..16594cff3ab 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -83,7 +83,6 @@ __FBSDID("$FreeBSD$");
 #define dbtokb(db)	(db)
 #endif
 
-const char *qfname = QUOTAFILENAME;
 const char *qfextension[] = INITQFNAMES;
 const char *quotagroup = QUOTAGROUP;
 char tmpfil[] = _PATH_TMP;
@@ -109,7 +108,6 @@ char *fmthumanvalinos(int64_t);
 void freeprivs(struct quotause *);
 int getentry(const char *, int);
 struct quotause *getprivs(long, int, char *);
-int hasquota(struct fstab *, int, char **);
 void putprivs(long, int, struct quotause *);
 int readprivs(struct quotause *, char *);
 int readtimes(struct quotause *, char *);
@@ -230,6 +228,8 @@ main(int argc, char *argv[])
 			if ((protoid = getentry(protoname, quotatype)) == -1)
 				exit(1);
 			protoprivs = getprivs(protoid, quotatype, fspath);
+			if (protoprivs == NULL)
+				exit(0);
 			for (qup = protoprivs; qup; qup = qup->next) {
 				qup->dqblk.dqb_btime = 0;
 				qup->dqblk.dqb_itime = 0;
@@ -259,20 +259,20 @@ main(int argc, char *argv[])
 						*argv);
 				if ((id = getentry(buf, quotatype)) < 0)
 					continue;
-				if (eflag) {
-					for (qup = protoprivs; qup;
-					    qup = qup->next) {
-						curprivs = getprivs(id,
-						    quotatype, qup->fsname);
-						if (curprivs == NULL)
-							continue;
-						strcpy(qup->qfname,
-						    curprivs->qfname);
-						strcpy(qup->fsname,
-						    curprivs->fsname);
-					}
+				if (!eflag) {
+					putprivs(id, quotatype, protoprivs);
+					continue;
+				}
+				for (qup = protoprivs; qup;
+				    qup = qup->next) {
+					curprivs = getprivs(id, quotatype,
+					    qup->fsname);
+					if (curprivs == NULL)
+						continue;
+					strcpy(qup->qfname, curprivs->qfname);
+					strcpy(qup->fsname, curprivs->fsname);
+					putprivs(id, quotatype, protoprivs);
 				}
-				putprivs(id, quotatype, protoprivs);
 			}
 		}
 		exit(0);
@@ -280,12 +280,12 @@ main(int argc, char *argv[])
 	tmpfd = mkstemp(tmpfil);
 	fchown(tmpfd, getuid(), getgid());
 	if (tflag) {
-		protoprivs = getprivs(0, quotatype, fspath);
-		if (writetimes(protoprivs, tmpfd, quotatype) == 0)
-			exit(1);
-		if (editit(tmpfil) && readtimes(protoprivs, tmpfil))
-			putprivs(0L, quotatype, protoprivs);
-		freeprivs(protoprivs);
+		if ((protoprivs = getprivs(0, quotatype, fspath)) != NULL) {
+			if (writetimes(protoprivs, tmpfd, quotatype) != 0 &&
+			    editit(tmpfil) && readtimes(protoprivs, tmpfil))
+				putprivs(0L, quotatype, protoprivs);
+			freeprivs(protoprivs);
+		}
 		close(tmpfd);
 		unlink(tmpfil);
 		exit(0);
@@ -293,7 +293,8 @@ main(int argc, char *argv[])
 	for ( ; argc > 0; argc--, argv++) {
 		if ((id = getentry(*argv, quotatype)) == -1)
 			continue;
-		curprivs = getprivs(id, quotatype, fspath);
+		if ((curprivs = getprivs(id, quotatype, fspath)) == NULL)
+			exit(1);
 		if (writeprivs(curprivs, tmpfd, *argv, quotatype) == 0)
 			continue;
 		if (editit(tmpfil) && readprivs(curprivs, tmpfil))
@@ -412,6 +413,9 @@ getprivs(long id, int quotatype, char *fspath)
 		quptail = qup;
 		qup->next = 0;
 	}
+	if (quphead == NULL) {
+		warnx("No quotas on %s", fspath ? fspath : "any filesystems");
+	}
 	endfsent();
 	return (quphead);
 }
@@ -462,7 +466,7 @@ putprivs(long id, int quotatype, struct quotause *quplist)
 			qup->dqblk.dqb_itime = 0;
 		qup->dqblk.dqb_curinodes = dqbuf.dqb_curinodes;
 		qup->dqblk.dqb_curblocks = dqbuf.dqb_curblocks;
-		if (quota_write(qf, &qup->dqblk, id) != 0) {
+		if (quota_write(qf, &qup->dqblk, id) == 0) {
 			warn("%s", qup->qfname);
 		}
 		quota_close(qf);
@@ -994,54 +998,3 @@ alldigits(const char *s)
 	} while ((c = *s++));
 	return (1);
 }
-
-/*
- * Check to see if a particular quota is to be enabled.
- */
-int
-hasquota(struct fstab *fs, int type, char **qfnamep)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s",
-		    qfextension[USRQUOTA], qfname);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], qfname);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    qfname, qfextension[type]);
-		*qfnamep = buf;
-	}
-	if (statfs(fs->fs_file, &sfb) != 0) {
-		warn("cannot statfs mount point %s", fs->fs_file);
-		sleep(3);
-		return (0);
-	}
-	if (strcmp(fs->fs_file, sfb.f_mntonname)) {
-		warnx("%s not mounted for %s quotas", fs->fs_file,
-		    type == USRQUOTA ? "user" : "group");
-		sleep(3);
-		return (0);
-	}
-	return (1);
-}

From 909d0c906fc23065abcf6f3535d7c81b729f74bf Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Fri, 13 Feb 2009 06:17:53 +0000
Subject: [PATCH 011/532] Move hasquota() function to libutil.

---
 lib/libutil/quotafile.3 | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index d819f619905..ac6af011703 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -50,6 +50,8 @@
 .Fn quota_write "struct quotafile *qf" "const struct dqblk *dqb" "int id"
 .Ft int
 .Fn quota_close "struct quotafile *qf"
+.Ft int
+.Fn hasquota "struct fstab *fs" "int type" "char **qfnamep"
 .Sh DESCRIPTION
 .Sh RETURN VALUES
 .Sh SEE ALSO

From a88984f248e353d85b18d6e609e7209fa9fb855f Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Fri, 13 Feb 2009 19:56:59 +0000
Subject: [PATCH 012/532] Make hasquota thread safe.

---
 lib/libutil/libutil.h   |  2 +-
 lib/libutil/quotafile.c | 17 ++++++++---------
 2 files changed, 9 insertions(+), 10 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index 7796de0ac78..b9aec78919d 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -148,7 +148,7 @@ struct quotafile *quota_create(const char *);
 void quota_close(struct quotafile *);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write(struct quotafile *, const struct dqblk *, int);
-int hasquota(struct fstab *, int, char **);
+int hasquota(struct fstab *, int, char *, int);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 2520e9fcc64..2b43b65a382 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -280,13 +280,13 @@ quota_write(struct quotafile *qf, const struct dqblk *dqb, int id)
  * Check to see if a particular quota is to be enabled.
  */
 int
-hasquota(struct fstab *fs, int type, char **qfnamep)
+hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
 {
 	char *opt;
 	char *cp;
 	struct statfs sfb;
+	char buf[BUFSIZ];
 	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
 
 	if (!initname) {
 		(void)snprintf(usrname, sizeof(usrname), "%s%s",
@@ -306,13 +306,6 @@ hasquota(struct fstab *fs, int type, char **qfnamep)
 	}
 	if (!opt)
 		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    QUOTAFILENAME, qfextension[type]);
-		*qfnamep = buf;
-	}
 	/*
 	 * Ensure that the filesystem is mounted.
 	 */
@@ -320,5 +313,11 @@ hasquota(struct fstab *fs, int type, char **qfnamep)
 	    strcmp(fs->fs_file, sfb.f_mntonname)) {
 		return (0);
 	}
+	if (cp) {
+		strncpy(qfnamep, cp, qfbufsize);
+	} else {
+		(void)snprintf(qfnamep, qfbufsize, "%s/%s.%s", fs->fs_file,
+		    QUOTAFILENAME, qfextension[type]);
+	}
 	return (1);
 }

From 8bd6a3ab4b32283502a1b58731eb0d0750b4f124 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Sat, 14 Feb 2009 08:08:08 +0000
Subject: [PATCH 013/532] Update the quotafile library to manage both active
 quotas via the quotactl(2) interface and inactive quotas by accessing the
 quota files directly.

Update the edquota program to use this new interface as proof of
concept.
---
 lib/libutil/libutil.h      |   7 +-
 lib/libutil/quotafile.3    | 121 +++++++++++++--
 lib/libutil/quotafile.c    | 303 +++++++++++++++++++++++++------------
 usr.sbin/edquota/edquota.c | 133 ++++++----------
 4 files changed, 359 insertions(+), 205 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index b9aec78919d..afd5db6d228 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -143,12 +143,11 @@ int pidfile_remove(struct pidfh *pfh);
 #ifdef _UFS_UFS_QUOTA_H_
 struct quotafile;
 struct fstab;
-struct quotafile *quota_open(const char *);
-struct quotafile *quota_create(const char *);
+struct quotafile *quota_open(struct fstab *, int, int);
 void quota_close(struct quotafile *);
 int quota_read(struct quotafile *, struct dqblk *, int);
-int quota_write(struct quotafile *, const struct dqblk *, int);
-int hasquota(struct fstab *, int, char *, int);
+int quota_write_limits(struct quotafile *, struct dqblk *, int);
+int quota_write_usage(struct quotafile *, struct dqblk *, int);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index ac6af011703..07a08a20c18 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -25,35 +25,132 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd November 4, 2008
+.Dd February 14, 2009
 .Dt QUOTAFILE 3
 .Os
 .Sh NAME
 .Nm quota_open
-.Nm quota_create
 .Nm quota_read
-.Nm quota_write
+.Nm quota_write_limits
+.Nm quota_write_usage
 .Nm quota_close
-.Nd "Manipulate quota files"
+.Nd "Manipulate quotas"
 .Sh LIBRARY
 .Lb libutil
 .Sh SYNOPSIS
 .In ufs/ufs/quota.h
 .In libutil.h
+.In fstab.h
 .Ft "struct quotafile *"
-.Fn quota_open "const char *path"
-.Ft "struct quotafile *"
-.Fn quota_create "const char *path"
+.Fn quota_open "struct fstab *fs" "int quotatype" "int openflags"
 .Ft int
 .Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
 .Ft int
-.Fn quota_write "struct quotafile *qf" "const struct dqblk *dqb" "int id"
+.Fn quota_write_limits "struct quotafile *qf" "struct dqblk *dqb" "int id"
+.Ft int
+.Fn quota_write_usage "struct quotafile *qf" "struct dqblk *dqb" "int id"
 .Ft int
 .Fn quota_close "struct quotafile *qf"
-.Ft int
-.Fn hasquota "struct fstab *fs" "int type" "char **qfnamep"
 .Sh DESCRIPTION
+These functions are designed to simplify access to filesystem quotas.
+If quotas are active on a filesystem,
+these functions will access them directly from the kernel using the
+.Fn quotactl
+system call.
+If quotas are not active,
+these functions will access them by reading and writing
+the quota files directly.
+.Pp
+The
+.Fn quota_open
+function takes a pointer to an
+.Vt fstab
+entry corresponding to the filesystem on which quotas
+are to be accessed.
+The
+.Va quotatype
+field indicates the type of quotas being sought, either
+.Dv USRQUOTA
+or
+.Dv GRPQUOTA .
+The
+.Va openflags
+are those used by the
+.Fn open
+system call, usually either
+.Dv O_RDONLY
+if the quotas are just to be read, or
+.Dv O_RDWR
+if the quotas are to be updated.
+The
+.Dv O_CREAT
+flag should be specified if a new quota file of the requested type should
+be created if it does not already exist.
+.Pp
+The
+.Fn quota_read
+function reads the quota from the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+into the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
+.Fn quota_write_limits
+function updates the limit fields (but not the usage fields)
+for the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+from the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
+.Fn quota_write_usage
+function updates the usage fields (but not the limit fields)
+for the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+from the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
+.Fn quota_close
+function closes any open file descriptors and frees any storage
+associated with the filesystem and quota type referenced by
+.Va qf .
 .Sh RETURN VALUES
+If the filesystem has quotas associated with it,
+.Fn quota_open
+returns a pointer to a 
+.Vt quotafile
+structure used in subsequent quota access calls.
+If the filesystem has no quotas, or access permission is denied
+.Dv NULL
+is returned and
+.Va errno
+is set to indicate the cause of failure.
+.Pp
+The
+.Fn quota_read ,
+.Fn quota_write_limits ,
+.Fn quota_write_usage ,
+and
+.Fn quota_close
+functions return zero on success.
+On error they return
+.Dv -1
+and set
+.Va errno
+to indicate the cause of failure.
 .Sh SEE ALSO
 .Xr quotactl 2 ,
 .Xr quota.user 5 ,
@@ -68,4 +165,6 @@ functions first appeared in
 The
 .Nm
 functions and this manual page were written by
-.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org .
+.An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org 
+and
+.An Marshall Kirk McKusick .
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 2b43b65a382..58c02b1c169 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -1,5 +1,6 @@
 /*-
  * Copyright (c) 2008 Dag-Erling Coïdan Smørgrav
+ * Copyright (c) 2008 Marshall Kirk McKusick
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -47,76 +48,132 @@
 #include <unistd.h>
 
 struct quotafile {
-	int fd;
-	int type; /* 32 or 64 */
+	int fd;				/* -1 means using quotactl for access */
+	int wordsize;			/* 32-bit or 64-bit limits */
+	int quotatype;			/* USRQUOTA or GRPQUOTA */
+	char fsname[MAXPATHLEN + 1];	/* mount point of filesystem */
+	char qfname[MAXPATHLEN + 1];	/* quota file if not using quotactl */
 };
 
 static const char *qfextension[] = INITQFNAMES;
 
-struct quotafile *
-quota_open(const char *fn)
+/*
+ * Check to see if a particular quota is to be enabled.
+ */
+static int
+hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
 {
-	struct quotafile *qf;
-	struct dqhdr64 dqh;
-	int serrno;
+	char *opt;
+	char *cp;
+	struct statfs sfb;
+	char buf[BUFSIZ];
+	static char initname, usrname[100], grpname[100];
 
-	if ((qf = calloc(1, sizeof(*qf))) == NULL)
-		return (NULL);
-	if ((qf->fd = open(fn, O_RDWR)) < 0) {
-		serrno = errno;
-		free(qf);
-		errno = serrno;
-		return (NULL);
+	if (!initname) {
+		(void)snprintf(usrname, sizeof(usrname), "%s%s",
+		    qfextension[USRQUOTA], QUOTAFILENAME);
+		(void)snprintf(grpname, sizeof(grpname), "%s%s",
+		    qfextension[GRPQUOTA], QUOTAFILENAME);
+		initname = 1;
 	}
-	qf->type = 32;
-	switch (read(qf->fd, &dqh, sizeof(dqh))) {
-	case -1:
-		serrno = errno;
-		close(qf->fd);
-		free(qf);
-		errno = serrno;
-		return (NULL);
-	case sizeof(dqh):
-		if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) != 0) {
-			/* no magic, assume 32 bits */
-			qf->type = 32;
-			return (qf);
-		}
-		if (be32toh(dqh.dqh_version) != Q_DQHDR64_VERSION ||
-		    be32toh(dqh.dqh_hdrlen) != sizeof(struct dqhdr64) ||
-		    be32toh(dqh.dqh_reclen) != sizeof(struct dqblk64)) {
-			/* correct magic, wrong version / lengths */
-			close(qf->fd);
-			free(qf);
-			errno = EINVAL;
-			return (NULL);
-		}
-		qf->type = 64;
-		return (qf);
-	default:
-		qf->type = 32;
-		return (qf);
+	strcpy(buf, fs->fs_mntops);
+	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
+		if ((cp = index(opt, '=')))
+			*cp++ = '\0';
+		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
+			break;
+		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
+			break;
 	}
-	/* not reached */
+	if (!opt)
+		return (0);
+	/*
+	 * Ensure that the filesystem is mounted.
+	 */
+	if (statfs(fs->fs_file, &sfb) != 0 ||
+	    strcmp(fs->fs_file, sfb.f_mntonname)) {
+		return (0);
+	}
+	if (cp) {
+		strncpy(qfnamep, cp, qfbufsize);
+	} else {
+		(void)snprintf(qfnamep, qfbufsize, "%s/%s.%s", fs->fs_file,
+		    QUOTAFILENAME, qfextension[type]);
+	}
+	return (1);
 }
 
 struct quotafile *
-quota_create(const char *fn)
+quota_open(struct fstab *fs, int quotatype, int openflags)
 {
 	struct quotafile *qf;
 	struct dqhdr64 dqh;
 	struct group *grp;
-	int serrno;
+	int qcmd, serrno;
 
 	if ((qf = calloc(1, sizeof(*qf))) == NULL)
 		return (NULL);
-	if ((qf->fd = open(fn, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0) {
+	qf->quotatype = quotatype;
+	strncpy(qf->fsname, fs->fs_file, sizeof(qf->fsname));
+	qcmd = QCMD(Q_GETQUOTA, quotatype);
+	if (quotactl(fs->fs_file, qcmd, 0, &dqh) == 0) {
+		qf->wordsize = 64;
+		qf->fd = -1;
+		return (qf);
+	}
+	if (!hasquota(fs, quotatype, qf->qfname, sizeof(qf->qfname))) {
+		free(qf);
+		errno = EOPNOTSUPP;
+		return (NULL);
+	}
+	if ((qf->fd = open(qf->qfname, openflags & O_ACCMODE)) < 0 &&
+	    (openflags & O_CREAT) == 0) {
 		serrno = errno;
 		free(qf);
 		errno = serrno;
 		return (NULL);
 	}
-	qf->type = 64;
+	/* File open worked, so process it */
+	if (qf->fd != -1) {
+		qf->wordsize = 32;
+		switch (read(qf->fd, &dqh, sizeof(dqh))) {
+		case -1:
+			serrno = errno;
+			close(qf->fd);
+			free(qf);
+			errno = serrno;
+			return (NULL);
+		case sizeof(dqh):
+			if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) != 0) {
+				/* no magic, assume 32 bits */
+				qf->wordsize = 32;
+				return (qf);
+			}
+			if (be32toh(dqh.dqh_version) != Q_DQHDR64_VERSION ||
+			    be32toh(dqh.dqh_hdrlen) != sizeof(struct dqhdr64) ||
+			    be32toh(dqh.dqh_reclen) != sizeof(struct dqblk64)) {
+				/* correct magic, wrong version / lengths */
+				close(qf->fd);
+				free(qf);
+				errno = EINVAL;
+				return (NULL);
+			}
+			qf->wordsize = 64;
+			return (qf);
+		default:
+			qf->wordsize = 32;
+			return (qf);
+		}
+		/* not reached */
+	}
+	/* Open failed above, but O_CREAT specified, so create a new file */
+	if ((qf->fd = open(qf->qfname, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0) {
+		serrno = errno;
+		free(qf);
+		errno = serrno;
+		return (NULL);
+	}
+	qf->wordsize = 64;
 	memset(&dqh, 0, sizeof(dqh));
 	memcpy(dqh.dqh_magic, Q_DQHDR64_MAGIC, sizeof(dqh.dqh_magic));
 	dqh.dqh_version = htobe32(Q_DQHDR64_VERSION);
@@ -124,7 +181,7 @@ quota_create(const char *fn)
 	dqh.dqh_reclen = htobe32(sizeof(struct dqblk64));
 	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh)) {
 		serrno = errno;
-		unlink(fn);
+		unlink(qf->qfname);
 		close(qf->fd);
 		free(qf);
 		errno = serrno;
@@ -140,7 +197,8 @@ void
 quota_close(struct quotafile *qf)
 {
 
-	close(qf->fd);
+	if (qf->fd != -1)
+		close(qf->fd);
 	free(qf);
 }
 
@@ -203,8 +261,13 @@ quota_read64(struct quotafile *qf, struct dqblk *dqb, int id)
 int
 quota_read(struct quotafile *qf, struct dqblk *dqb, int id)
 {
+	int qcmd;
 
-	switch (qf->type) {
+	if (qf->fd == -1) {
+		qcmd = QCMD(Q_GETQUOTA, qf->quotatype);
+		return (quotactl(qf->fsname, qcmd, id, dqb));
+	}
+	switch (qf->wordsize) {
 	case 32:
 		return quota_read32(qf, dqb, id);
 	case 64:
@@ -236,7 +299,9 @@ quota_write32(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = id * sizeof(struct dqblk32);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb32, sizeof(dqb32)) == sizeof(dqb32));
+	if (write(qf->fd, &dqb32, sizeof(dqb32)) == sizeof(dqb32))
+		return (0);
+	return (-1);
 }
 
 static int
@@ -257,14 +322,98 @@ quota_write64(struct quotafile *qf, const struct dqblk *dqb, int id)
 	off = sizeof(struct dqhdr64) + id * sizeof(struct dqblk64);
 	if (lseek(qf->fd, off, SEEK_SET) == -1)
 		return (-1);
-	return (write(qf->fd, &dqb64, sizeof(dqb64)) == sizeof(dqb64));
+	if (write(qf->fd, &dqb64, sizeof(dqb64)) == sizeof(dqb64))
+		return (0);
+	return (-1);
 }
 
 int
-quota_write(struct quotafile *qf, const struct dqblk *dqb, int id)
+quota_write_usage(struct quotafile *qf, struct dqblk *dqb, int id)
 {
+	struct dqblk dqbuf;
+	int qcmd;
 
-	switch (qf->type) {
+	if (qf->fd == -1) {
+		qcmd = QCMD(Q_SETUSE, qf->quotatype);
+		return (quotactl(qf->fsname, qcmd, id, dqb));
+	}
+	/*
+	 * Have to do read-modify-write of quota in file.
+	 */
+	if (quota_read(qf, &dqbuf, id) != 0)
+		return (-1);
+	/*
+	 * Reset time limit if have a soft limit and were
+	 * previously under it, but are now over it.
+	 */
+	if (dqbuf.dqb_bsoftlimit && id != 0 &&
+	    dqbuf.dqb_curblocks < dqbuf.dqb_bsoftlimit &&
+	    dqb->dqb_curblocks >= dqbuf.dqb_bsoftlimit)
+		dqbuf.dqb_btime = 0;
+	if (dqbuf.dqb_isoftlimit && id != 0 &&
+	    dqbuf.dqb_curinodes < dqbuf.dqb_isoftlimit &&
+	    dqb->dqb_curinodes >= dqbuf.dqb_isoftlimit)
+		dqbuf.dqb_itime = 0;
+	dqbuf.dqb_curinodes = dqb->dqb_curinodes;
+	dqbuf.dqb_curblocks = dqb->dqb_curblocks;
+	/*
+	 * Write it back.
+	 */
+	switch (qf->wordsize) {
+	case 32:
+		return quota_write32(qf, &dqbuf, id);
+	case 64:
+		return quota_write64(qf, &dqbuf, id);
+	default:
+		errno = EINVAL;
+		return (-1);
+	}
+	/* not reached */
+}
+
+int
+quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
+{
+	struct dqblk dqbuf;
+	int qcmd;
+
+	if (qf->fd == -1) {
+		qcmd = QCMD(Q_SETQUOTA, qf->quotatype);
+		return (quotactl(qf->fsname, qcmd, id, dqb));
+	}
+	/*
+	 * Have to do read-modify-write of quota in file.
+	 */
+	if (quota_read(qf, &dqbuf, id) != 0)
+		return (-1);
+	/*
+	 * Reset time limit if have a soft limit and were
+	 * previously under it, but are now over it
+	 * or if there previously was no soft limit, but
+	 * now have one and are over it.
+	 */
+	if (dqbuf.dqb_bsoftlimit && id != 0 &&
+	    dqbuf.dqb_curblocks < dqbuf.dqb_bsoftlimit &&
+	    dqbuf.dqb_curblocks >= dqb->dqb_bsoftlimit)
+		dqb->dqb_btime = 0;
+	if (dqbuf.dqb_bsoftlimit == 0 && id != 0 &&
+	    dqb->dqb_bsoftlimit > 0 &&
+	    dqbuf.dqb_curblocks >= dqb->dqb_bsoftlimit)
+		dqb->dqb_btime = 0;
+	if (dqbuf.dqb_isoftlimit && id != 0 &&
+	    dqbuf.dqb_curinodes < dqbuf.dqb_isoftlimit &&
+	    dqbuf.dqb_curinodes >= dqb->dqb_isoftlimit)
+		dqb->dqb_itime = 0;
+	if (dqbuf.dqb_isoftlimit == 0 && id !=0 &&
+	    dqb->dqb_isoftlimit > 0 &&
+	    dqbuf.dqb_curinodes >= dqb->dqb_isoftlimit)
+		dqb->dqb_itime = 0;
+	dqb->dqb_curinodes = dqbuf.dqb_curinodes;
+	dqb->dqb_curblocks = dqbuf.dqb_curblocks;
+	/*
+	 * Write it back.
+	 */
+	switch (qf->wordsize) {
 	case 32:
 		return quota_write32(qf, dqb, id);
 	case 64:
@@ -275,49 +424,3 @@ quota_write(struct quotafile *qf, const struct dqblk *dqb, int id)
 	}
 	/* not reached */
 }
-
-/*
- * Check to see if a particular quota is to be enabled.
- */
-int
-hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	char buf[BUFSIZ];
-	static char initname, usrname[100], grpname[100];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s",
-		    qfextension[USRQUOTA], QUOTAFILENAME);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], QUOTAFILENAME);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	/*
-	 * Ensure that the filesystem is mounted.
-	 */
-	if (statfs(fs->fs_file, &sfb) != 0 ||
-	    strcmp(fs->fs_file, sfb.f_mntonname)) {
-		return (0);
-	}
-	if (cp) {
-		strncpy(qfnamep, cp, qfbufsize);
-	} else {
-		(void)snprintf(qfnamep, qfbufsize, "%s/%s.%s", fs->fs_file,
-		    QUOTAFILENAME, qfextension[type]);
-	}
-	return (1);
-}
diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index 16594cff3ab..b2e57a155ed 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -84,16 +84,15 @@ __FBSDID("$FreeBSD$");
 #endif
 
 const char *qfextension[] = INITQFNAMES;
-const char *quotagroup = QUOTAGROUP;
 char tmpfil[] = _PATH_TMP;
 int hflag;
 
 struct quotause {
 	struct	quotause *next;
-	long	flags;
+	struct	quotafile *qf;
 	struct	dqblk dqblk;
+	int	flags;
 	char	fsname[MAXPATHLEN + 1];
-	char	qfname[1];	/* actually longer */
 };
 #define	FOUND	0x01
 
@@ -108,7 +107,7 @@ char *fmthumanvalinos(int64_t);
 void freeprivs(struct quotause *);
 int getentry(const char *, int);
 struct quotause *getprivs(long, int, char *);
-void putprivs(long, int, struct quotause *);
+void putprivs(long, struct quotause *);
 int readprivs(struct quotause *, char *);
 int readtimes(struct quotause *, char *);
 static void usage(void);
@@ -142,6 +141,11 @@ main(int argc, char *argv[])
 			fspath = optarg;
 			break;
 		case 'p':
+			if (eflag) {
+				warnx("cannot specify both -e and -p");
+				usage();
+				/* not reached */
+			}
 			protoname = optarg;
 			pflag++;
 			break;
@@ -158,7 +162,12 @@ main(int argc, char *argv[])
 			tflag++;
 			break;
 		case 'e':
-			if ((qup = calloc(1, sizeof(*qup) + BUFSIZ)) == NULL)
+			if (pflag) {
+				warnx("cannot specify both -e and -p");
+				usage();
+				/* not reached */
+			}
+			if ((qup = calloc(1, sizeof(*qup))) == NULL)
 				errx(2, "out of memory");
 			oldoptarg = optarg;
 			for (i = 0, cp = optarg;
@@ -214,7 +223,6 @@ main(int argc, char *argv[])
 				curprivs = qup;
 			}
 			eflag++;
-			pflag++;
 			break;
 		default:
 			usage();
@@ -223,8 +231,8 @@ main(int argc, char *argv[])
 	}
 	argc -= optind;
 	argv += optind;
-	if (pflag) {
-		if (protoprivs == NULL) {
+	if (pflag || eflag) {
+		if (pflag) {
 			if ((protoid = getentry(protoname, quotatype)) == -1)
 				exit(1);
 			protoprivs = getprivs(protoid, quotatype, fspath);
@@ -259,22 +267,23 @@ main(int argc, char *argv[])
 						*argv);
 				if ((id = getentry(buf, quotatype)) < 0)
 					continue;
-				if (!eflag) {
-					putprivs(id, quotatype, protoprivs);
+				if (pflag) {
+					putprivs(id, protoprivs);
 					continue;
 				}
-				for (qup = protoprivs; qup;
-				    qup = qup->next) {
+				for (qup = protoprivs; qup; qup = qup->next) {
 					curprivs = getprivs(id, quotatype,
 					    qup->fsname);
 					if (curprivs == NULL)
 						continue;
-					strcpy(qup->qfname, curprivs->qfname);
-					strcpy(qup->fsname, curprivs->fsname);
-					putprivs(id, quotatype, protoprivs);
+					curprivs->dqblk = qup->dqblk;
+					putprivs(id, curprivs);
+					freeprivs(curprivs);
 				}
 			}
 		}
+		if (pflag)
+			freeprivs(protoprivs);
 		exit(0);
 	}
 	tmpfd = mkstemp(tmpfil);
@@ -283,7 +292,7 @@ main(int argc, char *argv[])
 		if ((protoprivs = getprivs(0, quotatype, fspath)) != NULL) {
 			if (writetimes(protoprivs, tmpfd, quotatype) != 0 &&
 			    editit(tmpfil) && readtimes(protoprivs, tmpfil))
-				putprivs(0L, quotatype, protoprivs);
+				putprivs(0L, protoprivs);
 			freeprivs(protoprivs);
 		}
 		close(tmpfd);
@@ -298,7 +307,7 @@ main(int argc, char *argv[])
 		if (writeprivs(curprivs, tmpfd, *argv, quotatype) == 0)
 			continue;
 		if (editit(tmpfil) && readprivs(curprivs, tmpfil))
-			putprivs(id, quotatype, curprivs);
+			putprivs(id, curprivs);
 		freeprivs(curprivs);
 	}
 	close(tmpfd);
@@ -366,46 +375,29 @@ getprivs(long id, int quotatype, char *fspath)
 	struct fstab *fs;
 	struct quotause *qup, *quptail;
 	struct quotause *quphead;
-	int qcmd, qupsize;
-	char *qfpathname;
-	static int warned = 0;
 
 	setfsent();
 	quphead = quptail = NULL;
-	qcmd = QCMD(Q_GETQUOTA, quotatype);
 	while ((fs = getfsent())) {
 		if (fspath && *fspath && strcmp(fspath, fs->fs_spec) &&
 		    strcmp(fspath, fs->fs_file))
 			continue;
 		if (strcmp(fs->fs_vfstype, "ufs"))
 			continue;
-		if (!hasquota(fs, quotatype, &qfpathname))
+		if ((qf = quota_open(fs, quotatype, O_CREAT|O_RDWR)) == NULL) {
+			if (errno != EOPNOTSUPP)
+				warn("cannot open quotas on %s", fs->fs_file);
+			continue;
+		}
+		if ((qup = (struct quotause *)calloc(1, sizeof(*qup))) == NULL)
+			errx(2, "out of memory");
+		qup->qf = qf;
+		strncpy(qup->fsname, fs->fs_file, sizeof(qup->fsname));
+		if (quota_read(qf, &qup->dqblk, id) == -1) {
+			warn("cannot read quotas on %s", fs->fs_file);
+			freeprivs(qup);
 			continue;
-		qupsize = sizeof(*qup) + strlen(qfpathname);
-		if ((qup = (struct quotause *)malloc(qupsize)) == NULL)
-			errx(2, "out of memory");
-		if (quotactl(fs->fs_file, qcmd, id, &qup->dqblk) != 0) {
-			if (errno == EOPNOTSUPP && !warned) {
-				warned++;
-		warnx("warning: quotas are not compiled into this kernel");
-				sleep(3);
-			}
-			if ((qf = quota_open(qfpathname)) == NULL &&
-			    (qf = quota_create(qfpathname)) == NULL) {
-				warn("%s", qfpathname);
-				free(qup);
-				continue;
-			}
-			if (quota_read(qf, &qup->dqblk, id) != 0) {
-				warn("read error in %s", qfpathname);
-				quota_close(qf);
-				free(qup);
-				continue;
-			}
-			quota_close(qf);
 		}
-		strcpy(qup->qfname, qfpathname);
-		strcpy(qup->fsname, fs->fs_file);
 		if (quphead == NULL)
 			quphead = qup;
 		else
@@ -424,53 +416,13 @@ getprivs(long id, int quotatype, char *fspath)
  * Store the requested quota information.
  */
 void
-putprivs(long id, int quotatype, struct quotause *quplist)
+putprivs(long id, struct quotause *quplist)
 {
-	struct quotafile *qf;
 	struct quotause *qup;
-	int qcmd;
-	struct dqblk dqbuf;
 
-	qcmd = QCMD(Q_SETQUOTA, quotatype);
-	for (qup = quplist; qup; qup = qup->next) {
-		if (quotactl(qup->fsname, qcmd, id, &qup->dqblk) == 0)
-			continue;
-		if ((qf = quota_open(qup->qfname)) == NULL) {
-			warn("%s", qup->qfname);
-			continue;
-		}
-		if (quota_read(qf, &dqbuf, id) != 0) {
-			warn("read error in %s", qup->qfname);
-			quota_close(qf);
-			continue;
-		}
-		/*
-		 * Reset time limit if have a soft limit and were
-		 * previously under it, but are now over it
-		 * or if there previously was no soft limit, but
-		 * now have one and are over it.
-		 */
-		if (dqbuf.dqb_bsoftlimit && id != 0 &&
-		    dqbuf.dqb_curblocks < dqbuf.dqb_bsoftlimit &&
-		    dqbuf.dqb_curblocks >= qup->dqblk.dqb_bsoftlimit)
-			qup->dqblk.dqb_btime = 0;
-		if (dqbuf.dqb_bsoftlimit == 0 && id != 0 &&
-		    dqbuf.dqb_curblocks >= qup->dqblk.dqb_bsoftlimit)
-			qup->dqblk.dqb_btime = 0;
-		if (dqbuf.dqb_isoftlimit && id != 0 &&
-		    dqbuf.dqb_curinodes < dqbuf.dqb_isoftlimit &&
-		    dqbuf.dqb_curinodes >= qup->dqblk.dqb_isoftlimit)
-			qup->dqblk.dqb_itime = 0;
-		if (dqbuf.dqb_isoftlimit == 0 && id !=0 &&
-		    dqbuf.dqb_curinodes >= qup->dqblk.dqb_isoftlimit)
-			qup->dqblk.dqb_itime = 0;
-		qup->dqblk.dqb_curinodes = dqbuf.dqb_curinodes;
-		qup->dqblk.dqb_curblocks = dqbuf.dqb_curblocks;
-		if (quota_write(qf, &qup->dqblk, id) == 0) {
-			warn("%s", qup->qfname);
-		}
-		quota_close(qf);
-	}
+	for (qup = quplist; qup; qup = qup->next)
+		if (quota_write_limits(qup->qf, &qup->dqblk, id) == -1)
+			warn("%s", qup->fsname);
 }
 
 /*
@@ -978,6 +930,7 @@ freeprivs(struct quotause *quplist)
 	struct quotause *qup, *nextqup;
 
 	for (qup = quplist; qup; qup = nextqup) {
+		quota_close(qup->qf);
 		nextqup = qup->next;
 		free(qup);
 	}

From 91b2a3e2047d5ec6b2a753119be9ace40e4a8631 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Fri, 25 Sep 2009 18:44:34 +0000
Subject: [PATCH 014/532] Improve comments, and remove a bogus 0 id check.

---
 sys/ufs/ufs/ufs_quota.c | 51 ++++++++++++++++++++++++++++-------------
 1 file changed, 35 insertions(+), 16 deletions(-)

diff --git a/sys/ufs/ufs/ufs_quota.c b/sys/ufs/ufs/ufs_quota.c
index ddba6010100..04d28e434d2 100644
--- a/sys/ufs/ufs/ufs_quota.c
+++ b/sys/ufs/ufs/ufs_quota.c
@@ -60,6 +60,8 @@ __FBSDID("$FreeBSD$");
 #include <ufs/ufs/ufsmount.h>
 #include <ufs/ufs/ufs_extern.h>
 
+CTASSERT(sizeof(struct dqblk64) == sizeof(struct dqhdr64));
+
 static int unprivileged_get_quota = 0;
 SYSCTL_INT(_security_bsd, OID_AUTO, unprivileged_get_quota, CTLFLAG_RW,
     &unprivileged_get_quota, 0,
@@ -1124,6 +1126,14 @@ dqhashfind(struct dqhash *dqh, u_long id, struct vnode *dqvp)
 
 /*
  * Determine the quota file type.
+ *
+ * A 32-bit quota file is simply an array of struct dqblk32.
+ *
+ * A 64-bit quota file is a struct dqhdr64 followed by an array of struct
+ * dqblk64.  The header contains various magic bits which allow us to be
+ * reasonably confident that it is indeeda 64-bit quota file and not just
+ * a 32-bit quota file that just happens to "look right".
+ *
  */
 static int
 dqopen(struct vnode *vp, struct ufsmount *ump, int type)
@@ -1312,23 +1322,8 @@ hfound:		DQI_LOCK(dq);
 	/*
 	 * Read the requested quota record from the quota file, performing
 	 * any necessary conversions.
-	 *
-	 * The record's offset within the file depends on the size of the
-	 * record, which means we need to know whether it's a 32-bit file
-	 * or a 64-bit file.
-	 *
-	 * Luckily, root's record is always at offset 0, and most of it is
-	 * unused, so we can use it to store a magic number indicating the
-	 * file format.  Due to an acute lack of imagination, this magic
-	 * number, stored in the first byte of root's record and hence the
-	 * first byte of the file, is 64.
-	 *
-	 * Another lucky break is that quotaon() always loads root's
-	 * record, to get the default values for dq_btime and dq_itime, so
-	 * we will always have a chance to check the file format before
-	 * being asked for a "real" record.
 	 */
-	if (id == 0 || (ump->um_qflags[type] & QTF_64BIT)) {
+	if (ump->um_qflags[type] & QTF_64BIT) {
 		recsize = sizeof(struct dqblk64);
 		base = sizeof(struct dqhdr64);
 	} else {
@@ -1597,6 +1592,10 @@ dqflush(struct vnode *vp)
 
 #define CLIP32(u64) (u64 > UINT32_MAX ? UINT32_MAX : (uint32_t)u64)
 
+/*
+ * Convert on-disk 32-bit host-order structure to in-memory 64-bit
+ * host-order structure.
+ */
 static void
 dqb32_dq(const struct dqblk32 *dqb32, struct dquot *dq)
 {
@@ -1611,6 +1610,10 @@ dqb32_dq(const struct dqblk32 *dqb32, struct dquot *dq)
 	dq->dq_itime = dqb32->dqb_itime;
 }
 
+/*
+ * Convert on-disk 64-bit network-order structure to in-memory 64-bit
+ * host-order structure.
+ */
 static void
 dqb64_dq(const struct dqblk64 *dqb64, struct dquot *dq)
 {
@@ -1625,6 +1628,10 @@ dqb64_dq(const struct dqblk64 *dqb64, struct dquot *dq)
 	dq->dq_itime = be64toh(dqb64->dqb_itime);
 }
 
+/*
+ * Convert in-memory 64-bit host-order structure to on-disk 32-bit
+ * host-order structure.
+ */
 static void
 dq_dqb32(const struct dquot *dq, struct dqblk32 *dqb32)
 {
@@ -1639,6 +1646,10 @@ dq_dqb32(const struct dquot *dq, struct dqblk32 *dqb32)
 	dqb32->dqb_itime = CLIP32(dq->dq_itime);
 }
 
+/*
+ * Convert in-memory host-order 64-bit structure to on-disk 64-bit
+ * network-order structure.
+ */
 static void
 dq_dqb64(const struct dquot *dq, struct dqblk64 *dqb64)
 {
@@ -1653,6 +1664,10 @@ dq_dqb64(const struct dquot *dq, struct dqblk64 *dqb64)
 	dqb64->dqb_itime = htobe64(dq->dq_itime);
 }
 
+/*
+ * Convert in-memory 64-bit host-order structure to in-memory 32-bit
+ * host-order structure.
+ */
 static void
 dqb64_dqb32(const struct dqblk64 *dqb64, struct dqblk32 *dqb32)
 {
@@ -1667,6 +1682,10 @@ dqb64_dqb32(const struct dqblk64 *dqb64, struct dqblk32 *dqb32)
 	dqb32->dqb_itime = CLIP32(dqb64->dqb_itime);
 }
 
+/*
+ * Convert in-memory 32-bit host-order structure to in-memory 64-bit
+ * host-order structure.
+ */
 static void
 dqb32_dqb64(const struct dqblk32 *dqb32, struct dqblk64 *dqb64)
 {

From 0bb2e5de607991a2c9d5ed9e27956ddd8d60fb35 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Fri, 25 Sep 2009 18:50:33 +0000
Subject: [PATCH 015/532] Further improve comments.

---
 sys/ufs/ufs/ufs_quota.c | 18 ++++++------------
 1 file changed, 6 insertions(+), 12 deletions(-)

diff --git a/sys/ufs/ufs/ufs_quota.c b/sys/ufs/ufs/ufs_quota.c
index 04d28e434d2..5b1da50b73a 100644
--- a/sys/ufs/ufs/ufs_quota.c
+++ b/sys/ufs/ufs/ufs_quota.c
@@ -1593,8 +1593,7 @@ dqflush(struct vnode *vp)
 #define CLIP32(u64) (u64 > UINT32_MAX ? UINT32_MAX : (uint32_t)u64)
 
 /*
- * Convert on-disk 32-bit host-order structure to in-memory 64-bit
- * host-order structure.
+ * Convert 32-bit host-order structure to dquot.
  */
 static void
 dqb32_dq(const struct dqblk32 *dqb32, struct dquot *dq)
@@ -1611,8 +1610,7 @@ dqb32_dq(const struct dqblk32 *dqb32, struct dquot *dq)
 }
 
 /*
- * Convert on-disk 64-bit network-order structure to in-memory 64-bit
- * host-order structure.
+ * Convert 64-bit network-order structure to dquot.
  */
 static void
 dqb64_dq(const struct dqblk64 *dqb64, struct dquot *dq)
@@ -1629,8 +1627,7 @@ dqb64_dq(const struct dqblk64 *dqb64, struct dquot *dq)
 }
 
 /*
- * Convert in-memory 64-bit host-order structure to on-disk 32-bit
- * host-order structure.
+ * Convert dquot to 32-bit host-order structure.
  */
 static void
 dq_dqb32(const struct dquot *dq, struct dqblk32 *dqb32)
@@ -1647,8 +1644,7 @@ dq_dqb32(const struct dquot *dq, struct dqblk32 *dqb32)
 }
 
 /*
- * Convert in-memory host-order 64-bit structure to on-disk 64-bit
- * network-order structure.
+ * Convert dquot to 64-bit network-order structure.
  */
 static void
 dq_dqb64(const struct dquot *dq, struct dqblk64 *dqb64)
@@ -1665,8 +1661,7 @@ dq_dqb64(const struct dquot *dq, struct dqblk64 *dqb64)
 }
 
 /*
- * Convert in-memory 64-bit host-order structure to in-memory 32-bit
- * host-order structure.
+ * Convert 64-bit host-order structure to 32-bit host-order structure.
  */
 static void
 dqb64_dqb32(const struct dqblk64 *dqb64, struct dqblk32 *dqb32)
@@ -1683,8 +1678,7 @@ dqb64_dqb32(const struct dqblk64 *dqb64, struct dqblk32 *dqb32)
 }
 
 /*
- * Convert in-memory 32-bit host-order structure to in-memory 64-bit
- * host-order structure.
+ * Convert 32-bit host-order structure to 64-bit host-order structure.
  */
 static void
 dqb32_dqb64(const struct dqblk32 *dqb32, struct dqblk64 *dqb64)

From fdd356a8d1e79def831b31494da89efbac6bb028 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Fri, 25 Sep 2009 23:27:07 +0000
Subject: [PATCH 016/532] Style

---
 lib/libutil/quotafile.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 58c02b1c169..98f6f7c2d68 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -269,9 +269,9 @@ quota_read(struct quotafile *qf, struct dqblk *dqb, int id)
 	}
 	switch (qf->wordsize) {
 	case 32:
-		return quota_read32(qf, dqb, id);
+		return (quota_read32(qf, dqb, id));
 	case 64:
-		return quota_read64(qf, dqb, id);
+		return (quota_read64(qf, dqb, id));
 	default:
 		errno = EINVAL;
 		return (-1);
@@ -361,9 +361,9 @@ quota_write_usage(struct quotafile *qf, struct dqblk *dqb, int id)
 	 */
 	switch (qf->wordsize) {
 	case 32:
-		return quota_write32(qf, &dqbuf, id);
+		return (quota_write32(qf, &dqbuf, id));
 	case 64:
-		return quota_write64(qf, &dqbuf, id);
+		return (quota_write64(qf, &dqbuf, id));
 	default:
 		errno = EINVAL;
 		return (-1);
@@ -415,9 +415,9 @@ quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
 	 */
 	switch (qf->wordsize) {
 	case 32:
-		return quota_write32(qf, dqb, id);
+		return (quota_write32(qf, dqb, id));
 	case 64:
-		return quota_write64(qf, dqb, id);
+		return (quota_write64(qf, dqb, id));
 	default:
 		errno = EINVAL;
 		return (-1);

From 5666aadb3ddf92915fa484c460bcf9623d08fcf8 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Sat, 26 Sep 2009 23:16:06 +0000
Subject: [PATCH 017/532] Further extend the quotafile API.

---
 lib/libutil/Makefile    |   6 ++-
 lib/libutil/libutil.h   |   5 +-
 lib/libutil/quotafile.3 |  67 ++++++++++++++++++++++----
 lib/libutil/quotafile.c | 101 +++++++++++++++++++++++++++-------------
 4 files changed, 137 insertions(+), 42 deletions(-)

diff --git a/lib/libutil/Makefile b/lib/libutil/Makefile
index d7bd6dc145a..6b489d3b9d3 100644
--- a/lib/libutil/Makefile
+++ b/lib/libutil/Makefile
@@ -60,8 +60,12 @@ MLINKS+=pidfile.3 pidfile_open.3 \
 	pidfile.3 pidfile_close.3 \
 	pidfile.3 pidfile_remove.3
 MLINKS+=quotafile.3 quota_open.3 \
+	quotafile.3 quota_fsname.3 \
+	quotafile.3 quota_qfname.3 \
+	quotafile.3 quota_statfs.3 \
 	quotafile.3 quota_read.3 \
-	quotafile.3 quota_write.3 \
+	quotafile.3 quota_write_limits.3 \
+	quotafile.3 quota_write_usage.3 \
 	quotafile.3 quota_close.3
 
 .include <bsd.lib.mk>
diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index afd5db6d228..2190a532f79 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -144,10 +144,13 @@ int pidfile_remove(struct pidfh *pfh);
 struct quotafile;
 struct fstab;
 struct quotafile *quota_open(struct fstab *, int, int);
-void quota_close(struct quotafile *);
+const char *quota_fsname(const struct quotafile *);
+const char *quota_qfname(const struct quotafile *);
+int quota_check_path(const struct quotafile *, const char *path);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write_limits(struct quotafile *, struct dqblk *, int);
 int quota_write_usage(struct quotafile *, struct dqblk *, int);
+void quota_close(struct quotafile *);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index 07a08a20c18..4762bf92558 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -25,11 +25,14 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd February 14, 2009
+.Dd September 26, 2009
 .Dt QUOTAFILE 3
 .Os
 .Sh NAME
 .Nm quota_open
+.Nm quota_fsname
+.Nm quota_qfname
+.Nm quota_check_path
 .Nm quota_read
 .Nm quota_write_limits
 .Nm quota_write_usage
@@ -38,11 +41,19 @@
 .Sh LIBRARY
 .Lb libutil
 .Sh SYNOPSIS
+.In sys/param.h
+.In sys/mount.h
 .In ufs/ufs/quota.h
 .In libutil.h
 .In fstab.h
 .Ft "struct quotafile *"
 .Fn quota_open "struct fstab *fs" "int quotatype" "int openflags"
+.Ft "const char *"
+.Fn quota_fsname "const struct quotafile *qf"
+.Ft "const char *"
+.Fn quota_qfname "const struct quotafile *qf"
+.Ft int
+.Fn quota_check_path "const struct quotafile *qf" "const char *path"
 .Ft int
 .Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
 .Ft int
@@ -84,12 +95,42 @@ if the quotas are just to be read, or
 if the quotas are to be updated.
 The
 .Dv O_CREAT
-flag should be specified if a new quota file of the requested type should
-be created if it does not already exist.
+flag should be specified if a new quota file of the requested type
+should be created if it does not already exist.
+.Pp
+The
+.Fn quota_fsname
+function returns a pointer to a buffer containing the path to the root
+of the file system that corresponds to its
+.Va qf
+argument, as listed in
+.Pa /etc/fstab .
+Note that this may be a symbolic link to the actual directory.
+.Pp
+The
+.Fn quota_qfname
+function returns a pointer to a buffer containing the name of the
+quota file that corresponds to its
+.Va qf
+argument.
+Note that this may be a symbolic link to the actual file.
+.Pp
+The
+.Fn quota_check_path
+function checks if the specified path is within the filesystem that
+corresponds to its
+.Va qf
+argument.
+If the
+.Va path
+argument refers to a symbolic link,
+.Fn quota_check_path
+will follow it.
 .Pp
 The
 .Fn quota_read
-function reads the quota from the filesystem and quota type referenced by
+function reads the quota from the filesystem and quota type referenced
+by
 .Va qf
 for the user (or group) specified by
 .Va id
@@ -127,6 +168,9 @@ The
 function closes any open file descriptors and frees any storage
 associated with the filesystem and quota type referenced by
 .Va qf .
+.Sh IMPLEMENTATION NOTES
+If the underlying quota file is in the old 32-bit format, limit and
+usage values written to the quota file will be clipped to 32 bits.
 .Sh RETURN VALUES
 If the filesystem has quotas associated with it,
 .Fn quota_open
@@ -137,7 +181,15 @@ If the filesystem has no quotas, or access permission is denied
 .Dv NULL
 is returned and
 .Va errno
-is set to indicate the cause of failure.
+is set to indicate the error.
+.Pp
+The
+.Fn quota_check_path
+function returns\~1 for a positive result and\~0 for a negative
+result.
+If an error occurs, it returns\~-1 and sets
+.Va errno
+to indicate the error.
 .Pp
 The
 .Fn quota_read ,
@@ -146,11 +198,10 @@ The
 and
 .Fn quota_close
 functions return zero on success.
-On error they return
-.Dv -1
+On error they return\~-1
 and set
 .Va errno
-to indicate the cause of failure.
+to indicate the error.
 .Sh SEE ALSO
 .Xr quotactl 2 ,
 .Xr quota.user 5 ,
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 98f6f7c2d68..6dcb6979fb1 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -49,8 +49,10 @@
 
 struct quotafile {
 	int fd;				/* -1 means using quotactl for access */
+	int accmode;			/* access mode */
 	int wordsize;			/* 32-bit or 64-bit limits */
 	int quotatype;			/* USRQUOTA or GRPQUOTA */
+	dev_t dev;			/* device */
 	char fsname[MAXPATHLEN + 1];	/* mount point of filesystem */
 	char qfname[MAXPATHLEN + 1];	/* quota file if not using quotactl */
 };
@@ -59,6 +61,7 @@ static const char *qfextension[] = INITQFNAMES;
 
 /*
  * Check to see if a particular quota is to be enabled.
+ * XXX merge into quota_open
  */
 static int
 hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
@@ -69,6 +72,11 @@ hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
 	char buf[BUFSIZ];
 	static char initname, usrname[100], grpname[100];
 
+	/*
+	 * XXX
+	 * 1) we only need one of these
+	 * 2) fstab may specify a different filename
+	 */
 	if (!initname) {
 		(void)snprintf(usrname, sizeof(usrname), "%s%s",
 		    qfextension[USRQUOTA], QUOTAFILENAME);
@@ -109,12 +117,17 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 	struct quotafile *qf;
 	struct dqhdr64 dqh;
 	struct group *grp;
+	struct stat st;
 	int qcmd, serrno;
 
 	if ((qf = calloc(1, sizeof(*qf))) == NULL)
 		return (NULL);
+	qf->fd = -1;
 	qf->quotatype = quotatype;
 	strncpy(qf->fsname, fs->fs_file, sizeof(qf->fsname));
+	if (stat(qf->fsname, &st) != 0)
+		goto error;
+	qf->dev = st.st_dev;
 	qcmd = QCMD(Q_GETQUOTA, quotatype);
 	if (quotactl(fs->fs_file, qcmd, 0, &dqh) == 0) {
 		qf->wordsize = 64;
@@ -122,27 +135,19 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 		return (qf);
 	}
 	if (!hasquota(fs, quotatype, qf->qfname, sizeof(qf->qfname))) {
-		free(qf);
 		errno = EOPNOTSUPP;
-		return (NULL);
-	}
-	if ((qf->fd = open(qf->qfname, openflags & O_ACCMODE)) < 0 &&
-	    (openflags & O_CREAT) == 0) {
-		serrno = errno;
-		free(qf);
-		errno = serrno;
-		return (NULL);
+		goto error;
 	}
+	qf->accmode = openflags & O_ACCMODE;
+	if ((qf->fd = open(qf->qfname, qf->accmode)) < 0 &&
+	    (openflags & O_CREAT) != O_CREAT)
+		goto error;
 	/* File open worked, so process it */
 	if (qf->fd != -1) {
 		qf->wordsize = 32;
 		switch (read(qf->fd, &dqh, sizeof(dqh))) {
 		case -1:
-			serrno = errno;
-			close(qf->fd);
-			free(qf);
-			errno = serrno;
-			return (NULL);
+			goto error;
 		case sizeof(dqh):
 			if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) != 0) {
 				/* no magic, assume 32 bits */
@@ -153,10 +158,8 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 			    be32toh(dqh.dqh_hdrlen) != sizeof(struct dqhdr64) ||
 			    be32toh(dqh.dqh_reclen) != sizeof(struct dqblk64)) {
 				/* correct magic, wrong version / lengths */
-				close(qf->fd);
-				free(qf);
 				errno = EINVAL;
-				return (NULL);
+				goto error;
 			}
 			qf->wordsize = 64;
 			return (qf);
@@ -166,31 +169,33 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 		}
 		/* not reached */
 	}
-	/* Open failed above, but O_CREAT specified, so create a new file */
-	if ((qf->fd = open(qf->qfname, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0) {
-		serrno = errno;
-		free(qf);
-		errno = serrno;
-		return (NULL);
-	}
+	/* open failed, but O_CREAT was specified, so create a new file */
+	if ((qf->fd = open(qf->qfname, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0)
+		goto error;
 	qf->wordsize = 64;
 	memset(&dqh, 0, sizeof(dqh));
 	memcpy(dqh.dqh_magic, Q_DQHDR64_MAGIC, sizeof(dqh.dqh_magic));
 	dqh.dqh_version = htobe32(Q_DQHDR64_VERSION);
 	dqh.dqh_hdrlen = htobe32(sizeof(struct dqhdr64));
 	dqh.dqh_reclen = htobe32(sizeof(struct dqblk64));
-	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh)) {
-		serrno = errno;
-		unlink(qf->qfname);
-		close(qf->fd);
-		free(qf);
-		errno = serrno;
-		return (NULL);
-	}
+	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh))
+		goto error;
 	grp = getgrnam(QUOTAGROUP);
 	fchown(qf->fd, 0, grp ? grp->gr_gid : 0);
 	fchmod(qf->fd, 0640);
 	return (qf);
+error:
+	serrno = errno;
+	/* did we have an open file? */
+	if (qf->fd != -1) {
+		/* was it one we created ourselves? */
+		if ((openflags & O_CREAT) == O_CREAT)
+			unlink(qf->qfname);
+		close(qf->fd);
+	}
+	free(qf);
+	errno = serrno;
+	return (NULL);
 }
 
 void
@@ -202,6 +207,30 @@ quota_close(struct quotafile *qf)
 	free(qf);
 }
 
+const char *
+quota_fsname(const struct quotafile *qf)
+{
+
+	return (qf->fsname);
+}
+
+const char *
+quota_qfname(const struct quotafile *qf)
+{
+
+	return (qf->qfname);
+}
+
+int
+quota_check_path(const struct quotafile *qf, const char *path)
+{
+	struct stat st;
+
+	if (stat(path, &st) == -1)
+		return (-1);
+	return (st.st_dev == qf->dev);
+}
+
 static int
 quota_read32(struct quotafile *qf, struct dqblk *dqb, int id)
 {
@@ -333,6 +362,10 @@ quota_write_usage(struct quotafile *qf, struct dqblk *dqb, int id)
 	struct dqblk dqbuf;
 	int qcmd;
 
+	if ((qf->accmode & O_RDWR) != O_RDWR) {
+		errno = EBADF;
+		return (-1);
+	}
 	if (qf->fd == -1) {
 		qcmd = QCMD(Q_SETUSE, qf->quotatype);
 		return (quotactl(qf->fsname, qcmd, id, dqb));
@@ -377,6 +410,10 @@ quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
 	struct dqblk dqbuf;
 	int qcmd;
 
+	if ((qf->accmode & O_RDWR) != O_RDWR) {
+		errno = EBADF;
+		return (-1);
+	}
 	if (qf->fd == -1) {
 		qcmd = QCMD(Q_SETQUOTA, qf->quotatype);
 		return (quotactl(qf->fsname, qcmd, id, dqb));

From 6cfc1c1e184c27e3d9319b2bd9ebafcf50856071 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Sun, 27 Sep 2009 17:33:26 +0000
Subject: [PATCH 018/532] type / format fixup

---
 usr.sbin/edquota/edquota.c | 70 +++++++++++++++++++++-----------------
 1 file changed, 38 insertions(+), 32 deletions(-)

diff --git a/usr.sbin/edquota/edquota.c b/usr.sbin/edquota/edquota.c
index b2e57a155ed..69ea497fa52 100644
--- a/usr.sbin/edquota/edquota.c
+++ b/usr.sbin/edquota/edquota.c
@@ -97,10 +97,10 @@ struct quotause {
 #define	FOUND	0x01
 
 int alldigits(const char *s);
-int cvtatos(u_int64_t, char *, u_int64_t *);
-char *cvtstoa(u_int64_t);
-u_int64_t cvtblkval(u_int64_t, char, const char *);
-u_int64_t cvtinoval(u_int64_t, char, const char *);
+int cvtatos(uint64_t, char *, uint64_t *);
+char *cvtstoa(uint64_t);
+uint64_t cvtblkval(uint64_t, char, const char *);
+uint64_t cvtinoval(uint64_t, char, const char *);
 int editit(char *);
 char *fmthumanvalblks(int64_t);
 char *fmthumanvalinos(int64_t);
@@ -121,7 +121,7 @@ main(int argc, char *argv[])
 	long id, protoid;
 	int i, quotatype, range, tmpfd;
 	uid_t startuid, enduid;
-	u_int64_t lim;
+	uint64_t lim;
 	char *protoname, *cp, *endpt, *oldoptarg;
 	int eflag = 0, tflag = 0, pflag = 0, ch;
 	char *fspath = NULL;
@@ -509,7 +509,7 @@ fmthumanvalblks(int64_t blocks)
 		    dbtob(blocks), "", HN_AUTOSCALE, HN_NOSPACE);
 		return (numbuf);
 	}
-	snprintf(numbuf, sizeof(numbuf), "%lluk", dbtokb(blocks));
+	snprintf(numbuf, sizeof(numbuf), "%juk", (uintmax_t)dbtokb(blocks));
 	return(numbuf);
 }
 
@@ -523,7 +523,7 @@ fmthumanvalinos(int64_t inos)
 		    inos, "", HN_AUTOSCALE, HN_NOSPACE | HN_DIVISOR_1000);
 		return (numbuf);
 	}
-	snprintf(numbuf, sizeof(numbuf), "%llu", inos);
+	snprintf(numbuf, sizeof(numbuf), "%ju", (uintmax_t)inos);
 	return(numbuf);
 }
 
@@ -535,7 +535,7 @@ readprivs(struct quotause *quplist, char *inname)
 {
 	struct quotause *qup;
 	FILE *fd;
-	u_int64_t hardlimit, softlimit, curitems;
+	uintmax_t hardlimit, softlimit, curitems;
 	char hardunits, softunits, curitemunits;
 	int cnt;
 	char *cp;
@@ -562,7 +562,7 @@ readprivs(struct quotause *quplist, char *inname)
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    " in use: %llu%c, limits (soft = %llu%c, hard = %llu%c)",
+		    " in use: %ju%c, limits (soft = %ju%c, hard = %ju%c)",
 		    &curitems, &curitemunits, &softlimit, &softunits,
 		    &hardlimit, &hardunits);
 		/*
@@ -570,17 +570,17 @@ readprivs(struct quotause *quplist, char *inname)
 		 */
 		if (cnt != 6)
 			cnt = sscanf(cp,
-			 " in use: %llu%c, limits (soft = %llu%c hard = %llu%c",
+			 " in use: %ju%c, limits (soft = %ju%c hard = %ju%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(cp,
-			" in use: %llu%c, limits (soft = %llu%c hard = %llu%c)",
+			" in use: %ju%c, limits (soft = %ju%c hard = %ju%c)",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(cp,
-			" in use: %llu%c, limits (soft = %llu%c, hard = %llu%c",
+			" in use: %ju%c, limits (soft = %ju%c, hard = %ju%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6) {
@@ -598,7 +598,7 @@ readprivs(struct quotause *quplist, char *inname)
 			return (0);
 		}
 		cnt = sscanf(&cp[7],
-		    " in use: %llu%c limits (soft = %llu%c, hard = %llu%c)",
+		    " in use: %ju%c limits (soft = %ju%c, hard = %ju%c)",
 		    &curitems, &curitemunits, &softlimit,
 		    &softunits, &hardlimit, &hardunits);
 		/*
@@ -606,17 +606,17 @@ readprivs(struct quotause *quplist, char *inname)
 		 */
 		if (cnt != 6)
 			cnt = sscanf(&cp[7],
-			 " in use: %llu%c limits (soft = %llu%c hard = %llu%c",
+			 " in use: %ju%c limits (soft = %ju%c hard = %ju%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(&cp[7],
-			" in use: %llu%c limits (soft = %llu%c hard = %llu%c)",
+			" in use: %ju%c limits (soft = %ju%c hard = %ju%c)",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6)
 			cnt = sscanf(&cp[7],
-			" in use: %llu%c limits (soft = %llu%c, hard = %llu%c",
+			" in use: %ju%c limits (soft = %ju%c, hard = %ju%c",
 			    &curitems, &curitemunits, &softlimit,
 			    &softunits, &hardlimit, &hardunits);
 		if (cnt != 6) {
@@ -717,7 +717,7 @@ readtimes(struct quotause *quplist, char *inname)
 	FILE *fd;
 	int cnt;
 	char *cp;
-	u_int64_t itime, btime, iseconds, bseconds;
+	uintmax_t itime, btime, iseconds, bseconds;
 	char *fsp, bunits[10], iunits[10], line1[BUFSIZ];
 
 	fd = fopen(inname, "r");
@@ -740,7 +740,7 @@ readtimes(struct quotause *quplist, char *inname)
 			return (0);
 		}
 		cnt = sscanf(cp,
-		    " block grace period: %llu %s file grace period: %llu %s",
+		    " block grace period: %ju %s file grace period: %ju %s",
 		    &btime, bunits, &itime, iunits);
 		if (cnt != 4) {
 			warnx("%s:%s: bad format", fsp, cp);
@@ -779,21 +779,25 @@ readtimes(struct quotause *quplist, char *inname)
  * Convert seconds to ASCII times.
  */
 char *
-cvtstoa(u_int64_t secs)
+cvtstoa(uint64_t secs)
 {
 	static char buf[20];
 
 	if (secs % (24 * 60 * 60) == 0) {
 		secs /= 24 * 60 * 60;
-		sprintf(buf, "%llu day%s", secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%ju day%s", (uintmax_t)secs,
+		    secs == 1 ? "" : "s");
 	} else if (secs % (60 * 60) == 0) {
 		secs /= 60 * 60;
-		sprintf(buf, "%llu hour%s", secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%ju hour%s", (uintmax_t)secs,
+		    secs == 1 ? "" : "s");
 	} else if (secs % 60 == 0) {
 		secs /= 60;
-		sprintf(buf, "%llu minute%s", secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%ju minute%s", (uintmax_t)secs,
+		    secs == 1 ? "" : "s");
 	} else
-		sprintf(buf, "%llu second%s", secs, secs == 1 ? "" : "s");
+		sprintf(buf, "%ju second%s", (uintmax_t)secs,
+		    secs == 1 ? "" : "s");
 	return (buf);
 }
 
@@ -801,7 +805,7 @@ cvtstoa(u_int64_t secs)
  * Convert ASCII input times to seconds.
  */
 int
-cvtatos(u_int64_t period, char *units, u_int64_t *seconds)
+cvtatos(uint64_t period, char *units, uint64_t *seconds)
 {
 
 	if (bcmp(units, "second", 6) == 0)
@@ -823,8 +827,8 @@ cvtatos(u_int64_t period, char *units, u_int64_t *seconds)
 /*
  * Convert a limit to number of disk blocks.
  */
-u_int64_t
-cvtblkval(u_int64_t limit, char units, const char *itemname)
+uint64_t
+cvtblkval(uint64_t limit, char units, const char *itemname)
 {
 
 	switch(units) {
@@ -864,8 +868,9 @@ cvtblkval(u_int64_t limit, char units, const char *itemname)
 		    itemname);
 		break;
 	default:
-		errx(2, "%llu%c: unknown units for %s, specify none, K, M, G, T, P, or E\n",
-		    limit, units, itemname);
+		errx(2, "%ju%c: unknown units for %s, specify "
+		    "none, K, M, G, T, P, or E\n",
+		    (uintmax_t)limit, units, itemname);
 		break;
 	}
 	return (limit);
@@ -874,8 +879,8 @@ cvtblkval(u_int64_t limit, char units, const char *itemname)
 /*
  * Convert a limit to number of inodes.
  */
-u_int64_t
-cvtinoval(u_int64_t limit, char units, const char *itemname)
+uint64_t
+cvtinoval(uint64_t limit, char units, const char *itemname)
 {
 
 	switch(units) {
@@ -914,8 +919,9 @@ cvtinoval(u_int64_t limit, char units, const char *itemname)
 		    itemname);
 		break;
 	default:
-		errx(2, "%llu%c: unknown units for %s, specify none, K, M, G, T, P, or E\n",
-		    limit, units, itemname);
+		errx(2, "%ju%c: unknown units for %s, specify "
+		    "none, K, M, G, T, P, or E\n",
+		    (uintmax_t)limit, units, itemname);
 		break;
 	}
 	return (limit);

From 6197731d81b1001aa10bd4371c3f0e9e5796a915 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 20 Oct 2009 05:37:54 +0000
Subject: [PATCH 019/532] Add quota_maxid which returns the maximum user (or
 group) identifier in an associated quotafile. Needed by repquota.

Bug fix in quota_read.
---
 lib/libutil/libutil.h   |  1 +
 lib/libutil/quotafile.3 | 14 +++++++++++++-
 lib/libutil/quotafile.c | 26 ++++++++++++++++++++++----
 3 files changed, 36 insertions(+), 5 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index 2190a532f79..b5ff6fd9e07 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -146,6 +146,7 @@ struct fstab;
 struct quotafile *quota_open(struct fstab *, int, int);
 const char *quota_fsname(const struct quotafile *);
 const char *quota_qfname(const struct quotafile *);
+int quota_maxid(struct quotafile *);
 int quota_check_path(const struct quotafile *, const char *path);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write_limits(struct quotafile *, struct dqblk *, int);
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index 4762bf92558..d0c27d730b8 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -32,6 +32,7 @@
 .Nm quota_open
 .Nm quota_fsname
 .Nm quota_qfname
+.Nm quota_maxid
 .Nm quota_check_path
 .Nm quota_read
 .Nm quota_write_limits
@@ -44,8 +45,9 @@
 .In sys/param.h
 .In sys/mount.h
 .In ufs/ufs/quota.h
-.In libutil.h
+.In fcntl.h
 .In fstab.h
+.In libutil.h
 .Ft "struct quotafile *"
 .Fn quota_open "struct fstab *fs" "int quotatype" "int openflags"
 .Ft "const char *"
@@ -53,6 +55,8 @@
 .Ft "const char *"
 .Fn quota_qfname "const struct quotafile *qf"
 .Ft int
+.Fn quota_maxid "const struct quotafile *qf"
+.Ft int
 .Fn quota_check_path "const struct quotafile *qf" "const char *path"
 .Ft int
 .Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
@@ -116,6 +120,14 @@ argument.
 Note that this may be a symbolic link to the actual file.
 .Pp
 The
+.Fn quota_maxid
+function returns the maximum user (or group)
+.Va id
+contained in the quota file associated with its
+.Va qf
+argument.
+.Pp
+The
 .Fn quota_check_path
 function checks if the specified path is within the filesystem that
 corresponds to its
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 6dcb6979fb1..4c63a17b778 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -128,13 +128,13 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 	if (stat(qf->fsname, &st) != 0)
 		goto error;
 	qf->dev = st.st_dev;
+	serrno = hasquota(fs, quotatype, qf->qfname, sizeof(qf->qfname));
 	qcmd = QCMD(Q_GETQUOTA, quotatype);
 	if (quotactl(fs->fs_file, qcmd, 0, &dqh) == 0) {
 		qf->wordsize = 64;
-		qf->fd = -1;
 		return (qf);
 	}
-	if (!hasquota(fs, quotatype, qf->qfname, sizeof(qf->qfname))) {
+	if (serrno == 0) {
 		errno = EOPNOTSUPP;
 		goto error;
 	}
@@ -231,6 +231,24 @@ quota_check_path(const struct quotafile *qf, const char *path)
 	return (st.st_dev == qf->dev);
 }
 
+int
+quota_maxid(struct quotafile *qf)
+{
+	struct stat st;
+
+	if (stat(qf->qfname, &st) < 0)
+		return (0);
+	switch (qf->wordsize) {
+	case 32:
+		return (st.st_size / sizeof(struct dqblk32));
+	case 64:
+		return (st.st_size / sizeof(struct dqblk64) - 1);
+	default:
+		return (0);
+	}
+	/* not reached */
+}
+
 static int
 quota_read32(struct quotafile *qf, struct dqblk *dqb, int id)
 {
@@ -242,7 +260,7 @@ quota_read32(struct quotafile *qf, struct dqblk *dqb, int id)
 		return (-1);
 	switch (read(qf->fd, &dqb32, sizeof(dqb32))) {
 	case 0:
-		memset(&dqb, 0, sizeof(*dqb));
+		memset(dqb, 0, sizeof(*dqb));
 		return (0);
 	case sizeof(dqb32):
 		dqb->dqb_bhardlimit = dqb32.dqb_bhardlimit;
@@ -270,7 +288,7 @@ quota_read64(struct quotafile *qf, struct dqblk *dqb, int id)
 		return (-1);
 	switch (read(qf->fd, &dqb64, sizeof(dqb64))) {
 	case 0:
-		memset(&dqb, 0, sizeof(*dqb));
+		memset(dqb, 0, sizeof(*dqb));
 		return (0);
 	case sizeof(dqb64):
 		dqb->dqb_bhardlimit = be64toh(dqb64.dqb_bhardlimit);

From 0cf50fd7dc5610695e4246425de60a5f00f389ef Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 20 Oct 2009 05:40:51 +0000
Subject: [PATCH 020/532] Update to work with quotafile functions in -libutil.
 Add -h flag to provide humanized output.

---
 usr.sbin/repquota/repquota.8 |   7 +-
 usr.sbin/repquota/repquota.c | 196 ++++++++++++++---------------------
 2 files changed, 84 insertions(+), 119 deletions(-)

diff --git a/usr.sbin/repquota/repquota.8 b/usr.sbin/repquota/repquota.8
index b669fdf8266..8b5ab689ac7 100644
--- a/usr.sbin/repquota/repquota.8
+++ b/usr.sbin/repquota/repquota.8
@@ -39,12 +39,14 @@
 .Nd summarize quotas for a file system
 .Sh SYNOPSIS
 .Nm
+.Op Fl h
 .Op Fl g
 .Op Fl n
 .Op Fl u
 .Op Fl v
 .Ar filesystem Ar ...
 .Nm
+.Op Fl h
 .Op Fl g
 .Op Fl n
 .Op Fl u
@@ -64,6 +66,9 @@ Print the quotas of all the file systems listed in
 .It Fl g
 Print only group quotas (the default is to print both
 group and user quotas if they exist).
+.It Fl h
+Display information in a more human readable format
+rather than in historic kilobyte format.
 .It Fl n
 Display user and group IDs numerically rather than converting to
 a user or group name.
@@ -75,7 +80,7 @@ Print a header line before printing each file system quotas.
 .El
 .Pp
 For each user or group, the current
-number files and amount of space (in kilobytes) is
+number files and amount of space is
 printed, along with any quotas created with
 .Xr edquota 8 .
 .Pp
diff --git a/usr.sbin/repquota/repquota.c b/usr.sbin/repquota/repquota.c
index 19c185820b1..276a78dc65e 100644
--- a/usr.sbin/repquota/repquota.c
+++ b/usr.sbin/repquota/repquota.c
@@ -52,8 +52,10 @@ __FBSDID("$FreeBSD$");
 #include <ufs/ufs/quota.h>
 #include <err.h>
 #include <errno.h>
+#include <fcntl.h>
 #include <fstab.h>
 #include <grp.h>
+#include <libutil.h>
 #include <pwd.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -80,7 +82,6 @@ const char *qfextension[] = INITQFNAMES;
 
 struct fileusage {
 	struct	fileusage *fu_next;
-	struct	dqblk fu_dqblk;
 	u_long	fu_id;
 	char	fu_name[1];
 	/* actually bigger */
@@ -94,11 +95,12 @@ u_long highid[MAXQUOTAS];	/* highest addid()'ed identifier per type */
 int	vflag;			/* verbose */
 int	aflag;			/* all filesystems */
 int	nflag;			/* display user/group by id */
+int	hflag;			/* display in human readable format */
 
-int hasquota(struct fstab *, int, char **);
 int oneof(char *, char *[], int);
-int repquota(struct fstab *, int, char *);
+int repquota(struct fstab *, int);
 char *timeprt(time_t);
+static void prthumanval(int64_t bytes);
 static void usage(void);
 
 int
@@ -109,9 +111,8 @@ main(int argc, char *argv[])
 	struct group *gr;
 	int ch, gflag = 0, uflag = 0, errs = 0;
 	long i, argnum, done = 0;
-	char *qfnp;
 
-	while ((ch = getopt(argc, argv, "agnuv")) != -1) {
+	while ((ch = getopt(argc, argv, "aghnuv")) != -1) {
 		switch(ch) {
 		case 'a':
 			aflag++;
@@ -119,6 +120,9 @@ main(int argc, char *argv[])
 		case 'g':
 			gflag++;
 			break;
+		case 'h':
+			hflag++;
+			break;
 		case 'n':
 			nflag++;
 			break;
@@ -158,19 +162,19 @@ main(int argc, char *argv[])
 		if (strcmp(fs->fs_vfstype, "ufs"))
 			continue;
 		if (aflag) {
-			if (gflag && hasquota(fs, GRPQUOTA, &qfnp))
-				errs += repquota(fs, GRPQUOTA, qfnp);
-			if (uflag && hasquota(fs, USRQUOTA, &qfnp))
-				errs += repquota(fs, USRQUOTA, qfnp);
+			if (gflag)
+				errs += repquota(fs, GRPQUOTA);
+			if (uflag)
+				errs += repquota(fs, USRQUOTA);
 			continue;
 		}
 		if ((argnum = oneof(fs->fs_file, argv, argc)) >= 0 ||
 		    (argnum = oneof(fs->fs_spec, argv, argc)) >= 0) {
 			done |= 1 << argnum;
-			if (gflag && hasquota(fs, GRPQUOTA, &qfnp))
-				errs += repquota(fs, GRPQUOTA, qfnp);
-			if (uflag && hasquota(fs, USRQUOTA, &qfnp))
-				errs += repquota(fs, USRQUOTA, qfnp);
+			if (gflag)
+				errs += repquota(fs, GRPQUOTA);
+			if (uflag)
+				errs += repquota(fs, USRQUOTA);
 		}
 	}
 	endfsent();
@@ -184,87 +188,92 @@ static void
 usage(void)
 {
 	fprintf(stderr, "%s\n%s\n",
-		"usage: repquota [-v] [-g] [-n] [-u] -a",
-		"       repquota [-v] [-g] [-n] [-u] filesystem ...");
+		"usage: repquota [-h] [-v] [-g] [-n] [-u] -a",
+		"       repquota [-h] [-v] [-g] [-n] [-u] filesystem ...");
 	exit(1);
 }
 
 int
-repquota(struct fstab *fs, int type, char *qfpathname)
+repquota(struct fstab *fs, int type)
 {
 	struct fileusage *fup;
-	FILE *qf;
-	u_long id;
+	struct quotafile *qf;
+	u_long id, maxid;
 	struct dqblk dqbuf;
-	static struct dqblk zerodqblk;
-	static int warned = 0;
 	static int multiple = 0;
 
-	if (quotactl(fs->fs_file, QCMD(Q_SYNC, type), 0, 0) < 0 &&
-	    errno == EOPNOTSUPP && !warned && vflag) {
-		warned++;
-		fprintf(stdout,
-		    "*** Warning: Quotas are not compiled into this kernel\n");
+	if ((qf = quota_open(fs, type, O_RDONLY)) == NULL) {
+		if (vflag && !aflag) {
+			if (multiple++)
+				printf("\n");
+			fprintf(stdout, "*** No %s quotas on %s (%s)\n",
+			    qfextension[type], fs->fs_file, fs->fs_spec);
+		}
+		return(0);
 	}
 	if (multiple++)
 		printf("\n");
 	if (vflag)
 		fprintf(stdout, "*** Report for %s quotas on %s (%s)\n",
 		    qfextension[type], fs->fs_file, fs->fs_spec);
-	if ((qf = fopen(qfpathname, "r")) == NULL) {
-		warn("%s", qfpathname);
-		return (1);
-	}
-	for (id = 0; ; id++) {
-		fread(&dqbuf, sizeof(struct dqblk), 1, qf);
-		if (feof(qf))
+	printf("%*s           Block  limits                    File  limits\n",
+		max(UT_NAMESIZE,10), " ");
+	printf("User%*s  used   soft   hard  grace     used    soft    hard  grace\n",
+		max(UT_NAMESIZE,10), " ");
+	maxid = quota_maxid(qf);
+	for (id = 0; id <= maxid; id++) {
+		if (quota_read(qf, &dqbuf, id) != 0)
 			break;
 		if (dqbuf.dqb_curinodes == 0 && dqbuf.dqb_curblocks == 0)
 			continue;
 		if ((fup = lookup(id, type)) == 0)
 			fup = addid(id, type, (char *)0);
-		fup->fu_dqblk = dqbuf;
-	}
-	fclose(qf);
-	printf("%*s                 Block  limits                    File  limits\n",
-		max(UT_NAMESIZE,10), " ");
-	printf("%s%*s   used     soft     hard  grace     used    soft    hard  grace\n",
-		type == USRQUOTA ? "User " : "Group", max(UT_NAMESIZE,10), " ");
-	for (id = 0; id <= highid[type]; id++) {
-		fup = lookup(id, type);
-		if (fup == 0)
-			continue;
-		if (fup->fu_dqblk.dqb_curinodes == 0 &&
-		    fup->fu_dqblk.dqb_curblocks == 0)
-			continue;
 		printf("%-*s ", max(UT_NAMESIZE,10), fup->fu_name);
-		printf("%c%c %8lu %8lu %8lu %6s",
-			fup->fu_dqblk.dqb_bsoftlimit &&
-			    fup->fu_dqblk.dqb_curblocks >=
-			    fup->fu_dqblk.dqb_bsoftlimit ? '+' : '-',
-			fup->fu_dqblk.dqb_isoftlimit &&
-			    fup->fu_dqblk.dqb_curinodes >=
-			    fup->fu_dqblk.dqb_isoftlimit ? '+' : '-',
-			(u_long)(dbtokb(fup->fu_dqblk.dqb_curblocks)),
-			(u_long)(dbtokb(fup->fu_dqblk.dqb_bsoftlimit)),
-			(u_long)(dbtokb(fup->fu_dqblk.dqb_bhardlimit)),
-			fup->fu_dqblk.dqb_bsoftlimit &&
-			    fup->fu_dqblk.dqb_curblocks >=
-			    fup->fu_dqblk.dqb_bsoftlimit ?
-			    timeprt(fup->fu_dqblk.dqb_btime) : "-");
-		printf("  %7lu %7lu %7lu %6s\n",
-			(u_long)fup->fu_dqblk.dqb_curinodes,
-			(u_long)fup->fu_dqblk.dqb_isoftlimit,
-			(u_long)fup->fu_dqblk.dqb_ihardlimit,
-			fup->fu_dqblk.dqb_isoftlimit &&
-			    fup->fu_dqblk.dqb_curinodes >=
-			    fup->fu_dqblk.dqb_isoftlimit ?
-			    timeprt(fup->fu_dqblk.dqb_itime) : "-");
-		fup->fu_dqblk = zerodqblk;
+		printf("%c%c", 
+			dqbuf.dqb_bsoftlimit &&
+			    dqbuf.dqb_curblocks >=
+			    dqbuf.dqb_bsoftlimit ? '+' : '-',
+			dqbuf.dqb_isoftlimit &&
+			    dqbuf.dqb_curinodes >=
+			    dqbuf.dqb_isoftlimit ? '+' : '-');
+		prthumanval(dqbuf.dqb_curblocks);
+		prthumanval(dqbuf.dqb_bsoftlimit);
+		prthumanval(dqbuf.dqb_bhardlimit);
+		printf(" %6s",
+			dqbuf.dqb_bsoftlimit &&
+			    dqbuf.dqb_curblocks >=
+			    dqbuf.dqb_bsoftlimit ?
+			    timeprt(dqbuf.dqb_btime) : "-");
+		printf("  %7llu %7llu %7llu %6s\n",
+			dqbuf.dqb_curinodes,
+			dqbuf.dqb_isoftlimit,
+			dqbuf.dqb_ihardlimit,
+			dqbuf.dqb_isoftlimit &&
+			    dqbuf.dqb_curinodes >=
+			    dqbuf.dqb_isoftlimit ?
+			    timeprt(dqbuf.dqb_itime) : "-");
 	}
 	return (0);
 }
 
+static void
+prthumanval(int64_t blocks)
+{
+	char buf[7];
+	int flags;
+
+	if (!hflag) {
+		printf("%7llu", dbtokb(blocks));
+		return;
+	}
+	flags = HN_NOSPACE | HN_DECIMAL;
+	if (blocks != 0)
+		flags |= HN_B;
+	humanize_number(buf, sizeof(buf) - (blocks < 0 ? 0 : 1),
+	    dbtob(blocks), "", HN_AUTOSCALE, flags);
+	(void)printf("%7s", buf);
+}
+
 /*
  * Check to see if target appears in list of size cnt.
  */
@@ -279,55 +288,6 @@ oneof(char *target, char *list[], int cnt)
 	return (-1);
 }
 
-/*
- * Check to see if a particular quota is to be enabled.
- */
-int
-hasquota(struct fstab *fs, int type, char **qfnamep)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s",
-		    qfextension[USRQUOTA], qfname);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], qfname);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    qfname, qfextension[type]);
-		*qfnamep = buf;
-	}
-	if (statfs(fs->fs_file, &sfb) != 0) {
-		warn("cannot statfs mount point %s", fs->fs_file);
-		return (0);
-	}
-	if (strcmp(fs->fs_file, sfb.f_mntonname)) {
-		warnx("%s not mounted for %s quotas", fs->fs_file,
-		  type == USRQUOTA ? "user" : "group");
-		return (0);
-	}
-	return (1);
-}
-
 /*
  * Routines to manage the file usage table.
  *

From 88c01c773986ce45a687eea8837a888087fb3e38 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 20 Oct 2009 16:46:39 +0000
Subject: [PATCH 021/532] Update to work with quotafile functions in -libutil.

---
 usr.sbin/repquota/Makefile | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/usr.sbin/repquota/Makefile b/usr.sbin/repquota/Makefile
index 087e890ed35..9da22c2fc0f 100644
--- a/usr.sbin/repquota/Makefile
+++ b/usr.sbin/repquota/Makefile
@@ -3,6 +3,8 @@
 
 PROG=	repquota
 MAN=	repquota.8
+DPADD=  ${LIBUTIL}
+LDADD=  -lutil
 
 WARNS?=	4
 

From 4e717eeafc9593cbed1da2216e1bff6241bb25b6 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 20 Oct 2009 17:50:27 +0000
Subject: [PATCH 022/532] Claen up format when doing non-humanized output.

---
 usr.sbin/repquota/repquota.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/usr.sbin/repquota/repquota.c b/usr.sbin/repquota/repquota.c
index 276a78dc65e..aa822df09d2 100644
--- a/usr.sbin/repquota/repquota.c
+++ b/usr.sbin/repquota/repquota.c
@@ -208,6 +208,7 @@ repquota(struct fstab *fs, int type)
 				printf("\n");
 			fprintf(stdout, "*** No %s quotas on %s (%s)\n",
 			    qfextension[type], fs->fs_file, fs->fs_spec);
+			return(1);
 		}
 		return(0);
 	}
@@ -221,7 +222,7 @@ repquota(struct fstab *fs, int type)
 	printf("User%*s  used   soft   hard  grace     used    soft    hard  grace\n",
 		max(UT_NAMESIZE,10), " ");
 	maxid = quota_maxid(qf);
-	for (id = 0; id <= maxid; id++) {
+	for (id = 0; id < maxid; id++) {
 		if (quota_read(qf, &dqbuf, id) != 0)
 			break;
 		if (dqbuf.dqb_curinodes == 0 && dqbuf.dqb_curblocks == 0)
@@ -253,6 +254,7 @@ repquota(struct fstab *fs, int type)
 			    dqbuf.dqb_isoftlimit ?
 			    timeprt(dqbuf.dqb_itime) : "-");
 	}
+	quota_close(qf);
 	return (0);
 }
 
@@ -263,7 +265,7 @@ prthumanval(int64_t blocks)
 	int flags;
 
 	if (!hflag) {
-		printf("%7llu", dbtokb(blocks));
+		printf(" %6llu", dbtokb(blocks));
 		return;
 	}
 	flags = HN_NOSPACE | HN_DECIMAL;

From 37ed8e48db6cedb3822539f27e09f1d23d45ef8d Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Wed, 21 Oct 2009 04:35:09 +0000
Subject: [PATCH 023/532] Update to work with quotafile functions in -libutil.

---
 usr.bin/quota/quota.c | 231 +++++++++++++++---------------------------
 1 file changed, 82 insertions(+), 149 deletions(-)

diff --git a/usr.bin/quota/quota.c b/usr.bin/quota/quota.c
index f504e7c3159..24dcb494230 100644
--- a/usr.bin/quota/quota.c
+++ b/usr.bin/quota/quota.c
@@ -87,7 +87,7 @@ struct quotause {
 	char	fsname[MAXPATHLEN + 1];
 };
 
-static char *timeprt(time_t seconds);
+static char *timeprt(int64_t seconds, int *needfree);
 static struct quotause *getprivs(long id, int quotatype);
 static void usage(void);
 static int showuid(u_long uid);
@@ -97,7 +97,6 @@ static int showgrpname(char *name);
 static int showquotas(int type, u_long id, const char *name);
 static void showrawquotas(int type, u_long id, struct quotause *qup);
 static void heading(int type, u_long id, const char *name, const char *tag);
-static int ufshasquota(struct fstab *fs, int type, char **qfnamep);
 static int getufsquota(struct fstab *fs, struct quotause *qup, long id,
 	int quotatype);
 static int getnfsquota(struct statfs *fst, struct quotause *qup, long id,
@@ -117,8 +116,7 @@ int
 main(int argc, char *argv[])
 {
 	int ngroups; 
-	long ngroups_max;
-	gid_t mygid, *gidset;
+	gid_t mygid, gidset[NGROUPS];
 	int i, ch, gflag = 0, uflag = 0, errflag = 0;
 
 	while ((ch = getopt(argc, argv, "f:ghlrquv")) != -1) {
@@ -160,18 +158,13 @@ main(int argc, char *argv[])
 			errflag += showuid(getuid());
 		if (gflag) {
 			mygid = getgid();
-			ngroups_max = sysconf(_SC_NGROUPS_MAX) + 1;
-			if ((gidset = malloc(sizeof(gid_t) * ngroups_max))
-			    == NULL)
-				err(1, "malloc");
-			ngroups = getgroups(ngroups_max, gidset);
+			ngroups = getgroups(NGROUPS, gidset);
 			if (ngroups < 0)
 				err(1, "getgroups");
 			errflag += showgid(mygid);
 			for (i = 0; i < ngroups; i++)
 				if (gidset[i] != mygid)
 					errflag += showgid(gidset[i]);
-			free(gidset);
 		}
 		return(errflag);
 	}
@@ -271,7 +264,7 @@ showgrpname(char *name)
 }
 
 static void
-prthumanval(int len, int64_t bytes)
+prthumanval(int len, u_int64_t bytes)
 {
 	char buf[len + 1];
 
@@ -288,7 +281,8 @@ showquotas(int type, u_long id, const char *name)
 	struct quotause *quplist;
 	const char *msgi, *msgb;
 	const char *nam;
-	char *bgrace = NULL, *igrace = NULL;
+	char *bgrace, *igrace;
+	int bfree, ifree;
 	int lines = 0, overquota = 0;
 	static time_t now;
 
@@ -344,52 +338,46 @@ showquotas(int type, u_long id, const char *name)
 				printf("\t%s %s\n", msgb, qup->fsname);
 			continue;
 		}
-		if (vflag ||
-		    qup->dqblk.dqb_curblocks ||
-		    qup->dqblk.dqb_curinodes) {
-			if (lines++ == 0)
-				heading(type, id, name, "");
-			nam = qup->fsname;
-			if (strlen(qup->fsname) > 15) {
-				printf("%s\n", qup->fsname);
-				nam = "";
-			} 
-			printf("%15s", nam);
-			if (hflag) {
-				printf("   ");
-				prthumanval(4, dbtob(qup->dqblk.dqb_curblocks));
-				printf("%c  ", (msgb == (char *)0) ? ' ' : '*');
-				prthumanval(4, dbtob(qup->dqblk.dqb_bsoftlimit));
-				printf("   ");
-				prthumanval(4, dbtob(qup->dqblk.dqb_bhardlimit));
-			} else {
-				printf(" %7ju%c %6ju %7ju",
-				    (uintmax_t)(dbtob(qup->dqblk.dqb_curblocks)
-					/ 1024),
-				    (msgb == NULL) ? ' ' : '*',
-				    (uintmax_t)(dbtob(qup->dqblk.dqb_bsoftlimit)
-					/ 1024),
-				    (uintmax_t)(dbtob(qup->dqblk.dqb_bhardlimit)
-					/ 1024));
-			}
-			if (msgb != NULL)
-				bgrace = timeprt(qup->dqblk.dqb_btime);
-			if (msgi != NULL)
-				igrace = timeprt(qup->dqblk.dqb_itime);
-			printf(" %7s %7ju%c %6ju %7ju %7s\n",
-			    (msgb == NULL) ? "" : bgrace,
-			    (uintmax_t)qup->dqblk.dqb_curinodes,
-			    (msgi == NULL) ? ' ' : '*',
-			    (uintmax_t)qup->dqblk.dqb_isoftlimit,
-			    (uintmax_t)qup->dqblk.dqb_ihardlimit,
-			    (msgi == NULL) ? "" : igrace
-			);
-			if (msgb != NULL)
-				free(bgrace);
-			if (msgi != NULL)
-				free(igrace);
+		if (!vflag &&
+		    qup->dqblk.dqb_curblocks == 0 &&
+		    qup->dqblk.dqb_curinodes == 0)
 			continue;
+		if (lines++ == 0)
+			heading(type, id, name, "");
+		nam = qup->fsname;
+		if (strlen(qup->fsname) > 15) {
+			printf("%s\n", qup->fsname);
+			nam = "";
+		} 
+		printf("%-15s", nam);
+		if (hflag) {
+			prthumanval(7, dbtob(qup->dqblk.dqb_curblocks));
+			printf("%c", (msgb == (char *)0) ? ' ' : '*');
+			prthumanval(7, dbtob(qup->dqblk.dqb_bsoftlimit));
+			prthumanval(7, dbtob(qup->dqblk.dqb_bhardlimit));
+		} else {
+			printf(" %7ju%c %7ju %7ju",
+			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_curblocks,
+			    (msgb == (char *)0) ? ' ' : '*',
+			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_bsoftlimit,
+			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_bhardlimit);
 		}
+		if (msgb != NULL)
+			bgrace = timeprt(qup->dqblk.dqb_btime, &bfree);
+		if (msgi != NULL)
+			igrace = timeprt(qup->dqblk.dqb_itime, &ifree);
+		printf("%8s %6ju%c %6ju %6ju%8s\n"
+			, (msgb == (char *)0) ? "" : bgrace
+			, (uintmax_t)qup->dqblk.dqb_curinodes
+			, (msgi == (char *)0) ? ' ' : '*'
+			, (uintmax_t)qup->dqblk.dqb_isoftlimit
+			, (uintmax_t)qup->dqblk.dqb_ihardlimit
+			, (msgi == (char *)0) ? "" : igrace
+		);
+		if (msgb != NULL && bfree)
+			free(bgrace);
+		if (msgi != NULL && ifree)
+			free(igrace);
 	}
 	if (!qflag && !rflag && lines == 0)
 		heading(type, id, name, "none");
@@ -399,27 +387,37 @@ showquotas(int type, u_long id, const char *name)
 static void
 showrawquotas(int type, u_long id, struct quotause *qup)
 {
-	time_t tt;
+	time_t time;
+
 	printf("Raw %s quota information for id %lu on %s\n",
 	    type == USRQUOTA ? "user" : "group", id, qup->fsname);
-	printf("block hard limit:     %ju\n", (uintmax_t)qup->dqblk.dqb_bhardlimit);
-	printf("block soft limit:     %ju\n", (uintmax_t)qup->dqblk.dqb_bsoftlimit);
-	printf("current block count:  %ju\n", (uintmax_t)qup->dqblk.dqb_curblocks);
-	printf("i-node hard limit:    %ju\n", (uintmax_t)qup->dqblk.dqb_ihardlimit);
-	printf("i-node soft limit:    %ju\n", (uintmax_t)qup->dqblk.dqb_isoftlimit);
-	printf("current i-node count: %ju\n", (uintmax_t)qup->dqblk.dqb_curinodes);
-	printf("block grace time:     %jd", (intmax_t)qup->dqblk.dqb_btime);
+	printf("block hard limit:     %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_bhardlimit);
+	printf("block soft limit:     %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_bsoftlimit);
+	printf("current block count:  %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_curblocks);
+	printf("i-node hard limit:    %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_ihardlimit);
+	printf("i-node soft limit:    %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_isoftlimit);
+	printf("current i-node count: %ju\n",
+	    (uintmax_t)qup->dqblk.dqb_curinodes);
+	printf("block grace time:     %jd",
+	    (intmax_t)qup->dqblk.dqb_btime);
 	if (qup->dqblk.dqb_btime != 0) {
-		tt = qup->dqblk.dqb_btime;
-		printf(" %s", ctime(&tt));
-	} else
+		time = qup->dqblk.dqb_btime;
+		printf(" %s", ctime(&time));
+	} else {
 		printf("\n");
+	}
 	printf("i-node grace time:    %jd", (intmax_t)qup->dqblk.dqb_itime);
 	if (qup->dqblk.dqb_itime != 0) {
-		tt = qup->dqblk.dqb_itime;
-		printf(" %s", ctime(&tt));
-	} else
+		time = qup->dqblk.dqb_itime;
+		printf(" %s", ctime(&time));
+	} else {
 		printf("\n");
+	}
 }
 
 
@@ -430,7 +428,7 @@ heading(int type, u_long id, const char *name, const char *tag)
 	printf("Disk quotas for %s %s (%cid %lu): %s\n", qfextension[type],
 	    name, *qfextension[type], id, tag);
 	if (!qflag && tag[0] == '\0') {
-		printf("%15s %7s  %6s %7s %7s %7s  %6s %7s %7s\n"
+		printf("%-15s %7s %8s %7s %7s %6s %7s %6s%8s\n"
 			, "Filesystem"
 			, "usage"
 			, "quota"
@@ -448,7 +446,7 @@ heading(int type, u_long id, const char *name, const char *tag)
  * Calculate the grace period and return a printable string for it.
  */
 static char *
-timeprt(time_t seconds)
+timeprt(int64_t seconds, int *needfree)
 {
 	time_t hours, minutes;
 	char	*buf;
@@ -457,7 +455,8 @@ timeprt(time_t seconds)
 	if (now == 0)
 		time(&now);
 	if (now > seconds) {
-		return strdup("none");
+		*needfree = 0;
+		return ("none");
 	}
 	seconds -= now;
 	minutes = (seconds + 30) / 60;
@@ -465,16 +464,19 @@ timeprt(time_t seconds)
 	if (hours >= 36) {
 		if (asprintf(&buf, "%lddays", ((long)hours + 12) / 24) < 0)
 			errx(1, "asprintf failed in timeprt(1)");
+		*needfree = 1;
 		return (buf);
 	}
 	if (minutes >= 60) {
 		if (asprintf(&buf, "%2ld:%ld", (long)minutes / 60,
 		    (long)minutes % 60) < 0)
 			errx(1, "asprintf failed in timeprt(2)");
+		*needfree = 1;
 		return (buf);
 	}
 	if (asprintf(&buf, "%2ld", (long)minutes) < 0)
 		errx(1, "asprintf failed in timeprt(3)");
+	*needfree = 1;
 	return (buf);
 }
 
@@ -499,7 +501,7 @@ getprivs(long id, int quotatype)
 	if (nfst == 0)
 		errx(2, "no filesystems mounted!");
 	setfsent();
-	for (i=0; i<nfst; i++) {
+	for (i = 0; i < nfst; i++) {
 		if (qup == NULL) {
 			if ((qup = (struct quotause *)malloc(sizeof *qup))
 			    == NULL)
@@ -548,87 +550,18 @@ getprivs(long id, int quotatype)
 }
 
 /*
- * Check to see if a particular quota is to be enabled.
+ * Check to see if a particular quota is available.
  */
-static int
-ufshasquota(struct fstab *fs, int type, char **qfnamep)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s",
-		    qfextension[USRQUOTA], qfname);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], qfname);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    qfname, qfextension[type]);
-		*qfnamep = buf;
-	}
-	if (statfs(fs->fs_file, &sfb) != 0) {
-		warn("cannot statfs mount point %s", fs->fs_file);
-		return (0);
-	}
-	if (strcmp(fs->fs_file, sfb.f_mntonname)) {
-		warnx("%s not mounted for %s quotas", fs->fs_file,
-		    type == USRQUOTA ? "user" : "group");
-		return (0);
-	}
-	return (1);
-}
-
 static int
 getufsquota(struct fstab *fs, struct quotause *qup, long id, int quotatype)
 {
-	char *qfpathname;
-	int fd, qcmd;
+	struct quotafile *qf;
 
-	qcmd = QCMD(Q_GETQUOTA, quotatype);
-	if (!ufshasquota(fs, quotatype, &qfpathname))
+	if ((qf = quota_open(fs, quotatype, O_RDONLY)) == NULL)
 		return (0);
-
-	if (quotactl(fs->fs_file, qcmd, id, (char *)&qup->dqblk) != 0) {
-		if ((fd = open(qfpathname, O_RDONLY)) < 0) {
-			warn("%s", qfpathname);
-			return (0);
-		}
-		(void) lseek(fd, (off_t)(id * sizeof(struct dqblk)), L_SET);
-		switch (read(fd, &qup->dqblk, sizeof(struct dqblk))) {
-		case 0:				/* EOF */
-			/*
-			 * Convert implicit 0 quota (EOF)
-			 * into an explicit one (zero'ed dqblk)
-			 */
-			bzero((caddr_t)&qup->dqblk, sizeof(struct dqblk));
-			break;
-		case sizeof(struct dqblk):	/* OK */
-			break;
-		default:		/* ERROR */
-			warn("read error: %s", qfpathname);
-			close(fd);
-			return (0);
-		}
-		close(fd);
-	}
+	if (quota_read(qf, &qup->dqblk, id) != 0)
+		return (0);
+	quota_close(qf);
 	return (1);
 }
 

From 708a2897f1b3c36dacde8cf534288c05373f7dfd Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Mon, 16 Nov 2009 18:50:51 +0000
Subject: [PATCH 024/532] Convert quotaon/quotaoff to use the new quotafile
 functions.

---
 usr.sbin/quotaon/Makefile  |   3 +
 usr.sbin/quotaon/quotaon.c | 122 ++++++++-----------------------------
 2 files changed, 29 insertions(+), 96 deletions(-)

diff --git a/usr.sbin/quotaon/Makefile b/usr.sbin/quotaon/Makefile
index 3f88bd9da33..0a07f065194 100644
--- a/usr.sbin/quotaon/Makefile
+++ b/usr.sbin/quotaon/Makefile
@@ -8,4 +8,7 @@ MLINKS=	quotaon.8 quotaoff.8
 
 WARNS?=	4
 
+DPADD=  ${LIBUTIL}
+LDADD=  -lutil
+
 .include <bsd.prog.mk>
diff --git a/usr.sbin/quotaon/quotaon.c b/usr.sbin/quotaon/quotaon.c
index 314f596bc01..0f8ef98f7cd 100644
--- a/usr.sbin/quotaon/quotaon.c
+++ b/usr.sbin/quotaon/quotaon.c
@@ -53,6 +53,7 @@ __FBSDID("$FreeBSD$");
 #include <ufs/ufs/quota.h>
 #include <err.h>
 #include <fstab.h>
+#include <libutil.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
@@ -66,17 +67,15 @@ int	gflag;		/* operate on group quotas */
 int	uflag;		/* operate on user quotas */
 int	vflag;		/* verbose */
 
-int hasquota(struct fstab *, int, char **);
 int oneof(char *, char *[], int);
-int quotaonoff(struct fstab *fs, int, int, char *);
-int readonly(struct fstab *);
+int quotaonoff(struct fstab *fs, int, int);
 static void usage(void);
 
 int
 main(int argc, char **argv)
 {
 	struct fstab *fs;
-	char *qfnp, *whoami;
+	char *whoami;
 	long argnum, done = 0;
 	int ch, i, offmode = 0, errs = 0;
 
@@ -119,19 +118,19 @@ main(int argc, char **argv)
 		    strcmp(fs->fs_type, FSTAB_RW))
 			continue;
 		if (aflag) {
-			if (gflag && hasquota(fs, GRPQUOTA, &qfnp))
-				errs += quotaonoff(fs, offmode, GRPQUOTA, qfnp);
-			if (uflag && hasquota(fs, USRQUOTA, &qfnp))
-				errs += quotaonoff(fs, offmode, USRQUOTA, qfnp);
+			if (gflag)
+				errs += quotaonoff(fs, offmode, GRPQUOTA);
+			if (uflag)
+				errs += quotaonoff(fs, offmode, USRQUOTA);
 			continue;
 		}
 		if ((argnum = oneof(fs->fs_file, argv, argc)) >= 0 ||
 		    (argnum = oneof(fs->fs_spec, argv, argc)) >= 0) {
 			done |= 1 << argnum;
-			if (gflag && hasquota(fs, GRPQUOTA, &qfnp))
-				errs += quotaonoff(fs, offmode, GRPQUOTA, qfnp);
-			if (uflag && hasquota(fs, USRQUOTA, &qfnp))
-				errs += quotaonoff(fs, offmode, USRQUOTA, qfnp);
+			if (gflag)
+				errs += quotaonoff(fs, offmode, GRPQUOTA);
+			if (uflag)
+				errs += quotaonoff(fs, offmode, USRQUOTA);
 		}
 	}
 	endfsent();
@@ -154,32 +153,33 @@ usage(void)
 }
 
 int
-quotaonoff(fs, offmode, type, qfpathname)
-	register struct fstab *fs;
+quotaonoff(fs, offmode, type)
+	struct fstab *fs;
 	int offmode, type;
-	char *qfpathname;
 {
+	struct quotafile *qf;
 
-	if (strcmp(fs->fs_file, "/") && readonly(fs))
-		return (1);
+	if ((qf = quota_open(fs, type, O_RDONLY)) == NULL)
+		return (0);
 	if (offmode) {
-		if (quotactl(fs->fs_file, QCMD(Q_QUOTAOFF, type), 0, 0) < 0) {
-			warn("%s", fs->fs_file);
+		if (quota_off(qf) != 0) {
+			warn("%s", quota_fsname(qf));
 			return (1);
 		}
 		if (vflag)
-			printf("%s: quotas turned off\n", fs->fs_file);
-		return (0);
+			printf("%s: quotas turned off\n", quota_fsname(qf));
+		quota_close(qf);
+		return(0);
 	}
-	if (quotactl(fs->fs_file, QCMD(Q_QUOTAON, type), 0, qfpathname) < 0) {
-		warnx("using %s on", qfpathname);
-		warn("%s", fs->fs_file);
+	if (quota_on(qf) != 0) {
+		warn("using %s on %s", quota_qfname(qf), quota_fsname(qf));
 		return (1);
 	}
 	if (vflag)
 		printf("%s: %s quotas turned on with data file %s\n", 
-		    fs->fs_file, qfextension[type], qfpathname);
-	return (0);
+		    quota_fsname(qf), qfextension[type], quota_qfname(qf));
+	quota_close(qf);
+	return(0);
 }
 
 /*
@@ -195,73 +195,3 @@ oneof(char *target, char *list[], int cnt)
 			return (i);
 	return (-1);
 }
-
-/*
- * Check to see if a particular quota is to be enabled.
- */
-int
-hasquota(struct fstab *fs, int type, char **qfnamep)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s", 
-		    qfextension[USRQUOTA], qfname);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], qfname);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    qfname, qfextension[type]);
-		*qfnamep = buf;
-	}
-	if (statfs(fs->fs_file, &sfb) != 0) {
-		warn("cannot statfs mount point %s", fs->fs_file);
-		return (0);
-	}
-	if (strcmp(fs->fs_file, sfb.f_mntonname)) {
-		warnx("%s not mounted for %s quotas", fs->fs_file,
-		    type == USRQUOTA ? "user" : "group");
-		return (0);
-	}
-	return (1);
-}
-
-/*
- * Verify filesystem is mounted and not readonly.
- */
-int
-readonly(struct fstab *fs)
-{
-	struct statfs fsbuf;
-
-	if (statfs(fs->fs_file, &fsbuf) < 0 ||
-	    strcmp(fsbuf.f_mntonname, fs->fs_file) ||
-	    strcmp(fsbuf.f_mntfromname, fs->fs_spec)) {
-		printf("%s: not mounted\n", fs->fs_file);
-		return (1);
-	}
-	if (fsbuf.f_flags & MNT_RDONLY) {
-		printf("%s: mounted read-only\n", fs->fs_file);
-		return (1);
-	}
-	return (0);
-}

From e525d16a80319276894d6737ffb7c4b5a8fda142 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Mon, 16 Nov 2009 18:59:04 +0000
Subject: [PATCH 025/532] Add and document new quoat_on and quota_off
 functions.

---
 lib/libutil/libutil.h   |   4 +-
 lib/libutil/quotafile.3 | 149 +++++++++++++++++++++++++---------------
 lib/libutil/quotafile.c |  18 ++++-
 3 files changed, 113 insertions(+), 58 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index b5ff6fd9e07..3a05d9b56f6 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -144,6 +144,9 @@ int pidfile_remove(struct pidfh *pfh);
 struct quotafile;
 struct fstab;
 struct quotafile *quota_open(struct fstab *, int, int);
+void quota_close(struct quotafile *);
+int quota_on(struct quotafile *);
+int quota_off(struct quotafile *);
 const char *quota_fsname(const struct quotafile *);
 const char *quota_qfname(const struct quotafile *);
 int quota_maxid(struct quotafile *);
@@ -151,7 +154,6 @@ int quota_check_path(const struct quotafile *, const char *path);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write_limits(struct quotafile *, struct dqblk *, int);
 int quota_write_usage(struct quotafile *, struct dqblk *, int);
-void quota_close(struct quotafile *);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index d0c27d730b8..d2134c50530 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -30,14 +30,16 @@
 .Os
 .Sh NAME
 .Nm quota_open
+.Nm quota_close
+.Nm quota_on
+.Nm quota_off
+.Nm quota_read
+.Nm quota_write_limits
+.Nm quota_write_usage
 .Nm quota_fsname
 .Nm quota_qfname
 .Nm quota_maxid
 .Nm quota_check_path
-.Nm quota_read
-.Nm quota_write_limits
-.Nm quota_write_usage
-.Nm quota_close
 .Nd "Manipulate quotas"
 .Sh LIBRARY
 .Lb libutil
@@ -50,6 +52,18 @@
 .In libutil.h
 .Ft "struct quotafile *"
 .Fn quota_open "struct fstab *fs" "int quotatype" "int openflags"
+.Ft int
+.Fn quota_close "struct quotafile *qf"
+.Ft int
+.Fn quota_on "const struct quotafile *qf"
+.Ft int
+.Fn quota_off "const struct quotafile *qf"
+.Ft int
+.Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
+.Ft int
+.Fn quota_write_limits "struct quotafile *qf" "struct dqblk *dqb" "int id"
+.Ft int
+.Fn quota_write_usage "struct quotafile *qf" "struct dqblk *dqb" "int id"
 .Ft "const char *"
 .Fn quota_fsname "const struct quotafile *qf"
 .Ft "const char *"
@@ -58,14 +72,6 @@
 .Fn quota_maxid "const struct quotafile *qf"
 .Ft int
 .Fn quota_check_path "const struct quotafile *qf" "const char *path"
-.Ft int
-.Fn quota_read "struct quotafile *qf" "struct dqblk *dqb" "int id"
-.Ft int
-.Fn quota_write_limits "struct quotafile *qf" "struct dqblk *dqb" "int id"
-.Ft int
-.Fn quota_write_usage "struct quotafile *qf" "struct dqblk *dqb" "int id"
-.Ft int
-.Fn quota_close "struct quotafile *qf"
 .Sh DESCRIPTION
 These functions are designed to simplify access to filesystem quotas.
 If quotas are active on a filesystem,
@@ -103,6 +109,81 @@ flag should be specified if a new quota file of the requested type
 should be created if it does not already exist.
 .Pp
 The
+.Fn quota_close
+function closes any open file descriptors and frees any storage
+associated with the filesystem and quota type referenced by
+.Va qf .
+.Pp
+The
+.Fn quota_on
+function enables quotas for the filesystem associated with its
+.Va qf
+argument which may have been opened
+.Dv O_RDONLY
+or
+.Dv O_RDWR .
+The
+.Fn quota_on
+function returns 0 if successful;
+otherwise the value\~-1 is returned and the global variable
+.Va errno
+is set to indicate the error, see
+.Xr quotactl 2
+for the possible errors.
+.Pp
+The
+.Fn quota_off
+function disables quotas for the filesystem associated with its
+.Va qf
+argument which may have been opened
+.Dv O_RDONLY
+or
+.Dv O_RDWR .
+The
+.Fn quota_off
+function returns 0 if successful;
+otherwise the value\~-1 is returned and the global variable
+.Va errno
+is set to indicate the error, see
+.Xr quotactl 2
+for the possible errors.
+.Pp
+The
+.Fn quota_read
+function reads the quota from the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+into the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
+.Fn quota_write_limits
+function updates the limit fields (but not the usage fields)
+for the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+from the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
+.Fn quota_write_usage
+function updates the usage fields (but not the limit fields)
+for the filesystem and quota type referenced by
+.Va qf
+for the user (or group) specified by
+.Va id
+from the
+.Vt dqblk
+quota structure pointed to by
+.Va dqb .
+.Pp
+The
 .Fn quota_fsname
 function returns a pointer to a buffer containing the path to the root
 of the file system that corresponds to its
@@ -138,48 +219,6 @@ If the
 argument refers to a symbolic link,
 .Fn quota_check_path
 will follow it.
-.Pp
-The
-.Fn quota_read
-function reads the quota from the filesystem and quota type referenced
-by
-.Va qf
-for the user (or group) specified by
-.Va id
-into the
-.Vt dqblk
-quota structure pointed to by
-.Va dqb .
-.Pp
-The
-.Fn quota_write_limits
-function updates the limit fields (but not the usage fields)
-for the filesystem and quota type referenced by
-.Va qf
-for the user (or group) specified by
-.Va id
-from the
-.Vt dqblk
-quota structure pointed to by
-.Va dqb .
-.Pp
-The
-.Fn quota_write_usage
-function updates the usage fields (but not the limit fields)
-for the filesystem and quota type referenced by
-.Va qf
-for the user (or group) specified by
-.Va id
-from the
-.Vt dqblk
-quota structure pointed to by
-.Va dqb .
-.Pp
-The
-.Fn quota_close
-function closes any open file descriptors and frees any storage
-associated with the filesystem and quota type referenced by
-.Va qf .
 .Sh IMPLEMENTATION NOTES
 If the underlying quota file is in the old 32-bit format, limit and
 usage values written to the quota file will be clipped to 32 bits.
@@ -230,4 +269,4 @@ The
 functions and this manual page were written by
 .An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org 
 and
-.An Marshall Kirk McKusick .
+.An Marshall Kirk McKusick Aq mckusick@mckusick.com .
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 4c63a17b778..9017c4ecd80 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -61,7 +61,6 @@ static const char *qfextension[] = INITQFNAMES;
 
 /*
  * Check to see if a particular quota is to be enabled.
- * XXX merge into quota_open
  */
 static int
 hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
@@ -73,7 +72,6 @@ hasquota(struct fstab *fs, int type, char *qfnamep, int qfbufsize)
 	static char initname, usrname[100], grpname[100];
 
 	/*
-	 * XXX
 	 * 1) we only need one of these
 	 * 2) fstab may specify a different filename
 	 */
@@ -207,6 +205,22 @@ quota_close(struct quotafile *qf)
 	free(qf);
 }
 
+int
+quota_on(struct quotafile *qf)
+{
+	int qcmd;
+
+	qcmd = QCMD(Q_QUOTAON, qf->quotatype);
+	return (quotactl(qf->fsname, qcmd, 0, qf->qfname));
+}
+
+int
+quota_off(struct quotafile *qf)
+{
+
+	return (quotactl(qf->fsname, QCMD(Q_QUOTAOFF, qf->quotatype), 0, 0));
+}
+
 const char *
 quota_fsname(const struct quotafile *qf)
 {

From c3616249ab23a2cfd4caa843df95aeb9a17ef5b2 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Sun, 27 Dec 2009 06:26:04 +0000
Subject: [PATCH 026/532] Minor bugs turned up during conversion of quotacheck.

---
 lib/libutil/quotafile.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 9017c4ecd80..652d95a290f 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -118,6 +118,8 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 	struct stat st;
 	int qcmd, serrno;
 
+	if (strcmp(fs->fs_vfstype, "ufs"))
+		return (NULL);
 	if ((qf = calloc(1, sizeof(*qf))) == NULL)
 		return (NULL);
 	qf->fd = -1;
@@ -176,8 +178,11 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 	dqh.dqh_version = htobe32(Q_DQHDR64_VERSION);
 	dqh.dqh_hdrlen = htobe32(sizeof(struct dqhdr64));
 	dqh.dqh_reclen = htobe32(sizeof(struct dqblk64));
-	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh))
+	if (write(qf->fd, &dqh, sizeof(dqh)) != sizeof(dqh)) {
+		/* it was one we created ourselves */
+		unlink(qf->qfname);
 		goto error;
+	}
 	grp = getgrnam(QUOTAGROUP);
 	fchown(qf->fd, 0, grp ? grp->gr_gid : 0);
 	fchmod(qf->fd, 0640);
@@ -185,12 +190,8 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 error:
 	serrno = errno;
 	/* did we have an open file? */
-	if (qf->fd != -1) {
-		/* was it one we created ourselves? */
-		if ((openflags & O_CREAT) == O_CREAT)
-			unlink(qf->qfname);
+	if (qf->fd != -1)
 		close(qf->fd);
-	}
 	free(qf);
 	errno = serrno;
 	return (NULL);

From 3af26d4abba3ad1858920eedf0710663499167f1 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Sun, 27 Dec 2009 06:28:01 +0000
Subject: [PATCH 027/532] Convert quotacheck to use new quotafile functions in
 libutil. Still to come, conversion between 64-bit and 32-bit quotafile
 formats.

---
 sbin/quotacheck/Makefile     |   2 +
 sbin/quotacheck/preen.c      |  77 ++++++-----
 sbin/quotacheck/quotacheck.c | 254 +++++++++--------------------------
 sbin/quotacheck/quotacheck.h |  12 +-
 4 files changed, 109 insertions(+), 236 deletions(-)

diff --git a/sbin/quotacheck/Makefile b/sbin/quotacheck/Makefile
index 5184bf261f1..51a88b7681c 100644
--- a/sbin/quotacheck/Makefile
+++ b/sbin/quotacheck/Makefile
@@ -5,6 +5,8 @@ PROG=	quotacheck
 SRCS=	quotacheck.c preen.c fsutil.c utilities.c
 WARNS?=	2
 MAN=	quotacheck.8
+DPADD=  ${LIBUTIL}
+LDADD=  -lutil
 
 .PATH:	${.CURDIR}/../fsck ${.CURDIR}/../fsck_ffs
 
diff --git a/sbin/quotacheck/preen.c b/sbin/quotacheck/preen.c
index a6c4169ee55..1c285ccae9c 100644
--- a/sbin/quotacheck/preen.c
+++ b/sbin/quotacheck/preen.c
@@ -45,9 +45,13 @@ __RCSID("$NetBSD: preen.c,v 1.18 1998/07/26 20:02:36 mycroft Exp $");
 #include <sys/wait.h>
 #include <sys/queue.h>
 
+#include <ufs/ufs/quota.h>
+
 #include <err.h>
 #include <ctype.h>
+#include <fcntl.h>
 #include <fstab.h>
+#include <libutil.h>
 #include <string.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -58,9 +62,9 @@ __RCSID("$NetBSD: preen.c,v 1.18 1998/07/26 20:02:36 mycroft Exp $");
 struct partentry {
 	TAILQ_ENTRY(partentry)	 p_entries;
 	char			*p_devname;	/* device name */
-	char			*p_mntpt;	/* mount point */
-	char			*p_type;	/* file system type */
-	struct quotaname	*p_quota;	/* quota file info ptr */
+	const char		*p_mntpt;	/* mount point */
+	struct quotafile	*p_qfu;		/* user quota file info ptr */
+	struct quotafile	*p_qfg;		/* group quota file info */
 };
 
 TAILQ_HEAD(part, partentry) badh;
@@ -75,21 +79,19 @@ struct diskentry {
 TAILQ_HEAD(disk, diskentry) diskh;
 
 static struct diskentry *finddisk(const char *);
-static void addpart(const char *, const char *, const char *,
-    struct quotaname *);
+static void addpart(struct fstab *, struct quotafile *, struct quotafile *);
 static int startdisk(struct diskentry *);
 extern void *emalloc(size_t);
 extern char *estrdup(const char *);
 
 int
-checkfstab(void)
+checkfstab(int uflag, int gflag)
 {
 	struct fstab *fs;
 	struct diskentry *d, *nextdisk;
 	struct partentry *p;
 	int ret, pid, retcode, passno, sumstatus, status, nextpass;
-	char *name;
-	struct quotaname *qnp;
+	struct quotafile *qfu, *qfg;
 
 	TAILQ_INIT(&badh);
 	TAILQ_INIT(&diskh);
@@ -104,30 +106,32 @@ checkfstab(void)
 			return (8);
 		}
 		while ((fs = getfsent()) != 0) {
-			name = fs->fs_spec;
 			if (fs->fs_passno > passno && fs->fs_passno < nextpass)
 				nextpass = fs->fs_passno;
 
 			if (passno != fs->fs_passno)
 				continue;
 
-			if ((qnp = needchk(fs)) == NULL)
+			qfu = NULL;
+			if (uflag)
+				qfu = quota_open(fs, USRQUOTA, O_CREAT|O_RDWR);
+			qfg = NULL;
+			if (gflag)
+				qfg = quota_open(fs, GRPQUOTA, O_CREAT|O_RDWR);
+			if (qfu == NULL && qfg == NULL)
 				continue;
 
 			if (passno == 1) {
-				sumstatus = chkquota(name, fs->fs_file, qnp);
-
+				sumstatus = chkquota(fs->fs_spec, qfu, qfg);
+				if (qfu)
+					quota_close(qfu);
+				if (qfg)
+					quota_close(qfg);
 				if (sumstatus)
 					return (sumstatus);
 				continue;
 			}
-			if (name == NULL) {
-				(void) fprintf(stderr,
-				    "BAD DISK NAME %s\n", fs->fs_spec);
-				sumstatus |= 8;
-				continue;
-			}
-			addpart(fs->fs_vfstype, name, fs->fs_file, qnp);
+			addpart(fs, qfu, qfg);
 		}
 
 		if (passno == 1)
@@ -157,8 +161,8 @@ checkfstab(void)
 
 			if (WIFSIGNALED(status)) {
 				(void) fprintf(stderr,
-				    "%s: %s (%s): EXITED WITH SIGNAL %d\n",
-				    p->p_type, p->p_devname, p->p_mntpt,
+				    "%s: (%s): EXITED WITH SIGNAL %d\n",
+				    p->p_devname, p->p_mntpt,
 				    WTERMSIG(status));
 				retcode = 8;
 			}
@@ -169,8 +173,11 @@ checkfstab(void)
 				TAILQ_INSERT_TAIL(&badh, p, p_entries);
 				sumstatus |= retcode;
 			} else {
-				free(p->p_type);
 				free(p->p_devname);
+				if (p->p_qfu)
+					quota_close(p->p_qfu);
+				if (p->p_qfg)
+					quota_close(p->p_qfg);
 				free(p);
 			}
 			d->d_pid = 0;
@@ -196,8 +203,8 @@ checkfstab(void)
 
 		for (; p; p = TAILQ_NEXT(p, p_entries))
 			(void) fprintf(stderr,
-			    "%s: %s (%s)%s", p->p_type, p->p_devname,
-			    p->p_mntpt, TAILQ_NEXT(p, p_entries) ? ", " : "\n");
+			    "%s: (%s)%s", p->p_devname, p->p_mntpt,
+			    TAILQ_NEXT(p, p_entries) ? ", " : "\n");
 
 		return sumstatus;
 	}
@@ -242,23 +249,25 @@ finddisk(const char *name)
 }
 
 static void
-addpart(const char *type, const char *devname, const char *mntpt,
-    struct quotaname *qnp)
+addpart(struct fstab *fs, struct quotafile *qfu, struct quotafile *qfg)
 {
-	struct diskentry *d = finddisk(devname);
+	struct diskentry *d = finddisk(fs->fs_spec);
 	struct partentry *p;
 
 	TAILQ_FOREACH(p, &d->d_part, p_entries)
-		if (strcmp(p->p_devname, devname) == 0) {
-			warnx("%s in fstab more than once!\n", devname);
+		if (strcmp(p->p_devname, fs->fs_spec) == 0) {
+			warnx("%s in fstab more than once!\n", fs->fs_spec);
 			return;
 		}
 
 	p = emalloc(sizeof(*p));
-	p->p_devname = estrdup(devname);
-	p->p_mntpt = estrdup(mntpt);
-	p->p_type = estrdup(type);
-	p->p_quota = qnp;
+	p->p_devname = estrdup(blockcheck(fs->fs_spec));
+	if (qfu != NULL)
+		p->p_mntpt = quota_fsname(qfu);
+	else
+		p->p_mntpt = quota_fsname(qfg);
+	p->p_qfu = qfu;
+	p->p_qfg = qfg;
 
 	TAILQ_INSERT_TAIL(&d->d_part, p, p_entries);
 }
@@ -275,6 +284,6 @@ startdisk(struct diskentry *d)
 		return (8);
 	}
 	if (d->d_pid == 0)
-		exit(chkquota(p->p_devname, p->p_mntpt, p->p_quota));
+		exit(chkquota(p->p_devname, p->p_qfu, p->p_qfg));
 	return (0);
 }
diff --git a/sbin/quotacheck/quotacheck.c b/sbin/quotacheck/quotacheck.c
index 863932a12d2..026608a419b 100644
--- a/sbin/quotacheck/quotacheck.c
+++ b/sbin/quotacheck/quotacheck.c
@@ -61,6 +61,7 @@ __FBSDID("$FreeBSD$");
 #include <fcntl.h>
 #include <fstab.h>
 #include <grp.h>
+#include <libutil.h>
 #include <pwd.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -115,21 +116,19 @@ int	vflag;			/* verbose */
 int	fi;			/* open disk file descriptor */
 
 struct fileusage *
-	 addid(u_long, int, char *, char *);
-char	*blockcheck(char *);
+	 addid(u_long, int, char *, const char *);
 void	 bread(ufs2_daddr_t, char *, long);
 void	 freeinodebuf(void);
 union dinode *
 	 getnextinode(ino_t);
 int	 getquotagid(void);
-int	 hasquota(struct fstab *, int, char **);
 struct fileusage *
 	 lookup(u_long, int);
-struct quotaname *needchk(struct fstab *);
 int	 oneof(char *, char*[], int);
-void	 printchanges(char *, int, struct dqblk *, struct fileusage *, u_long);
+void	 printchanges(const char *, int, struct dqblk *, struct fileusage *,
+	    u_long);
 void	 setinodebuf(ino_t);
-int	 update(char *, char *, int);
+int	 update(const char *, struct quotafile *, int);
 void	 usage(void);
 
 int
@@ -138,7 +137,7 @@ main(int argc, char *argv[])
 	struct fstab *fs;
 	struct passwd *pw;
 	struct group *gr;
-	struct quotaname *qnp;
+	struct quotafile *qfu, *qfg;
 	int i, argnum, maxrun, errs, ch;
 	long done = 0;
 	char *name;
@@ -193,16 +192,27 @@ main(int argc, char *argv[])
 	if (maxrun > 0)
 		warnx("the -l option is now deprecated");
 	if (aflag)
-		exit(checkfstab());
+		exit(checkfstab(uflag, gflag));
 	if (setfsent() == 0)
 		errx(1, "%s: can't open", FSTAB);
 	while ((fs = getfsent()) != NULL) {
 		if (((argnum = oneof(fs->fs_file, argv, argc)) >= 0 ||
-		    (argnum = oneof(fs->fs_spec, argv, argc)) >= 0) &&
-		    (qnp = needchk(fs)) &&
+		     (argnum = oneof(fs->fs_spec, argv, argc)) >= 0) &&
 		    (name = blockcheck(fs->fs_spec))) {
 			done |= 1 << argnum;
-			errs += chkquota(name, fs->fs_file, qnp);
+			qfu = NULL;
+			if (uflag)
+				qfu = quota_open(fs, USRQUOTA, O_CREAT|O_RDWR);
+			qfg = NULL;
+			if (gflag)
+				qfg = quota_open(fs, GRPQUOTA, O_CREAT|O_RDWR);
+			if (qfu == NULL && qfg == NULL)
+				continue;
+			errs += chkquota(name, qfu, qfg);
+			if (qfu)
+				quota_close(qfu);
+			if (qfg)
+				quota_close(qfg);
 		}
 	}
 	endfsent();
@@ -222,32 +232,6 @@ usage(void)
 	exit(1);
 }
 
-struct quotaname *
-needchk(struct fstab *fs)
-{
-	struct quotaname *qnp;
-	char *qfnp;
-
-	if (strcmp(fs->fs_vfstype, "ufs") ||
-	    strcmp(fs->fs_type, FSTAB_RW))
-		return (NULL);
-	if ((qnp = malloc(sizeof(*qnp))) == NULL)
-		errx(1, "malloc failed");
-	qnp->flags = 0;
-	if (gflag && hasquota(fs, GRPQUOTA, &qfnp)) {
-		strcpy(qnp->grpqfname, qfnp);
-		qnp->flags |= HASGRP;
-	}
-	if (uflag && hasquota(fs, USRQUOTA, &qfnp)) {
-		strcpy(qnp->usrqfname, qfnp);
-		qnp->flags |= HASUSR;
-	}
-	if (qnp->flags)
-		return (qnp);
-	free(qnp);
-	return (NULL);
-}
-
 /*
  * Possible superblock locations ordered from most to least likely.
  */
@@ -257,20 +241,25 @@ static int sblock_try[] = SBLOCKSEARCH;
  * Scan the specified file system to check quota(s) present on it.
  */
 int
-chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
+chkquota(char *specname, struct quotafile *qfu, struct quotafile *qfg)
 {
 	struct fileusage *fup;
 	union dinode *dp;
 	int cg, i, mode, errs = 0;
 	ino_t ino, inosused, userino = 0, groupino = 0;
 	dev_t dev, userdev = 0, groupdev = 0;
-	char *cp;
 	struct stat sb;
+	const char *mntpt;
+	char *cp;
 
-	if (qnp == NULL)
-		err(1, "null quota information passed to chkquota()\n");
-	if ((fi = open(fsname, O_RDONLY, 0)) < 0) {
-		warn("%s", fsname);
+	if (qfu != NULL)
+		mntpt = quota_fsname(qfu);
+	else if (qfg != NULL)
+		mntpt = quota_fsname(qfg);
+	else
+		err(1, "null quotafile information passed to chkquota()\n");
+	if ((fi = open(specname, O_RDONLY, 0)) < 0) {
+		warn("%s", specname);
 		return (1);
 	}
 	if ((stat(mntpt, &sb)) < 0) {
@@ -280,21 +269,20 @@ chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
 	dev = sb.st_dev;
 	if (vflag) {
 		(void)printf("*** Checking ");
-		if (qnp->flags & HASUSR)
-			(void)printf("%s%s", qfextension[USRQUOTA],
-			    (qnp->flags & HASGRP) ? " and " : "");
-		if (qnp->flags & HASGRP)
-			(void)printf("%s", qfextension[GRPQUOTA]);
-		(void)printf(" quotas for %s (%s)\n", fsname, mntpt);
+		if (qfu)
+			(void)printf("user%s", qfg ? " and " : "");
+		if (qfg)
+			(void)printf("group");
+		(void)printf(" quotas for %s (%s)\n", specname, mntpt);
 	}
-	if (qnp->flags & HASUSR) {
-		if (stat(qnp->usrqfname, &sb) == 0) {
+	if (qfu) {
+		if (stat(quota_qfname(qfu), &sb) == 0) {
 			userino = sb.st_ino;
 			userdev = sb.st_dev;
 		}
 	}
-	if (qnp->flags & HASGRP) {
-		if (stat(qnp->grpqfname, &sb) == 0) {
+	if (qfg) {
+		if (stat(quota_qfname(qfg), &sb) == 0) {
 			groupino = sb.st_ino;
 			groupdev = sb.st_dev;
 		}
@@ -382,7 +370,7 @@ chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
 			if ((ino == userino && dev == userdev) ||
 			    (ino == groupino && dev == groupdev))
 				continue;
-			if (qnp->flags & HASGRP) {
+			if (qfg) {
 				fup = addid((u_long)DIP(dp, di_gid), GRPQUOTA,
 				    (char *)0, mntpt);
 				fup->fu_curinodes++;
@@ -390,7 +378,7 @@ chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
 				    mode == IFLNK)
 					fup->fu_curblocks += DIP(dp, di_blocks);
 			}
-			if (qnp->flags & HASUSR) {
+			if (qfu) {
 				fup = addid((u_long)DIP(dp, di_uid), USRQUOTA,
 				    (char *)0, mntpt);
 				fup->fu_curinodes++;
@@ -401,10 +389,10 @@ chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
 		}
 	}
 	freeinodebuf();
-	if (qnp->flags & HASUSR)
-		errs += update(mntpt, qnp->usrqfname, USRQUOTA);
-	if (qnp->flags & HASGRP)
-		errs += update(mntpt, qnp->grpqfname, GRPQUOTA);
+	if (qfu)
+		errs += update(mntpt, qfu, USRQUOTA);
+	if (qfg)
+		errs += update(mntpt, qfg, GRPQUOTA);
 	close(fi);
 	(void)fflush(stdout);
 	return (errs);
@@ -414,63 +402,20 @@ chkquota(char *fsname, char *mntpt, struct quotaname *qnp)
  * Update a specified quota file.
  */
 int
-update(char *fsname, char *quotafile, int type)
+update(const char *fsname, struct quotafile *qf, int type)
 {
 	struct fileusage *fup;
-	FILE *qfi, *qfo;
 	u_long id, lastid, highid = 0;
-	off_t offset;
-	int i;
 	struct dqblk dqbuf;
-	struct stat sb;
-	static int warned = 0;
 	static struct dqblk zerodqbuf;
 	static struct fileusage zerofileusage;
 
-	if ((qfo = fopen(quotafile, "r+")) == NULL) {
-		if (errno == ENOENT)
-			qfo = fopen(quotafile, "w+");
-		if (qfo) {
-			warnx("creating quota file %s", quotafile);
-#define	MODE	(S_IRUSR|S_IWUSR|S_IRGRP)
-			(void) fchown(fileno(qfo), getuid(), getquotagid());
-			(void) fchmod(fileno(qfo), MODE);
-		} else {
-			warn("%s", quotafile);
-			return (1);
-		}
-	}
-	if ((qfi = fopen(quotafile, "r")) == NULL) {
-		warn("%s", quotafile);
-		(void) fclose(qfo);
-		return (1);
-	}
-	if (quotactl(fsname, QCMD(Q_SYNC, type), (u_long)0, (caddr_t)0) < 0 &&
-	    errno == EOPNOTSUPP && !warned && vflag) {
-		warned++;
-		(void)printf("*** Warning: %s\n",
-		    "Quotas are not compiled into this kernel");
-	}
-	if (fstat(fileno(qfi), &sb) < 0) {
-		warn("Cannot fstat quota file %s\n", quotafile);
-		(void) fclose(qfo);
-		(void) fclose(qfi);
-		return (1);
-	}
-	if ((sb.st_size % sizeof(struct dqblk)) != 0)
-		warn("%s size is not a multiple of dqblk\n", quotafile);
-
 	/*
 	 * Scan the on-disk quota file and record any usage changes.
 	 */
-
-	if (sb.st_size != 0)
-		lastid = (sb.st_size / sizeof(struct dqblk)) - 1;
-	else
-		lastid = 0;
-	for (id = 0, offset = 0; id <= lastid;
-	    id++, offset += sizeof(struct dqblk)) {
-		if (fread((char *)&dqbuf, sizeof(struct dqblk), 1, qfi) == 0)
+	lastid = quota_maxid(qf);
+	for (id = 0; id <= lastid; id++) {
+		if (quota_read(qf, &dqbuf, id) < 0)
 			dqbuf = zerodqbuf;
 		if ((fup = lookup(id, type)) == NULL)
 			fup = &zerofileusage;
@@ -485,27 +430,9 @@ update(char *fsname, char *quotafile, int type)
 			continue;
 		}
 		printchanges(fsname, type, &dqbuf, fup, id);
-		/*
-		 * Reset time limit if have a soft limit and were
-		 * previously under it, but are now over it.
-		 */
-		if (dqbuf.dqb_bsoftlimit && id != 0 &&
-		    dqbuf.dqb_curblocks < dqbuf.dqb_bsoftlimit &&
-		    fup->fu_curblocks >= dqbuf.dqb_bsoftlimit)
-			dqbuf.dqb_btime = 0;
-		if (dqbuf.dqb_isoftlimit && id != 0 &&
-		    dqbuf.dqb_curinodes < dqbuf.dqb_isoftlimit &&
-		    fup->fu_curinodes >= dqbuf.dqb_isoftlimit)
-			dqbuf.dqb_itime = 0;
 		dqbuf.dqb_curinodes = fup->fu_curinodes;
 		dqbuf.dqb_curblocks = fup->fu_curblocks;
-		if (fseeko(qfo, offset, SEEK_SET) < 0) {
-			warn("%s: seek failed", quotafile);
-			return(1);
-		}
-		fwrite((char *)&dqbuf, sizeof(struct dqblk), 1, qfo);
-		(void) quotactl(fsname, QCMD(Q_SETUSE, type), id,
-		    (caddr_t)&dqbuf);
+		(void) quota_write_usage(qf, &dqbuf, id);
 		fup->fu_curinodes = 0;
 		fup->fu_curblocks = 0;
 	}
@@ -515,9 +442,8 @@ update(char *fsname, char *quotafile, int type)
 	 * that are not currently recorded in the quota file. E.g.
 	 * ids that are past the end of the current file.
 	 */
-
-	for (i = 0; i < FUHASH; i++) {
-		for (fup = fuhead[type][i]; fup != NULL; fup = fup->fu_next) {
+	for (id = 0; id < FUHASH; id++) {
+		for (fup = fuhead[type][id]; fup != NULL; fup = fup->fu_next) {
 			if (fup->fu_id <= lastid)
 				continue;
 			if (fup->fu_curinodes == 0 && fup->fu_curblocks == 0)
@@ -525,26 +451,17 @@ update(char *fsname, char *quotafile, int type)
 			bzero(&dqbuf, sizeof(struct dqblk));
 			if (fup->fu_id > highid)
 				highid = fup->fu_id;
-			printchanges(fsname, type, &dqbuf, fup, id);
+			printchanges(fsname, type, &dqbuf, fup, fup->fu_id);
 			dqbuf.dqb_curinodes = fup->fu_curinodes;
 			dqbuf.dqb_curblocks = fup->fu_curblocks;
-			offset = (off_t)fup->fu_id * sizeof(struct dqblk);
-			if (fseeko(qfo, offset, SEEK_SET) < 0) {
-				warn("%s: seek failed", quotafile);
-				return(1);
-			}
-			fwrite((char *)&dqbuf, sizeof(struct dqblk), 1, qfo);
-			(void) quotactl(fsname, QCMD(Q_SETUSE, type), id,
-			    (caddr_t)&dqbuf);
+			(void) quota_write_usage(qf, &dqbuf, fup->fu_id);
 			fup->fu_curinodes = 0;
 			fup->fu_curblocks = 0;
 		}
 	}
-	fclose(qfi);
-	fflush(qfo);
-	ftruncate(fileno(qfo),
-	    (((off_t)highid + 1) * sizeof(struct dqblk)));
-	fclose(qfo);
+	if (highid < lastid)
+		truncate(quota_qfname(qf),
+		    (((off_t)highid + 2) * sizeof(struct dqblk)));
 	return (0);
 }
 
@@ -575,55 +492,6 @@ getquotagid(void)
 	return (-1);
 }
 
-/*
- * Check to see if a particular quota is to be enabled.
- */
-int
-hasquota(struct fstab *fs, int type, char **qfnamep)
-{
-	char *opt;
-	char *cp;
-	struct statfs sfb;
-	static char initname, usrname[100], grpname[100];
-	static char buf[BUFSIZ];
-
-	if (!initname) {
-		(void)snprintf(usrname, sizeof(usrname), "%s%s",
-		    qfextension[USRQUOTA], qfname);
-		(void)snprintf(grpname, sizeof(grpname), "%s%s",
-		    qfextension[GRPQUOTA], qfname);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')) != NULL)
-			*cp++ = '\0';
-		if (type == USRQUOTA && strcmp(opt, usrname) == 0)
-			break;
-		if (type == GRPQUOTA && strcmp(opt, grpname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp)
-		*qfnamep = cp;
-	else {
-		(void)snprintf(buf, sizeof(buf), "%s/%s.%s", fs->fs_file,
-		    qfname, qfextension[type]);
-		*qfnamep = buf;
-	}
-	if (statfs(fs->fs_file, &sfb) != 0) {
-		warn("cannot statfs mount point %s", fs->fs_file);
-		return (0);
-	}
-	if (strcmp(fs->fs_file, sfb.f_mntonname)) {
-		warnx("%s not mounted for %s quotas", fs->fs_file,
-		    type == USRQUOTA ? "user" : "group");
-		return (0);
-	}
-	return (1);
-}
-
 /*
  * Routines to manage the file usage table.
  *
@@ -644,7 +512,7 @@ lookup(u_long id, int type)
  * Add a new file usage id if it does not already exist.
  */
 struct fileusage *
-addid(u_long id, int type, char *name, char *fsname)
+addid(u_long id, int type, char *name, const char *fsname)
 {
 	struct fileusage *fup, **fhp;
 	int len;
@@ -779,7 +647,7 @@ bread(ufs2_daddr_t bno, char *buf, long cnt)
  * Display updated block and i-node counts.
  */
 void
-printchanges(char *fsname, int type, struct dqblk *dp,
+printchanges(const char *fsname, int type, struct dqblk *dp,
     struct fileusage *fup, u_long id)
 {
 	if (!vflag)
diff --git a/sbin/quotacheck/quotacheck.h b/sbin/quotacheck/quotacheck.h
index aad9d30a76a..dcff75b68a8 100644
--- a/sbin/quotacheck/quotacheck.h
+++ b/sbin/quotacheck/quotacheck.h
@@ -32,12 +32,6 @@
  * $FreeBSD$
  */
 
-struct quotaname {
-	long	flags;
-	char	grpqfname[PATH_MAX];
-	char	usrqfname[PATH_MAX];
-};
-
-extern int checkfstab();
-extern int chkquota(char *, char *, struct quotaname *);
-extern struct quotaname *needchk(struct fstab *);
+extern char *blockcheck(char *);
+extern int checkfstab(int, int);
+extern int chkquota(char *, struct quotafile *, struct quotafile *);

From aee785babde44bcfef1d711653a85896dde7324b Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Mon, 28 Dec 2009 22:44:19 +0000
Subject: [PATCH 028/532] Add and document the quota_convert function which
 converts between the old 32-bit and the new 64-bit formats.

---
 lib/libutil/libutil.h   |  1 +
 lib/libutil/quotafile.3 | 34 ++++++++++----
 lib/libutil/quotafile.c | 99 ++++++++++++++++++++++++++++++++++++++++-
 3 files changed, 124 insertions(+), 10 deletions(-)

diff --git a/lib/libutil/libutil.h b/lib/libutil/libutil.h
index 3a05d9b56f6..7a4dc5670bc 100644
--- a/lib/libutil/libutil.h
+++ b/lib/libutil/libutil.h
@@ -154,6 +154,7 @@ int quota_check_path(const struct quotafile *, const char *path);
 int quota_read(struct quotafile *, struct dqblk *, int);
 int quota_write_limits(struct quotafile *, struct dqblk *, int);
 int quota_write_usage(struct quotafile *, struct dqblk *, int);
+int quota_convert(struct quotafile *, int);
 #endif
 
 __END_DECLS
diff --git a/lib/libutil/quotafile.3 b/lib/libutil/quotafile.3
index d2134c50530..5702cec9076 100644
--- a/lib/libutil/quotafile.3
+++ b/lib/libutil/quotafile.3
@@ -1,6 +1,6 @@
 .\"-
-.\" Copyright (c) 2008 Dag-Erling Coïdan Smørgrav
-.\" All rights reserved.
+.\" Copyright (c) 2009 Dag-Erling Coïdan Smørgrav and
+.\" Marshall Kirk McKusick. All rights reserved.
 .\"
 .\" Redistribution and use in source and binary forms, with or without
 .\" modification, are permitted provided that the following conditions
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd September 26, 2009
+.Dd December 28, 2009
 .Dt QUOTAFILE 3
 .Os
 .Sh NAME
@@ -40,6 +40,7 @@
 .Nm quota_qfname
 .Nm quota_maxid
 .Nm quota_check_path
+.Nm quota_convert
 .Nd "Manipulate quotas"
 .Sh LIBRARY
 .Lb libutil
@@ -72,6 +73,8 @@
 .Fn quota_maxid "const struct quotafile *qf"
 .Ft int
 .Fn quota_check_path "const struct quotafile *qf" "const char *path"
+.Ft int
+.Fn quota_convert "struct quotafile *qf" "int wordsize"
 .Sh DESCRIPTION
 These functions are designed to simplify access to filesystem quotas.
 If quotas are active on a filesystem,
@@ -219,9 +222,23 @@ If the
 argument refers to a symbolic link,
 .Fn quota_check_path
 will follow it.
+.Pp
+The
+.Fn quota_convert
+function converts the quota file associated with its
+.Va qf
+argument to the data size specified by its
+.Va wordsize
+argument.
+The supported wordsize arguments are 32 for the old 32-bit
+quota file format and 64 for the new 64-bit quota file format.
+The
+.Fn quota_convert
+function may only be called to operate on quota files that
+are not currently active.
 .Sh IMPLEMENTATION NOTES
-If the underlying quota file is in the old 32-bit format, limit and
-usage values written to the quota file will be clipped to 32 bits.
+If the underlying quota file is in or converted to the old 32-bit format,
+limit and usage values written to the quota file will be clipped to 32 bits.
 .Sh RETURN VALUES
 If the filesystem has quotas associated with it,
 .Fn quota_open
@@ -246,6 +263,7 @@ The
 .Fn quota_read ,
 .Fn quota_write_limits ,
 .Fn quota_write_usage ,
+.Fn quota_convert ,
 and
 .Fn quota_close
 functions return zero on success.
@@ -259,13 +277,13 @@ to indicate the error.
 .Xr quota.group 5
 .Sh HISTORY
 The
-.Nm
+.Nm quotafile
 functions first appeared in
-.Fx 8.0 .
+.Fx 8.1 .
 .Sh AUTHORS
 .An -nosplit
 The
-.Nm
+.Nm quotafile
 functions and this manual page were written by
 .An Dag-Erling Sm\(/orgrav Aq des@FreeBSD.org 
 and
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index 652d95a290f..ab57d662bdc 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -255,9 +255,9 @@ quota_maxid(struct quotafile *qf)
 		return (0);
 	switch (qf->wordsize) {
 	case 32:
-		return (st.st_size / sizeof(struct dqblk32));
+		return (st.st_size / sizeof(struct dqblk32) - 1);
 	case 64:
-		return (st.st_size / sizeof(struct dqblk64) - 1);
+		return (st.st_size / sizeof(struct dqblk64) - 2);
 	default:
 		return (0);
 	}
@@ -494,3 +494,98 @@ quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
 	}
 	/* not reached */
 }
+
+/*
+ * Convert a quota file from one format to another.
+ */
+int
+quota_convert(struct quotafile *qf, int wordsize)
+{
+	struct quotafile *newqf;
+	struct dqhdr64 dqh;
+	struct dqblk dqblk;
+	struct group *grp;
+	int serrno, maxid, id, fd;
+
+	/*
+	 * Quotas must not be active and quotafile must be open
+	 * for reading and writing.
+	 */
+	if ((qf->accmode & O_RDWR) != O_RDWR || qf->fd == -1) {
+		errno = EBADF;
+		return (-1);
+	}
+	if ((wordsize != 32 && wordsize != 64) ||
+	     wordsize == qf->wordsize) {
+		errno = EINVAL;
+		return (-1);
+	}
+	maxid = quota_maxid(qf);
+	if ((newqf = calloc(1, sizeof(*qf))) == NULL) {
+		errno = ENOMEM;
+		return (-1);
+	}
+	*newqf = *qf;
+	snprintf(newqf->qfname, MAXPATHLEN + 1, "%s_%d.orig", qf->qfname,
+	    qf->wordsize);
+	if (rename(qf->qfname, newqf->qfname) < 0) {
+		free(newqf);
+		return (-1);
+	}
+	if ((newqf->fd = open(qf->qfname, O_RDWR|O_CREAT|O_TRUNC, 0)) < 0) {
+		serrno = errno;
+		goto error;
+	}
+	newqf->wordsize = wordsize;
+	if (wordsize == 64) {
+		memset(&dqh, 0, sizeof(dqh));
+		memcpy(dqh.dqh_magic, Q_DQHDR64_MAGIC, sizeof(dqh.dqh_magic));
+		dqh.dqh_version = htobe32(Q_DQHDR64_VERSION);
+		dqh.dqh_hdrlen = htobe32(sizeof(struct dqhdr64));
+		dqh.dqh_reclen = htobe32(sizeof(struct dqblk64));
+		if (write(newqf->fd, &dqh, sizeof(dqh)) != sizeof(dqh)) {
+			serrno = errno;
+			goto error;
+		}
+	}
+	grp = getgrnam(QUOTAGROUP);
+	fchown(newqf->fd, 0, grp ? grp->gr_gid : 0);
+	fchmod(newqf->fd, 0640);
+	for (id = 0; id <= maxid; id++) {
+		if ((quota_read(qf, &dqblk, id)) < 0)
+			break;
+		switch (newqf->wordsize) {
+		case 32:
+			if ((quota_write32(newqf, &dqblk, id)) < 0)
+				break;
+			continue;
+		case 64:
+			if ((quota_write64(newqf, &dqblk, id)) < 0)
+				break;
+			continue;
+		default:
+			errno = EINVAL;
+			break;
+		}
+	}
+	if (id < maxid) {
+		serrno = errno;
+		goto error;
+	}
+	/*
+	 * Update the passed in quotafile to reference the new file
+	 * of the converted format size.
+	 */
+	fd = qf->fd;
+	qf->fd = newqf->fd;
+	newqf->fd = fd;
+	qf->wordsize = newqf->wordsize;
+	quota_close(newqf);
+	return (0);
+error:
+	/* put back the original file */
+	(void) rename(newqf->qfname, qf->qfname);
+	quota_close(newqf);
+	errno = serrno;
+	return (-1);
+}

From 46d2decf1ada39fefc050092012034fabcfaa061 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Mon, 28 Dec 2009 23:01:47 +0000
Subject: [PATCH 029/532] Add the -c option to quotacheck to use the
 quota_convert(3) function to convert between quota file formats.

---
 sbin/quotacheck/quotacheck.8 | 18 +++++++++++++
 sbin/quotacheck/quotacheck.c | 50 ++++++++++++++++++++++++++++++++----
 2 files changed, 63 insertions(+), 5 deletions(-)

diff --git a/sbin/quotacheck/quotacheck.8 b/sbin/quotacheck/quotacheck.8
index 5ba2ad76fa3..e8f4d9ca775 100644
--- a/sbin/quotacheck/quotacheck.8
+++ b/sbin/quotacheck/quotacheck.8
@@ -40,10 +40,12 @@
 .Sh SYNOPSIS
 .Nm
 .Op Fl guv
+.Op Fl c Ar 32 | 64
 .Op Fl l Ar maxrun
 .Fl a
 .Nm
 .Op Fl guv
+.Op Fl c Ar 32 | 64
 .Ar filesystem ...
 .Sh DESCRIPTION
 The
@@ -69,6 +71,22 @@ to be read-write with disk quotas.
 By default only the types of quotas listed in
 .Pa /etc/fstab
 are checked.
+.It Fl c Ar 32 | 64
+Before performing its checks,
+.Nm
+will convert the quota file to the specified word size.
+A conversion size of 64 is given to request conversion to
+the new 64-bit quota file format.
+A conversion size of 32 is given to request conversion back to
+the old 32-bit quota file format.
+The original quota file is left unchanged and moved aside with an
+underscore and its format size plus a
+.Pa .orig
+extension added to its name.
+Thus, the original 32-bit
+.Pa quota.user
+quota file converted to the 64-bit format quota file will be renamed to
+.Pa quota.user_32.orig .
 .It Fl g
 Only group quotas listed in
 .Pa /etc/fstab
diff --git a/sbin/quotacheck/quotacheck.c b/sbin/quotacheck/quotacheck.c
index 026608a419b..ed9ce753c55 100644
--- a/sbin/quotacheck/quotacheck.c
+++ b/sbin/quotacheck/quotacheck.c
@@ -110,6 +110,7 @@ struct fileusage {
 struct fileusage *fuhead[MAXQUOTAS][FUHASH];
 
 int	aflag;			/* all file systems */
+int	cflag;			/* convert format to 32 or 64 bit size */
 int	gflag;			/* check group quotas */
 int	uflag;			/* check user quotas */
 int	vflag;			/* verbose */
@@ -143,11 +144,16 @@ main(int argc, char *argv[])
 	char *name;
 
 	errs = maxrun = 0;
-	while ((ch = getopt(argc, argv, "aguvl:")) != -1) {
+	while ((ch = getopt(argc, argv, "ac:guvl:")) != -1) {
 		switch(ch) {
 		case 'a':
 			aflag++;
 			break;
+		case 'c':
+			if (cflag)
+				usage();
+			cflag = atoi(optarg);
+			break;
 		case 'g':
 			gflag++;
 			break;
@@ -168,6 +174,8 @@ main(int argc, char *argv[])
 	argv += optind;
 	if ((argc == 0 && !aflag) || (argc > 0 && aflag))
 		usage();
+	if (cflag && cflag != 32 && cflag != 64)
+		usage();
 	if (!gflag && !uflag) {
 		gflag++;
 		uflag++;
@@ -227,8 +235,8 @@ void
 usage(void)
 {
 	(void)fprintf(stderr, "%s\n%s\n",
-		"usage: quotacheck [-guv] [-l maxrun] -a",
-		"       quotacheck [-guv] filesystem ...");
+		"usage: quotacheck [-guv] [-c 32 | 64] [-l maxrun] -a",
+		"       quotacheck [-guv] [-c 32 | 64] filesystem ...");
 	exit(1);
 }
 
@@ -257,7 +265,31 @@ chkquota(char *specname, struct quotafile *qfu, struct quotafile *qfg)
 	else if (qfg != NULL)
 		mntpt = quota_fsname(qfg);
 	else
-		err(1, "null quotafile information passed to chkquota()\n");
+		errx(1, "null quotafile information passed to chkquota()\n");
+	if (cflag) {
+		if (vflag && qfu != NULL)
+			printf("%s: convert user quota to %d bits\n",
+			    mntpt, cflag);
+		if (qfu != NULL && quota_convert(qfu, cflag) < 0) {
+			if (errno == EBADF)
+				errx(1,
+				    "%s: cannot convert an active quota file",
+				    mntpt);
+			err(1, "user quota conversion to size %d failed",
+			    cflag);
+		}
+		if (vflag && qfg != NULL)
+			printf("%s: convert group quota to %d bits\n",
+			    mntpt, cflag);
+		if (qfg != NULL && quota_convert(qfg, cflag) < 0) {
+			if (errno == EBADF)
+				errx(1,
+				    "%s: cannot convert an active quota file",
+				    mntpt);
+			err(1, "group quota conversion to size %d failed",
+			    cflag);
+		}
+	}
 	if ((fi = open(specname, O_RDONLY, 0)) < 0) {
 		warn("%s", specname);
 		return (1);
@@ -407,6 +439,7 @@ update(const char *fsname, struct quotafile *qf, int type)
 	struct fileusage *fup;
 	u_long id, lastid, highid = 0;
 	struct dqblk dqbuf;
+	struct stat sb;
 	static struct dqblk zerodqbuf;
 	static struct fileusage zerofileusage;
 
@@ -459,7 +492,14 @@ update(const char *fsname, struct quotafile *qf, int type)
 			fup->fu_curblocks = 0;
 		}
 	}
-	if (highid < lastid)
+	/*
+	 * If this is old format file, then size may be smaller,
+	 * so ensure that we only truncate when it will make things
+	 * smaller, and not if it will grow an old format file.
+	 */
+	if (highid < lastid &&
+	    stat(quota_qfname(qf), &sb) == 0 &&
+	    sb.st_size > (((off_t)highid + 2) * sizeof(struct dqblk)))
 		truncate(quota_qfname(qf),
 		    (((off_t)highid + 2) * sizeof(struct dqblk)));
 	return (0);

From 9dc9c85fe9f1a06d23981220e362f04789a92d70 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Thu, 4 Mar 2010 00:47:24 +0000
Subject: [PATCH 030/532] WIP: the last missing piece of the quota64 puzzle. 
 Not quite there yet.

---
 libexec/rpc.rquotad/rquotad.c | 131 ++++++++--------------------------
 1 file changed, 28 insertions(+), 103 deletions(-)

diff --git a/libexec/rpc.rquotad/rquotad.c b/libexec/rpc.rquotad/rquotad.c
index a2fe04e6f98..3cd129239e7 100644
--- a/libexec/rpc.rquotad/rquotad.c
+++ b/libexec/rpc.rquotad/rquotad.c
@@ -36,20 +36,9 @@ static void rquota_service(struct svc_req *request, SVCXPRT *transp);
 static void sendquota(struct svc_req *request, SVCXPRT *transp);
 static void initfs(void);
 static int getfsquota(long id, char *path, struct dqblk *dqblk);
-static int hasquota(struct fstab *fs, char **qfnamep);
-
-/*
- * structure containing informations about ufs filesystems
- * initialised by initfs()
- */
-struct fs_stat {
-	struct fs_stat *fs_next;	/* next element */
-	char   *fs_file;		/* mount point of the filesystem */
-	char   *qfpathname;		/* pathname of the quota file */
-	dev_t   st_dev;			/* device of the filesystem */
-} fs_stat;
-static struct fs_stat *fs_begin = NULL;
 
+static struct quotafile **qfa;	/* array of qfs */
+static int nqf, szqf;		/* number of qfs and size of array */
 static int from_inetd = 1;
 
 static void
@@ -75,9 +64,7 @@ main(void)
 
 	if (!from_inetd) {
 		daemon(0, 0);
-
 		(void)rpcb_unset(RQUOTAPROG, RQUOTAVERS, NULL);
-
 		(void)signal(SIGINT, cleanup);
 		(void)signal(SIGTERM, cleanup);
 		(void)signal(SIGHUP, cleanup);
@@ -119,12 +106,10 @@ rquota_service(struct svc_req *request, SVCXPRT *transp)
 	case NULLPROC:
 		(void)svc_sendreply(transp, (xdrproc_t)xdr_void, (char *)NULL);
 		break;
-
 	case RQUOTAPROC_GETQUOTA:
 	case RQUOTAPROC_GETACTIVEQUOTA:
 		sendquota(request, transp);
 		break;
-
 	default:
 		svcerr_noproc(transp);
 		break;
@@ -183,39 +168,39 @@ sendquota(struct svc_req *request, SVCXPRT *transp)
 	}
 }
 
-/* initialise the fs_tab list from entries in /etc/fstab */
 static void
 initfs(void)
 {
-	struct fs_stat *fs_current = NULL;
-	struct fs_stat *fs_next = NULL;
-	char *qfpathname;
 	struct fstab *fs;
-	struct stat st;
 
 	setfsent();
+	szqf = 8;
+	if ((qfa = malloc(szqf * sizeof *qfa)) == NULL)
+		goto enomem;
 	while ((fs = getfsent())) {
 		if (strcmp(fs->fs_vfstype, "ufs"))
 			continue;
-		if (!hasquota(fs, &qfpathname))
+		if (nqf >= szqf) {
+			szqf *= 2;
+			if ((qfa = reallocf(qfa, szqf * sizeof *qfa)) == NULL)
+				goto enomem;
+		}
+		if ((qfa[nqf] = quota_open(fs, USRQUOTA, O_RDONLY)) == NULL) {
+			if (errno != EOPNOTSUPP)
+				goto fserr;
 			continue;
-
-		fs_current = malloc(sizeof(struct fs_stat));
-		fs_current->fs_next = fs_next;	/* next element */
-
-		fs_current->fs_file = malloc(strlen(fs->fs_file) + 1);
-		strcpy(fs_current->fs_file, fs->fs_file);
-
-		fs_current->qfpathname = malloc(strlen(qfpathname) + 1);
-		strcpy(fs_current->qfpathname, qfpathname);
-
-		stat(fs_current->fs_file, &st);
-		fs_current->st_dev = st.st_dev;
-
-		fs_next = fs_current;
+		}
+		++nqf;
+		/* XXX */
 	}
 	endfsent();
-	fs_begin = fs_current;
+	return;
+enomem:
+	syslog(LOG_ERR, "out of memory");
+	exit(1);
+fserr:
+	syslog(LOG_ERR, "%s: %s", fs->fs_file, strerror(errno));
+	exit(1);
 }
 
 /*
@@ -225,70 +210,10 @@ initfs(void)
 static int
 getfsquota(long id, char *path, struct dqblk *dqblk)
 {
-	struct quotafile *qf;
-	struct stat st_path;
-	struct fs_stat *fs;
-	int qcmd, ret = 0;
+	int i;
 
-	if (stat(path, &st_path) < 0)
-		return (0);
-
-	qcmd = QCMD(Q_GETQUOTA, USRQUOTA);
-
-	for (fs = fs_begin; fs != NULL; fs = fs->fs_next) {
-		/* where the device is the same as path */
-		if (fs->st_dev != st_path.st_dev)
-			continue;
-
-		/* find the specified filesystem. get and return quota */
-		if (quotactl(fs->fs_file, qcmd, id, dqblk) == 0)
-			return (1);
-
-		if ((qf = quota_open(fs->qfpathname)) == NULL) {
-			syslog(LOG_ERR, "open error: %s: %m", fs->qfpathname);
-			return (0);
-		}
-		if (quota_read(qf, dqblk, id) != 0) {
-			syslog(LOG_ERR, "read error: %s: %m", fs->qfpathname);
-			quota_close(qf);
-			return (0);
-		}
-		quota_close(qf);
-	}
-	return (ret);
-}
-
-/*
- * Check to see if a particular quota is to be enabled.
- * Comes from quota.c, NetBSD 0.9
- */
-static int
-hasquota(struct fstab *fs, char  **qfnamep)
-{
-	static char initname, usrname[100];
-	static char buf[BUFSIZ];
-	char *opt, *cp;
-	const char *qfextension[] = INITQFNAMES;
-
-	if (!initname) {
-		sprintf(usrname, "%s%s", qfextension[USRQUOTA], QUOTAFILENAME);
-		initname = 1;
-	}
-	strcpy(buf, fs->fs_mntops);
-	for (opt = strtok(buf, ","); opt; opt = strtok(NULL, ",")) {
-		if ((cp = index(opt, '=')))
-			*cp++ = '\0';
-		if (strcmp(opt, usrname) == 0)
-			break;
-	}
-	if (!opt)
-		return (0);
-	if (cp) {
-		*qfnamep = cp;
-		return (1);
-	}
-	sprintf(buf, "%s/%s.%s", fs->fs_file, QUOTAFILENAME,
-	    qfextension[USRQUOTA]);
-	*qfnamep = buf;
-	return (1);
+	for (i = 0; i < nqf; ++i)
+		if (quota_check_path(qfa[i], path) == 1)
+			return (quota_read(qfa[i], dqblk, id) == 0);
+	return (0);
 }

From 829b3f6be374043e042dd20b8b2097688820af22 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Thu, 4 Mar 2010 10:54:29 +0000
Subject: [PATCH 031/532] Fix warnings + minor style issues

---
 usr.bin/quota/quota.c | 63 +++++++++++++++++++++----------------------
 1 file changed, 30 insertions(+), 33 deletions(-)

diff --git a/usr.bin/quota/quota.c b/usr.bin/quota/quota.c
index 24dcb494230..240a7564bae 100644
--- a/usr.bin/quota/quota.c
+++ b/usr.bin/quota/quota.c
@@ -87,7 +87,7 @@ struct quotause {
 	char	fsname[MAXPATHLEN + 1];
 };
 
-static char *timeprt(int64_t seconds, int *needfree);
+static char *timeprt(int64_t seconds);
 static struct quotause *getprivs(long id, int quotatype);
 static void usage(void);
 static int showuid(u_long uid);
@@ -281,8 +281,7 @@ showquotas(int type, u_long id, const char *name)
 	struct quotause *quplist;
 	const char *msgi, *msgb;
 	const char *nam;
-	char *bgrace, *igrace;
-	int bfree, ifree;
+	char *bgrace = NULL, *igrace = NULL;
 	int lines = 0, overquota = 0;
 	static time_t now;
 
@@ -290,7 +289,7 @@ showquotas(int type, u_long id, const char *name)
 		time(&now);
 	quplist = getprivs(id, type);
 	for (qup = quplist; qup; qup = qup->next) {
-		msgi = (char *)0;
+		msgi = NULL;
 		if (qup->dqblk.dqb_ihardlimit &&
 		    qup->dqblk.dqb_curinodes >= qup->dqblk.dqb_ihardlimit) {
 			overquota++;
@@ -304,7 +303,7 @@ showquotas(int type, u_long id, const char *name)
 			else
 				msgi = "Over file quota on";
 		}
-		msgb = (char *)0;
+		msgb = NULL;
 		if (qup->dqblk.dqb_bhardlimit &&
 		    qup->dqblk.dqb_curblocks >= qup->dqblk.dqb_bhardlimit) {
 			overquota++;
@@ -329,12 +328,12 @@ showquotas(int type, u_long id, const char *name)
 		    qup->dqblk.dqb_bhardlimit == 0)
 			continue;
 		if (qflag) {
-			if ((msgi != (char *)0 || msgb != (char *)0) &&
+			if ((msgi != NULL || msgb != NULL) &&
 			    lines++ == 0)
 				heading(type, id, name, "");
-			if (msgi != (char *)0)
+			if (msgi != NULL)
 				printf("\t%s %s\n", msgi, qup->fsname);
-			if (msgb != (char *)0)
+			if (msgb != NULL)
 				printf("\t%s %s\n", msgb, qup->fsname);
 			continue;
 		}
@@ -352,42 +351,42 @@ showquotas(int type, u_long id, const char *name)
 		printf("%-15s", nam);
 		if (hflag) {
 			prthumanval(7, dbtob(qup->dqblk.dqb_curblocks));
-			printf("%c", (msgb == (char *)0) ? ' ' : '*');
+			printf("%c", (msgb == NULL) ? ' ' : '*');
 			prthumanval(7, dbtob(qup->dqblk.dqb_bsoftlimit));
 			prthumanval(7, dbtob(qup->dqblk.dqb_bhardlimit));
 		} else {
 			printf(" %7ju%c %7ju %7ju",
 			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_curblocks,
-			    (msgb == (char *)0) ? ' ' : '*',
+			    (msgb == NULL) ? ' ' : '*',
 			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_bsoftlimit,
 			    dbtob(1024) * (uintmax_t)qup->dqblk.dqb_bhardlimit);
 		}
 		if (msgb != NULL)
-			bgrace = timeprt(qup->dqblk.dqb_btime, &bfree);
+			bgrace = timeprt(qup->dqblk.dqb_btime);
 		if (msgi != NULL)
-			igrace = timeprt(qup->dqblk.dqb_itime, &ifree);
+			igrace = timeprt(qup->dqblk.dqb_itime);
 		printf("%8s %6ju%c %6ju %6ju%8s\n"
-			, (msgb == (char *)0) ? "" : bgrace
+			, (msgb == NULL) ? "" : bgrace
 			, (uintmax_t)qup->dqblk.dqb_curinodes
-			, (msgi == (char *)0) ? ' ' : '*'
+			, (msgi == NULL) ? ' ' : '*'
 			, (uintmax_t)qup->dqblk.dqb_isoftlimit
 			, (uintmax_t)qup->dqblk.dqb_ihardlimit
-			, (msgi == (char *)0) ? "" : igrace
+			, (msgi == NULL) ? "" : igrace
 		);
-		if (msgb != NULL && bfree)
+		if (msgb != NULL)
 			free(bgrace);
-		if (msgi != NULL && ifree)
+		if (msgi != NULL)
 			free(igrace);
 	}
 	if (!qflag && !rflag && lines == 0)
 		heading(type, id, name, "none");
-	return(overquota);
+	return (overquota);
 }
 
 static void
 showrawquotas(int type, u_long id, struct quotause *qup)
 {
-	time_t time;
+	time_t t;
 
 	printf("Raw %s quota information for id %lu on %s\n",
 	    type == USRQUOTA ? "user" : "group", id, qup->fsname);
@@ -406,15 +405,15 @@ showrawquotas(int type, u_long id, struct quotause *qup)
 	printf("block grace time:     %jd",
 	    (intmax_t)qup->dqblk.dqb_btime);
 	if (qup->dqblk.dqb_btime != 0) {
-		time = qup->dqblk.dqb_btime;
-		printf(" %s", ctime(&time));
+		t = qup->dqblk.dqb_btime;
+		printf(" %s", ctime(&t));
 	} else {
 		printf("\n");
 	}
 	printf("i-node grace time:    %jd", (intmax_t)qup->dqblk.dqb_itime);
 	if (qup->dqblk.dqb_itime != 0) {
-		time = qup->dqblk.dqb_itime;
-		printf(" %s", ctime(&time));
+		t = qup->dqblk.dqb_itime;
+		printf(" %s", ctime(&t));
 	} else {
 		printf("\n");
 	}
@@ -446,37 +445,35 @@ heading(int type, u_long id, const char *name, const char *tag)
  * Calculate the grace period and return a printable string for it.
  */
 static char *
-timeprt(int64_t seconds, int *needfree)
+timeprt(int64_t seconds)
 {
 	time_t hours, minutes;
-	char	*buf;
+	char *buf;
 	static time_t now;
 
 	if (now == 0)
 		time(&now);
 	if (now > seconds) {
-		*needfree = 0;
-		return ("none");
+		if ((buf = strdup("none")) == NULL)
+			errx(1, "strdup() failed in timeprt()");
+		return (buf);
 	}
 	seconds -= now;
 	minutes = (seconds + 30) / 60;
 	hours = (minutes + 30) / 60;
 	if (hours >= 36) {
 		if (asprintf(&buf, "%lddays", ((long)hours + 12) / 24) < 0)
-			errx(1, "asprintf failed in timeprt(1)");
-		*needfree = 1;
+			errx(1, "asprintf() failed in timeprt(1)");
 		return (buf);
 	}
 	if (minutes >= 60) {
 		if (asprintf(&buf, "%2ld:%ld", (long)minutes / 60,
 		    (long)minutes % 60) < 0)
-			errx(1, "asprintf failed in timeprt(2)");
-		*needfree = 1;
+			errx(1, "asprintf() failed in timeprt(2)");
 		return (buf);
 	}
 	if (asprintf(&buf, "%2ld", (long)minutes) < 0)
-		errx(1, "asprintf failed in timeprt(3)");
-	*needfree = 1;
+		errx(1, "asprintf() failed in timeprt(3)");
 	return (buf);
 }
 

From 7c54172a4f9db27283aaead2ee3c694f109ca497 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Thu, 4 Mar 2010 10:57:52 +0000
Subject: [PATCH 032/532] Fix warnings + indentation + utmpx

---
 usr.sbin/repquota/repquota.c | 48 +++++++++++++++++++-----------------
 1 file changed, 25 insertions(+), 23 deletions(-)

diff --git a/usr.sbin/repquota/repquota.c b/usr.sbin/repquota/repquota.c
index aa822df09d2..f90269f173c 100644
--- a/usr.sbin/repquota/repquota.c
+++ b/usr.sbin/repquota/repquota.c
@@ -49,7 +49,9 @@ __FBSDID("$FreeBSD$");
  */
 #include <sys/param.h>
 #include <sys/mount.h>
+
 #include <ufs/ufs/quota.h>
+
 #include <err.h>
 #include <errno.h>
 #include <fcntl.h>
@@ -57,12 +59,12 @@ __FBSDID("$FreeBSD$");
 #include <grp.h>
 #include <libutil.h>
 #include <pwd.h>
+#include <stdint.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <time.h>
 #include <unistd.h>
-#include <utmp.h>
 
 /* Let's be paranoid about block size */
 #if 10 > DEV_BSHIFT
@@ -218,9 +220,9 @@ repquota(struct fstab *fs, int type)
 		fprintf(stdout, "*** Report for %s quotas on %s (%s)\n",
 		    qfextension[type], fs->fs_file, fs->fs_spec);
 	printf("%*s           Block  limits                    File  limits\n",
-		max(UT_NAMESIZE,10), " ");
+		max(MAXLOGNAME - 1, 10), " ");
 	printf("User%*s  used   soft   hard  grace     used    soft    hard  grace\n",
-		max(UT_NAMESIZE,10), " ");
+		max(MAXLOGNAME - 1, 10), " ");
 	maxid = quota_maxid(qf);
 	for (id = 0; id < maxid; id++) {
 		if (quota_read(qf, &dqbuf, id) != 0)
@@ -229,30 +231,30 @@ repquota(struct fstab *fs, int type)
 			continue;
 		if ((fup = lookup(id, type)) == 0)
 			fup = addid(id, type, (char *)0);
-		printf("%-*s ", max(UT_NAMESIZE,10), fup->fu_name);
+		printf("%-*s ", max(MAXLOGNAME - 1, 10), fup->fu_name);
 		printf("%c%c", 
-			dqbuf.dqb_bsoftlimit &&
-			    dqbuf.dqb_curblocks >=
-			    dqbuf.dqb_bsoftlimit ? '+' : '-',
-			dqbuf.dqb_isoftlimit &&
-			    dqbuf.dqb_curinodes >=
-			    dqbuf.dqb_isoftlimit ? '+' : '-');
+		    dqbuf.dqb_bsoftlimit &&
+		    dqbuf.dqb_curblocks >=
+		    dqbuf.dqb_bsoftlimit ? '+' : '-',
+		    dqbuf.dqb_isoftlimit &&
+		    dqbuf.dqb_curinodes >=
+		    dqbuf.dqb_isoftlimit ? '+' : '-');
 		prthumanval(dqbuf.dqb_curblocks);
 		prthumanval(dqbuf.dqb_bsoftlimit);
 		prthumanval(dqbuf.dqb_bhardlimit);
 		printf(" %6s",
-			dqbuf.dqb_bsoftlimit &&
-			    dqbuf.dqb_curblocks >=
-			    dqbuf.dqb_bsoftlimit ?
-			    timeprt(dqbuf.dqb_btime) : "-");
-		printf("  %7llu %7llu %7llu %6s\n",
-			dqbuf.dqb_curinodes,
-			dqbuf.dqb_isoftlimit,
-			dqbuf.dqb_ihardlimit,
-			dqbuf.dqb_isoftlimit &&
-			    dqbuf.dqb_curinodes >=
-			    dqbuf.dqb_isoftlimit ?
-			    timeprt(dqbuf.dqb_itime) : "-");
+		    dqbuf.dqb_bsoftlimit &&
+		    dqbuf.dqb_curblocks >=
+		    dqbuf.dqb_bsoftlimit ?
+		    timeprt(dqbuf.dqb_btime) : "-");
+		printf("  %7ju %7ju %7ju %6s\n",
+		    (uintmax_t)dqbuf.dqb_curinodes,
+		    (uintmax_t)dqbuf.dqb_isoftlimit,
+		    (uintmax_t)dqbuf.dqb_ihardlimit,
+		    dqbuf.dqb_isoftlimit &&
+		    dqbuf.dqb_curinodes >=
+		    dqbuf.dqb_isoftlimit ?
+		    timeprt(dqbuf.dqb_itime) : "-");
 	}
 	quota_close(qf);
 	return (0);
@@ -265,7 +267,7 @@ prthumanval(int64_t blocks)
 	int flags;
 
 	if (!hflag) {
-		printf(" %6llu", dbtokb(blocks));
+		printf(" %6ju", (uintmax_t)dbtokb(blocks));
 		return;
 	}
 	flags = HN_NOSPACE | HN_DECIMAL;

From 516ad57b74be8c0b950e246197c1cd5e091995f3 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Tue, 16 Mar 2010 06:12:30 +0000
Subject: [PATCH 033/532] Debugging nits found while testing the new 64-bit
 quota code.

---
 lib/libc/sys/quotactl.2      | 22 ++++++++++++++++----
 lib/libutil/quotafile.c      | 34 ++++++++++++++++---------------
 sys/ufs/ufs/quota.h          |  2 ++
 sys/ufs/ufs/ufs_quota.c      | 39 +++++++++++++++++++++++++++++++++---
 sys/ufs/ufs/ufs_vfsops.c     |  4 ++++
 usr.bin/quota/Makefile       |  1 +
 usr.sbin/repquota/repquota.c |  2 +-
 7 files changed, 80 insertions(+), 24 deletions(-)

diff --git a/lib/libc/sys/quotactl.2 b/lib/libc/sys/quotactl.2
index 4b4dca5fbcb..ff3cb4b1756 100644
--- a/lib/libc/sys/quotactl.2
+++ b/lib/libc/sys/quotactl.2
@@ -84,7 +84,7 @@ and group identifiers (GRPQUOTA).
 The
 .Dq ufs
 specific commands are:
-.Bl -tag -width Q_QUOTAOFFxx
+.Bl -tag -width Q_GETQUOTASIZEx
 .It Dv Q_QUOTAON
 Enable disk quotas for the file system specified by
 .Fa path .
@@ -110,6 +110,17 @@ and
 .Fa id
 arguments are unused.
 Only the super-user may turn quotas off.
+.It Dv Q_GETQUOTASIZE
+Get the wordsize used to represent the quotas for the user or group
+(as determined by the command type).
+Possible values are 32 for the old-style quota file 
+and 64 for the new-style quota file.
+The
+.Fa addr
+argument is a pointer to an integer into which the size is stored.
+The identifier
+.Fa id
+is not used.
 .It Dv Q_GETQUOTA
 Get disk quota limits and current usage for the user or group
 (as determined by the command type) with identifier
@@ -177,9 +188,11 @@ The
 argument
 or the command type is invalid.
 In
-.Dv Q_GETQUOTA
-and
+.Dv Q_GETQUOTASIZE ,
+.Dv Q_GETQUOTA ,
 .Dv Q_SETQUOTA ,
+and
+.Dv Q_SETUSE ,
 quotas are not currently enabled for this file system.
 .Pp
 The
@@ -208,7 +221,8 @@ Too many symbolic links were encountered in translating a pathname.
 .It Bq Er EROFS
 In
 .Dv Q_QUOTAON ,
-the quota file resides on a read-only file system.
+either the file system on which quotas are to be enabled is mounted read-only
+or the quota file resides on a read-only file system.
 .It Bq Er EIO
 An
 .Tn I/O
diff --git a/lib/libutil/quotafile.c b/lib/libutil/quotafile.c
index ab57d662bdc..0fda5f62687 100644
--- a/lib/libutil/quotafile.c
+++ b/lib/libutil/quotafile.c
@@ -129,11 +129,9 @@ quota_open(struct fstab *fs, int quotatype, int openflags)
 		goto error;
 	qf->dev = st.st_dev;
 	serrno = hasquota(fs, quotatype, qf->qfname, sizeof(qf->qfname));
-	qcmd = QCMD(Q_GETQUOTA, quotatype);
-	if (quotactl(fs->fs_file, qcmd, 0, &dqh) == 0) {
-		qf->wordsize = 64;
+	qcmd = QCMD(Q_GETQUOTASIZE, quotatype);
+	if (quotactl(qf->fsname, qcmd, 0, &qf->wordsize) == 0)
 		return (qf);
-	}
 	if (serrno == 0) {
 		errno = EOPNOTSUPP;
 		goto error;
@@ -250,18 +248,22 @@ int
 quota_maxid(struct quotafile *qf)
 {
 	struct stat st;
+	int maxid;
 
 	if (stat(qf->qfname, &st) < 0)
 		return (0);
 	switch (qf->wordsize) {
 	case 32:
-		return (st.st_size / sizeof(struct dqblk32) - 1);
+		maxid = st.st_size / sizeof(struct dqblk32) - 1;
+		break;
 	case 64:
-		return (st.st_size / sizeof(struct dqblk64) - 2);
+		maxid = st.st_size / sizeof(struct dqblk64) - 2;
+		break;
 	default:
-		return (0);
+		maxid = 0;
+		break;
 	}
-	/* not reached */
+	return (maxid > 0 ? maxid : 0);
 }
 
 static int
@@ -395,10 +397,6 @@ quota_write_usage(struct quotafile *qf, struct dqblk *dqb, int id)
 	struct dqblk dqbuf;
 	int qcmd;
 
-	if ((qf->accmode & O_RDWR) != O_RDWR) {
-		errno = EBADF;
-		return (-1);
-	}
 	if (qf->fd == -1) {
 		qcmd = QCMD(Q_SETUSE, qf->quotatype);
 		return (quotactl(qf->fsname, qcmd, id, dqb));
@@ -406,6 +404,10 @@ quota_write_usage(struct quotafile *qf, struct dqblk *dqb, int id)
 	/*
 	 * Have to do read-modify-write of quota in file.
 	 */
+	if ((qf->accmode & O_RDWR) != O_RDWR) {
+		errno = EBADF;
+		return (-1);
+	}
 	if (quota_read(qf, &dqbuf, id) != 0)
 		return (-1);
 	/*
@@ -443,10 +445,6 @@ quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
 	struct dqblk dqbuf;
 	int qcmd;
 
-	if ((qf->accmode & O_RDWR) != O_RDWR) {
-		errno = EBADF;
-		return (-1);
-	}
 	if (qf->fd == -1) {
 		qcmd = QCMD(Q_SETQUOTA, qf->quotatype);
 		return (quotactl(qf->fsname, qcmd, id, dqb));
@@ -454,6 +452,10 @@ quota_write_limits(struct quotafile *qf, struct dqblk *dqb, int id)
 	/*
 	 * Have to do read-modify-write of quota in file.
 	 */
+	if ((qf->accmode & O_RDWR) != O_RDWR) {
+		errno = EBADF;
+		return (-1);
+	}
 	if (quota_read(qf, &dqbuf, id) != 0)
 		return (-1);
 	/*
diff --git a/sys/ufs/ufs/quota.h b/sys/ufs/ufs/quota.h
index 82b0170b036..ca0dcced7d2 100644
--- a/sys/ufs/ufs/quota.h
+++ b/sys/ufs/ufs/quota.h
@@ -88,6 +88,7 @@
 #define	Q_GETQUOTA	0x0700	/* get limits and usage (64-bit version) */
 #define	Q_SETQUOTA	0x0800	/* set limits and usage (64-bit version) */
 #define	Q_SETUSE	0x0900	/* set usage (64-bit version) */
+#define	Q_GETQUOTASIZE	0x0A00	/* get bit-size of quota file fields */
 
 /*
  * The following structure defines the format of the disk quota file
@@ -235,6 +236,7 @@ int	setuse32(struct thread *, struct mount *, u_long, int, void *);
 int	getquota(struct thread *, struct mount *, u_long, int, void *);
 int	setquota(struct thread *, struct mount *, u_long, int, void *);
 int	setuse(struct thread *, struct mount *, u_long, int, void *);
+int	getquotasize(struct thread *, struct mount *, u_long, int, void *);
 vfs_quotactl_t ufs_quotactl;
 
 #else /* !_KERNEL */
diff --git a/sys/ufs/ufs/ufs_quota.c b/sys/ufs/ufs/ufs_quota.c
index 5b1da50b73a..c5161899aa8 100644
--- a/sys/ufs/ufs/ufs_quota.c
+++ b/sys/ufs/ufs/ufs_quota.c
@@ -508,6 +508,9 @@ quotaon(struct thread *td, struct mount *mp, int type, void *fname)
 	if (error)
 		return (error);
 
+	if (mp->mnt_flag & MNT_RDONLY)
+		return (EROFS);
+
 	ump = VFSTOUFS(mp);
 	dq = NODQUOT;
 
@@ -534,7 +537,9 @@ quotaon(struct thread *td, struct mount *mp, int type, void *fname)
 		return (EALREADY);
 	}
 	ump->um_qflags[type] |= QTF_OPENING|QTF_CLOSING;
+	UFS_UNLOCK(ump);
 	if ((error = dqopen(vp, ump, type)) != 0) {
+		UFS_LOCK(ump);
 		ump->um_qflags[type] &= ~(QTF_OPENING|QTF_CLOSING);
 		UFS_UNLOCK(ump);
 		(void) vn_close(vp, FREAD|FWRITE, td->td_ucred, td);
@@ -544,7 +549,6 @@ quotaon(struct thread *td, struct mount *mp, int type, void *fname)
 	MNT_ILOCK(mp);
 	mp->mnt_flag |= MNT_QUOTA;
 	MNT_IUNLOCK(mp);
-	UFS_UNLOCK(ump);
 
 	vpp = &ump->um_quotas[type];
 	if (*vpp != vp)
@@ -988,6 +992,30 @@ setuse(struct thread *td, struct mount *mp, u_long id, int type, void *addr)
 	return (error);
 }
 
+/*
+ * Q_GETQUOTASIZE - get bit-size of quota file fields
+ */
+int
+getquotasize(struct thread *td, struct mount *mp, u_long id, int type,
+    void *sizep)
+{
+	struct ufsmount *ump = VFSTOUFS(mp);
+	int bitsize;
+
+	UFS_LOCK(ump);
+	if (ump->um_quotas[type] == NULLVP ||
+	    (ump->um_qflags[type] & QTF_CLOSING)) {
+		UFS_UNLOCK(ump);
+		return (EINVAL);
+	}
+	if ((ump->um_qflags[type] & QTF_64BIT) != 0)
+		bitsize = 64;
+	else
+		bitsize = 32;
+	UFS_UNLOCK(ump);
+	return (copyout(&bitsize, sizep, sizeof(int)));
+}
+
 /*
  * Q_SYNC - sync quota files to disk.
  */
@@ -1163,12 +1191,17 @@ dqopen(struct vnode *vp, struct ufsmount *ump, int type)
 		return (0);
 	}
 
+	UFS_LOCK(ump);
 	if (strcmp(dqh.dqh_magic, Q_DQHDR64_MAGIC) == 0 &&
 	    be32toh(dqh.dqh_version) == Q_DQHDR64_VERSION &&
 	    be32toh(dqh.dqh_hdrlen) == (uint32_t)sizeof(struct dqhdr64) &&
-	    be32toh(dqh.dqh_reclen) == (uint32_t)sizeof(struct dqblk64))
+	    be32toh(dqh.dqh_reclen) == (uint32_t)sizeof(struct dqblk64)) {
+		/* XXX: what if the magic matches, but the sizes are wrong? */
 		ump->um_qflags[type] |= QTF_64BIT;
-	/* XXX: what if the magic matches, but the sizes are wrong? */
+	} else {
+		ump->um_qflags[type] &= ~QTF_64BIT;
+	}
+	UFS_UNLOCK(ump);
 
 	return (0);
 }
diff --git a/sys/ufs/ufs/ufs_vfsops.c b/sys/ufs/ufs/ufs_vfsops.c
index 3abd5eb7839..0eeb14fc54e 100644
--- a/sys/ufs/ufs/ufs_vfsops.c
+++ b/sys/ufs/ufs/ufs_vfsops.c
@@ -151,6 +151,10 @@ ufs_quotactl(mp, cmds, id, arg)
 		error = getquota(td, mp, id, type, arg);
 		break;
 
+	case Q_GETQUOTASIZE:
+		error = getquotasize(td, mp, id, type, arg);
+		break;
+
 	case Q_SYNC:
 		error = qsync(mp);
 		break;
diff --git a/usr.bin/quota/Makefile b/usr.bin/quota/Makefile
index a479402878f..26585ae2b72 100644
--- a/usr.bin/quota/Makefile
+++ b/usr.bin/quota/Makefile
@@ -3,6 +3,7 @@
 
 PROG=	quota
 BINOWN=	root
+BINMODE=4555
 
 DPADD=	${LIBRPCSVC} ${LIBUTIL}
 LDADD=	-lrpcsvc -lutil
diff --git a/usr.sbin/repquota/repquota.c b/usr.sbin/repquota/repquota.c
index f90269f173c..8fdea201554 100644
--- a/usr.sbin/repquota/repquota.c
+++ b/usr.sbin/repquota/repquota.c
@@ -224,7 +224,7 @@ repquota(struct fstab *fs, int type)
 	printf("User%*s  used   soft   hard  grace     used    soft    hard  grace\n",
 		max(MAXLOGNAME - 1, 10), " ");
 	maxid = quota_maxid(qf);
-	for (id = 0; id < maxid; id++) {
+	for (id = 0; id <= maxid; id++) {
 		if (quota_read(qf, &dqbuf, id) != 0)
 			break;
 		if (dqbuf.dqb_curinodes == 0 && dqbuf.dqb_curblocks == 0)

From 1a0a89c328dc7176c2058e08087761d0b9957ff8 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Mon, 12 Apr 2010 17:43:58 +0000
Subject: [PATCH 034/532] Import zlib 1.2.4.3 (trimmed).

---
 ChangeLog  | 28 ++++++++++++++++++++++++++++
 README     |  7 ++++---
 deflate.c  |  2 +-
 gzguts.h   | 12 +++++++-----
 gzlib.c    |  6 ++++--
 inftrees.c |  4 ++--
 minigzip.c |  5 +++++
 zconf.h    | 11 ++++++++++-
 zlib.3     |  4 ++--
 zlib.h     | 44 ++++++++++++++++++++++++++------------------
 zutil.h    | 16 +++-------------
 11 files changed, 92 insertions(+), 47 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 9817f59b7a1..898c197f0cf 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,6 +1,34 @@
 
                 ChangeLog file for zlib
 
+Changes in 1.2.4.3 (10 Apr 2010)
+- Only use CROSS_PREFIX in configure for ar and ranlib if they exist
+- Use CROSS_PREFIX for nm [Bar-Lev]
+- Assume _LARGEFILE64_SOURCE defined is equivalent to true
+- Avoid use of undefined symbols in #if with && and ||
+- Make *64 prototypes in gzguts.h consistent with functions
+- Add -shared load option for MinGW in configure [Bowler]
+- Move z_off64_t to public interface, use instead of off64_t
+- Remove ! from shell test in configure (not portable to Solaris)
+- Change +0 macro tests to -0 for possibly increased portability
+
+Changes in 1.2.4.2 (9 Apr 2010)
+- Add consistent carriage returns to readme.txt's in masmx86 and masmx64
+- Really provide prototypes for *64 functions when building without LFS
+- Only define unlink() in minigzip.c if unistd.h not included
+- Update README to point to contrib/vstudio project files
+- Move projects/vc6 to old/ and remove projects/
+- Include stdlib.h in minigzip.c for setmode() definition under WinCE
+- Clean up assembler builds in win32/Makefile.msc [Rowe]
+- Include sys/types.h for Microsoft for off_t definition
+- Fix memory leak on error in gz_open()
+- Symbolize nm as $NM in configure [Weigelt]
+- Use TEST_LDSHARED instead of LDSHARED to link test programs [Weigelt]
+- Add +0 to _FILE_OFFSET_BITS and _LFS64_LARGEFILE in case not defined
+- Fix bug in gzeof() to take into account unused input data
+- Avoid initialization of structures with variables in puff.c
+- Updated win32/README-WIN32.txt [Rowe]
+
 Changes in 1.2.4.1 (28 Mar 2010)
 - Remove the use of [a-z] constructs for sed in configure [gentoo 310225]
 - Remove $(SHAREDLIB) from LIBS in Makefile.in [Creech]
diff --git a/README b/README
index a3302675ec2..68ff992369c 100644
--- a/README
+++ b/README
@@ -1,6 +1,6 @@
 ZLIB DATA COMPRESSION LIBRARY
 
-zlib 1.2.4.1 is a general purpose data compression library.  All the code is
+zlib 1.2.4.3 is a general purpose data compression library.  All the code is
 thread safe.  The data format used by the zlib library is described by RFCs
 (Request for Comments) 1950 to 1952 in the files
 http://www.ietf.org/rfc/rfc1950.txt (zlib format), rfc1951.txt (deflate format)
@@ -16,7 +16,8 @@ minigzip.c.
 To compile all files and run the test program, follow the instructions given at
 the top of Makefile.in.  In short "./configure; make test", and if that goes
 well, "make install" should work for most flavors of Unix.  For Windows, use one
-of the special makefiles in win32/ or projects/ .  For VMS, use make_vms.com.
+of the special makefiles in win32/ or contrib/vstudio/ .  For VMS, use
+make_vms.com.
 
 Questions about zlib should be sent to <zlib@gzip.org>, or to Gilles Vollant
 <info@winimage.com> for the Windows DLL version.  The zlib home page is
@@ -30,7 +31,7 @@ Mark Nelson <markn@ieee.org> wrote an article about zlib for the Jan.  1997
 issue of Dr.  Dobb's Journal; a copy of the article is available at
 http://marknelson.us/1997/01/01/zlib-engine/ .
 
-The changes made in version 1.2.4.1 are documented in the file ChangeLog.
+The changes made in version 1.2.4.3 are documented in the file ChangeLog.
 
 Unsupported third party contributions are provided in directory contrib/ .
 
diff --git a/deflate.c b/deflate.c
index b7fb9fc6982..efb26bf5d91 100644
--- a/deflate.c
+++ b/deflate.c
@@ -52,7 +52,7 @@
 #include "deflate.h"
 
 const char deflate_copyright[] =
-   " deflate 1.2.4.1 Copyright 1995-2010 Jean-loup Gailly and Mark Adler ";
+   " deflate 1.2.4.3 Copyright 1995-2010 Jean-loup Gailly and Mark Adler ";
 /*
   If you use the zlib library in a product, an acknowledgment is welcome
   in the documentation of your product. If for some reason you cannot
diff --git a/gzguts.h b/gzguts.h
index 53857e0a42b..b0a4cbdc336 100644
--- a/gzguts.h
+++ b/gzguts.h
@@ -3,7 +3,7 @@
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
-#if _LARGEFILE64_SOURCE == 1
+#ifdef _LARGEFILE64_SOURCE
 #  ifndef _LARGEFILE_SOURCE
 #    define _LARGEFILE_SOURCE 1
 #  endif
@@ -56,10 +56,12 @@
 #  endif
 #endif
 
-#if _LARGEFILE64_SOURCE == 1
-#  define z_off64_t off64_t
-#else
-#  define z_off64_t z_off_t
+/* provide prototypes for these when building zlib without LFS */
+#if !defined(_LARGEFILE64_SOURCE) || _LFS64_LARGEFILE-0 == 0
+    ZEXTERN gzFile ZEXPORT gzopen64 OF((const char *, const char *));
+    ZEXTERN z_off64_t ZEXPORT gzseek64 OF((gzFile, z_off64_t, int));
+    ZEXTERN z_off64_t ZEXPORT gztell64 OF((gzFile));
+    ZEXTERN z_off64_t ZEXPORT gzoffset64 OF((gzFile));
 #endif
 
 /* default i/o buffer size -- double this for output when reading */
diff --git a/gzlib.c b/gzlib.c
index 15999aee8b5..18390026905 100644
--- a/gzlib.c
+++ b/gzlib.c
@@ -5,7 +5,7 @@
 
 #include "gzguts.h"
 
-#if _LARGEFILE64_SOURCE == 1 && _LFS64_LARGEFILE == 1
+#if defined(_LARGEFILE64_SOURCE) && _LFS64_LARGEFILE-0
 #  define LSEEK lseek64
 #else
 #  define LSEEK lseek
@@ -172,6 +172,7 @@ local gzFile gz_open(path, fd, mode)
                         O_APPEND))),
             0666);
     if (state->fd == -1) {
+        free(state->path);
         free(state);
         return NULL;
     }
@@ -432,7 +433,8 @@ int ZEXPORT gzeof(file)
         return 0;
 
     /* return end-of-file state */
-    return state->mode == GZ_READ ? (state->eof && state->have == 0) : 0;
+    return state->mode == GZ_READ ?
+        (state->eof && state->strm.avail_in == 0 && state->have == 0) : 0;
 }
 
 /* -- see zlib.h -- */
diff --git a/inftrees.c b/inftrees.c
index 90de8b87f25..430b1741585 100644
--- a/inftrees.c
+++ b/inftrees.c
@@ -9,7 +9,7 @@
 #define MAXBITS 15
 
 const char inflate_copyright[] =
-   " inflate 1.2.4.1 Copyright 1995-2010 Mark Adler ";
+   " inflate 1.2.4.3 Copyright 1995-2010 Mark Adler ";
 /*
   If you use the zlib library in a product, an acknowledgment is welcome
   in the documentation of your product. If for some reason you cannot
@@ -62,7 +62,7 @@ unsigned short FAR *work;
         35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
     static const unsigned short lext[31] = { /* Length codes 257..285 extra */
         16, 16, 16, 16, 16, 16, 16, 16, 17, 17, 17, 17, 18, 18, 18, 18,
-        19, 19, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 16, 67, 206};
+        19, 19, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 16, 195, 66};
     static const unsigned short dbase[32] = { /* Distance codes 0..29 base */
         1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
         257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
diff --git a/minigzip.c b/minigzip.c
index 9677fa07f63..9825ccc3a71 100644
--- a/minigzip.c
+++ b/minigzip.c
@@ -32,6 +32,9 @@
 #if defined(MSDOS) || defined(OS2) || defined(WIN32) || defined(__CYGWIN__)
 #  include <fcntl.h>
 #  include <io.h>
+#  ifdef UNDER_CE
+#    include <stdlib.h>
+#  endif
 #  define SET_BINARY_MODE(file) setmode(fileno(file), O_BINARY)
 #else
 #  define SET_BINARY_MODE(file)
@@ -50,9 +53,11 @@
 #  include <unix.h> /* for fileno */
 #endif
 
+#if !defined(Z_HAVE_UNISTD_H) && !defined(_LARGEFILE64_SOURCE)
 #ifndef WIN32 /* unlink already in stdio.h for WIN32 */
   extern int unlink OF((const char *));
 #endif
+#endif
 
 #if defined(UNDER_CE)
 #  include <windows.h>
diff --git a/zconf.h b/zconf.h
index 6ad8a04edae..1988920109f 100644
--- a/zconf.h
+++ b/zconf.h
@@ -364,8 +364,11 @@ typedef uLong FAR uLongf;
 #  define Z_HAVE_UNISTD_H
 #endif
 
-#if defined(Z_HAVE_UNISTD_H) || _LARGEFILE64_SOURCE == 1
+#ifdef STDC
 #  include <sys/types.h>    /* for off_t */
+#endif
+
+#if defined(Z_HAVE_UNISTD_H) || defined(_LARGEFILE64_SOURCE)
 #  include <unistd.h>       /* for SEEK_* and off_t */
 #  ifdef VMS
 #    include <unixio.h>     /* for off_t */
@@ -385,6 +388,12 @@ typedef uLong FAR uLongf;
 #  define z_off_t long
 #endif
 
+#if defined(_LARGEFILE64_SOURCE) && _LFS64_LARGEFILE-0
+#  define z_off64_t off64_t
+#else
+#  define z_off64_t z_off_t
+#endif
+
 #if defined(__OS400__)
 #  define NO_vsnprintf
 #endif
diff --git a/zlib.3 b/zlib.3
index 2c1679d7b27..b5346168d0e 100644
--- a/zlib.3
+++ b/zlib.3
@@ -1,4 +1,4 @@
-.TH ZLIB 3 "28 Mar 2010"
+.TH ZLIB 3 "10 Apr 2010"
 .SH NAME
 zlib \- compression/decompression library
 .SH SYNOPSIS
@@ -125,7 +125,7 @@ before asking for help.
 Send questions and/or comments to zlib@gzip.org,
 or (for the Windows DLL version) to Gilles Vollant (info@winimage.com).
 .SH AUTHORS
-Version 1.2.4.1
+Version 1.2.4.3
 Copyright (C) 1995-2010 Jean-loup Gailly (jloup@gzip.org)
 and Mark Adler (madler@alumni.caltech.edu).
 .LP
diff --git a/zlib.h b/zlib.h
index 1fcf6300cc2..699630c1ddc 100644
--- a/zlib.h
+++ b/zlib.h
@@ -1,5 +1,5 @@
 /* zlib.h -- interface of the 'zlib' general purpose compression library
-  version 1.2.4.1, March 28th, 2010
+  version 1.2.4.3, April 10th, 2010
 
   Copyright (C) 1995-2010 Jean-loup Gailly and Mark Adler
 
@@ -37,12 +37,12 @@
 extern "C" {
 #endif
 
-#define ZLIB_VERSION "1.2.4.1"
-#define ZLIB_VERNUM 0x1241
+#define ZLIB_VERSION "1.2.4.3"
+#define ZLIB_VERNUM 0x1243
 #define ZLIB_VER_MAJOR 1
 #define ZLIB_VER_MINOR 2
 #define ZLIB_VER_REVISION 4
-#define ZLIB_VER_SUBREVISION 1
+#define ZLIB_VER_SUBREVISION 3
 
 /*
     The 'zlib' compression library provides in-memory compression and
@@ -1556,29 +1556,35 @@ ZEXTERN int ZEXPORT inflateBackInit_ OF((z_streamp strm, int windowBits,
         inflateBackInit_((strm), (windowBits), (window), \
                                             ZLIB_VERSION, sizeof(z_stream))
 
-#if _LARGEFILE64_SOURCE == 1 && _LFS64_LARGEFILE == 1
+/* provide 64-bit offset functions if _LARGEFILE64_SOURCE defined, and/or
+ * change the regular functions to 64 bits if _FILE_OFFSET_BITS is 64 (if
+ * both are true, the application gets the *64 functions, and the regular
+ * functions are changed to 64 bits) -- in case these are set on systems
+ * without large file support, _LFS64_LARGEFILE must also be true
+ */
+#if defined(_LARGEFILE64_SOURCE) && _LFS64_LARGEFILE-0
    ZEXTERN gzFile ZEXPORT gzopen64 OF((const char *, const char *));
-   ZEXTERN off64_t ZEXPORT gzseek64 OF((gzFile, off64_t, int));
-   ZEXTERN off64_t ZEXPORT gztell64 OF((gzFile));
-   ZEXTERN off64_t ZEXPORT gzoffset64 OF((gzFile));
-   ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, off64_t));
-   ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, off64_t));
+   ZEXTERN z_off64_t ZEXPORT gzseek64 OF((gzFile, z_off64_t, int));
+   ZEXTERN z_off64_t ZEXPORT gztell64 OF((gzFile));
+   ZEXTERN z_off64_t ZEXPORT gzoffset64 OF((gzFile));
+   ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, z_off64_t));
+   ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, z_off64_t));
 #endif
 
-#if !defined(ZLIB_INTERNAL) && _FILE_OFFSET_BITS == 64 && _LFS64_LARGEFILE == 1
+#if !defined(ZLIB_INTERNAL) && _FILE_OFFSET_BITS-0 == 64 && _LFS64_LARGEFILE-0
 #  define gzopen gzopen64
 #  define gzseek gzseek64
 #  define gztell gztell64
 #  define gzoffset gzoffset64
 #  define adler32_combine adler32_combine64
 #  define crc32_combine crc32_combine64
-#  if _LARGEFILE64_SOURCE != 1
+#  ifdef _LARGEFILE64_SOURCE
      ZEXTERN gzFile ZEXPORT gzopen64 OF((const char *, const char *));
-     ZEXTERN off_t ZEXPORT gzseek64 OF((gzFile, off_t, int));
-     ZEXTERN off_t ZEXPORT gztell64 OF((gzFile));
-     ZEXTERN off_t ZEXPORT gzoffset64 OF((gzFile));
-     ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, off_t));
-     ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, off_t));
+     ZEXTERN z_off_t ZEXPORT gzseek64 OF((gzFile, z_off_t, int));
+     ZEXTERN z_off_t ZEXPORT gztell64 OF((gzFile));
+     ZEXTERN z_off_t ZEXPORT gzoffset64 OF((gzFile));
+     ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, z_off_t));
+     ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, z_off_t));
 #  endif
 #else
    ZEXTERN gzFile ZEXPORT gzopen OF((const char *, const char *));
@@ -1589,10 +1595,12 @@ ZEXTERN int ZEXPORT inflateBackInit_ OF((z_streamp strm, int windowBits,
    ZEXTERN uLong ZEXPORT crc32_combine OF((uLong, uLong, z_off_t));
 #endif
 
+/* hack for buggy compilers */
 #if !defined(ZUTIL_H) && !defined(NO_DUMMY_DECL)
-    struct internal_state {int dummy;}; /* hack for buggy compilers */
+    struct internal_state {int dummy;};
 #endif
 
+/* undocumented functions */
 ZEXTERN const char   * ZEXPORT zError           OF((int));
 ZEXTERN int            ZEXPORT inflateSyncPoint OF((z_streamp));
 ZEXTERN const uLongf * ZEXPORT get_crc_table    OF((void));
diff --git a/zutil.h b/zutil.h
index b21a19cc043..a250088640a 100644
--- a/zutil.h
+++ b/zutil.h
@@ -154,20 +154,10 @@ extern const char * const z_errmsg[10]; /* indexed by 2-zlib_error */
   #pragma warn -8066
 #endif
 
-#if _LARGEFILE64_SOURCE == 1 && _LFS64_LARGEFILE == 1
-#  define z_off64_t off64_t
-#else
-#  define z_off64_t z_off_t
-#endif
-
 /* provide prototypes for these when building zlib without LFS */
-#if _LARGEFILE64_SOURCE != 1 || _LFS64_LARGEFILE != 1
-    ZEXTERN gzFile ZEXPORT gzopen64 OF((const char *, const char *));
-    ZEXTERN off_t ZEXPORT gzseek64 OF((gzFile, off_t, int));
-    ZEXTERN off_t ZEXPORT gztell64 OF((gzFile));
-    ZEXTERN off_t ZEXPORT gzoffset64 OF((gzFile));
-    ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, off_t));
-    ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, off_t));
+#if !defined(_LARGEFILE64_SOURCE) || _LFS64_LARGEFILE-0 == 0
+    ZEXTERN uLong ZEXPORT adler32_combine64 OF((uLong, uLong, z_off_t));
+    ZEXTERN uLong ZEXPORT crc32_combine64 OF((uLong, uLong, z_off_t));
 #endif
 
         /* common defaults */

From 0a5cd2607a04bc91bc394a08dd7b38b7ab75f0c2 Mon Sep 17 00:00:00 2001
From: Neel Natu <neel@FreeBSD.org>
Date: Wed, 14 Apr 2010 01:29:31 +0000
Subject: [PATCH 035/532] Revert the vm_machdep.c part of r205072.

This causes a panic in vm_thread_dispose() when it tries to add this kstack
to the kstack cache. This happens only when 'td_kstack' is not (PAGE_SIZE * 2)
bytes aligned and we have unmapped the page at that address in cpu_thread_alloc.

Pointed out by: nwhitehorn@
---
 sys/mips/mips/vm_machdep.c | 20 ++------------------
 1 file changed, 2 insertions(+), 18 deletions(-)

diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index 3f8e6cc7f29..57219f638f5 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -214,16 +214,6 @@ cpu_thread_swapin(struct thread *td)
 {
 	pt_entry_t *pte;
 	int i;
-	vm_offset_t unused_kstack_page;
-
-	/*
-	 * Unmap the unused kstack page.
-	 */
-	unused_kstack_page = td->td_kstack;
-	if (td->td_md.md_realstack == td->td_kstack)
-		unused_kstack_page += (KSTACK_PAGES - 1) * PAGE_SIZE;
-
-	pmap_kremove(unused_kstack_page);
 
 	/*
 	 * The kstack may be at a different physical address now.
@@ -249,19 +239,13 @@ cpu_thread_swapout(struct thread *td)
 void
 cpu_thread_alloc(struct thread *td)
 {
-	vm_offset_t unused_kstack_page;
 	pt_entry_t *pte;
 	int i;
 
-	if (td->td_kstack & (1 << PAGE_SHIFT)) {
+	if (td->td_kstack & (1 << PAGE_SHIFT))
 		td->td_md.md_realstack = td->td_kstack + PAGE_SIZE;
-		unused_kstack_page = td->td_kstack;
-	} else {
+	else
 		td->td_md.md_realstack = td->td_kstack;
-		unused_kstack_page = td->td_kstack +
-					(KSTACK_PAGES - 1) * PAGE_SIZE;
-	}
-	pmap_kremove(unused_kstack_page);
 
 	td->td_pcb = (struct pcb *)(td->td_md.md_realstack +
 	    (td->td_kstack_pages - 1) * PAGE_SIZE) - 1;

From 923ca9167bfee22ed4d3a749ae5e01064d0bae10 Mon Sep 17 00:00:00 2001
From: Neel Natu <neel@FreeBSD.org>
Date: Wed, 14 Apr 2010 01:57:53 +0000
Subject: [PATCH 036/532] Destroy the pmap 'pm_mutex' in pmap_release()
 otherwise we will panic subsequently in pmap_pinit() with the following
 signature:

panic: lock "pmap" 0xc7878bc8 already initialized

This bug was uncovered by the changes made to vm_map.c in r206140.
---
 sys/mips/mips/pmap.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 73f57e4fbee..d5765ea9b0e 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -1263,6 +1263,7 @@ pmap_release(pmap_t pmap)
 	ptdpg->wire_count--;
 	atomic_subtract_int(&cnt.v_wire_count, 1);
 	vm_page_free_zero(ptdpg);
+	PMAP_LOCK_DESTROY(pmap);
 }
 
 /*

From 547f27959bb05d2f5e998cb0fd3d9d68313c6800 Mon Sep 17 00:00:00 2001
From: Kevin Lo <kevlo@FreeBSD.org>
Date: Wed, 14 Apr 2010 07:07:43 +0000
Subject: [PATCH 037/532] Eliminate duplicate comment

---
 sys/dev/usb/usbdevs | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/usb/usbdevs b/sys/dev/usb/usbdevs
index 4cf2891af76..a7bac4405e8 100644
--- a/sys/dev/usb/usbdevs
+++ b/sys/dev/usb/usbdevs
@@ -2515,7 +2515,6 @@ product QUALCOMMINC E0086	0x0086	3G modem
 product QUALCOMMINC E2002	0x2002	3G modem
 product QUALCOMMINC E2003	0x2003	3G modem
 
-/* Quanta products */
 /* Quanta products */
 product QUANTA RW6815_1		0x00ce	HP iPAQ rw6815
 product QUANTA RT3070		0x0304	RT3070

From 0a6b4947d88bb216a16441ccaa4aa9164efbbbd3 Mon Sep 17 00:00:00 2001
From: Ed Maste <emaste@FreeBSD.org>
Date: Wed, 14 Apr 2010 13:44:22 +0000
Subject: [PATCH 038/532] Linux puts a blank line between each CPU.

---
 sys/compat/linprocfs/linprocfs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/compat/linprocfs/linprocfs.c b/sys/compat/linprocfs/linprocfs.c
index e6a4f16564a..58d897e9d08 100644
--- a/sys/compat/linprocfs/linprocfs.c
+++ b/sys/compat/linprocfs/linprocfs.c
@@ -274,7 +274,7 @@ linprocfs_docpuinfo(PFS_FILL_ARGS)
 		    "cpu family\t: %d\n"
 		    "model\t\t: %d\n"
 		    "model name\t: %s\n"
-		    "stepping\t: %d\n",
+		    "stepping\t: %d\n\n",
 		    i, cpu_vendor, class, cpu, model, cpu_id & 0xf);
 		/* XXX per-cpu vendor / class / model / id? */
 	}

From b9f59cca0d3f142de2acd5426da7bb950c85bb3e Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Wed, 14 Apr 2010 15:29:32 +0000
Subject: [PATCH 039/532] For early ALI chips do not announce I/O sizes that
 require unsupported 48bit DMA commands.

---
 sys/dev/ata/chipsets/ata-acerlabs.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/sys/dev/ata/chipsets/ata-acerlabs.c b/sys/dev/ata/chipsets/ata-acerlabs.c
index b7f11472da2..fee9692ebd5 100644
--- a/sys/dev/ata/chipsets/ata-acerlabs.c
+++ b/sys/dev/ata/chipsets/ata-acerlabs.c
@@ -184,8 +184,11 @@ ata_ali_ch_attach(device_t dev)
     if (ctlr->chip->cfg2 & ALI_NEW && ctlr->chip->chiprev < 0xc7)
 	ch->flags |= ATA_CHECKS_CABLE;
     /* older chips can't do 48bit DMA transfers */
-    if (ctlr->chip->chiprev <= 0xc4)
+    if (ctlr->chip->chiprev <= 0xc4) {
 	ch->flags |= ATA_NO_48BIT_DMA;
+	if (ch->dma.max_iosize > 256 * 512)
+		ch->dma.max_iosize = 256 * 512;
+    }
 
     return 0;
 }

From 7ea05bb9a5cf8d64f7479c30ecdc781028f09463 Mon Sep 17 00:00:00 2001
From: Jack F Vogel <jfv@FreeBSD.org>
Date: Wed, 14 Apr 2010 18:29:01 +0000
Subject: [PATCH 040/532] Remove multiqueue stack related stuff form lem, it is
 unneeded for legacy hardware. Also remove some TSO related cruft. Add some
 watchdog_time setting that was missing, thanks to Mikolaj Golub for pointing
 that out.

---
 sys/dev/e1000/if_lem.c | 164 ++---------------------------------------
 1 file changed, 5 insertions(+), 159 deletions(-)

diff --git a/sys/dev/e1000/if_lem.c b/sys/dev/e1000/if_lem.c
index cf71f2bb74b..825fb4ee6f0 100644
--- a/sys/dev/e1000/if_lem.c
+++ b/sys/dev/e1000/if_lem.c
@@ -39,9 +39,6 @@
 
 #include <sys/param.h>
 #include <sys/systm.h>
-#if __FreeBSD_version >= 800000
-#include <sys/buf_ring.h>
-#endif
 #include <sys/bus.h>
 #include <sys/endian.h>
 #include <sys/kernel.h>
@@ -94,7 +91,7 @@ int	lem_display_debug_stats = 0;
 /*********************************************************************
  *  Legacy Em Driver version:
  *********************************************************************/
-char lem_driver_version[] = "1.0.0";
+char lem_driver_version[] = "1.0.1";
 
 
 /*********************************************************************
@@ -177,11 +174,6 @@ static int	lem_suspend(device_t);
 static int	lem_resume(device_t);
 static void	lem_start(struct ifnet *);
 static void	lem_start_locked(struct ifnet *ifp);
-#if __FreeBSD_version >= 800000
-static int	lem_mq_start(struct ifnet *, struct mbuf *);
-static int	lem_mq_start_locked(struct ifnet *, struct mbuf *);
-static void	lem_qflush(struct ifnet *);
-#endif
 static int	lem_ioctl(struct ifnet *, u_long, caddr_t);
 static void	lem_init(void *);
 static void	lem_init_locked(struct adapter *);
@@ -304,12 +296,6 @@ MODULE_DEPEND(lem, ether, 1, 1, 1);
 
 #define EM_TICKS_TO_USECS(ticks)	((1024 * (ticks) + 500) / 1000)
 #define EM_USECS_TO_TICKS(usecs)	((1000 * (usecs) + 512) / 1024)
-#define M_TSO_LEN			66
-
-/* Allow common code without TSO */
-#ifndef CSUM_TSO
-#define CSUM_TSO	0
-#endif
 
 static int lem_tx_int_delay_dflt = EM_TICKS_TO_USECS(EM_TIDV);
 static int lem_rx_int_delay_dflt = EM_TICKS_TO_USECS(EM_RDTR);
@@ -827,118 +813,6 @@ lem_resume(device_t dev)
 }
 
 
-/*********************************************************************
- *  Transmit entry point
- *
- *  em_start is called by the stack to initiate a transmit.
- *  The driver will remain in this routine as long as there are
- *  packets to transmit and transmit resources are available.
- *  In case resources are not available stack is notified and
- *  the packet is requeued.
- **********************************************************************/
-
-#if __FreeBSD_version >= 800000
-static int
-lem_mq_start_locked(struct ifnet *ifp, struct mbuf *m)
-{
-	struct adapter	*adapter = ifp->if_softc;
-	struct mbuf	*next;
-	int error = E1000_SUCCESS;
-
-	EM_TX_LOCK_ASSERT(adapter);
-	/* To allow being called from a tasklet */
-	if (m == NULL)
-		goto process;
-
-	if (((ifp->if_drv_flags & (IFF_DRV_RUNNING|IFF_DRV_OACTIVE)) !=
-	    IFF_DRV_RUNNING)
-	    || (!adapter->link_active)) {
-		error = drbr_enqueue(ifp, adapter->br, m);
-		return (error);
-	} else if (drbr_empty(ifp, adapter->br) &&
-	    (adapter->num_tx_desc_avail > EM_TX_OP_THRESHOLD)) {
-		if ((error = lem_xmit(adapter, &m)) != 0) {
-			if (m)
-				error = drbr_enqueue(ifp, adapter->br, m);
-			return (error);
-		} else {
-			/*
-			 * We've bypassed the buf ring so we need to update
-			 * ifp directly
-			 */
-			drbr_stats_update(ifp, m->m_pkthdr.len, m->m_flags);
-			/*
-			** Send a copy of the frame to the BPF
-			** listener and set the watchdog on.
-			*/
-			ETHER_BPF_MTAP(ifp, m);
-			adapter->watchdog_check = TRUE;
-		}
-	} else if ((error = drbr_enqueue(ifp, adapter->br, m)) != 0)
-		return (error);
-	
-process:
-	if (drbr_empty(ifp, adapter->br))
-		return(error);
-        /* Process the queue */
-        while (TRUE) {
-                if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
-                        break;
-                next = drbr_dequeue(ifp, adapter->br);
-                if (next == NULL)
-                        break;
-                if ((error = lem_xmit(adapter, &next)) != 0) {
-			if (next != NULL)
-				error = drbr_enqueue(ifp, adapter->br, next);
-                        break;
-		}
-		drbr_stats_update(ifp, next->m_pkthdr.len, next->m_flags);
-                ETHER_BPF_MTAP(ifp, next);
-                /* Set the watchdog */
-		adapter->watchdog_check = TRUE;
-        }
-
-        if (adapter->num_tx_desc_avail <= EM_TX_OP_THRESHOLD)
-                ifp->if_drv_flags |= IFF_DRV_OACTIVE;
-
-	return (error);
-}
-
-/*
-** Multiqueue capable stack interface, this is not
-** yet truely multiqueue, but that is coming...
-*/
-static int
-lem_mq_start(struct ifnet *ifp, struct mbuf *m)
-{
-	
-	struct adapter *adapter = ifp->if_softc;
-	int error = 0;
-
-	if (EM_TX_TRYLOCK(adapter)) {
-		if (ifp->if_drv_flags & IFF_DRV_RUNNING)
-			error = lem_mq_start_locked(ifp, m);
-		EM_TX_UNLOCK(adapter);
-	} else 
-		error = drbr_enqueue(ifp, adapter->br, m);
-
-	return (error);
-}
-
-static void
-lem_qflush(struct ifnet *ifp)
-{
-	struct mbuf *m;
-	struct adapter *adapter = (struct adapter *)ifp->if_softc;
-
-	EM_TX_LOCK(adapter);
-	while ((m = buf_ring_dequeue_sc(adapter->br)) != NULL)
-		m_freem(m);
-	if_qflush(ifp);
-	EM_TX_UNLOCK(adapter);
-}
-#endif /* FreeBSD_version */
-
 static void
 lem_start_locked(struct ifnet *ifp)
 {
@@ -975,6 +849,7 @@ lem_start_locked(struct ifnet *ifp)
 
 		/* Set timeout in case hardware has problems transmitting. */
 		adapter->watchdog_check = TRUE;
+		adapter->watchdog_time = ticks;
 	}
 	if (adapter->num_tx_desc_avail <= EM_TX_OP_THRESHOLD)
 		ifp->if_drv_flags |= IFF_DRV_OACTIVE;
@@ -1151,12 +1026,6 @@ lem_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 			ifp->if_capenable ^= IFCAP_HWCSUM;
 			reinit = 1;
 		}
-#if __FreeBSD_version >= 700000
-		if (mask & IFCAP_TSO4) {
-			ifp->if_capenable ^= IFCAP_TSO4;
-			reinit = 1;
-		}
-#endif
 		if (mask & IFCAP_VLAN_HWTAGGING) {
 			ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
 			reinit = 1;
@@ -1279,10 +1148,6 @@ lem_init_locked(struct adapter *adapter)
 	if (adapter->hw.mac.type >= e1000_82543) {
 		if (ifp->if_capenable & IFCAP_TXCSUM)
 			ifp->if_hwassist |= (CSUM_TCP | CSUM_UDP);
-#if __FreeBSD_version >= 700000
-		if (ifp->if_capenable & IFCAP_TSO4)
-			ifp->if_hwassist |= CSUM_TSO;
-#endif
 	}
 
 	/* Configure for OS presence */
@@ -1394,13 +1259,8 @@ lem_poll(struct ifnet *ifp, enum poll_cmd cmd, int count)
 
 	EM_TX_LOCK(adapter);
 	lem_txeof(adapter);
-#if __FreeBSD_version >= 800000
-	if (!drbr_empty(ifp, adapter->br))
-		lem_mq_start_locked(ifp, NULL);
-#else
 	if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
 		lem_start_locked(ifp);
-#endif
 	EM_TX_UNLOCK(adapter);
 	return (rx_done);
 }
@@ -1494,14 +1354,8 @@ lem_handle_rxtx(void *context, int pending)
 			taskqueue_enqueue(adapter->tq, &adapter->rxtx_task);
 		EM_TX_LOCK(adapter);
 		lem_txeof(adapter);
-
-#if __FreeBSD_version >= 800000
-		if (!drbr_empty(ifp, adapter->br))
-			lem_mq_start_locked(ifp, NULL);
-#else
 		if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
 			lem_start_locked(ifp);
-#endif
 		EM_TX_UNLOCK(adapter);
 	}
 
@@ -1852,15 +1706,17 @@ lem_xmit(struct adapter *adapter, struct mbuf **m_headp)
         if (mtag != NULL) {
                 ctxd->upper.fields.special =
                     htole16(VLAN_TAG_VALUE(mtag));
+                ctxd->lower.data |= htole32(E1000_TXD_CMD_VLE);
+	}
 #else /* FreeBSD 7 */
 	if (m_head->m_flags & M_VLANTAG) {
 		/* Set the vlan id. */
 		ctxd->upper.fields.special =
 		    htole16(m_head->m_pkthdr.ether_vtag);
-#endif
                 /* Tell hardware to add tag */
                 ctxd->lower.data |= htole32(E1000_TXD_CMD_VLE);
         }
+#endif
 
         tx_buffer->m_head = m_head;
 	tx_buffer_mapped->map = tx_buffer->map;
@@ -2544,12 +2400,6 @@ lem_setup_interface(device_t dev, struct adapter *adapter)
 
 	ifp->if_capabilities = ifp->if_capenable = 0;
 
-#if __FreeBSD_version >= 800000
-	/* Multiqueue tx functions */
-	ifp->if_transmit = lem_mq_start;
-	ifp->if_qflush = lem_qflush;
-	adapter->br = buf_ring_alloc(4096, M_DEVBUF, M_WAITOK, &adapter->tx_mtx);
-#endif	
 	if (adapter->hw.mac.type >= e1000_82543) {
 		int version_cap;
 #if __FreeBSD_version < 700000
@@ -4549,10 +4399,6 @@ lem_print_hw_stats(struct adapter *adapter)
 	    (long long)adapter->stats.gprc);
 	device_printf(dev, "Good Packets Xmtd = %lld\n",
 	    (long long)adapter->stats.gptc);
-	device_printf(dev, "TSO Contexts Xmtd = %lld\n",
-	    (long long)adapter->stats.tsctc);
-	device_printf(dev, "TSO Contexts Failed = %lld\n",
-	    (long long)adapter->stats.tsctfc);
 }
 
 /**********************************************************************

From a17c1cc009da9c4c4535c5f1114fdd704f0ad9e9 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ulrich=20Sp=C3=B6rlein?= <uqs@FreeBSD.org>
Date: Wed, 14 Apr 2010 18:29:26 +0000
Subject: [PATCH 041/532] mdoc: remove .Pp where not needed

This trips up mdocml and can simply go away.

Reviewed by:	ru
Approved by:	philip, ed (mentors)
---
 lib/libc/gen/confstr.3    | 1 -
 lib/libc/gen/sysconf.3    | 4 ----
 lib/libc/gen/sysctl.3     | 5 -----
 lib/libc/locale/isalnum.3 | 1 -
 lib/libc/locale/isalpha.3 | 1 -
 5 files changed, 12 deletions(-)

diff --git a/lib/libc/gen/confstr.3 b/lib/libc/gen/confstr.3
index e97505f07de..3066aae363d 100644
--- a/lib/libc/gen/confstr.3
+++ b/lib/libc/gen/confstr.3
@@ -79,7 +79,6 @@ The copied value is always null terminated.
 The available values are as follows:
 .Pp
 .Bl -tag -width 6n
-.Pp
 .It Li _CS_PATH
 Return a value for the
 .Ev PATH
diff --git a/lib/libc/gen/sysconf.3 b/lib/libc/gen/sysconf.3
index b17d8a80d5e..05766ccef94 100644
--- a/lib/libc/gen/sysconf.3
+++ b/lib/libc/gen/sysconf.3
@@ -60,9 +60,7 @@ Shell programmers who need access to these parameters should use the
 utility.
 .Pp
 The available values are as follows:
-.Pp
 .Bl -tag -width 6n
-.Pp
 .It Li _SC_ARG_MAX
 The maximum bytes of argument to
 .Xr execve 2 .
@@ -165,9 +163,7 @@ otherwise \-1.
 .El
 .Pp
 These values also exist, but may not be standard:
-.Pp
 .Bl -tag -width 6n
-.Pp
 .It Li _SC_PHYS_PAGES
 The number of pages of physical memory.
 Note that it is possible that the product of this value and the value of
diff --git a/lib/libc/gen/sysctl.3 b/lib/libc/gen/sysctl.3
index 143be16c436..b7ed09139da 100644
--- a/lib/libc/gen/sysctl.3
+++ b/lib/libc/gen/sysctl.3
@@ -286,7 +286,6 @@ privilege may change the value.
 .It "HW_MACHINE_ARCH	string	no"
 .It "HW_REALMEM	integer	no"
 .El
-.Pp
 .Bl -tag -width 6n
 .It Li HW_MACHINE
 The machine class.
@@ -352,7 +351,6 @@ information.
 .It "KERN_VERSION	string	no"
 .It "KERN_VNODE	struct vnode	no"
 .El
-.Pp
 .Bl -tag -width 6n
 .It Li KERN_ARGMAX
 The maximum bytes of argument to
@@ -543,7 +541,6 @@ privilege may change the value.
 .It "PF_INET	IPv4 values	yes"
 .It "PF_INET6	IPv6 values	yes"
 .El
-.Pp
 .Bl -tag -width 6n
 .It Li PF_ROUTE
 Return the entire routing table or a subset of it.
@@ -650,7 +647,6 @@ privilege may change the value.
 .It "USER_TZNAME_MAX	integer	no"
 .El
 .Bl -tag -width 6n
-.Pp
 .It Li USER_BC_BASE_MAX
 The maximum ibase/obase values in the
 .Xr bc 1
@@ -740,7 +736,6 @@ privilege may change the value.
 .It "VM_V_INACTIVE_TARGET	integer	yes"
 .It "VM_V_PAGEOUT_FREE_MIN	integer	yes"
 .El
-.Pp
 .Bl -tag -width 6n
 .It Li VM_LOADAVG
 Return the load average history.
diff --git a/lib/libc/locale/isalnum.3 b/lib/libc/locale/isalnum.3
index 624a6e2335d..038c5cf413c 100644
--- a/lib/libc/locale/isalnum.3
+++ b/lib/libc/locale/isalnum.3
@@ -59,7 +59,6 @@ or the value of
 .Pp
 In the ASCII character set, this includes the following characters
 (with their numeric values shown in octal):
-.Pp
 .Bl -column \&000_``0''__ \&000_``0''__ \&000_``0''__ \&000_``0''__ \&000_``0''__
 .It "\&060\ ``0'' \t061\ ``1'' \t062\ ``2'' \t063\ ``3'' \t064\ ``4''"
 .It "\&065\ ``5'' \t066\ ``6'' \t067\ ``7'' \t070\ ``8'' \t071\ ``9''"
diff --git a/lib/libc/locale/isalpha.3 b/lib/libc/locale/isalpha.3
index 0bd3cda3c72..689b6cd54ad 100644
--- a/lib/libc/locale/isalpha.3
+++ b/lib/libc/locale/isalpha.3
@@ -59,7 +59,6 @@ or the value of
 .Pp
 In the ASCII character set, this includes the following characters
 (with their numeric values shown in octal):
-.Pp
 .Bl -column \&000_``0''__ \&000_``0''__ \&000_``0''__ \&000_``0''__ \&000_``0''__
 .It "\&101\ ``A'' \t102\ ``B'' \t103\ ``C'' \t104\ ``D'' \t105\ ``E''"
 .It "\&106\ ``F'' \t107\ ``G'' \t110\ ``H'' \t111\ ``I'' \t112\ ``J''"

From e03aa09b619c8844d1092202f0806464f10b98e5 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ulrich=20Sp=C3=B6rlein?= <uqs@FreeBSD.org>
Date: Wed, 14 Apr 2010 18:29:34 +0000
Subject: [PATCH 042/532] mdoc: don't abuse Bo/Pc to get what looks like an
 interval

Be explicit and use the general bracketing form plus symbols which are
to be interpreted mathematically in this case.

Complaint by:	mdocml
Reviewed by:	ru
Approved by:	philip, ed (mentors)
---
 lib/libc/gen/frexp.3 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/libc/gen/frexp.3 b/lib/libc/gen/frexp.3
index 390da750d79..ba33d3dfeba 100644
--- a/lib/libc/gen/frexp.3
+++ b/lib/libc/gen/frexp.3
@@ -70,7 +70,7 @@ such that
 is a
 .Vt double
 with magnitude in the interval
-.Bo 1/2 , 1 Pc
+.Eo [ 1/2 , 1 Ec )
 or zero, and
 .Fa value
 equals

From 20c3b3fa1c5dd984f6b4ef0a7f2b2b81142be023 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 14 Apr 2010 18:29:40 +0000
Subject: [PATCH 043/532] Make this code a little more portable by wrapping the
 mtx calls into macros.

MFC after:	1 week
---
 sys/net80211/ieee80211_freebsd.h  | 10 +++++++
 sys/net80211/ieee80211_scan_sta.c | 46 +++++++++++++++----------------
 2 files changed, 33 insertions(+), 23 deletions(-)

diff --git a/sys/net80211/ieee80211_freebsd.h b/sys/net80211/ieee80211_freebsd.h
index 861ac5153e3..abfc2c0a26e 100644
--- a/sys/net80211/ieee80211_freebsd.h
+++ b/sys/net80211/ieee80211_freebsd.h
@@ -147,6 +147,16 @@ typedef struct mtx acl_lock_t;
 #define	ACL_LOCK_ASSERT(_as) \
 	mtx_assert((&(_as)->as_lock), MA_OWNED)
 
+/*
+ * Scan table definitions.
+ */
+typedef struct mtx ieee80211_scan_table_lock_t;
+#define	IEEE80211_SCAN_TABLE_LOCK_INIT(_st, _name) \
+	mtx_init(&(_st)->st_lock, _name, "802.11 scan table", MTX_DEF)
+#define	IEEE80211_SCAN_TABLE_LOCK_DESTROY(_st)	mtx_destroy(&(_st)->st_lock)
+#define	IEEE80211_SCAN_TABLE_LOCK(_st)		mtx_lock(&(_st)->st_lock)
+#define	IEEE80211_SCAN_TABLE_UNLOCK(_st)	mtx_unlock(&(_st)->st_lock)
+
 /*
  * Node reference counting definitions.
  *
diff --git a/sys/net80211/ieee80211_scan_sta.c b/sys/net80211/ieee80211_scan_sta.c
index e697ad4a1a2..a4808e29671 100644
--- a/sys/net80211/ieee80211_scan_sta.c
+++ b/sys/net80211/ieee80211_scan_sta.c
@@ -97,7 +97,7 @@ struct sta_entry {
 CTASSERT(MAX_IEEE_CHAN >= 256);
 
 struct sta_table {
-	struct mtx	st_lock;		/* on scan table */
+	ieee80211_scan_table_lock_t st_lock;	/* on scan table */
 	TAILQ_HEAD(, sta_entry) st_entry;	/* all entries */
 	LIST_HEAD(, sta_entry) st_hash[STA_HASHSIZE];
 	struct mtx	st_scanlock;		/* on st_scaniter */
@@ -161,7 +161,7 @@ sta_attach(struct ieee80211_scan_state *ss)
 		M_80211_SCAN, M_NOWAIT | M_ZERO);
 	if (st == NULL)
 		return 0;
-	mtx_init(&st->st_lock, "scantable", "802.11 scan table", MTX_DEF);
+	IEEE80211_SCAN_TABLE_LOCK_INIT(st, "scantable");
 	mtx_init(&st->st_scanlock, "scangen", "802.11 scangen", MTX_DEF);
 	TAILQ_INIT(&st->st_entry);
 	ss->ss_priv = st;
@@ -179,7 +179,7 @@ sta_detach(struct ieee80211_scan_state *ss)
 
 	if (st != NULL) {
 		sta_flush_table(st);
-		mtx_destroy(&st->st_lock);
+		IEEE80211_SCAN_TABLE_LOCK_DESTROY(st);
 		mtx_destroy(&st->st_scanlock);
 		free(st, M_80211_SCAN);
 		KASSERT(nrefs > 0, ("imbalanced attach/detach"));
@@ -196,9 +196,9 @@ sta_flush(struct ieee80211_scan_state *ss)
 {
 	struct sta_table *st = ss->ss_priv;
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	sta_flush_table(st);
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 	ss->ss_last = 0;
 	return 0;
 }
@@ -244,14 +244,14 @@ sta_add(struct ieee80211_scan_state *ss,
 
 	hash = STA_HASH(macaddr);
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	LIST_FOREACH(se, &st->st_hash[hash], se_hash)
 		if (IEEE80211_ADDR_EQ(se->base.se_macaddr, macaddr))
 			goto found;
 	se = (struct sta_entry *) malloc(sizeof(struct sta_entry),
 		M_80211_SCAN, M_NOWAIT | M_ZERO);
 	if (se == NULL) {
-		mtx_unlock(&st->st_lock);
+		IEEE80211_SCAN_TABLE_UNLOCK(st);
 		return 0;
 	}
 	se->se_scangen = st->st_scaniter-1;
@@ -370,7 +370,7 @@ found:
 	if (rssi > st->st_maxrssi[sp->bchan])
 		st->st_maxrssi[sp->bchan] = rssi;
 
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 
 	/*
 	 * If looking for a quick choice and nothing's
@@ -1132,7 +1132,7 @@ sta_update_notseen(struct sta_table *st)
 {
 	struct sta_entry *se;
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	TAILQ_FOREACH(se, &st->st_entry, se_list) {
 		/*
 		 * If seen the reset and don't bump the count;
@@ -1146,7 +1146,7 @@ sta_update_notseen(struct sta_table *st)
 		else
 			se->se_notseen++;
 	}
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 }
 
 static void
@@ -1154,11 +1154,11 @@ sta_dec_fails(struct sta_table *st)
 {
 	struct sta_entry *se;
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	TAILQ_FOREACH(se, &st->st_entry, se_list)
 		if (se->se_fails)
 			se->se_fails--;
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 }
 
 static struct sta_entry *
@@ -1169,7 +1169,7 @@ select_bss(struct ieee80211_scan_state *ss, struct ieee80211vap *vap, int debug)
 
 	IEEE80211_DPRINTF(vap, debug, " %s\n",
 	    "macaddr          bssid         chan  rssi  rate flag  wep  essid");
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	TAILQ_FOREACH(se, &st->st_entry, se_list) {
 		ieee80211_ies_expand(&se->base.se_ies);
 		if (match_bss(vap, ss, se, debug) == 0) {
@@ -1179,7 +1179,7 @@ select_bss(struct ieee80211_scan_state *ss, struct ieee80211vap *vap, int debug)
 				selbs = se;
 		}
 	}
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 
 	return selbs;
 }
@@ -1258,11 +1258,11 @@ sta_lookup(struct sta_table *st, const uint8_t macaddr[IEEE80211_ADDR_LEN])
 	struct sta_entry *se;
 	int hash = STA_HASH(macaddr);
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	LIST_FOREACH(se, &st->st_hash[hash], se_hash)
 		if (IEEE80211_ADDR_EQ(se->base.se_macaddr, macaddr))
 			break;
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 
 	return se;		/* NB: unlocked */
 }
@@ -1382,18 +1382,18 @@ sta_iterate(struct ieee80211_scan_state *ss,
 	mtx_lock(&st->st_scanlock);
 	gen = st->st_scaniter++;
 restart:
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	TAILQ_FOREACH(se, &st->st_entry, se_list) {
 		if (se->se_scangen != gen) {
 			se->se_scangen = gen;
 			/* update public state */
 			se->base.se_age = ticks - se->se_lastupdate;
-			mtx_unlock(&st->st_lock);
+			IEEE80211_SCAN_TABLE_UNLOCK(st);
 			(*f)(arg, &se->base);
 			goto restart;
 		}
 	}
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 
 	mtx_unlock(&st->st_scanlock);
 }
@@ -1510,7 +1510,7 @@ adhoc_pick_channel(struct ieee80211_scan_state *ss, int flags)
 	bestchan = NULL;
 	bestrssi = -1;
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	for (i = 0; i < ss->ss_last; i++) {
 		c = ss->ss_chans[i];
 		/* never consider a channel with radar */
@@ -1532,7 +1532,7 @@ adhoc_pick_channel(struct ieee80211_scan_state *ss, int flags)
 		if (bestchan == NULL || maxrssi < bestrssi)
 			bestchan = c;
 	}
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 
 	return bestchan;
 }
@@ -1638,7 +1638,7 @@ adhoc_age(struct ieee80211_scan_state *ss)
 	struct sta_table *st = ss->ss_priv;
 	struct sta_entry *se, *next;
 
-	mtx_lock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_LOCK(st);
 	TAILQ_FOREACH_SAFE(se, &st->st_entry, se_list, next) {
 		if (se->se_notseen > STA_PURGE_SCANS) {
 			TAILQ_REMOVE(&st->st_entry, se, se_list);
@@ -1647,7 +1647,7 @@ adhoc_age(struct ieee80211_scan_state *ss)
 			free(se, M_80211_SCAN);
 		}
 	}
-	mtx_unlock(&st->st_lock);
+	IEEE80211_SCAN_TABLE_UNLOCK(st);
 }
 
 static const struct ieee80211_scanner adhoc_default = {

From f672e5fbe21dcf03aff493143e31d92fe6eb340d Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Wed, 14 Apr 2010 18:56:07 +0000
Subject: [PATCH 044/532] Add note about TARGET_ARCH

---
 Makefile.inc1 | 1 +
 1 file changed, 1 insertion(+)

diff --git a/Makefile.inc1 b/Makefile.inc1
index a772e2861e3..1c5d892468b 100644
--- a/Makefile.inc1
+++ b/Makefile.inc1
@@ -15,6 +15,7 @@
 #	-DNO_CTF do not run the DTrace CTF conversion tools on built objects
 #	LOCAL_DIRS="list of dirs" to add additional dirs to the SUBDIR list
 #	TARGET="machine" to crossbuild world for a different machine type
+#	TARGET_ARCH= may be required when a TARGET supports multiple endians
 
 #
 # The intended user-driven targets are:

From 8a07366ec284069c85972597ab0d305fad3e0d47 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Wed, 14 Apr 2010 19:03:27 +0000
Subject: [PATCH 045/532] Add armeb-*-freebsd* to the list of known
 architectures.  This is like arm-*-freebsd*, except it defaults to big endian
 builds instead of little endian builds.

---
 contrib/binutils/bfd/config.bfd | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/contrib/binutils/bfd/config.bfd b/contrib/binutils/bfd/config.bfd
index 034da18ef8b..5f115f0db2c 100755
--- a/contrib/binutils/bfd/config.bfd
+++ b/contrib/binutils/bfd/config.bfd
@@ -221,6 +221,10 @@ case "${targ}" in
     targ_defvec=bfd_elf32_littlearm_vec
     targ_selvecs=bfd_elf32_bigarm_vec
     ;;
+  armeb-*-freebsd*)
+    targ_defvec=bfd_elf32_bigarm_vec
+    targ_selvecs=bfd_elf32_littlearm_vec
+    ;;
   arm-*-elf | arm-*-freebsd* | arm*-*-linux-gnu* | arm*-*-conix* | \
   arm*-*-uclinux* | arm-*-kfreebsd*-gnu | arm-*-vxworks)
     targ_defvec=bfd_elf32_littlearm_vec

From c5ff2cdb99c0a70595a3d13be9c5d88741b5d3af Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ulrich=20Sp=C3=B6rlein?= <uqs@FreeBSD.org>
Date: Wed, 14 Apr 2010 19:06:39 +0000
Subject: [PATCH 046/532] Add and expand $FreeBSD$ keyword to allow committing
 to this file.

---
 usr.bin/csup/cpasswd.1 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.bin/csup/cpasswd.1 b/usr.bin/csup/cpasswd.1
index 946783f4c25..973dc808f28 100644
--- a/usr.bin/csup/cpasswd.1
+++ b/usr.bin/csup/cpasswd.1
@@ -27,7 +27,7 @@
 .\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .\"
 .\" $Id: cvpasswd.1,v 1.4 2003/03/04 18:24:42 jdp Exp $
-.\" $FreeBSD $
+.\" $FreeBSD$
 .\"
 .Dd June 27, 2007
 .Os FreeBSD

From aa12cea2ccc6e686d6d31cf67d6bc69cbc1ba744 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Ulrich=20Sp=C3=B6rlein?= <uqs@FreeBSD.org>
Date: Wed, 14 Apr 2010 19:08:06 +0000
Subject: [PATCH 047/532] mdoc: order prologue macros consistently by Dd/Dt/Os

Although groff_mdoc(7) gives another impression, this is the ordering
most widely used and also required by mdocml/mandoc.

Reviewed by:	ru
Approved by:	philip, ed (mentors)
---
 bin/pwait/pwait.1                          | 2 +-
 contrib/com_err/compile_et.1               | 2 +-
 gnu/usr.bin/gdb/kgdb/kgdb.1                | 2 +-
 lib/libc/gen/check_utility_compat.3        | 2 +-
 lib/libc/gen/dladdr.3                      | 2 +-
 lib/libc/gen/dlinfo.3                      | 2 +-
 lib/libc/gen/dllockinit.3                  | 2 +-
 lib/libc/gen/dlopen.3                      | 2 +-
 lib/libc/gen/fmtcheck.3                    | 2 +-
 lib/libc/gen/ftok.3                        | 2 +-
 lib/libc/gen/getutxent.3                   | 2 +-
 lib/libc/gen/setproctitle.3                | 2 +-
 lib/libc/gen/stringlist.3                  | 2 +-
 lib/libc/posix1e/mac_prepare.3             | 2 +-
 lib/libc/stdlib/hcreate.3                  | 2 +-
 lib/libc/stdlib/ptsname.3                  | 2 +-
 lib/libcam/cam.3                           | 2 +-
 lib/libcam/cam_cdbparse.3                  | 2 +-
 lib/libcompat/4.1/cftime.3                 | 2 +-
 lib/libcompat/4.4/cuserid.3                | 2 +-
 lib/libedit/editline.3                     | 2 +-
 lib/libedit/editrc.5                       | 2 +-
 lib/libelf/elf.3                           | 2 +-
 lib/libelf/elf_begin.3                     | 2 +-
 lib/libelf/elf_cntl.3                      | 2 +-
 lib/libelf/elf_end.3                       | 2 +-
 lib/libelf/elf_errmsg.3                    | 2 +-
 lib/libelf/elf_fill.3                      | 2 +-
 lib/libelf/elf_flagdata.3                  | 2 +-
 lib/libelf/elf_getarhdr.3                  | 2 +-
 lib/libelf/elf_getarsym.3                  | 2 +-
 lib/libelf/elf_getbase.3                   | 2 +-
 lib/libelf/elf_getdata.3                   | 2 +-
 lib/libelf/elf_getident.3                  | 2 +-
 lib/libelf/elf_getphnum.3                  | 2 +-
 lib/libelf/elf_getscn.3                    | 2 +-
 lib/libelf/elf_getshnum.3                  | 2 +-
 lib/libelf/elf_getshstrndx.3               | 2 +-
 lib/libelf/elf_hash.3                      | 2 +-
 lib/libelf/elf_kind.3                      | 2 +-
 lib/libelf/elf_memory.3                    | 2 +-
 lib/libelf/elf_next.3                      | 2 +-
 lib/libelf/elf_rand.3                      | 2 +-
 lib/libelf/elf_rawfile.3                   | 2 +-
 lib/libelf/elf_strptr.3                    | 2 +-
 lib/libelf/elf_update.3                    | 2 +-
 lib/libelf/elf_version.3                   | 2 +-
 lib/libelf/gelf.3                          | 2 +-
 lib/libelf/gelf_checksum.3                 | 2 +-
 lib/libelf/gelf_fsize.3                    | 2 +-
 lib/libelf/gelf_getcap.3                   | 2 +-
 lib/libelf/gelf_getclass.3                 | 2 +-
 lib/libelf/gelf_getdyn.3                   | 2 +-
 lib/libelf/gelf_getehdr.3                  | 2 +-
 lib/libelf/gelf_getmove.3                  | 2 +-
 lib/libelf/gelf_getphdr.3                  | 2 +-
 lib/libelf/gelf_getrel.3                   | 2 +-
 lib/libelf/gelf_getrela.3                  | 2 +-
 lib/libelf/gelf_getshdr.3                  | 2 +-
 lib/libelf/gelf_getsym.3                   | 2 +-
 lib/libelf/gelf_getsyminfo.3               | 2 +-
 lib/libelf/gelf_getsymshndx.3              | 2 +-
 lib/libelf/gelf_newehdr.3                  | 2 +-
 lib/libelf/gelf_newphdr.3                  | 2 +-
 lib/libelf/gelf_update_ehdr.3              | 2 +-
 lib/libelf/gelf_xlatetof.3                 | 2 +-
 lib/libgssapi/gss_accept_sec_context.3     | 2 +-
 lib/libgssapi/gss_acquire_cred.3           | 2 +-
 lib/libgssapi/gss_add_cred.3               | 2 +-
 lib/libgssapi/gss_add_oid_set_member.3     | 2 +-
 lib/libgssapi/gss_canonicalize_name.3      | 2 +-
 lib/libgssapi/gss_compare_name.3           | 2 +-
 lib/libgssapi/gss_context_time.3           | 2 +-
 lib/libgssapi/gss_create_empty_oid_set.3   | 2 +-
 lib/libgssapi/gss_delete_sec_context.3     | 2 +-
 lib/libgssapi/gss_display_name.3           | 2 +-
 lib/libgssapi/gss_display_status.3         | 2 +-
 lib/libgssapi/gss_duplicate_name.3         | 2 +-
 lib/libgssapi/gss_export_name.3            | 2 +-
 lib/libgssapi/gss_export_sec_context.3     | 2 +-
 lib/libgssapi/gss_get_mic.3                | 2 +-
 lib/libgssapi/gss_import_name.3            | 2 +-
 lib/libgssapi/gss_import_sec_context.3     | 2 +-
 lib/libgssapi/gss_indicate_mechs.3         | 2 +-
 lib/libgssapi/gss_init_sec_context.3       | 2 +-
 lib/libgssapi/gss_inquire_context.3        | 2 +-
 lib/libgssapi/gss_inquire_cred.3           | 2 +-
 lib/libgssapi/gss_inquire_cred_by_mech.3   | 2 +-
 lib/libgssapi/gss_inquire_mechs_for_name.3 | 2 +-
 lib/libgssapi/gss_inquire_names_for_mech.3 | 2 +-
 lib/libgssapi/gss_process_context_token.3  | 2 +-
 lib/libgssapi/gss_release_buffer.3         | 2 +-
 lib/libgssapi/gss_release_cred.3           | 2 +-
 lib/libgssapi/gss_release_name.3           | 2 +-
 lib/libgssapi/gss_release_oid_set.3        | 2 +-
 lib/libgssapi/gss_test_oid_set_member.3    | 2 +-
 lib/libgssapi/gss_unwrap.3                 | 2 +-
 lib/libgssapi/gss_verify_mic.3             | 2 +-
 lib/libgssapi/gss_wrap.3                   | 2 +-
 lib/libgssapi/gss_wrap_size_limit.3        | 2 +-
 lib/libmemstat/libmemstat.3                | 2 +-
 lib/libpmc/pmc.3                           | 2 +-
 lib/libpmc/pmc.atom.3                      | 2 +-
 lib/libpmc/pmc.core.3                      | 2 +-
 lib/libpmc/pmc.core2.3                     | 2 +-
 lib/libpmc/pmc.corei7.3                    | 2 +-
 lib/libpmc/pmc.corei7uc.3                  | 2 +-
 lib/libpmc/pmc.iaf.3                       | 2 +-
 lib/libpmc/pmc.k7.3                        | 2 +-
 lib/libpmc/pmc.k8.3                        | 2 +-
 lib/libpmc/pmc.p4.3                        | 2 +-
 lib/libpmc/pmc.p5.3                        | 2 +-
 lib/libpmc/pmc.p6.3                        | 2 +-
 lib/libpmc/pmc.tsc.3                       | 2 +-
 lib/libpmc/pmc.ucf.3                       | 2 +-
 lib/libpmc/pmc.westmere.3                  | 2 +-
 lib/libpmc/pmc.westmereuc.3                | 2 +-
 lib/libpmc/pmc_allocate.3                  | 2 +-
 lib/libpmc/pmc_attach.3                    | 2 +-
 lib/libpmc/pmc_capabilities.3              | 2 +-
 lib/libpmc/pmc_configure_logfile.3         | 2 +-
 lib/libpmc/pmc_disable.3                   | 2 +-
 lib/libpmc/pmc_event_names_of_class.3      | 2 +-
 lib/libpmc/pmc_get_driver_stats.3          | 2 +-
 lib/libpmc/pmc_get_msr.3                   | 2 +-
 lib/libpmc/pmc_init.3                      | 2 +-
 lib/libpmc/pmc_name_of_capability.3        | 2 +-
 lib/libpmc/pmc_read.3                      | 2 +-
 lib/libpmc/pmc_set.3                       | 2 +-
 lib/libpmc/pmc_start.3                     | 2 +-
 lib/libpmc/pmclog.3                        | 2 +-
 lib/libthr/libthr.3                        | 2 +-
 lib/libugidfw/bsde_get_rule.3              | 2 +-
 lib/libugidfw/bsde_get_rule_count.3        | 2 +-
 lib/libugidfw/bsde_parse_rule.3            | 2 +-
 lib/libugidfw/bsde_rule_to_string.3        | 2 +-
 lib/libugidfw/libugidfw.3                  | 2 +-
 lib/libulog/ulog_login.3                   | 2 +-
 lib/libulog/utempter_add_record.3          | 2 +-
 lib/libutil/_secure_path.3                 | 2 +-
 lib/libutil/auth.3                         | 2 +-
 lib/libutil/hexdump.3                      | 2 +-
 lib/libutil/kinfo_getfile.3                | 2 +-
 lib/libutil/kinfo_getvmmap.3               | 2 +-
 lib/libutil/kld.3                          | 2 +-
 lib/libutil/login_auth.3                   | 2 +-
 lib/libutil/login_cap.3                    | 2 +-
 lib/libutil/login_class.3                  | 2 +-
 lib/libutil/login_ok.3                     | 2 +-
 lib/libutil/login_times.3                  | 2 +-
 lib/libutil/login_tty.3                    | 2 +-
 lib/libutil/property.3                     | 2 +-
 lib/libutil/pty.3                          | 2 +-
 lib/libutil/realhostname.3                 | 2 +-
 lib/libutil/realhostname_sa.3              | 2 +-
 lib/libutil/trimdomain.3                   | 2 +-
 lib/libutil/uucplock.3                     | 2 +-
 sbin/iscontrol/iscsi.conf.5                | 2 +-
 sbin/spppcontrol/spppcontrol.8             | 2 +-
 share/man/man3/sysexits.3                  | 2 +-
 share/man/man3/tgmath.3                    | 2 +-
 share/man/man4/audit.4                     | 2 +-
 share/man/man4/auditpipe.4                 | 2 +-
 share/man/man4/coda.4                      | 2 +-
 share/man/man4/gbde.4                      | 2 +-
 share/man/man4/geom.4                      | 2 +-
 share/man/man4/geom_fox.4                  | 2 +-
 share/man/man4/geom_linux_lvm.4            | 2 +-
 share/man/man4/geom_uzip.4                 | 2 +-
 share/man/man4/ipw.4                       | 2 +-
 share/man/man4/iscsi_initiator.4           | 2 +-
 share/man/man4/iwi.4                       | 2 +-
 share/man/man4/iwn.4                       | 2 +-
 share/man/man4/kbdmux.4                    | 2 +-
 share/man/man4/lp.4                        | 2 +-
 share/man/man4/mac.4                       | 2 +-
 share/man/man4/mac_biba.4                  | 2 +-
 share/man/man4/mac_bsdextended.4           | 2 +-
 share/man/man4/mac_ifoff.4                 | 2 +-
 share/man/man4/mac_lomac.4                 | 2 +-
 share/man/man4/mac_mls.4                   | 2 +-
 share/man/man4/mac_none.4                  | 2 +-
 share/man/man4/mac_partition.4             | 2 +-
 share/man/man4/mac_seeotheruids.4          | 2 +-
 share/man/man4/mac_stub.4                  | 2 +-
 share/man/man4/mac_test.4                  | 2 +-
 share/man/man4/ng_netflow.4                | 2 +-
 share/man/man4/orm.4                       | 2 +-
 share/man/man4/ral.4                       | 2 +-
 share/man/man4/rp.4                        | 2 +-
 share/man/man4/rum.4                       | 2 +-
 share/man/man4/run.4                       | 2 +-
 share/man/man4/sched_4bsd.4                | 2 +-
 share/man/man4/sched_ule.4                 | 2 +-
 share/man/man4/si.4                        | 2 +-
 share/man/man4/tap.4                       | 2 +-
 share/man/man4/uhso.4                      | 2 +-
 share/man/man4/upgt.4                      | 2 +-
 share/man/man4/ural.4                      | 2 +-
 share/man/man4/vkbd.4                      | 2 +-
 share/man/man4/wpi.4                       | 2 +-
 share/man/man5/ar.5                        | 2 +-
 share/man/man7/clocks.7                    | 2 +-
 share/man/man7/maclabel.7                  | 2 +-
 share/man/man8/picobsd.8                   | 2 +-
 share/man/man8/rescue.8                    | 2 +-
 share/man/man9/CTASSERT.9                  | 2 +-
 share/man/man9/DELAY.9                     | 2 +-
 share/man/man9/KASSERT.9                   | 2 +-
 share/man/man9/VFS.9                       | 2 +-
 share/man/man9/VFS_CHECKEXP.9              | 2 +-
 share/man/man9/VFS_FHTOVP.9                | 2 +-
 share/man/man9/VFS_MOUNT.9                 | 2 +-
 share/man/man9/VFS_QUOTACTL.9              | 2 +-
 share/man/man9/VFS_ROOT.9                  | 2 +-
 share/man/man9/VFS_STATFS.9                | 2 +-
 share/man/man9/VFS_SYNC.9                  | 2 +-
 share/man/man9/VFS_UNMOUNT.9               | 2 +-
 share/man/man9/VFS_VGET.9                  | 2 +-
 share/man/man9/VOP_ACCESS.9                | 2 +-
 share/man/man9/VOP_ACLCHECK.9              | 2 +-
 share/man/man9/VOP_ADVLOCK.9               | 2 +-
 share/man/man9/VOP_ATTRIB.9                | 2 +-
 share/man/man9/VOP_BWRITE.9                | 2 +-
 share/man/man9/VOP_CREATE.9                | 2 +-
 share/man/man9/VOP_FSYNC.9                 | 2 +-
 share/man/man9/VOP_GETACL.9                | 2 +-
 share/man/man9/VOP_GETEXTATTR.9            | 2 +-
 share/man/man9/VOP_GETPAGES.9              | 2 +-
 share/man/man9/VOP_GETVOBJECT.9            | 2 +-
 share/man/man9/VOP_INACTIVE.9              | 2 +-
 share/man/man9/VOP_IOCTL.9                 | 2 +-
 share/man/man9/VOP_LINK.9                  | 2 +-
 share/man/man9/VOP_LISTEXTATTR.9           | 2 +-
 share/man/man9/VOP_LOCK.9                  | 2 +-
 share/man/man9/VOP_LOOKUP.9                | 2 +-
 share/man/man9/VOP_OPENCLOSE.9             | 2 +-
 share/man/man9/VOP_PATHCONF.9              | 2 +-
 share/man/man9/VOP_PRINT.9                 | 2 +-
 share/man/man9/VOP_RDWR.9                  | 2 +-
 share/man/man9/VOP_READDIR.9               | 2 +-
 share/man/man9/VOP_READLINK.9              | 2 +-
 share/man/man9/VOP_REALLOCBLKS.9           | 2 +-
 share/man/man9/VOP_REMOVE.9                | 2 +-
 share/man/man9/VOP_RENAME.9                | 2 +-
 share/man/man9/VOP_REVOKE.9                | 2 +-
 share/man/man9/VOP_SETACL.9                | 2 +-
 share/man/man9/VOP_SETEXTATTR.9            | 2 +-
 share/man/man9/VOP_STRATEGY.9              | 2 +-
 share/man/man9/VOP_VPTOCNP.9               | 2 +-
 share/man/man9/VOP_VPTOFH.9                | 2 +-
 share/man/man9/accept_filter.9             | 2 +-
 share/man/man9/accf_data.9                 | 2 +-
 share/man/man9/accf_dns.9                  | 2 +-
 share/man/man9/accf_http.9                 | 2 +-
 share/man/man9/acl.9                       | 2 +-
 share/man/man9/atomic.9                    | 2 +-
 share/man/man9/cr_cansee.9                 | 2 +-
 share/man/man9/cr_seeothergids.9           | 2 +-
 share/man/man9/cr_seeotheruids.9           | 2 +-
 share/man/man9/devfs_set_cdevpriv.9        | 2 +-
 share/man/man9/devtoname.9                 | 2 +-
 share/man/man9/extattr.9                   | 2 +-
 share/man/man9/firmware.9                  | 2 +-
 share/man/man9/hexdump.9                   | 2 +-
 share/man/man9/ifnet.9                     | 2 +-
 share/man/man9/make_dev.9                  | 2 +-
 share/man/man9/namei.9                     | 2 +-
 share/man/man9/p_candebug.9                | 2 +-
 share/man/man9/p_cansee.9                  | 2 +-
 share/man/man9/pfind.9                     | 2 +-
 share/man/man9/pgfind.9                    | 2 +-
 share/man/man9/prison_check.9              | 2 +-
 share/man/man9/random.9                    | 2 +-
 share/man/man9/rijndael.9                  | 2 +-
 share/man/man9/rtalloc.9                   | 2 +-
 share/man/man9/rtentry.9                   | 2 +-
 share/man/man9/sleep.9                     | 2 +-
 share/man/man9/spl.9                       | 2 +-
 share/man/man9/uio.9                       | 2 +-
 share/man/man9/usbdi.9                     | 2 +-
 share/man/man9/vaccess.9                   | 2 +-
 share/man/man9/vaccess_acl_nfs4.9          | 2 +-
 share/man/man9/vaccess_acl_posix1e.9       | 2 +-
 share/man/man9/vcount.9                    | 2 +-
 share/man/man9/vget.9                      | 2 +-
 share/man/man9/vm_map_entry_resize_free.9  | 2 +-
 share/man/man9/vnode.9                     | 2 +-
 share/man/man9/vput.9                      | 2 +-
 share/man/man9/vref.9                      | 2 +-
 share/man/man9/vrefcnt.9                   | 2 +-
 share/man/man9/vrele.9                     | 2 +-
 usr.bin/ar/ar.1                            | 2 +-
 usr.bin/c89/c89.1                          | 2 +-
 usr.bin/c99/c99.1                          | 2 +-
 usr.bin/column/column.1                    | 2 +-
 usr.bin/comm/comm.1                        | 2 +-
 usr.bin/csup/cpasswd.1                     | 2 +-
 usr.bin/csup/csup.1                        | 2 +-
 usr.bin/enigma/enigma.1                    | 2 +-
 usr.bin/hexdump/od.1                       | 2 +-
 usr.bin/killall/killall.1                  | 2 +-
 usr.bin/lockf/lockf.1                      | 2 +-
 usr.bin/wtmpcvt/wtmpcvt.1                  | 2 +-
 usr.sbin/asf/asf.8                         | 2 +-
 usr.sbin/burncd/burncd.8                   | 2 +-
 usr.sbin/ctm/ctm/ctm.1                     | 2 +-
 usr.sbin/ctm/ctm/ctm.5                     | 2 +-
 usr.sbin/devinfo/devinfo.8                 | 2 +-
 usr.sbin/fdformat/fdformat.1               | 2 +-
 usr.sbin/fdread/fdread.1                   | 2 +-
 usr.sbin/fdwrite/fdwrite.1                 | 2 +-
 usr.sbin/fifolog/fifolog_create/fifolog.1  | 2 +-
 usr.sbin/flowctl/flowctl.8                 | 2 +-
 usr.sbin/mtest/mtest.8                     | 2 +-
 usr.sbin/periodic/periodic.8               | 2 +-
 usr.sbin/pmcannotate/pmcannotate.8         | 2 +-
 usr.sbin/pmccontrol/pmccontrol.8           | 2 +-
 usr.sbin/pmcstat/pmcstat.8                 | 2 +-
 usr.sbin/pppctl/pppctl.8                   | 2 +-
 usr.sbin/setfmac/setfsmac.8                | 2 +-
 usr.sbin/setpmac/setpmac.8                 | 2 +-
 usr.sbin/uhsoctl/uhsoctl.1                 | 2 +-
 323 files changed, 323 insertions(+), 323 deletions(-)

diff --git a/bin/pwait/pwait.1 b/bin/pwait/pwait.1
index 132d83c98ce..f36090d1ebf 100644
--- a/bin/pwait/pwait.1
+++ b/bin/pwait/pwait.1
@@ -33,8 +33,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 1, 2009
-.Os
 .Dt PWAIT 1
+.Os
 .Sh NAME
 .Nm pwait
 .Nd wait for processes to terminate
diff --git a/contrib/com_err/compile_et.1 b/contrib/com_err/compile_et.1
index bdc608ba6dd..32f2eeece20 100644
--- a/contrib/com_err/compile_et.1
+++ b/contrib/com_err/compile_et.1
@@ -4,8 +4,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 22, 1988
-.Os
 .Dt COMPILE_ET 1
+.Os
 .Sh NAME
 .Nm compile_et
 .Nd error table compiler
diff --git a/gnu/usr.bin/gdb/kgdb/kgdb.1 b/gnu/usr.bin/gdb/kgdb/kgdb.1
index 5b6fea0bace..4073d705f8b 100644
--- a/gnu/usr.bin/gdb/kgdb/kgdb.1
+++ b/gnu/usr.bin/gdb/kgdb/kgdb.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 11, 2006
-.Os
 .Dt KGDB 1
+.Os
 .Sh NAME
 .Nm kgdb
 .Nd "kernel debugger"
diff --git a/lib/libc/gen/check_utility_compat.3 b/lib/libc/gen/check_utility_compat.3
index 7a96bac15c5..57638e50ead 100644
--- a/lib/libc/gen/check_utility_compat.3
+++ b/lib/libc/gen/check_utility_compat.3
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 27, 2002
-.Os
 .Dt CHECK_UTILITY_COMPAT 3
+.Os
 .Sh NAME
 .Nm check_utility_compat
 .Nd "determine whether a utility should be compatible"
diff --git a/lib/libc/gen/dladdr.3 b/lib/libc/gen/dladdr.3
index 414fa49a3e9..65662793b59 100644
--- a/lib/libc/gen/dladdr.3
+++ b/lib/libc/gen/dladdr.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 5, 1998
-.Os
 .Dt DLADDR 3
+.Os
 .Sh NAME
 .Nm dladdr
 .Nd find the shared object containing a given address
diff --git a/lib/libc/gen/dlinfo.3 b/lib/libc/gen/dlinfo.3
index 9433fb6e9d1..d00f07467af 100644
--- a/lib/libc/gen/dlinfo.3
+++ b/lib/libc/gen/dlinfo.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 14, 2003
-.Os
 .Dt DLINFO 3
+.Os
 .Sh NAME
 .Nm dlinfo
 .Nd information about dynamically loaded object
diff --git a/lib/libc/gen/dllockinit.3 b/lib/libc/gen/dllockinit.3
index 98d10749485..be3e2cab87f 100644
--- a/lib/libc/gen/dllockinit.3
+++ b/lib/libc/gen/dllockinit.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 5, 2000
-.Os
 .Dt DLLOCKINIT 3
+.Os
 .Sh NAME
 .Nm dllockinit
 .Nd register thread locking methods with the dynamic linker
diff --git a/lib/libc/gen/dlopen.3 b/lib/libc/gen/dlopen.3
index d4b111a1002..45b01397468 100644
--- a/lib/libc/gen/dlopen.3
+++ b/lib/libc/gen/dlopen.3
@@ -33,8 +33,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 7, 2009
-.Os
 .Dt DLOPEN 3
+.Os
 .Sh NAME
 .Nm dlopen ,
 .Nm dlsym ,
diff --git a/lib/libc/gen/fmtcheck.3 b/lib/libc/gen/fmtcheck.3
index ecd70fb8b30..2fa587b6bd9 100644
--- a/lib/libc/gen/fmtcheck.3
+++ b/lib/libc/gen/fmtcheck.3
@@ -26,8 +26,8 @@
 .\"
 .\" $FreeBSD$
 .Dd October 16, 2002
-.Os
 .Dt FMTCHECK 3
+.Os
 .Sh NAME
 .Nm fmtcheck
 .Nd sanitizes user-supplied
diff --git a/lib/libc/gen/ftok.3 b/lib/libc/gen/ftok.3
index 78bfe457c3d..996b2da41c7 100644
--- a/lib/libc/gen/ftok.3
+++ b/lib/libc/gen/ftok.3
@@ -25,8 +25,8 @@
 .\"
 .\" $FreeBSD$
 .Dd July 9, 2009
-.Os
 .Dt FTOK 3
+.Os
 .Sh NAME
 .Nm ftok
 .Nd create IPC identifier from path name
diff --git a/lib/libc/gen/getutxent.3 b/lib/libc/gen/getutxent.3
index 5ab21eede74..ccc5d9ed5c3 100644
--- a/lib/libc/gen/getutxent.3
+++ b/lib/libc/gen/getutxent.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 8, 2010
-.Os
 .Dt GETUTXENT 3
+.Os
 .Sh NAME
 .Nm endutxent ,
 .Nm getutxent ,
diff --git a/lib/libc/gen/setproctitle.3 b/lib/libc/gen/setproctitle.3
index fe9f19acde6..73dc8c00cc7 100644
--- a/lib/libc/gen/setproctitle.3
+++ b/lib/libc/gen/setproctitle.3
@@ -21,8 +21,8 @@
 .\"
 .\" The following requests are required for all man pages.
 .Dd December 16, 1995
-.Os
 .Dt SETPROCTITLE 3
+.Os
 .Sh NAME
 .Nm setproctitle
 .Nd set process title
diff --git a/lib/libc/gen/stringlist.3 b/lib/libc/gen/stringlist.3
index 44d2a36ffa4..c55b77d7802 100644
--- a/lib/libc/gen/stringlist.3
+++ b/lib/libc/gen/stringlist.3
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 28, 1999
-.Os
 .Dt STRINGLIST 3
+.Os
 .Sh NAME
 .Nm stringlist ,
 .Nm sl_init ,
diff --git a/lib/libc/posix1e/mac_prepare.3 b/lib/libc/posix1e/mac_prepare.3
index 1d8522908f9..8e694de133b 100644
--- a/lib/libc/posix1e/mac_prepare.3
+++ b/lib/libc/posix1e/mac_prepare.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 22, 2003
-.Os
 .Dt MAC_PREPARE 3
+.Os
 .Sh NAME
 .Nm mac_prepare , mac_prepare_type , mac_prepare_file_label ,
 .Nm mac_prepare_ifnet_label , mac_prepare_process_label
diff --git a/lib/libc/stdlib/hcreate.3 b/lib/libc/stdlib/hcreate.3
index cd827207694..2466c9f428a 100644
--- a/lib/libc/stdlib/hcreate.3
+++ b/lib/libc/stdlib/hcreate.3
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 6, 2008
-.Os
 .Dt HCREATE 3
+.Os
 .Sh NAME
 .Nm hcreate , hdestroy , hsearch
 .Nd manage hash search table
diff --git a/lib/libc/stdlib/ptsname.3 b/lib/libc/stdlib/ptsname.3
index f674a54bd90..48da3bd81c2 100644
--- a/lib/libc/stdlib/ptsname.3
+++ b/lib/libc/stdlib/ptsname.3
@@ -32,8 +32,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 20, 2008
-.Os
 .Dt PTSNAME 3
+.Os
 .Sh NAME
 .Nm grantpt ,
 .Nm ptsname ,
diff --git a/lib/libcam/cam.3 b/lib/libcam/cam.3
index 539709f265a..3a857b2fd79 100644
--- a/lib/libcam/cam.3
+++ b/lib/libcam/cam.3
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 10, 1998
-.Os
 .Dt CAM 3
+.Os
 .Sh NAME
 .Nm cam_open_device ,
 .Nm cam_open_spec_device ,
diff --git a/lib/libcam/cam_cdbparse.3 b/lib/libcam/cam_cdbparse.3
index 891652dca92..35a595151f3 100644
--- a/lib/libcam/cam_cdbparse.3
+++ b/lib/libcam/cam_cdbparse.3
@@ -62,8 +62,8 @@
 .\"
 .\"
 .Dd October 13, 1998
-.Os
 .Dt CAM_CDBPARSE 3
+.Os
 .Sh NAME
 .Nm csio_build ,
 .Nm csio_build_visit ,
diff --git a/lib/libcompat/4.1/cftime.3 b/lib/libcompat/4.1/cftime.3
index 116f5de5dfa..dd21e781ed7 100644
--- a/lib/libcompat/4.1/cftime.3
+++ b/lib/libcompat/4.1/cftime.3
@@ -35,8 +35,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 15, 1993
-.Os
 .Dt CFTIME 3
+.Os
 .Sh NAME
 .Nm cftime ,
 .Nm ascftime
diff --git a/lib/libcompat/4.4/cuserid.3 b/lib/libcompat/4.4/cuserid.3
index 72811d5196d..addd61608ff 100644
--- a/lib/libcompat/4.4/cuserid.3
+++ b/lib/libcompat/4.4/cuserid.3
@@ -32,8 +32,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 10, 1995
-.Os
 .Dt CUSERID 3
+.Os
 .Sh NAME
 .Nm cuserid
 .Nd get user name associated with effective UID
diff --git a/lib/libedit/editline.3 b/lib/libedit/editline.3
index d5dfff5e7a4..9794649a56e 100644
--- a/lib/libedit/editline.3
+++ b/lib/libedit/editline.3
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 12, 2007
-.Os
 .Dt EDITLINE 3
+.Os
 .Sh NAME
 .Nm editline ,
 .Nm el_init ,
diff --git a/lib/libedit/editrc.5 b/lib/libedit/editrc.5
index 4b1d4148b8d..2fa1a634dea 100644
--- a/lib/libedit/editrc.5
+++ b/lib/libedit/editrc.5
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 18, 2003
-.Os
 .Dt EDITRC 5
+.Os
 .Sh NAME
 .Nm editrc
 .Nd configuration file for editline library
diff --git a/lib/libelf/elf.3 b/lib/libelf/elf.3
index 854e1af1a2f..c649efe72be 100644
--- a/lib/libelf/elf.3
+++ b/lib/libelf/elf.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 21, 2007
-.Os
 .Dt ELF 3
+.Os
 .Sh NAME
 .Nm elf
 .Nd API for manipulating ELF objects
diff --git a/lib/libelf/elf_begin.3 b/lib/libelf/elf_begin.3
index c43766f7bcb..dac49102eaf 100644
--- a/lib/libelf/elf_begin.3
+++ b/lib/libelf/elf_begin.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 21, 2006
-.Os
 .Dt ELF_BEGIN 3
+.Os
 .Sh NAME
 .Nm elf_begin
 .Nd open an ELF file or ar(1) archive
diff --git a/lib/libelf/elf_cntl.3 b/lib/libelf/elf_cntl.3
index 73dca213944..52771701181 100644
--- a/lib/libelf/elf_cntl.3
+++ b/lib/libelf/elf_cntl.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 9, 2006
-.Os
 .Dt ELF_CNTL 3
+.Os
 .Sh NAME
 .Nm elf_cntl
 .Nd control an elf file descriptor
diff --git a/lib/libelf/elf_end.3 b/lib/libelf/elf_end.3
index 0f74a898472..3622fb2e670 100644
--- a/lib/libelf/elf_end.3
+++ b/lib/libelf/elf_end.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 29, 2006
-.Os
 .Dt ELF_END 3
+.Os
 .Sh NAME
 .Nm elf_end
 .Nd release an ELF descriptor
diff --git a/lib/libelf/elf_errmsg.3 b/lib/libelf/elf_errmsg.3
index 1ed70bf9ee6..f01416804b5 100644
--- a/lib/libelf/elf_errmsg.3
+++ b/lib/libelf/elf_errmsg.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 11, 2006
-.Os
 .Dt ELF_ERRMSG 3
+.Os
 .Sh NAME
 .Nm elf_errmsg ,
 .Nm elf_errno
diff --git a/lib/libelf/elf_fill.3 b/lib/libelf/elf_fill.3
index 33446b99c48..a635b435a6e 100644
--- a/lib/libelf/elf_fill.3
+++ b/lib/libelf/elf_fill.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 11, 2006
-.Os
 .Dt ELF_FILL 3
+.Os
 .Sh NAME
 .Nm elf_fill
 .Nd set fill byte for inter-section padding
diff --git a/lib/libelf/elf_flagdata.3 b/lib/libelf/elf_flagdata.3
index 30c1111904b..b07e5242681 100644
--- a/lib/libelf/elf_flagdata.3
+++ b/lib/libelf/elf_flagdata.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 22, 2007
-.Os
 .Dt ELF_FLAGDATA 3
+.Os
 .Sh NAME
 .Nm elf_flagdata ,
 .Nm elf_flagehdr ,
diff --git a/lib/libelf/elf_getarhdr.3 b/lib/libelf/elf_getarhdr.3
index f78fc80bc33..03f911adf3a 100644
--- a/lib/libelf/elf_getarhdr.3
+++ b/lib/libelf/elf_getarhdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 15, 2006
-.Os
 .Dt ELF_GETARHDR 3
+.Os
 .Sh NAME
 .Nm elf_getarhdr
 .Nd retrieve ar(1) header for an archive member
diff --git a/lib/libelf/elf_getarsym.3 b/lib/libelf/elf_getarsym.3
index 1030ea2b6f5..44b28c8c957 100644
--- a/lib/libelf/elf_getarsym.3
+++ b/lib/libelf/elf_getarsym.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 15, 2006
-.Os
 .Dt ELF_GETARSYM 3
+.Os
 .Sh NAME
 .Nm elf_getarsym
 .Nd retrieve the symbol table of an archive
diff --git a/lib/libelf/elf_getbase.3 b/lib/libelf/elf_getbase.3
index 30be6c5c259..be9d103f14e 100644
--- a/lib/libelf/elf_getbase.3
+++ b/lib/libelf/elf_getbase.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 11, 2006
-.Os
 .Dt ELF_GETBASE 3
+.Os
 .Sh NAME
 .Nm elf_getbase
 .Nd get the base offset for an object file
diff --git a/lib/libelf/elf_getdata.3 b/lib/libelf/elf_getdata.3
index 64da4ad9f8c..10f37639563 100644
--- a/lib/libelf/elf_getdata.3
+++ b/lib/libelf/elf_getdata.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 26, 2006
-.Os
 .Dt ELF_GETDATA 3
+.Os
 .Sh NAME
 .Nm elf_getdata ,
 .Nm elf_newdata ,
diff --git a/lib/libelf/elf_getident.3 b/lib/libelf/elf_getident.3
index f34ab70e373..1031e1d0ebc 100644
--- a/lib/libelf/elf_getident.3
+++ b/lib/libelf/elf_getident.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 3, 2006
-.Os
 .Dt ELF_GETIDENT 3
+.Os
 .Sh NAME
 .Nm elf_getident
 .Nd return the initial bytes of a file
diff --git a/lib/libelf/elf_getphnum.3 b/lib/libelf/elf_getphnum.3
index 606ceb0e755..308f87cd4cc 100644
--- a/lib/libelf/elf_getphnum.3
+++ b/lib/libelf/elf_getphnum.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 16, 2006
-.Os
 .Dt ELF_GETPHNUM 3
+.Os
 .Sh NAME
 .Nm elf_getphnum
 .Nd return the number of program headers in an ELF file
diff --git a/lib/libelf/elf_getscn.3 b/lib/libelf/elf_getscn.3
index 5b9de412a01..1b1b5f9ef73 100644
--- a/lib/libelf/elf_getscn.3
+++ b/lib/libelf/elf_getscn.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 22, 2007
-.Os
 .Dt ELF_GETSCN 3
+.Os
 .Sh NAME
 .Nm elf_getscn ,
 .Nm elf_ndxscn ,
diff --git a/lib/libelf/elf_getshnum.3 b/lib/libelf/elf_getshnum.3
index 10a370ce6db..7c8a04c26a5 100644
--- a/lib/libelf/elf_getshnum.3
+++ b/lib/libelf/elf_getshnum.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 31, 2006
-.Os
 .Dt ELF_GETSHNUM 3
+.Os
 .Sh NAME
 .Nm elf_getshnum
 .Nd return the number of sections in an ELF file
diff --git a/lib/libelf/elf_getshstrndx.3 b/lib/libelf/elf_getshstrndx.3
index 115b2dac1c6..71df3ac5192 100644
--- a/lib/libelf/elf_getshstrndx.3
+++ b/lib/libelf/elf_getshstrndx.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 31, 2006
-.Os
 .Dt ELF_GETSHSTRNDX 3
+.Os
 .Sh NAME
 .Nm elf_getshstrndx ,
 .Nm elf_setshstrndx
diff --git a/lib/libelf/elf_hash.3 b/lib/libelf/elf_hash.3
index 02095988959..0df2804e2fa 100644
--- a/lib/libelf/elf_hash.3
+++ b/lib/libelf/elf_hash.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 15, 2006
-.Os
 .Dt ELF_HASH 3
+.Os
 .Sh NAME
 .Nm elf_hash
 .Nd compute a hash value for a string
diff --git a/lib/libelf/elf_kind.3 b/lib/libelf/elf_kind.3
index def1582b229..de74e574993 100644
--- a/lib/libelf/elf_kind.3
+++ b/lib/libelf/elf_kind.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 1, 2006
-.Os
 .Dt ELF_KIND 3
+.Os
 .Sh NAME
 .Nm elf_kind
 .Nd determine ELF file type
diff --git a/lib/libelf/elf_memory.3 b/lib/libelf/elf_memory.3
index 9af6c5e30fa..83de5124b21 100644
--- a/lib/libelf/elf_memory.3
+++ b/lib/libelf/elf_memory.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 28, 2006
-.Os
 .Dt ELF_MEMORY 3
+.Os
 .Sh NAME
 .Nm elf_memory
 .Nd process an ELF or ar(1) archive mapped into memory
diff --git a/lib/libelf/elf_next.3 b/lib/libelf/elf_next.3
index da995fc253b..184ab9f44ee 100644
--- a/lib/libelf/elf_next.3
+++ b/lib/libelf/elf_next.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 17, 2006
-.Os
 .Dt ELF_NEXT 3
+.Os
 .Sh NAME
 .Nm elf_next
 .Nd provide sequential access to the next archive member
diff --git a/lib/libelf/elf_rand.3 b/lib/libelf/elf_rand.3
index 46cec6d2895..e4d23d761f1 100644
--- a/lib/libelf/elf_rand.3
+++ b/lib/libelf/elf_rand.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 17, 2006
-.Os
 .Dt ELF_RAND 3
+.Os
 .Sh NAME
 .Nm elf_rand
 .Nd provide sequential access to the next archive member
diff --git a/lib/libelf/elf_rawfile.3 b/lib/libelf/elf_rawfile.3
index 76026e7971b..4458148830b 100644
--- a/lib/libelf/elf_rawfile.3
+++ b/lib/libelf/elf_rawfile.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 3, 2006
-.Os
 .Dt ELF_RAWFILE 3
+.Os
 .Sh NAME
 .Nm elf_rawfile
 .Nd return uninterpreted contents of an ELF file
diff --git a/lib/libelf/elf_strptr.3 b/lib/libelf/elf_strptr.3
index bf24cc1f655..a0a9e649bb1 100644
--- a/lib/libelf/elf_strptr.3
+++ b/lib/libelf/elf_strptr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 16, 2006
-.Os
 .Dt ELF_STRPTR 3
+.Os
 .Sh NAME
 .Nm elf_strptr
 .Nd retrieve a string pointer in a string table
diff --git a/lib/libelf/elf_update.3 b/lib/libelf/elf_update.3
index 33a14d192bc..7f6f4f09ba5 100644
--- a/lib/libelf/elf_update.3
+++ b/lib/libelf/elf_update.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 19, 2008
-.Os
 .Dt ELF_UPDATE 3
+.Os
 .Sh NAME
 .Nm elf_update
 .Nd update an ELF descriptor
diff --git a/lib/libelf/elf_version.3 b/lib/libelf/elf_version.3
index 25ea62b231c..e9187106174 100644
--- a/lib/libelf/elf_version.3
+++ b/lib/libelf/elf_version.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 1, 2006
-.Os
 .Dt ELF_VERSION 3
+.Os
 .Sh NAME
 .Nm elf_version
 .Nd retrieve or set ELF library operating version
diff --git a/lib/libelf/gelf.3 b/lib/libelf/gelf.3
index 4ab5ff2867c..d541388aee3 100644
--- a/lib/libelf/gelf.3
+++ b/lib/libelf/gelf.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 1, 2006
-.Os
 .Dt GELF 3
+.Os
 .Sh NAME
 .Nm GElf
 .Nd class-independent API for ELF manipulation
diff --git a/lib/libelf/gelf_checksum.3 b/lib/libelf/gelf_checksum.3
index a6f19784b3c..99870f0af13 100644
--- a/lib/libelf/gelf_checksum.3
+++ b/lib/libelf/gelf_checksum.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_CHECKSUM 3
+.Os
 .Sh NAME
 .Nm elf32_checksum ,
 .Nm elf64_checksum ,
diff --git a/lib/libelf/gelf_fsize.3 b/lib/libelf/gelf_fsize.3
index 915d840b638..caefae755fe 100644
--- a/lib/libelf/gelf_fsize.3
+++ b/lib/libelf/gelf_fsize.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 5, 2008
-.Os
 .Dt GELF_FSIZE 3
+.Os
 .Sh NAME
 .Nm gelf_fsize ,
 .Nm elf32_fsize ,
diff --git a/lib/libelf/gelf_getcap.3 b/lib/libelf/gelf_getcap.3
index f6262d81a94..e63a2633753 100644
--- a/lib/libelf/gelf_getcap.3
+++ b/lib/libelf/gelf_getcap.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETCAP 3
+.Os
 .Sh NAME
 .Nm gelf_getcap ,
 .Nm gelf_update_cap
diff --git a/lib/libelf/gelf_getclass.3 b/lib/libelf/gelf_getclass.3
index cab3e24b832..7f7bc30ce20 100644
--- a/lib/libelf/gelf_getclass.3
+++ b/lib/libelf/gelf_getclass.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 3, 2006
-.Os
 .Dt GELF_GETCLASS 3
+.Os
 .Sh NAME
 .Nm gelf_getclass
 .Nd retrieve the class of an ELF descriptor
diff --git a/lib/libelf/gelf_getdyn.3 b/lib/libelf/gelf_getdyn.3
index 11244440ff8..a755d63a09f 100644
--- a/lib/libelf/gelf_getdyn.3
+++ b/lib/libelf/gelf_getdyn.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETDYN 3
+.Os
 .Sh NAME
 .Nm gelf_getdyn ,
 .Nm gelf_update_dyn
diff --git a/lib/libelf/gelf_getehdr.3 b/lib/libelf/gelf_getehdr.3
index d4c8cb34e53..8e71c2d4945 100644
--- a/lib/libelf/gelf_getehdr.3
+++ b/lib/libelf/gelf_getehdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 16, 2006
-.Os
 .Dt GELF_GETEHDR 3
+.Os
 .Sh NAME
 .Nm elf32_getehdr ,
 .Nm elf64_getehdr ,
diff --git a/lib/libelf/gelf_getmove.3 b/lib/libelf/gelf_getmove.3
index 11fd9cad955..abe398ab374 100644
--- a/lib/libelf/gelf_getmove.3
+++ b/lib/libelf/gelf_getmove.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETMOVE 3
+.Os
 .Sh NAME
 .Nm gelf_getmove ,
 .Nm gelf_update_move
diff --git a/lib/libelf/gelf_getphdr.3 b/lib/libelf/gelf_getphdr.3
index d6ffb906147..53fab17bba8 100644
--- a/lib/libelf/gelf_getphdr.3
+++ b/lib/libelf/gelf_getphdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 21, 2007
-.Os
 .Dt GELF_GETPHDR 3
+.Os
 .Sh NAME
 .Nm elf32_getphdr ,
 .Nm elf64_getphdr ,
diff --git a/lib/libelf/gelf_getrel.3 b/lib/libelf/gelf_getrel.3
index 75fa79e9822..6e650a5b7d2 100644
--- a/lib/libelf/gelf_getrel.3
+++ b/lib/libelf/gelf_getrel.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETREL 3
+.Os
 .Sh NAME
 .Nm gelf_getrel ,
 .Nm gelf_update_rel
diff --git a/lib/libelf/gelf_getrela.3 b/lib/libelf/gelf_getrela.3
index d8968f56129..dc7e5795cbe 100644
--- a/lib/libelf/gelf_getrela.3
+++ b/lib/libelf/gelf_getrela.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETRELA 3
+.Os
 .Sh NAME
 .Nm gelf_getrela ,
 .Nm gelf_update_rela
diff --git a/lib/libelf/gelf_getshdr.3 b/lib/libelf/gelf_getshdr.3
index c1823b54d52..851f51f00ef 100644
--- a/lib/libelf/gelf_getshdr.3
+++ b/lib/libelf/gelf_getshdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 27, 2006
-.Os
 .Dt GELF_GETSHDR 3
+.Os
 .Sh NAME
 .Nm elf32_getshdr ,
 .Nm elf64_getshdr ,
diff --git a/lib/libelf/gelf_getsym.3 b/lib/libelf/gelf_getsym.3
index 4e35133648d..56a6783acb1 100644
--- a/lib/libelf/gelf_getsym.3
+++ b/lib/libelf/gelf_getsym.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETSYM 3
+.Os
 .Sh NAME
 .Nm gelf_getsym ,
 .Nm gelf_update_sym
diff --git a/lib/libelf/gelf_getsyminfo.3 b/lib/libelf/gelf_getsyminfo.3
index 904e6490a2e..12da80f544e 100644
--- a/lib/libelf/gelf_getsyminfo.3
+++ b/lib/libelf/gelf_getsyminfo.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2006
-.Os
 .Dt GELF_GETSYMINFO 3
+.Os
 .Sh NAME
 .Nm gelf_getsyminfo ,
 .Nm gelf_update_syminfo
diff --git a/lib/libelf/gelf_getsymshndx.3 b/lib/libelf/gelf_getsymshndx.3
index c254a438259..d79c392959f 100644
--- a/lib/libelf/gelf_getsymshndx.3
+++ b/lib/libelf/gelf_getsymshndx.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 5, 2006
-.Os
 .Dt GELF_GETSYMSHNDX 3
+.Os
 .Sh NAME
 .Nm gelf_getsymshndx ,
 .Nm gelf_update_symshndx
diff --git a/lib/libelf/gelf_newehdr.3 b/lib/libelf/gelf_newehdr.3
index 98820e2e06e..6a2edf8f042 100644
--- a/lib/libelf/gelf_newehdr.3
+++ b/lib/libelf/gelf_newehdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 22, 2007
-.Os
 .Dt GELF_NEWEHDR 3
+.Os
 .Sh NAME
 .Nm elf32_newehdr ,
 .Nm elf64_newehdr ,
diff --git a/lib/libelf/gelf_newphdr.3 b/lib/libelf/gelf_newphdr.3
index 3d669b63aa1..bd13e5df53a 100644
--- a/lib/libelf/gelf_newphdr.3
+++ b/lib/libelf/gelf_newphdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 22, 2007
-.Os
 .Dt GELF_NEWPHDR 3
+.Os
 .Sh NAME
 .Nm elf32_newphdr ,
 .Nm elf64_newphdr ,
diff --git a/lib/libelf/gelf_update_ehdr.3 b/lib/libelf/gelf_update_ehdr.3
index 82098314e03..df23a0ceb0d 100644
--- a/lib/libelf/gelf_update_ehdr.3
+++ b/lib/libelf/gelf_update_ehdr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 27, 2006
-.Os
 .Dt GELF_UPDATE_EHDR 3
+.Os
 .Sh NAME
 .Nm gelf_update_ehdr ,
 .Nm gelf_update_phdr ,
diff --git a/lib/libelf/gelf_xlatetof.3 b/lib/libelf/gelf_xlatetof.3
index b7699305ea8..453f56fcc40 100644
--- a/lib/libelf/gelf_xlatetof.3
+++ b/lib/libelf/gelf_xlatetof.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 2006
-.Os
 .Dt GELF_XLATETOF 3
+.Os
 .Sh NAME
 .Nm elf32_xlate ,
 .Nm elf64_xlate ,
diff --git a/lib/libgssapi/gss_accept_sec_context.3 b/lib/libgssapi/gss_accept_sec_context.3
index 8957831e145..a4c219c164e 100644
--- a/lib/libgssapi/gss_accept_sec_context.3
+++ b/lib/libgssapi/gss_accept_sec_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_ACCEPT_SEC_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_accept_sec_context
 .Nd Accept a security context initiated by a peer application
diff --git a/lib/libgssapi/gss_acquire_cred.3 b/lib/libgssapi/gss_acquire_cred.3
index c48a46838d0..5de9ea0362a 100644
--- a/lib/libgssapi/gss_acquire_cred.3
+++ b/lib/libgssapi/gss_acquire_cred.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_ACQUIRE_CRED 3 PRM
+.Os
 .Sh NAME
 .Nm gss_acquire_cred
 .Nd Obtain a GSS-API credential handle for pre-existing credentials
diff --git a/lib/libgssapi/gss_add_cred.3 b/lib/libgssapi/gss_add_cred.3
index 53df140d555..67ff1c3dc74 100644
--- a/lib/libgssapi/gss_add_cred.3
+++ b/lib/libgssapi/gss_add_cred.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_ADD_CRED 3 PRM
+.Os
 .Sh NAME
 .Nm gss_add_cred
 .Nd Construct credentials incrementally
diff --git a/lib/libgssapi/gss_add_oid_set_member.3 b/lib/libgssapi/gss_add_oid_set_member.3
index fb2119cefe8..5839da6717c 100644
--- a/lib/libgssapi/gss_add_oid_set_member.3
+++ b/lib/libgssapi/gss_add_oid_set_member.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_ADD_OID_SET_MEMBER 3 PRM
+.Os
 .Sh NAME
 .Nm gss_add_oid_set_member
 .Nd Add an object identifier to a set
diff --git a/lib/libgssapi/gss_canonicalize_name.3 b/lib/libgssapi/gss_canonicalize_name.3
index 86957dc0584..036c0eff824 100644
--- a/lib/libgssapi/gss_canonicalize_name.3
+++ b/lib/libgssapi/gss_canonicalize_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_CANONICALIZE_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_canonicalize_name
 .Nd Convert an internal name to an MN
diff --git a/lib/libgssapi/gss_compare_name.3 b/lib/libgssapi/gss_compare_name.3
index 7f69d430b27..0bafb596408 100644
--- a/lib/libgssapi/gss_compare_name.3
+++ b/lib/libgssapi/gss_compare_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_COMPARE_NAME PRM
+.Os
 .Sh NAME
 .Nm gss_compare_name
 .Nd Compare two internal-form names
diff --git a/lib/libgssapi/gss_context_time.3 b/lib/libgssapi/gss_context_time.3
index 9fb43eacab0..b6b4157dd94 100644
--- a/lib/libgssapi/gss_context_time.3
+++ b/lib/libgssapi/gss_context_time.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_CONTEXT_TIME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_context_time
 .Nd Determine for how long a context will remain valid
diff --git a/lib/libgssapi/gss_create_empty_oid_set.3 b/lib/libgssapi/gss_create_empty_oid_set.3
index 08d82f6b2ba..3a84f855c15 100644
--- a/lib/libgssapi/gss_create_empty_oid_set.3
+++ b/lib/libgssapi/gss_create_empty_oid_set.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_CREATE_EMPTY_OID_SET 3 PRM
+.Os
 .Sh NAME
 .Nm gss_create_empty_oid_set
 .Nd Create a set containing no object identifiers
diff --git a/lib/libgssapi/gss_delete_sec_context.3 b/lib/libgssapi/gss_delete_sec_context.3
index a0c26df7f3e..1c4fa7b6690 100644
--- a/lib/libgssapi/gss_delete_sec_context.3
+++ b/lib/libgssapi/gss_delete_sec_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_DELETE_SEC_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_delete_sec_context
 .Nd Discard a security context
diff --git a/lib/libgssapi/gss_display_name.3 b/lib/libgssapi/gss_display_name.3
index 2441eef4c49..4944995c467 100644
--- a/lib/libgssapi/gss_display_name.3
+++ b/lib/libgssapi/gss_display_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_DISPLAY_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_display_name
 .Nd Convert internal-form name to text
diff --git a/lib/libgssapi/gss_display_status.3 b/lib/libgssapi/gss_display_status.3
index 2b1affd4e35..4bf908bb1d0 100644
--- a/lib/libgssapi/gss_display_status.3
+++ b/lib/libgssapi/gss_display_status.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_DISPLAY_STATUS 3 PRM
+.Os
 .Sh NAME
 .Nm gss_display_status
 .Nd Convert a GSS-API status code to text
diff --git a/lib/libgssapi/gss_duplicate_name.3 b/lib/libgssapi/gss_duplicate_name.3
index 334a1268410..5eb4a6284c4 100644
--- a/lib/libgssapi/gss_duplicate_name.3
+++ b/lib/libgssapi/gss_duplicate_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_DUPLICATE_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_duplicate_name
 .Nd Create a copy of an internal name
diff --git a/lib/libgssapi/gss_export_name.3 b/lib/libgssapi/gss_export_name.3
index e1feee22bd6..5cbf803b949 100644
--- a/lib/libgssapi/gss_export_name.3
+++ b/lib/libgssapi/gss_export_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_EXPORT_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_export_name
 .Nd Convert an MN to export form
diff --git a/lib/libgssapi/gss_export_sec_context.3 b/lib/libgssapi/gss_export_sec_context.3
index 1f7f01df9d3..7ecbef31717 100644
--- a/lib/libgssapi/gss_export_sec_context.3
+++ b/lib/libgssapi/gss_export_sec_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_EXPORT_SEC_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_export_sec_context
 .Nd Transfer a security context to another process
diff --git a/lib/libgssapi/gss_get_mic.3 b/lib/libgssapi/gss_get_mic.3
index 8892808f022..e5d81fa77f4 100644
--- a/lib/libgssapi/gss_get_mic.3
+++ b/lib/libgssapi/gss_get_mic.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_GET_MIC 3 PRM
+.Os
 .Sh NAME
 .Nm gss_get_mic ,
 .Nm gss_sign
diff --git a/lib/libgssapi/gss_import_name.3 b/lib/libgssapi/gss_import_name.3
index 3119f04534b..aef1ae8cb10 100644
--- a/lib/libgssapi/gss_import_name.3
+++ b/lib/libgssapi/gss_import_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_IMPORT_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_import_name
 .Nd Convert a contiguous string name to internal-form
diff --git a/lib/libgssapi/gss_import_sec_context.3 b/lib/libgssapi/gss_import_sec_context.3
index a889e2b7fa4..87aaa7872c7 100644
--- a/lib/libgssapi/gss_import_sec_context.3
+++ b/lib/libgssapi/gss_import_sec_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_IMPORT_SEC_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_import_sec_context
 .Nd Import a transferred context
diff --git a/lib/libgssapi/gss_indicate_mechs.3 b/lib/libgssapi/gss_indicate_mechs.3
index 26ffaa60162..3eec8615a17 100644
--- a/lib/libgssapi/gss_indicate_mechs.3
+++ b/lib/libgssapi/gss_indicate_mechs.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INDICATE_MECHS 3 PRM
+.Os
 .Sh NAME
 .Nm gss_indicate_mechs
 .Nd Determine available underlying authentication mechanisms
diff --git a/lib/libgssapi/gss_init_sec_context.3 b/lib/libgssapi/gss_init_sec_context.3
index 278dd17979b..9d14859fb41 100644
--- a/lib/libgssapi/gss_init_sec_context.3
+++ b/lib/libgssapi/gss_init_sec_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INIT_SEC_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_init_sec_context
 .Nd Initiate a security context with a peer application
diff --git a/lib/libgssapi/gss_inquire_context.3 b/lib/libgssapi/gss_inquire_context.3
index 51fa18db829..ba5d649facd 100644
--- a/lib/libgssapi/gss_inquire_context.3
+++ b/lib/libgssapi/gss_inquire_context.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INQUIRE_CONTEXT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_inquire_context
 .Nd Obtain information about a security context
diff --git a/lib/libgssapi/gss_inquire_cred.3 b/lib/libgssapi/gss_inquire_cred.3
index 99c20ce01fb..923b3734bea 100644
--- a/lib/libgssapi/gss_inquire_cred.3
+++ b/lib/libgssapi/gss_inquire_cred.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INQUIRE_CRED 3 PRM
+.Os
 .Sh NAME
 .Nm gss_inquire_cred
 .Nd Obtain information about a credential
diff --git a/lib/libgssapi/gss_inquire_cred_by_mech.3 b/lib/libgssapi/gss_inquire_cred_by_mech.3
index 22319bd4613..c316930cffb 100644
--- a/lib/libgssapi/gss_inquire_cred_by_mech.3
+++ b/lib/libgssapi/gss_inquire_cred_by_mech.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INQUIRE_CRED_BY_MECH 3 PRM
+.Os
 .Sh NAME
 .Nm gss_inquire_cred_by_mech
 .Nd Obtain per-mechanism information about a credential
diff --git a/lib/libgssapi/gss_inquire_mechs_for_name.3 b/lib/libgssapi/gss_inquire_mechs_for_name.3
index b56e663a18d..d8d94edb482 100644
--- a/lib/libgssapi/gss_inquire_mechs_for_name.3
+++ b/lib/libgssapi/gss_inquire_mechs_for_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INQUIRE_MECHS_FOR_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_inquire_mechs_for_name
 .Nd List mechanisms that support the specified name-type
diff --git a/lib/libgssapi/gss_inquire_names_for_mech.3 b/lib/libgssapi/gss_inquire_names_for_mech.3
index 8b7f069f915..c1b7528b56e 100644
--- a/lib/libgssapi/gss_inquire_names_for_mech.3
+++ b/lib/libgssapi/gss_inquire_names_for_mech.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_INQUIRE_NAMES_FOR_MECH 3 PRM
+.Os
 .Sh NAME
 .Nm gss_inquire_names_for_mech
 .Nd List the name-types supported by the specified mechanism
diff --git a/lib/libgssapi/gss_process_context_token.3 b/lib/libgssapi/gss_process_context_token.3
index 85e1029aab0..53f5ebfcc53 100644
--- a/lib/libgssapi/gss_process_context_token.3
+++ b/lib/libgssapi/gss_process_context_token.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_PROCESS_CONTEXT_TOKEN 3 PRM
+.Os
 .Sh NAME
 .Nm gss_process_context_token
 .Nd Process a token on a security context from a peer application
diff --git a/lib/libgssapi/gss_release_buffer.3 b/lib/libgssapi/gss_release_buffer.3
index 432d42276c5..259db3ba6b8 100644
--- a/lib/libgssapi/gss_release_buffer.3
+++ b/lib/libgssapi/gss_release_buffer.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_RELEASE_BUFFER 3 PRM
+.Os
 .Sh NAME
 .Nm gss_release_buffer
 .Nd Discard a buffer
diff --git a/lib/libgssapi/gss_release_cred.3 b/lib/libgssapi/gss_release_cred.3
index ca99fc40a8f..38fc7844bea 100644
--- a/lib/libgssapi/gss_release_cred.3
+++ b/lib/libgssapi/gss_release_cred.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_RELEASE_CRED 3 PRM
+.Os
 .Sh NAME
 .Nm gss_release_cred
 .Nd Discard a credential handle
diff --git a/lib/libgssapi/gss_release_name.3 b/lib/libgssapi/gss_release_name.3
index 7bf1181f22a..ca737e0af3a 100644
--- a/lib/libgssapi/gss_release_name.3
+++ b/lib/libgssapi/gss_release_name.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_RELEASE_NAME 3 PRM
+.Os
 .Sh NAME
 .Nm gss_release_name
 .Nd Discard an internal-form name
diff --git a/lib/libgssapi/gss_release_oid_set.3 b/lib/libgssapi/gss_release_oid_set.3
index 0142e2feec1..391ad283c79 100644
--- a/lib/libgssapi/gss_release_oid_set.3
+++ b/lib/libgssapi/gss_release_oid_set.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_RELEASE_OID_SET 3 PRM
+.Os
 .Sh NAME
 .Nm gss_release_oid_set
 .Nd Discard a set of object identifiers
diff --git a/lib/libgssapi/gss_test_oid_set_member.3 b/lib/libgssapi/gss_test_oid_set_member.3
index 86e06b35bb7..c6dad0f2fe7 100644
--- a/lib/libgssapi/gss_test_oid_set_member.3
+++ b/lib/libgssapi/gss_test_oid_set_member.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_TEST_OID_SET_MEMBER 3 PRM
+.Os
 .Sh NAME
 .Nm gss_test_oid_set_member
 .Nd Determines whether an object identifier is a member of a set
diff --git a/lib/libgssapi/gss_unwrap.3 b/lib/libgssapi/gss_unwrap.3
index cc75782310b..94537a49650 100644
--- a/lib/libgssapi/gss_unwrap.3
+++ b/lib/libgssapi/gss_unwrap.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_UNWRAP 3 PRM
+.Os
 .Sh NAME
 .Nm gss_unwrap ,
 .Nm gss_unseal
diff --git a/lib/libgssapi/gss_verify_mic.3 b/lib/libgssapi/gss_verify_mic.3
index b72f468ce7d..e8a46119422 100644
--- a/lib/libgssapi/gss_verify_mic.3
+++ b/lib/libgssapi/gss_verify_mic.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_VERIFY_MIC 3 PRM
+.Os
 .Sh NAME
 .Nm gss_verify_mic ,
 .Nm gss_verify
diff --git a/lib/libgssapi/gss_wrap.3 b/lib/libgssapi/gss_wrap.3
index c432b482c1b..5732dd24cf8 100644
--- a/lib/libgssapi/gss_wrap.3
+++ b/lib/libgssapi/gss_wrap.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_WRAP 3 PRM
+.Os
 .Sh NAME
 .Nm gss_wrap ,
 .Nm gss_seal
diff --git a/lib/libgssapi/gss_wrap_size_limit.3 b/lib/libgssapi/gss_wrap_size_limit.3
index c1d40da014d..7a1cb599ad2 100644
--- a/lib/libgssapi/gss_wrap_size_limit.3
+++ b/lib/libgssapi/gss_wrap_size_limit.3
@@ -28,8 +28,8 @@
 .\"
 .\" The following commands are required for all man pages.
 .Dd January 26, 2010
-.Os
 .Dt GSS_WRAP_SIZE_LIMIT 3 PRM
+.Os
 .Sh NAME
 .Nm gss_wrap_size_limit
 .Nd Determine maximum message sizes
diff --git a/lib/libmemstat/libmemstat.3 b/lib/libmemstat/libmemstat.3
index 82ec53abee6..9a4877b0aa7 100644
--- a/lib/libmemstat/libmemstat.3
+++ b/lib/libmemstat/libmemstat.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 27, 2005
-.Os
 .Dt LIBMEMSTAT 3
+.Os
 .Sh NAME
 .Nm libmemstat
 .Nd "library interface to retrieve kernel memory allocator statistics"
diff --git a/lib/libpmc/pmc.3 b/lib/libpmc/pmc.3
index f48fe7985e9..5e3236baf9b 100644
--- a/lib/libpmc/pmc.3
+++ b/lib/libpmc/pmc.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 24, 2008
-.Os
 .Dt PMC 3
+.Os
 .Sh NAME
 .Nm pmc
 .Nd library for accessing hardware performance monitoring counters
diff --git a/lib/libpmc/pmc.atom.3 b/lib/libpmc/pmc.atom.3
index bd081f9d27a..a54d1db2b1c 100644
--- a/lib/libpmc/pmc.atom.3
+++ b/lib/libpmc/pmc.atom.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 12, 2008
-.Os
 .Dt PMC.ATOM 3
+.Os
 .Sh NAME
 .Nm pmc.atom
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.core.3 b/lib/libpmc/pmc.core.3
index 3c97e092594..3feadaef2c5 100644
--- a/lib/libpmc/pmc.core.3
+++ b/lib/libpmc/pmc.core.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 12, 2008
-.Os
 .Dt PMC.CORE 3
+.Os
 .Sh NAME
 .Nm pmc.core
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.core2.3 b/lib/libpmc/pmc.core2.3
index 41c16750d62..3dbc0c850e5 100644
--- a/lib/libpmc/pmc.core2.3
+++ b/lib/libpmc/pmc.core2.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 8, 2009
-.Os
 .Dt PMC.CORE2 3
+.Os
 .Sh NAME
 .Nm pmc.core2
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.corei7.3 b/lib/libpmc/pmc.corei7.3
index 1a8b9d9d419..f90b0d3ead8 100644
--- a/lib/libpmc/pmc.corei7.3
+++ b/lib/libpmc/pmc.corei7.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 24, 2010
-.Os
 .Dt PMC.COREI7 3
+.Os
 .Sh NAME
 .Nm pmc.corei7
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.corei7uc.3 b/lib/libpmc/pmc.corei7uc.3
index 2c1b3dde07f..fc3435b698a 100644
--- a/lib/libpmc/pmc.corei7uc.3
+++ b/lib/libpmc/pmc.corei7uc.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 24, 2010
-.Os
 .Dt PMC.COREI7UC 3
+.Os
 .Sh NAME
 .Nm pmc.corei7uc
 .Nd uncore measurement events for
diff --git a/lib/libpmc/pmc.iaf.3 b/lib/libpmc/pmc.iaf.3
index c6d01eaaa72..ec9f21cb276 100644
--- a/lib/libpmc/pmc.iaf.3
+++ b/lib/libpmc/pmc.iaf.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 14, 2008
-.Os
 .Dt PMC.IAF 3
+.Os
 .Sh NAME
 .Nm pmc.iaf
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.k7.3 b/lib/libpmc/pmc.k7.3
index 8db9020dead..2775d4fd922 100644
--- a/lib/libpmc/pmc.k7.3
+++ b/lib/libpmc/pmc.k7.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC.K7 3
+.Os
 .Sh NAME
 .Nm pmc.k7
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.k8.3 b/lib/libpmc/pmc.k8.3
index 3e4afb1ff5b..995bfac958a 100644
--- a/lib/libpmc/pmc.k8.3
+++ b/lib/libpmc/pmc.k8.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC.K8 3
+.Os
 .Sh NAME
 .Nm pmc.k8
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.p4.3 b/lib/libpmc/pmc.p4.3
index ecfc6287db8..e13fa6ec5ea 100644
--- a/lib/libpmc/pmc.p4.3
+++ b/lib/libpmc/pmc.p4.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC.P4 3
+.Os
 .Sh NAME
 .Nm pmc.p4
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.p5.3 b/lib/libpmc/pmc.p5.3
index 2930f16caac..36ab917dc25 100644
--- a/lib/libpmc/pmc.p5.3
+++ b/lib/libpmc/pmc.p5.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC 3
+.Os
 .Sh NAME
 .Nm pmc
 .Nd library for accessing hardware performance monitoring counters
diff --git a/lib/libpmc/pmc.p6.3 b/lib/libpmc/pmc.p6.3
index 15a1101bf67..d8cde6427b6 100644
--- a/lib/libpmc/pmc.p6.3
+++ b/lib/libpmc/pmc.p6.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC.P6 3
+.Os
 .Sh NAME
 .Nm pmc.p6
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.tsc.3 b/lib/libpmc/pmc.tsc.3
index 3dd0b8859ca..144ff35c3f0 100644
--- a/lib/libpmc/pmc.tsc.3
+++ b/lib/libpmc/pmc.tsc.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 4, 2008
-.Os
 .Dt PMC.TSC 3
+.Os
 .Sh NAME
 .Nm pmc.tsc
 .Nd measurements using the i386 timestamp counter
diff --git a/lib/libpmc/pmc.ucf.3 b/lib/libpmc/pmc.ucf.3
index c8f2468cdf8..230a1108578 100644
--- a/lib/libpmc/pmc.ucf.3
+++ b/lib/libpmc/pmc.ucf.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 30, 2010
-.Os
 .Dt PMC.UCF 3
+.Os
 .Sh NAME
 .Nm pmc.ucf
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.westmere.3 b/lib/libpmc/pmc.westmere.3
index 0b87e50230a..98adce6f3af 100644
--- a/lib/libpmc/pmc.westmere.3
+++ b/lib/libpmc/pmc.westmere.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 24, 2010
-.Os
 .Dt PMC.WESTMERE 3
+.Os
 .Sh NAME
 .Nm pmc.westmere
 .Nd measurement events for
diff --git a/lib/libpmc/pmc.westmereuc.3 b/lib/libpmc/pmc.westmereuc.3
index c5fd12dfa66..c77ef9bfa48 100644
--- a/lib/libpmc/pmc.westmereuc.3
+++ b/lib/libpmc/pmc.westmereuc.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 24, 2010
-.Os
 .Dt PMC.WESTMEREUC 3
+.Os
 .Sh NAME
 .Nm pmc.westmere
 .Nd uncore measurement events for
diff --git a/lib/libpmc/pmc_allocate.3 b/lib/libpmc/pmc_allocate.3
index 03f907811a7..6a2a6c0d0f0 100644
--- a/lib/libpmc/pmc_allocate.3
+++ b/lib/libpmc/pmc_allocate.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 22, 2008
-.Os
 .Dt PMC_ALLOCATE 3
+.Os
 .Sh NAME
 .Nm pmc_allocate ,
 .Nm pmc_release
diff --git a/lib/libpmc/pmc_attach.3 b/lib/libpmc/pmc_attach.3
index be340a77598..ca725112b0b 100644
--- a/lib/libpmc/pmc_attach.3
+++ b/lib/libpmc/pmc_attach.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 25, 2007
-.Os
 .Dt PMC_ATTACH 3
+.Os
 .Sh NAME
 .Nm pmc_attach ,
 .Nm pmc_detach
diff --git a/lib/libpmc/pmc_capabilities.3 b/lib/libpmc/pmc_capabilities.3
index 0d1a99dd602..6aee17f8bfb 100644
--- a/lib/libpmc/pmc_capabilities.3
+++ b/lib/libpmc/pmc_capabilities.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 22, 2008
-.Os
 .Dt PMC_CAPABILITIES 3
+.Os
 .Sh NAME
 .Nm pmc_capabilities ,
 .Nm pmc_cpuinfo ,
diff --git a/lib/libpmc/pmc_configure_logfile.3 b/lib/libpmc/pmc_configure_logfile.3
index 95c6c28d438..a33688c222c 100644
--- a/lib/libpmc/pmc_configure_logfile.3
+++ b/lib/libpmc/pmc_configure_logfile.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 24, 2007
-.Os
 .Dt PMC_CONFIGURE_LOGFILE 3
+.Os
 .Sh NAME
 .Nm pmc_configure_logfile ,
 .Nm pmc_flush_logfile ,
diff --git a/lib/libpmc/pmc_disable.3 b/lib/libpmc/pmc_disable.3
index cd837d3a075..a6902ffd463 100644
--- a/lib/libpmc/pmc_disable.3
+++ b/lib/libpmc/pmc_disable.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 22, 2008
-.Os
 .Dt PMC_ENABLE 3
+.Os
 .Sh NAME
 .Nm pmc_disable ,
 .Nm pmc_enable
diff --git a/lib/libpmc/pmc_event_names_of_class.3 b/lib/libpmc/pmc_event_names_of_class.3
index fc1c63c3120..5b2c89d1b80 100644
--- a/lib/libpmc/pmc_event_names_of_class.3
+++ b/lib/libpmc/pmc_event_names_of_class.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 23, 2007
-.Os
 .Dt PMC_EVENT_NAMES_OF_CLASS 3
+.Os
 .Sh NAME
 .Nm pmc_event_names_of_class
 .Nd return a list of event names supported by a PMC class.
diff --git a/lib/libpmc/pmc_get_driver_stats.3 b/lib/libpmc/pmc_get_driver_stats.3
index d8bfe233b60..fa214b321a3 100644
--- a/lib/libpmc/pmc_get_driver_stats.3
+++ b/lib/libpmc/pmc_get_driver_stats.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 25, 2007
-.Os
 .Dt PMC_GET_DRIVER_STATS 3
+.Os
 .Sh NAME
 .Nm pmc_get_driver_stats
 .Nd retrieve driver statistics
diff --git a/lib/libpmc/pmc_get_msr.3 b/lib/libpmc/pmc_get_msr.3
index 6a2b94274bd..6361d3adab2 100644
--- a/lib/libpmc/pmc_get_msr.3
+++ b/lib/libpmc/pmc_get_msr.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 25, 2007
-.Os
 .Dt PMC_GET_MSR 3
+.Os
 .Sh NAME
 .Nm pmc_get_msr
 .Nd x86 architecture-specific PMC operations
diff --git a/lib/libpmc/pmc_init.3 b/lib/libpmc/pmc_init.3
index 5113b4aa951..655bfb6f78b 100644
--- a/lib/libpmc/pmc_init.3
+++ b/lib/libpmc/pmc_init.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 24, 2007
-.Os
 .Dt PMC_INIT 3
+.Os
 .Sh NAME
 .Nm pmc_init
 .Nd initialize library
diff --git a/lib/libpmc/pmc_name_of_capability.3 b/lib/libpmc/pmc_name_of_capability.3
index 53ddfdf9075..78efeafd6ac 100644
--- a/lib/libpmc/pmc_name_of_capability.3
+++ b/lib/libpmc/pmc_name_of_capability.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 24, 2007
-.Os
 .Dt PMC_NAME_OF_CAPABILITY 3
+.Os
 .Sh NAME
 .Nm pmc_name_of_capability ,
 .Nm pmc_name_of_class ,
diff --git a/lib/libpmc/pmc_read.3 b/lib/libpmc/pmc_read.3
index d9fa3ce8b9c..8d718ca3fd1 100644
--- a/lib/libpmc/pmc_read.3
+++ b/lib/libpmc/pmc_read.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 25, 2007
-.Os
 .Dt PMC_READ 3
+.Os
 .Sh NAME
 .Nm pmc_read ,
 .Nm pmc_rw ,
diff --git a/lib/libpmc/pmc_set.3 b/lib/libpmc/pmc_set.3
index a9e438af9df..e8d65971dd2 100644
--- a/lib/libpmc/pmc_set.3
+++ b/lib/libpmc/pmc_set.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 25, 2007
-.Os
 .Dt PMC_SET 3
+.Os
 .Sh NAME
 .Nm pmc_set
 .Nd set the reload count of a sampling PMC
diff --git a/lib/libpmc/pmc_start.3 b/lib/libpmc/pmc_start.3
index 7fb474a18bf..2272122dd19 100644
--- a/lib/libpmc/pmc_start.3
+++ b/lib/libpmc/pmc_start.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 22, 2008
-.Os
 .Dt PMC_START 3
+.Os
 .Sh NAME
 .Nm pmc_start ,
 .Nm pmc_stop
diff --git a/lib/libpmc/pmclog.3 b/lib/libpmc/pmclog.3
index 688445a4524..4438f106d0b 100644
--- a/lib/libpmc/pmclog.3
+++ b/lib/libpmc/pmclog.3
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 26, 2006
-.Os
 .Dt PMCLOG 3
+.Os
 .Sh NAME
 .Nm pmclog_open ,
 .Nm pmclog_close ,
diff --git a/lib/libthr/libthr.3 b/lib/libthr/libthr.3
index 87e5e6475d7..696da17bdcb 100644
--- a/lib/libthr/libthr.3
+++ b/lib/libthr/libthr.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 19, 2007
-.Os
 .Dt LIBTHR 3
+.Os
 .Sh NAME
 .Nm libthr
 .Nd "1:1 POSIX threads library"
diff --git a/lib/libugidfw/bsde_get_rule.3 b/lib/libugidfw/bsde_get_rule.3
index a3b6691eaa9..1f37b62613d 100644
--- a/lib/libugidfw/bsde_get_rule.3
+++ b/lib/libugidfw/bsde_get_rule.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 24, 2004
-.Os
 .Dt BSDE_GET_RULE 3
+.Os
 .Sh NAME
 .Nm bsde_add_rule ,
 .Nm bsde_get_rule ,
diff --git a/lib/libugidfw/bsde_get_rule_count.3 b/lib/libugidfw/bsde_get_rule_count.3
index bf441e150d9..2ca965f8fd4 100644
--- a/lib/libugidfw/bsde_get_rule_count.3
+++ b/lib/libugidfw/bsde_get_rule_count.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 7, 2003
-.Os
 .Dt BSDE_GET_RULE_COUNT 3
+.Os
 .Sh NAME
 .Nm bsde_get_rule_count ,
 .Nm bsde_get_rule_slots
diff --git a/lib/libugidfw/bsde_parse_rule.3 b/lib/libugidfw/bsde_parse_rule.3
index d510cab26a2..4b177af22ed 100644
--- a/lib/libugidfw/bsde_parse_rule.3
+++ b/lib/libugidfw/bsde_parse_rule.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 7, 2003
-.Os
 .Dt BSDE_PARSE_RULE 3
+.Os
 .Sh NAME
 .Nm bsde_parse_rule ,
 .Nm bsde_parse_rule_string
diff --git a/lib/libugidfw/bsde_rule_to_string.3 b/lib/libugidfw/bsde_rule_to_string.3
index 44144069a52..6279dcfa371 100644
--- a/lib/libugidfw/bsde_rule_to_string.3
+++ b/lib/libugidfw/bsde_rule_to_string.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 7, 2003
-.Os
 .Dt BSDE_RULE_TO_STRING 3
+.Os
 .Sh NAME
 .Nm bsde_rule_to_string
 .Nd "convert a ugidfw rule into its text representation"
diff --git a/lib/libugidfw/libugidfw.3 b/lib/libugidfw/libugidfw.3
index 3ff407c1609..42d23c68039 100644
--- a/lib/libugidfw/libugidfw.3
+++ b/lib/libugidfw/libugidfw.3
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 25, 2004
-.Os
 .Dt LIBUGIDFW 3
+.Os
 .Sh NAME
 .Nm libugidfw
 .Nd "library interface to the file system firewall MAC policy"
diff --git a/lib/libulog/ulog_login.3 b/lib/libulog/ulog_login.3
index ec397c8c0cd..451a4f7a31f 100644
--- a/lib/libulog/ulog_login.3
+++ b/lib/libulog/ulog_login.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 5, 2009
-.Os
 .Dt ULOG_LOGIN 3
+.Os
 .Sh NAME
 .Nm ulog_login ,
 .Nm ulog_login_pseudo ,
diff --git a/lib/libulog/utempter_add_record.3 b/lib/libulog/utempter_add_record.3
index 1553d68c26b..cd8e8f1e69f 100644
--- a/lib/libulog/utempter_add_record.3
+++ b/lib/libulog/utempter_add_record.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 6, 2009
-.Os
 .Dt UTEMPTER_ADD_RECORD 3
+.Os
 .Sh NAME
 .Nm utempter_add_record ,
 .Nm utempter_remove_added_record ,
diff --git a/lib/libutil/_secure_path.3 b/lib/libutil/_secure_path.3
index fe23d12b2b9..cf89315a89e 100644
--- a/lib/libutil/_secure_path.3
+++ b/lib/libutil/_secure_path.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 2, 1997
-.Os
 .Dt _SECURE_PATH 3
+.Os
 .Sh NAME
 .Nm _secure_path
 .Nd determine if a file appears to be secure
diff --git a/lib/libutil/auth.3 b/lib/libutil/auth.3
index 3a1bb4dd4db..247a0985914 100644
--- a/lib/libutil/auth.3
+++ b/lib/libutil/auth.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd October 7, 1998
-.Os
 .Dt AUTH_GETVAL 3
+.Os
 .Sh NAME
 .Nm auth_getval
 .Nd functions for reading values from
diff --git a/lib/libutil/hexdump.3 b/lib/libutil/hexdump.3
index cc004f61bb3..8e997ac24b4 100644
--- a/lib/libutil/hexdump.3
+++ b/lib/libutil/hexdump.3
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 1, 2008
-.Os
 .Dt HEXDUMP 3
+.Os
 .Sh NAME
 .Nm hexdump
 .Nd "dump a block of bytes to standard out in hexadecimal form"
diff --git a/lib/libutil/kinfo_getfile.3 b/lib/libutil/kinfo_getfile.3
index 5705bf774d6..c0b0b022021 100644
--- a/lib/libutil/kinfo_getfile.3
+++ b/lib/libutil/kinfo_getfile.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 6, 2008
-.Os
 .Dt KINFO_GETFILE 3
+.Os
 .Sh NAME
 .Nm kinfo_getfile
 .Nd function for getting per-process file descriptor information
diff --git a/lib/libutil/kinfo_getvmmap.3 b/lib/libutil/kinfo_getvmmap.3
index 0f12b675ac3..7aa46fc102e 100644
--- a/lib/libutil/kinfo_getvmmap.3
+++ b/lib/libutil/kinfo_getvmmap.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 6, 2008
-.Os
 .Dt KINFO_GETVMMAP 3
+.Os
 .Sh NAME
 .Nm kinfo_getvmmap
 .Nd function for getting per-process memory map information
diff --git a/lib/libutil/kld.3 b/lib/libutil/kld.3
index cc34fa1a040..6b77c47fc0d 100644
--- a/lib/libutil/kld.3
+++ b/lib/libutil/kld.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 18, 2006
-.Os
 .Dt KLD 3
+.Os
 .Sh NAME
 .Nm kld_isloaded ,
 .Nm kld_load
diff --git a/lib/libutil/login_auth.3 b/lib/libutil/login_auth.3
index 3750bd69182..003f6e92ac0 100644
--- a/lib/libutil/login_auth.3
+++ b/lib/libutil/login_auth.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 29, 1996
-.Os
 .Dt LOGIN_AUTH 3
+.Os
 .Sh NAME
 .\" .Nm authenticate
 .\" .Nm auth_script
diff --git a/lib/libutil/login_cap.3 b/lib/libutil/login_cap.3
index 7ecefffb2a3..bdab56eba50 100644
--- a/lib/libutil/login_cap.3
+++ b/lib/libutil/login_cap.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 14, 2007
-.Os
 .Dt LOGIN_CAP 3
+.Os
 .Sh NAME
 .Nm login_close ,
 .Nm login_getcapbool ,
diff --git a/lib/libutil/login_class.3 b/lib/libutil/login_class.3
index 75ccccc1449..083e94289d1 100644
--- a/lib/libutil/login_class.3
+++ b/lib/libutil/login_class.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 20, 2008
-.Os
 .Dt LOGIN_CLASS 3
+.Os
 .Sh NAME
 .Nm setclasscontext ,
 .Nm setclasscpumask ,
diff --git a/lib/libutil/login_ok.3 b/lib/libutil/login_ok.3
index 1bb6236c243..9022ff55d3a 100644
--- a/lib/libutil/login_ok.3
+++ b/lib/libutil/login_ok.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 2, 1997
-.Os
 .Dt LOGIN_OK 3
+.Os
 .Sh NAME
 .Nm auth_ttyok ,
 .Nm auth_hostok ,
diff --git a/lib/libutil/login_times.3 b/lib/libutil/login_times.3
index a1444843d42..9d1f88d858c 100644
--- a/lib/libutil/login_times.3
+++ b/lib/libutil/login_times.3
@@ -20,8 +20,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 20, 2008
-.Os
 .Dt LOGIN_TIMES 3
+.Os
 .Sh NAME
 .Nm parse_lt ,
 .Nm in_lt ,
diff --git a/lib/libutil/login_tty.3 b/lib/libutil/login_tty.3
index 6303d685e03..907b97cd5cf 100644
--- a/lib/libutil/login_tty.3
+++ b/lib/libutil/login_tty.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd December 29, 1996
-.Os
 .Dt LOGIN_TTY 3
+.Os
 .Sh NAME
 .Nm login_tty
 .Nd prepare a tty for a new login session
diff --git a/lib/libutil/property.3 b/lib/libutil/property.3
index 61eb1539909..51f7e14928f 100644
--- a/lib/libutil/property.3
+++ b/lib/libutil/property.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd October 7, 1998
-.Os
 .Dt PROPERTIES 3
+.Os
 .Sh NAME
 .Nm properties_read ,
 .Nm propery_find ,
diff --git a/lib/libutil/pty.3 b/lib/libutil/pty.3
index 08a5985b684..f71cc5323d7 100644
--- a/lib/libutil/pty.3
+++ b/lib/libutil/pty.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd December 29, 1996
-.Os
 .Dt PTY 3
+.Os
 .Sh NAME
 .Nm openpty ,
 .Nm forkpty
diff --git a/lib/libutil/realhostname.3 b/lib/libutil/realhostname.3
index fd64d4c6966..9f5a6c5ddc6 100644
--- a/lib/libutil/realhostname.3
+++ b/lib/libutil/realhostname.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 6, 1999
-.Os
 .Dt REALHOSTNAME 3
+.Os
 .Sh NAME
 .Nm realhostname
 .Nd "convert an IP number to the real host name"
diff --git a/lib/libutil/realhostname_sa.3 b/lib/libutil/realhostname_sa.3
index 9f7f0dfb022..3fd44a442b1 100644
--- a/lib/libutil/realhostname_sa.3
+++ b/lib/libutil/realhostname_sa.3
@@ -52,8 +52,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 11, 2000
-.Os
 .Dt REALHOSTNAME_SA 3
+.Os
 .Sh NAME
 .Nm realhostname_sa
 .Nd "convert a"
diff --git a/lib/libutil/trimdomain.3 b/lib/libutil/trimdomain.3
index f01e884e787..8d600c01921 100644
--- a/lib/libutil/trimdomain.3
+++ b/lib/libutil/trimdomain.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 7, 1999
-.Os
 .Dt TRIMDOMAIN 3
+.Os
 .Sh NAME
 .Nm trimdomain
 .Nd "trim the current domain name from a host name"
diff --git a/lib/libutil/uucplock.3 b/lib/libutil/uucplock.3
index 5f75c7f94e9..5d8647d0319 100644
--- a/lib/libutil/uucplock.3
+++ b/lib/libutil/uucplock.3
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd March 30, 1997
-.Os
 .Dt UUCPLOCK 3
+.Os
 .Sh NAME
 .Nm uu_lock ,
 .Nm uu_unlock ,
diff --git a/sbin/iscontrol/iscsi.conf.5 b/sbin/iscontrol/iscsi.conf.5
index 2edcd35742a..0de5122914f 100644
--- a/sbin/iscontrol/iscsi.conf.5
+++ b/sbin/iscontrol/iscsi.conf.5
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 5, 2007
-.Os
 .Dt ISCSI.CONF 5
+.Os
 .Sh NAME
 .Nm iscsi.conf
 .Nd key options to be negotiated in an iSCSI session
diff --git a/sbin/spppcontrol/spppcontrol.8 b/sbin/spppcontrol/spppcontrol.8
index 4389c8f66ec..4d948a60651 100644
--- a/sbin/spppcontrol/spppcontrol.8
+++ b/sbin/spppcontrol/spppcontrol.8
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 30, 2001
-.Os
 .Dt SPPPCONTROL 8
+.Os
 .Sh NAME
 .Nm spppcontrol
 .Nd display or set parameters for an sppp interface
diff --git a/share/man/man3/sysexits.3 b/share/man/man3/sysexits.3
index 4661c9d9017..2ef3407b325 100644
--- a/share/man/man3/sysexits.3
+++ b/share/man/man3/sysexits.3
@@ -27,8 +27,8 @@
 .\"
 .\" "
 .Dd March 31, 1996
-.Os
 .Dt SYSEXITS 3
+.Os
 .Sh NAME
 .Nm sysexits
 .Nd preferable exit codes for programs
diff --git a/share/man/man3/tgmath.3 b/share/man/man3/tgmath.3
index 4957c9b2649..f8faf269a7a 100644
--- a/share/man/man3/tgmath.3
+++ b/share/man/man3/tgmath.3
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 14, 2004
-.Os
 .Dt TGMATH 3
+.Os
 .Sh NAME
 .Nm tgmath
 .Nd "type-generic macros"
diff --git a/share/man/man4/audit.4 b/share/man/man4/audit.4
index 82bb23fdcac..167248cec5d 100644
--- a/share/man/man4/audit.4
+++ b/share/man/man4/audit.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 31, 2009
-.Os
 .Dt AUDIT 4
+.Os
 .Sh NAME
 .Nm audit
 .Nd Security Event Audit
diff --git a/share/man/man4/auditpipe.4 b/share/man/man4/auditpipe.4
index 7ba35afcb47..19db8565b70 100644
--- a/share/man/man4/auditpipe.4
+++ b/share/man/man4/auditpipe.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 5, 2006
-.Os
 .Dt AUDITPIPE 4
+.Os
 .Sh NAME
 .Nm auditpipe
 .Nd "pseudo-device for live audit event tracking"
diff --git a/share/man/man4/coda.4 b/share/man/man4/coda.4
index 9fd6cbb8be1..94b12aec4d5 100644
--- a/share/man/man4/coda.4
+++ b/share/man/man4/coda.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 21, 2008
-.Os
 .Dt CODA 4
+.Os
 .Sh NAME
 .Nm coda 
 .Nd Coda Distributed File System
diff --git a/share/man/man4/gbde.4 b/share/man/man4/gbde.4
index e3bc6c7fb2a..edac1a07316 100644
--- a/share/man/man4/gbde.4
+++ b/share/man/man4/gbde.4
@@ -32,8 +32,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 19, 2002
-.Os
 .Dt GBDE 4
+.Os
 .Sh NAME
 .Nm gbde
 .Nd Geom Based Disk Encryption
diff --git a/share/man/man4/geom.4 b/share/man/man4/geom.4
index 38573893357..0a1b24e5ec3 100644
--- a/share/man/man4/geom.4
+++ b/share/man/man4/geom.4
@@ -35,8 +35,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 25, 2006
-.Os
 .Dt GEOM 4
+.Os
 .Sh NAME
 .Nm GEOM
 .Nd "modular disk I/O request transformation framework"
diff --git a/share/man/man4/geom_fox.4 b/share/man/man4/geom_fox.4
index 588e1d7c932..8e262b94ae5 100644
--- a/share/man/man4/geom_fox.4
+++ b/share/man/man4/geom_fox.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 2, 2005
-.Os
 .Dt GEOM_FOX 4
+.Os
 .Sh NAME
 .Nm geom_fox
 .Nd "GEOM based basic disk multipathing"
diff --git a/share/man/man4/geom_linux_lvm.4 b/share/man/man4/geom_linux_lvm.4
index e41703a489b..30b875edd3e 100644
--- a/share/man/man4/geom_linux_lvm.4
+++ b/share/man/man4/geom_linux_lvm.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 20, 2008
-.Os
 .Dt GEOM_LINUX_LVM 4
+.Os
 .Sh NAME
 .Nm geom_linux_lvm
 .Nd "GEOM based Linux LVM logical volume mapping"
diff --git a/share/man/man4/geom_uzip.4 b/share/man/man4/geom_uzip.4
index 30d5ab03e81..7275029d062 100644
--- a/share/man/man4/geom_uzip.4
+++ b/share/man/man4/geom_uzip.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 9, 2006
-.Os
 .Dt GEOM_UZIP 4
+.Os
 .Sh NAME
 .Nm geom_uzip
 .Nd "GEOM based compressed disk images"
diff --git a/share/man/man4/ipw.4 b/share/man/man4/ipw.4
index 30293f0cafe..060058a0981 100644
--- a/share/man/man4/ipw.4
+++ b/share/man/man4/ipw.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 13, 2008
-.Os
 .Dt IPW 4
+.Os
 .Sh NAME
 .Nm ipw
 .Nd "Intel PRO/Wireless 2100 IEEE 802.11 driver"
diff --git a/share/man/man4/iscsi_initiator.4 b/share/man/man4/iscsi_initiator.4
index 484a2135531..b6c47f0769b 100644
--- a/share/man/man4/iscsi_initiator.4
+++ b/share/man/man4/iscsi_initiator.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 23, 2007
-.Os
 .Dt ISCSI_INITIATOR 4
+.Os
 .Sh NAME
 .Nm iscsi_initiator
 .Nd kernel driver for the iSCSI protocol
diff --git a/share/man/man4/iwi.4 b/share/man/man4/iwi.4
index 3c9bdd46608..64f5c0ed13a 100644
--- a/share/man/man4/iwi.4
+++ b/share/man/man4/iwi.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 13, 2008
-.Os
 .Dt IWI 4
+.Os
 .Sh NAME
 .Nm iwi
 .Nd "Intel PRO/Wireless 2200BG/2225BG/2915ABG IEEE 802.11 driver"
diff --git a/share/man/man4/iwn.4 b/share/man/man4/iwn.4
index b73331a04b5..a229e57c843 100644
--- a/share/man/man4/iwn.4
+++ b/share/man/man4/iwn.4
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 8, 2010
-.Os
 .Dt IWN 4
+.Os
 .Sh NAME
 .Nm iwn
 .Nd Intel Wireless WiFi Link 4965/1000/5000/5150/5300/6000/6050 
diff --git a/share/man/man4/kbdmux.4 b/share/man/man4/kbdmux.4
index e29f8cd766e..f92fa4584e9 100644
--- a/share/man/man4/kbdmux.4
+++ b/share/man/man4/kbdmux.4
@@ -2,8 +2,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 12, 2005
-.Os
 .Dt KBDMUX 4
+.Os
 .Sh NAME
 .Nm kbdmux
 .Nd "keyboard multiplexer"
diff --git a/share/man/man4/lp.4 b/share/man/man4/lp.4
index 6976101af2c..f5b628e6ff0 100644
--- a/share/man/man4/lp.4
+++ b/share/man/man4/lp.4
@@ -35,8 +35,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 4, 1996
-.Os
 .Dt LP 4
+.Os
 .Sh NAME
 .Nm lp
 .Nd printer port Internet Protocol driver
diff --git a/share/man/man4/mac.4 b/share/man/man4/mac.4
index 82d7a165aa1..b1ff1d951d8 100644
--- a/share/man/man4/mac.4
+++ b/share/man/man4/mac.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 30, 2007
-.Os
 .Dt MAC 4
+.Os
 .Sh NAME
 .Nm mac
 .Nd Mandatory Access Control
diff --git a/share/man/man4/mac_biba.4 b/share/man/man4/mac_biba.4
index 20a66615e5f..265f5710edc 100644
--- a/share/man/man4/mac_biba.4
+++ b/share/man/man4/mac_biba.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 18, 2002
-.Os
 .Dt MAC_BIBA 4
+.Os
 .Sh NAME
 .Nm mac_biba
 .Nd "Biba data integrity policy"
diff --git a/share/man/man4/mac_bsdextended.4 b/share/man/man4/mac_bsdextended.4
index 899990c34a1..23b89c04d04 100644
--- a/share/man/man4/mac_bsdextended.4
+++ b/share/man/man4/mac_bsdextended.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 21, 2005
-.Os
 .Dt MAC_BSDEXTENDED 4
+.Os
 .Sh NAME
 .Nm mac_bsdextended
 .Nd "file system firewall policy"
diff --git a/share/man/man4/mac_ifoff.4 b/share/man/man4/mac_ifoff.4
index ede13b0fac5..87c73b25088 100644
--- a/share/man/man4/mac_ifoff.4
+++ b/share/man/man4/mac_ifoff.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 10, 2002
-.Os
 .Dt MAC_IFOFF 4
+.Os
 .Sh NAME
 .Nm mac_ifoff
 .Nd "interface silencing policy"
diff --git a/share/man/man4/mac_lomac.4 b/share/man/man4/mac_lomac.4
index 8133ad8a833..8ae8875acb4 100644
--- a/share/man/man4/mac_lomac.4
+++ b/share/man/man4/mac_lomac.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 11, 2002
-.Os
 .Dt MAC_LOMAC 4
+.Os
 .Sh NAME
 .Nm mac_lomac
 .Nd "Low-watermark Mandatory Access Control data integrity policy"
diff --git a/share/man/man4/mac_mls.4 b/share/man/man4/mac_mls.4
index 921781adf1e..b314fb5ca3a 100644
--- a/share/man/man4/mac_mls.4
+++ b/share/man/man4/mac_mls.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 1, 2002
-.Os
 .Dt MAC_MLS 4
+.Os
 .Sh NAME
 .Nm mac_mls
 .Nd "Multi-Level Security confidentiality policy"
diff --git a/share/man/man4/mac_none.4 b/share/man/man4/mac_none.4
index 159151e5338..8f4602935ac 100644
--- a/share/man/man4/mac_none.4
+++ b/share/man/man4/mac_none.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 1, 2002
-.Os
 .Dt MAC_NONE 4
+.Os
 .Sh NAME
 .Nm mac_none
 .Nd "null MAC policy module"
diff --git a/share/man/man4/mac_partition.4 b/share/man/man4/mac_partition.4
index 54abee5ae7b..296635edaf0 100644
--- a/share/man/man4/mac_partition.4
+++ b/share/man/man4/mac_partition.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 9, 2002
-.Os
 .Dt MAC_PARTITION 4
+.Os
 .Sh NAME
 .Nm mac_partition
 .Nd "process partition policy"
diff --git a/share/man/man4/mac_seeotheruids.4 b/share/man/man4/mac_seeotheruids.4
index 4eeb7bba9eb..c870ca00d3e 100644
--- a/share/man/man4/mac_seeotheruids.4
+++ b/share/man/man4/mac_seeotheruids.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 6, 2005
-.Os
 .Dt MAC_SEEOTHERUIDS 4
+.Os
 .Sh NAME
 .Nm mac_seeotheruids
 .Nd "simple policy controlling whether users see other users"
diff --git a/share/man/man4/mac_stub.4 b/share/man/man4/mac_stub.4
index ded95e2c081..79326df3649 100644
--- a/share/man/man4/mac_stub.4
+++ b/share/man/man4/mac_stub.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 1, 2002
-.Os
 .Dt MAC_STUB 4
+.Os
 .Sh NAME
 .Nm mac_stub
 .Nd "MAC policy stub module"
diff --git a/share/man/man4/mac_test.4 b/share/man/man4/mac_test.4
index dde744589a2..e86d4bd0f14 100644
--- a/share/man/man4/mac_test.4
+++ b/share/man/man4/mac_test.4
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 1, 2002
-.Os
 .Dt MAC_TEST 4
+.Os
 .Sh NAME
 .Nm mac_test
 .Nd MAC framework testing policy
diff --git a/share/man/man4/ng_netflow.4 b/share/man/man4/ng_netflow.4
index d1ef2048226..5c5c8e0ccfd 100644
--- a/share/man/man4/ng_netflow.4
+++ b/share/man/man4/ng_netflow.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 8, 2008
-.Os
 .Dt NG_NETFLOW 4
+.Os
 .Sh NAME
 .Nm ng_netflow
 .Nd Cisco's NetFlow implementation
diff --git a/share/man/man4/orm.4 b/share/man/man4/orm.4
index 2ef54a0b5d3..2922cb68ead 100644
--- a/share/man/man4/orm.4
+++ b/share/man/man4/orm.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 15, 2000
-.Os
 .Dt ORM 4
+.Os
 .Sh NAME
 .Nm orm
 .Nd ISA I/O space option ROM(s) driver
diff --git a/share/man/man4/ral.4 b/share/man/man4/ral.4
index 2fd8e3355cb..25fb343a0b4 100644
--- a/share/man/man4/ral.4
+++ b/share/man/man4/ral.4
@@ -16,8 +16,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 8, 2009
-.Os
 .Dt RAL 4
+.Os
 .Sh NAME
 .Nm ral
 .Nd "Ralink Technology IEEE 802.11 wireless network driver"
diff --git a/share/man/man4/rp.4 b/share/man/man4/rp.4
index 8f0f425c662..a894f1242c7 100644
--- a/share/man/man4/rp.4
+++ b/share/man/man4/rp.4
@@ -3,8 +3,8 @@
 .\"
 .\" $FreeBSD$
 .Dd November 15, 1995
-.Os
 .Dt RP 4
+.Os
 .Sh NAME
 .Nm rp
 .Nd "driver for Comtrol RocketPort Intelligent Serial Port Cards"
diff --git a/share/man/man4/rum.4 b/share/man/man4/rum.4
index cbbbcd0962b..82e25f5c3b9 100644
--- a/share/man/man4/rum.4
+++ b/share/man/man4/rum.4
@@ -17,8 +17,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 13, 2008
-.Os
 .Dt RUM 4
+.Os
 .Sh NAME
 .Nm rum
 .Nd Ralink Technology USB IEEE 802.11a/b/g wireless network device
diff --git a/share/man/man4/run.4 b/share/man/man4/run.4
index e613649e853..a7c6f4deeaf 100644
--- a/share/man/man4/run.4
+++ b/share/man/man4/run.4
@@ -17,8 +17,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 29, 2010
-.Os
 .Dt RUN 4
+.Os
 .Sh NAME
 .Nm run
 .Nd Ralink Technology USB IEEE 802.11a/g/n wireless network device
diff --git a/share/man/man4/sched_4bsd.4 b/share/man/man4/sched_4bsd.4
index 9533cae0055..a02c40eb58a 100644
--- a/share/man/man4/sched_4bsd.4
+++ b/share/man/man4/sched_4bsd.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 21, 2008
-.Os
 .Dt SCHED_4BSD 4
+.Os
 .Sh NAME
 .Nm sched_4bsd
 .Nd "4.4BSD scheduler"
diff --git a/share/man/man4/sched_ule.4 b/share/man/man4/sched_ule.4
index c5ce58b57c7..fdb96f1c828 100644
--- a/share/man/man4/sched_ule.4
+++ b/share/man/man4/sched_ule.4
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 21, 2008
-.Os
 .Dt SCHED_ULE 4
+.Os
 .Sh NAME
 .Nm sched_ule
 .Nd ULE scheduler
diff --git a/share/man/man4/si.4 b/share/man/man4/si.4
index 65176bdb78b..5eb7629d979 100644
--- a/share/man/man4/si.4
+++ b/share/man/man4/si.4
@@ -1,7 +1,7 @@
 .\" $FreeBSD$
 .Dd September 16, 1995
-.Os
 .Dt SI 4
+.Os
 .Sh NAME
 .Nm si
 .Nd "driver for Specialix International SI/XIO or SX intelligent serial card"
diff --git a/share/man/man4/tap.4 b/share/man/man4/tap.4
index 9f047257289..a306108af63 100644
--- a/share/man/man4/tap.4
+++ b/share/man/man4/tap.4
@@ -2,8 +2,8 @@
 .\" Based on PR#2411
 .\"
 .Dd September 8, 2008
-.Os
 .Dt TAP 4
+.Os
 .Sh NAME
 .Nm tap
 .Nd Ethernet tunnel software network interface
diff --git a/share/man/man4/uhso.4 b/share/man/man4/uhso.4
index 7f10c6212a3..e4d6df5b2aa 100644
--- a/share/man/man4/uhso.4
+++ b/share/man/man4/uhso.4
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 14, 2010
-.Os
 .Dt UHSO 4
+.Os
 .Sh NAME
 .Nm uhso
 .Nd support for several HSxPA devices from Option N.V.
diff --git a/share/man/man4/upgt.4 b/share/man/man4/upgt.4
index 39dee55c716..06c5eb57f76 100644
--- a/share/man/man4/upgt.4
+++ b/share/man/man4/upgt.4
@@ -50,8 +50,8 @@
 .\" SUCH DAMAGE.
 .\"
 .Dd April 17, 2008
-.Os
 .Dt UPGT 4
+.Os
 .Sh NAME
 .Nm upgt
 .Nd Conexant/Intersil PrismGT SoftMAC USB IEEE 802.11b/g wireless network
diff --git a/share/man/man4/ural.4 b/share/man/man4/ural.4
index 053ae704a3c..3825166ffa8 100644
--- a/share/man/man4/ural.4
+++ b/share/man/man4/ural.4
@@ -16,8 +16,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 13, 2008
-.Os
 .Dt URAL 4
+.Os
 .Sh NAME
 .Nm ural
 .Nd "Ralink Technology RT2500USB IEEE 802.11 driver"
diff --git a/share/man/man4/vkbd.4 b/share/man/man4/vkbd.4
index bdaa941531c..079688f6bb0 100644
--- a/share/man/man4/vkbd.4
+++ b/share/man/man4/vkbd.4
@@ -2,8 +2,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 12, 2004
-.Os
 .Dt VKBD 4
+.Os
 .Sh NAME
 .Nm vkbd
 .Nd the virtual AT keyboard interface
diff --git a/share/man/man4/wpi.4 b/share/man/man4/wpi.4
index f424d74a3c7..07e88169827 100644
--- a/share/man/man4/wpi.4
+++ b/share/man/man4/wpi.4
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 5, 2008
-.Os
 .Dt WPI 4
+.Os
 .Sh NAME
 .Nm wpi
 .Nd "Intel 3945ABG Wireless LAN IEEE 802.11 driver"
diff --git a/share/man/man5/ar.5 b/share/man/man5/ar.5
index 58d2d788701..dc802cee79c 100644
--- a/share/man/man5/ar.5
+++ b/share/man/man5/ar.5
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 7, 2007
-.Os
 .Dt AR 5
+.Os
 .Sh NAME
 .Nm ar
 .Nd format of archives managed by ar(1) and ranlib(1)
diff --git a/share/man/man7/clocks.7 b/share/man/man7/clocks.7
index 700a79f33da..692796842ed 100644
--- a/share/man/man7/clocks.7
+++ b/share/man/man7/clocks.7
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd January 18, 2008
-.Os
 .Dt CLOCKS 7
+.Os
 .Sh NAME
 .Nm clocks
 .Nd various system timers
diff --git a/share/man/man7/maclabel.7 b/share/man/man7/maclabel.7
index 05c3654aa9e..0648fb565c9 100644
--- a/share/man/man7/maclabel.7
+++ b/share/man/man7/maclabel.7
@@ -34,8 +34,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 25, 2002
-.Os
 .Dt MACLABEL 7
+.Os
 .Sh NAME
 .Nm maclabel
 .Nd Mandatory Access Control label format
diff --git a/share/man/man8/picobsd.8 b/share/man/man8/picobsd.8
index 1ce50b3386b..c6c6c948ea8 100644
--- a/share/man/man8/picobsd.8
+++ b/share/man/man8/picobsd.8
@@ -1,8 +1,8 @@
 .\" -*- nroff-fill -*-
 .\" $FreeBSD$
 .Dd June 25, 2009
-.Os
 .Dt PICOBSD 8
+.Os
 .Sh NAME
 .Nm picobsd
 .Nd building small FreeBSD disk images
diff --git a/share/man/man8/rescue.8 b/share/man/man8/rescue.8
index 685827f3d59..29218874d67 100644
--- a/share/man/man8/rescue.8
+++ b/share/man/man8/rescue.8
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 23, 2003
-.Os
 .Dt RESCUE 8
+.Os
 .Sh NAME
 .Nm rescue
 .Nd rescue utilities in
diff --git a/share/man/man9/CTASSERT.9 b/share/man/man9/CTASSERT.9
index db1ad49ee77..4885d4248b8 100644
--- a/share/man/man9/CTASSERT.9
+++ b/share/man/man9/CTASSERT.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 5, 2008
-.Os
 .Dt CTASSERT 9
+.Os
 .Sh NAME
 .Nm CTASSERT
 .Nd compile time assertion macro
diff --git a/share/man/man9/DELAY.9 b/share/man/man9/DELAY.9
index ac6231c2bc7..58b2a07e8b6 100644
--- a/share/man/man9/DELAY.9
+++ b/share/man/man9/DELAY.9
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 21, 2000
-.Os
 .Dt DELAY 9
+.Os
 .Sh NAME
 .Nm DELAY
 .Nd busy loop for an interval
diff --git a/share/man/man9/KASSERT.9 b/share/man/man9/KASSERT.9
index e53e9b605a8..62c97c44bd5 100644
--- a/share/man/man9/KASSERT.9
+++ b/share/man/man9/KASSERT.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 14, 2000
-.Os
 .Dt KASSERT 9
+.Os
 .Sh NAME
 .Nm KASSERT
 .Nd kernel expression verification macro
diff --git a/share/man/man9/VFS.9 b/share/man/man9/VFS.9
index 30f89189bf2..ca3109b5b48 100644
--- a/share/man/man9/VFS.9
+++ b/share/man/man9/VFS.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 9, 2010
-.Os
 .Dt VFS 9
+.Os
 .Sh NAME
 .Nm VFS
 .Nd kernel interface to file systems
diff --git a/share/man/man9/VFS_CHECKEXP.9 b/share/man/man9/VFS_CHECKEXP.9
index 92afda86d69..62d545c753e 100644
--- a/share/man/man9/VFS_CHECKEXP.9
+++ b/share/man/man9/VFS_CHECKEXP.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 4, 2010
-.Os
 .Dt VFS_CHECKEXP 9
+.Os
 .Sh NAME
 .Nm VFS_CHECKEXP
 .Nd check if a file system is exported to a client
diff --git a/share/man/man9/VFS_FHTOVP.9 b/share/man/man9/VFS_FHTOVP.9
index f6f88d204ee..39a6168f51b 100644
--- a/share/man/man9/VFS_FHTOVP.9
+++ b/share/man/man9/VFS_FHTOVP.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 4, 2010
-.Os
 .Dt VFS_FHTOVP 9
+.Os
 .Sh NAME
 .Nm VFS_FHTOVP
 .Nd turn an NFS filehandle into a vnode
diff --git a/share/man/man9/VFS_MOUNT.9 b/share/man/man9/VFS_MOUNT.9
index 4856dc49ea8..e46c71a5825 100644
--- a/share/man/man9/VFS_MOUNT.9
+++ b/share/man/man9/VFS_MOUNT.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_MOUNT 9
+.Os
 .Sh NAME
 .Nm VFS_MOUNT
 .Nd mount a file system
diff --git a/share/man/man9/VFS_QUOTACTL.9 b/share/man/man9/VFS_QUOTACTL.9
index 48ff866c211..fded4d6b883 100644
--- a/share/man/man9/VFS_QUOTACTL.9
+++ b/share/man/man9/VFS_QUOTACTL.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_QUOTACTL 9
+.Os
 .Sh NAME
 .Nm VFS_QUOTACTL
 .Nd manipulate file system quotas
diff --git a/share/man/man9/VFS_ROOT.9 b/share/man/man9/VFS_ROOT.9
index 2490cfc0bda..f81a500b3ab 100644
--- a/share/man/man9/VFS_ROOT.9
+++ b/share/man/man9/VFS_ROOT.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_ROOT 9
+.Os
 .Sh NAME
 .Nm VFS_ROOT
 .Nd return the root vnode of a file system
diff --git a/share/man/man9/VFS_STATFS.9 b/share/man/man9/VFS_STATFS.9
index 80d0d6652a5..7a7e3714112 100644
--- a/share/man/man9/VFS_STATFS.9
+++ b/share/man/man9/VFS_STATFS.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_STATFS 9
+.Os
 .Sh NAME
 .Nm VFS_STATFS
 .Nd return file system status
diff --git a/share/man/man9/VFS_SYNC.9 b/share/man/man9/VFS_SYNC.9
index 60947a24709..ed510dd7b66 100644
--- a/share/man/man9/VFS_SYNC.9
+++ b/share/man/man9/VFS_SYNC.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_SYNC 9
+.Os
 .Sh NAME
 .Nm VFS_SYNC
 .Nd flush unwritten data
diff --git a/share/man/man9/VFS_UNMOUNT.9 b/share/man/man9/VFS_UNMOUNT.9
index 676685d4a4b..59113ee174d 100644
--- a/share/man/man9/VFS_UNMOUNT.9
+++ b/share/man/man9/VFS_UNMOUNT.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd May 23, 2009
-.Os
 .Dt VFS_UNMOUNT 9
+.Os
 .Sh NAME
 .Nm VFS_UNMOUNT
 .Nd unmount a file system
diff --git a/share/man/man9/VFS_VGET.9 b/share/man/man9/VFS_VGET.9
index 02b55364127..22c902fd694 100644
--- a/share/man/man9/VFS_VGET.9
+++ b/share/man/man9/VFS_VGET.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 7, 2005
-.Os
 .Dt VFS_VGET 9
+.Os
 .Sh NAME
 .Nm VFS_VGET
 .Nd convert an inode number to a vnode
diff --git a/share/man/man9/VOP_ACCESS.9 b/share/man/man9/VOP_ACCESS.9
index 927c1187ce1..005f2e20e57 100644
--- a/share/man/man9/VOP_ACCESS.9
+++ b/share/man/man9/VOP_ACCESS.9
@@ -30,8 +30,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 18, 2009
-.Os
 .Dt VOP_ACCESS 9
+.Os
 .Sh NAME
 .Nm VOP_ACCESS ,
 .Nm VOP_ACCESSX
diff --git a/share/man/man9/VOP_ACLCHECK.9 b/share/man/man9/VOP_ACLCHECK.9
index 62102281ce7..36381121337 100644
--- a/share/man/man9/VOP_ACLCHECK.9
+++ b/share/man/man9/VOP_ACLCHECK.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt VOP_ACLCHECK 9
+.Os
 .Sh NAME
 .Nm VOP_ACLCHECK
 .Nd check an access control list for a vnode
diff --git a/share/man/man9/VOP_ADVLOCK.9 b/share/man/man9/VOP_ADVLOCK.9
index bdb00b7070e..3dabc59ea49 100644
--- a/share/man/man9/VOP_ADVLOCK.9
+++ b/share/man/man9/VOP_ADVLOCK.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 30, 1999
-.Os
 .Dt VOP_ADVLOCK 9
+.Os
 .Sh NAME
 .Nm VOP_ADVLOCK
 .Nd advisory record locking
diff --git a/share/man/man9/VOP_ATTRIB.9 b/share/man/man9/VOP_ATTRIB.9
index f14b1a5d20b..e48e4eb9a25 100644
--- a/share/man/man9/VOP_ATTRIB.9
+++ b/share/man/man9/VOP_ATTRIB.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 29, 2008
-.Os
 .Dt VOP_ATTRIB 9
+.Os
 .Sh NAME
 .Nm VOP_GETATTR ,
 .Nm VOP_SETATTR
diff --git a/share/man/man9/VOP_BWRITE.9 b/share/man/man9/VOP_BWRITE.9
index 0b8f7e0e1c6..d2ee5105166 100644
--- a/share/man/man9/VOP_BWRITE.9
+++ b/share/man/man9/VOP_BWRITE.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_BWRITE 9
+.Os
 .Sh NAME
 .Nm VOP_BWRITE
 .Nd write a file system buffer
diff --git a/share/man/man9/VOP_CREATE.9 b/share/man/man9/VOP_CREATE.9
index 2f30328d2b0..b86a88dce11 100644
--- a/share/man/man9/VOP_CREATE.9
+++ b/share/man/man9/VOP_CREATE.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_CREATE 9
+.Os
 .Sh NAME
 .Nm VOP_CREATE ,
 .Nm VOP_MKNOD ,
diff --git a/share/man/man9/VOP_FSYNC.9 b/share/man/man9/VOP_FSYNC.9
index 54398137c44..e457f846640 100644
--- a/share/man/man9/VOP_FSYNC.9
+++ b/share/man/man9/VOP_FSYNC.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_FSYNC 9
+.Os
 .Sh NAME
 .Nm VOP_FSYNC
 .Nd flush file system buffers for a file
diff --git a/share/man/man9/VOP_GETACL.9 b/share/man/man9/VOP_GETACL.9
index ba819085bb7..d2303408101 100644
--- a/share/man/man9/VOP_GETACL.9
+++ b/share/man/man9/VOP_GETACL.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt VOP_GETACL 9
+.Os
 .Sh NAME
 .Nm VOP_GETACL
 .Nd retrieve access control list for a vnode
diff --git a/share/man/man9/VOP_GETEXTATTR.9 b/share/man/man9/VOP_GETEXTATTR.9
index 0971c27a295..cf00d26f48f 100644
--- a/share/man/man9/VOP_GETEXTATTR.9
+++ b/share/man/man9/VOP_GETEXTATTR.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt VOP_GETEXTATTR 9
+.Os
 .Sh NAME
 .Nm VOP_GETEXTATTR
 .Nd retrieve named extended attribute from a vnode
diff --git a/share/man/man9/VOP_GETPAGES.9 b/share/man/man9/VOP_GETPAGES.9
index ffac609de13..ab0afe8be09 100644
--- a/share/man/man9/VOP_GETPAGES.9
+++ b/share/man/man9/VOP_GETPAGES.9
@@ -30,8 +30,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 27, 2003
-.Os
 .Dt VOP_GETPAGES 9
+.Os
 .Sh NAME
 .Nm VOP_GETPAGES ,
 .Nm VOP_PUTPAGES
diff --git a/share/man/man9/VOP_GETVOBJECT.9 b/share/man/man9/VOP_GETVOBJECT.9
index e2e0ba471a8..2490da710b5 100644
--- a/share/man/man9/VOP_GETVOBJECT.9
+++ b/share/man/man9/VOP_GETVOBJECT.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 10, 2000
-.Os
 .Dt VOP_CREATEVOBJECT 9
+.Os
 .Sh NAME
 .Nm VOP_CREATEVOBJECT ,
 .Nm VOP_DESTROYVOBJECT ,
diff --git a/share/man/man9/VOP_INACTIVE.9 b/share/man/man9/VOP_INACTIVE.9
index 75bd53b8919..2f1e4ed60bd 100644
--- a/share/man/man9/VOP_INACTIVE.9
+++ b/share/man/man9/VOP_INACTIVE.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_INACTIVE 9
+.Os
 .Sh NAME
 .Nm VOP_INACTIVE ,
 .Nm VOP_RECLAIM
diff --git a/share/man/man9/VOP_IOCTL.9 b/share/man/man9/VOP_IOCTL.9
index 7b45484056c..ba1208179fc 100644
--- a/share/man/man9/VOP_IOCTL.9
+++ b/share/man/man9/VOP_IOCTL.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_IOCTL 9
+.Os
 .Sh NAME
 .Nm VOP_IOCTL
 .Nd device specific control
diff --git a/share/man/man9/VOP_LINK.9 b/share/man/man9/VOP_LINK.9
index b7fbb7bd227..f5b7c68e704 100644
--- a/share/man/man9/VOP_LINK.9
+++ b/share/man/man9/VOP_LINK.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_LINK 9
+.Os
 .Sh NAME
 .Nm VOP_LINK
 .Nd create a new name for a file
diff --git a/share/man/man9/VOP_LISTEXTATTR.9 b/share/man/man9/VOP_LISTEXTATTR.9
index 4c8b744efcf..cca875d68ee 100644
--- a/share/man/man9/VOP_LISTEXTATTR.9
+++ b/share/man/man9/VOP_LISTEXTATTR.9
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 19, 2005
-.Os
 .Dt VOP_LISTEXTATTR 9
+.Os
 .Sh NAME
 .Nm VOP_LISTEXTATTR
 .Nd retrieve a list of named extended attribute from a vnode
diff --git a/share/man/man9/VOP_LOCK.9 b/share/man/man9/VOP_LOCK.9
index 6c54a38df5b..b4fdc7ac72b 100644
--- a/share/man/man9/VOP_LOCK.9
+++ b/share/man/man9/VOP_LOCK.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 25, 2008
-.Os
 .Dt VOP_LOCK 9
+.Os
 .Sh NAME
 .Nm VOP_LOCK ,
 .Nm VOP_UNLOCK ,
diff --git a/share/man/man9/VOP_LOOKUP.9 b/share/man/man9/VOP_LOOKUP.9
index ba5a260d36b..64023f4a9e6 100644
--- a/share/man/man9/VOP_LOOKUP.9
+++ b/share/man/man9/VOP_LOOKUP.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 24, 1997
-.Os
 .Dt VOP_LOOKUP 9
+.Os
 .Sh NAME
 .Nm VOP_LOOKUP
 .Nd lookup a component of a pathname
diff --git a/share/man/man9/VOP_OPENCLOSE.9 b/share/man/man9/VOP_OPENCLOSE.9
index 7430da70f3e..5442e3277f4 100644
--- a/share/man/man9/VOP_OPENCLOSE.9
+++ b/share/man/man9/VOP_OPENCLOSE.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 2, 2009
-.Os
 .Dt VOP_OPEN 9
+.Os
 .Sh NAME
 .Nm VOP_OPEN ,
 .Nm VOP_CLOSE
diff --git a/share/man/man9/VOP_PATHCONF.9 b/share/man/man9/VOP_PATHCONF.9
index 056f6479a83..38fed835002 100644
--- a/share/man/man9/VOP_PATHCONF.9
+++ b/share/man/man9/VOP_PATHCONF.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_PATHCONF 9
+.Os
 .Sh NAME
 .Nm VOP_PATHCONF
 .Nd return POSIX pathconf information
diff --git a/share/man/man9/VOP_PRINT.9 b/share/man/man9/VOP_PRINT.9
index 2da85f9a9d7..c8188fc0263 100644
--- a/share/man/man9/VOP_PRINT.9
+++ b/share/man/man9/VOP_PRINT.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_PRINT 9
+.Os
 .Sh NAME
 .Nm VOP_PRINT
 .Nd print debugging information
diff --git a/share/man/man9/VOP_RDWR.9 b/share/man/man9/VOP_RDWR.9
index b0105236785..eb1af784207 100644
--- a/share/man/man9/VOP_RDWR.9
+++ b/share/man/man9/VOP_RDWR.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_RDWR 9
+.Os
 .Sh NAME
 .Nm VOP_READ ,
 .Nm VOP_WRITE
diff --git a/share/man/man9/VOP_READDIR.9 b/share/man/man9/VOP_READDIR.9
index 5b944381162..057600e914d 100644
--- a/share/man/man9/VOP_READDIR.9
+++ b/share/man/man9/VOP_READDIR.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_READDIR 9
+.Os
 .Sh NAME
 .Nm VOP_READDIR
 .Nd read contents of a directory
diff --git a/share/man/man9/VOP_READLINK.9 b/share/man/man9/VOP_READLINK.9
index f98222f2fb1..c1883c1e99c 100644
--- a/share/man/man9/VOP_READLINK.9
+++ b/share/man/man9/VOP_READLINK.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_READLINK 9
+.Os
 .Sh NAME
 .Nm VOP_READLINK
 .Nd read the target of a symbolic link
diff --git a/share/man/man9/VOP_REALLOCBLKS.9 b/share/man/man9/VOP_REALLOCBLKS.9
index 9209844eb8d..d38caf16d72 100644
--- a/share/man/man9/VOP_REALLOCBLKS.9
+++ b/share/man/man9/VOP_REALLOCBLKS.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_REALLOCBLKS 9
+.Os
 .Sh NAME
 .Nm VOP_REALLOCBLKS
 .Nd rearrange blocks in a file to be contiguous
diff --git a/share/man/man9/VOP_REMOVE.9 b/share/man/man9/VOP_REMOVE.9
index 0a10ed222dc..38e3b060bc5 100644
--- a/share/man/man9/VOP_REMOVE.9
+++ b/share/man/man9/VOP_REMOVE.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_REMOVE 9
+.Os
 .Sh NAME
 .Nm VOP_REMOVE ,
 .Nm VOP_RMDIR
diff --git a/share/man/man9/VOP_RENAME.9 b/share/man/man9/VOP_RENAME.9
index 87975852959..76bb0dc1a1c 100644
--- a/share/man/man9/VOP_RENAME.9
+++ b/share/man/man9/VOP_RENAME.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_RENAME 9
+.Os
 .Sh NAME
 .Nm VOP_RENAME
 .Nd rename a file
diff --git a/share/man/man9/VOP_REVOKE.9 b/share/man/man9/VOP_REVOKE.9
index 6e040143534..3557ce1109e 100644
--- a/share/man/man9/VOP_REVOKE.9
+++ b/share/man/man9/VOP_REVOKE.9
@@ -35,8 +35,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 5, 2002
-.Os
 .Dt VOP_REVOKE 9
+.Os
 .Sh NAME
 .Nm VOP_REVOKE
 .Nd "revoke access to a device and its aliases"
diff --git a/share/man/man9/VOP_SETACL.9 b/share/man/man9/VOP_SETACL.9
index b6a7003a86b..482d2898034 100644
--- a/share/man/man9/VOP_SETACL.9
+++ b/share/man/man9/VOP_SETACL.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt VOP_SETACL 9
+.Os
 .Sh NAME
 .Nm VOP_SETACL
 .Nd set the access control list for a vnode
diff --git a/share/man/man9/VOP_SETEXTATTR.9 b/share/man/man9/VOP_SETEXTATTR.9
index ccb8aaad0fb..4bc53877b5a 100644
--- a/share/man/man9/VOP_SETEXTATTR.9
+++ b/share/man/man9/VOP_SETEXTATTR.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt VOP_SETEXTATTR 9
+.Os
 .Sh NAME
 .Nm VOP_SETEXTATTR
 .Nd set named extended attribute for a vnode
diff --git a/share/man/man9/VOP_STRATEGY.9 b/share/man/man9/VOP_STRATEGY.9
index 8c4fbace42a..33d9fb8c146 100644
--- a/share/man/man9/VOP_STRATEGY.9
+++ b/share/man/man9/VOP_STRATEGY.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VOP_STRATEGY 9
+.Os
 .Sh NAME
 .Nm VOP_STRATEGY
 .Nd read or write a file system buffer
diff --git a/share/man/man9/VOP_VPTOCNP.9 b/share/man/man9/VOP_VPTOCNP.9
index 892b4a64ae0..78582151381 100644
--- a/share/man/man9/VOP_VPTOCNP.9
+++ b/share/man/man9/VOP_VPTOCNP.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 7, 2008
-.Os
 .Dt VOP_VPTOCNP 9
+.Os
 .Sh NAME
 .Nm VOP_VPTOCNP
 .Nd translate a vnode to its component name
diff --git a/share/man/man9/VOP_VPTOFH.9 b/share/man/man9/VOP_VPTOFH.9
index 36e703d7958..3619a7b065a 100644
--- a/share/man/man9/VOP_VPTOFH.9
+++ b/share/man/man9/VOP_VPTOFH.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 16, 2007
-.Os
 .Dt VOP_VPTOFH 9
+.Os
 .Sh NAME
 .Nm VOP_VPTOFH
 .Nd turn a vnode into an NFS filehandle
diff --git a/share/man/man9/accept_filter.9 b/share/man/man9/accept_filter.9
index 4a7d239f86d..3f89a0cdf7d 100644
--- a/share/man/man9/accept_filter.9
+++ b/share/man/man9/accept_filter.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd June 25, 2000
-.Os
 .Dt ACCEPT_FILTER 9
+.Os
 .Sh NAME
 .Nm accept_filter ,
 .Nm accept_filt_add ,
diff --git a/share/man/man9/accf_data.9 b/share/man/man9/accf_data.9
index 836deba17a6..2d3b7a24d69 100644
--- a/share/man/man9/accf_data.9
+++ b/share/man/man9/accf_data.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd November 15, 2000
-.Os
 .Dt ACCF_DATA 9
+.Os
 .Sh NAME
 .Nm accf_data
 .Nd buffer incoming connections until data arrives
diff --git a/share/man/man9/accf_dns.9 b/share/man/man9/accf_dns.9
index f4b1563bb73..4a2dad25066 100644
--- a/share/man/man9/accf_dns.9
+++ b/share/man/man9/accf_dns.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd July 16, 2008
-.Os
 .Dt ACCF_DNS 9
+.Os
 .Sh NAME
 .Nm accf_dns
 .Nd buffer incoming DNS requests until the whole first request is present
diff --git a/share/man/man9/accf_http.9 b/share/man/man9/accf_http.9
index a75321c9b1b..2a340605ecf 100644
--- a/share/man/man9/accf_http.9
+++ b/share/man/man9/accf_http.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd November 15, 2000
-.Os
 .Dt ACCF_HTTP 9
+.Os
 .Sh NAME
 .Nm accf_http
 .Nd "buffer incoming connections until a certain complete HTTP requests arrive"
diff --git a/share/man/man9/acl.9 b/share/man/man9/acl.9
index c5192fc8b14..21640bd7b86 100644
--- a/share/man/man9/acl.9
+++ b/share/man/man9/acl.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 18, 2009
-.Os
 .Dt ACL 9
+.Os
 .Sh NAME
 .Nm acl
 .Nd virtual file system access control lists
diff --git a/share/man/man9/atomic.9 b/share/man/man9/atomic.9
index 0bc567c4b9b..94c4ba4d0e0 100644
--- a/share/man/man9/atomic.9
+++ b/share/man/man9/atomic.9
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 27, 2005
-.Os
 .Dt ATOMIC 9
+.Os
 .Sh NAME
 .Nm atomic_add ,
 .Nm atomic_clear ,
diff --git a/share/man/man9/cr_cansee.9 b/share/man/man9/cr_cansee.9
index 53823c16371..c4cc2d4a861 100644
--- a/share/man/man9/cr_cansee.9
+++ b/share/man/man9/cr_cansee.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 19, 2006
-.Os
 .Dt CR_CANSEE 9
+.Os
 .Sh NAME
 .Nm cr_cansee
 .Nd "determine visibility of objects given their user credentials"
diff --git a/share/man/man9/cr_seeothergids.9 b/share/man/man9/cr_seeothergids.9
index 1f5f4ee05db..87fd1396226 100644
--- a/share/man/man9/cr_seeothergids.9
+++ b/share/man/man9/cr_seeothergids.9
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 11, 2003
-.Os
 .Dt CR_SEEOTHERGIDS 9
+.Os
 .Sh NAME
 .Nm cr_seeothergids
 .Nd determine visibility of objects given their group memberships
diff --git a/share/man/man9/cr_seeotheruids.9 b/share/man/man9/cr_seeotheruids.9
index e79363f225e..2066c405e55 100644
--- a/share/man/man9/cr_seeotheruids.9
+++ b/share/man/man9/cr_seeotheruids.9
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 11, 2003
-.Os
 .Dt CR_SEEOTHERUIDS 9
+.Os
 .Sh NAME
 .Nm cr_seeotheruids
 .Nd determine visibility of objects given their user credentials
diff --git a/share/man/man9/devfs_set_cdevpriv.9 b/share/man/man9/devfs_set_cdevpriv.9
index 0896d4de081..fcdc5c4fd41 100644
--- a/share/man/man9/devfs_set_cdevpriv.9
+++ b/share/man/man9/devfs_set_cdevpriv.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 8, 2008
-.Os
 .Dt DEVFS_CDEVPRIV
+.Os
 .Sh NAME
 .Nm devfs_set_cdevpriv ,
 .Nm devfs_get_cdevpriv ,
diff --git a/share/man/man9/devtoname.9 b/share/man/man9/devtoname.9
index 9a62a5ededb..f93e77bccc0 100644
--- a/share/man/man9/devtoname.9
+++ b/share/man/man9/devtoname.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 25, 1999
-.Os
 .Dt DEVTONAME 9
+.Os
 .Sh NAME
 .Nm devtoname
 .Nd "converts dev_t data into a string indicating the device name"
diff --git a/share/man/man9/extattr.9 b/share/man/man9/extattr.9
index 9c74a60bf9b..bf79487f8bf 100644
--- a/share/man/man9/extattr.9
+++ b/share/man/man9/extattr.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 23, 1999
-.Os
 .Dt EXTATTR 9
+.Os
 .Sh NAME
 .Nm extattr
 .Nd virtual file system named extended attributes
diff --git a/share/man/man9/firmware.9 b/share/man/man9/firmware.9
index 87d43e3205a..5254f2448ad 100644
--- a/share/man/man9/firmware.9
+++ b/share/man/man9/firmware.9
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 2, 2008
-.Os
 .Dt FIRMWARE 9
+.Os
 .Sh NAME
 .Nm firmware_register ,
 .Nm firmware_unregister ,
diff --git a/share/man/man9/hexdump.9 b/share/man/man9/hexdump.9
index 656299cd561..a731ab51ab9 100644
--- a/share/man/man9/hexdump.9
+++ b/share/man/man9/hexdump.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 7, 2003
-.Os
 .Dt HEXDUMP 9
+.Os
 .Sh NAME
 .Nm hexdump
 .Nd "dump a block of bytes to the console in hexadecimal form"
diff --git a/share/man/man9/ifnet.9 b/share/man/man9/ifnet.9
index 7a3a3d9d9e9..55e75418536 100644
--- a/share/man/man9/ifnet.9
+++ b/share/man/man9/ifnet.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 1, 2009
-.Os
 .Dt IFNET 9
+.Os
 .Sh NAME
 .Nm ifnet ,
 .Nm ifaddr ,
diff --git a/share/man/man9/make_dev.9 b/share/man/man9/make_dev.9
index b39ca8b833e..50f7d46fc6c 100644
--- a/share/man/man9/make_dev.9
+++ b/share/man/man9/make_dev.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 28, 2008
-.Os
 .Dt MAKE_DEV 9
+.Os
 .Sh NAME
 .Nm make_dev ,
 .Nm make_dev_cred ,
diff --git a/share/man/man9/namei.9 b/share/man/man9/namei.9
index 0325c36fa58..33f753f1449 100644
--- a/share/man/man9/namei.9
+++ b/share/man/man9/namei.9
@@ -34,8 +34,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 21, 2005
-.Os
 .Dt NAMEI 9
+.Os
 .Sh NAME
 .Nm namei ,
 .Nm NDINIT ,
diff --git a/share/man/man9/p_candebug.9 b/share/man/man9/p_candebug.9
index 2127fbc4a28..6249a4de410 100644
--- a/share/man/man9/p_candebug.9
+++ b/share/man/man9/p_candebug.9
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 19, 2006
-.Os
 .Dt P_CANDEBUG 9
+.Os
 .Sh NAME
 .Nm p_candebug
 .Nd determine debuggability of a process
diff --git a/share/man/man9/p_cansee.9 b/share/man/man9/p_cansee.9
index 92ffeeaaa78..fafe9d3a1fb 100644
--- a/share/man/man9/p_cansee.9
+++ b/share/man/man9/p_cansee.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 19, 2006
-.Os
 .Dt P_CANSEE 9
+.Os
 .Sh NAME
 .Nm p_cansee
 .Nd determine visibility of a process
diff --git a/share/man/man9/pfind.9 b/share/man/man9/pfind.9
index 58adab3fad3..dfd15b8dcf7 100644
--- a/share/man/man9/pfind.9
+++ b/share/man/man9/pfind.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 11, 2001
-.Os
 .Dt PFIND 9
+.Os
 .Sh NAME
 .Nm pfind , zpfind
 .Nd locate a process by number
diff --git a/share/man/man9/pgfind.9 b/share/man/man9/pgfind.9
index 96abd0ae521..a3049876d72 100644
--- a/share/man/man9/pgfind.9
+++ b/share/man/man9/pgfind.9
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 8, 2001
-.Os
 .Dt PGFIND 9
+.Os
 .Sh NAME
 .Nm pgfind
 .Nd "locate a process group by number"
diff --git a/share/man/man9/prison_check.9 b/share/man/man9/prison_check.9
index a8ba040131a..68e90fb7fef 100644
--- a/share/man/man9/prison_check.9
+++ b/share/man/man9/prison_check.9
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 11, 2003
-.Os
 .Dt PRISON_CHECK 9
+.Os
 .Sh NAME
 .Nm prison_check
 .Nd determine if two credentials belong to the same jail
diff --git a/share/man/man9/random.9 b/share/man/man9/random.9
index 426774e9083..1d7c7daf089 100644
--- a/share/man/man9/random.9
+++ b/share/man/man9/random.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd September 25, 2000
-.Os
 .Dt RANDOM 9
+.Os
 .Sh NAME
 .Nm arc4rand ,
 .Nm arc4random ,
diff --git a/share/man/man9/rijndael.9 b/share/man/man9/rijndael.9
index 5ee423971de..48b402bc0f1 100644
--- a/share/man/man9/rijndael.9
+++ b/share/man/man9/rijndael.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd February 6, 2002
-.Os
 .Dt RIJNDAEL 9
+.Os
 .Sh NAME
 .Nm rijndael_makeKey ,
 .Nm rijndael_cipherInit ,
diff --git a/share/man/man9/rtalloc.9 b/share/man/man9/rtalloc.9
index cbd2c5eaab5..31d5ebbdf67 100644
--- a/share/man/man9/rtalloc.9
+++ b/share/man/man9/rtalloc.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 11, 2008
-.Os
 .Dt RTALLOC 9
+.Os
 .Sh NAME
 .Nm rtalloc ,
 .Nm rtalloc_ign ,
diff --git a/share/man/man9/rtentry.9 b/share/man/man9/rtentry.9
index 0e25b21acfc..d7e6953f5f8 100644
--- a/share/man/man9/rtentry.9
+++ b/share/man/man9/rtentry.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 11, 2008
-.Os
 .Dt RTENTRY 9
+.Os
 .Sh NAME
 .Nm rtentry
 .Nd structure of an entry in the kernel routing table
diff --git a/share/man/man9/sleep.9 b/share/man/man9/sleep.9
index 8fb7f88a320..91a212166bd 100644
--- a/share/man/man9/sleep.9
+++ b/share/man/man9/sleep.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 12, 2009
-.Os
 .Dt SLEEP 9
+.Os
 .Sh NAME
 .Nm msleep ,
 .Nm msleep_spin ,
diff --git a/share/man/man9/spl.9 b/share/man/man9/spl.9
index fe4bec27f57..95a5fc25aff 100644
--- a/share/man/man9/spl.9
+++ b/share/man/man9/spl.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 21, 1996
-.Os
 .Dt SPL 9
+.Os
 .Sh NAME
 .Nm splbio ,
 .Nm splclock ,
diff --git a/share/man/man9/uio.9 b/share/man/man9/uio.9
index b5c12cbe3bd..0ddb2d39f99 100644
--- a/share/man/man9/uio.9
+++ b/share/man/man9/uio.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 21, 2010
-.Os
 .Dt UIO 9
+.Os
 .Sh NAME
 .Nm uio ,
 .Nm uiomove
diff --git a/share/man/man9/usbdi.9 b/share/man/man9/usbdi.9
index 100c53c439b..a26e026640c 100644
--- a/share/man/man9/usbdi.9
+++ b/share/man/man9/usbdi.9
@@ -25,8 +25,8 @@
 .\"
 .\"	$FreeBSD$
 .Dd June 24, 2009
-.Os
 .Dt USBDI 9
+.Os
 .Sh NAME
 .Nm usb_fifo_alloc_buffer ,
 .Nm usb_fifo_attach ,
diff --git a/share/man/man9/vaccess.9 b/share/man/man9/vaccess.9
index a2e2ec5b72a..4bf37740690 100644
--- a/share/man/man9/vaccess.9
+++ b/share/man/man9/vaccess.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 18, 2009
-.Os
 .Dt VACCESS 9
+.Os
 .Sh NAME
 .Nm vaccess
 .Nd generate an access control decision using vnode parameters
diff --git a/share/man/man9/vaccess_acl_nfs4.9 b/share/man/man9/vaccess_acl_nfs4.9
index 32cbdb3a940..27b4cfa644b 100644
--- a/share/man/man9/vaccess_acl_nfs4.9
+++ b/share/man/man9/vaccess_acl_nfs4.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 18, 2009
-.Os
 .Dt VACCESS_ACL_NFS4 9
+.Os
 .Sh NAME
 .Nm vaccess_acl_nfs4
 .Nd generate a NFSv4 ACL access control decision using vnode parameters
diff --git a/share/man/man9/vaccess_acl_posix1e.9 b/share/man/man9/vaccess_acl_posix1e.9
index 06105113260..dc857a8ee61 100644
--- a/share/man/man9/vaccess_acl_posix1e.9
+++ b/share/man/man9/vaccess_acl_posix1e.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 22, 2001
-.Os
 .Dt VACCESS_ACL_POSIX1E 9
+.Os
 .Sh NAME
 .Nm vaccess_acl_posix1e
 .Nd generate a POSIX.1e ACL access control decision using vnode parameters
diff --git a/share/man/man9/vcount.9 b/share/man/man9/vcount.9
index 610891b6a8d..41263b9a1a3 100644
--- a/share/man/man9/vcount.9
+++ b/share/man/man9/vcount.9
@@ -33,8 +33,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 6, 2001
-.Os
 .Dt VCOUNT 9
+.Os
 .Sh NAME
 .Nm vcount ,
 .Nm count_dev
diff --git a/share/man/man9/vget.9 b/share/man/man9/vget.9
index ef8fb10dd05..1a7d16cdc37 100644
--- a/share/man/man9/vget.9
+++ b/share/man/man9/vget.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VGET 9
+.Os
 .Sh NAME
 .Nm vget
 .Nd get a vnode from the free list
diff --git a/share/man/man9/vm_map_entry_resize_free.9 b/share/man/man9/vm_map_entry_resize_free.9
index b037bcb4562..eaaec650fb6 100644
--- a/share/man/man9/vm_map_entry_resize_free.9
+++ b/share/man/man9/vm_map_entry_resize_free.9
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 17, 2004
-.Os
 .Dt VM_MAP_ENTRY_RESIZE_FREE 9
+.Os
 .Sh NAME
 .Nm vm_map_entry_resize_free
 .Nd "vm map free space algorithm"
diff --git a/share/man/man9/vnode.9 b/share/man/man9/vnode.9
index 49a2cc77d1c..cfc6f7d5f7f 100644
--- a/share/man/man9/vnode.9
+++ b/share/man/man9/vnode.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 13, 2010
-.Os
 .Dt VNODE 9
+.Os
 .Sh NAME
 .Nm vnode
 .Nd internal representation of a file or directory
diff --git a/share/man/man9/vput.9 b/share/man/man9/vput.9
index 47bee09b75e..d915539d4b7 100644
--- a/share/man/man9/vput.9
+++ b/share/man/man9/vput.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VPUT 9
+.Os
 .Sh NAME
 .Nm vput
 .Nd "decrement the use count for a vnode and unlock it"
diff --git a/share/man/man9/vref.9 b/share/man/man9/vref.9
index d979fb3f2c7..f21b9f269f3 100644
--- a/share/man/man9/vref.9
+++ b/share/man/man9/vref.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VREF 9
+.Os
 .Sh NAME
 .Nm vref
 .Nd increment the use count for a vnode
diff --git a/share/man/man9/vrefcnt.9 b/share/man/man9/vrefcnt.9
index fbf43df7eed..de7af26fb9c 100644
--- a/share/man/man9/vrefcnt.9
+++ b/share/man/man9/vrefcnt.9
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 10, 2008
-.Os
 .Dt VREFCNT 9
+.Os
 .Sh NAME
 .Nm vrefcnt
 .Nd returns the use count of a vnode
diff --git a/share/man/man9/vrele.9 b/share/man/man9/vrele.9
index 249114900b1..be5d5f7ade7 100644
--- a/share/man/man9/vrele.9
+++ b/share/man/man9/vrele.9
@@ -29,8 +29,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 24, 1996
-.Os
 .Dt VRELE 9
+.Os
 .Sh NAME
 .Nm vrele
 .Nd decrement the use count for a vnode
diff --git a/usr.bin/ar/ar.1 b/usr.bin/ar/ar.1
index cb9414e4b7b..250cd20f90b 100644
--- a/usr.bin/ar/ar.1
+++ b/usr.bin/ar/ar.1
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 31, 2007
-.Os
 .Dt AR 1
+.Os
 .Sh NAME
 .Nm ar ,
 .Nm ranlib
diff --git a/usr.bin/c89/c89.1 b/usr.bin/c89/c89.1
index 302aa321dad..82acf37fbca 100644
--- a/usr.bin/c89/c89.1
+++ b/usr.bin/c89/c89.1
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 17, 1997
-.Os
 .Dt C89 1
+.Os
 .Sh NAME
 .Nm c89
 .Nd POSIX.2 C language compiler
diff --git a/usr.bin/c99/c99.1 b/usr.bin/c99/c99.1
index c56e5856e35..e5207e5a058 100644
--- a/usr.bin/c99/c99.1
+++ b/usr.bin/c99/c99.1
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd October 7, 2002
-.Os
 .Dt C99 1
+.Os
 .Sh NAME
 .Nm c99
 .Nd standard C language compiler
diff --git a/usr.bin/column/column.1 b/usr.bin/column/column.1
index 5c5a100fe3b..d4418c5ec75 100644
--- a/usr.bin/column/column.1
+++ b/usr.bin/column/column.1
@@ -33,8 +33,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 29, 2004
-.Os
 .Dt COLUMN 1
+.Os
 .Sh NAME
 .Nm column
 .Nd columnate lists
diff --git a/usr.bin/comm/comm.1 b/usr.bin/comm/comm.1
index 02b508cf626..67b5eecca57 100644
--- a/usr.bin/comm/comm.1
+++ b/usr.bin/comm/comm.1
@@ -36,8 +36,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 12, 2009
-.Os
 .Dt COMM 1
+.Os
 .Sh NAME
 .Nm comm
 .Nd select or reject lines common to two files
diff --git a/usr.bin/csup/cpasswd.1 b/usr.bin/csup/cpasswd.1
index 973dc808f28..1f486d18dcd 100644
--- a/usr.bin/csup/cpasswd.1
+++ b/usr.bin/csup/cpasswd.1
@@ -30,8 +30,8 @@
 .\" $FreeBSD$
 .\"
 .Dd June 27, 2007
-.Os FreeBSD
 .Dt CPASSWD 1
+.Os FreeBSD
 .Sh NAME
 .Nm cpasswd
 .Nd scramble passwords for csup authentication
diff --git a/usr.bin/csup/csup.1 b/usr.bin/csup/csup.1
index 2690863bde0..02e2edb168f 100644
--- a/usr.bin/csup/csup.1
+++ b/usr.bin/csup/csup.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 1, 2006
-.Os FreeBSD
 .Dt CSUP 1
+.Os FreeBSD
 .Sh NAME
 .Nm csup
 .Nd network distribution package for CVS repositories
diff --git a/usr.bin/enigma/enigma.1 b/usr.bin/enigma/enigma.1
index 4fc0b206fe9..e5f65ac4a4d 100644
--- a/usr.bin/enigma/enigma.1
+++ b/usr.bin/enigma/enigma.1
@@ -7,8 +7,8 @@
 .\" $FreeBSD$
 .\" "
 .Dd May 14, 2004
-.Os
 .Dt ENIGMA 1
+.Os
 .Sh NAME
 .Nm enigma ,
 .Nm crypt
diff --git a/usr.bin/hexdump/od.1 b/usr.bin/hexdump/od.1
index dd67ff58b23..b5b5494a47a 100644
--- a/usr.bin/hexdump/od.1
+++ b/usr.bin/hexdump/od.1
@@ -33,8 +33,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 18, 2010
-.Os
 .Dt OD 1
+.Os
 .Sh NAME
 .Nm od
 .Nd octal, decimal, hex, ASCII dump
diff --git a/usr.bin/killall/killall.1 b/usr.bin/killall/killall.1
index 19362bc6d65..58ffa068b9f 100644
--- a/usr.bin/killall/killall.1
+++ b/usr.bin/killall/killall.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 25, 2009
-.Os
 .Dt KILLALL 1
+.Os
 .Sh NAME
 .Nm killall
 .Nd kill processes by name
diff --git a/usr.bin/lockf/lockf.1 b/usr.bin/lockf/lockf.1
index 6e2e06eae58..255a8283a1f 100644
--- a/usr.bin/lockf/lockf.1
+++ b/usr.bin/lockf/lockf.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd July 7, 1998
-.Os
 .Dt LOCKF 1
+.Os
 .Sh NAME
 .Nm lockf
 .Nd execute a command while holding a file lock
diff --git a/usr.bin/wtmpcvt/wtmpcvt.1 b/usr.bin/wtmpcvt/wtmpcvt.1
index fdc995b4d5c..50f89eeb5c5 100644
--- a/usr.bin/wtmpcvt/wtmpcvt.1
+++ b/usr.bin/wtmpcvt/wtmpcvt.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 14, 2010
-.Os
 .Dt WTMPCVT 1
+.Os
 .Sh NAME
 .Nm wtmpcvt
 .Nd convert wtmp files to the utmpx format
diff --git a/usr.sbin/asf/asf.8 b/usr.sbin/asf/asf.8
index 25b03089c7a..92edb4b790d 100644
--- a/usr.sbin/asf/asf.8
+++ b/usr.sbin/asf/asf.8
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 20, 2006
-.Os
 .Dt ASF 8
+.Os
 .Sh NAME
 .Nm asf
 .Nd add symbol files
diff --git a/usr.sbin/burncd/burncd.8 b/usr.sbin/burncd/burncd.8
index 2501e7209c7..a9c53e62774 100644
--- a/usr.sbin/burncd/burncd.8
+++ b/usr.sbin/burncd/burncd.8
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 21, 2009
-.Os
 .Dt BURNCD 8
+.Os
 .Sh NAME
 .Nm burncd
 .Nd control the ATAPI CD-R/RW driver
diff --git a/usr.sbin/ctm/ctm/ctm.1 b/usr.sbin/ctm/ctm/ctm.1
index 732b0b9a968..d5cc6144d1d 100644
--- a/usr.sbin/ctm/ctm/ctm.1
+++ b/usr.sbin/ctm/ctm/ctm.1
@@ -13,8 +13,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 25, 1995
-.Os
 .Dt CTM 1
+.Os
 .Sh NAME
 .Nm ctm
 .Nd source code mirror program
diff --git a/usr.sbin/ctm/ctm/ctm.5 b/usr.sbin/ctm/ctm/ctm.5
index 678837e6182..a51c93b0069 100644
--- a/usr.sbin/ctm/ctm/ctm.5
+++ b/usr.sbin/ctm/ctm/ctm.5
@@ -13,8 +13,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 25, 1995
-.Os
 .Dt CTM 5
+.Os
 .Sh NAME
 .Nm ctm
 .Nd source code mirror system
diff --git a/usr.sbin/devinfo/devinfo.8 b/usr.sbin/devinfo/devinfo.8
index 6c893dc66a5..ecb3b012637 100644
--- a/usr.sbin/devinfo/devinfo.8
+++ b/usr.sbin/devinfo/devinfo.8
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 28, 2005
-.Os
 .Dt DEVINFO 8
+.Os
 .Sh NAME
 .Nm devinfo
 .Nd print information about system device configuration
diff --git a/usr.sbin/fdformat/fdformat.1 b/usr.sbin/fdformat/fdformat.1
index 25891c0f003..596b09b6cc0 100644
--- a/usr.sbin/fdformat/fdformat.1
+++ b/usr.sbin/fdformat/fdformat.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd December 25, 2001
-.Os
 .Dt FDFORMAT 1
+.Os
 .Sh NAME
 .Nm fdformat
 .Nd format floppy disks
diff --git a/usr.sbin/fdread/fdread.1 b/usr.sbin/fdread/fdread.1
index 20a79fc1453..d29a4a44dd6 100644
--- a/usr.sbin/fdread/fdread.1
+++ b/usr.sbin/fdread/fdread.1
@@ -27,8 +27,8 @@
 .\"
 .\"
 .Dd May 14, 2001
-.Os
 .Dt FDREAD 1
+.Os
 .Sh NAME
 .Nm fdread
 .Nd read floppy disks
diff --git a/usr.sbin/fdwrite/fdwrite.1 b/usr.sbin/fdwrite/fdwrite.1
index 28df514f264..e8473983bff 100644
--- a/usr.sbin/fdwrite/fdwrite.1
+++ b/usr.sbin/fdwrite/fdwrite.1
@@ -10,8 +10,8 @@
 .\"
 .\"
 .Dd September 16, 1993
-.Os
 .Dt FDWRITE 1
+.Os
 .Sh NAME
 .Nm fdwrite
 .Nd format and write floppy disks
diff --git a/usr.sbin/fifolog/fifolog_create/fifolog.1 b/usr.sbin/fifolog/fifolog_create/fifolog.1
index 7db58e71840..d0be0ce9bb3 100644
--- a/usr.sbin/fifolog/fifolog_create/fifolog.1
+++ b/usr.sbin/fifolog/fifolog_create/fifolog.1
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 9, 2008
-.Os
 .Dt FIFOLOG 1
+.Os
 .Sh NAME
 .Nm fifolog_create , fifolog_writer , fifolog_reader
 .Nd "initialize, write, seek and extract data from a fifolog"
diff --git a/usr.sbin/flowctl/flowctl.8 b/usr.sbin/flowctl/flowctl.8
index e72486bbf80..75304c45910 100644
--- a/usr.sbin/flowctl/flowctl.8
+++ b/usr.sbin/flowctl/flowctl.8
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd March 23, 2005
-.Os
 .Dt FLOWCTL 8
+.Os
 .Sh NAME
 .Nm flowctl
 .Nd
diff --git a/usr.sbin/mtest/mtest.8 b/usr.sbin/mtest/mtest.8
index b4d3514582c..9a961e01846 100644
--- a/usr.sbin/mtest/mtest.8
+++ b/usr.sbin/mtest/mtest.8
@@ -27,8 +27,8 @@
 .\" $FreeBSD$
 .\"
 .Dd April 29, 2009
-.Os
 .Dt MTEST 8
+.Os
 .Sh NAME
 .Nm mtest
 .Nd test multicast socket operations
diff --git a/usr.sbin/periodic/periodic.8 b/usr.sbin/periodic/periodic.8
index 63359308656..22a96a2a647 100644
--- a/usr.sbin/periodic/periodic.8
+++ b/usr.sbin/periodic/periodic.8
@@ -25,8 +25,8 @@
 .\" $FreeBSD$
 .\"
 .Dd August 30, 2007
-.Os
 .Dt PERIODIC 8
+.Os
 .Sh NAME
 .Nm periodic
 .Nd run periodic system functions
diff --git a/usr.sbin/pmcannotate/pmcannotate.8 b/usr.sbin/pmcannotate/pmcannotate.8
index adadae437fa..17c7c0ed5d6 100644
--- a/usr.sbin/pmcannotate/pmcannotate.8
+++ b/usr.sbin/pmcannotate/pmcannotate.8
@@ -28,8 +28,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 20, 2008
-.Os
 .Dt PMCANNOTATE 8
+.Os
 .Sh NAME
 .Nm pmcannotate
 .Nd "sources printout with inlined profiling"
diff --git a/usr.sbin/pmccontrol/pmccontrol.8 b/usr.sbin/pmccontrol/pmccontrol.8
index e521a1f4702..f436753dd37 100644
--- a/usr.sbin/pmccontrol/pmccontrol.8
+++ b/usr.sbin/pmccontrol/pmccontrol.8
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd November 9, 2008
-.Os
 .Dt PMCCONTROL 8
+.Os
 .Sh NAME
 .Nm pmccontrol
 .Nd "control hardware performance monitoring counters"
diff --git a/usr.sbin/pmcstat/pmcstat.8 b/usr.sbin/pmcstat/pmcstat.8
index 309eb3eae2d..b033357fa42 100644
--- a/usr.sbin/pmcstat/pmcstat.8
+++ b/usr.sbin/pmcstat/pmcstat.8
@@ -26,8 +26,8 @@
 .\" $FreeBSD$
 .\"
 .Dd September 19, 2008
-.Os
 .Dt PMCSTAT 8
+.Os
 .Sh NAME
 .Nm pmcstat
 .Nd "performance measurement with performance monitoring hardware"
diff --git a/usr.sbin/pppctl/pppctl.8 b/usr.sbin/pppctl/pppctl.8
index 245690f17f1..431b7d76c17 100644
--- a/usr.sbin/pppctl/pppctl.8
+++ b/usr.sbin/pppctl/pppctl.8
@@ -1,7 +1,7 @@
 .\" $FreeBSD$
 .Dd June 26, 1997
-.Os
 .Dt PPPCTL 8
+.Os
 .Sh NAME
 .Nm pppctl
 .Nd PPP control program
diff --git a/usr.sbin/setfmac/setfsmac.8 b/usr.sbin/setfmac/setfsmac.8
index 7d4150f9447..9bdb0e8f7ec 100644
--- a/usr.sbin/setfmac/setfsmac.8
+++ b/usr.sbin/setfmac/setfsmac.8
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd February 17, 2004
-.Os
 .Dt SETFSMAC 8
+.Os
 .Sh NAME
 .Nm setfsmac
 .Nd set MAC label for a file hierarchy
diff --git a/usr.sbin/setpmac/setpmac.8 b/usr.sbin/setpmac/setpmac.8
index e84a8bc6322..c5c2963c557 100644
--- a/usr.sbin/setpmac/setpmac.8
+++ b/usr.sbin/setpmac/setpmac.8
@@ -31,8 +31,8 @@
 .\" $FreeBSD$
 .\"
 .Dd January 14, 2003
-.Os
 .Dt SETPMAC 8
+.Os
 .Sh NAME
 .Nm setpmac
 .Nd "run a command with a different MAC process label"
diff --git a/usr.sbin/uhsoctl/uhsoctl.1 b/usr.sbin/uhsoctl/uhsoctl.1
index e01c1a2af49..cd0c5ca66f8 100644
--- a/usr.sbin/uhsoctl/uhsoctl.1
+++ b/usr.sbin/uhsoctl/uhsoctl.1
@@ -24,8 +24,8 @@
 .\" $FreeBSD$
 .\"
 .Dd Aug 12, 2009
-.Os
 .Dt UHSOCTL 1
+.Os
 .Sh NAME
 .Nm uhsoctl
 .Nd connection utility for Option based devices

From b71e04d3a8557b54d33fcc1bd4e4fa07cc7f856b Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:04:55 +0000
Subject: [PATCH 048/532] ld_gs_base is executing with stack containing only
 the frame, temporary pushed %rflags has been popped already.

Pointy hat to:	kib
MFC after:	3 days
---
 sys/amd64/amd64/exception.S | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/amd64/amd64/exception.S b/sys/amd64/amd64/exception.S
index bbc304a48d6..0d5a50ddc0c 100644
--- a/sys/amd64/amd64/exception.S
+++ b/sys/amd64/amd64/exception.S
@@ -815,7 +815,6 @@ fsbase_load_fault:
 	ALIGN_TEXT
 	.globl	gsbase_load_fault
 gsbase_load_fault:
-	popfq
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
 	movq	%rsp, %rdi
 	call	trap

From d578fbb66e2c49ea820bdd44730717709c4217b9 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:31:06 +0000
Subject: [PATCH 049/532] Make this directory more regular.  Since it is one we
 control, use the freebsd-based names for filenames.  This allows us to
 eliminate almost all of the uses of ${MACHINE_ARCH} here to do special
 things, and instead we use it to include filenames.  This makes new
 architectures easier to support.

---
 gnu/usr.bin/gdb/gdbserver/Makefile              | 17 +++--------------
 .../{fbsd-ppc-low.c => fbsd-powerpc-low.c}      |  0
 .../gdb/gdbserver/{reg-x86-64.c => reg-amd64.c} |  0
 .../gdb/gdbserver/{reg-ppc.c => reg-powerpc.c}  |  0
 4 files changed, 3 insertions(+), 14 deletions(-)
 rename gnu/usr.bin/gdb/gdbserver/{fbsd-ppc-low.c => fbsd-powerpc-low.c} (100%)
 rename gnu/usr.bin/gdb/gdbserver/{reg-x86-64.c => reg-amd64.c} (100%)
 rename gnu/usr.bin/gdb/gdbserver/{reg-ppc.c => reg-powerpc.c} (100%)

diff --git a/gnu/usr.bin/gdb/gdbserver/Makefile b/gnu/usr.bin/gdb/gdbserver/Makefile
index 390ef32692b..08f25ae61a5 100644
--- a/gnu/usr.bin/gdb/gdbserver/Makefile
+++ b/gnu/usr.bin/gdb/gdbserver/Makefile
@@ -14,20 +14,9 @@ SRCS=	inferiors.c mem-break.c regcache.c remote-utils.c	\
 	server.c signals.c target.c utils.c
 SRCS+=	fbsd-low.c
 
-.if ${MACHINE_ARCH} == "amd64"
-SRCS+= fbsd-amd64-low.c i387-fp.c reg-x86-64.c
-.endif
-
-.if ${MACHINE_ARCH} == "arm"
-SRCS+= fbsd-arm-low.c reg-arm.c
-.endif
-
-.if ${MACHINE_ARCH} == "i386"
-SRCS+= fbsd-i386-low.c i387-fp.c reg-i386.c
-.endif
-
-.if ${MACHINE_ARCH} == "powerpc"
-SRCS+= fbsd-ppc-low.c reg-ppc.c
+SRCS+= fbsd-${MACHINE_ARCH}-low.c reg-${MACHINE_ARCH}.c
+.if ${MACHINE_ARCH} == "amd64" || ${MACHINE_ARCH} == "i386"
+SRCS+= i387-fp.c
 .endif
 
 #CFLAGS+=	-I${.CURDIR}/../arch/${MACHINE_ARCH}
diff --git a/gnu/usr.bin/gdb/gdbserver/fbsd-ppc-low.c b/gnu/usr.bin/gdb/gdbserver/fbsd-powerpc-low.c
similarity index 100%
rename from gnu/usr.bin/gdb/gdbserver/fbsd-ppc-low.c
rename to gnu/usr.bin/gdb/gdbserver/fbsd-powerpc-low.c
diff --git a/gnu/usr.bin/gdb/gdbserver/reg-x86-64.c b/gnu/usr.bin/gdb/gdbserver/reg-amd64.c
similarity index 100%
rename from gnu/usr.bin/gdb/gdbserver/reg-x86-64.c
rename to gnu/usr.bin/gdb/gdbserver/reg-amd64.c
diff --git a/gnu/usr.bin/gdb/gdbserver/reg-ppc.c b/gnu/usr.bin/gdb/gdbserver/reg-powerpc.c
similarity index 100%
rename from gnu/usr.bin/gdb/gdbserver/reg-ppc.c
rename to gnu/usr.bin/gdb/gdbserver/reg-powerpc.c

From d193ed0bedd2c0f08defb084b803662f8ce6f2cf Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:45:33 +0000
Subject: [PATCH 050/532] Add driver for Silicon Integrated Systems SiS190/191
 Fast/Gigabit Ethernet. This driver was written by Alexander Pohoyda and
 greatly enhanced by Nikolay Denev. I don't have these hardwares but this
 driver was tested by Nikolay Denev and xclin.

Because SiS didn't release data sheet for this controller, programming
information came from Linux driver and OpenSolaris. Unlike other open
source driver for SiS190/191, sge(4) takes full advantage of TX/RX
checksum offloading and does not require additional copy operation in
RX handler.
The controller seems to have advanced offloading features like VLAN
hardware tag insertion/stripping, TCP segmentation offload(TSO) as
well as jumbo frame support but these features are not available
yet. Special thanks to xclin <xclin<> cs dot nctu dot edu dot tw>
who sent fix for receiving VLAN oversized frames.
---
 sys/amd64/conf/GENERIC     |    1 +
 sys/boot/forth/loader.conf |    1 +
 sys/conf/NOTES             |    2 +
 sys/conf/files             |    1 +
 sys/dev/sge/if_sge.c       | 1745 ++++++++++++++++++++++++++++++++++++
 sys/dev/sge/if_sgereg.h    |  350 ++++++++
 sys/i386/conf/GENERIC      |    1 +
 sys/modules/Makefile       |    1 +
 sys/modules/sge/Makefile   |    8 +
 9 files changed, 2110 insertions(+)
 create mode 100644 sys/dev/sge/if_sge.c
 create mode 100644 sys/dev/sge/if_sgereg.h
 create mode 100644 sys/modules/sge/Makefile

diff --git a/sys/amd64/conf/GENERIC b/sys/amd64/conf/GENERIC
index 0d5cad781e4..f6d0a600bf2 100644
--- a/sys/amd64/conf/GENERIC
+++ b/sys/amd64/conf/GENERIC
@@ -224,6 +224,7 @@ device		pcn		# AMD Am79C97x PCI 10/100 (precedence over 'le')
 device		re		# RealTek 8139C+/8169/8169S/8110S
 device		rl		# RealTek 8129/8139
 device		sf		# Adaptec AIC-6915 (``Starfire'')
+device		sge		# Silicon Integrated Systems SiS190/191
 device		sis		# Silicon Integrated Systems SiS 900/SiS 7016
 device		sk		# SysKonnect SK-984x & SK-982x gigabit Ethernet
 device		ste		# Sundance ST201 (D-Link DFE-550TX)
diff --git a/sys/boot/forth/loader.conf b/sys/boot/forth/loader.conf
index d7f1323be9b..e0e4409f03b 100644
--- a/sys/boot/forth/loader.conf
+++ b/sys/boot/forth/loader.conf
@@ -271,6 +271,7 @@ if_rl_load="NO"			# RealTek 8129/8139
 if_rue_load="NO"		# RealTek RTL8150 USB to Fast Ethernet
 if_sbni_load="NO"		# Granch SBNI12 leased line adapters
 if_sf_load="NO"			# Adaptec Duralink PCI (AIC-6915 "starfire")
+if_sge_load="NO"		# Silicon Integrated Systems SiS190/191
 if_sis_load="NO"		# Silicon Integrated Systems SiS 900/7016
 if_sk_load="NO"			# SysKonnect SK-984x series PCI Gigabit Ethernet
 if_sn_load="NO"			# SMC 91Cxx
diff --git a/sys/conf/NOTES b/sys/conf/NOTES
index 8a2023ddc1c..f0cc9a0bc10 100644
--- a/sys/conf/NOTES
+++ b/sys/conf/NOTES
@@ -1952,6 +1952,7 @@ device		xmphy		# XaQti XMAC II
 #       This includes dual and quad port cards, as well as one 100baseFX card.
 #       Most of these are 64-bit PCI devices, except for one single port
 #       card which is 32-bit.
+# sge:  Silicon Integrated Systems SiS190/191 Fast/Gigabit Ethernet adapter
 # sis:  Support for NICs based on the Silicon Integrated Systems SiS 900,
 #       SiS 7016 and NS DP83815 PCI fast ethernet controller chips.
 # sk:   Support for the SysKonnect SK-984x series PCI gigabit ethernet NICs.
@@ -2046,6 +2047,7 @@ device		re		# RealTek 8139C+/8169/8169S/8110S
 device		rl		# RealTek 8129/8139
 device		pcn		# AMD Am79C97x PCI 10/100 NICs
 device		sf		# Adaptec AIC-6915 (``Starfire'')
+device		sge		# Silicon Integrated Systems SiS190/191
 device		sis		# Silicon Integrated Systems SiS 900/SiS 7016
 device		sk		# SysKonnect SK-984x & SK-982x gigabit Ethernet
 device		ste		# Sundance ST201 (D-Link DFE-550TX)
diff --git a/sys/conf/files b/sys/conf/files
index 7dd1b0ed3cd..6bb1355b919 100644
--- a/sys/conf/files
+++ b/sys/conf/files
@@ -1490,6 +1490,7 @@ dev/scd/scd.c			optional scd isa
 dev/scd/scd_isa.c		optional scd isa
 dev/sdhci/sdhci.c		optional sdhci pci
 dev/sf/if_sf.c			optional sf pci
+dev/sge/if_sge.c		optional sge pci
 dev/si/si.c			optional si
 dev/si/si2_z280.c		optional si
 dev/si/si3_t225.c		optional si
diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
new file mode 100644
index 00000000000..a26a6a46359
--- /dev/null
+++ b/sys/dev/sge/if_sge.c
@@ -0,0 +1,1745 @@
+/*-
+ * Copyright (c) 2008-2010 Nikolay Denev <ndenev@gmail.com>
+ * Copyright (c) 2007-2008 Alexander Pohoyda <alexander.pohoyda@gmx.net>
+ * Copyright (c) 1997, 1998, 1999
+ *	Bill Paul <wpaul@ctr.columbia.edu>.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *	This product includes software developed by Bill Paul.
+ * 4. Neither the name of the author nor the names of any co-contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS''
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+ * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+ * PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL AUTHORS OR
+ * THE VOICES IN THEIR HEADS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+/*
+ * SiS 190/191 PCI Ethernet NIC driver.
+ *
+ * Adapted to SiS 190 NIC by Alexander Pohoyda based on the original
+ * SiS 900 driver by Bill Paul, using SiS 190/191 Solaris driver by
+ * Masayuki Murayama and SiS 190/191 GNU/Linux driver by K.M. Liu
+ * <kmliu@sis.com>.  Thanks to Pyun YongHyeon <pyunyh@gmail.com> for
+ * review and very useful comments.
+ *
+ * Adapted to SiS 191 NIC by Nikolay Denev with further ideas from the
+ * Linux and Solaris drivers.
+ */
+
+#include <sys/param.h>
+#include <sys/systm.h>
+#include <sys/bus.h>
+#include <sys/endian.h>
+#include <sys/kernel.h>
+#include <sys/lock.h>
+#include <sys/malloc.h>
+#include <sys/mbuf.h>
+#include <sys/module.h>
+#include <sys/mutex.h>
+#include <sys/rman.h>
+#include <sys/socket.h>
+#include <sys/sockio.h>
+
+#include <net/bpf.h>
+#include <net/if.h>
+#include <net/if_arp.h>
+#include <net/ethernet.h>
+#include <net/if_dl.h>
+#include <net/if_media.h>
+#include <net/if_types.h>
+#include <net/if_vlan_var.h>
+
+#include <machine/bus.h>
+#include <machine/resource.h>
+
+#include <dev/mii/mii.h>
+#include <dev/mii/miivar.h>
+
+#include <dev/pci/pcireg.h>
+#include <dev/pci/pcivar.h>
+
+#include "if_sgereg.h"
+
+MODULE_DEPEND(sge, pci, 1, 1, 1);
+MODULE_DEPEND(sge, ether, 1, 1, 1);
+MODULE_DEPEND(sge, miibus, 1, 1, 1);
+
+/* "device miibus0" required.  See GENERIC if you get errors here. */
+#include "miibus_if.h"
+
+/*
+ * Various supported device vendors/types and their names.
+ */
+static struct sge_type sge_devs[] = {
+	{ SIS_VENDORID, SIS_DEVICEID_190, "SiS190 Fast Ethernet" },
+	{ SIS_VENDORID, SIS_DEVICEID_191, "SiS191 Fast/Gigabit Ethernet" },
+	{ 0, 0, NULL }
+};
+
+static int	sge_probe(device_t);
+static int	sge_attach(device_t);
+static int	sge_detach(device_t);
+static int	sge_shutdown(device_t);
+static int	sge_suspend(device_t);
+static int	sge_resume(device_t);
+
+static int	sge_miibus_readreg(device_t, int, int);
+static int	sge_miibus_writereg(device_t, int, int, int);
+static void	sge_miibus_statchg(device_t);
+
+static int	sge_newbuf(struct sge_softc *, int);
+static int	sge_encap(struct sge_softc *, struct mbuf **);
+#ifndef __NO_STRICT_ALIGNMENT
+static __inline void
+		sge_fixup_rx(struct mbuf *);
+#endif
+static __inline void
+		sge_discard_rxbuf(struct sge_softc *, int);
+static void	sge_rxeof(struct sge_softc *);
+static void	sge_txeof(struct sge_softc *);
+static void	sge_intr(void *);
+static void	sge_tick(void *);
+static void	sge_start(struct ifnet *);
+static void	sge_start_locked(struct ifnet *);
+static int	sge_ioctl(struct ifnet *, u_long, caddr_t);
+static void	sge_init(void *);
+static void	sge_init_locked(struct sge_softc *);
+static void	sge_stop(struct sge_softc *);
+static void	sge_watchdog(struct sge_softc *);
+static int	sge_ifmedia_upd(struct ifnet *);
+static void	sge_ifmedia_sts(struct ifnet *, struct ifmediareq *);
+
+static int	sge_get_mac_addr_apc(struct sge_softc *, uint8_t *);
+static int	sge_get_mac_addr_eeprom(struct sge_softc *, uint8_t *);
+static uint16_t	sge_read_eeprom(struct sge_softc *, int);
+
+static void	sge_rxfilter(struct sge_softc *);
+static void	sge_reset(struct sge_softc *);
+static int	sge_list_rx_init(struct sge_softc *);
+static int	sge_list_rx_free(struct sge_softc *);
+static int	sge_list_tx_init(struct sge_softc *);
+static int	sge_list_tx_free(struct sge_softc *);
+
+static int	sge_dma_alloc(struct sge_softc *);
+static void	sge_dma_free(struct sge_softc *);
+static void	sge_dma_map_addr(void *, bus_dma_segment_t *, int, int);
+
+static device_method_t sge_methods[] = {
+	/* Device interface */
+	DEVMETHOD(device_probe,		sge_probe),
+	DEVMETHOD(device_attach,	sge_attach),
+	DEVMETHOD(device_detach,	sge_detach),
+	DEVMETHOD(device_suspend,	sge_suspend),
+	DEVMETHOD(device_resume,	sge_resume),
+	DEVMETHOD(device_shutdown,	sge_shutdown),
+
+	/* Bus interface */
+	DEVMETHOD(bus_print_child,	bus_generic_print_child),
+	DEVMETHOD(bus_driver_added,	bus_generic_driver_added),
+
+	/* MII interface */
+	DEVMETHOD(miibus_readreg,	sge_miibus_readreg),
+	DEVMETHOD(miibus_writereg,	sge_miibus_writereg),
+	DEVMETHOD(miibus_statchg,	sge_miibus_statchg),
+
+	KOBJMETHOD_END
+};
+
+static driver_t sge_driver = {
+	"sge", sge_methods, sizeof(struct sge_softc)
+};
+
+static devclass_t sge_devclass;
+
+DRIVER_MODULE(sge, pci, sge_driver, sge_devclass, 0, 0);
+DRIVER_MODULE(miibus, sge, miibus_driver, miibus_devclass, 0, 0);
+
+/*
+ * Register space access macros.
+ */
+#define	CSR_WRITE_4(sc, reg, val)	bus_write_4(sc->sge_res, reg, val)
+#define	CSR_WRITE_2(sc, reg, val)	bus_write_2(sc->sge_res, reg, val)
+#define	CSR_WRITE_1(cs, reg, val)	bus_write_1(sc->sge_res, reg, val)
+
+#define	CSR_READ_4(sc, reg)		bus_read_4(sc->sge_res, reg)
+#define	CSR_READ_2(sc, reg)		bus_read_2(sc->sge_res, reg)
+#define	CSR_READ_1(sc, reg)		bus_read_1(sc->sge_res, reg)
+
+/* Define to show Tx/Rx error status. */
+#undef SGE_SHOW_ERRORS
+
+#define	SGE_CSUM_FEATURES	(CSUM_IP | CSUM_TCP | CSUM_UDP)
+
+static void
+sge_dma_map_addr(void *arg, bus_dma_segment_t *segs, int nseg, int error)
+{
+	bus_addr_t *p;
+
+	if (error != 0)
+		return;
+	KASSERT(nseg == 1, ("too many DMA segments, %d should be 1", nseg));
+	p  = arg;
+	*p = segs->ds_addr;
+}
+
+/*
+ * Read a sequence of words from the EEPROM.
+ */
+static uint16_t
+sge_read_eeprom(struct sge_softc *sc, int offset)
+{
+	uint32_t val;
+	int i;
+
+	KASSERT(offset <= EI_OFFSET, ("EEPROM offset too big"));
+	CSR_WRITE_4(sc, ROMInterface,
+	    EI_REQ | EI_OP_RD | (offset << EI_OFFSET_SHIFT));
+	DELAY(500);
+	for (i = 0; i < SGE_TIMEOUT; i++) {
+		val = CSR_READ_4(sc, ROMInterface);
+		if ((val & EI_REQ) == 0)
+			break;
+		DELAY(100);
+	}
+	if (i == SGE_TIMEOUT) {
+		device_printf(sc->sge_dev,
+		    "EEPROM read timeout : 0x%08x\n", val);
+		return (0xffff);
+	}
+
+	return ((val & EI_DATA) >> EI_DATA_SHIFT);
+}
+
+static int
+sge_get_mac_addr_eeprom(struct sge_softc *sc, uint8_t *dest)
+{
+	uint16_t val;
+	int i;
+
+	val = sge_read_eeprom(sc, EEPROMSignature);
+	if (val == 0xffff || val == 0) {
+		device_printf(sc->sge_dev,
+		    "invalid EEPROM signature : 0x%04x\n", val);
+		return (EINVAL);
+	}
+
+	for (i = 0; i < ETHER_ADDR_LEN; i += 2) {
+		val = sge_read_eeprom(sc, EEPROMMACAddr + i / 2);
+		dest[i + 0] = (uint8_t)val;
+		dest[i + 1] = (uint8_t)(val >> 8);
+	}
+
+	if ((sge_read_eeprom(sc, EEPROMInfo) & 0x80) != 0)
+		sc->sge_flags |= SGE_FLAG_RGMII;
+	return (0);
+}
+
+/*
+ * For SiS96x, APC CMOS RAM is used to store ethernet address.
+ * APC CMOS RAM is accessed through ISA bridge.
+ */
+static int
+sge_get_mac_addr_apc(struct sge_softc *sc, uint8_t *dest)
+{
+#if defined(__amd64__) || defined(__i386__)
+	devclass_t pci;
+	device_t bus, dev = NULL;
+	device_t *kids;
+	struct apc_tbl {
+		uint16_t vid;
+		uint16_t did;
+	} *tp, apc_tbls[] = {
+		{ SIS_VENDORID, 0x0965 },
+		{ SIS_VENDORID, 0x0966 },
+		{ SIS_VENDORID, 0x0968 }
+	};
+	uint8_t reg;
+	int busnum, cnt, i, j, numkids;
+
+	cnt = sizeof(apc_tbls) / sizeof(apc_tbls[0]);
+	pci = devclass_find("pci");
+	for (busnum = 0; busnum < devclass_get_maxunit(pci); busnum++) {
+		bus = devclass_get_device(pci, busnum);
+		if (!bus)
+			continue;
+		if (device_get_children(bus, &kids, &numkids) != 0)
+			continue;
+		for (i = 0; i < numkids; i++) {
+			dev = kids[i];
+			if (pci_get_class(dev) == PCIC_BRIDGE &&
+			    pci_get_subclass(dev) == PCIS_BRIDGE_ISA) {
+				tp = apc_tbls;
+				for (j = 0; j < cnt; j++) {
+					if (pci_get_vendor(dev) == tp->vid &&
+					    pci_get_device(dev) == tp->did) {
+						free(kids, M_TEMP);
+						goto apc_found;
+					}
+					tp++;
+				}
+			}
+                }
+		free(kids, M_TEMP);
+	}
+	device_printf(sc->sge_dev, "couldn't find PCI-ISA bridge\n");
+	return (EINVAL);
+apc_found:
+	/* Enable port 0x78 and 0x79 to access APC registers. */
+	reg = pci_read_config(dev, 0x48, 1);
+	pci_write_config(dev, 0x48, reg & ~0x02, 1);
+	DELAY(50);
+	pci_read_config(dev, 0x48, 1);
+	/* Read stored ethernet address. */
+	for (i = 0; i < ETHER_ADDR_LEN; i++) {
+		outb(0x78, 0x09 + i);
+		dest[i] = inb(0x79);
+	}
+	outb(0x78, 0x12);
+	if ((inb(0x79) & 0x80) != 0)
+		sc->sge_flags |= SGE_FLAG_RGMII;
+	/* Restore access to APC registers. */
+	pci_write_config(dev, 0x48, reg, 1);
+
+	return (0);
+#else
+	return (EINVAL);
+#endif
+}
+
+static int
+sge_miibus_readreg(device_t dev, int phy, int reg)
+{
+	struct sge_softc *sc;
+	uint32_t val;
+	int i;
+
+	sc = device_get_softc(dev);
+	CSR_WRITE_4(sc, GMIIControl, (phy << GMI_PHY_SHIFT) |
+	    (reg << GMI_REG_SHIFT) | GMI_OP_RD | GMI_REQ);
+	DELAY(10);
+	for (i = 0; i < SGE_TIMEOUT; i++) {
+		val = CSR_READ_4(sc, GMIIControl);
+		if ((val & GMI_REQ) == 0)
+			break;
+		DELAY(10);
+	}
+	if (i == SGE_TIMEOUT) {
+		device_printf(sc->sge_dev, "PHY read timeout : %d\n", reg);
+		return (0);
+	}
+	return ((val & GMI_DATA) >> GMI_DATA_SHIFT);
+}
+
+static int
+sge_miibus_writereg(device_t dev, int phy, int reg, int data)
+{
+	struct sge_softc *sc;
+	uint32_t val;
+	int i;
+
+	sc = device_get_softc(dev);
+	CSR_WRITE_4(sc, GMIIControl, (phy << GMI_PHY_SHIFT) |
+	    (reg << GMI_REG_SHIFT) | (data << GMI_DATA_SHIFT) |
+	    GMI_OP_WR | GMI_REQ);
+	DELAY(10);
+	for (i = 0; i < SGE_TIMEOUT; i++) {
+		val = CSR_READ_4(sc, GMIIControl);
+		if ((val & GMI_REQ) == 0)
+			break;
+		DELAY(10);
+	}
+	if (i == SGE_TIMEOUT)
+		device_printf(sc->sge_dev, "PHY write timeout : %d\n", reg);
+	return (0);
+}
+
+static void
+sge_miibus_statchg(device_t dev)
+{
+	struct sge_softc *sc;
+	struct mii_data *mii;
+	struct ifnet *ifp;
+	uint32_t ctl, speed;
+
+	sc = device_get_softc(dev);
+	mii = device_get_softc(sc->sge_miibus);
+	ifp = sc->sge_ifp;
+	if (mii == NULL || ifp == NULL ||
+	    (ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
+		return;
+	speed = 0;
+	sc->sge_flags &= ~SGE_FLAG_LINK;
+	if ((mii->mii_media_status & (IFM_ACTIVE | IFM_AVALID)) ==
+	    (IFM_ACTIVE | IFM_AVALID)) {
+		switch (IFM_SUBTYPE(mii->mii_media_active)) {
+		case IFM_10_T:
+			sc->sge_flags |= SGE_FLAG_LINK;
+			speed = SC_SPEED_10;
+			break;
+		case IFM_100_TX:
+			sc->sge_flags |= SGE_FLAG_LINK;
+			speed = SC_SPEED_100;
+			break;
+		case IFM_1000_T:
+			if ((sc->sge_flags & SGE_FLAG_FASTETHER) == 0) {
+				sc->sge_flags |= SGE_FLAG_LINK;
+				speed = SC_SPEED_1000;
+			}
+			break;
+		default:
+			break;
+                }
+        }
+	if ((sc->sge_flags & SGE_FLAG_LINK) == 0)
+		return;
+	/* Reprogram MAC to resolved speed/duplex/flow-control parameters. */
+	ctl = CSR_READ_4(sc, StationControl);
+	ctl &= ~(0x0f000000 | SC_FDX | SC_SPEED_MASK);
+	if (speed == SC_SPEED_1000) {
+		ctl |= 0x07000000;
+		sc->sge_flags |= SGE_FLAG_SPEED_1000;
+	} else {
+		ctl |= 0x04000000;
+		sc->sge_flags &= ~SGE_FLAG_SPEED_1000;
+	}
+#ifdef notyet
+	if ((sc->sge_flags & SGE_FLAG_GMII) != 0)
+		ctl |= 0x03000000;
+#endif
+	ctl |= speed;
+	if ((IFM_OPTIONS(mii->mii_media_active) & IFM_FDX) != 0) {
+		ctl |= SC_FDX;
+		sc->sge_flags |= SGE_FLAG_FDX;
+	} else
+		sc->sge_flags &= ~SGE_FLAG_FDX;
+	CSR_WRITE_4(sc, StationControl, ctl);
+	if ((sc->sge_flags & SGE_FLAG_RGMII) != 0) {
+		CSR_WRITE_4(sc, RGMIIDelay, 0x0441);
+		CSR_WRITE_4(sc, RGMIIDelay, 0x0440);
+	}
+}
+
+static void
+sge_rxfilter(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+	struct ifmultiaddr *ifma;
+	uint32_t crc, hashes[2];
+	uint16_t rxfilt;
+
+	SGE_LOCK_ASSERT(sc);
+
+	ifp = sc->sge_ifp;
+	hashes[0] = hashes[1] = 0;
+	rxfilt = AcceptMyPhys;
+	if ((ifp->if_flags & IFF_BROADCAST) != 0)
+		rxfilt |= AcceptBroadcast;
+	if ((ifp->if_flags & (IFF_PROMISC | IFF_ALLMULTI)) != 0) {
+		if ((ifp->if_flags & IFF_PROMISC) != 0)
+			rxfilt |= AcceptAllPhys;
+		rxfilt |= AcceptMulticast;
+		hashes[0] = 0xFFFFFFFF;
+		hashes[1] = 0xFFFFFFFF;
+		goto done;
+	}
+	rxfilt |= AcceptMulticast;
+	/* Now program new ones. */
+	if_maddr_rlock(ifp);
+	TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
+		if (ifma->ifma_addr->sa_family != AF_LINK)
+			continue;
+		crc = ether_crc32_be(LLADDR((struct sockaddr_dl *)
+		    ifma->ifma_addr), ETHER_ADDR_LEN);
+		hashes[crc >> 31] |= 1 << ((crc >> 26) & 0x1f);
+	}
+	if_maddr_runlock(ifp);
+done:
+	CSR_WRITE_2(sc, RxMacControl, rxfilt | 0x02);
+	CSR_WRITE_4(sc, RxHashTable, hashes[0]);
+	CSR_WRITE_4(sc, RxHashTable2, hashes[1]);
+}
+
+static void
+sge_reset(struct sge_softc *sc)
+{
+
+	CSR_WRITE_4(sc, IntrMask, 0);
+	CSR_WRITE_4(sc, IntrStatus, 0xffffffff);
+
+	/* Soft reset. */
+	CSR_WRITE_4(sc, IntrControl, 0x8000);
+	CSR_READ_4(sc, IntrControl);
+	DELAY(100);
+	CSR_WRITE_4(sc, IntrControl, 0);
+	/* Stop MAC. */
+	CSR_WRITE_4(sc, TX_CTL, 0x1a00);
+	CSR_WRITE_4(sc, RX_CTL, 0x1a00);
+
+	CSR_WRITE_4(sc, IntrMask, 0);
+	CSR_WRITE_4(sc, IntrStatus, 0xffffffff);
+
+	CSR_WRITE_4(sc, GMIIControl, 0);
+}
+
+/*
+ * Probe for an SiS chip. Check the PCI vendor and device
+ * IDs against our list and return a device name if we find a match.
+ */
+static int
+sge_probe(device_t dev)
+{
+	struct sge_type *t;
+
+	t = sge_devs;
+	while (t->sge_name != NULL) {
+		if ((pci_get_vendor(dev) == t->sge_vid) &&
+		    (pci_get_device(dev) == t->sge_did)) {
+			device_set_desc(dev, t->sge_name);
+			return (BUS_PROBE_DEFAULT);
+		}
+		t++;
+	}
+
+	return (ENXIO);
+}
+
+/*
+ * Attach the interface.  Allocate softc structures, do ifmedia
+ * setup and ethernet/BPF attach.
+ */
+static int
+sge_attach(device_t dev)
+{
+	struct sge_softc *sc;
+	struct ifnet *ifp;
+	uint8_t eaddr[ETHER_ADDR_LEN];
+	int error = 0, rid;
+
+	sc = device_get_softc(dev);
+	sc->sge_dev = dev;
+
+	mtx_init(&sc->sge_mtx, device_get_nameunit(dev), MTX_NETWORK_LOCK,
+	    MTX_DEF);
+        callout_init_mtx(&sc->sge_stat_ch, &sc->sge_mtx, 0);
+
+	/*
+	 * Map control/status registers.
+	 */
+	pci_enable_busmaster(dev);
+
+	/* Allocate resources. */
+	sc->sge_res_id = PCIR_BAR(0);
+	sc->sge_res_type = SYS_RES_MEMORY;
+	sc->sge_res = bus_alloc_resource_any(dev, sc->sge_res_type,
+	    &sc->sge_res_id, RF_ACTIVE);
+	if (sc->sge_res == NULL) {
+		device_printf(dev, "couldn't allocate resource\n");
+		error = ENXIO;
+		goto fail;
+	}
+
+	rid = 0;
+	sc->sge_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
+	    RF_SHAREABLE | RF_ACTIVE);
+	if (sc->sge_irq == NULL) {
+		device_printf(dev, "couldn't allocate IRQ resources\n");
+		error = ENXIO;
+		goto fail;
+	}
+	sc->sge_rev = pci_get_revid(dev);
+	if (pci_get_device(dev) == SIS_DEVICEID_190)
+		sc->sge_flags |= SGE_FLAG_FASTETHER;
+	/* Reset the adapter. */
+	sge_reset(sc);
+
+	/* Get MAC address from the EEPROM. */
+	if ((pci_read_config(dev, 0x73, 1) & 0x01) != 0)
+		sge_get_mac_addr_apc(sc, eaddr);
+	else
+		sge_get_mac_addr_eeprom(sc, eaddr);
+
+	if ((error = sge_dma_alloc(sc)) != 0)
+		goto fail;
+
+	ifp = sc->sge_ifp = if_alloc(IFT_ETHER);
+	if (ifp == NULL) {
+		device_printf(dev, "cannot allocate ifnet structure.\n");
+		error = ENOSPC;
+		goto fail;
+	}
+	ifp->if_softc = sc;
+	if_initname(ifp, device_get_name(dev), device_get_unit(dev));
+	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
+	ifp->if_ioctl = sge_ioctl;
+	ifp->if_start = sge_start;
+	ifp->if_init = sge_init;
+	ifp->if_snd.ifq_drv_maxlen = SGE_TX_RING_CNT - 1;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifp->if_snd.ifq_drv_maxlen);
+	IFQ_SET_READY(&ifp->if_snd);
+	ifp->if_capabilities = IFCAP_TXCSUM | IFCAP_RXCSUM;
+	ifp->if_hwassist = SGE_CSUM_FEATURES;
+	ifp->if_capenable = ifp->if_capabilities;
+	/*
+	 * Do MII setup.
+	 */
+	if (mii_phy_probe(dev, &sc->sge_miibus, sge_ifmedia_upd,
+	    sge_ifmedia_sts)) {
+		device_printf(dev, "no PHY found!\n");
+		error = ENXIO;
+		goto fail;
+	}
+
+	/*
+	 * Call MI attach routine.
+	 */
+	ether_ifattach(ifp, eaddr);
+
+	/* VLAN setup. */
+	ifp->if_capabilities |= IFCAP_VLAN_MTU;
+	ifp->if_capenable = ifp->if_capabilities;
+	/* Tell the upper layer(s) we support long frames. */
+	ifp->if_data.ifi_hdrlen = sizeof(struct ether_vlan_header);
+
+	/* Hook interrupt last to avoid having to lock softc */
+	error = bus_setup_intr(dev, sc->sge_irq, INTR_TYPE_NET | INTR_MPSAFE,
+	    NULL, sge_intr, sc, &sc->sge_intrhand);
+	if (error) {
+		device_printf(dev, "couldn't set up irq\n");
+		ether_ifdetach(ifp);
+		goto fail;
+	}
+
+fail:
+	if (error)
+		sge_detach(dev);
+
+	return (error);
+}
+
+/*
+ * Shutdown hardware and free up resources.  This can be called any
+ * time after the mutex has been initialized.  It is called in both
+ * the error case in attach and the normal detach case so it needs
+ * to be careful about only freeing resources that have actually been
+ * allocated.
+ */
+static int
+sge_detach(device_t dev)
+{
+	struct sge_softc *sc;
+	struct ifnet *ifp;
+
+	sc = device_get_softc(dev);
+	ifp = sc->sge_ifp;
+	/* These should only be active if attach succeeded. */
+	if (device_is_attached(dev)) {
+		ether_ifdetach(ifp);
+		SGE_LOCK(sc);
+		sge_stop(sc);
+		SGE_UNLOCK(sc);
+		callout_drain(&sc->sge_stat_ch);
+	}
+	if (sc->sge_miibus)
+		device_delete_child(dev, sc->sge_miibus);
+	bus_generic_detach(dev);
+
+	if (sc->sge_intrhand)
+		bus_teardown_intr(dev, sc->sge_irq, sc->sge_intrhand);
+	if (sc->sge_irq)
+		bus_release_resource(dev, SYS_RES_IRQ, 0, sc->sge_irq);
+	if (sc->sge_res)
+		bus_release_resource(dev, sc->sge_res_type, sc->sge_res_id,
+		    sc->sge_res);
+	if (ifp)
+		if_free(ifp);
+	sge_dma_free(sc);
+	mtx_destroy(&sc->sge_mtx);
+
+	return (0);
+}
+
+/*
+ * Stop all chip I/O so that the kernel's probe routines don't
+ * get confused by errant DMAs when rebooting.
+ */
+static int
+sge_shutdown(device_t dev)
+{
+	struct sge_softc *sc;
+
+	sc = device_get_softc(dev);
+	SGE_LOCK(sc);
+	sge_stop(sc);
+	SGE_UNLOCK(sc);
+	return (0);
+}
+
+static int
+sge_suspend(device_t dev)
+{
+	struct sge_softc *sc;
+	struct ifnet *ifp;
+
+	sc = device_get_softc(dev);
+	SGE_LOCK(sc);
+	ifp = sc->sge_ifp;
+	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0)
+		sge_stop(sc);
+	SGE_UNLOCK(sc);
+	return (0);
+}
+
+static int
+sge_resume(device_t dev)
+{
+	struct sge_softc *sc;
+	struct ifnet *ifp;
+
+	sc = device_get_softc(dev);
+	SGE_LOCK(sc);
+	ifp = sc->sge_ifp;
+	if ((ifp->if_flags & IFF_UP) != 0)
+		sge_init_locked(sc);
+	SGE_UNLOCK(sc);
+	return (0);
+}
+
+static int
+sge_dma_alloc(struct sge_softc *sc)
+{
+	struct sge_chain_data *cd;
+	struct sge_list_data *ld;
+	int error, i;
+
+	cd = &sc->sge_cdata;
+	ld = &sc->sge_ldata;
+	error = bus_dma_tag_create(bus_get_dma_tag(sc->sge_dev),
+	    1, 0,			/* alignment, boundary */
+	    BUS_SPACE_MAXADDR_32BIT,	/* lowaddr */
+	    BUS_SPACE_MAXADDR,		/* highaddr */
+	    NULL, NULL,			/* filter, filterarg */
+	    BUS_SPACE_MAXSIZE_32BIT,	/* maxsize */
+	    1,				/* nsegments */
+	    BUS_SPACE_MAXSIZE_32BIT,	/* maxsegsize */
+	    0,				/* flags */
+	    NULL,			/* lockfunc */
+	    NULL,			/* lockarg */
+	    &cd->sge_tag);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create parent DMA tag.\n");
+		goto fail;
+	}
+
+	/* RX descriptor ring */
+	error = bus_dma_tag_create(cd->sge_tag,
+	    SGE_DESC_ALIGN, 0,		/* alignment, boundary */
+	    BUS_SPACE_MAXADDR,		/* lowaddr */
+	    BUS_SPACE_MAXADDR,		/* highaddr */
+	    NULL, NULL,			/* filter, filterarg */
+	    SGE_RX_RING_SZ, 1,		/* maxsize,nsegments */
+	    SGE_RX_RING_SZ,		/* maxsegsize */
+	    0,				/* flags */
+	    NULL,			/* lockfunc */
+	    NULL,			/* lockarg */
+	    &cd->sge_rx_tag);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create Rx ring DMA tag.\n");
+		goto fail;
+	}
+	/* Allocate DMA'able memory and load DMA map for RX ring. */
+	error = bus_dmamem_alloc(cd->sge_rx_tag, (void **)&ld->sge_rx_ring,
+	    BUS_DMA_NOWAIT | BUS_DMA_ZERO | BUS_DMA_COHERENT,
+	    &cd->sge_rx_dmamap);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not allocate DMA'able memory for Rx ring.\n");
+		goto fail;
+	}
+	error = bus_dmamap_load(cd->sge_rx_tag, cd->sge_rx_dmamap,
+	    ld->sge_rx_ring, SGE_RX_RING_SZ, sge_dma_map_addr,
+	    &ld->sge_rx_paddr, BUS_DMA_NOWAIT);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not load DMA'able memory for Rx ring.\n");
+	}
+
+	/* TX descriptor ring */
+	error = bus_dma_tag_create(cd->sge_tag,
+	    SGE_DESC_ALIGN, 0,		/* alignment, boundary */
+	    BUS_SPACE_MAXADDR,		/* lowaddr */
+	    BUS_SPACE_MAXADDR,		/* highaddr */
+	    NULL, NULL,			/* filter, filterarg */
+	    SGE_TX_RING_SZ, 1,		/* maxsize,nsegments */
+	    SGE_TX_RING_SZ,		/* maxsegsize */
+	    0,				/* flags */
+	    NULL,			/* lockfunc */
+	    NULL,			/* lockarg */
+	    &cd->sge_tx_tag);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create Rx ring DMA tag.\n");
+		goto fail;
+	}
+	/* Allocate DMA'able memory and load DMA map for TX ring. */
+	error = bus_dmamem_alloc(cd->sge_tx_tag, (void **)&ld->sge_tx_ring,
+	    BUS_DMA_NOWAIT | BUS_DMA_ZERO | BUS_DMA_COHERENT,
+	    &cd->sge_tx_dmamap);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not allocate DMA'able memory for Tx ring.\n");
+		goto fail;
+	}
+	error = bus_dmamap_load(cd->sge_tx_tag, cd->sge_tx_dmamap,
+	    ld->sge_tx_ring, SGE_TX_RING_SZ, sge_dma_map_addr,
+	    &ld->sge_tx_paddr, BUS_DMA_NOWAIT);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not load DMA'able memory for Rx ring.\n");
+		goto fail;
+	}
+
+	/* Create DMA tag for Tx buffers. */
+	error = bus_dma_tag_create(cd->sge_tag, 1, 0, BUS_SPACE_MAXADDR,
+	    BUS_SPACE_MAXADDR, NULL, NULL, MCLBYTES * SGE_MAXTXSEGS,
+	    SGE_MAXTXSEGS, MCLBYTES, 0, NULL, NULL, &cd->sge_txmbuf_tag);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create Tx mbuf DMA tag.\n");
+		goto fail;
+	}
+
+	/* Create DMA tag for Rx buffers. */
+	error = bus_dma_tag_create(cd->sge_tag, SGE_RX_BUF_ALIGN, 0,
+	    BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR, NULL, NULL, MCLBYTES, 1,
+	    MCLBYTES, 0, NULL, NULL, &cd->sge_rxmbuf_tag);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create Rx mbuf DMA tag.\n");
+		goto fail;
+	}
+
+	/* Create DMA maps for Tx buffers. */
+	for (i = 0; i < SGE_TX_RING_CNT; i++) {
+		error = bus_dmamap_create(cd->sge_txmbuf_tag, 0,
+		    &cd->sge_tx_map[i]);
+		if (error != 0) {
+			device_printf(sc->sge_dev,
+			    "could not create Tx DMA map.\n");
+			goto fail;
+		}
+	}
+	/* Create spare DMA map for Rx buffer. */
+	error = bus_dmamap_create(cd->sge_rxmbuf_tag, 0, &cd->sge_rx_spare_map);
+	if (error != 0) {
+		device_printf(sc->sge_dev,
+		    "could not create spare Rx DMA map.\n");
+		goto fail;
+	}
+	/* Create DMA maps for Rx buffers. */
+	for (i = 0; i < SGE_RX_RING_CNT; i++) {
+		error = bus_dmamap_create(cd->sge_rxmbuf_tag, 0,
+		    &cd->sge_rx_map[i]);
+		if (error) {
+			device_printf(sc->sge_dev,
+			    "could not create Rx DMA map.\n");
+			goto fail;
+		}
+	}
+fail:
+	return (error);
+}
+
+static void
+sge_dma_free(struct sge_softc *sc)
+{
+	struct sge_chain_data *cd;
+	struct sge_list_data *ld;
+	int i;
+
+	cd = &sc->sge_cdata;
+	ld = &sc->sge_ldata;
+	/* Rx ring. */
+	if (cd->sge_rx_tag != NULL) {
+		if (cd->sge_rx_dmamap != NULL)
+			bus_dmamap_unload(cd->sge_rx_tag, cd->sge_rx_dmamap);
+		if (cd->sge_rx_dmamap != NULL && ld->sge_rx_ring != NULL)
+			bus_dmamem_free(cd->sge_rx_tag, ld->sge_rx_ring,
+			    cd->sge_rx_dmamap);
+		ld->sge_rx_ring = NULL;
+		cd->sge_rx_dmamap = NULL;
+		bus_dma_tag_destroy(cd->sge_rx_tag);
+		cd->sge_rx_tag = NULL;
+	}
+	/* Tx ring. */
+	if (cd->sge_tx_tag != NULL) {
+		if (cd->sge_tx_dmamap != NULL)
+			bus_dmamap_unload(cd->sge_tx_tag, cd->sge_tx_dmamap);
+		if (cd->sge_tx_dmamap != NULL && ld->sge_tx_ring != NULL)
+			bus_dmamem_free(cd->sge_tx_tag, ld->sge_tx_ring,
+			    cd->sge_tx_dmamap);
+		ld->sge_tx_ring = NULL;
+		cd->sge_tx_dmamap = NULL;
+		bus_dma_tag_destroy(cd->sge_tx_tag);
+		cd->sge_tx_tag = NULL;
+	}
+	/* Rx buffers. */
+	if (cd->sge_rxmbuf_tag != NULL) {
+		for (i = 0; i < SGE_RX_RING_CNT; i++) {
+			if (cd->sge_rx_map[i] != NULL) {
+				bus_dmamap_destroy(cd->sge_rxmbuf_tag,
+				    cd->sge_rx_map[i]);
+				cd->sge_rx_map[i] = NULL;
+			}
+		}
+		if (cd->sge_rx_spare_map != NULL) {
+			bus_dmamap_destroy(cd->sge_rxmbuf_tag,
+			    cd->sge_rx_spare_map);
+			cd->sge_rx_spare_map = NULL;
+		}
+		bus_dma_tag_destroy(cd->sge_rxmbuf_tag);
+		cd->sge_rxmbuf_tag = NULL;
+	}
+	/* Tx buffers. */
+	if (cd->sge_txmbuf_tag != NULL) {
+		for (i = 0; i < SGE_TX_RING_CNT; i++) {
+			if (cd->sge_tx_map[i] != NULL) {
+				bus_dmamap_destroy(cd->sge_txmbuf_tag,
+				    cd->sge_tx_map[i]);
+				cd->sge_tx_map[i] = NULL;
+			}
+		}
+		bus_dma_tag_destroy(cd->sge_txmbuf_tag);
+		cd->sge_txmbuf_tag = NULL;
+	}
+	if (cd->sge_tag != NULL)
+		bus_dma_tag_destroy(cd->sge_tag);
+	cd->sge_tag = NULL;
+}
+
+/*
+ * Initialize the TX descriptors.
+ */
+static int
+sge_list_tx_init(struct sge_softc *sc)
+{
+	struct sge_list_data *ld;
+	struct sge_chain_data *cd;
+
+	SGE_LOCK_ASSERT(sc);
+	ld = &sc->sge_ldata;
+	cd = &sc->sge_cdata;
+	bzero(ld->sge_tx_ring, SGE_TX_RING_SZ);
+	ld->sge_tx_ring[SGE_TX_RING_CNT - 1].sge_flags = htole32(RING_END);
+	bus_dmamap_sync(cd->sge_tx_tag, cd->sge_tx_dmamap,
+	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
+	cd->sge_tx_prod = 0;
+	cd->sge_tx_cons = 0;
+	cd->sge_tx_cnt = 0;
+	return (0);
+}
+
+static int
+sge_list_tx_free(struct sge_softc *sc)
+{
+	struct sge_chain_data *cd;
+	int i;
+
+	SGE_LOCK_ASSERT(sc);
+	cd = &sc->sge_cdata;
+	for (i = 0; i < SGE_TX_RING_CNT; i++) {
+		if (cd->sge_tx_mbuf[i] != NULL) {
+			bus_dmamap_sync(cd->sge_txmbuf_tag,
+			    cd->sge_tx_map[i], BUS_DMASYNC_POSTWRITE);
+			bus_dmamap_unload(cd->sge_txmbuf_tag,
+			    cd->sge_tx_map[i]);
+			m_free(cd->sge_tx_mbuf[i]);
+			cd->sge_tx_mbuf[i] = NULL;
+		}
+	}
+
+	return (0);
+}
+
+/*
+ * Initialize the RX descriptors and allocate mbufs for them.  Note that
+ * we arrange the descriptors in a closed ring, so that the last descriptor
+ * has RING_END flag set.
+ */
+static int
+sge_list_rx_init(struct sge_softc *sc)
+{
+	struct sge_chain_data *cd;
+	int i;
+
+	SGE_LOCK_ASSERT(sc);
+	cd = &sc->sge_cdata;
+	cd->sge_rx_cons = 0;
+	bzero(sc->sge_ldata.sge_rx_ring, SGE_RX_RING_SZ);
+	for (i = 0; i < SGE_RX_RING_CNT; i++) {
+		if (sge_newbuf(sc, i) != 0)
+			return (ENOBUFS);
+	}
+	bus_dmamap_sync(cd->sge_rx_tag, cd->sge_rx_dmamap,
+	    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
+	return (0);
+}
+
+static int
+sge_list_rx_free(struct sge_softc *sc)
+{
+	struct sge_chain_data *cd;
+	int i;
+
+	SGE_LOCK_ASSERT(sc);
+	cd = &sc->sge_cdata;
+	for (i = 0; i < SGE_RX_RING_CNT; i++) {
+		if (cd->sge_rx_mbuf[i] != NULL) {
+			bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[i],
+			    BUS_DMASYNC_POSTREAD);
+			bus_dmamap_unload(cd->sge_rxmbuf_tag,
+			    cd->sge_rx_map[i]);
+			m_free(cd->sge_rx_mbuf[i]);
+			cd->sge_rx_mbuf[i] = NULL;
+		}
+	}
+	return (0);
+}
+
+/*
+ * Initialize an RX descriptor and attach an MBUF cluster.
+ */
+static int
+sge_newbuf(struct sge_softc *sc, int prod)
+{
+	struct mbuf *m;
+	struct sge_desc *desc;
+	struct sge_chain_data *cd;
+	bus_dma_segment_t segs[1];
+	bus_dmamap_t map;
+	int error, nsegs;
+
+	SGE_LOCK_ASSERT(sc);
+
+	cd = &sc->sge_cdata;
+	m = m_getcl(M_DONTWAIT, MT_DATA, M_PKTHDR);
+	if (m == NULL)
+		return (ENOBUFS);
+	m->m_len = m->m_pkthdr.len = MCLBYTES;
+	m_adj(m, SGE_RX_BUF_ALIGN);
+	error = bus_dmamap_load_mbuf_sg(cd->sge_rxmbuf_tag,
+	    cd->sge_rx_spare_map, m, segs, &nsegs, 0);
+	if (error != 0) {
+		m_freem(m);
+		return (error);
+	}
+	KASSERT(nsegs == 1, ("%s: %d segments returned!", __func__, nsegs));
+	if (cd->sge_rx_mbuf[prod] != NULL) {
+		bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod],
+		    BUS_DMASYNC_POSTREAD);
+		bus_dmamap_unload(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod]);
+	}
+	map = cd->sge_rx_map[prod];
+	cd->sge_rx_map[prod] =  cd->sge_rx_spare_map;
+	cd->sge_rx_spare_map = map;
+	bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod],
+	    BUS_DMASYNC_PREREAD);
+	cd->sge_rx_mbuf[prod] = m;
+
+	desc = &sc->sge_ldata.sge_rx_ring[prod];
+	desc->sge_sts_size = 0;
+	desc->sge_ptr = htole32(SGE_ADDR_LO(segs[0].ds_addr));
+	desc->sge_flags = htole32(segs[0].ds_len);
+	if (prod == SGE_RX_RING_CNT - 1)
+		desc->sge_flags |= htole32(RING_END);
+	desc->sge_cmdsts = htole32(RDC_OWN | RDC_INTR | RDC_IP_CSUM |
+	    RDC_TCP_CSUM | RDC_UDP_CSUM);
+	return (0);
+}
+
+#ifndef __NO_STRICT_ALIGNMENT
+static __inline void
+sge_fixup_rx(struct mbuf *m)
+{
+        int i;
+        uint16_t *src, *dst;
+
+	src = mtod(m, uint16_t *);
+	dst = src - 3;
+
+	for (i = 0; i < (m->m_len / sizeof(uint16_t) + 1); i++)
+		*dst++ = *src++;
+
+	m->m_data -= (SGE_RX_BUF_ALIGN - ETHER_ALIGN);
+}
+#endif
+
+static __inline void
+sge_discard_rxbuf(struct sge_softc *sc, int index)
+{
+	struct sge_desc *desc;
+
+	desc = &sc->sge_ldata.sge_rx_ring[index];
+	desc->sge_sts_size = 0;
+	desc->sge_flags = htole32(MCLBYTES - SGE_RX_BUF_ALIGN);
+	if (index == SGE_RX_RING_CNT - 1)
+		desc->sge_flags |= htole32(RING_END);
+	desc->sge_cmdsts = htole32(RDC_OWN | RDC_INTR | RDC_IP_CSUM |
+	    RDC_TCP_CSUM | RDC_UDP_CSUM);
+}
+
+/*
+ * A frame has been uploaded: pass the resulting mbuf chain up to
+ * the higher level protocols.
+ */
+static void
+sge_rxeof(struct sge_softc *sc)
+{
+        struct ifnet *ifp;
+        struct mbuf *m;
+	struct sge_chain_data *cd;
+	struct sge_desc	*cur_rx;
+	uint32_t rxinfo, rxstat;
+	int cons, prog;
+
+	SGE_LOCK_ASSERT(sc);
+
+	ifp = sc->sge_ifp;
+	cd = &sc->sge_cdata;
+
+	bus_dmamap_sync(cd->sge_rx_tag, cd->sge_rx_dmamap,
+	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
+	cons = cd->sge_rx_cons;
+	for (prog = 0; prog < SGE_RX_RING_CNT; prog++,
+	    SGE_INC(cons, SGE_RX_RING_CNT)) {
+		if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
+			break;
+		cur_rx = &sc->sge_ldata.sge_rx_ring[cons];
+		rxinfo = le32toh(cur_rx->sge_cmdsts);
+		if ((rxinfo & RDC_OWN) != 0)
+			break;
+		rxstat = le32toh(cur_rx->sge_sts_size);
+		if (SGE_RX_ERROR(rxstat) != 0 || SGE_RX_NSEGS(rxstat) != 1) {
+			/* XXX We don't support multi-segment frames yet. */
+#ifdef SGE_SHOW_ERRORS
+			device_printf(sc->sge_dev, "Rx error : 0x%b\n", rxstat,
+			    RX_ERR_BITS);
+#endif
+			sge_discard_rxbuf(sc, cons);
+			ifp->if_ierrors++;
+			continue;
+		}
+		m = cd->sge_rx_mbuf[cons];
+		if (sge_newbuf(sc, cons) != 0) {
+			sge_discard_rxbuf(sc, cons);
+			ifp->if_iqdrops++;
+			continue;
+		}
+		if ((ifp->if_capenable & IFCAP_RXCSUM) != 0) {
+			if ((rxinfo & RDC_IP_CSUM) != 0 &&
+			    (rxinfo & RDC_IP_CSUM_OK) != 0)
+				m->m_pkthdr.csum_flags |=
+				    CSUM_IP_CHECKED | CSUM_IP_VALID;
+			if (((rxinfo & RDC_TCP_CSUM) != 0 &&
+			    (rxinfo & RDC_TCP_CSUM_OK) != 0) ||
+			    ((rxinfo & RDC_UDP_CSUM) != 0 &&
+			    (rxinfo & RDC_UDP_CSUM_OK) != 0)) {
+				m->m_pkthdr.csum_flags |=
+				    CSUM_DATA_VALID | CSUM_PSEUDO_HDR;
+				m->m_pkthdr.csum_data = 0xffff;
+			}
+		}
+		/*
+		 * TODO : VLAN hardware tag stripping.
+		 */
+		m->m_pkthdr.len = m->m_len =
+		    SGE_RX_BYTES(rxstat) - ETHER_CRC_LEN;
+#ifndef __NO_STRICT_ALIGNMENT
+		sge_fixup_rx(m);
+#endif
+		m->m_pkthdr.rcvif = ifp;
+		ifp->if_ipackets++;
+		SGE_UNLOCK(sc);
+		(*ifp->if_input)(ifp, m);
+		SGE_LOCK(sc);
+	}
+
+	if (prog > 0) {
+		bus_dmamap_sync(cd->sge_rx_tag, cd->sge_rx_dmamap,
+		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
+		cd->sge_rx_cons = cons;
+	}
+}
+
+/*
+ * A frame was downloaded to the chip.  It's safe for us to clean up
+ * the list buffers.
+ */
+static void
+sge_txeof(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+	struct sge_list_data *ld;
+	struct sge_chain_data *cd;
+	uint32_t txstat;
+	int cons, prod;
+
+	SGE_LOCK_ASSERT(sc);
+
+	ifp = sc->sge_ifp;
+	ld = &sc->sge_ldata;
+	cd = &sc->sge_cdata;
+
+	if (cd->sge_tx_cnt == 0)
+		return;
+	bus_dmamap_sync(cd->sge_tx_tag, cd->sge_tx_dmamap,
+	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
+	cons = cd->sge_tx_cons;
+	prod = cd->sge_tx_prod;
+	for (; cons != prod; SGE_INC(cons, SGE_TX_RING_CNT)) {
+		txstat = le32toh(ld->sge_tx_ring[cons].sge_cmdsts);
+		if ((txstat & TDC_OWN) != 0)
+			break;
+		cd->sge_tx_cnt--;
+		ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
+		if (cd->sge_tx_mbuf[cons] != NULL) {
+			bus_dmamap_sync(cd->sge_txmbuf_tag,
+			    cd->sge_tx_map[cons], BUS_DMASYNC_POSTWRITE);
+			bus_dmamap_unload(cd->sge_txmbuf_tag,
+			    cd->sge_tx_map[cons]);
+			m_freem(cd->sge_tx_mbuf[cons]);
+			cd->sge_tx_mbuf[cons] = NULL;
+			if (SGE_TX_ERROR(txstat) != 0) {
+#ifdef SGE_SHOW_ERRORS
+				device_printf(sc->sge_dev, "Tx error : 0x%b\n",
+				    txstat, TX_ERR_BITS);
+#endif
+				ifp->if_oerrors++;
+			} else {
+#ifdef notyet
+				ifp->if_collisions += (txstat & 0xFFFF) - 1;
+#endif
+				ifp->if_opackets++;
+			}
+		}
+
+	}
+	cd->sge_tx_cons = cons;
+	if (cd->sge_tx_cnt == 0)
+		sc->sge_timer = 0;
+}
+
+static void
+sge_tick(void *arg)
+{
+	struct sge_softc *sc;
+	struct mii_data *mii;
+	struct ifnet *ifp;
+
+	sc = arg;
+	SGE_LOCK_ASSERT(sc);
+
+	ifp = sc->sge_ifp;
+	mii = device_get_softc(sc->sge_miibus);
+	mii_tick(mii);
+	if ((sc->sge_flags & SGE_FLAG_LINK) == 0) {
+		sge_miibus_statchg(sc->sge_dev);
+		if ((sc->sge_flags & SGE_FLAG_LINK) != 0 &&
+		    !IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+			sge_start_locked(ifp);
+	}
+	/*
+	 * Reclaim transmitted frames here as we do not request
+	 * Tx completion interrupt for every queued frames to
+	 * reduce excessive interrupts.
+	 */
+	sge_txeof(sc);
+	sge_watchdog(sc);
+	callout_reset(&sc->sge_stat_ch, hz, sge_tick, sc);
+}
+
+static void
+sge_intr(void *arg)
+{
+	struct sge_softc *sc;
+	struct ifnet *ifp;
+	uint32_t status;
+
+	sc = arg;
+	SGE_LOCK(sc);
+	ifp = sc->sge_ifp;
+
+	status = CSR_READ_4(sc, IntrStatus);
+	if (status == 0xFFFFFFFF || (status & SGE_INTRS) == 0) {
+		/* Not ours. */
+		SGE_UNLOCK(sc);
+		return;
+	}
+	/* Acknowledge interrupts. */
+	CSR_WRITE_4(sc, IntrStatus, status);
+	/* Disable further interrupts. */
+	CSR_WRITE_4(sc, IntrMask, 0);
+	/*
+	 * It seems the controller supports some kind of interrupt
+	 * moderation mechanism but we still don't know how to
+	 * enable that.  To reduce number of generated interrupts
+	 * under load we check pending interrupts in a loop.  This
+	 * will increase number of register access and is not correct
+	 * way to handle interrupt moderation but there seems to be
+	 * no other way at this time.
+	 */
+	for (;;) {
+		if ((ifp->if_drv_flags & IFF_DRV_RUNNING) == 0)
+			break;
+		if ((status & (INTR_RX_DONE | INTR_RX_IDLE)) != 0) {
+			sge_rxeof(sc);
+			/* Wakeup Rx MAC. */
+			if ((status & INTR_RX_IDLE) != 0)
+				CSR_WRITE_4(sc, RX_CTL,
+				    0x1a00 | 0x000c | RX_CTL_POLL | RX_CTL_ENB);
+		}
+		if ((status & (INTR_TX_DONE | INTR_TX_IDLE)) != 0)
+			sge_txeof(sc);
+		status = CSR_READ_4(sc, IntrStatus);
+		if ((status & SGE_INTRS) == 0)
+			break;
+		/* Acknowledge interrupts. */
+		CSR_WRITE_4(sc, IntrStatus, status);
+	}
+	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0) {
+		/* Re-enable interrupts */
+		CSR_WRITE_4(sc, IntrMask, SGE_INTRS);
+		if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
+			sge_start_locked(ifp);
+	}
+	SGE_UNLOCK(sc);
+}
+
+/*
+ * Encapsulate an mbuf chain in a descriptor by coupling the mbuf data
+ * pointers to the fragment pointers.
+ */
+static int
+sge_encap(struct sge_softc *sc, struct mbuf **m_head)
+{
+	struct mbuf *m;
+	struct sge_desc *desc;
+	bus_dma_segment_t txsegs[SGE_MAXTXSEGS];
+	bus_dmamap_t map;
+	uint32_t cflags;
+	int error, nsegs, prod;
+
+	SGE_LOCK_ASSERT(sc);
+
+	prod = sc->sge_cdata.sge_tx_prod;
+	map = sc->sge_cdata.sge_tx_map[prod];
+	/*
+	 * Reading Windows inf file indicates SiS controller supports
+	 * TSO, VLAN hardware tag insertion/stripping, interrupt
+	 * moderation and Tx/Rx checksum offloading.  Unfortunately
+	 * vendor didn't release these information so we're guessing
+	 * descriptor usage with trial and errors.
+	 *
+	 * Controller seems to support multi-fragmented buffers but
+	 * don't know how to enable that feature so limit number of
+	 * fragmented Tx buffers to single buffer until we understand
+	 * the controller internals.
+	 * I assume the controller can pad zero bytes if frame length
+	 * is less than 60 bytes and I also think the controller has
+	 * no Tx buffer alignment limitation. - Need testing!
+	 */
+	if ((*m_head)->m_next != NULL) {
+		m = m_defrag(*m_head, M_DONTWAIT);
+		if (m == NULL) {
+			m_freem(*m_head);
+			*m_head = NULL;
+			return (ENOBUFS);
+		}
+		*m_head = m;
+	}
+	error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_tx_tag, map,
+	    *m_head, txsegs, &nsegs, 0);
+	if (error != 0) {
+		m_freem(*m_head);
+		*m_head = NULL;
+		return (error);
+	}
+	/* Check descriptor overrun. */
+	if (sc->sge_cdata.sge_tx_cnt + nsegs >= SGE_TX_RING_CNT) {
+		bus_dmamap_unload(sc->sge_cdata.sge_tx_tag, map);
+		return (ENOBUFS);
+	}
+	bus_dmamap_sync(sc->sge_cdata.sge_tx_tag, map, BUS_DMASYNC_PREWRITE);
+
+	cflags = 0;
+	if ((*m_head)->m_pkthdr.csum_flags & CSUM_IP)
+		cflags |= TDC_IP_CSUM;
+	if ((*m_head)->m_pkthdr.csum_flags & CSUM_TCP)
+		cflags |= TDC_TCP_CSUM;
+	if ((*m_head)->m_pkthdr.csum_flags & CSUM_UDP)
+		cflags |= TDC_UDP_CSUM;
+	desc = &sc->sge_ldata.sge_tx_ring[prod];
+	desc->sge_sts_size = htole32((*m_head)->m_pkthdr.len);
+	desc->sge_ptr = htole32(SGE_ADDR_LO(txsegs[0].ds_addr));
+	desc->sge_flags = htole32(txsegs[0].ds_len);
+	if (prod == SGE_TX_RING_CNT - 1)
+		desc->sge_flags |= htole32(RING_END);
+	desc->sge_cmdsts = htole32(TDC_DEF | TDC_CRC | TDC_PAD | cflags);
+#if 1
+	if ((sc->sge_flags & SGE_FLAG_SPEED_1000) != 0)
+		desc->sge_cmdsts |= htole32(TDC_BST);
+#else
+	if ((sc->sge_flags & SGE_FLAG_FDX) == 0) {
+		desc->sge_cmdsts |= htole32(TDC_COL | TDC_CRS | TDC_BKF);
+		if ((sc->sge_flags & SGE_FLAG_SPEED_1000) != 0)
+			desc->sge_cmdsts |= htole32(TDC_EXT | TDC_BST);
+	}
+#endif
+	/* Request interrupt and give ownership to controller. */
+	if ((prod % SGE_TX_INTR_FRAMES) == 0)
+		desc->sge_cmdsts |= htole32(TDC_OWN | TDC_INTR);
+	else
+		desc->sge_cmdsts |= htole32(TDC_OWN);
+	sc->sge_cdata.sge_tx_mbuf[prod] = *m_head;
+	sc->sge_cdata.sge_tx_cnt++;
+	SGE_INC(sc->sge_cdata.sge_tx_prod, SGE_TX_RING_CNT);
+	return (0);
+}
+
+static void
+sge_start(struct ifnet *ifp)
+{
+	struct sge_softc *sc;
+
+	sc = ifp->if_softc;
+	SGE_LOCK(sc);
+	sge_start_locked(ifp);
+	SGE_UNLOCK(sc);
+}
+
+static void
+sge_start_locked(struct ifnet *ifp)
+{
+	struct sge_softc *sc;
+	struct mbuf *m_head;
+	int queued = 0;
+
+	sc = ifp->if_softc;
+	SGE_LOCK_ASSERT(sc);
+
+	if ((sc->sge_flags & SGE_FLAG_LINK) == 0 ||
+	    (ifp->if_drv_flags & (IFF_DRV_RUNNING | IFF_DRV_OACTIVE)) !=
+	    IFF_DRV_RUNNING)
+		return;
+
+	for (queued = 0; !IFQ_DRV_IS_EMPTY(&ifp->if_snd); ) {
+		if (sc->sge_cdata.sge_tx_cnt == SGE_TX_RING_CNT - 1) {
+			ifp->if_drv_flags |= IFF_DRV_OACTIVE;
+			break;
+		}
+		IFQ_DRV_DEQUEUE(&ifp->if_snd, m_head);
+		if (m_head == NULL)
+			break;
+		if (sge_encap(sc, &m_head)) {
+			IFQ_DRV_PREPEND(&ifp->if_snd, m_head);
+			ifp->if_drv_flags |= IFF_DRV_OACTIVE;
+			break;
+		}
+		queued++;
+		/*
+		 * If there's a BPF listener, bounce a copy of this frame
+		 * to him.
+		 */
+		BPF_MTAP(ifp, m_head);
+	}
+
+	if (queued > 0) {
+		bus_dmamap_sync(sc->sge_cdata.sge_tx_tag,
+		    sc->sge_cdata.sge_tx_dmamap,
+		    BUS_DMASYNC_PREREAD | BUS_DMASYNC_PREWRITE);
+		CSR_WRITE_4(sc, TX_CTL, 0x1a00 | TX_CTL_ENB | TX_CTL_POLL);
+		sc->sge_timer = 5;
+	}
+}
+
+static void
+sge_init(void *arg)
+{
+	struct sge_softc *sc;
+
+	sc = arg;
+	SGE_LOCK(sc);
+	sge_init_locked(sc);
+	SGE_UNLOCK(sc);
+}
+
+static void
+sge_init_locked(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+	struct mii_data *mii;
+	int i;
+
+	SGE_LOCK_ASSERT(sc);
+	ifp = sc->sge_ifp;
+	mii = device_get_softc(sc->sge_miibus);
+	if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0)
+		return;
+	/*
+	 * Cancel pending I/O and free all RX/TX buffers.
+	 */
+	sge_stop(sc);
+	sge_reset(sc);
+
+	/* Init circular RX list. */
+	if (sge_list_rx_init(sc) == ENOBUFS) {
+		device_printf(sc->sge_dev, "no memory for Rx buffers\n");
+		sge_stop(sc);
+		return;
+	}
+	/* Init TX descriptors. */
+	sge_list_tx_init(sc);
+	/*
+	 * Load the address of the RX and TX lists.
+	 */
+	CSR_WRITE_4(sc, TX_DESC, SGE_ADDR_LO(sc->sge_ldata.sge_tx_paddr));
+	CSR_WRITE_4(sc, RX_DESC, SGE_ADDR_LO(sc->sge_ldata.sge_rx_paddr));
+
+	CSR_WRITE_4(sc, TxMacControl, 0x60);
+	CSR_WRITE_4(sc, 0x6c, 0);
+	CSR_WRITE_4(sc, RxWakeOnLan, 0);
+	CSR_WRITE_4(sc, RxWakeOnLanData, 0);
+	/* Allow receiving VLAN frames. */
+	CSR_WRITE_2(sc, RxMPSControl, ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN);
+
+	for (i = 0; i < ETHER_ADDR_LEN; i++)
+		CSR_WRITE_1(sc, RxMacAddr + i, IF_LLADDR(ifp)[i]);
+	sge_rxfilter(sc);
+
+	/* Initialize default speed/duplex information. */
+	if ((sc->sge_flags & SGE_FLAG_FASTETHER) == 0)
+		sc->sge_flags |= SGE_FLAG_SPEED_1000;
+	sc->sge_flags |= SGE_FLAG_FDX;
+	if ((sc->sge_flags & SGE_FLAG_RGMII) != 0)
+		CSR_WRITE_4(sc, StationControl, 0x04008001);
+	else
+		CSR_WRITE_4(sc, StationControl, 0x04000001);
+	/*
+	 * XXX Try to mitigate interrupts.
+	 */
+	if (sc->sge_intrcontrol != 0)
+		CSR_WRITE_4(sc, IntrControl, sc->sge_intrcontrol);
+	if (sc->sge_intrtimer != 0)
+		CSR_WRITE_4(sc, IntrTimer, sc->sge_intrtimer);
+
+	/*
+	 * Clear and enable interrupts.
+	 */
+	CSR_WRITE_4(sc, IntrStatus, 0xFFFFFFFF);
+	CSR_WRITE_4(sc, IntrMask, SGE_INTRS);
+
+	/* Enable receiver and transmitter. */
+	CSR_WRITE_4(sc, TX_CTL, 0x1a00 | TX_CTL_ENB);
+	CSR_WRITE_4(sc, RX_CTL, 0x1a00 | 0x000c | RX_CTL_POLL | RX_CTL_ENB);
+
+	ifp->if_drv_flags |= IFF_DRV_RUNNING;
+	ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
+
+	sc->sge_flags &= ~SGE_FLAG_LINK;
+	mii_mediachg(mii);
+	callout_reset(&sc->sge_stat_ch, hz, sge_tick, sc);
+}
+
+/*
+ * Set media options.
+ */
+static int
+sge_ifmedia_upd(struct ifnet *ifp)
+{
+	struct sge_softc *sc;
+	struct mii_data *mii;
+	int error;
+
+	sc = ifp->if_softc;
+	SGE_LOCK(sc);
+	mii = device_get_softc(sc->sge_miibus);
+	sc->sge_flags &= ~SGE_FLAG_LINK;
+	if (mii->mii_instance) {
+		struct mii_softc *miisc;
+		LIST_FOREACH(miisc, &mii->mii_phys, mii_list)
+			mii_phy_reset(miisc);
+	}
+	error = mii_mediachg(mii);
+	SGE_UNLOCK(sc);
+
+	return (error);
+}
+
+/*
+ * Report current media status.
+ */
+static void
+sge_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr)
+{
+	struct sge_softc *sc;
+	struct mii_data *mii;
+
+	sc = ifp->if_softc;
+	SGE_LOCK(sc);
+	mii = device_get_softc(sc->sge_miibus);
+	if ((ifp->if_flags & IFF_UP) == 0) {
+		SGE_UNLOCK(sc);
+		return;
+	}
+	mii_pollstat(mii);
+	SGE_UNLOCK(sc);
+	ifmr->ifm_active = mii->mii_media_active;
+	ifmr->ifm_status = mii->mii_media_status;
+}
+
+static int
+sge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
+{
+	struct sge_softc *sc;
+	struct ifreq *ifr;
+	struct mii_data *mii;
+	int error = 0, mask;
+
+	sc = ifp->if_softc;
+	ifr = (struct ifreq *)data;
+
+	switch(command) {
+	case SIOCSIFFLAGS:
+		SGE_LOCK(sc);
+		if ((ifp->if_flags & IFF_UP) != 0) {
+			if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0 &&
+			    ((ifp->if_flags ^ sc->sge_if_flags) &
+			    (IFF_PROMISC | IFF_ALLMULTI)) != 0)
+				sge_rxfilter(sc);
+			else
+				sge_init_locked(sc);
+		} else if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0)
+			sge_stop(sc);
+		sc->sge_if_flags = ifp->if_flags;
+		SGE_UNLOCK(sc);
+		break;
+	case SIOCSIFCAP:
+		SGE_LOCK(sc);
+		mask = ifr->ifr_reqcap ^ ifp->if_capenable;
+		if ((mask & IFCAP_TXCSUM) != 0 &&
+		    (ifp->if_capabilities & IFCAP_TXCSUM) != 0) {
+			ifp->if_capenable ^= IFCAP_TXCSUM;
+			if ((ifp->if_capenable & IFCAP_TXCSUM) != 0)
+				ifp->if_hwassist |= SGE_CSUM_FEATURES;
+			else
+				ifp->if_hwassist &= ~SGE_CSUM_FEATURES;
+		}
+		if ((mask & IFCAP_RXCSUM) != 0 &&
+		    (ifp->if_capabilities & IFCAP_RXCSUM) != 0)
+			ifp->if_capenable ^= IFCAP_RXCSUM;
+		SGE_UNLOCK(sc);
+		break;
+	case SIOCADDMULTI:
+	case SIOCDELMULTI:
+		SGE_LOCK(sc);
+		if ((ifp->if_drv_flags & IFF_DRV_RUNNING) != 0)
+			sge_rxfilter(sc);
+		SGE_UNLOCK(sc);
+		break;
+	case SIOCGIFMEDIA:
+	case SIOCSIFMEDIA:
+		mii = device_get_softc(sc->sge_miibus);
+		error = ifmedia_ioctl(ifp, ifr, &mii->mii_media, command);
+		break;
+	default:
+		error = ether_ioctl(ifp, command, data);
+		break;
+	}
+
+	return (error);
+}
+
+static void
+sge_watchdog(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+
+	SGE_LOCK_ASSERT(sc);
+	if (sc->sge_timer == 0 || --sc->sge_timer > 0)
+		return;
+
+	ifp = sc->sge_ifp;
+	if ((sc->sge_flags & SGE_FLAG_LINK) == 0) {
+		if (1 || bootverbose)
+			device_printf(sc->sge_dev,
+			    "watchdog timeout (lost link)\n");
+		ifp->if_oerrors++;
+		ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
+		sge_init_locked(sc);
+		return;
+	}
+	device_printf(sc->sge_dev, "watchdog timeout\n");
+	ifp->if_oerrors++;
+
+	ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
+	sge_init_locked(sc);
+	if (!IFQ_DRV_IS_EMPTY(&sc->sge_ifp->if_snd))
+		sge_start_locked(ifp);
+}
+
+/*
+ * Stop the adapter and free any mbufs allocated to the
+ * RX and TX lists.
+ */
+static void
+sge_stop(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+
+	ifp = sc->sge_ifp;
+
+	SGE_LOCK_ASSERT(sc);
+
+	sc->sge_timer = 0;
+	callout_stop(&sc->sge_stat_ch);
+	ifp->if_drv_flags &= ~(IFF_DRV_RUNNING | IFF_DRV_OACTIVE);
+
+	CSR_WRITE_4(sc, IntrMask, 0);
+	CSR_READ_4(sc, IntrMask);
+	CSR_WRITE_4(sc, IntrStatus, 0xffffffff);
+	/* Stop TX/RX MAC. */
+	CSR_WRITE_4(sc, TX_CTL, 0x1a00);
+	CSR_WRITE_4(sc, RX_CTL, 0x1a00);
+	/* XXX Can we assume active DMA cycles gone? */
+	DELAY(2000);
+	CSR_WRITE_4(sc, IntrMask, 0);
+	CSR_WRITE_4(sc, IntrStatus, 0xffffffff);
+
+	sc->sge_flags &= ~SGE_FLAG_LINK;
+	sge_list_rx_free(sc);
+	sge_list_tx_free(sc);
+}
diff --git a/sys/dev/sge/if_sgereg.h b/sys/dev/sge/if_sgereg.h
new file mode 100644
index 00000000000..29253e0f892
--- /dev/null
+++ b/sys/dev/sge/if_sgereg.h
@@ -0,0 +1,350 @@
+/*-
+ * Copyright (c) 2008, 2009, 2010 Nikolay Denev <ndenev@gmail.com>
+ * Copyright (c) 2007, 2008 Alexander Pohoyda <alexander.pohoyda@gmx.net>
+ * Copyright (c) 1997, 1998, 1999
+ *      Bill Paul <wpaul@ctr.columbia.edu>.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *      This product includes software developed by Bill Paul.
+ * 4. Neither the name of the author nor the names of any co-contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY Bill Paul AND CONTRIBUTORS ``AS IS''
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
+ * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
+ * PARTICULAR PURPOSE ARE DISCLAIMED.  IN NO EVENT SHALL AUTHORS OR
+ * THE VOICES IN THEIR HEADS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
+ * (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
+ * SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT,
+ * STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
+ * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED
+ * OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _IF_SGEREG_H
+#define	_IF_SGEREG_H
+
+/*
+ * SiS PCI vendor ID.
+ */
+#define	SIS_VENDORID		0x1039
+
+/*
+ * SiS PCI device IDs
+ */
+#define	SIS_DEVICEID_190	0x0190
+#define	SIS_DEVICEID_191	0x0191
+
+#define	TX_CTL			0x00
+#define	TX_DESC			0x04
+#define	Reserved0		0x08
+#define	TX_NEXT			0x0c
+
+#define	RX_CTL			0x10
+#define	RX_DESC			0x14
+#define	Reserved1		0x18
+#define	RX_NEXT			0x1c
+
+#define	IntrStatus		0x20
+#define	IntrMask		0x24
+#define	IntrControl		0x28
+#define	IntrTimer		0x2c
+
+#define	PMControl		0x30
+#define	Reserved2		0x34
+#define	ROMControl		0x38
+#define	ROMInterface		0x3c
+#define	StationControl		0x40
+#define	GMIIControl		0x44
+#define	GMacIOCR		0x48
+#define	GMacIOCTL		0x4c
+#define	TxMacControl		0x50
+#define	TxMacTimeLimit		0x54
+#define	RGMIIDelay		0x58
+#define	Reserved3		0x5c
+#define	RxMacControl		0x60	/* 1  WORD */
+#define	RxMacAddr		0x62	/* 6x BYTE */
+#define	RxHashTable		0x68	/* 1 LONG */
+#define	RxHashTable2		0x6c	/* 1 LONG */
+#define	RxWakeOnLan		0x70
+#define	RxWakeOnLanData		0x74
+#define	RxMPSControl		0x78
+#define	Reserved4		0x7c
+
+/*
+ * IntrStatus Register Content
+ */
+#define	INTR_SOFT		0x40000000
+#define	INTR_TIMER		0x20000000
+#define	INTR_PAUSE_FRAME	0x00080000
+#define	INTR_MAGIC_FRAME	0x00040000
+#define	INTR_WAKE_FRAME		0x00020000
+#define	INTR_LINK		0x00010000
+#define	INTR_RX_IDLE		0x00000080	
+#define	INTR_RX_DONE		0x00000040
+#define	INTR_TXQ1_IDLE		0x00000020
+#define	INTR_TXQ1_DONE		0x00000010
+#define	INTR_TX_IDLE		0x00000008
+#define	INTR_TX_DONE		0x00000004
+#define	INTR_RX_HALT		0x00000002
+#define	INTR_TX_HALT		0x00000001
+
+#define	SGE_INTRS							\
+	(INTR_RX_IDLE | INTR_RX_DONE | INTR_TXQ1_IDLE |			\
+	 INTR_TXQ1_DONE |INTR_TX_IDLE | INTR_TX_DONE |			\
+	 INTR_TX_HALT | INTR_RX_HALT)
+
+/*
+ * RxStatusDesc Register Content
+ */
+#define	RxRES			0x00200000
+#define	RxCRC			0x00080000
+#define	RxRUNT			0x00100000
+#define	RxRWT			0x00400000
+
+/*
+ * RX_CTL Register Content
+ */
+#define	RX_CTL_POLL		0x00000010
+#define	RX_CTL_ENB		0x00000001
+
+/*
+ * TX_CTL Register Content
+ */
+#define	TX_CTL_POLL		0x00000010
+#define	TX_CTL_ENB		0x00000001
+
+/*
+ * RxMacControl Register Content
+ */
+#define	AcceptBroadcast		0x0800
+#define	AcceptMulticast		0x0400
+#define	AcceptMyPhys		0x0200
+#define	AcceptAllPhys		0x0100
+#define	AcceptErr		0x0020
+#define	AcceptRunt		0x0010
+
+/* Station control register. */
+#define	SC_LOOPBACK		0x80000000
+#define	SC_RGMII		0x00008000
+#define	SC_FDX			0x00001000
+#define	SC_SPEED_MASK		0x00000c00
+#define	SC_SPEED_10		0x00000400
+#define	SC_SPEED_100		0x00000800
+#define	SC_SPEED_1000		0x00000c00
+
+/*
+ * Gigabit Media Independent Interface CTL register
+ */
+#define	GMI_DATA		0xffff0000
+#define	GMI_DATA_SHIFT		16
+#define	GMI_REG			0x0000f800
+#define	GMI_REG_SHIFT		11
+#define	GMI_PHY			0x000007c0
+#define	GMI_PHY_SHIFT		6
+#define	GMI_OP_WR		0x00000020
+#define	GMI_OP_RD		0x00000000
+#define	GMI_REQ			0x00000010
+#define	GMI_MDIO		0x00000008
+#define	GMI_MDDIR		0x00000004
+#define	GMI_MDC			0x00000002
+#define	GMI_MDEN		0x00000001
+
+/* Tx descriptor command bits. */
+#define	TDC_OWN			0x80000000
+#define	TDC_INTR		0x40000000
+#define	TDC_THOL3		0x30000000
+#define	TDC_THOL2		0x20000000
+#define	TDC_THOL1		0x10000000
+#define	TDC_THOL0		0x00000000
+#define	TDC_LS			0x08000000
+#define	TDC_IP_CSUM		0x04000000
+#define	TDC_TCP_CSUM		0x02000000
+#define	TDC_UDP_CSUM		0x01000000
+#define	TDC_BST			0x00800000
+#define	TDC_EXT			0x00400000
+#define	TDC_DEF			0x00200000
+#define	TDC_BKF			0x00100000
+#define	TDC_CRS			0x00080000
+#define	TDC_COL			0x00040000
+#define	TDC_CRC			0x00020000
+#define	TDC_PAD			0x00010000
+
+#define	SGE_TX_INTR_FRAMES	32
+
+/*
+ * TX descriptor status bits.
+ */
+#define	TDS_OWC			0x00080000
+#define	TDS_ABT			0x00040000
+#define	TDS_FIFO		0x00020000
+#define	TDS_CRS			0x00010000
+#define	TDS_COLLS		0x0000ffff
+#define	SGE_TX_ERROR(x)		((x) & (TDS_OWC | TDS_ABT | TDS_FIFO | TDS_CRS))
+#define	TX_ERR_BITS		"\20"				\
+				"\21CRS\22FIFO\23ABT\24OWC"
+
+/* Rx descriptor command bits. */
+#define	RDC_OWN			0x80000000
+#define	RDC_INTR		0x40000000
+#define	RDC_IP_CSUM		0x20000000
+#define	RDC_TCP_CSUM		0x10000000
+#define	RDC_UDP_CSUM		0x08000000
+#define	RDC_IP_CSUM_OK		0x04000000
+#define	RDC_TCP_CSUM_OK		0x02000000
+#define	RDC_UDP_CSUM_OK		0x01000000
+#define	RDC_WAKEUP		0x00400000
+#define	RDC_MAGIC		0x00200000
+#define	RDC_PAUSE		0x00100000
+#define	RDC_BCAST		0x000c0000
+#define	RDC_MCAST		0x00080000
+#define	RDC_UCAST		0x00040000
+#define	RDC_CRCOFF		0x00020000
+#define	RDC_PREADD		0x00010000
+
+/*
+ * RX descriptor status bits
+ */
+#define	RDS_TAGON		0x80000000
+#define	RDS_DESCS		0x3f000000
+#define	RDS_ABORT		0x00800000
+#define	RDS_SHORT		0x00400000
+#define	RDS_LIMIT		0x00200000
+#define	RDS_MIIER		0x00100000
+#define	RDS_OVRUN		0x00080000
+#define	RDS_NIBON		0x00040000
+#define	RDS_COLON		0x00020000
+#define	RDS_CRCOK		0x00010000
+#define	SGE_RX_ERROR(x)							\
+        ((x) & (RDS_COLON | RDS_NIBON | RDS_OVRUN | RDS_MIIER |		\
+	RDS_LIMIT | RDS_SHORT | RDS_ABORT))
+#define	SGE_RX_NSEGS(x)		(((x) & RDS_DESCS) >> 24)
+#define	RX_ERR_BITS 		"\20"					\
+				"\21CRCOK\22COLON\23NIBON\24OVRUN"	\
+				"\25MIIER\26LIMIT\27SHORT\30ABORT"	\
+				"\40TAGON"
+
+#define	RING_END		0x80000000
+#define	SGE_RX_BYTES(x)		((x) & 0xFFFF)
+#define	SGE_INC(x, y)		(x) = (((x) + 1) % y)
+
+/* Taken from Solaris driver */
+#define	EI_DATA			0xffff0000
+#define	EI_DATA_SHIFT		16
+#define	EI_OFFSET		0x0000fc00
+#define	EI_OFFSET_SHIFT		10
+#define	EI_OP			0x00000300
+#define	EI_OP_SHIFT		8
+#define	EI_OP_RD		(2 << EI_OP_SHIFT)
+#define	EI_OP_WR		(1 << EI_OP_SHIFT)
+#define	EI_REQ			0x00000080
+#define	EI_DO			0x00000008
+#define	EI_DI			0x00000004
+#define	EI_CLK			0x00000002
+#define	EI_CS			0x00000001
+
+/*
+ * EEPROM Addresses
+ */
+#define	EEPROMSignature		0x00
+#define	EEPROMCLK		0x01
+#define	EEPROMInfo		0x02
+#define	EEPROMMACAddr		0x03
+
+struct sge_desc {
+	uint32_t	sge_sts_size;
+	uint32_t	sge_cmdsts;
+	uint32_t	sge_ptr;
+	uint32_t	sge_flags;
+};
+
+#define	SGE_RX_RING_CNT		256 /* [8, 1024] */
+#define	SGE_TX_RING_CNT		256 /* [8, 8192] */
+#define	SGE_DESC_ALIGN		16
+#define	SGE_MAXTXSEGS		1
+#define	SGE_RX_BUF_ALIGN	sizeof(uint64_t)
+
+#define	SGE_RX_RING_SZ		(SGE_RX_RING_CNT * sizeof(struct sge_desc))
+#define	SGE_TX_RING_SZ		(SGE_TX_RING_CNT * sizeof(struct sge_desc))
+#define	SGE_ADDR_LO(x)		((uint64_t) (x) & 0xFFFFFFFF)
+
+struct sge_list_data {
+	struct sge_desc		*sge_rx_ring;
+	struct sge_desc		*sge_tx_ring;
+	/* physical bus addresses of sge_rx_ring/sge_tx_ring */
+	bus_addr_t		sge_rx_paddr;
+	bus_addr_t		sge_tx_paddr;
+};
+
+struct sge_chain_data {
+	bus_dma_tag_t		sge_tag;
+	bus_dma_tag_t		sge_rx_tag;
+	bus_dma_tag_t		sge_tx_tag;
+	bus_dmamap_t		sge_rx_dmamap;
+	bus_dmamap_t		sge_tx_dmamap;
+	bus_dma_tag_t		sge_txmbuf_tag;
+	bus_dma_tag_t		sge_rxmbuf_tag;
+	struct mbuf		*sge_rx_mbuf[SGE_RX_RING_CNT];
+	struct mbuf		*sge_tx_mbuf[SGE_TX_RING_CNT];
+	bus_dmamap_t		sge_rx_map[SGE_RX_RING_CNT];
+	bus_dmamap_t		sge_rx_spare_map;
+	bus_dmamap_t		sge_tx_map[SGE_TX_RING_CNT];
+	int			sge_rx_cons;
+	int			sge_tx_prod;
+	int			sge_tx_cons;
+	int			sge_tx_cnt;
+};
+
+struct sge_type {
+	uint16_t		sge_vid;
+	uint16_t		sge_did;
+	char			*sge_name;
+};
+
+struct sge_softc {
+	struct ifnet		*sge_ifp;	/* interface info */
+	struct resource		*sge_res;
+	int			sge_res_id;
+	int			sge_res_type;
+	struct resource		*sge_irq;
+	void			*sge_intrhand;
+	device_t		sge_dev;
+	device_t		sge_miibus;
+	uint8_t			sge_rev;
+	struct sge_list_data	sge_ldata;
+	struct sge_chain_data	sge_cdata;
+	struct callout		sge_stat_ch;
+	int			sge_timer;
+	int			sge_flags;
+#define	SGE_FLAG_FASTETHER	0x0001
+#define	SGE_FLAG_RGMII		0x0010
+#define	SGE_FLAG_SPEED_1000	0x2000
+#define	SGE_FLAG_FDX		0x4000
+#define	SGE_FLAG_LINK		0x8000
+	int			sge_if_flags;
+	int			sge_intrcontrol;
+	int			sge_intrtimer;
+	struct mtx		sge_mtx;
+};
+
+#define	SGE_LOCK(_sc)		mtx_lock(&(_sc)->sge_mtx)
+#define	SGE_UNLOCK(_sc)		mtx_unlock(&(_sc)->sge_mtx)
+#define	SGE_LOCK_ASSERT(_sc)	mtx_assert(&(_sc)->sge_mtx, MA_OWNED)
+
+#define	SGE_TIMEOUT		1000
+
+#endif /* _IF_SGEREG_H */
diff --git a/sys/i386/conf/GENERIC b/sys/i386/conf/GENERIC
index d40d7c0fa9f..2bc296342c5 100644
--- a/sys/i386/conf/GENERIC
+++ b/sys/i386/conf/GENERIC
@@ -235,6 +235,7 @@ device		pcn		# AMD Am79C97x PCI 10/100 (precedence over 'le')
 device		re		# RealTek 8139C+/8169/8169S/8110S
 device		rl		# RealTek 8129/8139
 device		sf		# Adaptec AIC-6915 (``Starfire'')
+device		sge		# Silicon Integrated Systems SiS190/191
 device		sis		# Silicon Integrated Systems SiS 900/SiS 7016
 device		sk		# SysKonnect SK-984x & SK-982x gigabit Ethernet
 device		ste		# Sundance ST201 (D-Link DFE-550TX)
diff --git a/sys/modules/Makefile b/sys/modules/Makefile
index 3af154c750b..83c7757c7e2 100644
--- a/sys/modules/Makefile
+++ b/sys/modules/Makefile
@@ -252,6 +252,7 @@ SUBDIR=	${_3dfx} \
 	sdhci \
 	sem \
 	sf \
+	sge \
 	siba_bwn \
 	siis \
 	sis \
diff --git a/sys/modules/sge/Makefile b/sys/modules/sge/Makefile
new file mode 100644
index 00000000000..5f8c5879554
--- /dev/null
+++ b/sys/modules/sge/Makefile
@@ -0,0 +1,8 @@
+# $FreeBSD$
+
+.PATH: ${.CURDIR}/../../dev/sge
+
+KMOD=	if_sge
+SRCS=	if_sge.c device_if.h bus_if.h pci_if.h miibus_if.h
+
+.include <bsd.kmod.mk>

From dec52eb94cc638ab4e4211d366ecde29b2fbc2b5 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:50:07 +0000
Subject: [PATCH 051/532] We don't need the definition for in_cksum repeated
 here since we get it from machine/in_cksum.h.  This definition prevents us
 from using hand-tuned assembler versions of in_cksum.

# this fixes the modules build on arm for ipfilter.
---
 sys/contrib/ipfilter/netinet/ip_compat.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/contrib/ipfilter/netinet/ip_compat.h b/sys/contrib/ipfilter/netinet/ip_compat.h
index 31e5b114e00..35c9f68ffc0 100644
--- a/sys/contrib/ipfilter/netinet/ip_compat.h
+++ b/sys/contrib/ipfilter/netinet/ip_compat.h
@@ -975,7 +975,6 @@ typedef	u_int32_t	u_32_t;
 #   define	SPL_NET(x)	;
 #   define	SPL_IMP(x)	;
 #   define	SPL_SCHED(x)	;
-extern	int	in_cksum __P((struct mbuf *, int));
 #  else
 #   define	SPL_SCHED(x)	x = splhigh()
 #  endif /* __FreeBSD_version >= 500043 */

From 451f0bf403f4a5bf37fdffac4da1ecd4aef7228b Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:54:23 +0000
Subject: [PATCH 052/532] Add sge(4) to the list of supported network
 interface.

---
 usr.sbin/sysinstall/devices.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/usr.sbin/sysinstall/devices.c b/usr.sbin/sysinstall/devices.c
index a2b7939f048..8cf1fb30c49 100644
--- a/usr.sbin/sysinstall/devices.c
+++ b/usr.sbin/sysinstall/devices.c
@@ -150,6 +150,7 @@ static struct _devname {
     NETWORK("rue",	"RealTek USB Ethernet card"),
     NETWORK("rum",	"Ralink Technology USB IEEE 802.11 wireless adapter"),
     NETWORK("sf",	"Adaptec AIC-6915 PCI Ethernet card"),
+    NETWORK("sge",	"Silicon Integrated Systems SiS190/191 Ethernet"),
     NETWORK("sis",	"SiS 900/SiS 7016 PCI Ethernet card"),
 #ifdef PC98
     NETWORK("snc",	"SONIC Ethernet card"),

From d43a11879733cf48553f00c7b0029f57a3a59aa5 Mon Sep 17 00:00:00 2001
From: Jack F Vogel <jfv@FreeBSD.org>
Date: Wed, 14 Apr 2010 20:55:33 +0000
Subject: [PATCH 053/532] Add a missing fragment in the tx msix handler to
 invoke another if all work is not done.

Sync the igb driver with changes suggested by yongari and
made in em, these made sense to be in both drivers.
---
 sys/dev/e1000/if_em.c  | 11 ++++++++---
 sys/dev/e1000/if_igb.c | 36 +++++++++++++++++++++++-------------
 2 files changed, 31 insertions(+), 16 deletions(-)

diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
index 834e676a6af..1db90121ddb 100644
--- a/sys/dev/e1000/if_em.c
+++ b/sys/dev/e1000/if_em.c
@@ -93,7 +93,7 @@ int	em_display_debug_stats = 0;
 /*********************************************************************
  *  Driver version:
  *********************************************************************/
-char em_driver_version[] = "7.0.4";
+char em_driver_version[] = "7.0.5";
 
 
 /*********************************************************************
@@ -1484,12 +1484,17 @@ em_msix_tx(void *arg)
 {
 	struct tx_ring *txr = arg;
 	struct adapter *adapter = txr->adapter;
+	bool		more;
 
 	++txr->tx_irq;
 	EM_TX_LOCK(txr);
-	em_txeof(txr);
+	more = em_txeof(txr);
 	EM_TX_UNLOCK(txr);
-	E1000_WRITE_REG(&adapter->hw, E1000_IMS, txr->ims);
+	if (more)
+		taskqueue_enqueue(txr->tq, &txr->tx_task);
+	else
+		/* Reenable this interrupt */
+		E1000_WRITE_REG(&adapter->hw, E1000_IMS, txr->ims);
 	return;
 }
 
diff --git a/sys/dev/e1000/if_igb.c b/sys/dev/e1000/if_igb.c
index 267590faf0e..e901bc24e2f 100644
--- a/sys/dev/e1000/if_igb.c
+++ b/sys/dev/e1000/if_igb.c
@@ -99,7 +99,7 @@ int	igb_display_debug_stats = 0;
 /*********************************************************************
  *  Driver version:
  *********************************************************************/
-char igb_driver_version[] = "version - 1.9.4";
+char igb_driver_version[] = "version - 1.9.5";
 
 
 /*********************************************************************
@@ -758,8 +758,15 @@ igb_start_locked(struct tx_ring *txr, struct ifnet *ifp)
 	if (!adapter->link_active)
 		return;
 
-	while (!IFQ_DRV_IS_EMPTY(&ifp->if_snd)) {
+	/* Call cleanup if number of TX descriptors low */
+	if (txr->tx_avail <= IGB_TX_CLEANUP_THRESHOLD)
+		igb_txeof(txr);
 
+	while (!IFQ_DRV_IS_EMPTY(&ifp->if_snd)) {
+		if (txr->tx_avail <= IGB_TX_OP_THRESHOLD) {
+			ifp->if_drv_flags |= IFF_DRV_OACTIVE;
+			break;
+		}
 		IFQ_DRV_DEQUEUE(&ifp->if_snd, m_head);
 		if (m_head == NULL)
 			break;
@@ -779,6 +786,7 @@ igb_start_locked(struct tx_ring *txr, struct ifnet *ifp)
 		ETHER_BPF_MTAP(ifp, m_head);
 
 		/* Set watchdog on */
+		txr->watchdog_time = ticks;
 		txr->watchdog_check = TRUE;
 	}
 }
@@ -817,8 +825,6 @@ igb_mq_start(struct ifnet *ifp, struct mbuf *m)
 	/* Which queue to use */
 	if ((m->m_flags & M_FLOWID) != 0)
 		i = m->m_pkthdr.flowid % adapter->num_queues;
-	else
-		i = curcpu % adapter->num_queues;
 
 	txr = &adapter->tx_rings[i];
 
@@ -847,6 +853,10 @@ igb_mq_start_locked(struct ifnet *ifp, struct tx_ring *txr, struct mbuf *m)
 		return (err);
 	}
 
+	/* Call cleanup if number of TX descriptors low */
+	if (txr->tx_avail <= IGB_TX_CLEANUP_THRESHOLD)
+		igb_txeof(txr);
+
 	enq = 0;
 	if (m == NULL) {
 		next = drbr_dequeue(ifp, txr->br);
@@ -856,6 +866,7 @@ igb_mq_start_locked(struct ifnet *ifp, struct tx_ring *txr, struct mbuf *m)
 		next = drbr_dequeue(ifp, txr->br);
 	} else
 		next = m;
+
 	/* Process the queue */
 	while (next != NULL) {
 		if ((err = igb_xmit(txr, &next)) != 0) {
@@ -877,6 +888,7 @@ igb_mq_start_locked(struct ifnet *ifp, struct tx_ring *txr, struct mbuf *m)
 	if (enq > 0) {
 		/* Set the watchdog */
 		txr->watchdog_check = TRUE;
+		txr->watchdog_time = ticks;
 	}
 	return (err);
 }
@@ -1248,19 +1260,13 @@ igb_handle_que(void *context, int pending)
 	struct adapter *adapter = que->adapter;
 	struct tx_ring *txr = que->txr;
 	struct ifnet	*ifp = adapter->ifp;
-	u32		loop = IGB_MAX_LOOP;
 	bool		more;
 
-	/* RX first */
-	do {
+	if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
 		more = igb_rxeof(que, -1);
-	} while (loop-- && more);
 
-	if (IGB_TX_TRYLOCK(txr)) {
-		loop = IGB_MAX_LOOP;
-		do {
-			more = igb_txeof(txr);
-		} while (loop-- && more);
+		IGB_TX_LOCK(txr);
+		igb_txeof(txr);
 #if __FreeBSD_version >= 800000
 		igb_mq_start_locked(ifp, txr, NULL);
 #else
@@ -1268,6 +1274,10 @@ igb_handle_que(void *context, int pending)
 			igb_start_locked(txr, ifp);
 #endif
 		IGB_TX_UNLOCK(txr);
+		if (more) {
+			taskqueue_enqueue(que->tq, &que->que_task);
+			return;
+		}
 	}
 
 	/* Reenable this interrupt */

From 9b00b6e1ee532360ba70ef40fe80232e3719b699 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Wed, 14 Apr 2010 21:27:48 +0000
Subject: [PATCH 054/532] Add sge(4) man page and hook up sge(4) to the build.
 Also add Xr to appropriate man pages.

---
 share/man/man4/Makefile |   1 +
 share/man/man4/altq.4   |   3 +-
 share/man/man4/miibus.4 |   5 +-
 share/man/man4/sge.4    | 120 ++++++++++++++++++++++++++++++++++++++++
 share/man/man4/vlan.4   |   3 +-
 5 files changed, 129 insertions(+), 3 deletions(-)
 create mode 100644 share/man/man4/sge.4

diff --git a/share/man/man4/Makefile b/share/man/man4/Makefile
index 4c499847a02..c71a09740e7 100644
--- a/share/man/man4/Makefile
+++ b/share/man/man4/Makefile
@@ -350,6 +350,7 @@ MAN=	aac.4 \
 	sem.4 \
 	ses.4 \
 	sf.4 \
+	sge.4 \
 	si.4 \
 	siba.4 \
 	sio.4 \
diff --git a/share/man/man4/altq.4 b/share/man/man4/altq.4
index 8a568428d40..a618c966e89 100644
--- a/share/man/man4/altq.4
+++ b/share/man/man4/altq.4
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd July 26, 2009
+.Dd April 14, 2010
 .Dt ALTQ 4
 .Os
 .Sh NAME
@@ -151,6 +151,7 @@ They have been applied to the following hardware drivers:
 .Xr rl 4 ,
 .Xr rum 4 ,
 .Xr sf 4 ,
+.Xr sge 4 ,
 .Xr sis 4 ,
 .Xr sk 4 ,
 .Xr ste 4 ,
diff --git a/share/man/man4/miibus.4 b/share/man/man4/miibus.4
index ec939671677..5a385957231 100644
--- a/share/man/man4/miibus.4
+++ b/share/man/man4/miibus.4
@@ -8,7 +8,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd June 14, 2009
+.Dd April 14, 2010
 .Dt MIIBUS 4
 .Os
 .Sh NAME
@@ -97,6 +97,8 @@ RealTek 8129/8139
 RealTek RTL8150 USB To Fast Ethernet
 .It Xr sf 4
 Adaptec AIC-6915
+.It Xr sge 4
+Silicon Integrated Systems SiS190/191 Ethernet
 .It Xr sis 4
 Silicon Integrated Systems SiS 900/SiS 7016
 .It Xr sk 4
@@ -158,6 +160,7 @@ but as a result are not well behaved newbus device drivers.
 .Xr rl 4 ,
 .Xr rue 4 ,
 .Xr sf 4 ,
+.Xr sge 4 ,
 .Xr sis 4 ,
 .Xr sk 4 ,
 .Xr ste 4 ,
diff --git a/share/man/man4/sge.4 b/share/man/man4/sge.4
new file mode 100644
index 00000000000..a3d057d386c
--- /dev/null
+++ b/share/man/man4/sge.4
@@ -0,0 +1,120 @@
+.\" Copyright (c) 2010 Pyun YongHyeon
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in the
+.\"    documentation and/or other materials provided with the distribution.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+.\" SUCH DAMAGE.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd April 14, 2010
+.Dt SGE 4
+.Os
+.Sh NAME
+.Nm sge
+.Nd Silicon Integrated Systems SiS190/191 Fast/Gigabit Ethernet driver
+.Sh SYNOPSIS
+To compile this driver into the kernel,
+place the following lines in your
+kernel configuration file:
+.Bd -ragged -offset indent
+.Cd "device miibus"
+.Cd "device sge"
+.Ed
+.Pp
+Alternatively, to load the driver as a
+module at boot time, place the following line in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+if_sge="YES"
+.Ed
+.Sh DESCRIPTION
+The
+.Nm
+device driver provides support for SiS190 Fast Ethernet
+controllers and SiS191 Fast/Gigabit Ethernet controllers.
+.Pp
+All LOMs supported by the
+.Nm
+driver have TCP/UDP/IP checksum offload for transmit and receive.
+Due to lack of documentation more offloading features like TCP
+segmentation offload (TSO), hardware VLAN tag stripping/insertion
+features, Wake On Lan (WOL), Jumbo frame and an interrupt moderation
+mechanism are not supported yet.
+.Pp
+The
+.Nm
+driver supports the following media types:
+.Bl -tag -width ".Cm 10baseT/UTP"
+.It Cm autoselect
+Enable autoselection of the media type and options.
+The user can manually override
+the autoselected mode by adding media options to
+.Xr rc.conf 5 .
+.It Cm 10baseT/UTP
+Set 10Mbps operation.
+.It Cm 100baseTX
+Set 100Mbps (Fast Ethernet) operation.
+.It Cm 1000baseTX
+Set 1000baseTX operation over twisted pair.
+.El
+.Pp
+The
+.Nm
+driver supports the following media options:
+.Bl -tag -width ".Cm full-duplex"
+.It Cm full-duplex
+Force full duplex operation.
+.It Cm half-duplex
+Force half duplex operation.
+.El
+.Pp
+For more information on configuring this device, see
+.Xr ifconfig 8 .
+.Sh HARDWARE
+The
+.Nm
+device driver provides support for the following Ethernet controllers:
+.Pp
+.Bl -bullet -compact
+.It
+SiS190 Fast Ethernet controller
+.It
+SiS191 Fast/Gigabit Ethernet controller
+.El
+.Sh SEE ALSO
+.Xr altq 4 ,
+.Xr arp 4 ,
+.Xr miibus 4 ,
+.Xr netintro 4 ,
+.Xr ng_ether 4 ,
+.Xr vlan 4 ,
+.Xr ifconfig 8
+.Sh HISTORY
+The
+.Nm
+driver was written by
+.An Alexander Pohoyda
+.Aq alexander.pohoyda@gmx.net .
+And enhanced by
+.An Nikolay Denev
+.Aq ndenev@gmail.com .
+It first appeared in
+.Fx 8.1 .
diff --git a/share/man/man4/vlan.4 b/share/man/man4/vlan.4
index 149d4bc5fcd..f7f7807a86d 100644
--- a/share/man/man4/vlan.4
+++ b/share/man/man4/vlan.4
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd June 14, 2009
+.Dd April 14, 2010
 .Dt VLAN 4
 .Os
 .Sh NAME
@@ -172,6 +172,7 @@ natively:
 .Xr nve 4 ,
 .Xr rl 4 ,
 .Xr sf 4 ,
+.Xr sge 4 ,
 .Xr sis 4 ,
 .Xr sk 4 ,
 .Xr ste 4 ,

From 94a04272dd05c37153769d2283b6d189fbfbf60d Mon Sep 17 00:00:00 2001
From: Julian Elischer <julian@FreeBSD.org>
Date: Wed, 14 Apr 2010 21:42:29 +0000
Subject: [PATCH 055/532] Change the semantics of the debug.ktr.alq_enable 
 control so that when you disable alq, it acts as if alq had not been enabled
 in the build. in other words, the rest of ktr is still available for use. If
 you really don't want that as well, set the mask to 0.

MFC after:3 weeks
---
 sys/kern/kern_ktr.c | 54 +++++++++++++++++++++++++--------------------
 1 file changed, 30 insertions(+), 24 deletions(-)

diff --git a/sys/kern/kern_ktr.c b/sys/kern/kern_ktr.c
index 6117cd6cbba..9586ae65278 100644
--- a/sys/kern/kern_ktr.c
+++ b/sys/kern/kern_ktr.c
@@ -199,9 +199,8 @@ ktr_tracepoint(u_int mask, const char *file, int line, const char *format,
 	struct ktr_entry *entry;
 #ifdef KTR_ALQ
 	struct ale *ale = NULL;
-#else
-	int newindex, saveindex;
 #endif
+	int newindex, saveindex;
 #if defined(KTR_VERBOSE) || defined(KTR_ALQ)
 	struct thread *td;
 #endif
@@ -221,27 +220,30 @@ ktr_tracepoint(u_int mask, const char *file, int line, const char *format,
 	td->td_pflags |= TDP_INKTR;
 #endif
 #ifdef KTR_ALQ
-	if (ktr_alq_enabled &&
-	    td->td_critnest == 0 &&
-	    (td->td_flags & TDF_IDLETD) == 0 &&
-	    td != ald_thread) {
-		if (ktr_alq_max && ktr_alq_cnt > ktr_alq_max)
-			goto done;
-		if ((ale = alq_get(ktr_alq, ALQ_NOWAIT)) == NULL) {
-			ktr_alq_failed++;
+	if (ktr_alq_enabled) {
+		if (td->td_critnest == 0 &&
+		    (td->td_flags & TDF_IDLETD) == 0 &&
+		    td != ald_thread) {
+			if (ktr_alq_max && ktr_alq_cnt > ktr_alq_max)
+				goto done;
+			if ((ale = alq_get(ktr_alq, ALQ_NOWAIT)) == NULL) {
+				ktr_alq_failed++;
+				goto done;
+			}
+			ktr_alq_cnt++;
+			entry = (struct ktr_entry *)ale->ae_data;
+		} else {
 			goto done;
 		}
-		ktr_alq_cnt++;
-		entry = (struct ktr_entry *)ale->ae_data;
 	} else
-		goto done;
-#else
-	do {
-		saveindex = ktr_idx;
-		newindex = (saveindex + 1) & (KTR_ENTRIES - 1);
-	} while (atomic_cmpset_rel_int(&ktr_idx, saveindex, newindex) == 0);
-	entry = &ktr_buf[saveindex];
 #endif
+	{
+		do {
+			saveindex = ktr_idx;
+			newindex = (saveindex + 1) & (KTR_ENTRIES - 1);
+		} while (atomic_cmpset_rel_int(&ktr_idx, saveindex, newindex) == 0);
+		entry = &ktr_buf[saveindex];
+	}
 	entry->ktr_timestamp = KTR_TIME;
 	entry->ktr_cpu = cpu;
 	entry->ktr_thread = curthread;
@@ -271,7 +273,7 @@ ktr_tracepoint(u_int mask, const char *file, int line, const char *format,
 	entry->ktr_parms[4] = arg5;
 	entry->ktr_parms[5] = arg6;
 #ifdef KTR_ALQ
-	if (ale)
+	if (ktr_alq_enabled && ale)
 		alq_post(ktr_alq, ale);
 done:
 #endif
@@ -295,7 +297,9 @@ DB_SHOW_COMMAND(ktr, db_ktr_all)
 	
 	tstate.cur = (ktr_idx - 1) & (KTR_ENTRIES - 1);
 	tstate.first = -1;
-	db_ktr_verbose = index(modif, 'v') != NULL;
+	db_ktr_verbose = 0;
+	db_ktr_verbose |= (index(modif, 'v') != NULL) ? 2 : 0;
+	db_ktr_verbose |= (index(modif, 'V') != NULL) ? 1 : 0; /* just timestap please */
 	if (index(modif, 'a') != NULL) {
 		db_disable_pager();
 		while (cncheckc() != -1)
@@ -329,9 +333,11 @@ db_mach_vtrace(void)
 	db_printf(":cpu%d", kp->ktr_cpu);
 #endif
 	db_printf(")");
-	if (db_ktr_verbose) {
-		db_printf(" %10.10lld %s.%d", (long long)kp->ktr_timestamp,
-		    kp->ktr_file, kp->ktr_line);
+	if (db_ktr_verbose >= 1) {
+		db_printf(" %10.10lld", (long long)kp->ktr_timestamp);
+	}
+	if (db_ktr_verbose >= 2) {
+		db_printf(" %s.%d", kp->ktr_file, kp->ktr_line);
 	}
 	db_printf(": ");
 	db_printf(kp->ktr_desc, kp->ktr_parms[0], kp->ktr_parms[1],

From f6b93040fd95c76758b371e160dbf6d996b14336 Mon Sep 17 00:00:00 2001
From: Julian Elischer <julian@FreeBSD.org>
Date: Wed, 14 Apr 2010 21:44:14 +0000
Subject: [PATCH 056/532] Man page change to documant slight change in ktr_alq
 behaviour.

---
 share/man/man4/ktr.4 | 10 ++++++++--
 1 file changed, 8 insertions(+), 2 deletions(-)

diff --git a/share/man/man4/ktr.4 b/share/man/man4/ktr.4
index 89f530596f5..7da4e373051 100644
--- a/share/man/man4/ktr.4
+++ b/share/man/man4/ktr.4
@@ -122,7 +122,7 @@ option sets the flag to one.
 The KTR buffer can be examined from within
 .Xr ddb 4
 via the
-.Ic show ktr Op Cm /v
+.Ic show ktr Op Cm /vV
 command.
 This command displays the contents of the trace buffer one page at a time.
 At the
@@ -136,6 +136,10 @@ If the
 .Cm /v
 modifier is specified, then they are displayed in addition to the normal
 output.
+If the
+.Cm /V
+modifier is specified, then just the timestamp is displayed in
+addition to the normal output.
 Note that the events are displayed in reverse chronological order.
 That is, the most recent events are displayed first.
 .Ss Logging ktr to Disk
@@ -167,7 +171,9 @@ the next invocation.
 enables logging of
 .Nm
 entries to disk if it is set to one.
-Setting this to 0 will terminate logging.
+Setting this to 0 will terminate logging to disk and revert to 
+logging to the normal ktr ring buffer.
+Data is not sent to the ring buffer while logging to disk.
 .It Va debug.ktr.alq_max
 is the maximum number of entries that will be recorded to disk, or 0 for
 infinite.

From 410c7662749344645598007a28d70d283541347d Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Wed, 14 Apr 2010 21:53:27 +0000
Subject: [PATCH 057/532] Move fatal error at the right place. Fix exit from
 top mode when checking if PMC is available.

MFC after: 3 days
---
 usr.sbin/pmcstat/pmcpl_callgraph.c | 2 ++
 usr.sbin/pmcstat/pmcpl_gprof.c     | 2 ++
 usr.sbin/pmcstat/pmcstat_log.c     | 2 --
 3 files changed, 4 insertions(+), 2 deletions(-)

diff --git a/usr.sbin/pmcstat/pmcpl_callgraph.c b/usr.sbin/pmcstat/pmcpl_callgraph.c
index 53d342bfdc1..d948b77f43b 100644
--- a/usr.sbin/pmcstat/pmcpl_callgraph.c
+++ b/usr.sbin/pmcstat/pmcpl_callgraph.c
@@ -581,6 +581,8 @@ pmcpl_cg_topdisplay(void)
 	struct pmcstat_pmcrecord *pmcr;
 
 	pmcr = pmcstat_pmcindex_to_pmcr(pmcstat_pmcinfilter);
+	if (!pmcr)
+		err(EX_SOFTWARE, "ERROR: invalid pmcindex");
 
 	/*
 	 * We pull out all callgraph nodes in the top-level hash table
diff --git a/usr.sbin/pmcstat/pmcpl_gprof.c b/usr.sbin/pmcstat/pmcpl_gprof.c
index 9327eb95b62..2027ecf38e7 100644
--- a/usr.sbin/pmcstat/pmcpl_gprof.c
+++ b/usr.sbin/pmcstat/pmcpl_gprof.c
@@ -171,6 +171,8 @@ pmcstat_gmon_create_name(const char *samplesdir, struct pmcstat_image *image,
 	char fullpath[PATH_MAX];
 
 	pmcname = pmcstat_pmcid_to_name(pmcid);
+	if (!pmcname)
+		err(EX_SOFTWARE, "ERROR: cannot find pmcid");
 
 	(void) snprintf(fullpath, sizeof(fullpath),
 	    "%s/%s/%s", samplesdir, pmcname,
diff --git a/usr.sbin/pmcstat/pmcstat_log.c b/usr.sbin/pmcstat/pmcstat_log.c
index 51f66ca9e25..6d8c57a760a 100644
--- a/usr.sbin/pmcstat/pmcstat_log.c
+++ b/usr.sbin/pmcstat/pmcstat_log.c
@@ -1050,7 +1050,6 @@ pmcstat_pmcid_to_name(pmc_id_t pmcid)
 	    if (pr->pr_pmcid == pmcid)
 		    return (pmcstat_string_unintern(pr->pr_pmcname));
 
-	err(EX_SOFTWARE, "ERROR: cannot find pmcid");
 	return NULL;
 }
 
@@ -1083,7 +1082,6 @@ pmcstat_pmcindex_to_pmcr(int pmcin)
 		if (pr->pr_pmcin == pmcin)
 			return pr;
 
-	err(EX_SOFTWARE, "ERROR: invalid pmcindex");
 	return NULL;
 }
 

From 57d848483eb9cb8a270e4355e4634ae940a7d7dd Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Wed, 14 Apr 2010 22:02:19 +0000
Subject: [PATCH 058/532] When an underlying ioctl(2) handler returns an error,
 our ioctl(2) interface considers that it hits a fatal error, and will not
 copyout the request structure back for _IOW and _IOWR ioctls, keeping them
 untouched.

The previous implementation of the SIOCGIFDESCR ioctl intends to
feed the buffer length back to userland.  However, if we return
an error, the feedback would be defeated and ifconfig(8) would
trap into an infinite loop.

This commit changes SIOCGIFDESCR to set buffer field to NULL to
indicate the previous ENAMETOOLONG case.

Reported by:	bschmidt
MFC after:	2 weeks
---
 sbin/ifconfig/ifconfig.c  | 23 ++++++++++++-----------
 share/man/man4/netintro.4 |  9 ++++++---
 sys/net/if.c              |  7 +++----
 3 files changed, 21 insertions(+), 18 deletions(-)

diff --git a/sbin/ifconfig/ifconfig.c b/sbin/ifconfig/ifconfig.c
index 7af3306d68f..aa961755722 100644
--- a/sbin/ifconfig/ifconfig.c
+++ b/sbin/ifconfig/ifconfig.c
@@ -922,19 +922,20 @@ status(const struct afswtch *afp, const struct sockaddr_dl *sdl,
 			ifr.ifr_buffer.buffer = descr;
 			ifr.ifr_buffer.length = descrlen;
 			if (ioctl(s, SIOCGIFDESCR, &ifr) == 0) {
-				if (strlen(descr) > 0)
-					printf("\tdescription: %s\n", descr);
-				break;
-			} else if (errno == ENAMETOOLONG)
-				descrlen = ifr.ifr_buffer.length;
-			else
-				break;
-		} else {
+				if (ifr.ifr_buffer.buffer == descr) {
+					if (strlen(descr) > 0)
+						printf("\tdescription: %s\n",
+						    descr);
+				} else if (ifr.ifr_buffer.length > descrlen) {
+					descrlen = ifr.ifr_buffer.length;
+					continue;
+				}
+			}
+		} else
 			warn("unable to allocate memory for interface"
 			    "description");
-			break;
-		}
-	};
+		break;
+	}
 
 	if (ioctl(s, SIOCGIFCAP, (caddr_t)&ifr) == 0) {
 		if (ifr.ifr_curcap != 0) {
diff --git a/share/man/man4/netintro.4 b/share/man/man4/netintro.4
index 348a13e68fd..3e989401777 100644
--- a/share/man/man4/netintro.4
+++ b/share/man/man4/netintro.4
@@ -32,7 +32,7 @@
 .\"     @(#)netintro.4	8.2 (Berkeley) 11/30/93
 .\" $FreeBSD$
 .\"
-.Dd January 26, 2010
+.Dd April 14, 2010
 .Dt NETINTRO 4
 .Os
 .Sh NAME
@@ -292,8 +292,11 @@ field of
 struct passed in as parameter, and the length would include
 the terminating nul character.
 If there is not enough space to hold the interface length,
-no copy would be done and an
-error would be returned.
+no copy would be done and the
+.Va buffer
+field of
+.Va ifru_buffer
+would be set to NULL.
 The kernel will store the buffer length in the
 .Va length
 field upon return, regardless whether the buffer itself is
diff --git a/sys/net/if.c b/sys/net/if.c
index e4a20054963..98c8afa4a2b 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -2049,14 +2049,13 @@ ifhwioctl(u_long cmd, struct ifnet *ifp, caddr_t data, struct thread *td)
 	case SIOCGIFDESCR:
 		error = 0;
 		sx_slock(&ifdescr_sx);
-		if (ifp->if_description == NULL) {
-			ifr->ifr_buffer.length = 0;
+		if (ifp->if_description == NULL)
 			error = ENOMSG;
-		} else {
+		else {
 			/* space for terminating nul */
 			descrlen = strlen(ifp->if_description) + 1;
 			if (ifr->ifr_buffer.length < descrlen)
-				error = ENAMETOOLONG;
+				ifr->ifr_buffer.buffer = NULL;
 			else
 				error = copyout(ifp->if_description,
 				    ifr->ifr_buffer.buffer, descrlen);

From 2d8d6ee437981b190cc77ac7fade864800754239 Mon Sep 17 00:00:00 2001
From: Gavin Atkinson <gavin@FreeBSD.org>
Date: Wed, 14 Apr 2010 22:03:48 +0000
Subject: [PATCH 059/532] Use the UIPROTO_BOOT_KEYBOARD #define from usb.h
 rather than a local (almost identically named) local #define.

Reviewed by:	hselasky
---
 sys/dev/usb/input/ukbd.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/sys/dev/usb/input/ukbd.c b/sys/dev/usb/input/ukbd.c
index 8584f8d505d..f446db7a28c 100644
--- a/sys/dev/usb/input/ukbd.c
+++ b/sys/dev/usb/input/ukbd.c
@@ -102,8 +102,6 @@ TUNABLE_INT("hw.usb.ukbd.debug", &ukbd_debug);
 TUNABLE_INT("hw.usb.ukbd.no_leds", &ukbd_no_leds);
 #endif
 
-#define	UPROTO_BOOT_KEYBOARD 1
-
 #define	UKBD_EMULATE_ATSCANCODE	       1
 #define	UKBD_DRIVER_NAME          "ukbd"
 #define	UKBD_NMOD                     8	/* units */
@@ -770,7 +768,7 @@ ukbd_probe(device_t dev)
 		return (ENXIO);
 
 	if ((uaa->info.bInterfaceSubClass == UISUBCLASS_BOOT) &&
-	    (uaa->info.bInterfaceProtocol == UPROTO_BOOT_KEYBOARD)) {
+	    (uaa->info.bInterfaceProtocol == UIPROTO_BOOT_KEYBOARD)) {
 		if (usb_test_quirk(uaa, UQ_KBD_IGNORE))
 			return (ENXIO);
 		else

From 7a90b21212384b51a24738d850302b5cade42b90 Mon Sep 17 00:00:00 2001
From: Julian Elischer <julian@FreeBSD.org>
Date: Wed, 14 Apr 2010 23:06:07 +0000
Subject: [PATCH 060/532] Move two copies of the same definition to a common
 include file.

MFC after: 3 weeks
---
 sys/ddb/db_sym.c | 6 ------
 sys/net/vnet.c   | 9 ---------
 sys/net/vnet.h   | 9 +++++++++
 3 files changed, 9 insertions(+), 15 deletions(-)

diff --git a/sys/ddb/db_sym.c b/sys/ddb/db_sym.c
index 99209a89c83..04af1eba535 100644
--- a/sys/ddb/db_sym.c
+++ b/sys/ddb/db_sym.c
@@ -64,12 +64,6 @@ static boolean_t	db_line_at_pc(c_db_sym_t, char **, int *, db_expr_t);
 static int db_cpu = -1;
 
 #ifdef VIMAGE
-extern uintptr_t	*__start_set_vnet;
-extern uintptr_t	*__stop_set_vnet;
-
-#define	VNET_START	(uintptr_t)&__start_set_vnet
-#define	VNET_STOP	(uintptr_t)&__stop_set_vnet
-
 static void *db_vnet = NULL;
 #endif
 
diff --git a/sys/net/vnet.c b/sys/net/vnet.c
index 1dac03abf1d..8013f5ed594 100644
--- a/sys/net/vnet.c
+++ b/sys/net/vnet.c
@@ -153,15 +153,6 @@ struct vnet *vnet0;
  * module will find every network stack instance with proper default values.
  */
 
-/*
- * Location of the kernel's 'set_vnet' linker set.
- */
-extern uintptr_t	*__start_set_vnet;
-extern uintptr_t	*__stop_set_vnet;
-
-#define	VNET_START	(uintptr_t)&__start_set_vnet
-#define	VNET_STOP	(uintptr_t)&__stop_set_vnet
-
 /*
  * Number of bytes of data in the 'set_vnet' linker set, and hence the total
  * size of all kernel virtualized global variables, and the malloc(9) type
diff --git a/sys/net/vnet.h b/sys/net/vnet.h
index fb2cc393cb7..4cdfdef3063 100644
--- a/sys/net/vnet.h
+++ b/sys/net/vnet.h
@@ -91,6 +91,15 @@ struct vnet {
 #include <sys/rwlock.h>
 #include <sys/sx.h>
 
+/*
+ * Location of the kernel's 'set_vnet' linker set.
+ */
+extern uintptr_t	*__start_set_vnet;
+extern uintptr_t	*__stop_set_vnet;
+
+#define	VNET_START	(uintptr_t)&__start_set_vnet
+#define	VNET_STOP	(uintptr_t)&__stop_set_vnet
+
 /*
  * Functions to allocate and destroy virtual network stacks.
  */

From a3a86fc3973f7a598beba6b0c04837d095d05a12 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 15 Apr 2010 08:29:14 +0000
Subject: [PATCH 061/532] scsi_cd: CD_FLAG_VALID_MEDIA is sufficient to set
 d_sectorsize and d_mediasize

CD_FLAG_VALID_TOC is not required for setting those media properties.

PR:		kern/145385
Submitted by:	Juergen Lock <nox@jelal.kn-bremen.de>
		a slightly different version
Tested by:	Pavel Sukhoy <sukhoy@ripn.net>,
		Markus Wild <m.wild@cybernet.ch>,
		Juergen Lock <nox@jelal.kn-bremen.de>,
		uqs
MFC after:	1 week
---
 sys/cam/scsi/scsi_cd.c | 9 +++++----
 1 file changed, 5 insertions(+), 4 deletions(-)

diff --git a/sys/cam/scsi/scsi_cd.c b/sys/cam/scsi/scsi_cd.c
index 1e5be3c5886..d09d129a947 100644
--- a/sys/cam/scsi/scsi_cd.c
+++ b/sys/cam/scsi/scsi_cd.c
@@ -2773,8 +2773,12 @@ cdcheckmedia(struct cam_periph *periph)
 		softc->flags &= ~(CD_FLAG_VALID_MEDIA|CD_FLAG_VALID_TOC);
 		cdprevent(periph, PR_ALLOW);
 		return (error);
-	} else
+	} else {
 		softc->flags |= CD_FLAG_VALID_MEDIA;
+		softc->disk->d_sectorsize = softc->params.blksize;
+		softc->disk->d_mediasize =
+		    (off_t)softc->params.blksize * softc->params.disksize;
+	}
 
 	/*
 	 * Now we check the table of contents.  This (currently) is only
@@ -2863,9 +2867,6 @@ cdcheckmedia(struct cam_periph *periph)
 	}
 
 	softc->flags |= CD_FLAG_VALID_TOC;
-	softc->disk->d_sectorsize = softc->params.blksize;
-	softc->disk->d_mediasize =
-	    (off_t)softc->params.blksize * softc->params.disksize;
 
 bailout:
 

From fdfa00ba835ea7dd8128a8ae0ea4f191197a7d06 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 15 Apr 2010 08:32:50 +0000
Subject: [PATCH 062/532] Still reference struct __sigaction with clarification
 when this form of argument declaration is needed.

Discussed with:	bde
MFC after:	3 days
---
 lib/libc/sys/sigaction.2 | 10 ++++++++++
 1 file changed, 10 insertions(+)

diff --git a/lib/libc/sys/sigaction.2 b/lib/libc/sys/sigaction.2
index 9bd4c9e4f36..51b6c47c78f 100644
--- a/lib/libc/sys/sigaction.2
+++ b/lib/libc/sys/sigaction.2
@@ -496,6 +496,16 @@ or
 .Dv SIG_IGN
 this way.
 .Pp
+If preprocessing symbol
+.Va _POSIX_C_SOURCE
+with the value >= 199309 is not defined, the following declaration for
+the handler shall be used:
+.Bl -tag -offset indent -width short
+.It Tn POSIX Dv SA_SIGINFO :
+.Ft void
+.Fn handler int "struct __sigaction *" "void *" ;
+.El
+.Pp
 If the
 .Dv SA_SIGINFO
 flag is not set, the handler function should match

From 2a842317eb386798cb7a8e14aad7068d5546e259 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 15 Apr 2010 08:39:56 +0000
Subject: [PATCH 063/532] g_io_check: respond to zero pp->mediasize with ENXIO

Previsouly this condition was reported with EIO by bio_offset > mediasize
check.
Perhaps that check should be extended to bio_offset+bio_length > mediasize.

MFC after:	1 week
---
 sys/geom/geom_io.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/geom/geom_io.c b/sys/geom/geom_io.c
index 931b7c30338..1090c0bbf66 100644
--- a/sys/geom/geom_io.c
+++ b/sys/geom/geom_io.c
@@ -309,8 +309,8 @@ g_io_check(struct bio *bp)
 	case BIO_READ:
 	case BIO_WRITE:
 	case BIO_DELETE:
-		/* Zero sectorsize is a probably lack of media */
-		if (pp->sectorsize == 0)
+		/* Zero sectorsize or mediasize is probably a lack of media. */
+		if (pp->sectorsize == 0 || pp->mediasize == 0)
 			return (ENXIO);
 		/* Reject I/O not on sector boundary */
 		if (bp->bio_offset % pp->sectorsize)

From 94410af77428e042c74266d1a6d6a3e02a11acd5 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Thu, 15 Apr 2010 11:17:33 +0000
Subject: [PATCH 065/532] Explicitly enable PCI busmastering on attach. Now
 SiI3124 with siis(4) successfully works on sparc64 (SunBlade 100).

H/W donated by:	Gheorghe Ardelean
---
 sys/dev/siis/siis.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/dev/siis/siis.c b/sys/dev/siis/siis.c
index f03fba98027..c9c157226fa 100644
--- a/sys/dev/siis/siis.c
+++ b/sys/dev/siis/siis.c
@@ -164,6 +164,7 @@ siis_attach(device_t dev)
 		rman_fini(&ctlr->sc_iomem);
 		return (error);
 	}
+	pci_enable_busmaster(dev);
 	/* Reset controller */
 	siis_resume(dev);
 	/* Number of HW channels */

From 61f73308d437a7ce14a7c0d1f59dc6d1141a4912 Mon Sep 17 00:00:00 2001
From: VANHULLEBUS Yvan <vanhu@FreeBSD.org>
Date: Thu, 15 Apr 2010 12:40:33 +0000
Subject: [PATCH 066/532] Locks SPTREE when setting some SP entries to state
 DEAD. This can prevent kernel panics when updating SPs while there is some
 traffic for them.

Obtained from: NETASQ
MFC after: 1m
---
 sys/netipsec/key.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/sys/netipsec/key.c b/sys/netipsec/key.c
index e3a61aced00..2b17fb0d91b 100644
--- a/sys/netipsec/key.c
+++ b/sys/netipsec/key.c
@@ -1882,7 +1882,9 @@ key_spdadd(so, m, mhp)
 	newsp = key_getsp(&spidx);
 	if (mhp->msg->sadb_msg_type == SADB_X_SPDUPDATE) {
 		if (newsp) {
+			SPTREE_LOCK();
 			newsp->state = IPSEC_SPSTATE_DEAD;
+			SPTREE_UNLOCK();
 			KEY_FREESP(&newsp);
 		}
 	} else {
@@ -2117,7 +2119,9 @@ key_spddelete(so, m, mhp)
 	/* save policy id to buffer to be returned. */
 	xpl0->sadb_x_policy_id = sp->id;
 
+	SPTREE_LOCK();
 	sp->state = IPSEC_SPSTATE_DEAD;
+	SPTREE_UNLOCK();
 	KEY_FREESP(&sp);
 
     {
@@ -2184,7 +2188,9 @@ key_spddelete2(so, m, mhp)
 		return key_senderror(so, m, EINVAL);
 	}
 
+	SPTREE_LOCK();
 	sp->state = IPSEC_SPSTATE_DEAD;
+	SPTREE_UNLOCK();
 	KEY_FREESP(&sp);
 
     {

From c406ad2ecc2152da2888704513b0a2107de3be4a Mon Sep 17 00:00:00 2001
From: Andrew Gallatin <gallatin@FreeBSD.org>
Date: Thu, 15 Apr 2010 13:50:55 +0000
Subject: [PATCH 067/532] Cleanup if_media handling in mxge(4)

- Re-probe xfp / sfp+ socket on link events, in case user
    has changed transceiver
- correctly report current media to avoid confusing lagg (reported by Panasas)
- Report link speed  (submitted by yongari)

Reviewed by: yongari (earlier version)

MFC after: 7 days
---
 sys/dev/mxge/if_mxge.c     | 115 +++++++++++++++++++++++--------------
 sys/dev/mxge/if_mxge_var.h |   8 +++
 2 files changed, 79 insertions(+), 44 deletions(-)

diff --git a/sys/dev/mxge/if_mxge.c b/sys/dev/mxge/if_mxge.c
index d7f58255241..50c9af65794 100644
--- a/sys/dev/mxge/if_mxge.c
+++ b/sys/dev/mxge/if_mxge.c
@@ -883,6 +883,9 @@ mxge_send_cmd(mxge_softc_t *sc, uint32_t cmd, mxge_cmd_t *data)
 		case MXGEFW_CMD_ERROR_BUSY:
 			err = EBUSY;
 			break;
+		case MXGEFW_CMD_ERROR_I2C_ABSENT:
+			err = ENXIO;
+			break;
 		default:
 			device_printf(sc->dev, 
 				      "mxge: command %d "
@@ -2782,37 +2785,25 @@ static struct mxge_media_type mxge_sfp_media_types[] =
 };
 
 static void
-mxge_set_media(mxge_softc_t *sc, int type)
+mxge_media_set(mxge_softc_t *sc, int media_type)
 {
-	sc->media_flags |= type;
-	ifmedia_add(&sc->media, sc->media_flags, 0, NULL);
-	ifmedia_set(&sc->media, sc->media_flags);
+
+	
+	ifmedia_add(&sc->media, IFM_ETHER | IFM_FDX | media_type, 
+		    0, NULL);
+	ifmedia_set(&sc->media, IFM_ETHER | IFM_FDX | media_type);
+	sc->current_media = media_type;
+	sc->media.ifm_media = sc->media.ifm_cur->ifm_media;
 }
 
-
-/*
- * Determine the media type for a NIC.  Some XFPs will identify
- * themselves only when their link is up, so this is initiated via a
- * link up interrupt.  However, this can potentially take up to
- * several milliseconds, so it is run via the watchdog routine, rather
- * than in the interrupt handler itself.   This need only be done
- * once, not each time the link is up.
- */
 static void
-mxge_media_probe(mxge_softc_t *sc)
+mxge_media_init(mxge_softc_t *sc)
 {
-	mxge_cmd_t cmd;
-	char *cage_type;
 	char *ptr;
-	struct mxge_media_type *mxge_media_types = NULL;
-	int i, err, ms, mxge_media_type_entries;
-	uint32_t byte;
+	int i;
 
-	sc->need_media_probe = 0;
-
-	/* if we've already set a media type, we're done */
-	if (sc->media_flags  != (IFM_ETHER | IFM_AUTO))
-		return;
+	ifmedia_removeall(&sc->media);
+	mxge_media_set(sc, IFM_AUTO);
 
 	/* 
 	 * parse the product code to deterimine the interface type
@@ -2823,6 +2814,7 @@ mxge_media_probe(mxge_softc_t *sc)
 	ptr = sc->product_code_string;
 	if (ptr == NULL) {
 		device_printf(sc->dev, "Missing product code\n");
+		return;
 	}
 
 	for (i = 0; i < 3; i++, ptr++) {
@@ -2835,17 +2827,44 @@ mxge_media_probe(mxge_softc_t *sc)
 	}
 	if (*ptr == 'C') {
 		/* -C is CX4 */
-		mxge_set_media(sc, IFM_10G_CX4);
-		return;
-	}
-	else if (*ptr == 'Q') {
+		sc->connector = MXGE_CX4;
+		mxge_media_set(sc, IFM_10G_CX4);
+	} else if (*ptr == 'Q') {
 		/* -Q is Quad Ribbon Fiber */
+		sc->connector = MXGE_QRF;
 		device_printf(sc->dev, "Quad Ribbon Fiber Media\n");
 		/* FreeBSD has no media type for Quad ribbon fiber */
-		return;
+	} else if (*ptr == 'R') {
+		/* -R is XFP */
+		sc->connector = MXGE_XFP;
+	} else if (*ptr == 'S' || *(ptr +1) == 'S') {
+		/* -S or -2S is SFP+ */
+		sc->connector = MXGE_SFP;
+	} else {
+		device_printf(sc->dev, "Unknown media type: %c\n", *ptr);
 	}
+}
 
-	if (*ptr == 'R') {
+/*
+ * Determine the media type for a NIC.  Some XFPs will identify
+ * themselves only when their link is up, so this is initiated via a
+ * link up interrupt.  However, this can potentially take up to
+ * several milliseconds, so it is run via the watchdog routine, rather
+ * than in the interrupt handler itself. 
+ */
+static void
+mxge_media_probe(mxge_softc_t *sc)
+{
+	mxge_cmd_t cmd;
+	char *cage_type;
+
+	struct mxge_media_type *mxge_media_types = NULL;
+	int i, err, ms, mxge_media_type_entries;
+	uint32_t byte;
+
+	sc->need_media_probe = 0;
+
+	if (sc->connector == MXGE_XFP) {
 		/* -R is XFP */
 		mxge_media_types = mxge_xfp_media_types;
 		mxge_media_type_entries = 
@@ -2853,9 +2872,7 @@ mxge_media_probe(mxge_softc_t *sc)
 			sizeof (mxge_xfp_media_types[0]);
 		byte = MXGE_XFP_COMPLIANCE_BYTE;
 		cage_type = "XFP";
-	}
-
-	if (*ptr == 'S' || *(ptr +1) == 'S') {
+	} else 	if (sc->connector == MXGE_SFP) {
 		/* -S or -2S is SFP+ */
 		mxge_media_types = mxge_sfp_media_types;
 		mxge_media_type_entries = 
@@ -2863,10 +2880,8 @@ mxge_media_probe(mxge_softc_t *sc)
 			sizeof (mxge_sfp_media_types[0]);
 		cage_type = "SFP+";
 		byte = 3;
-	}
-
-	if (mxge_media_types == NULL) {
-		device_printf(sc->dev, "Unknown media type: %c\n", *ptr);
+	} else {
+		/* nothing to do; media type cannot change */
 		return;
 	}
 
@@ -2909,7 +2924,10 @@ mxge_media_probe(mxge_softc_t *sc)
 		if (mxge_verbose)
 			device_printf(sc->dev, "%s:%s\n", cage_type,
 				      mxge_media_types[0].name);
-		mxge_set_media(sc, mxge_media_types[0].flag);
+		if (sc->current_media != mxge_media_types[0].flag) {
+			mxge_media_init(sc);
+			mxge_media_set(sc, mxge_media_types[0].flag);
+		}
 		return;
 	}
 	for (i = 1; i < mxge_media_type_entries; i++) {
@@ -2919,12 +2937,16 @@ mxge_media_probe(mxge_softc_t *sc)
 					      cage_type,
 					      mxge_media_types[i].name);
 
-			mxge_set_media(sc, mxge_media_types[i].flag);
+			if (sc->current_media != mxge_media_types[i].flag) {
+				mxge_media_init(sc);
+				mxge_media_set(sc, mxge_media_types[i].flag);
+			}
 			return;
 		}
 	}
-	device_printf(sc->dev, "%s media 0x%x unknown\n", cage_type,
-		      cmd.data0);
+	if (mxge_verbose)
+		device_printf(sc->dev, "%s media 0x%x unknown\n",
+			      cage_type, cmd.data0);
 
 	return;
 }
@@ -2988,10 +3010,12 @@ mxge_intr(void *arg)
 			sc->link_state = stats->link_up;
 			if (sc->link_state) {
 				if_link_state_change(sc->ifp, LINK_STATE_UP);
+				 sc->ifp->if_baudrate = IF_Gbps(10UL);
 				if (mxge_verbose)
 					device_printf(sc->dev, "link up\n");
 			} else {
 				if_link_state_change(sc->ifp, LINK_STATE_DOWN);
+				sc->ifp->if_baudrate = 0;
 				if (mxge_verbose)
 					device_printf(sc->dev, "link down\n");
 			}
@@ -4026,9 +4050,9 @@ mxge_media_status(struct ifnet *ifp, struct ifmediareq *ifmr)
 	if (sc == NULL)
 		return;
 	ifmr->ifm_status = IFM_AVALID;
+	ifmr->ifm_active = IFM_ETHER | IFM_FDX;
 	ifmr->ifm_status |= sc->link_state ? IFM_ACTIVE : 0;
-	ifmr->ifm_active = IFM_AUTO | IFM_ETHER;
-	ifmr->ifm_active |= sc->link_state ? IFM_FDX : 0;
+	ifmr->ifm_active |= sc->current_media;
 }
 
 static int
@@ -4135,6 +4159,9 @@ mxge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 		break;
 
 	case SIOCGIFMEDIA:
+		mtx_lock(&sc->driver_mtx);
+		mxge_media_probe(sc);
+		mtx_unlock(&sc->driver_mtx);
 		err = ifmedia_ioctl(ifp, (struct ifreq *)data, 
 				    &sc->media, command);
                 break;
@@ -4766,7 +4793,7 @@ mxge_attach(device_t dev)
 	/* Initialise the ifmedia structure */
 	ifmedia_init(&sc->media, 0, mxge_media_change, 
 		     mxge_media_status);
-	mxge_set_media(sc, IFM_ETHER | IFM_AUTO);
+	mxge_media_init(sc);
 	mxge_media_probe(sc);
 	sc->dying = 0;
 	ether_ifattach(ifp, sc->mac_addr);
diff --git a/sys/dev/mxge/if_mxge_var.h b/sys/dev/mxge/if_mxge_var.h
index 5c1627f71e5..c85a29bfb58 100644
--- a/sys/dev/mxge/if_mxge_var.h
+++ b/sys/dev/mxge/if_mxge_var.h
@@ -268,6 +268,8 @@ struct mxge_softc {
 	int num_slices;
 	int rx_ring_size;
 	int dying;
+	int connector;
+	int current_media;
 	mxge_dma_t dmabench_dma;
 	struct callout co_hdl;
 	struct taskqueue *tq;
@@ -293,6 +295,12 @@ struct mxge_softc {
 #define MXGE_MIN_THROTTLE	416
 #define MXGE_MAX_THROTTLE	4096
 
+/* Types of connectors on NICs supported by this driver */
+#define MXGE_CX4 0
+#define MXGE_XFP 1
+#define MXGE_SFP 2
+#define MXGE_QRF 3
+
 #define MXGE_HIGHPART_TO_U32(X) \
 (sizeof (X) == 8) ? ((uint32_t)((uint64_t)(X) >> 32)) : (0)
 #define MXGE_LOWPART_TO_U32(X) ((uint32_t)(X))

From 2a29e09f38ba94e8dfa61195ef79d1c45ecc4a6a Mon Sep 17 00:00:00 2001
From: Andrew Gallatin <gallatin@FreeBSD.org>
Date: Thu, 15 Apr 2010 14:26:52 +0000
Subject: [PATCH 068/532] Add missing IFCAP_LINKSTATE to mxge

Submitted by: yongari
---
 sys/dev/mxge/if_mxge.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/mxge/if_mxge.c b/sys/dev/mxge/if_mxge.c
index 50c9af65794..956b1aa2dbb 100644
--- a/sys/dev/mxge/if_mxge.c
+++ b/sys/dev/mxge/if_mxge.c
@@ -4759,7 +4759,7 @@ mxge_attach(device_t dev)
 
 	ifp->if_baudrate = IF_Gbps(10UL);
 	ifp->if_capabilities = IFCAP_RXCSUM | IFCAP_TXCSUM | IFCAP_TSO4 |
-		IFCAP_VLAN_MTU;
+		IFCAP_VLAN_MTU | IFCAP_LINKSTATE;
 #ifdef INET
 	ifp->if_capabilities |= IFCAP_LRO;
 #endif

From e1a4e3fa37d5e5fd069f75f0c5842714d9b9728b Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Thu, 15 Apr 2010 15:10:46 +0000
Subject: [PATCH 069/532] Allow option aliasing.  Lines of the form:

OLD_OPT = NEW_OPT

in options* files will now map OLD_OPT to NEW_OPT with a friendly
message.  This is indented for situations where we need to preserve an
interface in the config file in an upwards compatible fashion on a
stable branch.

Reviewed by:	nwhitehorn@
MFC after:	3 days
---
 usr.sbin/config/config.h    |  2 ++
 usr.sbin/config/mkoptions.c | 27 ++++++++++++++++++++++++++-
 2 files changed, 28 insertions(+), 1 deletion(-)

diff --git a/usr.sbin/config/config.h b/usr.sbin/config/config.h
index d55c96b00a9..ec199862ae7 100644
--- a/usr.sbin/config/config.h
+++ b/usr.sbin/config/config.h
@@ -129,6 +129,8 @@ SLIST_HEAD(opt_head, opt) opt, mkopt, rmopts;
 struct opt_list {
 	char *o_name;
 	char *o_file;
+	int o_flags;
+#define OL_ALIAS	1
 	SLIST_ENTRY(opt_list) o_next;
 };
 
diff --git a/usr.sbin/config/mkoptions.c b/usr.sbin/config/mkoptions.c
index 5cd9d61a824..1a6ccc8fec0 100644
--- a/usr.sbin/config/mkoptions.c
+++ b/usr.sbin/config/mkoptions.c
@@ -94,6 +94,17 @@ options(void)
 	SLIST_INSERT_HEAD(&opt, op, op_next);
 
 	read_options();
+	SLIST_FOREACH(op, &opt, op_next) {
+		SLIST_FOREACH(ol, &otab, o_next) {
+			if (eq(op->op_name, ol->o_name) &&
+			    (ol->o_flags & OL_ALIAS)) {
+				printf("Mapping option %s to %s.\n",
+				    op->op_name, ol->o_file);
+				op->op_name = ol->o_file;
+				break;
+			}
+		}
+	}
 	SLIST_FOREACH(ol, &otab, o_next)
 		do_option(ol->o_name);
 	SLIST_FOREACH(op, &opt, op_next) {
@@ -124,7 +135,6 @@ do_option(char *name)
 	int tidy;
 
 	file = tooption(name);
-
 	/*
 	 * Check to see if the option was specified..
 	 */
@@ -292,6 +302,7 @@ read_options(void)
 	struct opt_list *po;
 	int first = 1;
 	char genopt[MAXPATHLEN];
+	int flags = 0;
 
 	SLIST_INIT(&otab);
 	(void) snprintf(fname, sizeof(fname), "../../conf/options");
@@ -301,6 +312,7 @@ openit:
 		return;
 	}
 next:
+	flags = 0;
 	wd = get_word(fp);
 	if (wd == (char *)EOF) {
 		(void) fclose(fp);
@@ -332,6 +344,18 @@ next:
 		(void) snprintf(genopt, sizeof(genopt), "opt_%s.h", lower(s));
 		val = genopt;
 		free(s);
+	} else if (eq(val, "=")) {
+		val = get_word(fp);
+		if (val == (char *)EOF) {
+			printf("%s: unexpected end of file\n", fname);
+			exit(1);
+		}
+		if (val == 0) {
+			printf("%s: Expected a right hand side at %s\n", fname,
+			    this);
+			exit(1);
+		}
+		flags |= OL_ALIAS;
 	}
 	val = ns(val);
 
@@ -348,6 +372,7 @@ next:
 		err(EXIT_FAILURE, "calloc");
 	po->o_name = this;
 	po->o_file = val;
+	po->o_flags = flags;
 	SLIST_INSERT_HEAD(&otab, po, o_next);
 
 	goto next;

From 31c4cef715a139f573d9fc3da1c93c63308f5426 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 15 Apr 2010 16:34:06 +0000
Subject: [PATCH 070/532] Use lower priority for GELI worker threads. This
 improves system responsiveness under heavy GELI load.

MFC after:	3 days
---
 sys/geom/eli/g_eli.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/sys/geom/eli/g_eli.c b/sys/geom/eli/g_eli.c
index 2a405c5b502..a3746c09e1a 100644
--- a/sys/geom/eli/g_eli.c
+++ b/sys/geom/eli/g_eli.c
@@ -340,7 +340,7 @@ g_eli_worker(void *arg)
 	}
 #endif
 	thread_lock(curthread);
-	sched_prio(curthread, PRIBIO);
+	sched_prio(curthread, PUSER);
 	if (sc->sc_crypto == G_ELI_CRYPTO_SW && g_eli_threads == 0)
 		sched_bind(curthread, wr->w_number);
 	thread_unlock(curthread);
@@ -361,8 +361,7 @@ g_eli_worker(void *arg)
 				mtx_unlock(&sc->sc_queue_mtx);
 				kproc_exit(0);
 			}
-			msleep(sc, &sc->sc_queue_mtx, PRIBIO | PDROP,
-			    "geli:w", 0);
+			msleep(sc, &sc->sc_queue_mtx, PDROP, "geli:w", 0);
 			continue;
 		}
 		mtx_unlock(&sc->sc_queue_mtx);

From 54ddff9dfa43eafbbb9b99bd6e9f612d6fea6908 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 15 Apr 2010 16:35:34 +0000
Subject: [PATCH 071/532] Flush disk write cache after storing and clearing
 metadata.

---
 sbin/geom/misc/subr.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sbin/geom/misc/subr.c b/sbin/geom/misc/subr.c
index 64df7c6845e..21deac8bf52 100644
--- a/sbin/geom/misc/subr.c
+++ b/sbin/geom/misc/subr.c
@@ -236,6 +236,7 @@ g_metadata_store(const char *name, u_char *md, size_t size)
 		error = errno;
 		goto out;
 	}
+	(void)ioctl(fd, DIOCGFLUSH, NULL);
 out:
 	if (sector != NULL)
 		free(sector);
@@ -293,6 +294,7 @@ g_metadata_clear(const char *name, const char *magic)
 		error = errno;
 		goto out;
 	}
+	(void)ioctl(fd, DIOCGFLUSH, NULL);
 out:
 	if (sector != NULL)
 		free(sector);

From 40c7da090ff2d263824c0089a508868452977c87 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 15 Apr 2010 16:40:54 +0000
Subject: [PATCH 072/532] Fix 3-way deadlock that can happen because of ZFS and
 vnode lock order reversal.

thread0 (vfs_fhtovp)	thread1 (vop_getattr)	thread2 (zfs_recv)
--------------------	---------------------	------------------
			vn_lock
rrw_enter_read
						rrw_enter_write (hangs)
			rrw_enter_read (hangs)
vn_lock (hangs)

Submitted by:	Attila Nagy <bra@fsn.hu>
MFC after:	3 days
---
 .../opensolaris/uts/common/fs/zfs/zfs_vfsops.c        | 11 +++++++----
 .../contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c |  6 +++---
 2 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c
index ad8165b3157..07c9b61bd32 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vfsops.c
@@ -868,13 +868,15 @@ zfs_root(vfs_t *vfsp, int flags, vnode_t **vpp)
 	ZFS_ENTER_NOERROR(zfsvfs);
 
 	error = zfs_zget(zfsvfs, zfsvfs->z_root, &rootzp);
+
+	ZFS_EXIT(zfsvfs);
+
 	if (error == 0) {
 		*vpp = ZTOV(rootzp);
 		error = vn_lock(*vpp, flags);
 		(*vpp)->v_vflag |= VV_ROOT;
 	}
 
-	ZFS_EXIT(zfsvfs);
 	return (error);
 }
 
@@ -1143,13 +1145,13 @@ zfs_vget(vfs_t *vfsp, ino_t ino, int flags, vnode_t **vpp)
 		VN_RELE(ZTOV(zp));
 		err = EINVAL;
 	}
+	ZFS_EXIT(zfsvfs);
 	if (err != 0)
 		*vpp = NULL;
 	else {
 		*vpp = ZTOV(zp);
 		vn_lock(*vpp, flags);
 	}
-	ZFS_EXIT(zfsvfs);
 	return (err);
 }
 
@@ -1237,8 +1239,8 @@ zfs_fhtovp(vfs_t *vfsp, fid_t *fidp, vnode_t **vpp)
 		} else {
 			VN_HOLD(*vpp);
 		}
-		vn_lock(*vpp, LK_EXCLUSIVE | LK_RETRY);
 		ZFS_EXIT(zfsvfs);
+		vn_lock(*vpp, LK_EXCLUSIVE | LK_RETRY);
 		return (0);
 	}
 
@@ -1259,10 +1261,11 @@ zfs_fhtovp(vfs_t *vfsp, fid_t *fidp, vnode_t **vpp)
 		return (EINVAL);
 	}
 
+	ZFS_EXIT(zfsvfs);
+
 	*vpp = ZTOV(zp);
 	vn_lock(*vpp, LK_EXCLUSIVE | LK_RETRY);
 	vnode_create_vobject(*vpp, zp->z_phys->zp_size, curthread);
-	ZFS_EXIT(zfsvfs);
 	return (0);
 }
 
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
index 4f61f5f3f5e..59a58dd7a5a 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
@@ -1209,15 +1209,17 @@ zfs_lookup(vnode_t *dvp, char *nm, vnode_t **vpp, struct componentname *cnp,
 			ltype = VOP_ISLOCKED(dvp);
 			VOP_UNLOCK(dvp, 0);
 		}
+		ZFS_EXIT(zfsvfs);
 		error = vn_lock(*vpp, cnp->cn_lkflags);
 		if (cnp->cn_flags & ISDOTDOT)
 			vn_lock(dvp, ltype | LK_RETRY);
 		if (error != 0) {
 			VN_RELE(*vpp);
 			*vpp = NULL;
-			ZFS_EXIT(zfsvfs);
 			return (error);
 		}
+	} else {
+		ZFS_EXIT(zfsvfs);
 	}
 
 #ifdef FREEBSD_NAMECACHE
@@ -1237,8 +1239,6 @@ zfs_lookup(vnode_t *dvp, char *nm, vnode_t **vpp, struct componentname *cnp,
 	}
 #endif
 
-	ZFS_EXIT(zfsvfs);
-
 	return (error);
 }
 

From ceab1828a000fa0f165997499325670c7ac8b320 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 15 Apr 2010 16:43:43 +0000
Subject: [PATCH 073/532] Use spaces instead of tab for indent here.

---
 etc/mtree/BSD.include.dist | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/etc/mtree/BSD.include.dist b/etc/mtree/BSD.include.dist
index 78240112c9b..1ec71719ac4 100644
--- a/etc/mtree/BSD.include.dist
+++ b/etc/mtree/BSD.include.dist
@@ -104,8 +104,8 @@
         ..
         lmc
         ..
-	mfi
-	..
+        mfi
+        ..
         mpt
             mpilib
             ..

From 20b77db949cfb020f933039e76bbf06667b93c71 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 15 Apr 2010 17:04:08 +0000
Subject: [PATCH 074/532] Increase ggate queue size to maximum value. HAST was
 not able to stand heavy random load.

Reported by:	Hiroyuki Yamagami
MFC after:	3 days
---
 sbin/hastd/primary.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sbin/hastd/primary.c b/sbin/hastd/primary.c
index 20af0b983c1..09151546772 100644
--- a/sbin/hastd/primary.c
+++ b/sbin/hastd/primary.c
@@ -682,7 +682,7 @@ init_ggate(struct hast_resource *res)
 	ggiocreate.gctl_mediasize = res->hr_datasize;
 	ggiocreate.gctl_sectorsize = res->hr_local_sectorsize;
 	ggiocreate.gctl_flags = 0;
-	ggiocreate.gctl_maxcount = 128;
+	ggiocreate.gctl_maxcount = G_GATE_MAX_QUEUE_SIZE;
 	ggiocreate.gctl_timeout = 0;
 	ggiocreate.gctl_unit = G_GATE_NAME_GIVEN;
 	snprintf(ggiocreate.gctl_name, sizeof(ggiocreate.gctl_name), "hast/%s",

From 3e22320c434f2f2c6e442c4d8c5cb11a9810f9e2 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 15 Apr 2010 17:17:02 +0000
Subject: [PATCH 075/532] Fix typo.

MFC after:	3 days
---
 sys/kern/vfs_cache.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/vfs_cache.c b/sys/kern/vfs_cache.c
index 78548033ea2..533ec97875d 100644
--- a/sys/kern/vfs_cache.c
+++ b/sys/kern/vfs_cache.c
@@ -610,7 +610,7 @@ cache_enter(dvp, vp, cnp)
 
 	CTR3(KTR_VFS, "cache_enter(%p, %p, %s)", dvp, vp, cnp->cn_nameptr);
 	VNASSERT(vp == NULL || (vp->v_iflag & VI_DOOMED) == 0, vp,
-	    ("cahe_enter: Adding a doomed vnode"));
+	    ("cache_enter: Adding a doomed vnode"));
 
 	if (!doingcache)
 		return;

From c6491946d820c02b41e8438e82a6a4089b21075a Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 15 Apr 2010 17:24:21 +0000
Subject: [PATCH 076/532] Fix include path.

---
 sys/dev/sge/if_sge.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index a26a6a46359..c4fb5c9bc8b 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -81,7 +81,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/pci/pcireg.h>
 #include <dev/pci/pcivar.h>
 
-#include "if_sgereg.h"
+#include <dev/sge/if_sgereg.h>
 
 MODULE_DEPEND(sge, pci, 1, 1, 1);
 MODULE_DEPEND(sge, ether, 1, 1, 1);

From 55e999851d94e1b109c0f83cd1d84911775c17e2 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 15 Apr 2010 19:15:05 +0000
Subject: [PATCH 077/532] Diff reduction with OpenBSD:

 - Remove unused locally added variable;
 - Deprecate -o: it's the default behavior on OpenBSD.

X-MFC:	along with nc 4.7.
---
 contrib/netcat/nc.1     | 13 ++-----------
 contrib/netcat/netcat.c | 18 ++++++------------
 2 files changed, 8 insertions(+), 23 deletions(-)

diff --git a/contrib/netcat/nc.1 b/contrib/netcat/nc.1
index fc2c8af3857..ac6aa475fbd 100644
--- a/contrib/netcat/nc.1
+++ b/contrib/netcat/nc.1
@@ -27,7 +27,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd February 23, 2010
+.Dd April 15, 2010
 .Dt NC 1
 .Os
 .Sh NAME
@@ -36,7 +36,7 @@
 .Sh SYNOPSIS
 .Nm nc
 .Bk -words
-.Op Fl 46DdEhklnorStUuvz
+.Op Fl 46DdEhklnrStUuvz
 .Op Fl e Ar IPsec_policy
 .Op Fl I Ar length
 .Op Fl i Ar interval
@@ -159,15 +159,6 @@ socket option.
 .It Fl O Ar length
 Specifies the size of the TCP send buffer.
 When
-.It Fl o
-.Dq Once-only mode .
-By default,
-.Nm
-does not terminate on EOF condition on input,
-but continues until the network side has been closed down.
-Specifying
-.Fl o
-will make it terminate on EOF as well.
 .It Fl P Ar proxy_username
 Specifies a username to present to a proxy server that requires authentication.
 If no username is specified then authentication will not be attempted.
diff --git a/contrib/netcat/netcat.c b/contrib/netcat/netcat.c
index c83f62293d3..8faaeb06851 100644
--- a/contrib/netcat/netcat.c
+++ b/contrib/netcat/netcat.c
@@ -72,14 +72,12 @@
 #define PORT_MAX_LEN	6
 
 /* Command Line Options */
-int	Eflag;					/* Use IPsec ESP */
 int	dflag;					/* detached, no stdin */
 unsigned int iflag;				/* Interval Flag */
 int	jflag;					/* use jumbo frames if we can */
 int	kflag;					/* More than one connect */
 int	lflag;					/* Bind to local port */
 int	nflag;					/* Don't do name look up */
-int	oflag;					/* Once only: stop on EOF */
 int	FreeBSD_Oflag;				/* Do not use TCP options */
 char   *Pflag;					/* Proxy username */
 char   *pflag;					/* Localport flag */
@@ -151,7 +149,7 @@ main(int argc, char *argv[])
 	sv = NULL;
 
 	while ((ch = getopt_long(argc, argv,
-	    "46DdEe:hI:i:jklnO:oP:p:rSs:tT:UuV:vw:X:x:z",
+	    "46DdEe:hI:i:jklnoO:P:p:rSs:tT:UuV:vw:X:x:z",
 	    longopts, NULL)) != -1) {
 		switch (ch) {
 		case '4':
@@ -214,7 +212,7 @@ main(int argc, char *argv[])
 			nflag = 1;
 			break;
 		case 'o':
-			oflag = 1;
+			fprintf(stderr, "option -o is deprecated.\n");
 			break;
 		case 'P':
 			Pflag = optarg;
@@ -282,8 +280,6 @@ main(int argc, char *argv[])
 		case 'T':
 			Tflag = parse_iptos(optarg);
 			break;
-		case 0:
-			break;
 		default:
 			usage(1);
 		}
@@ -729,10 +725,9 @@ readwrite(int nfd)
 		}
 
 		if (!dflag && pfd[1].revents & POLLIN) {
-			if ((n = read(wfd, buf, plen)) < 0 ||
-			    (oflag && n == 0)) {
+			if ((n = read(wfd, buf, plen)) < 0)
 				return;
-			} else if (n == 0) {
+			else if (n == 0) {
 				shutdown(nfd, SHUT_WR);
 				pfd[1].fd = -1;
 				pfd[1].events = 0;
@@ -945,7 +940,6 @@ help(void)
 	\t-n		Suppress name/port resolutions\n\
 	\t--no-tcpopt	Disable TCP options\n\
 	\t-O length	TCP send buffer length\n\
-	\t-o		Terminate on EOF on input\n\
 	\t-P proxyuser\tUsername for proxy authentication\n\
 	\t-p port\t	Specify local port for remote connects\n\
 	\t-r		Randomize remote ports\n\
@@ -995,9 +989,9 @@ usage(int ret)
 {
 	fprintf(stderr,
 #ifdef IPSEC
-	    "usage: nc [-46DdEhklnorStUuvz] [-e policy] [-I length] [-i interval] [-O length]\n"
+	    "usage: nc [-46DdEhklnrStUuvz] [-e policy] [-I length] [-i interval] [-O length]\n"
 #else
-	    "usage: nc [-46DdhklnorStUuvz] [-I length] [-i interval] [-O length]\n"
+	    "usage: nc [-46DdhklnrStUuvz] [-I length] [-i interval] [-O length]\n"
 #endif
 	    "\t  [-P proxy_username] [-p source_port] [-s source_ip_address] [-T ToS]\n"
 	    "\t  [-V fib] [-w timeout] [-X proxy_protocol]\n"

From 5d0848e9c67a570760d6420d9a3f6df99ec20d6d Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Thu, 15 Apr 2010 19:45:03 +0000
Subject: [PATCH 078/532] - Fix a typo OFFCORE_REQUESTS.ANY.RFO is B0H10H and
 not 80H10H. - Enable missing PARTIAL_ADDRESS_ALIAS for Core i7.

MFC after: 3 days
---
 sys/dev/hwpmc/hwpmc_core.c | 3 ++-
 sys/dev/hwpmc/pmc_events.h | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/sys/dev/hwpmc/hwpmc_core.c b/sys/dev/hwpmc/hwpmc_core.c
index 90d7c8bc23f..e7de09915b3 100644
--- a/sys/dev/hwpmc/hwpmc_core.c
+++ b/sys/dev/hwpmc/hwpmc_core.c
@@ -603,7 +603,7 @@ static struct iap_event_descr iap_events[] = {
     IAPDESCR(06H_0FH, 0x06, 0x0F, IAP_F_FM | IAP_F_I7O),
 
     IAPDESCR(07H_00H, 0x07, 0x00, IAP_F_FM | IAP_F_CC | IAP_F_CC2),
-    IAPDESCR(07H_01H, 0x07, 0x01, IAP_F_FM | IAP_F_ALLCPUSCORE2 | IAP_F_WM),
+    IAPDESCR(07H_01H, 0x07, 0x01, IAP_F_FM | IAP_F_ALLCPUSCORE2 | IAP_F_I7 | IAP_F_WM),
     IAPDESCR(07H_02H, 0x07, 0x02, IAP_F_FM | IAP_F_ALLCPUSCORE2),
     IAPDESCR(07H_03H, 0x07, 0x03, IAP_F_FM | IAP_F_ALLCPUSCORE2),
     IAPDESCR(07H_06H, 0x07, 0x06, IAP_F_FM | IAP_F_CA),
@@ -1053,6 +1053,7 @@ static struct iap_event_descr iap_events[] = {
     IAPDESCR(B0H_02H, 0xB0, 0x02, IAP_F_FM | IAP_F_WM | IAP_F_I7O),
     IAPDESCR(B0H_04H, 0xB0, 0x04, IAP_F_FM | IAP_F_WM | IAP_F_I7O),
     IAPDESCR(B0H_08H, 0xB0, 0x08, IAP_F_FM | IAP_F_WM | IAP_F_I7O),
+    IAPDESCR(B0H_10H, 0xB0, 0x10, IAP_F_FM | IAP_F_WM | IAP_F_I7O),
     IAPDESCR(B0H_20H, 0xB0, 0x20, IAP_F_FM | IAP_F_I7O),
     IAPDESCR(B0H_40H, 0xB0, 0x40, IAP_F_FM | IAP_F_I7 | IAP_F_WM),
     IAPDESCR(B0H_80H, 0xB0, 0x80, IAP_F_FM | IAP_F_CA | IAP_F_WM | IAP_F_I7O),
diff --git a/sys/dev/hwpmc/pmc_events.h b/sys/dev/hwpmc/pmc_events.h
index e38772cd95d..df2d3efd7bf 100644
--- a/sys/dev/hwpmc/pmc_events.h
+++ b/sys/dev/hwpmc/pmc_events.h
@@ -2207,7 +2207,7 @@ __PMC_EV_ALIAS("OFFCORE_REQUESTS.DEMAND.READ_DATA", IAP_EVENT_B0H_01H)	\
 __PMC_EV_ALIAS("OFFCORE_REQUESTS.DEMAND.READ_CODE", IAP_EVENT_B0H_02H)	\
 __PMC_EV_ALIAS("OFFCORE_REQUESTS.DEMAND.RFO", IAP_EVENT_B0H_04H)	\
 __PMC_EV_ALIAS("OFFCORE_REQUESTS.ANY.READ", IAP_EVENT_B0H_08H)		\
-__PMC_EV_ALIAS("OFFCORE_REQUESTS.ANY.RFO", IAP_EVENT_80H_10H)		\
+__PMC_EV_ALIAS("OFFCORE_REQUESTS.ANY.RFO", IAP_EVENT_B0H_10H)		\
 __PMC_EV_ALIAS("OFFCORE_REQUESTS.L1D_WRITEBACK", IAP_EVENT_B0H_40H)	\
 __PMC_EV_ALIAS("OFFCORE_REQUESTS.ANY", IAP_EVENT_B0H_80H)		\
 __PMC_EV_ALIAS("UOPS_EXECUTED.PORT0", IAP_EVENT_B1H_01H)		\

From d52bba9301b45a85101914647396c9a92e26aad8 Mon Sep 17 00:00:00 2001
From: Doug Barton <dougb@FreeBSD.org>
Date: Thu, 15 Apr 2010 21:18:24 +0000
Subject: [PATCH 079/532] If a service is running, make 'stop' work even if
 ${name}_enable is not set.

PR:		conf/130414
Submitted by:	Dominic Fandrey <kamikaze@bsdforen.de>
Reviewed by:	freebsd-rc@
---
 etc/rc.subr | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/etc/rc.subr b/etc/rc.subr
index 18273be83ab..397af1d7907 100644
--- a/etc/rc.subr
+++ b/etc/rc.subr
@@ -646,12 +646,12 @@ run_rc_command()
 		if [ "$_elem" != "$rc_arg" ]; then
 			continue
 		fi
-					# if ${rcvar} is set, and $1 is not
-					# "rcvar", then run
+					# if ${rcvar} is set, $1 is not "rcvar"
+					# and ${rc_pid} is not set, then run
 					#	checkyesno ${rcvar}
 					# and return if that failed
 					#
-		if [ -n "${rcvar}" -a "$rc_arg" != "rcvar" ]; then
+		if [ -n "${rcvar}" -a "$rc_arg" != "rcvar" -a -z "${rc_pid}" ]; then
 			if ! checkyesno ${rcvar}; then
 				if [ -n "${rc_quiet}" ]; then
 					return 0

From c27b391b3f5fece33f763f040e229ed603e560bb Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 15 Apr 2010 21:41:07 +0000
Subject: [PATCH 080/532] indent(1): don't treat bare '_t' as a type name when
 -ta is used

It seems that identifier "_t" is sometimes used as a variable name,
even in our tree.  Not that I endorse that, but still it's better
to require at least one character before _t suffix to consider
an identifier to be a type name.

Reported by:	Alex Vasylenko <lxv@omut.org>
MFC after:	1 week
---
 usr.bin/indent/lexi.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/usr.bin/indent/lexi.c b/usr.bin/indent/lexi.c
index 3d415d074dc..b3604c64205 100644
--- a/usr.bin/indent/lexi.c
+++ b/usr.bin/indent/lexi.c
@@ -251,9 +251,10 @@ lexi(void)
 
 	if (auto_typedefs) {
 	    const char *q = s_token;
+	    size_t q_len = strlen(q);
 	    /* Check if we have an "_t" in the end */
-	    if (q[0] && q[1] &&
-	        (strcmp(q + strlen(q) - 2, "_t") == 0)) {
+	    if (q_len > 2 &&
+	        (strcmp(q + q_len - 2, "_t") == 0)) {
 	        ps.its_a_keyword = true;
 		ps.last_u_d = true;
 	        goto found_auto_typedef;

From 55909abf07e4069f4c4e176c320dbec548dd60ab Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Thu, 15 Apr 2010 22:57:30 +0000
Subject: [PATCH 081/532] The experimental NFS client was not filling in
 recovery credentials for opens done locally in the client when a delegation
 for the file was held. This could cause the client to crash in crsetgroups()
 when recovering from a server crash/reboot. This patch fills in the recovery
 credentials for this case, in order to avoid the client crash. Also, add
 KASSERT()s to the credential copy functions, to catch any other cases where
 the credentials aren't filled in correctly.

MFC after:	1 week
---
 sys/fs/nfs/nfs_commonport.c     | 2 ++
 sys/fs/nfs/nfsclstate.h         | 1 +
 sys/fs/nfsclient/nfs_clport.c   | 2 ++
 sys/fs/nfsclient/nfs_clrpcops.c | 8 +++++++-
 sys/fs/nfsclient/nfs_clstate.c  | 9 +++++++--
 5 files changed, 19 insertions(+), 3 deletions(-)

diff --git a/sys/fs/nfs/nfs_commonport.c b/sys/fs/nfs/nfs_commonport.c
index a65ebde881c..2849163837d 100644
--- a/sys/fs/nfs/nfs_commonport.c
+++ b/sys/fs/nfs/nfs_commonport.c
@@ -225,6 +225,8 @@ void
 newnfs_copycred(struct nfscred *nfscr, struct ucred *cr)
 {
 
+	KASSERT(nfscr->nfsc_ngroups >= 0,
+	    ("newnfs_copycred: negative nfsc_ngroups"));
 	cr->cr_uid = nfscr->nfsc_uid;
 	crsetgroups(cr, nfscr->nfsc_ngroups, nfscr->nfsc_groups);
 }
diff --git a/sys/fs/nfs/nfsclstate.h b/sys/fs/nfs/nfsclstate.h
index 10747af5a64..edd479cf2ce 100644
--- a/sys/fs/nfs/nfsclstate.h
+++ b/sys/fs/nfs/nfsclstate.h
@@ -140,6 +140,7 @@ struct nfsclopen {
 #define	NFSCLOPEN_OK		0
 #define	NFSCLOPEN_DOOPEN	1
 #define	NFSCLOPEN_DOOPENDOWNGRADE 2
+#define	NFSCLOPEN_SETCRED	3
 
 struct nfscllockowner {
 	LIST_ENTRY(nfscllockowner) nfsl_list;
diff --git a/sys/fs/nfsclient/nfs_clport.c b/sys/fs/nfsclient/nfs_clport.c
index e81c3bf805b..f39666db68e 100644
--- a/sys/fs/nfsclient/nfs_clport.c
+++ b/sys/fs/nfsclient/nfs_clport.c
@@ -978,6 +978,8 @@ newnfs_copyincred(struct ucred *cr, struct nfscred *nfscr)
 {
 	int i;
 
+	KASSERT(cr->cr_ngroups >= 0,
+	    ("newnfs_copyincred: negative cr_ngroups"));
 	nfscr->nfsc_uid = cr->cr_uid;
 	nfscr->nfsc_ngroups = MIN(cr->cr_ngroups, NFS_MAXGRPS + 1);
 	for (i = 0; i < nfscr->nfsc_ngroups; i++)
diff --git a/sys/fs/nfsclient/nfs_clrpcops.c b/sys/fs/nfsclient/nfs_clrpcops.c
index 95943a9660b..a7015f5226b 100644
--- a/sys/fs/nfsclient/nfs_clrpcops.c
+++ b/sys/fs/nfsclient/nfs_clrpcops.c
@@ -278,7 +278,13 @@ else printf(" fhl=0\n");
 			error = EIO;
 		}
 		newnfs_copyincred(cred, &op->nfso_cred);
-	    }
+	    } else if (ret == NFSCLOPEN_SETCRED)
+		/*
+		 * This is a new local open on a delegation. It needs
+		 * to have credentials so that an open can be done
+		 * against the server during recovery.
+		 */
+		newnfs_copyincred(cred, &op->nfso_cred);
 
 	    /*
 	     * nfso_opencnt is the count of how many VOP_OPEN()s have
diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index 568c5de640a..08150a7228e 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -274,8 +274,13 @@ nfscl_open(vnode_t vp, u_int8_t *nfhp, int fhlen, u_int32_t amode, int usedeleg,
 		*owpp = owp;
 	if (opp != NULL)
 		*opp = op;
-	if (retp != NULL)
-		*retp = NFSCLOPEN_OK;
+	if (retp != NULL) {
+		if (nfhp != NULL && dp != NULL && nop == NULL)
+			/* new local open on delegation */
+			*retp = NFSCLOPEN_SETCRED;
+		else
+			*retp = NFSCLOPEN_OK;
+	}
 
 	/*
 	 * Now, check the mode on the open and return the appropriate

From 6856a5e345c29381f081beef5529bbd35f769b86 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 15 Apr 2010 23:21:24 +0000
Subject: [PATCH 082/532] Utilize IP_BINDANY which provided the same semantics
 of OpenBSD's SO_BINDANY.

---
 contrib/netcat/netcat.c | 6 ++----
 1 file changed, 2 insertions(+), 4 deletions(-)

diff --git a/contrib/netcat/netcat.c b/contrib/netcat/netcat.c
index 8faaeb06851..28a98238676 100644
--- a/contrib/netcat/netcat.c
+++ b/contrib/netcat/netcat.c
@@ -570,10 +570,8 @@ remote_connect(const char *host, const char *port, struct addrinfo hints)
 		if (sflag || pflag) {
 			struct addrinfo ahints, *ares;
 
-#ifdef SO_BINDANY
-			/* try SO_BINDANY, but don't insist */
-			setsockopt(s, SOL_SOCKET, SO_BINDANY, &on, sizeof(on));
-#endif
+			/* try IP_BINDANY, but don't insist */
+			setsockopt(s, IPPROTO_IP, IP_BINDANY, &on, sizeof(on));
 			memset(&ahints, 0, sizeof(struct addrinfo));
 			ahints.ai_family = res0->ai_family;
 			ahints.ai_socktype = uflag ? SOCK_DGRAM : SOCK_STREAM;

From 0ac68bd339575e74d26a008fe6cf913e38651b9e Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Thu, 15 Apr 2010 23:56:05 +0000
Subject: [PATCH 083/532] Add mutex lock calls to 2 cases in the experimental
 NFS client's renew thread where they were missing.

MFC after:	1 week
---
 sys/fs/nfsclient/nfs_clstate.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index 08150a7228e..7bccbfcc5d7 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -2318,7 +2318,9 @@ nfscl_renewthread(struct nfsclclient *clp, NFSPROC_T *p)
 	int error, cbpathdown, islept, igotlock, ret, clearok;
 
 	cred = newnfs_getcred();
+	NFSLOCKCLSTATE();
 	clp->nfsc_flags |= NFSCLFLAGS_HASTHREAD;
+	NFSUNLOCKCLSTATE();
 	for(;;) {
 		newnfs_setroot(cred);
 		cbpathdown = 0;
@@ -2331,9 +2333,11 @@ nfscl_renewthread(struct nfsclclient *clp, NFSPROC_T *p)
 			error = nfsrpc_renew(clp, cred, p);
 			if (error == NFSERR_CBPATHDOWN)
 			    cbpathdown = 1;
-			else if (error == NFSERR_STALECLIENTID)
+			else if (error == NFSERR_STALECLIENTID) {
+			    NFSLOCKCLSTATE();
 			    clp->nfsc_flags |= NFSCLFLAGS_RECOVER;
-			else if (error == NFSERR_EXPIRED)
+			    NFSUNLOCKCLSTATE();
+			} else if (error == NFSERR_EXPIRED)
 			    (void) nfscl_hasexpired(clp, clidrev, p);
 		}
 

From 4076170459d487877d3c99cdabb26b861bc6033b Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 16 Apr 2010 02:56:24 +0000
Subject: [PATCH 084/532] Remove some unused header files.

---
 ObsoleteFiles.inc           |   8 ++
 sys/mips/include/archtype.h |  49 -------
 sys/mips/include/defs.h     | 256 ------------------------------------
 sys/mips/include/queue.h    | 171 ------------------------
 sys/mips/include/rm7000.h   |  95 -------------
 sys/mips/include/segments.h |  40 ------
 sys/mips/mips/trap.c        |   2 -
 7 files changed, 8 insertions(+), 613 deletions(-)
 delete mode 100644 sys/mips/include/archtype.h
 delete mode 100644 sys/mips/include/defs.h
 delete mode 100644 sys/mips/include/queue.h
 delete mode 100644 sys/mips/include/rm7000.h
 delete mode 100644 sys/mips/include/segments.h

diff --git a/ObsoleteFiles.inc b/ObsoleteFiles.inc
index b18c0ea52d3..0844d714693 100644
--- a/ObsoleteFiles.inc
+++ b/ObsoleteFiles.inc
@@ -14,6 +14,14 @@
 # The file is partitioned: OLD_FILES first, then OLD_LIBS and OLD_DIRS last.
 #
 
+# 20100415: [mips] removed unused headers
+.if ${TARGET_ARCH} == "mips"
+OLD_FILES+=usr/include/machine/archtype.h
+OLD_FILES+=usr/include/machine/segments.h
+OLD_FILES+=usr/include/machine/rm7000.h
+OLD_FILES+=usr/include/machine/defs.h
+OLD_FILES+=usr/include/machine/queue.h
+.endif
 # 20100326: [ia64] removed <machine/nexusvar.h>
 .if ${TARGET_ARCH} == "ia64"
 OLD_FILES+=usr/include/machine/nexusvar.h
diff --git a/sys/mips/include/archtype.h b/sys/mips/include/archtype.h
deleted file mode 100644
index ed1b5ea6b8f..00000000000
--- a/sys/mips/include/archtype.h
+++ /dev/null
@@ -1,49 +0,0 @@
-/*	$OpenBSD: archtype.h,v 1.6 1999/01/27 04:46:04 imp Exp $	*/
-/*
- * Copyright (c) 1997 Per Fogelstrom
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *	This product includes software developed under OpenBSD by
- *	Per Fogelstrom.
- * 4. The name of the author may not be used to endorse or promote products
- *    derived from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS
- * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
- * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *	JNPR: archtype.h,v 1.6 2007/08/09 11:23:32 katta
- *
- * $FreeBSD$
- */
-
-#ifndef _MACHINE_ARCHTYPE_H_
-#define	_MACHINE_ARCHTYPE_H_
-/*
- * Define architectural identitys for the different Mips machines.
- */
-
-/*
- * FREEBSD_DEVELOPERS_FIXME
- * Define constants for the supported MIPS CPU's
- */
-#define	MIPS_CLASS_UNKNOWN	0x00
-
-#endif /* !_MACHINE_ARCHTYPE_H_ */
diff --git a/sys/mips/include/defs.h b/sys/mips/include/defs.h
deleted file mode 100644
index 20d093e029b..00000000000
--- a/sys/mips/include/defs.h
+++ /dev/null
@@ -1,256 +0,0 @@
-/*
- * Copyright (c) 1996, 2001-2003, 2005, Juniper Networks, Inc.
- * All rights reserved.
- *
- * defs.h -- Simple universal types and definitions for use by the microkernel
- * Jim Hayes, November 1996
- *
- *	JNPR: defs.h,v 1.3.2.1 2007/09/10 08:16:32 girish
- * $FreeBSD$
- */
-
-#ifndef __DEFS_H__
-#define	__DEFS_H__
-
-/*
- * Paranoid compilation. If defined, the PARANOID flag will enable asserts,
- * data structure magic stamping and a suite of other debug tools. To disable
- * it, comment out its definition.
- */
-#define	PARANOID
-
-/*
- * This is the ONLY place you should see hardware specific information
- * encoded as #ifdefs. (Well, except for stdarg.h, perhaps.)
- * I apologize in advance!
- */
-#include <machine/defs_mips.h>
-#define	CPU_GOT_ONE
-
-#if !defined(CPU_GOT_ONE)
-#error "YOU NEED TO SPECIFY ONE CPU TYPE TO USE THIS FILE"
-#endif
-
-#ifdef	TRUE
-#undef	TRUE
-#endif
-
-#ifdef	FALSE
-#undef	FALSE
-#endif
-
-typedef enum boolean_
-{
-	FALSE = 0,
-	TRUE = 1
-} boolean;
-
-/*
- * Make NULL a pointer within the microkernel environment to catch
- * pointer semantic miscreants.
- *
- * The reason it's conditional here is that some of the BSD includes
- * define it multiple times as a straight integer and GCC barfs on
- * the alternative prototypes.
- */
-
-#ifndef NULL
-#define	NULL (void *)0
-#endif
-
-/*
- * Define some standard sized types.  (Defined in cpu-specific type files
- * included above.)
- */
-
-#define	MAX_U8  255
-#define	MAX_S8  128
-#define	MIN_S8 -127
-
-#define	MAX_U16  0xffff
-#define	MIN_S16  ((int16_t)(1 << 15))
-#define	MAX_S16  ((int16_t)~MIN_S16)
-
-#define	MAX_U32  0xffffffff
-#define	MIN_S32  ((int32_t)(1 << 31))
-#define	MAX_S32  ((int32_t)~MIN_S32)
-
-#define	MAX_U64  ((u_int64_t)0 - 1)
-#define	MAX_S64  ((int64_t)(MAX_U64 >> 1))
-#define	MIN_S64  (-MAX_S64-1)
-
-/*
- * Solaris uses _SIZE_T to mark the fact that "size_t" has already
- * been defined.  _SYS_TYPES_H_ is used by BSD.
- * 
- */
-#if !defined(_SYS_TYPES_H_) && !defined(_SIZE_T)
-typedef UNSIGNED_32  size_t;
-#define	_SIZE_T
-#endif
-
-#if !defined(_SYS_TYPES_H_)
-typedef char *		caddr_t;
-
-typedef UNSIGNED_8	u_int8_t;
-typedef SIGNED_8	int8_t;
-
-typedef UNSIGNED_16	u_int16_t;
-typedef SIGNED_16	int16_t;
-
-typedef UNSIGNED_32	u_int32_t;
-typedef SIGNED_32	int32_t;
-
-typedef UNSIGNED_64	u_int64_t;
-typedef SIGNED_64	int64_t;
-
-typedef UNSIGNED_32	u_long;
-typedef UNSIGNED_16	u_short;
-typedef UNSIGNED_8	u_char;
-
-
-/*
- * Define the standard terminology used in the diag software
- * with regards to bytes, words, etc.
- * BYTE = 8 bits
- * HWORD (halfword) = 2 bytes or 16 bits
- * WORD = 4 bytes or 32 bits
- * QUAD = 8 bytes or 64 bits
- *
- * (The term QUAD seems less-than-intuitive here, but it is
- * derived from BSD sources where it is defined as int64_t.)
- *
- * For consistency use the following defines wherever appropriate.
- */
-
-typedef enum {
-	NBI_BYTE  = (sizeof(u_int8_t) * 8),
-	NBI_HWORD = (sizeof(u_int16_t) * 8),
-	NBI_WORD  = (sizeof(u_int32_t) * 8),
-	NBI_QUAD  = (sizeof(u_int64_t) * 8)
-} num_bits_t;
-
-typedef enum {
-	NBY_BYTE  = sizeof(u_int8_t),
-	NBY_HWORD = sizeof(u_int16_t),
-	NBY_WORD  = sizeof(u_int32_t),
-	NBY_QUAD  = sizeof(u_int64_t)
-} num_bytes_t;
-
-/*
- * We assume that pid values are 16 bit integers
- */
-
-typedef u_int16_t pid_t;
-
-#endif /* _SYS_TYPES_H_ */
-
-typedef UNSIGNED_32	magic_t;
-typedef int		status_t;
-
-#define	BITS_IN_BYTE	8
-
-/*
- * Packed definition. We use this for fields in network frames where we
- * don't want the compiler to pack out to even alignment
- */
-
-#ifdef PACKED
-#undef PACKED
-#endif
-#define	PACKED(x)	x __attribute__ ((packed))
-
-/*
- * __unused is a FreeBSDism that prevents the compiler from choking
- * on function parameters that remain unused through the life of a
- * function.  This is not an issue for the Cygnus toolchain.  In general
- * it SHOULD NOT BE USED in the martini embedded software repository.
- * It should only be used inside of shared code.
- */
-#ifndef __unused
-#define	__unused	__attribute__ ((__unused__))
-#endif
-
-/*
- * Basic memory multiples
- */
-
-#define	SIZE_1K		0x00000400
-#define	SIZE_2K		0x00000800
-#define	SIZE_4K		0x00001000
-#define	SIZE_8K		0x00002000
-#define	SIZE_16K	0x00004000
-#define	SIZE_32K	0x00008000
-#define	SIZE_64K	0x00010000
-#define	SIZE_128K	0x00020000
-#define	SIZE_256K	0x00040000
-#define	SIZE_512K	0x00080000
-#define	SIZE_1M		0x00100000
-#define	SIZE_2M		0x00200000
-#define	SIZE_4M		0x00400000
-#define	SIZE_8M		0x00800000
-#define	SIZE_16M	0x01000000
-#define	SIZE_32M	0x02000000
-#define	SIZE_64M	0x04000000
-#define	SIZE_128M	0x08000000
-#define	SIZE_256M	0x10000000
-#define	SIZE_512M	0x20000000
-#define	SIZE_1G		0x40000000
-#define	SIZE_2G		0x80000000
-
-/*
- * swap16_inline
- * swap32_inline
- *
- * Byteswap a 16 and 32 bit quantities
- */
-
-static inline u_int16_t
-swap16_inline(u_int16_t data)
-{
-	return(((data & 0x00ff) << 8) |
-	       ((data & 0xff00) >> 8));
-}
-
-static inline u_int32_t
-swap32_inline(u_int32_t data)
-{
-	return(((data & 0x000000ff) << 24) |
-	       ((data & 0x0000ff00) << 8) |
-	       ((data & 0x00ff0000) >> 8) |
-	       ((data & 0xff000000) >> 24));
-}
-
-/*
- * Define errno_t here as it is needed by the rom and ukernel
- */
-typedef u_int32_t errno_t;
-
-#define	EOK	0
-
-/*
- * Define the main communication structure used for passing
- * information from the rom to the ukernel (done here as it is
- * used by them both)
- */
-typedef struct rom_info_ rom_info_t;
-
-/*
- * Typedef the return code from the ukernel to the ROM
- */
-typedef u_int32_t rom_return_t;
-
-/*
- * Pull in the relevant global environment header file
- *
- * This file is shared by the uKernel and the system simulation effort.
- */
-#if defined(ENV_UKERN) || defined (ENV_SYS_SIM)
-#include "ukern.h"
-#endif /* ENV_UKERN */
-
-#if defined(ENV_ROM)
-#include "rom.h"
-#endif
- 
-#endif /* __DEFS_H__ */
diff --git a/sys/mips/include/queue.h b/sys/mips/include/queue.h
deleted file mode 100644
index d992332c84e..00000000000
--- a/sys/mips/include/queue.h
+++ /dev/null
@@ -1,171 +0,0 @@
-/*-
- * Copyright (c) 1996-1997, 2001, 2005, Juniper Networks, Inc.
- * All rights reserved.
- * Jim Hayes, November 1996
- *
- * queue.h - Description of uKernel queues, for the Juniper Kernel
- *
- *	JNPR: queue.h,v 1.1 2006/08/07 05:38:57 katta
- * $FreeBSD$
- *
- */
-
-#ifndef __QUEUE_H__
-#define	__QUEUE_H__
-
-/*---------------------------------------------------------------------------
- * QUEUE MANAGEMENT DOCUMENTATION
- */
-
-/*
-   --------
-   Q_INIT()
-   --------
-
-   void q_init(void)
-
-   Initialize the queue management system for the microkernel.
-   This initializes the debugging flags and sets up accounting.
-
-   ---------
-   Q_ALLOC()
-   ---------
-
-   queue_t *q_alloc()
-
-   Allocates a queue from kernel memory, and initializes it for you.
-
-   The default initialization provides a queue that is unbounded.
-
-   If you want to be bounded with special features, use q_control
-   after initialization.
-
-   q_alloc() returns NULL in the face of peril or low memory.
-
-   --------
-   Q_FREE()
-   --------
-
-   void *q_free(queue_t *queue_pointer)
-
-   Returns a queue to kernel memory, and frees the queue contents
-   for you using free() and complains (with a traceback) that you
-   tried to kill of a non-empty queue.
-
-   If any threads are waiting on the queue, wake them up.
-
-   -----------
-   Q_CONTROL()
-   -----------
-   void q_control(queue_t *queue_pointer, queue_size_t max_queue_size);
-
-   For now, allows you to limit queue growth.
-
-   ----------------
-   Q_DEQUEUE_WAIT() ** MAY CAUSE THREAD TO BLOCK/CANNOT BE CALLED FROM ISRs **
-   ----------------
-
-   void *q_dequeue_wait(queue_t *queue_pointer, wakeup_mask_t *mask)
-
-   Removes and returns a pointer to the next message in the specified
-   queue.  If the queue is empty, the calling thread goes to sleep
-   until something is queued to the queue.  If this call returns NULL,
-   then an extraordinary event requires this thread's attention--
-   check errno in this case.
-
-   ---------
-   Q_DEQUEUE	 ** CAN BE CALLED FROM ISRs **
-   ---------
-
-   void *q_dequeue(queue_t *queue_pointer)
-
-   Just like q_dequeue_wait(), but instead of blocking, return NULL.
-
-   -----------
-   Q_ENQUEUE()		   ** CAN BE CALLED FROM ISRs **
-   -----------
-
-   boolean q_enqueue(queue_t *queue_pointer, void *element_pointer)
-
-   Add the element to the end of the named queue.  If the add fails
-   because a limit has been reached, return TRUE.  Otherwise return
-   FALSE if everything went OK.
-
-   ----------
-   Q_URGENT()
-   ----------
-
-   boolean q_urgent(queue_t *queue_pointer, void *element_pointer)
-
-   Same as q_enqueue(), except this element will be placed at the top
-   of the queue, and will be picked off at the next q_dequeue_wait()
-   operation.
-
-   --------
-   Q_PEEK()		   ** CAN BE CALLED FROM ISRs **
-   --------
-
-   void *q_peek(queue_t *queue_pointer)
-
-   Returns a pointer to the top element of the queue without actually
-   dequeuing it.  Returns NULL of the queue is empty.
-
-   This routine will never block.
-
-   ----------
-   Q_DELETE()
-   ----------
-
-   void q_delete(queue_t *queue_pointer, void *element_pointer)
-
-   Delete the element_pointer from the queue, if it exists.  This
-   isn't speedy, and isn't meant for tasks requiring performance.
-   It's primary use is to pull something off the queue when you know
-   in the common case that it's gonna be at or near the top of the
-   list.  (I.e. waking a thread from a wake list when extraordinary
-   conditions exist, and you have to pluck it from the middle of the
-   list.)
-
-   This routine does not block or return anything.
-
-   --------
-   Q_SIZE()
-   --------
-
-   queue_size_t q_size(queue_t *queue_pointer)
-
-   Returns the number of elements in the queue.
-
-   ------------
-   Q_MAX_SIZE()
-   ------------
-
-   queue_size_t q_max_size(queue_t *queue_pointer);
-
-   Returns the maximum size of this queue, or 0 if this queue is
-   unbounded.
-
-*/
-
-/*-------------------------------------------------------------------------
- * Basic queue management structures.
- */
-
-/*
- * Typedefs
- */
-
-typedef u_int32_t queue_size_t;
-
-/*
- * Prototypes
- */
-
-void		 q_init(void);
-queue_t		*q_alloc(void);
-void		*q_peek(queue_t *queue);
-void		*q_dequeue(queue_t *queue);
-boolean		 q_enqueue(queue_t *queue, void *item);
-boolean		 q_urgent(queue_t *queue, void *item);
-
-#endif /* __QUEUE_H__ */
diff --git a/sys/mips/include/rm7000.h b/sys/mips/include/rm7000.h
deleted file mode 100644
index f1c0c44a0db..00000000000
--- a/sys/mips/include/rm7000.h
+++ /dev/null
@@ -1,95 +0,0 @@
-/*	$OpenBSD$ */
-
-/*
- * Copyright (c) 2000 Opsycon Open System Consulting AB	 (www.opsycon.se)
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. All advertising materials mentioning features or use of this software
- *    must display the following acknowledgement:
- *	This product includes software developed by Opsycon Open System
- *	Consulting AB, Sweden under contract to QED, Inc.
- * 4. The name of the author may not be used to endorse or promote products
- *    derived from this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS
- * OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY
- * DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *	JNPR: rm7000.h,v 1.2.4.1 2007/08/29 12:06:30 girish
- * $FreeBSD$
- */
-
-#ifndef _MACHINE_RM7000_H_
-#define	_MACHINE_RM7000_H_
-
-/*
- *  QED RM7000 specific defines.
- */
-
-/*
- *  Performance counters.
- */
-
-#define	PCNT_SRC_CLOCKS		0x00	/* Clock cycles */
-#define	PCNT_SRC_INSTR		0x01	/* Total instructions issued */
-#define	PCNT_SRC_FPINSTR	0x02	/* Float instructions issued */
-#define	PCNT_SRC_IINSTR		0x03	/* Integer instructions issued */
-#define	PCNT_SRC_LOAD		0x04	/* Load instructions issued */
-#define	PCNT_SRC_STORE		0x05	/* Store instructions issued */
-#define	PCNT_SRC_DUAL		0x06	/* Dual issued pairs */
-#define	PCNT_SRC_BRPREF		0x07	/* Branch prefetches */
-#define	PCNT_SRC_EXTMISS	0x08	/* External cache misses */
-#define	PCNT_SRC_STALL		0x09	/* Stall cycles */
-#define	PCNT_SRC_SECMISS	0x0a	/* Secondary cache misses */
-#define	PCNT_SRC_INSMISS	0x0b	/* Instruction cache misses */
-#define	PCNT_SRC_DTAMISS	0x0c	/* Data cache misses */
-#define	PCNT_SRC_DTLBMISS	0x0d	/* Data TLB misses */
-#define	PCNT_SRC_ITLBMISS	0x0e	/* Instruction TLB misses */
-#define	PCNT_SRC_JTLBIMISS	0x0f	/* Joint TLB instruction misses */
-#define	PCNT_SRC_JTLBDMISS	0x10	/* Joint TLB data misses */
-#define	PCNT_SRC_BRTAKEN	0x11	/* Branches taken */
-#define	PCNT_SRC_BRISSUED	0x12	/* Branches issued */
-#define	PCNT_SRC_SECWBACK	0x13	/* Secondary cache writebacks */
-#define	PCNT_SRC_PRIWBACK	0x14	/* Primary cache writebacks */
-#define	PCNT_SRC_DCSTALL	0x15	/* Dcache miss stall cycles */
-#define	PCNT_SRC_MISS		0x16	/* Cache misses */
-#define	PCNT_SRC_FPEXC		0x17	/* FP possible execption cycles */
-#define	PCNT_SRC_MULSLIP	0x18	/* Slip cycles due to mult. busy */
-#define	PCNT_SRC_CP0SLIP	0x19	/* CP0 Slip cycles */
-#define	PCNT_SRC_LDSLIP		0x1a	/* Slip cycles	due to pend. non-b ld */
-#define	PCNT_SRC_WBFULL		0x1b	/* Write buffer full stall cycles  */
-#define	PCNT_SRC_CISTALL	0x1c	/* Cache instruction stall cycles  */
-#define	PCNT_SRC_MULSTALL	0x1d	/* Multiplier stall cycles  */
-#define	PCNT_SRC_ELDSTALL	0x1d	/* Excepion stall due to non-b ld */
-#define	PCNT_SRC_MAX		0x1d	/* Maximum PCNT select code */
-
-/*
- *  Counter control bits.
- */
-
-#define	PCNT_CE			0x0400	/* Count enable */
-#define	PCNT_UM			0x0200	/* Count in User mode */
-#define	PCNT_KM			0x0100	/* Count in kernel mode */
-
-/*
- *  Performance counter system call function codes.
- */
-#define	PCNT_FNC_SELECT		0x0001	/* Select counter source */
-#define	PCNT_FNC_READ		0x0002	/* Read current value of counter */
-
-#endif /* _MACHINE_RM7000_H_ */
diff --git a/sys/mips/include/segments.h b/sys/mips/include/segments.h
deleted file mode 100644
index 406b965f2e0..00000000000
--- a/sys/mips/include/segments.h
+++ /dev/null
@@ -1,40 +0,0 @@
-/*-
- * Copyright (c) 1989, 1990 William F. Jolitz
- * Copyright (c) 1990 The Regents of the University of California.
- * All rights reserved.
- *
- * This code is derived from software contributed to Berkeley by
- * William Jolitz.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 4. Neither the name of the University nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *	from: @(#)segments.h	7.1 (Berkeley) 5/9/91
- * $FreeBSD$
- */
-
-#ifndef _MACHINE_SEGMENTS_H_
-#define	_MACHINE_SEGMENTS_H_
-
-#endif /* !_MACHINE_SEGMENTS_H_ */
diff --git a/sys/mips/mips/trap.c b/sys/mips/mips/trap.c
index 124087cc039..8bfc6e7f4d5 100644
--- a/sys/mips/mips/trap.c
+++ b/sys/mips/mips/trap.c
@@ -83,8 +83,6 @@ __FBSDID("$FreeBSD$");
 #include <machine/mips_opcode.h>
 #include <machine/frame.h>
 #include <machine/regnum.h>
-#include <machine/rm7000.h>
-#include <machine/archtype.h>
 #include <machine/asm.h>
 
 #ifdef DDB

From 09398e9bd42e66e6e6693460f5cd995c1d794da8 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Fri, 16 Apr 2010 06:47:29 +0000
Subject: [PATCH 085/532] Fix control socket leak when worker process exits.

Submitted by:	Mikolaj Golub <to.my.trociny@gmail.com>
MFC after:	3 days
---
 sbin/hastd/hastd.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sbin/hastd/hastd.c b/sbin/hastd/hastd.c
index 19f08936c2d..957885d7ac1 100644
--- a/sbin/hastd/hastd.c
+++ b/sbin/hastd/hastd.c
@@ -137,6 +137,7 @@ child_exit(void)
 			pjdlog_error("Worker process failed (pid=%u, status=%d).",
 			    (unsigned int)pid, WEXITSTATUS(status));
 		}
+		proto_close(res->hr_ctrl);
 		res->hr_workerpid = 0;
 		if (res->hr_role == HAST_ROLE_PRIMARY) {
 			sleep(1);

From 20ec52dc4bee544f948a29e8a31976302216cb69 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Fri, 16 Apr 2010 06:49:12 +0000
Subject: [PATCH 086/532] Fix log size calculation which caused message
 truncation.

Submitted by:	Mikolaj Golub <to.my.trociny@gmail.com>
MFC after:	3 days
---
 sbin/hastd/pjdlog.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sbin/hastd/pjdlog.c b/sbin/hastd/pjdlog.c
index 38c5539ad0c..9f8b3f499de 100644
--- a/sbin/hastd/pjdlog.c
+++ b/sbin/hastd/pjdlog.c
@@ -228,7 +228,7 @@ pjdlogv_common(int loglevel, int debuglevel, int error, const char *fmt,
 
 		len = snprintf(log, sizeof(log), "%s", pjdlog_prefix);
 		if ((size_t)len < sizeof(log))
-			len = vsnprintf(log + len, sizeof(log) - len, fmt, ap);
+			len += vsnprintf(log + len, sizeof(log) - len, fmt, ap);
 		if (error != -1 && (size_t)len < sizeof(log)) {
 			(void)snprintf(log + len, sizeof(log) - len, ": %s.",
 			    strerror(error));

From 17f820725e8ef038ae1d08cb75a1781eb8ad8cc3 Mon Sep 17 00:00:00 2001
From: Jaakko Heinonen <jh@FreeBSD.org>
Date: Fri, 16 Apr 2010 07:02:28 +0000
Subject: [PATCH 087/532] Revert r206560. The change doesn't work correctly in
 all cases with multiple devfs mounts.

---
 sys/fs/devfs/devfs_devs.c | 23 -----------------------
 sys/fs/devfs/devfs_int.h  |  1 -
 2 files changed, 24 deletions(-)

diff --git a/sys/fs/devfs/devfs_devs.c b/sys/fs/devfs/devfs_devs.c
index 79037ba6313..81b2752583e 100644
--- a/sys/fs/devfs/devfs_devs.c
+++ b/sys/fs/devfs/devfs_devs.c
@@ -408,9 +408,6 @@ devfs_populate_loop(struct devfs_mount *dm, int cleanup)
 			continue;
 		KASSERT((cdp->cdp_flags & CDP_ACTIVE), ("Bogons, I tell ya'!"));
 
-		if (cdp->cdp_flags & CDP_INVALID)
-			continue;
-
 		if (dm->dm_idx <= cdp->cdp_maxdirent &&
 		    cdp->cdp_dirents[dm->dm_idx] != NULL) {
 			de = cdp->cdp_dirents[dm->dm_idx];
@@ -428,8 +425,6 @@ devfs_populate_loop(struct devfs_mount *dm, int cleanup)
 		dd = dm->dm_rootdir;
 		s = cdp->cdp_c.si_name;
 		for (;;) {
-			while (*s == '/')
-				s++;
 			for (q = s; *q != '/' && *q != '\0'; q++)
 				continue;
 			if (*q != '/')
@@ -439,24 +434,6 @@ devfs_populate_loop(struct devfs_mount *dm, int cleanup)
 				de = devfs_vmkdir(dm, s, q - s, dd, 0);
 			s = q + 1;
 			dd = de;
-			if (dd->de_flags & (DE_DOT | DE_DOTDOT))
-				break;
-		}
-
-		/*
-		 * XXX: Ignore duplicate and empty device names.
-		 * XXX: Currently there is no way to report the error to
-		 * XXX: the make_dev(9) caller.
-		 */
-		if (dd->de_dirent->d_type != DT_DIR ||
-		    dd->de_flags & (DE_DOT | DE_DOTDOT) || q - s < 1 ||
-		    devfs_find(dd, s, q - s) != NULL) {
-			dev_lock();
-			cdp->cdp_flags |= CDP_INVALID;
-			dev_unlock();
-			printf("%s: %s: invalid or duplicate device name\n",
-			    __func__, cdp->cdp_c.si_name);
-			return (1);
 		}
 
 		de = devfs_newdirent(s, q - s);
diff --git a/sys/fs/devfs/devfs_int.h b/sys/fs/devfs/devfs_int.h
index a998061513f..5a61dd4a79b 100644
--- a/sys/fs/devfs/devfs_int.h
+++ b/sys/fs/devfs/devfs_int.h
@@ -55,7 +55,6 @@ struct cdev_priv {
 	u_int			cdp_flags;
 #define CDP_ACTIVE		(1 << 0)
 #define CDP_SCHED_DTR		(1 << 1)
-#define CDP_INVALID		(1 << 2)
 
 	u_int			cdp_inuse;
 	u_int			cdp_maxdirent;

From c99cdece4efc95b389db21957dc2a9c39e744299 Mon Sep 17 00:00:00 2001
From: Jack F Vogel <jfv@FreeBSD.org>
Date: Fri, 16 Apr 2010 16:33:05 +0000
Subject: [PATCH 088/532] Remove the tx queue selection based on the cpu whe no
 flowid is present, this was causing some bad reordering, now just use 0.

Also, add a few watchdog bits, and tx handler bits
that were corrected in igb.
---
 sys/dev/ixgbe/ixgbe.c | 28 +++++++++++++++-------------
 1 file changed, 15 insertions(+), 13 deletions(-)

diff --git a/sys/dev/ixgbe/ixgbe.c b/sys/dev/ixgbe/ixgbe.c
index f8338e9aee4..9cc392db955 100644
--- a/sys/dev/ixgbe/ixgbe.c
+++ b/sys/dev/ixgbe/ixgbe.c
@@ -759,6 +759,7 @@ ixgbe_start_locked(struct tx_ring *txr, struct ifnet * ifp)
 
 		/* Set watchdog on */
 		txr->watchdog_check = TRUE;
+		txr->watchdog_time = ticks;
 
 	}
 	return;
@@ -798,8 +799,6 @@ ixgbe_mq_start(struct ifnet *ifp, struct mbuf *m)
 	/* Which queue to use */
 	if ((m->m_flags & M_FLOWID) != 0)
 		i = m->m_pkthdr.flowid % adapter->num_queues;
-	else	/* use the cpu we're on */
-		i = curcpu % adapter->num_queues;
 
 	txr = &adapter->tx_rings[i];
 
@@ -856,8 +855,11 @@ ixgbe_mq_start_locked(struct ifnet *ifp, struct tx_ring *txr, struct mbuf *m)
 		next = drbr_dequeue(ifp, txr->br);
 	}
 
-	if (enqueued > 0) 
+	if (enqueued > 0) {
+		/* Set watchdog on */
 		txr->watchdog_check = TRUE;
+		txr->watchdog_time = ticks;
+	}
 
 	return (err);
 }
@@ -1251,16 +1253,12 @@ ixgbe_handle_que(void *context, int pending)
 	struct adapter  *adapter = que->adapter;
 	struct tx_ring  *txr = que->txr;
 	struct ifnet    *ifp = adapter->ifp;
-	u32		loop = MAX_LOOP;
-	bool		more_rx, more_tx;
-
-	IXGBE_TX_LOCK(txr);
-	do {
-		more_rx = ixgbe_rxeof(que, adapter->rx_process_limit);
-		more_tx = ixgbe_txeof(txr);
-	} while (loop-- && (more_rx || more_tx));
+	bool		more;
 
 	if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
+		more = ixgbe_rxeof(que, adapter->rx_process_limit);
+		IXGBE_TX_LOCK(txr);
+		ixgbe_txeof(txr);
 #if __FreeBSD_version >= 800000
 		if (!drbr_empty(ifp, txr->br))
 			ixgbe_mq_start_locked(ifp, txr, NULL);
@@ -1268,11 +1266,16 @@ ixgbe_handle_que(void *context, int pending)
 		if (!IFQ_DRV_IS_EMPTY(&ifp->if_snd))
 			ixgbe_start_locked(txr, ifp);
 #endif
+		IXGBE_TX_UNLOCK(txr);
+		if (more) {
+			taskqueue_enqueue(que->tq, &que->que_task);
+			return;
+		}
 	}
 
-	IXGBE_TX_UNLOCK(txr);
 	/* Reenable this interrupt */
 	ixgbe_enable_queue(adapter, que->msix);
+	return;
 }
 
 
@@ -1718,7 +1721,6 @@ ixgbe_xmit(struct tx_ring *txr, struct mbuf **m_headp)
 	 * hardware that this frame is available to transmit.
 	 */
 	++txr->total_packets;
-	txr->watchdog_time = ticks;
 	IXGBE_WRITE_REG(&adapter->hw, IXGBE_TDT(txr->me), i);
 
 	/* Do a clean if descriptors are low */

From 7a246d0b1457b7cb94c0393b9924940ffa6d4360 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Fri, 16 Apr 2010 16:49:42 +0000
Subject: [PATCH 089/532] Add ubthidhci.

MFC after:	2 days
---
 etc/rc.d/Makefile | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/etc/rc.d/Makefile b/etc/rc.d/Makefile
index 17f76348664..802d7179b8e 100755
--- a/etc/rc.d/Makefile
+++ b/etc/rc.d/Makefile
@@ -50,6 +50,10 @@ FILES+=	sshd
 FILES+= nscd
 .endif
 
+.if ${MK_BLUETOOTH} != "no"
+FILES+=	ubthidhci
+.endif
+
 FILESDIR=	/etc/rc.d
 FILESMODE=	${BINMODE}
 

From 0c5aed59ee17fba47e089a089a822ac07af83009 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Fri, 16 Apr 2010 20:07:24 +0000
Subject: [PATCH 090/532] Expose a few symbols as public interface rather than
 private.

Note: the *64 interfaces are no longer exposed via zlib.h but were keep
as public interfaces;

Note 2: this commit would break applications that uses the moved symbols
directly.
---
 lib/libz/Symbol.map | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/lib/libz/Symbol.map b/lib/libz/Symbol.map
index 714606b5949..2651efc70fa 100644
--- a/lib/libz/Symbol.map
+++ b/lib/libz/Symbol.map
@@ -4,11 +4,13 @@
 
 ZLIB_1.2.4.0 {
 	adler32;
+	adler32_combine;
 	adler32_combine64;
 	compress;
 	compress2;
 	compressBound;
 	crc32;
+	crc32_combine;
 	crc32_combine64;
 	deflate;
 	deflateBound;
@@ -35,15 +37,19 @@ ZLIB_1.2.4.0 {
 	gzflush;
 	gzgetc;
 	gzgets;
+	gzoffset;
 	gzoffset64;
+	gzopen;
 	gzopen64;
 	gzprintf;
 	gzputc;
 	gzputs;
 	gzread;
 	gzrewind;
+	gzseek;
 	gzseek64;
 	gzsetparams;
+	gztell;
 	gztell64;
 	gzungetc;
 	gzwrite;
@@ -80,13 +86,7 @@ ZLIBprivate_1.0 {
 	_tr_init;
 	_tr_stored_block;
 	_tr_tally;
-	adler32_combine;
-	crc32_combine;
 	gz_error;
-	gzoffset;
-	gzopen;
-	gzseek;
-	gztell;
 	inflate_fast;
 	inflate_table;
 	longest_match;

From bfbe51816a5449beec27c58e62cf0cb50e44bce7 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 16 Apr 2010 22:15:26 +0000
Subject: [PATCH 091/532] Add some tests for fnmatch(3).

MFC after:	1 week
---
 tools/regression/lib/libc/gen/Makefile       |   2 +-
 tools/regression/lib/libc/gen/test-fnmatch.c | 335 +++++++++++++++++++
 2 files changed, 336 insertions(+), 1 deletion(-)
 create mode 100644 tools/regression/lib/libc/gen/test-fnmatch.c

diff --git a/tools/regression/lib/libc/gen/Makefile b/tools/regression/lib/libc/gen/Makefile
index 3dbe803a7be..3e911e4fc42 100644
--- a/tools/regression/lib/libc/gen/Makefile
+++ b/tools/regression/lib/libc/gen/Makefile
@@ -1,6 +1,6 @@
 # $FreeBSD$
 
-TESTS=	test-fmtcheck test-fpclassify test-wordexp
+TESTS=	test-fmtcheck test-fnmatch test-fpclassify test-wordexp
 
 .PHONY: tests
 tests: ${TESTS}
diff --git a/tools/regression/lib/libc/gen/test-fnmatch.c b/tools/regression/lib/libc/gen/test-fnmatch.c
new file mode 100644
index 00000000000..ffb8c6c7800
--- /dev/null
+++ b/tools/regression/lib/libc/gen/test-fnmatch.c
@@ -0,0 +1,335 @@
+/*-
+ * Copyright (c) 2010 Jilles Tjoelker
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <fnmatch.h>
+
+struct testcase {
+	const char *pattern;
+	const char *string;
+	int flags;
+	int result;
+} testcases[] = {
+	"", "", 0, 0,
+	"a", "a", 0, 0,
+	"a", "b", 0, FNM_NOMATCH,
+	"a", "A", 0, FNM_NOMATCH,
+	"*", "a", 0, 0,
+	"*", "aa", 0, 0,
+	"*a", "a", 0, 0,
+	"*a", "b", 0, FNM_NOMATCH,
+	"*a*", "b", 0, FNM_NOMATCH,
+	"*a*b*", "ab", 0, 0,
+	"*a*b*", "qaqbq", 0, 0,
+	"*a*bb*", "qaqbqbbq", 0, 0,
+	"*a*bc*", "qaqbqbcq", 0, 0,
+	"*a*bb*", "qaqbqbb", 0, 0,
+	"*a*bc*", "qaqbqbc", 0, 0,
+	"*a*bb", "qaqbqbb", 0, 0,
+	"*a*bc", "qaqbqbc", 0, 0,
+	"*a*bb", "qaqbqbbq", 0, FNM_NOMATCH,
+	"*a*bc", "qaqbqbcq", 0, FNM_NOMATCH,
+	"*a*a*a*a*a*a*a*a*a*a*", "aaaaaaaaa", 0, FNM_NOMATCH,
+	"*a*a*a*a*a*a*a*a*a*a*", "aaaaaaaaaa", 0, 0,
+	"*a*a*a*a*a*a*a*a*a*a*", "aaaaaaaaaaa", 0, 0,
+	".*.*.*.*.*.*.*.*.*.*", ".........", 0, FNM_NOMATCH,
+	".*.*.*.*.*.*.*.*.*.*", "..........", 0, 0,
+	".*.*.*.*.*.*.*.*.*.*", "...........", 0, 0,
+	"*?*?*?*?*?*?*?*?*?*?*", "123456789", 0, FNM_NOMATCH,
+	"??????????*", "123456789", 0, FNM_NOMATCH,
+	"*??????????", "123456789", 0, FNM_NOMATCH,
+	"*?*?*?*?*?*?*?*?*?*?*", "1234567890", 0, 0,
+	"??????????*", "1234567890", 0, 0,
+	"*??????????", "1234567890", 0, 0,
+	"*?*?*?*?*?*?*?*?*?*?*", "12345678901", 0, 0,
+	"??????????*", "12345678901", 0, 0,
+	"*??????????", "12345678901", 0, 0,
+	"[x]", "x", 0, 0,
+	"[*]", "*", 0, 0,
+	"[?]", "?", 0, 0,
+	"[", "[", 0, 0,
+	"[[]", "[", 0, 0,
+	"[[]", "x", 0, FNM_NOMATCH,
+	"[*]", "", 0, FNM_NOMATCH,
+	"[*]", "x", 0, FNM_NOMATCH,
+	"[?]", "x", 0, FNM_NOMATCH,
+	"*[*]*", "foo*foo", 0, 0,
+	"*[*]*", "foo", 0, FNM_NOMATCH,
+	"[0-9]", "0", 0, 0,
+	"[0-9]", "5", 0, 0,
+	"[0-9]", "9", 0, 0,
+	"[0-9]", "/", 0, FNM_NOMATCH,
+	"[0-9]", ":", 0, FNM_NOMATCH,
+	"[0-9]", "*", 0, FNM_NOMATCH,
+	"[!0-9]", "0", 0, FNM_NOMATCH,
+	"[!0-9]", "5", 0, FNM_NOMATCH,
+	"[!0-9]", "9", 0, FNM_NOMATCH,
+	"[!0-9]", "/", 0, 0,
+	"[!0-9]", ":", 0, 0,
+	"[!0-9]", "*", 0, 0,
+	"*[0-9]", "a0", 0, 0,
+	"*[0-9]", "a5", 0, 0,
+	"*[0-9]", "a9", 0, 0,
+	"*[0-9]", "a/", 0, FNM_NOMATCH,
+	"*[0-9]", "a:", 0, FNM_NOMATCH,
+	"*[0-9]", "a*", 0, FNM_NOMATCH,
+	"*[!0-9]", "a0", 0, FNM_NOMATCH,
+	"*[!0-9]", "a5", 0, FNM_NOMATCH,
+	"*[!0-9]", "a9", 0, FNM_NOMATCH,
+	"*[!0-9]", "a/", 0, 0,
+	"*[!0-9]", "a:", 0, 0,
+	"*[!0-9]", "a*", 0, 0,
+	"*[0-9]", "a00", 0, 0,
+	"*[0-9]", "a55", 0, 0,
+	"*[0-9]", "a99", 0, 0,
+	"*[0-9]", "a0a0", 0, 0,
+	"*[0-9]", "a5a5", 0, 0,
+	"*[0-9]", "a9a9", 0, 0,
+	"\\*", "*", 0, 0,
+	"\\?", "?", 0, 0,
+	"\\[x]", "[x]", 0, 0,
+	"\\[", "[", 0, 0,
+	"\\\\", "\\", 0, 0,
+	"*\\**", "foo*foo", 0, 0,
+	"*\\**", "foo", 0, FNM_NOMATCH,
+	"*\\\\*", "foo\\foo", 0, 0,
+	"*\\\\*", "foo", 0, FNM_NOMATCH,
+	"\\(", "(", 0, 0,
+	"\\a", "a", 0, 0,
+	"\\*", "a", 0, FNM_NOMATCH,
+	"\\?", "a", 0, FNM_NOMATCH,
+	"\\*", "\\*", 0, FNM_NOMATCH,
+	"\\?", "\\?", 0, FNM_NOMATCH,
+	"\\[x]", "\\[x]", 0, FNM_NOMATCH,
+	"\\[x]", "\\x", 0, FNM_NOMATCH,
+	"\\[", "\\[", 0, FNM_NOMATCH,
+	"\\(", "\\(", 0, FNM_NOMATCH,
+	"\\a", "\\a", 0, FNM_NOMATCH,
+	"\\*", "\\*", FNM_NOESCAPE, 0,
+	"\\?", "\\?", FNM_NOESCAPE, 0,
+	"\\", "\\", FNM_NOESCAPE, 0,
+	"\\\\", "\\", FNM_NOESCAPE, FNM_NOMATCH,
+	"\\\\", "\\\\", FNM_NOESCAPE, 0,
+	"*\\*", "foo\\foo", FNM_NOESCAPE, 0,
+	"*\\*", "foo", FNM_NOESCAPE, FNM_NOMATCH,
+	"*", ".", FNM_PERIOD, FNM_NOMATCH,
+	"?", ".", FNM_PERIOD, FNM_NOMATCH,
+	".*", ".", 0, 0,
+	".*", "..", 0, 0,
+	".*", ".a", 0, 0,
+	"[0-9]", ".", FNM_PERIOD, FNM_NOMATCH,
+	"a*", "a.", 0, 0,
+	"a/a", "a/a", FNM_PATHNAME, 0,
+	"a/*", "a/a", FNM_PATHNAME, 0,
+	"*/a", "a/a", FNM_PATHNAME, 0,
+	"*/*", "a/a", FNM_PATHNAME, 0,
+	"a*b/*", "abbb/x", FNM_PATHNAME, 0,
+	"a*b/*", "abbb/.x", FNM_PATHNAME, 0,
+	"*", "a/a", FNM_PATHNAME, FNM_NOMATCH,
+	"*/*", "a/a/a", FNM_PATHNAME, FNM_NOMATCH,
+	"b/*", "b/.x", FNM_PATHNAME | FNM_PERIOD, FNM_NOMATCH,
+	"b*/*", "a/.x", FNM_PATHNAME | FNM_PERIOD, FNM_NOMATCH,
+	"b/.*", "b/.x", FNM_PATHNAME | FNM_PERIOD, 0,
+	"b*/.*", "b/.x", FNM_PATHNAME | FNM_PERIOD, 0,
+	"a", "A", FNM_CASEFOLD, 0,
+	"A", "a", FNM_CASEFOLD, 0,
+	"[a]", "A", FNM_CASEFOLD, 0,
+	"[A]", "a", FNM_CASEFOLD, 0,
+	"a", "b", FNM_CASEFOLD, FNM_NOMATCH,
+	"a", "a/b", FNM_PATHNAME, FNM_NOMATCH,
+	"*", "a/b", FNM_PATHNAME, FNM_NOMATCH,
+	"*b", "a/b", FNM_PATHNAME, FNM_NOMATCH,
+	"a", "a/b", FNM_PATHNAME | FNM_LEADING_DIR, 0,
+	"*", "a/b", FNM_PATHNAME | FNM_LEADING_DIR, 0,
+	"*", ".a/b", FNM_PATHNAME | FNM_LEADING_DIR, 0,
+	"*a", ".a/b", FNM_PATHNAME | FNM_LEADING_DIR, 0,
+	"*", ".a/b", FNM_PATHNAME | FNM_PERIOD | FNM_LEADING_DIR, FNM_NOMATCH,
+	"*a", ".a/b", FNM_PATHNAME | FNM_PERIOD | FNM_LEADING_DIR, FNM_NOMATCH,
+};
+
+static const char *
+flags_to_string(int flags)
+{
+	static const int flagvalues[] = { FNM_NOESCAPE, FNM_PATHNAME,
+		FNM_PERIOD, FNM_LEADING_DIR, FNM_CASEFOLD, 0 };
+	static const char flagnames[] = "FNM_NOESCAPE\0FNM_PATHNAME\0FNM_PERIOD\0FNM_LEADING_DIR\0FNM_CASEFOLD\0";
+	static char result[sizeof(flagnames) + 3 * sizeof(int) + 2];
+	char *p;
+	size_t i, len;
+	const char *fp;
+
+	p = result;
+	fp = flagnames;
+	for (i = 0; flagvalues[i] != 0; i++) {
+		len = strlen(fp);
+		if (flags & flagvalues[i]) {
+			if (p != result)
+				*p++ = '|';
+			memcpy(p, fp, len);
+			p += len;
+			flags &= ~flagvalues[i];
+		}
+		fp += len + 1;
+	}
+	if (p == result)
+		memcpy(p, "0", 2);
+	else if (flags != 0)
+		sprintf(p, "%d", flags);
+	else
+		*p = '\0';
+	return result;
+}
+
+int
+main(int argc, char *argv[])
+{
+	size_t i, n;
+	int flags, result, extra, errors;
+	struct testcase *t;
+
+	n = sizeof(testcases) / sizeof(testcases[0]);
+	errors = 0;
+	printf("1..%zu\n", n);
+	for (i = 0; i < n; i++) {
+		t = &testcases[i];
+		flags = t->flags;
+		extra = 0;
+		do {
+			result = fnmatch(t->pattern, t->string, flags);
+			if (result != t->result)
+				break;
+			if (strchr(t->pattern, '\\') == NULL &&
+			    !(flags & FNM_NOESCAPE)) {
+				flags |= FNM_NOESCAPE;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if (strchr(t->pattern, '\\') != NULL &&
+			    strchr(t->string, '\\') == NULL &&
+			    t->result == FNM_NOMATCH &&
+			    !(flags & (FNM_NOESCAPE | FNM_LEADING_DIR))) {
+				flags |= FNM_NOESCAPE;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if ((t->string[0] != '.' || t->pattern[0] == '.' ||
+			    t->result == FNM_NOMATCH) &&
+			    !(flags & (FNM_PATHNAME | FNM_PERIOD))) {
+				flags |= FNM_PERIOD;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if ((strchr(t->string, '/') == NULL ||
+			    t->result == FNM_NOMATCH) &&
+			    !(flags & FNM_PATHNAME)) {
+				flags |= FNM_PATHNAME;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if ((((t->string[0] != '.' || t->pattern[0] == '.') &&
+			    strstr(t->string, "/.") == NULL) ||
+			    t->result == FNM_NOMATCH) &&
+			    flags & FNM_PATHNAME && !(flags & FNM_PERIOD)) {
+				flags |= FNM_PERIOD;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if ((((t->string[0] != '.' || t->pattern[0] == '.') &&
+			    strchr(t->string, '/') == NULL) ||
+			    t->result == FNM_NOMATCH) &&
+			    !(flags & (FNM_PATHNAME | FNM_PERIOD))) {
+				flags |= FNM_PATHNAME | FNM_PERIOD;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if ((strchr(t->string, '/') == NULL || t->result == 0)
+			    && !(flags & FNM_LEADING_DIR)) {
+				flags |= FNM_LEADING_DIR;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if (t->result == 0 && !(flags & FNM_CASEFOLD)) {
+				flags |= FNM_CASEFOLD;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+			if (strchr(t->pattern, '\\') == NULL &&
+			    t->result == 0 &&
+			    !(flags & (FNM_NOESCAPE | FNM_CASEFOLD))) {
+				flags |= FNM_NOESCAPE | FNM_CASEFOLD;
+				result = fnmatch(t->pattern, t->string, flags);
+				if (result != t->result)
+					break;
+				flags = t->flags;
+				extra++;
+			}
+		} while (0);
+		if (result == t->result)
+			printf("ok %zu - fnmatch(\"%s\", \"%s\", %s) = %d (+%d)\n",
+			    i + 1, t->pattern, t->string,
+			    flags_to_string(flags),
+			    result, extra);
+		else {
+			printf("not ok %zu - fnmatch(\"%s\", \"%s\", %s) = %d != %d\n",
+			    i + 1, t->pattern, t->string,
+			    flags_to_string(flags),
+			    result, t->result);
+			errors = 1;
+		}
+	}
+
+	return (errors);
+}

From 139ac6b23976252cbbc61df7ac4a49afa3431d1c Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 16 Apr 2010 22:29:24 +0000
Subject: [PATCH 092/532] fnmatch: Fix bad FNM_PERIOD disabling if an asterisk
 has been seen.

Example: fnmatch("a*b/*", "abbb/.x", FNM_PATHNAME | FNM_PERIOD)

PR:		116074
MFC after:	1 week
---
 lib/libc/gen/fnmatch.c                       | 16 ++++++++--------
 tools/regression/lib/libc/gen/test-fnmatch.c |  1 +
 2 files changed, 9 insertions(+), 8 deletions(-)

diff --git a/lib/libc/gen/fnmatch.c b/lib/libc/gen/fnmatch.c
index e697bdcedf2..3acbc3b1647 100644
--- a/lib/libc/gen/fnmatch.c
+++ b/lib/libc/gen/fnmatch.c
@@ -67,7 +67,8 @@ __FBSDID("$FreeBSD$");
 #define RANGE_ERROR     (-1)
 
 static int rangematch(const char *, wchar_t, int, char **, mbstate_t *);
-static int fnmatch1(const char *, const char *, int, mbstate_t, mbstate_t);
+static int fnmatch1(const char *, const char *, const char *, int, mbstate_t,
+		mbstate_t);
 
 int
 fnmatch(pattern, string, flags)
@@ -76,22 +77,21 @@ fnmatch(pattern, string, flags)
 {
 	static const mbstate_t initial;
 
-	return (fnmatch1(pattern, string, flags, initial, initial));
+	return (fnmatch1(pattern, string, string, flags, initial, initial));
 }
 
 static int
-fnmatch1(pattern, string, flags, patmbs, strmbs)
-	const char *pattern, *string;
+fnmatch1(pattern, string, stringstart, flags, patmbs, strmbs)
+	const char *pattern, *string, *stringstart;
 	int flags;
 	mbstate_t patmbs, strmbs;
 {
-	const char *stringstart;
 	char *newp;
 	char c;
 	wchar_t pc, sc;
 	size_t pclen, sclen;
 
-	for (stringstart = string;;) {
+	for (;;) {
 		pclen = mbrtowc(&pc, pattern, MB_LEN_MAX, &patmbs);
 		if (pclen == (size_t)-1 || pclen == (size_t)-2)
 			return (FNM_NOMATCH);
@@ -145,8 +145,8 @@ fnmatch1(pattern, string, flags, patmbs, strmbs)
 
 			/* General case, use recursion. */
 			while (sc != EOS) {
-				if (!fnmatch1(pattern, string,
-				    flags & ~FNM_PERIOD, patmbs, strmbs))
+				if (!fnmatch1(pattern, string, stringstart,
+				    flags, patmbs, strmbs))
 					return (0);
 				sclen = mbrtowc(&sc, string, MB_LEN_MAX,
 				    &strmbs);
diff --git a/tools/regression/lib/libc/gen/test-fnmatch.c b/tools/regression/lib/libc/gen/test-fnmatch.c
index ffb8c6c7800..6cbf8eeef70 100644
--- a/tools/regression/lib/libc/gen/test-fnmatch.c
+++ b/tools/regression/lib/libc/gen/test-fnmatch.c
@@ -174,6 +174,7 @@ struct testcase {
 	"*a", ".a/b", FNM_PATHNAME | FNM_LEADING_DIR, 0,
 	"*", ".a/b", FNM_PATHNAME | FNM_PERIOD | FNM_LEADING_DIR, FNM_NOMATCH,
 	"*a", ".a/b", FNM_PATHNAME | FNM_PERIOD | FNM_LEADING_DIR, FNM_NOMATCH,
+	"a*b/*", "abbb/.x", FNM_PATHNAME | FNM_PERIOD, FNM_NOMATCH,
 };
 
 static const char *

From 4e859425ae287fde522d93b98f5df53814d1a9d9 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 16 Apr 2010 23:42:19 +0000
Subject: [PATCH 093/532] Set KERNLOADADDR and TARGET_BIG_ENDIAN for SWARM.

---
 sys/mips/conf/SWARM | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sys/mips/conf/SWARM b/sys/mips/conf/SWARM
index 3189bcbcb9c..455e2986f9c 100644
--- a/sys/mips/conf/SWARM
+++ b/sys/mips/conf/SWARM
@@ -19,8 +19,13 @@ options		CFE_CONSOLE
 options		CFE_ENV
 options		ALT_BREAK_TO_DEBUGGER
 
+# Don't build any modules yet.
+makeoptions	MODULES_OVERRIDE=""
+makeoptions	TARGET_BIG_ENDIAN=defined
 makeoptions	LDSCRIPT_NAME=	ldscript.mips.cfe
 
+makeoptions	KERNLOADADDR=0x81000000
+
 #cpu		CPU_MIPS64
 #options 	ISA_MIPS64
 #makeoptions	ARCH_FLAGS="-march=mips64 -mgp64 -mabi=o64"

From d0985cfb4121e3e5552101aa5b1789431ee70910 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 16 Apr 2010 23:46:30 +0000
Subject: [PATCH 094/532] o) Fix the intr_* functions to not spam the whole
 status register, just the IE    bit. o) Remove some unused inlines. o)
 Generate CP0 access functions for 64-bit TLB registers when building for   
 n64. o) Add an inline function version of the COP0_SYNC macro.

---
 sys/mips/include/cpufunc.h | 101 +++++++++++++++++--------------------
 1 file changed, 45 insertions(+), 56 deletions(-)

diff --git a/sys/mips/include/cpufunc.h b/sys/mips/include/cpufunc.h
index f9100eaeb32..9c34e3152d8 100644
--- a/sys/mips/include/cpufunc.h
+++ b/sys/mips/include/cpufunc.h
@@ -1,5 +1,29 @@
 /*	$OpenBSD: pio.h,v 1.2 1998/09/15 10:50:12 pefo Exp $	*/
 
+/*-
+ * Copyright (c) 2002-2004 Juli Mallett.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
 /*
  * Copyright (c) 1995-1999 Per Fogelstrom.  All rights reserved.
  *
@@ -58,15 +82,17 @@ mips_barrier(void)
 			  : : : "memory");
 }
 
+static __inline void
+mips_cp0_sync(void)
+{
+	__asm __volatile (__XSTRING(COP0_SYNC));
+}
+
 static __inline void
 mips_wbflush(void)
 {
 	__asm __volatile ("sync" : : : "memory");
 	mips_barrier();
-#if 0
-	__asm __volatile("mtc0 %0, $12\n" /* MIPS_COP_0_STATUS */
-	   : : "r" (flag));
-#endif
 }
 
 static __inline void
@@ -82,54 +108,7 @@ mips_write_membar(void)
 }
 
 #ifdef _KERNEL
-
-static __inline void
-mips_tlbp(void)
-{
-	__asm __volatile ("tlbp");
-	mips_barrier();
-#if 0
-	register_t ret;
-	register_t tmp;
-
-	__asm __volatile("mfc0	%0, $12\n" /* MIPS_COP_0_STATUS */
-	 		 "and	%1, %0, $~1\n" /* MIPS_SR_INT_IE */
-			 "mtc0	%1, $12\n" /* MIPS_COP_0_STATUS */
-			 : "=r" (ret), "=r" (tmp));
-	return (ret);
-#endif
-}
-
-static __inline void
-mips_tlbr(void)
-{
-	__asm __volatile ("tlbr");
-	mips_barrier();
-}
-
-static __inline void
-mips_tlbwi(void)
-{
-	__asm __volatile ("tlbwi");
-	mips_barrier();
-#if 0
-	__asm __volatile("mfc %0, $12\n" /* MIPS_COP_0_STATUS */
-	    		 "or  %0, %0, $1\n" /* MIPS_SR_INT_IE */
-			 "mtc0 %0, $12\n" /* MIPS_COP_0_STATUS */
-			 : "=r" (tmp));
-#endif
-}
-
-static __inline void
-mips_tlbwr(void)
-{
-	__asm __volatile ("tlbwr");
-	mips_barrier();
-}
-
-
-#if 0	/* XXX mips64 */
-
+#if defined(__mips_n32) || defined(__mips_n64)
 #define	MIPS_RDRW64_COP0(n,r)					\
 static __inline uint64_t					\
 mips_rd_ ## n (void)						\
@@ -152,10 +131,12 @@ mips_wr_ ## n (uint64_t a0)					\
 	mips_barrier();						\
 } struct __hack
 
+#if defined(__mips_n64)
 MIPS_RDRW64_COP0(entrylo0, MIPS_COP_0_TLB_LO0);
 MIPS_RDRW64_COP0(entrylo1, MIPS_COP_0_TLB_LO1);
 MIPS_RDRW64_COP0(entryhi, MIPS_COP_0_TLB_HI);
 MIPS_RDRW64_COP0(pagemask, MIPS_COP_0_TLB_PG_MASK);
+#endif
 MIPS_RDRW64_COP0(xcontext, MIPS_COP_0_TLB_XCONTEXT);
 
 #undef	MIPS_RDRW64_COP0
@@ -230,12 +211,15 @@ MIPS_RDRW32_COP0(cause, MIPS_COP_0_CAUSE);
 MIPS_RDRW32_COP0(status, MIPS_COP_0_STATUS);
 
 /* XXX: Some of these registers are specific to MIPS32. */
+#if !defined(__mips_n64)
 MIPS_RDRW32_COP0(entrylo0, MIPS_COP_0_TLB_LO0);
 MIPS_RDRW32_COP0(entrylo1, MIPS_COP_0_TLB_LO1);
-MIPS_RDRW32_COP0(entrylow, MIPS_COP_0_TLB_LOW);
 MIPS_RDRW32_COP0(entryhi, MIPS_COP_0_TLB_HI);
 MIPS_RDRW32_COP0(pagemask, MIPS_COP_0_TLB_PG_MASK);
+#endif
 MIPS_RDRW32_COP0(prid, MIPS_COP_0_PRID);
+/* XXX 64-bit?  */
+MIPS_RDRW32_COP0_SEL(prid, MIPS_COP_0_PRID, 1);
 MIPS_RDRW32_COP0(watchlo, MIPS_COP_0_WATCH_LO);
 MIPS_RDRW32_COP0_SEL(watchlo, MIPS_COP_0_WATCH_LO, 1);
 MIPS_RDRW32_COP0_SEL(watchlo, MIPS_COP_0_WATCH_LO, 2);
@@ -250,7 +234,6 @@ MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 1);
 MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 2);
 MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 3);
 
-
 #undef	MIPS_RDRW32_COP0
 
 static __inline register_t
@@ -261,7 +244,7 @@ intr_disable(void)
 	s = mips_rd_status();
 	mips_wr_status(s & ~MIPS_SR_INT_IE);
 
-	return (s);
+	return (s & MIPS_SR_INT_IE);
 }
 
 static __inline register_t
@@ -275,7 +258,13 @@ intr_enable(void)
 	return (s);
 }
 
-#define	intr_restore(s)	mips_wr_status((s))
+static __inline void
+intr_restore(register_t ie)
+{
+	if (ie == MIPS_SR_INT_IE) {
+		intr_enable();
+	}
+}
 
 static __inline void
 breakpoint(void)

From 11484eb34fb91bbaa34cb1b811443f2a6107ec70 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 16 Apr 2010 23:48:28 +0000
Subject: [PATCH 095/532] o) Use the direct map where possible for
 uiomove_fromphys, based on code from    sparc64. o) Use uiomove_fromphys
 rather than the broken fpage mechanism for /dev/mem. o) Update sf_buf
 allocator to not share buffers and to do a pmap_qremove when    done with an
 sf_buf so as to better track valid mappings.

---
 sys/mips/include/sf_buf.h   |  15 +--
 sys/mips/mips/mem.c         | 196 ++++++++++++++++--------------------
 sys/mips/mips/uio_machdep.c |  34 ++++---
 sys/mips/mips/vm_machdep.c  | 117 ++++++++-------------
 4 files changed, 150 insertions(+), 212 deletions(-)

diff --git a/sys/mips/include/sf_buf.h b/sys/mips/include/sf_buf.h
index 0a9980cf3d7..b6ee1cc20b6 100644
--- a/sys/mips/include/sf_buf.h
+++ b/sys/mips/include/sf_buf.h
@@ -1,5 +1,5 @@
 /*-
- * Copyright (c) 2003, 2005 Alan L. Cox <alc@cs.rice.edu>
+ * Copyright (c) 2003 Alan L. Cox <alc@cs.rice.edu>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -23,29 +23,20 @@
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  *
- *	from: src/sys/i386/include/sf_buf.h,v 1.4 2005/02/13 06:23:13 alc
  * $FreeBSD$
  */
 
 #ifndef _MACHINE_SF_BUF_H_
-#define	_MACHINE_SF_BUF_H_
+#define _MACHINE_SF_BUF_H_
 
 #include <sys/queue.h>
-#include <vm/vm.h>
-#include <vm/vm_param.h>
-#include <vm/vm_page.h>
 
 struct vm_page;
 
 struct sf_buf {
-	LIST_ENTRY(sf_buf) list_entry;	/* list of buffers */
-	TAILQ_ENTRY(sf_buf) free_entry; /* list of buffers */
+	SLIST_ENTRY(sf_buf) free_list;	/* list of free buffer slots */
 	struct		vm_page *m;	/* currently mapped page */
 	vm_offset_t	kva;		/* va of mapping */
-	int		ref_count;	/* usage of this mapping */
-#ifdef SMP
-	cpumask_t	cpumask;	/* cpus on which mapping is valid */
-#endif
 };
 
 static __inline vm_offset_t
diff --git a/sys/mips/mips/mem.c b/sys/mips/mips/mem.c
index c0e88e08e0c..a7f8244385a 100644
--- a/sys/mips/mips/mem.c
+++ b/sys/mips/mips/mem.c
@@ -1,13 +1,12 @@
-/*	$OpenBSD: mem.c,v 1.2 1998/08/31 17:42:34 millert Exp $ */
-/*	$NetBSD: mem.c,v 1.6 1995/04/10 11:55:03 mycroft Exp $	*/
-/*
+/*-
  * Copyright (c) 1988 University of Utah.
- * Copyright (c) 1982, 1986, 1990, 1993
- *	The Regents of the University of California.  All rights reserved.
+ * Copyright (c) 1982, 1986, 1990 The Regents of the University of California.
+ * All rights reserved.
  *
  * This code is derived from software contributed to Berkeley by
  * the Systems Programming Group of the University of Utah Computer
- * Science Department and Ralph Campbell.
+ * Science Department, and code derived from software contributed to
+ * Berkeley by William Jolitz.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
@@ -33,161 +32,136 @@
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  *
- *	@(#)mem.c	8.3 (Berkeley) 1/12/94
- *	JNPR: mem.c,v 1.3 2007/08/09 11:23:32 katta Exp $
- */
-
-/*
- * Memory special file
+ *	from: Utah $Hdr: mem.c 1.13 89/10/08$
+ *	from: @(#)mem.c	7.2 (Berkeley) 5/9/91
  */
 
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
+/*
+ * Memory special file
+ */
+
 #include <sys/param.h>
-#include <sys/kernel.h>
 #include <sys/conf.h>
+#include <sys/fcntl.h>
+#include <sys/kernel.h>
+#include <sys/lock.h>
+#include <sys/malloc.h>
+#include <sys/memrange.h>
+#include <sys/module.h>
+#include <sys/mutex.h>
 #include <sys/proc.h>
-#include <sys/signalvar.h>
-#include <vm/vm.h>
-#include <vm/vm_extern.h>
-#include <vm/pmap.h>
-#include <vm/vm_map.h>
-#include <sys/user.h>
 #include <sys/msgbuf.h>
 #include <sys/systm.h>
-#include <sys/systm.h>
-#include <sys/buf.h>
+#include <sys/signalvar.h>
 #include <sys/uio.h>
-#include <sys/sched.h>
-#include <sys/malloc.h>
-#include <machine/pte.h>
-#include <machine/cpu.h>
+
 #include <machine/md_var.h>
-#include <machine/atomic.h>
+#include <machine/vmparam.h>
+
+#include <vm/vm.h>
+#include <vm/pmap.h>
+#include <vm/vm_extern.h>
+#include <vm/vm_page.h>
+
 #include <machine/memdev.h>
 
+struct mem_range_softc mem_range_softc;
 
-extern struct sysmaps sysmaps_pcpu[];
-/*ARGSUSED*/
+/* ARGSUSED */
 int
-memrw(dev, uio, flags)
-	struct cdev *dev;
-	struct uio *uio;
-	int flags;
+memrw(struct cdev *dev, struct uio *uio, int flags)
 {
-	register vm_offset_t v;
-	register int c;
-	register struct iovec *iov;
+	struct iovec *iov;
 	int error = 0;
+	vm_offset_t va, eva, off, v;
+	vm_prot_t prot;
+	struct vm_page m;
+	vm_page_t marr;
+	vm_size_t cnt;
 
-	while (uio->uio_resid > 0 && error == 0) {
+	cnt = 0;
+	error = 0;
+
+	GIANT_REQUIRED;
+
+	while (uio->uio_resid > 0 && !error) {
 		iov = uio->uio_iov;
 		if (iov->iov_len == 0) {
 			uio->uio_iov++;
 			uio->uio_iovcnt--;
 			if (uio->uio_iovcnt < 0)
-				panic("mmrw");
+				panic("memrw");
 			continue;
 		}
-
-		/* minor device 0 is physical memory */
 		if (dev2unit(dev) == CDEV_MINOR_MEM) {
 			v = uio->uio_offset;
-			c = iov->iov_len;
 
-			vm_offset_t va;
-			vm_paddr_t pa;
-			register int o;
+			off = uio->uio_offset & PAGE_MASK;
+			cnt = PAGE_SIZE - ((vm_offset_t)iov->iov_base &
+			    PAGE_MASK);
+			cnt = min(cnt, PAGE_SIZE - off);
+			cnt = min(cnt, iov->iov_len);
 
-			if (is_cacheable_mem(v) &&
-			    is_cacheable_mem(v + c - 1)) {
-				struct fpage *fp;
-				struct sysmaps *sysmaps;
-
-				sysmaps = &sysmaps_pcpu[PCPU_GET(cpuid)];
-				mtx_lock(&sysmaps->lock);
-				sched_pin();
-
-				fp = &sysmaps->fp[PMAP_FPAGE1];
-				pa = uio->uio_offset & ~PAGE_MASK;
-				va = pmap_map_fpage(pa, fp, FALSE);
-				o = (int)uio->uio_offset & PAGE_MASK;
-				c = (u_int)(PAGE_SIZE -
-					    ((uintptr_t)iov->iov_base & PAGE_MASK));
-				c = min(c, (u_int)(PAGE_SIZE - o));
-				c = min(c, (u_int)iov->iov_len);
-				error = uiomove((caddr_t)(va + o), (int)c, uio);
-				pmap_unmap_fpage(pa, fp);
-				sched_unpin();
-				mtx_unlock(&sysmaps->lock);
-			} else
-				return (EFAULT);
-			continue;
+			m.phys_addr = trunc_page(v);
+			marr = &m;
+			error = uiomove_fromphys(&marr, off, cnt, uio);
 		}
-
-		/* minor device 1 is kernel memory */
 		else if (dev2unit(dev) == CDEV_MINOR_KMEM) {
-			v = uio->uio_offset;
-			c = min(iov->iov_len, MAXPHYS);
+			va = uio->uio_offset;
 
-			vm_offset_t addr, eaddr;
-			vm_offset_t wired_tlb_virtmem_end;
+			va = trunc_page(uio->uio_offset);
+			eva = round_page(uio->uio_offset
+			    + iov->iov_len);
 
-			/*
-			 * Make sure that all of the pages are currently
-			 * resident so that we don't create any zero-fill pages.
+			/* 
+			 * Make sure that all the pages are currently resident
+			 * so that we don't create any zero-fill pages.
 			 */
-			addr = trunc_page(uio->uio_offset);
-			eaddr = round_page(uio->uio_offset + c);
 
-			if (addr > (vm_offset_t) VM_MIN_KERNEL_ADDRESS) {
-				wired_tlb_virtmem_end = VM_MIN_KERNEL_ADDRESS +
-				    VM_KERNEL_ALLOC_OFFSET;
-				if ((addr < wired_tlb_virtmem_end) &&
-				    (eaddr >= wired_tlb_virtmem_end))
-					addr = wired_tlb_virtmem_end;
-
-				if (addr >= wired_tlb_virtmem_end) {
-					for (; addr < eaddr; addr += PAGE_SIZE) 
-						if (pmap_extract(kernel_pmap,
-						    addr) == 0)
-							return EFAULT;
-
-					if (!kernacc(
-					    (caddr_t)(uintptr_t)uio->uio_offset, c,
-					    uio->uio_rw == UIO_READ ?
-					    VM_PROT_READ : VM_PROT_WRITE))
-						return (EFAULT);
-				}
-			}
-			else if (MIPS_IS_KSEG0_ADDR(v)) {
-				if (MIPS_KSEG0_TO_PHYS(v + c) >= ctob(physmem))
+			for (; va < eva; va += PAGE_SIZE)
+				if (pmap_extract(kernel_pmap, va) == 0)
 					return (EFAULT);
-			}
-			else if (MIPS_IS_KSEG1_ADDR(v)) {
-				if (MIPS_KSEG1_TO_PHYS(v + c) >= ctob(physmem))
-					return (EFAULT);
-			}
-			else
+
+			prot = (uio->uio_rw == UIO_READ)
+			    ? VM_PROT_READ : VM_PROT_WRITE;
+
+			va = uio->uio_offset;
+			if (kernacc((void *) va, iov->iov_len, prot)
+			    == FALSE)
 				return (EFAULT);
 
+			error = uiomove((void *)va, iov->iov_len, uio);
 
-			error = uiomove((caddr_t)v, c, uio);
 			continue;
 		}
-
 	}
+
 	return (error);
 }
 
-/*ARGSUSED*/
+/*
+ * allow user processes to MMAP some memory sections
+ * instead of going through read/write
+ */
 int
-memmmap(struct cdev *dev, vm_ooffset_t off, vm_paddr_t *paddr,
+memmmap(struct cdev *dev, vm_ooffset_t offset, vm_paddr_t *paddr,
     int prot, vm_memattr_t *memattr)
 {
+	/*
+	 * /dev/mem is the only one that makes sense through this
+	 * interface.  For /dev/kmem any physaddr we return here
+	 * could be transient and hence incorrect or invalid at
+	 * a later time.
+	 */
+	if (dev2unit(dev) != CDEV_MINOR_MEM)
+		return (-1);
 
-	return (EOPNOTSUPP);
+	*paddr = offset;
+
+	return (0);
 }
 
 void
diff --git a/sys/mips/mips/uio_machdep.c b/sys/mips/mips/uio_machdep.c
index 0872b4d388e..10deff667e5 100644
--- a/sys/mips/mips/uio_machdep.c
+++ b/sys/mips/mips/uio_machdep.c
@@ -32,8 +32,7 @@
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
  *
- *	@(#)kern_subr.c 8.3 (Berkeley) 1/21/94
- *	from: src/sys/i386/i386/uio_machdep.c,v 1.8 2005/02/13 23:09:36 alc
+ *	@(#)kern_subr.c	8.3 (Berkeley) 1/21/94
  */
 
 #include <sys/cdefs.h>
@@ -44,17 +43,18 @@ __FBSDID("$FreeBSD$");
 #include <sys/lock.h>
 #include <sys/mutex.h>
 #include <sys/proc.h>
-#include <sys/sched.h>
 #include <sys/sf_buf.h>
 #include <sys/systm.h>
 #include <sys/uio.h>
 
 #include <vm/vm.h>
 #include <vm/vm_page.h>
+#include <vm/vm_param.h>
 
 /*
- * Implement uiomove(9) from physical memory using sf_bufs to reduce
- * the creation and destruction of ephemeral mappings.
+ * Implement uiomove(9) from physical memory using a combination
+ * of the direct mapping and sf_bufs to reduce the creation and
+ * destruction of ephemeral mappings.  
  */
 int
 uiomove_fromphys(vm_page_t ma[], vm_offset_t offset, int n, struct uio *uio)
@@ -64,6 +64,8 @@ uiomove_fromphys(vm_page_t ma[], vm_offset_t offset, int n, struct uio *uio)
 	struct iovec *iov;
 	void *cp;
 	vm_offset_t page_offset;
+	vm_paddr_t pa;
+	vm_page_t m;
 	size_t cnt;
 	int error = 0;
 	int save = 0;
@@ -85,10 +87,16 @@ uiomove_fromphys(vm_page_t ma[], vm_offset_t offset, int n, struct uio *uio)
 		if (cnt > n)
 			cnt = n;
 		page_offset = offset & PAGE_MASK;
-		cnt = min(cnt, PAGE_SIZE - page_offset);
-		sched_pin();
-		sf = sf_buf_alloc(ma[offset >> PAGE_SHIFT], SFB_CPUPRIVATE);
-		cp = (char *)sf_buf_kva(sf) + page_offset;
+		cnt = ulmin(cnt, PAGE_SIZE - page_offset);
+		m = ma[offset >> PAGE_SHIFT];
+		pa = VM_PAGE_TO_PHYS(m);
+		if (pa < MIPS_KSEG0_LARGEST_PHYS) {
+			cp = (char *)MIPS_PHYS_TO_KSEG0(pa);
+			sf = NULL;
+		} else {
+			sf = sf_buf_alloc(m, 0);
+			cp = (char *)sf_buf_kva(sf) + page_offset;
+		}
 		switch (uio->uio_segflg) {
 		case UIO_USERSPACE:
 			if (ticks - PCPU_GET(switchticks) >= hogticks)
@@ -98,8 +106,8 @@ uiomove_fromphys(vm_page_t ma[], vm_offset_t offset, int n, struct uio *uio)
 			else
 				error = copyin(iov->iov_base, cp, cnt);
 			if (error) {
-				sf_buf_free(sf);
-				sched_unpin();
+				if (sf != NULL)
+					sf_buf_free(sf);
 				goto out;
 			}
 			break;
@@ -112,8 +120,8 @@ uiomove_fromphys(vm_page_t ma[], vm_offset_t offset, int n, struct uio *uio)
 		case UIO_NOCOPY:
 			break;
 		}
-		sf_buf_free(sf);
-		sched_unpin();
+		if (sf != NULL)
+			sf_buf_free(sf);
 		iov->iov_base = (char *)iov->iov_base + cnt;
 		iov->iov_len -= cnt;
 		uio->uio_resid -= cnt;
diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index 57219f638f5..3b7a1a1b18c 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -56,6 +56,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/sysctl.h>
 #include <sys/unistd.h>
 
+#include <machine/asm.h>
 #include <machine/cache.h>
 #include <machine/clock.h>
 #include <machine/cpu.h>
@@ -63,12 +64,15 @@ __FBSDID("$FreeBSD$");
 #include <machine/pcb.h>
 
 #include <vm/vm.h>
-#include <vm/vm_param.h>
-#include <sys/lock.h>
-#include <vm/vm_kern.h>
-#include <vm/vm_page.h>
-#include <vm/vm_map.h>
 #include <vm/vm_extern.h>
+#include <vm/pmap.h>
+#include <vm/vm_kern.h>
+#include <vm/vm_map.h>
+#include <vm/vm_page.h>
+#include <vm/vm_pageout.h>
+#include <vm/vm_param.h>
+#include <vm/uma.h>
+#include <vm/uma_int.h>
 
 #include <sys/user.h>
 #include <sys/mbuf.h>
@@ -81,25 +85,17 @@ __FBSDID("$FreeBSD$");
 static void	sf_buf_init(void *arg);
 SYSINIT(sock_sf, SI_SUB_MBUF, SI_ORDER_ANY, sf_buf_init, NULL);
 
-LIST_HEAD(sf_head, sf_buf);
-
-
 /*
- * A hash table of active sendfile(2) buffers
+ * Expanded sf_freelist head.  Really an SLIST_HEAD() in disguise, with the
+ * sf_freelist head with the sf_lock mutex.
  */
-static struct sf_head *sf_buf_active;
-static u_long sf_buf_hashmask;
+static struct {
+	SLIST_HEAD(, sf_buf) sf_head;
+	struct mtx sf_lock;
+} sf_freelist;
 
-#define	SF_BUF_HASH(m)	(((m) - vm_page_array) & sf_buf_hashmask)
-
-static TAILQ_HEAD(, sf_buf) sf_buf_freelist;
 static u_int	sf_buf_alloc_want;
 
-/*
- * A lock used to synchronize access to the hash table and free list
- */
-static struct mtx sf_buf_lock;
-
 /*
  * Finish a fork operation, with process p2 nearly set up.
  * Copy and update the pcb, set up the stack so that the child
@@ -471,56 +467,34 @@ sf_buf_init(void *arg)
 	nsfbufs = NSFBUFS;
 	TUNABLE_INT_FETCH("kern.ipc.nsfbufs", &nsfbufs);
 
-	sf_buf_active = hashinit(nsfbufs, M_TEMP, &sf_buf_hashmask);
-	TAILQ_INIT(&sf_buf_freelist);
+	mtx_init(&sf_freelist.sf_lock, "sf_bufs list lock", NULL, MTX_DEF);
+	SLIST_INIT(&sf_freelist.sf_head);
 	sf_base = kmem_alloc_nofault(kernel_map, nsfbufs * PAGE_SIZE);
 	sf_bufs = malloc(nsfbufs * sizeof(struct sf_buf), M_TEMP,
 	    M_NOWAIT | M_ZERO);
 	for (i = 0; i < nsfbufs; i++) {
 		sf_bufs[i].kva = sf_base + i * PAGE_SIZE;
-		TAILQ_INSERT_TAIL(&sf_buf_freelist, &sf_bufs[i], free_entry);
+		SLIST_INSERT_HEAD(&sf_freelist.sf_head, &sf_bufs[i], free_list);
 	}
 	sf_buf_alloc_want = 0;
-	mtx_init(&sf_buf_lock, "sf_buf", NULL, MTX_DEF);
 }
 
 /*
- * Allocate an sf_buf for the given vm_page.  On this machine, however, there
- * is no sf_buf object.	 Instead, an opaque pointer to the given vm_page is
- * returned.
+ * Get an sf_buf from the freelist.  Will block if none are available.
  */
 struct sf_buf *
 sf_buf_alloc(struct vm_page *m, int flags)
 {
-	struct sf_head *hash_list;
 	struct sf_buf *sf;
 	int error;
 
-	hash_list = &sf_buf_active[SF_BUF_HASH(m)];
-	mtx_lock(&sf_buf_lock);
-	LIST_FOREACH(sf, hash_list, list_entry) {
-		if (sf->m == m) {
-			sf->ref_count++;
-			if (sf->ref_count == 1) {
-				TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
-				nsfbufsused++;
-				nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
-			}
-			/*
-			 * Flush all mappings in order to have up to date 
-			 * physycal memory
-			 */
-			pmap_flush_pvcache(sf->m);
-			mips_dcache_inv_range(sf->kva, PAGE_SIZE);
-			goto done;
-		}
-	}
-	while ((sf = TAILQ_FIRST(&sf_buf_freelist)) == NULL) {
+	mtx_lock(&sf_freelist.sf_lock);
+	while ((sf = SLIST_FIRST(&sf_freelist.sf_head)) == NULL) {
 		if (flags & SFB_NOWAIT)
-			goto done;
+			break;
 		sf_buf_alloc_want++;
 		mbstat.sf_allocwait++;
-		error = msleep(&sf_buf_freelist, &sf_buf_lock,
+		error = msleep(&sf_freelist, &sf_freelist.sf_lock,
 		    (flags & SFB_CATCH) ? PCATCH | PVM : PVM, "sfbufa", 0);
 		sf_buf_alloc_want--;
 
@@ -528,42 +502,33 @@ sf_buf_alloc(struct vm_page *m, int flags)
 		 * If we got a signal, don't risk going back to sleep.
 		 */
 		if (error)
-			goto done;
+			break;
 	}
-	TAILQ_REMOVE(&sf_buf_freelist, sf, free_entry);
-	if (sf->m != NULL)
-		LIST_REMOVE(sf, list_entry);
-	LIST_INSERT_HEAD(hash_list, sf, list_entry);
-	sf->ref_count = 1;
-	sf->m = m;
-	nsfbufsused++;
-	nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
-	pmap_qenter(sf->kva, &sf->m, 1);
-done:
-	mtx_unlock(&sf_buf_lock);
+	if (sf != NULL) {
+		SLIST_REMOVE_HEAD(&sf_freelist.sf_head, free_list);
+		sf->m = m;
+		nsfbufsused++;
+		nsfbufspeak = imax(nsfbufspeak, nsfbufsused);
+		pmap_qenter(sf->kva, &sf->m, 1);
+	}
+	mtx_unlock(&sf_freelist.sf_lock);
 	return (sf);
 }
 
 /*
- * Free the sf_buf.  In fact, do nothing because there are no resources
- * associated with the sf_buf.
+ * Release resources back to the system.
  */
 void
 sf_buf_free(struct sf_buf *sf)
 {
-	mtx_lock(&sf_buf_lock);
-	sf->ref_count--;
-	/*
-	 * Make sure all changes in KVA end up in physical memory
-	 */
-	mips_dcache_wbinv_range(sf->kva, PAGE_SIZE);
-	if (sf->ref_count == 0) {
-		TAILQ_INSERT_TAIL(&sf_buf_freelist, sf, free_entry);
-		nsfbufsused--;
-		if (sf_buf_alloc_want > 0)
-			wakeup_one(&sf_buf_freelist);
-	}
-	mtx_unlock(&sf_buf_lock);
+
+	pmap_qremove(sf->kva, 1);
+	mtx_lock(&sf_freelist.sf_lock);
+	SLIST_INSERT_HEAD(&sf_freelist.sf_head, sf, free_list);
+	nsfbufsused--;
+	if (sf_buf_alloc_want > 0)
+		wakeup_one(&sf_freelist);
+	mtx_unlock(&sf_freelist.sf_lock);
 }
 
 /*

From 2ae5ecf8a2070f3235255bcb39df8d241eef834f Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 16 Apr 2010 23:54:56 +0000
Subject: [PATCH 096/532] Adjust limits and formats for ABIs with 64-bit longs.

---
 sys/mips/include/_inttypes.h | 94 +++++++++++++++++++-----------------
 sys/mips/include/_limits.h   |  8 ++-
 2 files changed, 56 insertions(+), 46 deletions(-)

diff --git a/sys/mips/include/_inttypes.h b/sys/mips/include/_inttypes.h
index 7b14b9b83cb..79664e4f383 100644
--- a/sys/mips/include/_inttypes.h
+++ b/sys/mips/include/_inttypes.h
@@ -38,177 +38,183 @@
  * Macros for format specifiers.
  */
 
+#if defined(__mips_n64)
+#define	PRI64		"l"
+#else
+#define	PRI64		"ll"
+#endif
+
 /* fprintf(3) macros for signed integers. */
 
 #define	PRId8		"d"	/* int8_t */
 #define	PRId16		"d"	/* int16_t */
 #define	PRId32		"d"	/* int32_t */
-#define	PRId64		"lld"	/* int64_t */
+#define	PRId64		PRI64"d"	/* int64_t */
 #define	PRIdLEAST8	"d"	/* int_least8_t */
 #define	PRIdLEAST16	"d"	/* int_least16_t */
 #define	PRIdLEAST32	"d"	/* int_least32_t */
-#define	PRIdLEAST64	"lld"	/* int_least64_t */
+#define	PRIdLEAST64	PRI64"d"	/* int_least64_t */
 #define	PRIdFAST8	"d"	/* int_fast8_t */
 #define	PRIdFAST16	"d"	/* int_fast16_t */
 #define	PRIdFAST32	"d"	/* int_fast32_t */
-#define	PRIdFAST64	"lld"	/* int_fast64_t */
+#define	PRIdFAST64	PRI64"d"	/* int_fast64_t */
 #define	PRIdMAX		"jd"	/* intmax_t */
-#define	PRIdPTR		"d"	/* intptr_t */
+#define	PRIdPTR		"ld"	/* intptr_t */
 
 #define	PRIi8		"i"	/* int8_t */
 #define	PRIi16		"i"	/* int16_t */
 #define	PRIi32		"i"	/* int32_t */
-#define	PRIi64		"lli"	/* int64_t */
+#define	PRIi64		PRI64"i"	/* int64_t */
 #define	PRIiLEAST8	"i"	/* int_least8_t  */
 #define	PRIiLEAST16	"i"	/* int_least16_t */
 #define	PRIiLEAST32	"i"	/* int_least32_t */
-#define	PRIiLEAST64	"lli"	/* int_least64_t */
+#define	PRIiLEAST64	PRI64"i"	/* int_least64_t */
 #define	PRIiFAST8	"i"	/* int_fast8_t */
 #define	PRIiFAST16	"i"	/* int_fast16_t */
 #define	PRIiFAST32	"i"	/* int_fast32_t */
-#define	PRIiFAST64	"lli"	/* int_fast64_t */
+#define	PRIiFAST64	PRI64"i"	/* int_fast64_t */
 #define	PRIiMAX		"ji"	/* intmax_t */
-#define	PRIiPTR		"i"	/* intptr_t */
+#define	PRIiPTR		"li"	/* intptr_t */
 
 /* fprintf(3) macros for unsigned integers. */
 
 #define	PRIo8		"o"	/* uint8_t */
 #define	PRIo16		"o"	/* uint16_t */
 #define	PRIo32		"o"	/* uint32_t */
-#define	PRIo64		"llo"	/* uint64_t */
+#define	PRIo64		PRI64"o"	/* uint64_t */
 #define	PRIoLEAST8	"o"	/* uint_least8_t */
 #define	PRIoLEAST16	"o"	/* uint_least16_t */
 #define	PRIoLEAST32	"o"	/* uint_least32_t */
-#define	PRIoLEAST64	"llo"	/* uint_least64_t */
+#define	PRIoLEAST64	PRI64"o"	/* uint_least64_t */
 #define	PRIoFAST8	"o"	/* uint_fast8_t */
 #define	PRIoFAST16	"o"	/* uint_fast16_t */
 #define	PRIoFAST32	"o"	/* uint_fast32_t */
-#define	PRIoFAST64	"llo"	/* uint_fast64_t */
+#define	PRIoFAST64	PRI64"o"	/* uint_fast64_t */
 #define	PRIoMAX		"jo"	/* uintmax_t */
-#define	PRIoPTR		"o"	/* uintptr_t */
+#define	PRIoPTR		"lo"	/* uintptr_t */
 
 #define	PRIu8		"u"	/* uint8_t */
 #define	PRIu16		"u"	/* uint16_t */
 #define	PRIu32		"u"	/* uint32_t */
-#define	PRIu64		"llu"	/* uint64_t */
+#define	PRIu64		PRI64"u"	/* uint64_t */
 #define	PRIuLEAST8	"u"	/* uint_least8_t */
 #define	PRIuLEAST16	"u"	/* uint_least16_t */
 #define	PRIuLEAST32	"u"	/* uint_least32_t */
-#define	PRIuLEAST64	"llu"	/* uint_least64_t */
+#define	PRIuLEAST64	PRI64"u"	/* uint_least64_t */
 #define	PRIuFAST8	"u"	/* uint_fast8_t */
 #define	PRIuFAST16	"u"	/* uint_fast16_t */
 #define	PRIuFAST32	"u"	/* uint_fast32_t */
-#define	PRIuFAST64	"llu"	/* uint_fast64_t */
+#define	PRIuFAST64	PRI64"u"	/* uint_fast64_t */
 #define	PRIuMAX		"ju"	/* uintmax_t */
-#define	PRIuPTR		"u"	/* uintptr_t */
+#define	PRIuPTR		"lu"	/* uintptr_t */
 
 #define	PRIx8		"x"	/* uint8_t */
 #define	PRIx16		"x"	/* uint16_t */
 #define	PRIx32		"x"	/* uint32_t */
-#define	PRIx64		"llx"	/* uint64_t */
+#define	PRIx64		PRI64"x"	/* uint64_t */
 #define	PRIxLEAST8	"x"	/* uint_least8_t */
 #define	PRIxLEAST16	"x"	/* uint_least16_t */
 #define	PRIxLEAST32	"x"	/* uint_least32_t */
-#define	PRIxLEAST64	"llx"	/* uint_least64_t */
+#define	PRIxLEAST64	PRI64"x"	/* uint_least64_t */
 #define	PRIxFAST8	"x"	/* uint_fast8_t */
 #define	PRIxFAST16	"x"	/* uint_fast16_t */
 #define	PRIxFAST32	"x"	/* uint_fast32_t */
-#define	PRIxFAST64	"llx"	/* uint_fast64_t */
+#define	PRIxFAST64	PRI64"x"	/* uint_fast64_t */
 #define	PRIxMAX		"jx"	/* uintmax_t */
-#define	PRIxPTR		"x"	/* uintptr_t */
+#define	PRIxPTR		"lx"	/* uintptr_t */
 
 #define	PRIX8		"X"	/* uint8_t */
 #define	PRIX16		"X"	/* uint16_t */
 #define	PRIX32		"X"	/* uint32_t */
-#define	PRIX64		"llX"	/* uint64_t */
+#define	PRIX64		PRI64"X"	/* uint64_t */
 #define	PRIXLEAST8	"X"	/* uint_least8_t */
 #define	PRIXLEAST16	"X"	/* uint_least16_t */
 #define	PRIXLEAST32	"X"	/* uint_least32_t */
-#define	PRIXLEAST64	"llX"	/* uint_least64_t */
+#define	PRIXLEAST64	PRI64"X"	/* uint_least64_t */
 #define	PRIXFAST8	"X"	/* uint_fast8_t */
 #define	PRIXFAST16	"X"	/* uint_fast16_t */
 #define	PRIXFAST32	"X"	/* uint_fast32_t */
-#define	PRIXFAST64	"llX"	/* uint_fast64_t */
+#define	PRIXFAST64	PRI64"X"	/* uint_fast64_t */
 #define	PRIXMAX		"jX"	/* uintmax_t */
-#define	PRIXPTR		"X"	/* uintptr_t */
+#define	PRIXPTR		"lX"	/* uintptr_t */
 
 /* fscanf(3) macros for signed integers. */
 
 #define	SCNd8		"hhd"	/* int8_t */
 #define	SCNd16		"hd"	/* int16_t */
 #define	SCNd32		"d"	/* int32_t */
-#define	SCNd64		"lld"	/* int64_t */
+#define	SCNd64		PRI64"d"	/* int64_t */
 #define	SCNdLEAST8	"hhd"	/* int_least8_t */
 #define	SCNdLEAST16	"hd"	/* int_least16_t */
 #define	SCNdLEAST32	"d"	/* int_least32_t */
-#define	SCNdLEAST64	"lld"	/* int_least64_t */
+#define	SCNdLEAST64	PRI64"d"	/* int_least64_t */
 #define	SCNdFAST8	"d"	/* int_fast8_t */
 #define	SCNdFAST16	"d"	/* int_fast16_t */
 #define	SCNdFAST32	"d"	/* int_fast32_t */
-#define	SCNdFAST64	"lld"	/* int_fast64_t */
+#define	SCNdFAST64	PRI64"d"	/* int_fast64_t */
 #define	SCNdMAX		"jd"	/* intmax_t */
-#define	SCNdPTR		"d"	/* intptr_t */
+#define	SCNdPTR		"ld"	/* intptr_t */
 
 #define	SCNi8		"hhi"	/* int8_t */
 #define	SCNi16		"hi"	/* int16_t */
 #define	SCNi32		"i"	/* int32_t */
-#define	SCNi64		"lli"	/* int64_t */
+#define	SCNi64		PRI64"i"	/* int64_t */
 #define	SCNiLEAST8	"hhi"	/* int_least8_t */
 #define	SCNiLEAST16	"hi"	/* int_least16_t */
 #define	SCNiLEAST32	"i"	/* int_least32_t */
-#define	SCNiLEAST64	"lli"	/* int_least64_t */
+#define	SCNiLEAST64	PRI64"i"	/* int_least64_t */
 #define	SCNiFAST8	"i"	/* int_fast8_t */
 #define	SCNiFAST16	"i"	/* int_fast16_t */
 #define	SCNiFAST32	"i"	/* int_fast32_t */
-#define	SCNiFAST64	"lli"	/* int_fast64_t */
+#define	SCNiFAST64	PRI64"i"	/* int_fast64_t */
 #define	SCNiMAX		"ji"	/* intmax_t */
-#define	SCNiPTR		"i"	/* intptr_t */
+#define	SCNiPTR		"li"	/* intptr_t */
 
 /* fscanf(3) macros for unsigned integers. */
 
 #define	SCNo8		"hho"	/* uint8_t */
 #define	SCNo16		"ho"	/* uint16_t */
 #define	SCNo32		"o"	/* uint32_t */
-#define	SCNo64		"llo"	/* uint64_t */
+#define	SCNo64		PRI64"o"	/* uint64_t */
 #define	SCNoLEAST8	"hho"	/* uint_least8_t */
 #define	SCNoLEAST16	"ho"	/* uint_least16_t */
 #define	SCNoLEAST32	"o"	/* uint_least32_t */
-#define	SCNoLEAST64	"llo"	/* uint_least64_t */
+#define	SCNoLEAST64	PRI64"o"	/* uint_least64_t */
 #define	SCNoFAST8	"o"	/* uint_fast8_t */
 #define	SCNoFAST16	"o"	/* uint_fast16_t */
 #define	SCNoFAST32	"o"	/* uint_fast32_t */
-#define	SCNoFAST64	"llo"	/* uint_fast64_t */
+#define	SCNoFAST64	PRI64"o"	/* uint_fast64_t */
 #define	SCNoMAX		"jo"	/* uintmax_t */
-#define	SCNoPTR		"o"	/* uintptr_t */
+#define	SCNoPTR		"lo"	/* uintptr_t */
 
 #define	SCNu8		"hhu"	/* uint8_t */
 #define	SCNu16		"hu"	/* uint16_t */
 #define	SCNu32		"u"	/* uint32_t */
-#define	SCNu64		"llu"	/* uint64_t */
+#define	SCNu64		PRI64"u"	/* uint64_t */
 #define	SCNuLEAST8	"hhu"	/* uint_least8_t */
 #define	SCNuLEAST16	"hu"	/* uint_least16_t */
 #define	SCNuLEAST32	"u"	/* uint_least32_t */
-#define	SCNuLEAST64	"llu"	/* uint_least64_t */
+#define	SCNuLEAST64	PRI64"u"	/* uint_least64_t */
 #define	SCNuFAST8	"u"	/* uint_fast8_t */
 #define	SCNuFAST16	"u"	/* uint_fast16_t */
 #define	SCNuFAST32	"u"	/* uint_fast32_t */
-#define	SCNuFAST64	"llu"	/* uint_fast64_t */
+#define	SCNuFAST64	PRI64"u"	/* uint_fast64_t */
 #define	SCNuMAX		"ju"	/* uintmax_t */
-#define	SCNuPTR		"u"	/* uintptr_t */
+#define	SCNuPTR		"lu"	/* uintptr_t */
 
 #define	SCNx8		"hhx"	/* uint8_t */
 #define	SCNx16		"hx"	/* uint16_t */
 #define	SCNx32		"x"	/* uint32_t */
-#define	SCNx64		"llx"	/* uint64_t */
+#define	SCNx64		PRI64"x"	/* uint64_t */
 #define	SCNxLEAST8	"hhx"	/* uint_least8_t */
 #define	SCNxLEAST16	"hx"	/* uint_least16_t */
 #define	SCNxLEAST32	"x"	/* uint_least32_t */
-#define	SCNxLEAST64	"llx"	/* uint_least64_t */
+#define	SCNxLEAST64	PRI64"x"	/* uint_least64_t */
 #define	SCNxFAST8	"x"	/* uint_fast8_t */
 #define	SCNxFAST16	"x"	/* uint_fast16_t */
 #define	SCNxFAST32	"x"	/* uint_fast32_t */
-#define	SCNxFAST64	"llx"	/* uint_fast64_t */
+#define	SCNxFAST64	PRI64"x"	/* uint_fast64_t */
 #define	SCNxMAX		"jx"	/* uintmax_t */
-#define	SCNxPTR		"x"	/* uintptr_t */
+#define	SCNxPTR		"lx"	/* uintptr_t */
 
 #endif /* !_MACHINE_INTTYPES_H_ */
diff --git a/sys/mips/include/_limits.h b/sys/mips/include/_limits.h
index d544305ce8e..e160a7f4a1d 100644
--- a/sys/mips/include/_limits.h
+++ b/sys/mips/include/_limits.h
@@ -34,6 +34,10 @@
 #ifndef _MACHINE__LIMITS_H_
 #define	_MACHINE__LIMITS_H_
 
+#if _MIPS_SZLONG == 64
+#define	_LARGE_LONG
+#endif
+
 /*
  * According to ANSI (section 2.2.4.2), the values below must be usable by
  * #if preprocessing directives.  Additionally, the expression must have the
@@ -76,9 +80,9 @@
 #define	__LLONG_MAX	0x7fffffffffffffffLL	/* max value for a long long */
 #define	__LLONG_MIN	(-0x7fffffffffffffffLL - 1)  /* min for a long long */
 
-#define	__SSIZE_MAX	__INT_MAX	/* max value for a ssize_t */
+#define	__SSIZE_MAX	__LONG_MAX	/* max value for a ssize_t */
 
-#define	__SIZE_T_MAX	__UINT_MAX	/* max value for a size_t */
+#define	__SIZE_T_MAX	__ULONG_MAX	/* max value for a size_t */
 
 #define	__OFF_MAX	__LLONG_MAX	/* max value for an off_t */
 #define	__OFF_MIN	__LLONG_MIN	/* min value for an off_t */

From e586cf9a2847c8f7f1d61cfef3b531536e0bb219 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 00:05:22 +0000
Subject: [PATCH 097/532] o) Remove code related to VM_ALLOC_WIRED_TLB_PG_POOL,
 VM_KERNEL_ALLOC_OFFSET    and floating pages.  They are unused and
 unsupported.

---
 sys/mips/include/pmap.h    |  37 ----
 sys/mips/include/vmparam.h |   6 +-
 sys/mips/mips/genassym.c   |   1 -
 sys/mips/mips/machdep.c    |   4 -
 sys/mips/mips/pmap.c       | 351 +++++++------------------------------
 sys/mips/mips/swtch.S      |   2 +-
 6 files changed, 61 insertions(+), 340 deletions(-)

diff --git a/sys/mips/include/pmap.h b/sys/mips/include/pmap.h
index 51ff5512e79..80772d97ee9 100644
--- a/sys/mips/include/pmap.h
+++ b/sys/mips/include/pmap.h
@@ -160,14 +160,8 @@ typedef struct pv_entry {
 extern vm_offset_t phys_avail[PHYS_AVAIL_ENTRIES + 2];
 extern vm_offset_t physmem_desc[PHYS_AVAIL_ENTRIES + 2];
 
-extern char *ptvmmap;		/* poor name! */
 extern vm_offset_t virtual_avail;
 extern vm_offset_t virtual_end;
-extern pd_entry_t *segbase;
-
-extern vm_paddr_t mips_wired_tlb_physmem_start;
-extern vm_paddr_t mips_wired_tlb_physmem_end;
-extern u_int need_wired_tlb_page_pool;
 
 #define	pmap_page_get_memattr(m)	VM_MEMATTR_DEFAULT
 #define	pmap_page_is_mapped(m)	(!TAILQ_EMPTY(&(m)->md.pv_list))
@@ -188,37 +182,6 @@ int pmap_compute_pages_to_dump(void);
 void pmap_update_page(pmap_t pmap, vm_offset_t va, pt_entry_t pte);
 void pmap_flush_pvcache(vm_page_t m);
 
-/*
- * floating virtual pages (FPAGES)
- *
- * These are the reserved virtual memory areas which can be
- * mapped to any physical memory.
- */
-#define	FPAGES			2
-#define	FPAGES_SHARED		2
-#define	FSPACE			((FPAGES * MAXCPU + FPAGES_SHARED)  * PAGE_SIZE)
-#define	PMAP_FPAGE1		0x00	/* Used by pmap_zero_page &
-					 * pmap_copy_page */
-#define	PMAP_FPAGE2		0x01	/* Used by pmap_copy_page */
-
-#define	PMAP_FPAGE3		0x00	/* Used by pmap_zero_page_idle */
-#define	PMAP_FPAGE_KENTER_TEMP	0x01	/* Used by coredump */
-
-struct fpage {
-	vm_offset_t kva;
-	u_int state;
-};
-
-struct sysmaps {
-	struct mtx lock;
-	struct fpage fp[FPAGES];
-};
-
-vm_offset_t 
-pmap_map_fpage(vm_paddr_t pa, struct fpage *fp,
-    boolean_t check_unmaped);
-void pmap_unmap_fpage(vm_paddr_t pa, struct fpage *fp);
-
 /*
  * Function to save TLB contents so that they may be inspected in the debugger.
  */
diff --git a/sys/mips/include/vmparam.h b/sys/mips/include/vmparam.h
index a5242934019..d4bc7fa07b0 100644
--- a/sys/mips/include/vmparam.h
+++ b/sys/mips/include/vmparam.h
@@ -101,12 +101,8 @@
 #define	VM_MAX_MMAP_ADDR	VM_MAXUSER_ADDRESS
 #define	VM_MAX_ADDRESS		((vm_offset_t)0x80000000)
 
-#ifndef VM_KERNEL_ALLOC_OFFSET
-#define	VM_KERNEL_ALLOC_OFFSET	((vm_offset_t)0x00000000)
-#endif
-
 #define	VM_MIN_KERNEL_ADDRESS		((vm_offset_t)0xC0000000)
-#define	VM_KERNEL_WIRED_ADDR_END	(VM_MIN_KERNEL_ADDRESS + VM_KERNEL_ALLOC_OFFSET)
+#define	VM_KERNEL_WIRED_ADDR_END	(VM_MIN_KERNEL_ADDRESS)
 #define	VM_MAX_KERNEL_ADDRESS	((vm_offset_t)0xFFFFC000)
 
 /*
diff --git a/sys/mips/mips/genassym.c b/sys/mips/mips/genassym.c
index 62da8f1f501..f59ef6854ba 100644
--- a/sys/mips/mips/genassym.c
+++ b/sys/mips/mips/genassym.c
@@ -87,7 +87,6 @@ ASSYM(PC_CURPMAP, offsetof(struct pcpu, pc_curpmap));
 
 ASSYM(VM_MAX_KERNEL_ADDRESS, VM_MAX_KERNEL_ADDRESS);
 ASSYM(VM_MAXUSER_ADDRESS, VM_MAXUSER_ADDRESS);
-ASSYM(VM_KERNEL_ALLOC_OFFSET, VM_KERNEL_ALLOC_OFFSET);
 ASSYM(SIGF_UC, offsetof(struct sigframe, sf_uc));
 ASSYM(SIGFPE, SIGFPE);
 ASSYM(PAGE_SHIFT, PAGE_SHIFT);
diff --git a/sys/mips/mips/machdep.c b/sys/mips/mips/machdep.c
index ad1dab52206..4f728d56c05 100644
--- a/sys/mips/mips/machdep.c
+++ b/sys/mips/mips/machdep.c
@@ -142,10 +142,6 @@ vm_offset_t physmem_desc[PHYS_AVAIL_ENTRIES + 2];
 struct platform platform;
 #endif
 
-vm_paddr_t	mips_wired_tlb_physmem_start;
-vm_paddr_t	mips_wired_tlb_physmem_end;
-u_int		need_wired_tlb_page_pool;
-
 static void cpu_startup(void *);
 SYSINIT(cpu, SI_SUB_CPU, SI_ORDER_FIRST, cpu_startup, NULL);
 
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index d5765ea9b0e..9f01f8a88c4 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -161,10 +161,6 @@ static uma_zone_t pvzone;
 static struct vm_object pvzone_obj;
 static int pv_entry_count = 0, pv_entry_max = 0, pv_entry_high_water = 0;
 
-struct fpage fpages_shared[FPAGES_SHARED];
-
-struct sysmaps sysmaps_pcpu[MAXCPU];
-
 static PMAP_INLINE void free_pv_entry(pv_entry_t pv);
 static pv_entry_t get_pv_entry(pmap_t locked_pmap);
 static __inline void pmap_changebit(vm_page_t m, int bit, boolean_t setem);
@@ -188,7 +184,6 @@ static int pmap_unuse_pt(pmap_t, vm_offset_t, vm_page_t);
 static int init_pte_prot(vm_offset_t va, vm_page_t m, vm_prot_t prot);
 static void pmap_TLB_invalidate_kernel(vm_offset_t);
 static void pmap_TLB_update_kernel(vm_offset_t, pt_entry_t);
-static void pmap_init_fpage(void);
 
 #ifdef SMP
 static void pmap_invalidate_page_action(void *arg);
@@ -355,7 +350,7 @@ again:
 	kstack0 = pmap_steal_memory(KSTACK_PAGES << PAGE_SHIFT);
 
 
-	virtual_avail = VM_MIN_KERNEL_ADDRESS + VM_KERNEL_ALLOC_OFFSET;
+	virtual_avail = VM_MIN_KERNEL_ADDRESS;
 	virtual_end = VM_MAX_KERNEL_ADDRESS;
 
 #ifdef SMP
@@ -480,8 +475,6 @@ void
 pmap_init(void)
 {
 
-	if (need_wired_tlb_page_pool)
-		pmap_init_fpage();
 	/*
 	 * Initialize the address space (zone) for the pv entries.  Set a
 	 * high water mark so that the system can recover from excessive
@@ -805,136 +798,6 @@ pmap_qremove(vm_offset_t va, int count)
  * Page table page management routines.....
  ***************************************************/
 
-/*
- * floating pages (FPAGES) management routines
- *
- * FPAGES are the reserved virtual memory areas which can be
- * mapped to any physical memory. This gets used typically
- * in the following functions:
- *
- * pmap_zero_page
- * pmap_copy_page
- */
-
-/*
- * Create the floating pages, aka FPAGES!
- */
-static void
-pmap_init_fpage()
-{
-	vm_offset_t kva;
-	int i, j;
-	struct sysmaps *sysmaps;
-
-	/*
-	 * We allocate a total of (FPAGES*MAXCPU + FPAGES_SHARED + 1) pages
-	 * at first. FPAGES & FPAGES_SHARED should be EVEN Then we'll adjust
-	 * 'kva' to be even-page aligned so that the fpage area can be wired
-	 * in the TLB with a single TLB entry.
-	 */
-	kva = kmem_alloc_nofault(kernel_map,
-	    (FPAGES * MAXCPU + 1 + FPAGES_SHARED) * PAGE_SIZE);
-	if ((void *)kva == NULL)
-		panic("pmap_init_fpage: fpage allocation failed");
-
-	/*
-	 * Make up start at an even page number so we can wire down the
-	 * fpage area in the tlb with a single tlb entry.
-	 */
-	if ((((vm_offset_t)kva) >> PGSHIFT) & 1) {
-		/*
-		 * 'kva' is not even-page aligned. Adjust it and free the
-		 * first page which is unused.
-		 */
-		kmem_free(kernel_map, (vm_offset_t)kva, NBPG);
-		kva = ((vm_offset_t)kva) + NBPG;
-	} else {
-		/*
-		 * 'kva' is even page aligned. We don't need the last page,
-		 * free it.
-		 */
-		kmem_free(kernel_map, ((vm_offset_t)kva) + FSPACE, NBPG);
-	}
-
-	for (i = 0; i < MAXCPU; i++) {
-		sysmaps = &sysmaps_pcpu[i];
-		mtx_init(&sysmaps->lock, "SYSMAPS", NULL, MTX_DEF);
-
-		/* Assign FPAGES pages to the CPU */
-		for (j = 0; j < FPAGES; j++)
-			sysmaps->fp[j].kva = kva + (j) * PAGE_SIZE;
-		kva = ((vm_offset_t)kva) + (FPAGES * PAGE_SIZE);
-	}
-
-	/*
-	 * An additional 2 pages are needed, one for pmap_zero_page_idle()
-	 * and one for coredump. These pages are shared by all cpu's
-	 */
-	fpages_shared[PMAP_FPAGE3].kva = kva;
-	fpages_shared[PMAP_FPAGE_KENTER_TEMP].kva = kva + PAGE_SIZE;
-}
-
-/*
- * Map the page to the fpage virtual address as specified thru' fpage id
- */
-vm_offset_t
-pmap_map_fpage(vm_paddr_t pa, struct fpage *fp, boolean_t check_unmaped)
-{
-	vm_offset_t kva;
-	register pt_entry_t *pte;
-	pt_entry_t npte;
-
-	KASSERT(curthread->td_pinned > 0, ("curthread not pinned"));
-	/*
-	 * Check if the fpage is free
-	 */
-	if (fp->state) {
-		if (check_unmaped == TRUE)
-			pmap_unmap_fpage(pa, fp);
-		else
-			panic("pmap_map_fpage: fpage is busy");
-	}
-	fp->state = TRUE;
-	kva = fp->kva;
-
-	npte = mips_paddr_to_tlbpfn(pa) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-	pte = pmap_pte(kernel_pmap, kva);
-	*pte = npte;
-
-	pmap_TLB_update_kernel(kva, npte);
-
-	return (kva);
-}
-
-/*
- * Unmap the page from the fpage virtual address as specified thru' fpage id
- */
-void
-pmap_unmap_fpage(vm_paddr_t pa, struct fpage *fp)
-{
-	vm_offset_t kva;
-	register pt_entry_t *pte;
-
-	KASSERT(curthread->td_pinned > 0, ("curthread not pinned"));
-	/*
-	 * Check if the fpage is busy
-	 */
-	if (!(fp->state)) {
-		panic("pmap_unmap_fpage: fpage is free");
-	}
-	kva = fp->kva;
-
-	pte = pmap_pte(kernel_pmap, kva);
-	*pte = PTE_G;
-	pmap_TLB_invalidate_kernel(kva);
-
-	fp->state = FALSE;
-
-	/*
-	 * Should there be any flush operation at the end?
-	 */
-}
-
 /*  Revision 1.507
  *
  * Simplify the reference counting of page table pages.	 Specifically, use
@@ -1051,10 +914,6 @@ pmap_pinit(pmap_t pmap)
 	req = VM_ALLOC_NOOBJ | VM_ALLOC_NORMAL | VM_ALLOC_WIRED |
 	    VM_ALLOC_ZERO;
 
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool)
-		req |= VM_ALLOC_WIRED_TLB_PG_POOL;
-#endif
 	/*
 	 * allocate the page directory page
 	 */
@@ -1105,10 +964,6 @@ _pmap_allocpte(pmap_t pmap, unsigned ptepindex, int flags)
 	    ("_pmap_allocpte: flags is neither M_NOWAIT nor M_WAITOK"));
 
 	req = VM_ALLOC_WIRED | VM_ALLOC_ZERO | VM_ALLOC_NOOBJ;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool)
-		req |= VM_ALLOC_WIRED_TLB_PG_POOL;
-#endif
 	/*
 	 * Find or fabricate a new pagetable page
 	 */
@@ -1279,7 +1134,7 @@ pmap_growkernel(vm_offset_t addr)
 
 	mtx_assert(&kernel_map->system_mtx, MA_OWNED);
 	if (kernel_vm_end == 0) {
-		kernel_vm_end = VM_MIN_KERNEL_ADDRESS + VM_KERNEL_ALLOC_OFFSET;
+		kernel_vm_end = VM_MIN_KERNEL_ADDRESS;
 		nkpt = 0;
 		while (segtab_pde(kernel_segmap, kernel_vm_end)) {
 			kernel_vm_end = (kernel_vm_end + PAGE_SIZE * NPTEPG) &
@@ -1308,10 +1163,6 @@ pmap_growkernel(vm_offset_t addr)
 		 * This index is bogus, but out of the way
 		 */
 		req = VM_ALLOC_INTERRUPT | VM_ALLOC_WIRED | VM_ALLOC_NOOBJ;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-		if (need_wired_tlb_page_pool)
-			req |= VM_ALLOC_WIRED_TLB_PG_POOL;
-#endif
 		nkpg = vm_page_alloc(NULL, nkpt, req);
 		if (!nkpg)
 			panic("pmap_growkernel: no memory to grow kernel");
@@ -2155,12 +2006,6 @@ pmap_kenter_temporary(vm_paddr_t pa, int i)
 		printf("%s: ERROR!!! More than one page of virtual address mapping not supported\n",
 		    __func__);
 
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool) {
-		va = pmap_map_fpage(pa, &fpages_shared[PMAP_FPAGE_KENTER_TEMP],
-		    TRUE);
-	} else
-#endif
 	if (pa < MIPS_KSEG0_LARGEST_PHYS) {
 		va = MIPS_PHYS_TO_KSEG0(pa);
 	} else {
@@ -2312,26 +2157,7 @@ pmap_zero_page(vm_page_t m)
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
 	int int_level;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool) {
-		struct fpage *fp1;
-		struct sysmaps *sysmaps;
 
-		sysmaps = &sysmaps_pcpu[PCPU_GET(cpuid)];
-		mtx_lock(&sysmaps->lock);
-		sched_pin();
-
-		fp1 = &sysmaps->fp[PMAP_FPAGE1];
-		va = pmap_map_fpage(phys, fp1, FALSE);
-		bzero((caddr_t)va, PAGE_SIZE);
-		pmap_unmap_fpage(phys, fp1);
-		sched_unpin();
-		mtx_unlock(&sysmaps->lock);
-		/*
-		 * Should you do cache flush?
-		 */
-	} else
-#endif
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
 
 		va = MIPS_PHYS_TO_KSEG0(phys);
@@ -2358,7 +2184,6 @@ pmap_zero_page(vm_page_t m)
 		sched_unpin();
 		PMAP_LGMEM_UNLOCK(sysm);
 	}
-
 }
 
 /*
@@ -2373,24 +2198,7 @@ pmap_zero_page_area(vm_page_t m, int off, int size)
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
 	int int_level;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool) {
-		struct fpage *fp1;
-		struct sysmaps *sysmaps;
 
-		sysmaps = &sysmaps_pcpu[PCPU_GET(cpuid)];
-		mtx_lock(&sysmaps->lock);
-		sched_pin();
-
-		fp1 = &sysmaps->fp[PMAP_FPAGE1];
-		va = pmap_map_fpage(phys, fp1, FALSE);
-		bzero((caddr_t)va + off, size);
-		pmap_unmap_fpage(phys, fp1);
-
-		sched_unpin();
-		mtx_unlock(&sysmaps->lock);
-	} else
-#endif
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
 		va = MIPS_PHYS_TO_KSEG0(phys);
 		bzero((char *)(caddr_t)va + off, size);
@@ -2423,15 +2231,7 @@ pmap_zero_page_idle(vm_page_t m)
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
 	int int_level;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool) {
-		sched_pin();
-		va = pmap_map_fpage(phys, &fpages_shared[PMAP_FPAGE3], FALSE);
-		bzero((caddr_t)va, PAGE_SIZE);
-		pmap_unmap_fpage(phys, &fpages_shared[PMAP_FPAGE3]);
-		sched_unpin();
-	} else
-#endif
+
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
 		va = MIPS_PHYS_TO_KSEG0(phys);
 		bzero((caddr_t)va, PAGE_SIZE);
@@ -2472,95 +2272,67 @@ pmap_copy_page(vm_page_t src, vm_page_t dst)
 	vm_paddr_t phy_src = VM_PAGE_TO_PHYS(src);
 	vm_paddr_t phy_dst = VM_PAGE_TO_PHYS(dst);
 	int int_level;
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	if (need_wired_tlb_page_pool) {
-		struct fpage *fp1, *fp2;
-		struct sysmaps *sysmaps;
-
-		sysmaps = &sysmaps_pcpu[PCPU_GET(cpuid)];
-		mtx_lock(&sysmaps->lock);
-		sched_pin();
-
-		fp1 = &sysmaps->fp[PMAP_FPAGE1];
-		fp2 = &sysmaps->fp[PMAP_FPAGE2];
-
-		va_src = pmap_map_fpage(phy_src, fp1, FALSE);
-		va_dst = pmap_map_fpage(phy_dst, fp2, FALSE);
-
-		bcopy((caddr_t)va_src, (caddr_t)va_dst, PAGE_SIZE);
-
-		pmap_unmap_fpage(phy_src, fp1);
-		pmap_unmap_fpage(phy_dst, fp2);
-		sched_unpin();
-		mtx_unlock(&sysmaps->lock);
 
+	if ((phy_src < MIPS_KSEG0_LARGEST_PHYS) && (phy_dst < MIPS_KSEG0_LARGEST_PHYS)) {
+		/* easy case, all can be accessed via KSEG0 */
 		/*
-		 * Should you flush the cache?
+		 * Flush all caches for VA that are mapped to this page
+		 * to make sure that data in SDRAM is up to date
 		 */
-	} else
-#endif
-	{
-		if ((phy_src < MIPS_KSEG0_LARGEST_PHYS) && (phy_dst < MIPS_KSEG0_LARGEST_PHYS)) {
-			/* easy case, all can be accessed via KSEG0 */
-			/*
-			 * Flush all caches for VA that are mapped to this page
-			 * to make sure that data in SDRAM is up to date
-			 */
-			pmap_flush_pvcache(src);
-			mips_dcache_wbinv_range_index(
-			    MIPS_PHYS_TO_KSEG0(phy_dst), NBPG);
-			va_src = MIPS_PHYS_TO_KSEG0(phy_src);
-			va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
-			bcopy((caddr_t)va_src, (caddr_t)va_dst, PAGE_SIZE);
-			mips_dcache_wbinv_range(va_dst, PAGE_SIZE);
-		} else {
-			int cpu;
-			struct local_sysmaps *sysm;
+		pmap_flush_pvcache(src);
+		mips_dcache_wbinv_range_index(
+		    MIPS_PHYS_TO_KSEG0(phy_dst), NBPG);
+		va_src = MIPS_PHYS_TO_KSEG0(phy_src);
+		va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
+		bcopy((caddr_t)va_src, (caddr_t)va_dst, PAGE_SIZE);
+		mips_dcache_wbinv_range(va_dst, PAGE_SIZE);
+	} else {
+		int cpu;
+		struct local_sysmaps *sysm;
 
-			cpu = PCPU_GET(cpuid);
-			sysm = &sysmap_lmem[cpu];
-			PMAP_LGMEM_LOCK(sysm);
-			sched_pin();
-			int_level = disableintr();
-			if (phy_src < MIPS_KSEG0_LARGEST_PHYS) {
-				/* one side needs mapping - dest */
-				va_src = MIPS_PHYS_TO_KSEG0(phy_src);
-				sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-				pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
-				sysm->valid2 = 1;
-				va_dst = (vm_offset_t)sysm->CADDR2;
-			} else if (phy_dst < MIPS_KSEG0_LARGEST_PHYS) {
-				/* one side needs mapping - src */
-				va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
-				sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-				pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-				va_src = (vm_offset_t)sysm->CADDR1;
-				sysm->valid1 = 1;
-			} else {
-				/* all need mapping */
-				sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-				sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-				pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-				pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
-				sysm->valid1 = sysm->valid2 = 1;
-				va_src = (vm_offset_t)sysm->CADDR1;
-				va_dst = (vm_offset_t)sysm->CADDR2;
-			}
-			bcopy((void *)va_src, (void *)va_dst, PAGE_SIZE);
-			if (sysm->valid1) {
-				pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-				sysm->CMAP1 = 0;
-				sysm->valid1 = 0;
-			}
-			if (sysm->valid2) {
-				pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR2);
-				sysm->CMAP2 = 0;
-				sysm->valid2 = 0;
-			}
-			restoreintr(int_level);
-			sched_unpin();
-			PMAP_LGMEM_UNLOCK(sysm);
+		cpu = PCPU_GET(cpuid);
+		sysm = &sysmap_lmem[cpu];
+		PMAP_LGMEM_LOCK(sysm);
+		sched_pin();
+		int_level = disableintr();
+		if (phy_src < MIPS_KSEG0_LARGEST_PHYS) {
+			/* one side needs mapping - dest */
+			va_src = MIPS_PHYS_TO_KSEG0(phy_src);
+			sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
+			sysm->valid2 = 1;
+			va_dst = (vm_offset_t)sysm->CADDR2;
+		} else if (phy_dst < MIPS_KSEG0_LARGEST_PHYS) {
+			/* one side needs mapping - src */
+			va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
+			sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
+			va_src = (vm_offset_t)sysm->CADDR1;
+			sysm->valid1 = 1;
+		} else {
+			/* all need mapping */
+			sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+			sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
+			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
+			sysm->valid1 = sysm->valid2 = 1;
+			va_src = (vm_offset_t)sysm->CADDR1;
+			va_dst = (vm_offset_t)sysm->CADDR2;
 		}
+		bcopy((void *)va_src, (void *)va_dst, PAGE_SIZE);
+		if (sysm->valid1) {
+			pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
+			sysm->CMAP1 = 0;
+			sysm->valid1 = 0;
+		}
+		if (sysm->valid2) {
+			pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR2);
+			sysm->CMAP2 = 0;
+			sysm->valid2 = 0;
+		}
+		restoreintr(int_level);
+		sched_unpin();
+		PMAP_LGMEM_UNLOCK(sysm);
 	}
 }
 
@@ -3308,11 +3080,6 @@ pmap_kextract(vm_offset_t va)
 	else if (va >= MIPS_KSEG1_START &&
 	    va < MIPS_KSEG2_START)
 		pa = MIPS_KSEG1_TO_PHYS(va);
-#ifdef VM_ALLOC_WIRED_TLB_PG_POOL
-	else if (need_wired_tlb_page_pool && ((va >= VM_MIN_KERNEL_ADDRESS) &&
-	    (va < (VM_MIN_KERNEL_ADDRESS + VM_KERNEL_ALLOC_OFFSET))))
-		pa = MIPS_KSEG0_TO_PHYS(va);
-#endif
 	else if (va >= MIPS_KSEG2_START && va < VM_MAX_KERNEL_ADDRESS) {
 		pt_entry_t *ptep;
 
diff --git a/sys/mips/mips/swtch.S b/sys/mips/mips/swtch.S
index dd66ecec684..4fc3cbd6cb9 100644
--- a/sys/mips/mips/swtch.S
+++ b/sys/mips/mips/swtch.S
@@ -340,7 +340,7 @@ blocked_loop:
 	lw	a2, TD_PCB(a1)
 	sw	a2, PC_CURPCB(a3)
 	lw	v0, TD_REALKSTACK(a1)
-	li	s0, (MIPS_KSEG2_START+VM_KERNEL_ALLOC_OFFSET)		# If Uarea addr is below kseg2,
+	li	s0, MIPS_KSEG2_START		# If Uarea addr is below kseg2,
 	bltu	v0, s0, sw2			# no need to insert in TLB.
 	lw	a1, TD_UPTE+0(s7)		# t0 = first u. pte
 	lw	a2, TD_UPTE+4(s7)		# t1 = 2nd u. pte

From b92f01b79c1c926a46299ca5f4c85574acc7da3e Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 01:17:31 +0000
Subject: [PATCH 098/532] o) Use inline functions to access coprocessor 0
 registers rather than external    ones implemented using assembly. o) Use
 TRAPF_USERMODE() consistently rather than USERMODE().  Eliminate   
 <machine/psl.h> as a result. o) Use intr_*() rather than *intr(),
 consistently. o) Use register_t instead of u_int in some trap code. o) Merge
 some more endian-related macros to machine/asm.h from NetBSD. o) Add PTR_LI
 macro, which loads an address with the correct sign-extension for    a
 pointer. o) Restore interrupts when bailing out due to an excessive IRQ in   
 nexus_setup_intr(). o) Remove unused functions from psraccess.S. o) Enter
 temporary virtual entries for large memory access into the page tables   
 rather than simply hoping they stay resident in the TLB and we don't need to 
   do a refill for them. o) Abstract out large memory mapping setup/teardown
 using some macros. o) Do mips_dcache_wbinv_range() when using temporary
 virtual addresses just    like we do when we can use the direct map.

---
 ObsoleteFiles.inc             |   4 +
 sys/mips/include/asm.h        |  84 +++++++----
 sys/mips/include/cpu.h        |  21 +--
 sys/mips/include/db_machdep.h |   1 -
 sys/mips/include/param.h      |   4 +-
 sys/mips/include/profile.h    |   8 +-
 sys/mips/include/psl.h        |  50 -------
 sys/mips/include/trap.h       |  20 +--
 sys/mips/mips/machdep.c       |  20 +--
 sys/mips/mips/mp_machdep.c    |   2 +-
 sys/mips/mips/nexus.c         |   9 +-
 sys/mips/mips/pmap.c          | 254 ++++++++++++++--------------------
 sys/mips/mips/psraccess.S     | 117 ----------------
 sys/mips/mips/tick.c          |   8 +-
 sys/mips/mips/trap.c          |  17 ++-
 sys/mips/rmi/clock.c          |  12 +-
 16 files changed, 214 insertions(+), 417 deletions(-)
 delete mode 100644 sys/mips/include/psl.h

diff --git a/ObsoleteFiles.inc b/ObsoleteFiles.inc
index 0844d714693..2d4624dc984 100644
--- a/ObsoleteFiles.inc
+++ b/ObsoleteFiles.inc
@@ -14,6 +14,10 @@
 # The file is partitioned: OLD_FILES first, then OLD_LIBS and OLD_DIRS last.
 #
 
+# 20100416: [mips] removed <machine/psl.h>
+.if ${TARGET_ARCH} == "mips"
+OLD_FILES+=usr/include/machine/psl.h
+.endif
 # 20100415: [mips] removed unused headers
 .if ${TARGET_ARCH} == "mips"
 OLD_FILES+=usr/include/machine/archtype.h
diff --git a/sys/mips/include/asm.h b/sys/mips/include/asm.h
index 0a9c518e929..23bfde148d6 100644
--- a/sys/mips/include/asm.h
+++ b/sys/mips/include/asm.h
@@ -98,23 +98,6 @@
 
 #define	_C_LABEL(x)	x
 
-/* 
- *  Endian-independent assembly-code aliases for unaligned memory accesses.
- */
-#if BYTE_ORDER == LITTLE_ENDIAN
-#define	LWLO	lwl
-#define	LWHI	lwr
-#define	SWLO	swl
-#define	SWHI	swr
-#endif
-
-#if BYTE_ORDER == BIG_ENDIAN
-#define	LWLO	lwr
-#define	LWHI	lwl
-#define	SWLO	swr
-#define	SWHI	swl
-#endif
-
 #ifdef USE_AENT
 #define	AENT(x)		\
 	.aent	x, 0
@@ -306,28 +289,32 @@ _C_LABEL(x):
 
 /*
  * Call ast if required
+ *
+ * XXX Do we really need to disable interrupts?
  */
 #define DO_AST				             \
 44:				                     \
-	PTR_LA	s0, _C_LABEL(disableintr)           ;\
-	jalr	s0                                  ;\
-	nop                                         ;\
-	move	a0, v0                              ;\
+	mfc0	t0, MIPS_COP_0_STATUS               ;\
+	and	a0, t0, MIPS_SR_INT_IE              ;\
+	xor	t0, a0, t0                          ;\
+	mtc0	t0, MIPS_COP_0_STATUS               ;\
+	COP0_SYNC                                   ;\
 	GET_CPU_PCPU(s1)                            ;\
-	lw	s3, PC_CURPCB(s1)                   ;\
-	lw	s1, PC_CURTHREAD(s1)                ;\
+	PTR_L	s3, PC_CURPCB(s1)                   ;\
+	PTR_L	s1, PC_CURTHREAD(s1)                ;\
 	lw	s2, TD_FLAGS(s1)                    ;\
 	li	s0, TDF_ASTPENDING | TDF_NEEDRESCHED;\
 	and	s2, s0                              ;\
-	PTR_LA	s0, _C_LABEL(restoreintr)           ;\
-	jalr	s0                                  ;\
-	nop                                         ;\
+	mfc0	t0, MIPS_COP_0_STATUS               ;\
+	or	t0, a0, t0                          ;\
+	mtc0	t0, MIPS_COP_0_STATUS               ;\
+	COP0_SYNC                                   ;\
 	beq	s2, zero, 4f                        ;\
 	nop                                         ;\
 	PTR_LA	s0, _C_LABEL(ast)                   ;\
 	jalr	s0                                  ;\
 	PTR_ADDU a0, s3, U_PCB_REGS                 ;\
-	j 44b			                    ;\
+	j	44b		                    ;\
         nop                                         ;\
 4:
 
@@ -382,6 +369,45 @@ _C_LABEL(x):
 #define	CALLFRAME_SP	(CALLFRAME_SIZ - 2 * SZREG)
 #define	CALLFRAME_RA	(CALLFRAME_SIZ - 1 * SZREG)
 
+/*
+ *   Endian-independent assembly-code aliases for unaligned memory accesses.
+ */
+#if _BYTE_ORDER == _LITTLE_ENDIAN
+# define LWHI lwr
+# define LWLO lwl
+# define SWHI swr
+# define SWLO swl
+# if SZREG == 4
+#  define REG_LHI   lwr
+#  define REG_LLO   lwl
+#  define REG_SHI   swr
+#  define REG_SLO   swl
+# else
+#  define REG_LHI   ldr
+#  define REG_LLO   ldl
+#  define REG_SHI   sdr
+#  define REG_SLO   sdl
+# endif
+#endif
+
+#if _BYTE_ORDER == _BIG_ENDIAN
+# define LWHI lwl
+# define LWLO lwr
+# define SWHI swl
+# define SWLO swr
+# if SZREG == 4
+#  define REG_LHI   lwl
+#  define REG_LLO   lwr
+#  define REG_SHI   swl
+#  define REG_SLO   swr
+# else
+#  define REG_LHI   ldl
+#  define REG_LLO   ldr
+#  define REG_SHI   sdl
+#  define REG_SLO   sdr
+# endif
+#endif
+
 /*
  * While it would be nice to be compatible with the SGI
  * REG_L and REG_S macros, because they do not take parameters, it
@@ -402,6 +428,7 @@ _C_LABEL(x):
 #define	PTR_SUBIU	subu
 #define	PTR_L		lw
 #define	PTR_LA		la
+#define	PTR_LI		li
 #define	PTR_S		sw
 #define	PTR_SLL		sll
 #define	PTR_SLLV	sllv
@@ -424,6 +451,7 @@ _C_LABEL(x):
 #define	PTR_SUBIU	dsubu
 #define	PTR_L		ld
 #define	PTR_LA		dla
+#define	PTR_LI		dli
 #define	PTR_S		sd
 #define	PTR_SLL		dsll
 #define	PTR_SLLV	dsllv
@@ -765,7 +793,7 @@ _C_LABEL(x):
 #endif
 
 #define	GET_CPU_PCPU(reg)		\
-	lw	reg, _C_LABEL(pcpup);
+	PTR_L	reg, _C_LABEL(pcpup);
 
 /*
  * Description of the setjmp buffer
diff --git a/sys/mips/include/cpu.h b/sys/mips/include/cpu.h
index 1ec1cfc8a86..5c5e5e5e92b 100644
--- a/sys/mips/include/cpu.h
+++ b/sys/mips/include/cpu.h
@@ -47,7 +47,6 @@
 #ifndef _MACHINE_CPU_H_
 #define	_MACHINE_CPU_H_
 
-#include <machine/psl.h>
 #include <machine/endian.h>
 
 #define MIPS_KSEG0_LARGEST_PHYS         0x20000000
@@ -334,6 +333,7 @@
 #define	cpu_swapout(p)		panic("cpu_swapout: can't get here");
 
 #ifndef _LOCORE
+#include <machine/cpufunc.h>
 #include <machine/frame.h>
 /*
  * Arguments to hardclock and gatherstats encapsulate the previous
@@ -342,7 +342,6 @@
 #define	clockframe trapframe	/* Use normal trap frame */
 
 #define	CLKF_USERMODE(framep)	((framep)->sr & SR_KSU_USER)
-#define	CLKF_BASEPRI(framep)	((framep)->cpl == 0)
 #define	CLKF_PC(framep)		((framep)->pc)
 #define	CLKF_INTR(framep)	(0)
 #define	MIPS_CLKF_INTR()	(intr_nesting_level >= 1)
@@ -350,6 +349,11 @@
 #define	TRAPF_PC(framep)	((framep)->pc)
 #define	cpu_getstack(td)	((td)->td_frame->sp)
 
+/*
+ * A machine-independent interface to the CPU's counter.
+ */
+#define	get_cyclecount()	mips_rd_count()
+
 /*
  * CPU identification, from PRID register.
  */
@@ -542,18 +546,6 @@ extern int intr_nesting_level;
  *  Low level access routines to CPU registers
  */
 
-void setsoftintr0(void);
-void clearsoftintr0(void);
-void setsoftintr1(void);
-void clearsoftintr1(void);
-
-
-u_int32_t mips_cp0_status_read(void);
-void mips_cp0_status_write(u_int32_t);
-
-int disableintr(void);
-void restoreintr(int);
-int enableintr(void);
 int Mips_TLBGetPID(void);
 
 void swi_vm(void *);
@@ -562,7 +554,6 @@ void cpu_reset(void);
 
 u_int32_t set_intr_mask(u_int32_t);
 u_int32_t get_intr_mask(void);
-u_int32_t get_cyclecount(void);
 
 #define	cpu_spinwait()		/* nothing */
 
diff --git a/sys/mips/include/db_machdep.h b/sys/mips/include/db_machdep.h
index d7bf69a427d..f0a84d2ee9a 100644
--- a/sys/mips/include/db_machdep.h
+++ b/sys/mips/include/db_machdep.h
@@ -38,7 +38,6 @@
 #define	_MIPS_DB_MACHDEP_H_
 
 #include <machine/frame.h>
-#include <machine/psl.h>
 #include <machine/trap.h>
 #include <machine/endian.h>
 
diff --git a/sys/mips/include/param.h b/sys/mips/include/param.h
index 9d487022439..0f380852609 100644
--- a/sys/mips/include/param.h
+++ b/sys/mips/include/param.h
@@ -46,9 +46,7 @@
 
 #include <sys/cdefs.h>
 #ifdef _KERNEL
-#ifdef _LOCORE
-#include <machine/psl.h>
-#else
+#ifndef _LOCORE
 #include <machine/cpu.h>
 #endif
 #endif
diff --git a/sys/mips/include/profile.h b/sys/mips/include/profile.h
index 728a468dbf4..6d976d90e77 100644
--- a/sys/mips/include/profile.h
+++ b/sys/mips/include/profile.h
@@ -84,17 +84,17 @@
 #ifdef SMP
 extern int	mcount_lock;
 #define	MCOUNT_ENTER(s)	{					\
-	s = disable_intr();					\
+	s = intr_disable();					\
 	while (!atomic_cmpset_acq_int(&mcount_lock, 0, 1))	\
 		/* nothing */ ;					\
 }
 #define	MCOUNT_EXIT(s)	{					\
 	atomic_store_rel_int(&mcount_lock, 0);			\
-	enableintr(s);						\
+	intr_restore(s);						\
 }
 #else
-#define	MCOUNT_ENTER(s)	{ s = disable_intr(); }
-#define	MCOUNT_EXIT(s)	(enableintr(s))
+#define	MCOUNT_ENTER(s)	{ s = intr_disable(); }
+#define	MCOUNT_EXIT(s)	(intr_restore(s))
 #endif
 
 /* REVISIT for mips */
diff --git a/sys/mips/include/psl.h b/sys/mips/include/psl.h
deleted file mode 100644
index f02a1a95fa9..00000000000
--- a/sys/mips/include/psl.h
+++ /dev/null
@@ -1,50 +0,0 @@
-/*	$OpenBSD: psl.h,v 1.2 1998/01/28 13:46:25 pefo Exp $	*/
-
-/*-
- * Copyright (c) 1992, 1993
- *	The Regents of the University of California.  All rights reserved.
- *
- * This code is derived from software contributed to Berkeley by
- * Ralph Campbell.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 4. Neither the name of the University nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- *	from: @(#)psl.h 8.1 (Berkeley) 6/10/93
- *	JNPR: psl.h,v 1.1 2006/08/07 05:38:57 katta
- * $FreeBSD$
- */
-
-#ifndef _MACHINE_PSL_H_
-#define	_MACHINE_PSL_H_
-
-#include <machine/cpu.h>
-
-/*
- * Macros to decode processor status word.
- */
-#define	USERMODE(ps)	(((ps) & SR_KSU_MASK) == SR_KSU_USER)
-#define	BASEPRI(ps)	(((ps) & (INT_MASK | SR_INT_ENA_PREV)) \
-			    == (INT_MASK | SR_INT_ENA_PREV))
-#endif	/* _MACHINE_PSL_H_ */
diff --git a/sys/mips/include/trap.h b/sys/mips/include/trap.h
index f382e70f488..d8042382157 100644
--- a/sys/mips/include/trap.h
+++ b/sys/mips/include/trap.h
@@ -74,17 +74,17 @@
 #if !defined(SMP) && (defined(DDB) || defined(DEBUG))
 
 struct trapdebug {		/* trap history buffer for debugging */
-	u_int	status;
-	u_int	cause;
-	u_int	vadr;
-	u_int	pc;
-	u_int	ra;
-	u_int	sp;
-	u_int	code;
+	register_t	status;
+	register_t	cause;
+	register_t	vadr;
+	register_t	pc;
+	register_t	ra;
+	register_t	sp;
+	register_t	code;
 };
 
 #define	trapdebug_enter(x, cd) {	\
-	intrmask_t s = disableintr();	\
+	register_t s = intr_disable();	\
 	trp->status = x->sr;		\
 	trp->cause = x->cause;		\
 	trp->vadr = x->badvaddr;	\
@@ -94,7 +94,7 @@ struct trapdebug {		/* trap history buffer for debugging */
 	trp->code = cd;			\
 	if (++trp == &trapdebug[TRAPSIZE])	\
 		trp = trapdebug;	\
-	restoreintr(s);			\
+	intr_restore(s);		\
 }
 
 #define	TRAPSIZE 10		/* Trap log buffer length */
@@ -116,7 +116,7 @@ void MipsTLBMissException(void);
 void MipsUserGenException(void);
 void MipsUserIntr(void);
 
-u_int trap(struct trapframe *);
+register_t trap(struct trapframe *);
 
 #ifndef LOCORE /* XXX */
 int check_address(void *);
diff --git a/sys/mips/mips/machdep.c b/sys/mips/mips/machdep.c
index 4f728d56c05..6ca4ec9409a 100644
--- a/sys/mips/mips/machdep.c
+++ b/sys/mips/mips/machdep.c
@@ -370,11 +370,10 @@ mips_vector_init(void)
 	 * when handler is installed for it
 	 */
 	set_intr_mask(ALL_INT_MASK);
-	enableintr();
+	intr_enable();
 
 	/* Clear BEV in SR so we start handling our own exceptions */
-	mips_cp0_status_write(mips_cp0_status_read() & ~SR_BOOT_EXC_VEC);
-
+	mips_wr_status(mips_rd_status() & ~SR_BOOT_EXC_VEC);
 }
 
 /*
@@ -471,7 +470,7 @@ spinlock_enter(void)
 
 	td = curthread;
 	if (td->td_md.md_spinlock_count == 0)
-		td->td_md.md_saved_intr = disableintr();
+		td->td_md.md_saved_intr = intr_disable();
 	td->td_md.md_spinlock_count++;
 	critical_enter();
 }
@@ -485,16 +484,7 @@ spinlock_exit(void)
 	critical_exit();
 	td->td_md.md_spinlock_count--;
 	if (td->td_md.md_spinlock_count == 0)
-		restoreintr(td->td_md.md_saved_intr);
-}
-
-u_int32_t
-get_cyclecount(void)
-{
-	u_int32_t count;
-
-	mfc0_macro(count, 9);
-	return (count);
+		intr_restore(td->td_md.md_saved_intr);
 }
 
 /*
@@ -503,7 +493,7 @@ get_cyclecount(void)
 void
 cpu_idle(int busy)
 {
-	if (mips_cp0_status_read() & SR_INT_ENAB)
+	if (mips_rd_status() & SR_INT_ENAB)
 		__asm __volatile ("wait");
 	else
 		panic("ints disabled in idleproc!");
diff --git a/sys/mips/mips/mp_machdep.c b/sys/mips/mips/mp_machdep.c
index f65db9bd6b4..52d05fad711 100644
--- a/sys/mips/mips/mp_machdep.c
+++ b/sys/mips/mips/mp_machdep.c
@@ -296,7 +296,7 @@ smp_init_secondary(u_int32_t cpuid)
 	 */
 	mips_wr_compare(mips_rd_count() + counter_freq / hz);
 
-	enableintr();
+	intr_enable();
 
 	/* enter the scheduler */
 	sched_throw(NULL);
diff --git a/sys/mips/mips/nexus.c b/sys/mips/mips/nexus.c
index 6600980bb02..10b38b5ff86 100644
--- a/sys/mips/mips/nexus.c
+++ b/sys/mips/mips/nexus.c
@@ -166,16 +166,19 @@ static int
 nexus_setup_intr(device_t dev, device_t child, struct resource *res, int flags,
     driver_filter_t *filt, driver_intr_t *intr, void *arg, void **cookiep)
 {
+	register_t s;
 	int irq;
 
-	intrmask_t s = disableintr();
+	s = intr_disable();
 	irq = rman_get_start(res);
-	if (irq >= NUM_MIPS_IRQS)
+	if (irq >= NUM_MIPS_IRQS) {
+		intr_restore(s);
 		return (0);
+	}
 
 	cpu_establish_hardintr(device_get_nameunit(child), filt, intr, arg,
 	    irq, flags, cookiep);
-	restoreintr(s);
+	intr_restore(s);
 	return (0);
 }
 
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 9f01f8a88c4..f1056fdcb46 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -194,10 +194,7 @@ static void pmap_update_page_action(void *arg);
 
 struct local_sysmaps {
 	struct mtx lock;
-	pt_entry_t CMAP1;
-	pt_entry_t CMAP2;
-	caddr_t CADDR1;
-	caddr_t CADDR2;
+	vm_offset_t base;
 	uint16_t valid1, valid2;
 };
 
@@ -211,6 +208,59 @@ struct local_sysmaps {
 static struct local_sysmaps sysmap_lmem[MAXCPU];
 caddr_t virtual_sys_start = (caddr_t)0;
 
+#define	PMAP_LMEM_MAP1(va, phys)					\
+	int cpu;							\
+	struct local_sysmaps *sysm;					\
+	pt_entry_t *pte, npte;						\
+									\
+	cpu = PCPU_GET(cpuid);						\
+	sysm = &sysmap_lmem[cpu];					\
+	PMAP_LGMEM_LOCK(sysm);						\
+	intr = intr_disable();						\
+	sched_pin();							\
+	va = sysm->base;						\
+	npte = mips_paddr_to_tlbpfn(phys) |				\
+	    PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;			\
+	pte = pmap_pte(kernel_pmap, va);				\
+	*pte = npte;							\
+	sysm->valid1 = 1;
+
+#define	PMAP_LMEM_MAP2(va1, phys1, va2, phys2)				\
+	int cpu;							\
+	struct local_sysmaps *sysm;					\
+	pt_entry_t *pte, npte;						\
+									\
+	cpu = PCPU_GET(cpuid);						\
+	sysm = &sysmap_lmem[cpu];					\
+	PMAP_LGMEM_LOCK(sysm);						\
+	intr = intr_disable();						\
+	sched_pin();							\
+	va1 = sysm->base;						\
+	va2 = sysm->base + PAGE_SIZE;					\
+	npte = mips_paddr_to_tlbpfn(phys2) |				\
+	    PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;			\
+	pte = pmap_pte(kernel_pmap, va1);				\
+	*pte = npte;							\
+	npte = mips_paddr_to_tlbpfn(phys2) |				\
+	    PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;			\
+	pte = pmap_pte(kernel_pmap, va2);				\
+	*pte = npte;							\
+	sysm->valid1 = 1;						\
+	sysm->valid2 = 1;
+
+#define	PMAP_LMEM_UNMAP()						\
+	pte = pmap_pte(kernel_pmap, sysm->base);			\
+	*pte = PTE_G;							\
+	pmap_invalidate_page(kernel_pmap, sysm->base);			\
+	sysm->valid1 = 0;						\
+	pte = pmap_pte(kernel_pmap, sysm->base + PAGE_SIZE);		\
+	*pte = PTE_G;							\
+	pmap_invalidate_page(kernel_pmap, sysm->base + PAGE_SIZE);	\
+	sysm->valid2 = 0;						\
+	sched_unpin();							\
+	intr_restore(intr);						\
+	PMAP_LGMEM_UNLOCK(sysm);
+
 pd_entry_t
 pmap_segmap(pmap_t pmap, vm_offset_t va)
 {
@@ -382,12 +432,8 @@ again:
 	 */
 	if (memory_larger_than_512meg) {
 		for (i = 0; i < MAXCPU; i++) {
-			sysmap_lmem[i].CMAP1 = PTE_G;
-			sysmap_lmem[i].CMAP2 = PTE_G;
-			sysmap_lmem[i].CADDR1 = (caddr_t)virtual_avail;
-			virtual_avail += PAGE_SIZE;
-			sysmap_lmem[i].CADDR2 = (caddr_t)virtual_avail;
-			virtual_avail += PAGE_SIZE;
+			sysmap_lmem[i].base = virtual_avail;
+			virtual_avail += PAGE_SIZE * 2;
 			sysmap_lmem[i].valid1 = sysmap_lmem[i].valid2 = 0;
 			PMAP_LGMEM_LOCK_INIT(&sysmap_lmem[i]);
 		}
@@ -2001,7 +2047,7 @@ void *
 pmap_kenter_temporary(vm_paddr_t pa, int i)
 {
 	vm_offset_t va;
-	int int_level;
+	register_t intr;
 	if (i != 0)
 		printf("%s: ERROR!!! More than one page of virtual address mapping not supported\n",
 		    __func__);
@@ -2011,20 +2057,24 @@ pmap_kenter_temporary(vm_paddr_t pa, int i)
 	} else {
 		int cpu;
 		struct local_sysmaps *sysm;
+		pt_entry_t *pte, npte;
+
 		/* If this is used other than for dumps, we may need to leave
 		 * interrupts disasbled on return. If crash dumps don't work when
 		 * we get to this point, we might want to consider this (leaving things
 		 * disabled as a starting point ;-)
 	 	 */
-		int_level = disableintr();
+		intr = intr_disable();
 		cpu = PCPU_GET(cpuid);
 		sysm = &sysmap_lmem[cpu];
 		/* Since this is for the debugger, no locks or any other fun */
-		sysm->CMAP1 = mips_paddr_to_tlbpfn(pa) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+		npte = mips_paddr_to_tlbpfn(pa) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
+		pte = pmap_pte(kernel_pmap, sysm->base);
+		*pte = npte;
 		sysm->valid1 = 1;
-		pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-		va = (vm_offset_t)sysm->CADDR1;
-		restoreintr(int_level);
+		pmap_update_page(kernel_pmap, sysm->base, npte);
+		va = sysm->base;
+		intr_restore(intr);
 	}
 	return ((void *)va);
 }
@@ -2033,7 +2083,7 @@ void
 pmap_kenter_temporary_free(vm_paddr_t pa)
 {
 	int cpu;
-	int int_level;
+	register_t intr;
 	struct local_sysmaps *sysm;
 
 	if (pa < MIPS_KSEG0_LARGEST_PHYS) {
@@ -2043,10 +2093,13 @@ pmap_kenter_temporary_free(vm_paddr_t pa)
 	cpu = PCPU_GET(cpuid);
 	sysm = &sysmap_lmem[cpu];
 	if (sysm->valid1) {
-		int_level = disableintr();
-		pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-		restoreintr(int_level);
-		sysm->CMAP1 = 0;
+		pt_entry_t *pte;
+
+		intr = intr_disable();
+		pte = pmap_pte(kernel_pmap, sysm->base);
+		*pte = PTE_G;
+		pmap_invalidate_page(kernel_pmap, sysm->base);
+		intr_restore(intr);
 		sysm->valid1 = 0;
 	}
 }
@@ -2156,33 +2209,20 @@ pmap_zero_page(vm_page_t m)
 {
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
-	int int_level;
+	register_t intr;
 
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
-
 		va = MIPS_PHYS_TO_KSEG0(phys);
 
 		bzero((caddr_t)va, PAGE_SIZE);
 		mips_dcache_wbinv_range(va, PAGE_SIZE);
 	} else {
-		int cpu;
-		struct local_sysmaps *sysm;
+		PMAP_LMEM_MAP1(va, phys);
 
-		cpu = PCPU_GET(cpuid);
-		sysm = &sysmap_lmem[cpu];
-		PMAP_LGMEM_LOCK(sysm);
-		sched_pin();
-		int_level = disableintr();
-		sysm->CMAP1 = mips_paddr_to_tlbpfn(phys) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-		sysm->valid1 = 1;
-		pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-		bzero(sysm->CADDR1, PAGE_SIZE);
-		pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-		restoreintr(int_level);
-		sysm->CMAP1 = 0;
-		sysm->valid1 = 0;
-		sched_unpin();
-		PMAP_LGMEM_UNLOCK(sysm);
+		bzero((caddr_t)va, PAGE_SIZE);
+		mips_dcache_wbinv_range(va, PAGE_SIZE);
+
+		PMAP_LMEM_UNMAP();
 	}
 }
 
@@ -2197,31 +2237,19 @@ pmap_zero_page_area(vm_page_t m, int off, int size)
 {
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
-	int int_level;
+	register_t intr;
 
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
 		va = MIPS_PHYS_TO_KSEG0(phys);
 		bzero((char *)(caddr_t)va + off, size);
 		mips_dcache_wbinv_range(va + off, size);
 	} else {
-		int cpu;
-		struct local_sysmaps *sysm;
+		PMAP_LMEM_MAP1(va, phys);
 
-		cpu = PCPU_GET(cpuid);
-		sysm = &sysmap_lmem[cpu];
-		PMAP_LGMEM_LOCK(sysm);
-		int_level = disableintr();
-		sched_pin();
-		sysm->CMAP1 = mips_paddr_to_tlbpfn(phys) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-		sysm->valid1 = 1;
-		pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-		bzero((char *)sysm->CADDR1 + off, size);
-		pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-		restoreintr(int_level);
-		sysm->CMAP1 = 0;
-		sysm->valid1 = 0;
-		sched_unpin();
-		PMAP_LGMEM_UNLOCK(sysm);
+		bzero((char *)va + off, size);
+		mips_dcache_wbinv_range(va + off, size);
+
+		PMAP_LMEM_UNMAP();
 	}
 }
 
@@ -2230,33 +2258,20 @@ pmap_zero_page_idle(vm_page_t m)
 {
 	vm_offset_t va;
 	vm_paddr_t phys = VM_PAGE_TO_PHYS(m);
-	int int_level;
+	register_t intr;
 
 	if (phys < MIPS_KSEG0_LARGEST_PHYS) {
 		va = MIPS_PHYS_TO_KSEG0(phys);
 		bzero((caddr_t)va, PAGE_SIZE);
 		mips_dcache_wbinv_range(va, PAGE_SIZE);
 	} else {
-		int cpu;
-		struct local_sysmaps *sysm;
+		PMAP_LMEM_MAP1(va, phys);
 
-		cpu = PCPU_GET(cpuid);
-		sysm = &sysmap_lmem[cpu];
-		PMAP_LGMEM_LOCK(sysm);
-		int_level = disableintr();
-		sched_pin();
-		sysm->CMAP1 = mips_paddr_to_tlbpfn(phys) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-		sysm->valid1 = 1;
-		pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-		bzero(sysm->CADDR1, PAGE_SIZE);
-		pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-		restoreintr(int_level);
-		sysm->CMAP1 = 0;
-		sysm->valid1 = 0;
-		sched_unpin();
-		PMAP_LGMEM_UNLOCK(sysm);
+		bzero((caddr_t)va, PAGE_SIZE);
+		mips_dcache_wbinv_range(va, PAGE_SIZE);
+
+		PMAP_LMEM_UNMAP();
 	}
-
 }
 
 /*
@@ -2271,7 +2286,7 @@ pmap_copy_page(vm_page_t src, vm_page_t dst)
 	vm_offset_t va_src, va_dst;
 	vm_paddr_t phy_src = VM_PAGE_TO_PHYS(src);
 	vm_paddr_t phy_dst = VM_PAGE_TO_PHYS(dst);
-	int int_level;
+	register_t intr;
 
 	if ((phy_src < MIPS_KSEG0_LARGEST_PHYS) && (phy_dst < MIPS_KSEG0_LARGEST_PHYS)) {
 		/* easy case, all can be accessed via KSEG0 */
@@ -2281,58 +2296,18 @@ pmap_copy_page(vm_page_t src, vm_page_t dst)
 		 */
 		pmap_flush_pvcache(src);
 		mips_dcache_wbinv_range_index(
-		    MIPS_PHYS_TO_KSEG0(phy_dst), NBPG);
+		    MIPS_PHYS_TO_KSEG0(phy_dst), PAGE_SIZE);
 		va_src = MIPS_PHYS_TO_KSEG0(phy_src);
 		va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
 		bcopy((caddr_t)va_src, (caddr_t)va_dst, PAGE_SIZE);
 		mips_dcache_wbinv_range(va_dst, PAGE_SIZE);
 	} else {
-		int cpu;
-		struct local_sysmaps *sysm;
+		PMAP_LMEM_MAP2(va_src, phy_src, va_dst, phy_dst);
 
-		cpu = PCPU_GET(cpuid);
-		sysm = &sysmap_lmem[cpu];
-		PMAP_LGMEM_LOCK(sysm);
-		sched_pin();
-		int_level = disableintr();
-		if (phy_src < MIPS_KSEG0_LARGEST_PHYS) {
-			/* one side needs mapping - dest */
-			va_src = MIPS_PHYS_TO_KSEG0(phy_src);
-			sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
-			sysm->valid2 = 1;
-			va_dst = (vm_offset_t)sysm->CADDR2;
-		} else if (phy_dst < MIPS_KSEG0_LARGEST_PHYS) {
-			/* one side needs mapping - src */
-			va_dst = MIPS_PHYS_TO_KSEG0(phy_dst);
-			sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-			va_src = (vm_offset_t)sysm->CADDR1;
-			sysm->valid1 = 1;
-		} else {
-			/* all need mapping */
-			sysm->CMAP1 = mips_paddr_to_tlbpfn(phy_src) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-			sysm->CMAP2 = mips_paddr_to_tlbpfn(phy_dst) | PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;
-			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR1, sysm->CMAP1);
-			pmap_TLB_update_kernel((vm_offset_t)sysm->CADDR2, sysm->CMAP2);
-			sysm->valid1 = sysm->valid2 = 1;
-			va_src = (vm_offset_t)sysm->CADDR1;
-			va_dst = (vm_offset_t)sysm->CADDR2;
-		}
 		bcopy((void *)va_src, (void *)va_dst, PAGE_SIZE);
-		if (sysm->valid1) {
-			pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR1);
-			sysm->CMAP1 = 0;
-			sysm->valid1 = 0;
-		}
-		if (sysm->valid2) {
-			pmap_TLB_invalidate_kernel((vm_offset_t)sysm->CADDR2);
-			sysm->CMAP2 = 0;
-			sysm->valid2 = 0;
-		}
-		restoreintr(int_level);
-		sched_unpin();
-		PMAP_LGMEM_UNLOCK(sysm);
+		mips_dcache_wbinv_range(va_dst, PAGE_SIZE);
+
+		PMAP_LMEM_UNMAP();
 	}
 }
 
@@ -3085,34 +3060,11 @@ pmap_kextract(vm_offset_t va)
 
 		/* Is the kernel pmap initialized? */
 		if (kernel_pmap->pm_active) {
-			if (va >= (vm_offset_t)virtual_sys_start) {
-				/* Its inside the virtual address range */
-				ptep = pmap_pte(kernel_pmap, va);
-				if (ptep)
-					pa = mips_tlbpfn_to_paddr(*ptep) |
-					    (va & PAGE_MASK);
-			} else {
-				int i;
-
-				/*
-				 * its inside the special mapping area, I
-				 * don't think this should happen, but if it
-				 * does I want it toa all work right :-)
-				 * Note if it does happen, we assume the
-				 * caller has the lock? FIXME, this needs to
-				 * be checked FIXEM - RRS.
-				 */
-				for (i = 0; i < MAXCPU; i++) {
-					if ((sysmap_lmem[i].valid1) && ((vm_offset_t)sysmap_lmem[i].CADDR1 == va)) {
-						pa = mips_tlbpfn_to_paddr(sysmap_lmem[i].CMAP1);
-						break;
-					}
-					if ((sysmap_lmem[i].valid2) && ((vm_offset_t)sysmap_lmem[i].CADDR2 == va)) {
-						pa = mips_tlbpfn_to_paddr(sysmap_lmem[i].CMAP2);
-						break;
-					}
-				}
-			}
+			/* Its inside the virtual address range */
+			ptep = pmap_pte(kernel_pmap, va);
+			if (ptep)
+				pa = mips_tlbpfn_to_paddr(*ptep) |
+				    (va & PAGE_MASK);
 		}
 	}
 	return pa;
diff --git a/sys/mips/mips/psraccess.S b/sys/mips/mips/psraccess.S
index 0bcb04d3d33..9b40a234458 100644
--- a/sys/mips/mips/psraccess.S
+++ b/sys/mips/mips/psraccess.S
@@ -53,109 +53,6 @@
 
 	.set	noreorder		# Noreorder is default style!
 
-/*
- * Set/clear software interrupt.
- */
-
-LEAF(setsoftintr0)
-	mfc0	v0, COP_0_CAUSE_REG	# read cause register
-	nop
-	or	v0, v0, SOFT_INT_MASK_0 # set soft clock interrupt
-	mtc0	v0, COP_0_CAUSE_REG	# save it
-	j	ra
-	nop
-END(setsoftintr0)
-
-LEAF(clearsoftintr0)
-	mfc0	v0, COP_0_CAUSE_REG	# read cause register
-	nop
-	and	v0, v0, ~SOFT_INT_MASK_0  # clear soft clock interrupt
-	mtc0	v0, COP_0_CAUSE_REG	# save it
-	j	ra
-	nop
-END(clearsoftintr0)
-
-LEAF(setsoftintr1)
-	mfc0	v0, COP_0_CAUSE_REG	# read cause register
-	nop
-	or	v0, v0, SOFT_INT_MASK_1 # set soft net interrupt
-	mtc0	v0, COP_0_CAUSE_REG	# save it
-	j	ra
-	nop
-END(setsoftintr1)
-
-LEAF(clearsoftintr1)
-	mfc0	v0, COP_0_CAUSE_REG	# read cause register
-	nop
-	and	v0, v0, ~SOFT_INT_MASK_1  # clear soft net interrupt
-	mtc0	v0, COP_0_CAUSE_REG	# save it
-	j	ra
-	nop
-END(clearsoftintr1)
-
-/*
- * Set/change interrupt priority routines.
- * These routines return the previous state.
- */
-LEAF(restoreintr)
-	mfc0	t0,COP_0_STATUS_REG
-	and	t1,t0,SR_INT_ENAB
-	beq	a0,t1,1f
-	xor	t0,SR_INT_ENAB
-  
-	.set	noreorder
-
-	mtc0	t0,COP_0_STATUS_REG
-	nop
-	nop
-	nop
-	nop
-1:
-	j	ra
-	nop
-END(restoreintr)
-
-/*
- * Set/change interrupt priority routines.
- * These routines return the previous state.
- */
- 
-LEAF(enableintr)
-#ifdef TARGET_OCTEON
-	.set mips64r2
-	ei     v0
-	and	v0, SR_INT_ENAB		# return old interrupt enable bit
-        .set	mips0
-#else		
-	mfc0	v0, COP_0_STATUS_REG	# read status register
-	nop
-	or	v1, v0, SR_INT_ENAB
-	mtc0	v1, COP_0_STATUS_REG	# enable all interrupts
-	and	v0, SR_INT_ENAB		# return old interrupt enable
-#endif	
-	j	ra
-	nop
-END(enableintr)
-
-
-LEAF(disableintr)
-#ifdef TARGET_OCTEON
-	.set mips64r2
-	di     v0
-	and	v0, SR_INT_ENAB		# return old interrupt enable bit
-	.set	mips0
-#else		
-	mfc0	v0, COP_0_STATUS_REG	# read status register
-	nop
-	and	v1, v0, ~SR_INT_ENAB
-	mtc0	v1, COP_0_STATUS_REG	# disable all interrupts
-	MIPS_CPU_NOP_DELAY
-	and	v0, SR_INT_ENAB		# return old interrupt enable
-#endif	
-	j	ra
-	nop
-END(disableintr)
-
 LEAF(set_intr_mask)
 	li	t0, SR_INT_MASK		# 1 means masked so invert.
 	not	a0, a0			# 1 means masked so invert.
@@ -182,17 +79,3 @@ LEAF(get_intr_mask)
 	nop
 
 END(get_intr_mask)
-
-/*
- * u_int32_t mips_cp0_config1_read(void)
- *
- *	Return the current value of the CP0 Config (Select 1) register.
- */
-LEAF(mips_cp0_config1_read)
-	.set	push
-	.set	mips32
-	mfc0	v0, COP_0_CONFIG, 1
-	j	ra
-	nop
-	.set	pop
-END(mips_cp0_config1_read)
diff --git a/sys/mips/mips/tick.c b/sys/mips/mips/tick.c
index 60b3511e0c0..d2678b1f637 100644
--- a/sys/mips/mips/tick.c
+++ b/sys/mips/mips/tick.c
@@ -301,16 +301,16 @@ clock_intr(void *arg)
 	if (cpu_ticks->hard_ticks >= cycles_per_hz) {
 	        cpu_ticks->hard_ticks -= cycles_per_hz;
 		if (PCPU_GET(cpuid) == 0)
-			hardclock(USERMODE(tf->sr), tf->pc);
+			hardclock(TRAPF_USERMODE(tf), tf->pc);
 		else
-			hardclock_cpu(USERMODE(tf->sr));
+			hardclock_cpu(TRAPF_USERMODE(tf));
 	}
 
 	/* Fire statclock at stathz. */
 	cpu_ticks->stat_ticks += delta;
 	if (cpu_ticks->stat_ticks >= cycles_per_stathz) {
 		cpu_ticks->stat_ticks -= cycles_per_stathz;
-		statclock(USERMODE(tf->sr));
+		statclock(TRAPF_USERMODE(tf));
 	}
 
 	/* Fire profclock at profhz, but only when needed. */
@@ -318,7 +318,7 @@ clock_intr(void *arg)
 	if (cpu_ticks->prof_ticks >= cycles_per_profhz) {
 		cpu_ticks->prof_ticks -= cycles_per_profhz;
 		if (profprocs != 0)
-			profclock(USERMODE(tf->sr), tf->pc);
+			profclock(TRAPF_USERMODE(tf), tf->pc);
 	}
 	critical_exit();
 #if 0 /* TARGET_OCTEON */
diff --git a/sys/mips/mips/trap.c b/sys/mips/mips/trap.c
index 8bfc6e7f4d5..de4bd01bf29 100644
--- a/sys/mips/mips/trap.c
+++ b/sys/mips/mips/trap.c
@@ -75,7 +75,6 @@ __FBSDID("$FreeBSD$");
 #include <net/netisr.h>
 
 #include <machine/trap.h>
-#include <machine/psl.h>
 #include <machine/cpu.h>
 #include <machine/pte.h>
 #include <machine/pmap.h>
@@ -272,7 +271,7 @@ extern char *syscallnames[];
  * In the case of a kernel trap, we return the pc where to resume if
  * p->p_addr->u_pcb.pcb_onfault is set, otherwise, return old pc.
  */
-u_int
+register_t
 trap(struct trapframe *trapframe)
 {
 	int type, usermode;
@@ -293,7 +292,7 @@ trap(struct trapframe *trapframe)
 	trapdebug_enter(trapframe, 0);
 	
 	type = (trapframe->cause & CR_EXC_CODE) >> CR_EXC_CODE_SHIFT;
-	if (USERMODE(trapframe->sr)) {
+	if (TRAPF_USERMODE(trapframe)) {
 		type |= T_USER;
 		usermode = 1;
 	} else {
@@ -307,9 +306,9 @@ trap(struct trapframe *trapframe)
 	 */
 	if (trapframe->sr & SR_INT_ENAB) {
 		set_intr_mask(~(trapframe->sr & ALL_INT_MASK));
-		enableintr();
+		intr_enable();
 	} else {
-		disableintr();
+		intr_disable();
 	}
 
 #ifdef TRAP_DEBUG
@@ -983,9 +982,10 @@ out:
 void
 trapDump(char *msg)
 {
-	int i, s;
+	register_t s;
+	int i;
 
-	s = disableintr();
+	s = intr_disable();
 	printf("trapDump(%s)\n", msg);
 	for (i = 0; i < TRAPSIZE; i++) {
 		if (trp == trapdebug) {
@@ -1003,9 +1003,8 @@ trapDump(char *msg)
 
 		printf("   RA %x SP %x code %d\n", trp->ra, trp->sp, trp->code);
 	}
-	restoreintr(s);
+	intr_restore(s);
 }
-
 #endif
 
 
diff --git a/sys/mips/rmi/clock.c b/sys/mips/rmi/clock.c
index 6964522e9c9..c123edf589e 100644
--- a/sys/mips/rmi/clock.c
+++ b/sys/mips/rmi/clock.c
@@ -116,11 +116,11 @@ count_compare_clockhandler(struct trapframe *tf)
 		cycles += XLR_CPU_HZ / hz;
 		mips_wr_compare(cycles);
 
-		hardclock_cpu(USERMODE(tf->sr));
+		hardclock_cpu(TRAPF_USERMODE(tf));
 		if (count_scale_factor[cpu] == STAT_PROF_CLOCK_SCALE_FACTOR) {
-			statclock(USERMODE(tf->sr));
+			statclock(TRAPF_USERMODE(tf));
 			if (profprocs != 0) {
-				profclock(USERMODE(tf->sr), tf->pc);
+				profclock(TRAPF_USERMODE(tf), tf->pc);
 			}
 			count_scale_factor[cpu] = 0;
 		}
@@ -148,11 +148,11 @@ pic_hardclockhandler(struct trapframe *tf)
 			printf("Clock tick foo at %ld\n", clock_tick_foo);
 		}
 */
-		hardclock(USERMODE(tf->sr), tf->pc);
+		hardclock(TRAPF_USERMODE(tf), tf->pc);
 		if (scale_factor == STAT_PROF_CLOCK_SCALE_FACTOR) {
-			statclock(USERMODE(tf->sr));
+			statclock(TRAPF_USERMODE(tf));
 			if (profprocs != 0) {
-				profclock(USERMODE(tf->sr), tf->pc);
+				profclock(TRAPF_USERMODE(tf), tf->pc);
 			}
 			scale_factor = 0;
 		}

From a97a1ee3d98b706bc84c8d89ff76ffb316f9fb1c Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 01:49:50 +0000
Subject: [PATCH 099/532] o) Back out my previous change to SWARM; some of it
 was to address an issue    that turned out to be unrelated, and the rest was,
 as pointed out by Neel,    just wrong-headed. o) Tweak mem.c to fix use of
 /dev/kmem for direct-mapped addresses.

---
 sys/mips/conf/SWARM |  5 -----
 sys/mips/mips/mem.c | 22 ++++++++++++----------
 2 files changed, 12 insertions(+), 15 deletions(-)

diff --git a/sys/mips/conf/SWARM b/sys/mips/conf/SWARM
index 455e2986f9c..3189bcbcb9c 100644
--- a/sys/mips/conf/SWARM
+++ b/sys/mips/conf/SWARM
@@ -19,13 +19,8 @@ options		CFE_CONSOLE
 options		CFE_ENV
 options		ALT_BREAK_TO_DEBUGGER
 
-# Don't build any modules yet.
-makeoptions	MODULES_OVERRIDE=""
-makeoptions	TARGET_BIG_ENDIAN=defined
 makeoptions	LDSCRIPT_NAME=	ldscript.mips.cfe
 
-makeoptions	KERNLOADADDR=0x81000000
-
 #cpu		CPU_MIPS64
 #options 	ISA_MIPS64
 #makeoptions	ARCH_FLAGS="-march=mips64 -mgp64 -mabi=o64"
diff --git a/sys/mips/mips/mem.c b/sys/mips/mips/mem.c
index a7f8244385a..9db282e4206 100644
--- a/sys/mips/mips/mem.c
+++ b/sys/mips/mips/mem.c
@@ -120,21 +120,23 @@ memrw(struct cdev *dev, struct uio *uio, int flags)
 			 * Make sure that all the pages are currently resident
 			 * so that we don't create any zero-fill pages.
 			 */
+			if (va >= VM_MIN_KERNEL_ADDRESS &&
+			    eva <= VM_MAX_KERNEL_ADDRESS) {
+				for (; va < eva; va += PAGE_SIZE)
+					if (pmap_extract(kernel_pmap, va) == 0)
+						return (EFAULT);
 
-			for (; va < eva; va += PAGE_SIZE)
-				if (pmap_extract(kernel_pmap, va) == 0)
+				prot = (uio->uio_rw == UIO_READ)
+				    ? VM_PROT_READ : VM_PROT_WRITE;
+
+				va = uio->uio_offset;
+				if (kernacc((void *) va, iov->iov_len, prot)
+				    == FALSE)
 					return (EFAULT);
-
-			prot = (uio->uio_rw == UIO_READ)
-			    ? VM_PROT_READ : VM_PROT_WRITE;
+			}
 
 			va = uio->uio_offset;
-			if (kernacc((void *) va, iov->iov_len, prot)
-			    == FALSE)
-				return (EFAULT);
-
 			error = uiomove((void *)va, iov->iov_len, uio);
-
 			continue;
 		}
 	}

From b6777295a6a08f4e5d76402dc98804aaa7cd76ee Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 03:08:13 +0000
Subject: [PATCH 100/532] o) Add SMP support for Octeon using U-Boot to launch
 all the processors at the    same time. o) Remove some unused trivial uart
 functions from octeon_machdep now that the    uart part is fully working and
 they are unused. o) Use __func__ instead of __FUNCTION__. o) Use intr_*()
 instead of other routines that do the same thing. o) Remove some duplicate
 printfs from the Octeon port, as well as duplicate    setting of Maxmem. o)
 Use the right frequency divider on Octeon. o) Use PCPU_GET(cpuid)
 consistently to get the cpuid of the running core. o) Remove some unused
 macros in the Octeon port. o) Use mips_sync() around use of the global dpcpu,
 whose value may not be    visible to APs at first. o) When loading the first
 thread's stack, use macros to make the code correct    for n64 as well. o)
 Remove stub, do-nothing FAU init/enable/disable functions from the RGMX   
 driver.

---
 sys/mips/cavium/asm_octeon.S            | 216 ++++++------------------
 sys/mips/cavium/dev/rgmii/octeon_fau.c  |  83 ---------
 sys/mips/cavium/dev/rgmii/octeon_fau.h  |   5 -
 sys/mips/cavium/dev/rgmii/octeon_rgmx.c |   2 +-
 sys/mips/cavium/files.octeon1           |   4 +-
 sys/mips/cavium/octeon_machdep.c        | 184 +++-----------------
 sys/mips/cavium/octeon_mp.c             | 102 +++++++++++
 sys/mips/cavium/octeon_pcmap_regs.h     |  90 ----------
 sys/mips/mips/locore.S                  |  18 +-
 sys/mips/mips/mp_machdep.c              |   4 +
 sys/mips/mips/mpboot.S                  |  15 +-
 11 files changed, 210 insertions(+), 513 deletions(-)
 delete mode 100644 sys/mips/cavium/dev/rgmii/octeon_fau.c
 create mode 100644 sys/mips/cavium/octeon_mp.c

diff --git a/sys/mips/cavium/asm_octeon.S b/sys/mips/cavium/asm_octeon.S
index d9f79f134ff..94ac875a199 100644
--- a/sys/mips/cavium/asm_octeon.S
+++ b/sys/mips/cavium/asm_octeon.S
@@ -1,182 +1,66 @@
-/***********************license start***************
- *  Copyright (c) 2003-2008 Cavium Networks (support@cavium.com). All rights
- *  reserved.
+/*-
+ * Copyright (c) 2004-2010 Juli Mallett <jmallett@FreeBSD.org>
+ * All rights reserved.
  *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions are
- *  met:
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
- *      * Redistributions of source code must retain the above copyright
- *        notice, this list of conditions and the following disclaimer.
- *
- *      * Redistributions in binary form must reproduce the above
- *        copyright notice, this list of conditions and the following
- *        disclaimer in the documentation and/or other materials provided
- *        with the distribution.
- *
- *      * Neither the name of Cavium Networks nor the names of
- *        its contributors may be used to endorse or promote products
- *        derived from this software without specific prior written
- *        permission.
- *
- *  TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE SOFTWARE IS PROVIDED "AS IS"
- *  AND WITH ALL FAULTS AND CAVIUM NETWORKS MAKES NO PROMISES, REPRESENTATIONS
- *  OR WARRANTIES, EITHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, WITH
- *  RESPECT TO THE SOFTWARE, INCLUDING ITS CONDITION, ITS CONFORMITY TO ANY
- *  REPRESENTATION OR DESCRIPTION, OR THE EXISTENCE OF ANY LATENT OR PATENT
- *  DEFECTS, AND CAVIUM SPECIFICALLY DISCLAIMS ALL IMPLIED (IF ANY) WARRANTIES
- *  OF TITLE, MERCHANTABILITY, NONINFRINGEMENT, FITNESS FOR A PARTICULAR
- *  PURPOSE, LACK OF VIRUSES, ACCURACY OR COMPLETENESS, QUIET ENJOYMENT, QUIET
- *  POSSESSION OR CORRESPONDENCE TO DESCRIPTION.  THE ENTIRE RISK ARISING OUT
- *  OF USE OR PERFORMANCE OF THE SOFTWARE LIES WITH YOU.
- *
- *
- *  For any questions regarding licensing please contact marketing@caviumnetworks.com
- *
- ***********************license end**************************************/
-
-/* $FreeBSD$ */
+ * $FreeBSD$
+ */
 
 #include <machine/asm.h>
-#include <machine/cache_r4k.h>
-#include <machine/cpuregs.h>
-#include <machine/param.h>
-#include <machine/pte.h>
 
-#include "assym.s"
-	
+	.set noreorder
 
-
-#define CPU_DISABLE_INTERRUPTS(reg, reg2, reg3) \
-        mfc0    reg, MIPS_COP_0_STATUS; \
-        nop; \
-        move    reg3, reg; \
-        li      reg2, ~MIPS_SR_INT_IE; \
-        and     reg, reg2, reg; \
-        mtc0    reg, MIPS_COP_0_STATUS; \
-        COP0_SYNC
-
-
-
-#define CPU_ENABLE_INTERRUPTS(reg, reg3) \
-        mfc0    reg, MIPS_COP_0_STATUS; \
-        nop; \
-        or      reg, reg, reg3; \
-        mtc0    reg, MIPS_COP_0_STATUS; \
-        COP0_SYNC
-
-
-#define PUSHR(reg) \
-        addiu   sp,sp,-16        ; \
-        sd      reg, 8(sp)      ; \
-        nop                     ; 
-
-#define POPR(reg) \
-        ld      reg, 8(sp)      ; \
-        addiu   sp,sp,16        ; \
-        nop                     ; 
-
-
-
-                
+#ifdef SMP
 /*
- * octeon_ciu_get_interrupt_reg_addr
- *
- * Given  Int-X, En-X combination, return the CIU Interrupt Enable Register addr
- * a0 = ciu Int-X:  0/1
- * a1 = ciu EN-0:   0/1
+ * This function must be implemented in assembly because it is called early
+ * in AP boot without a valid stack.
  */
-LEAF(octeon_ciu_get_interrupt_reg_addr)
-        .set    noreorder
-        .set    mips3
+LEAF(platform_processor_id)
+	.set push
+	.set mips32r2
+	jr	ra
+	rdhwr	v0, $0
+	.set pop
+END(platform_processor_id)
 
-        beqz    a0, ciu_get_interrupt_reg_addr_Int_0
-        nop
-
-ciu_get_interrupt_reg_addr_Int_1:
-        beqz    a1, ciu_get_interrupt_reg_addr_Int_1_En_0
-        nop
-
-ciu_get_interrupt_reg_addr_Int_1_En1:
-        li      a0, OCTEON_CIU_ADDR_HI
-        dsll32  a0, a0, 0
-        nop
-        ori      a0, OCTEON_CIU_EN1_INT1_LO
-        j       ciu_get_interrupt_reg_addr_ret
-        nop
-
-ciu_get_interrupt_reg_addr_Int_1_En_0:
-        li      a0, OCTEON_CIU_ADDR_HI
-        dsll32  a0, a0, 0
-        nop
-        ori     a0, OCTEON_CIU_EN0_INT1_LO
-        j       ciu_get_interrupt_reg_addr_ret
-        nop
-
-ciu_get_interrupt_reg_addr_Int_0:
-        beqz    a1, ciu_get_interrupt_reg_addr_Int_0_En_0
-        nop
-
-ciu_get_interrupt_reg_addr_Int_0_En_1:
-        li      a0, OCTEON_CIU_ADDR_HI
-        dsll32  a0, a0, 0
-        nop
-        ori     a0, OCTEON_CIU_EN1_INT0_LO
-        j       ciu_get_interrupt_reg_addr_ret
-        nop
-
-ciu_get_interrupt_reg_addr_Int_0_En_0:
-        li      a0, OCTEON_CIU_ADDR_HI
-        dsll32  a0, a0, 0
-        nop
-        ori     a0, OCTEON_CIU_EN0_INT0_LO
-                
-        
-ciu_get_interrupt_reg_addr_ret: 
-        j       ra
-        nop
-        
-        .set	mips0
-        .set    reorder
-END(octeon_ciu_get_interrupt_reg_addr)
-
-
-                
 /*
- * octeon_ciu_mask_all_interrupts
- *
- * a0 = ciu Interrupt-X:  0/1
- * a1 = ciu Enable-X:   0/1
+ * Called on APs to wait until they are told to launch.
  */
-LEAF(octeon_ciu_mask_all_interrupts)
-	.set    noreorder
-        .set    mips3
+LEAF(octeon_ap_wait)
+	jal	platform_processor_id
+	nop
 
-        PUSHR(ra)
-        PUSHR(s0)
-        
-        move    t0, a0
-        move    t1, a1
-        li      a0, MIPS_SR_INT_IE
-        CPU_DISABLE_INTERRUPTS(a2, a1, s0)
-        move    a0, t0
-        move    t1, a1
-        jal     octeon_ciu_get_interrupt_reg_addr
-        nop
-        ld      a2, 0(a0)       # Dummy read
-        nop
-        move    a2, zero        # Clear all
-        sd      a2, 0(a0)       # Write new Enable bits
-        nop
-        CPU_ENABLE_INTERRUPTS(a2, s0)
+1:	ll	t0, octeon_ap_boot
+	bne	v0, t0, 1b
+	nop
 
-        POPR(s0)
-        POPR(ra)
-	j	ra			# Return
-	nop				# (bd slot)
+	move	t0, zero
+	sc	t0, octeon_ap_boot
 
-        .set	mips0
-	.set	reorder
-END(octeon_ciu_mask_all_interrupts)
+	beqz	t0, 1b
+	nop
 
+	j	mpentry
+	nop
+END(octeon_ap_wait)
+#endif
diff --git a/sys/mips/cavium/dev/rgmii/octeon_fau.c b/sys/mips/cavium/dev/rgmii/octeon_fau.c
deleted file mode 100644
index fb4c0ad3a80..00000000000
--- a/sys/mips/cavium/dev/rgmii/octeon_fau.c
+++ /dev/null
@@ -1,83 +0,0 @@
-/***********************license start***************
- *  Copyright (c) 2003-2008 Cavium Networks (support@cavium.com). All rights
- *  reserved.
- *
- *
- *  Redistribution and use in source and binary forms, with or without
- *  modification, are permitted provided that the following conditions are
- *  met:
- *
- *      * Redistributions of source code must retain the above copyright
- *        notice, this list of conditions and the following disclaimer.
- *
- *      * Redistributions in binary form must reproduce the above
- *        copyright notice, this list of conditions and the following
- *        disclaimer in the documentation and/or other materials provided
- *        with the distribution.
- *
- *      * Neither the name of Cavium Networks nor the names of
- *        its contributors may be used to endorse or promote products
- *        derived from this software without specific prior written
- *        permission.
- *
- *  TO THE MAXIMUM EXTENT PERMITTED BY LAW, THE SOFTWARE IS PROVIDED "AS IS"
- *  AND WITH ALL FAULTS AND CAVIUM NETWORKS MAKES NO PROMISES, REPRESENTATIONS
- *  OR WARRANTIES, EITHER EXPRESS, IMPLIED, STATUTORY, OR OTHERWISE, WITH
- *  RESPECT TO THE SOFTWARE, INCLUDING ITS CONDITION, ITS CONFORMITY TO ANY
- *  REPRESENTATION OR DESCRIPTION, OR THE EXISTENCE OF ANY LATENT OR PATENT
- *  DEFECTS, AND CAVIUM SPECIFICALLY DISCLAIMS ALL IMPLIED (IF ANY) WARRANTIES
- *  OF TITLE, MERCHANTABILITY, NONINFRINGEMENT, FITNESS FOR A PARTICULAR
- *  PURPOSE, LACK OF VIRUSES, ACCURACY OR COMPLETENESS, QUIET ENJOYMENT, QUIET
- *  POSSESSION OR CORRESPONDENCE TO DESCRIPTION.  THE ENTIRE RISK ARISING OUT
- *  OF USE OR PERFORMANCE OF THE SOFTWARE LIES WITH YOU.
- *
- *
- *  For any questions regarding licensing please contact marketing@caviumnetworks.com
- *
- ***********************license end**************************************/
-
-/*------------------------------------------------------------------
- * octeon_fau.c        Fetch & Add Block
- *
- *------------------------------------------------------------------
- */
-
-#include <sys/cdefs.h>
-__FBSDID("$FreeBSD$");
-
-#include <sys/param.h>
-#include <sys/systm.h>
-
-#include <mips/cavium/octeon_pcmap_regs.h>
-#include "octeon_fau.h"
-
-/*
- * oct_fau_init
- *
- * How do we initialize FAU unit. I don't even think we can reset it.
- */
-void octeon_fau_init (void)
-{
-}
-
-
-/*
- * oct_fau_enable
- *
- * Let the Fetch/Add unit roll
- */
-void octeon_fau_enable (void)
-{
-}
-
-
-/*
- * oct_fau_disable
- *
- * disable fau
- *
- * Don't know if we can even do that.
- */
-void octeon_fau_disable (void)
-{
-}
diff --git a/sys/mips/cavium/dev/rgmii/octeon_fau.h b/sys/mips/cavium/dev/rgmii/octeon_fau.h
index 1683ff55e00..acf5132e2e8 100644
--- a/sys/mips/cavium/dev/rgmii/octeon_fau.h
+++ b/sys/mips/cavium/dev/rgmii/octeon_fau.h
@@ -217,9 +217,4 @@ static inline void octeon_fau_atomic_add64 (octeon_fau_reg_64_t reg, int64_t val
 }
 
 
-extern void octeon_fau_init(void);
-extern void octeon_fau_enable(void);
-extern void octeon_fau_disable(void);
-
-
 #endif  /* ___OCTEON_FAU__H___ */
diff --git a/sys/mips/cavium/dev/rgmii/octeon_rgmx.c b/sys/mips/cavium/dev/rgmii/octeon_rgmx.c
index ada09461778..5b84af5dcaa 100644
--- a/sys/mips/cavium/dev/rgmii/octeon_rgmx.c
+++ b/sys/mips/cavium/dev/rgmii/octeon_rgmx.c
@@ -1481,7 +1481,7 @@ static void octeon_config_hw_units_post_ports (void)
         oct_write64(OCTEON_POW_WORKQUEUE_INT_THRESHOLD(OCTEON_POW_RX_GROUP_NUM), thr.word64);
 #endif
 
-        ciu_enable_interrupts(OCTEON_CORE_ID, OCTEON_RGMX_CIU_INTX, OCTEON_RGMX_CIU_ENX,
+        ciu_enable_interrupts(PCPU_GET(cpuid), OCTEON_RGMX_CIU_INTX, OCTEON_RGMX_CIU_ENX,
                               (OCTEON_POW_RX_GROUP_MASK |
                                CIU_GENTIMER_BITS_ENABLE(CIU_GENTIMER_NUM_1)), CIU_MIPS_IP2);
 
diff --git a/sys/mips/cavium/files.octeon1 b/sys/mips/cavium/files.octeon1
index c10988c46c1..4cf31f84310 100644
--- a/sys/mips/cavium/files.octeon1
+++ b/sys/mips/cavium/files.octeon1
@@ -1,8 +1,7 @@
 # $FreeBSD$
 # Octeon Support Files
 #
-mips/mips/mp_machdep.c				optional smp
-mips/cavium/dev/rgmii/octeon_fau.c		optional rgmii
+mips/cavium/asm_octeon.S			optional smp
 mips/cavium/dev/rgmii/octeon_fpa.c		optional rgmii
 mips/cavium/dev/rgmii/octeon_ipd.c 		optional rgmii
 mips/cavium/dev/rgmii/octeon_pko.c		optional rgmii
@@ -10,6 +9,7 @@ mips/cavium/dev/rgmii/octeon_rgmx.c		optional rgmii
 mips/cavium/obio.c				optional uart
 mips/cavium/octeon_ebt3000_cf.c			optional cf
 mips/cavium/octeon_machdep.c			standard
+mips/cavium/octeon_mp.c				optional smp
 mips/cavium/uart_bus_octeonusart.c		optional uart
 mips/cavium/uart_cpu_octeonusart.c		optional uart
 mips/cavium/uart_dev_oct16550.c			optional uart
diff --git a/sys/mips/cavium/octeon_machdep.c b/sys/mips/cavium/octeon_machdep.c
index f55c7ea4dfb..c796071a385 100644
--- a/sys/mips/cavium/octeon_machdep.c
+++ b/sys/mips/cavium/octeon_machdep.c
@@ -111,25 +111,6 @@ platform_reset(void)
 	oct_write64(OCTEON_CIU_SOFT_RST, 1);
 }
 
-
-static inline uint32_t
-octeon_disable_interrupts(void)
-{
-	uint32_t status_bits;
-
-	status_bits = mips_rd_status();
-	mips_wr_status(status_bits & ~MIPS_SR_INT_IE);
-	return (status_bits);
-}
-
-
-static inline void
-octeon_set_interrupts(uint32_t status_bits)
-{
-	mips_wr_status(status_bits);
-}
-
-
 void
 octeon_led_write_char(int char_position, char val)
 {
@@ -203,82 +184,6 @@ octeon_led_run_wheel(int *prog_count, int led_position)
 	*prog_count &= 0x7;
 }
 
-#define LSR_DATAREADY        0x01    /* Data ready */
-#define LSR_THRE             0x20    /* Transmit holding register empty */
-#define LSR_TEMT	     0x40    /* Transmitter Empty. THR, TSR & FIFO */
-#define USR_TXFIFO_NOTFULL   0x02    /* Uart TX FIFO Not full */
-
-/*
- * octeon_uart_write_byte
- * 
- * Put out a single byte off of uart port.
- */
-
-void
-octeon_uart_write_byte(int uart_index, uint8_t ch)
-{
-	uint64_t val, val2;
-	if (uart_index < 0 || uart_index > 1)
-		return;
-
-	while (1) {
-		val = oct_read64(OCTEON_MIO_UART0_LSR + (uart_index * 0x400));
-		val2 = oct_read64(OCTEON_MIO_UART0_USR + (uart_index * 0x400));
-		if ((((uint8_t) val) & LSR_THRE) ||
-		    (((uint8_t) val2) & USR_TXFIFO_NOTFULL)) {
-			break;
-		}
-	}
-
-	/* Write the byte */
-	oct_write8(OCTEON_MIO_UART0_THR + (uart_index * 0x400), (uint64_t) ch);
-
-	/* Force Flush the IOBus */
-	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);
-}
-
-
-void
-octeon_uart_write_byte0(uint8_t ch)
-{
-	uint64_t val, val2;
-
-	while (1) {
-		val = oct_read64(OCTEON_MIO_UART0_LSR);
-		val2 = oct_read64(OCTEON_MIO_UART0_USR);
-		if ((((uint8_t) val) & LSR_THRE) ||
-		    (((uint8_t) val2) & USR_TXFIFO_NOTFULL)) {
-			break;
-		}
-	}
-
-	/* Write the byte */
-	oct_write8(OCTEON_MIO_UART0_THR, (uint64_t) ch);
-
-	/* Force Flush the IOBus */
-	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);
-}
-
-/*
- * octeon_uart_write_string
- * 
- */
-void
-octeon_uart_write_string(int uart_index, const char *str)
-{
-	/* Just loop writing one byte at a time */
-    
-	while (*str) {
-		octeon_uart_write_byte(uart_index, *str);
-		if (*str == '\n') {
-			octeon_uart_write_byte(uart_index, '\r');
-		}
-		str++;
-	}
-}
-
-static char wstr[30];
-
 void
 octeon_led_write_hex(uint32_t wl)
 {
@@ -289,44 +194,6 @@ octeon_led_write_hex(uint32_t wl)
 }
 
 
-void octeon_uart_write_hex2(uint32_t wl, uint32_t wh)
-{
-	sprintf(wstr, "0x%X-0x%X  ", wh, wl);
-	octeon_uart_write_string(0, wstr);
-}
-
-void
-octeon_uart_write_hex(uint32_t wl)
-{
-	sprintf(wstr, " 0x%X  ", wl);
-	octeon_uart_write_string(0, wstr);
-}
-
-/*
- * octeon_wait_uart_flush
- */
-void
-octeon_wait_uart_flush(int uart_index, uint8_t ch)
-{
-	uint64_t val;
-	int64_t val3;
-	uint32_t cpu_status_bits;
-
-	if (uart_index < 0 || uart_index > 1)
-		return;
-
-	cpu_status_bits = octeon_disable_interrupts();
-	/* Force Flush the IOBus */
-	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);
-	for (val3 = 0xfffffffff; val3 > 0; val3--) {
-		val = oct_read64(OCTEON_MIO_UART0_LSR + (uart_index * 0x400));
-		if (((uint8_t) val) & LSR_TEMT)
-			break;
-	}
-	octeon_set_interrupts(cpu_status_bits);
-}
-
-
 /*
  * octeon_debug_symbol
  *
@@ -450,17 +317,17 @@ ciu_get_en_reg_addr_new(int corenum, int intx, int enx, int ciu_ip)
 	/* XXX kasserts? */
 	if (enx < CIU_EN_0 || enx > CIU_EN_1) {
 		printf("%s: invalid enx value %d, should be %d or %d\n",
-		    __FUNCTION__, enx, CIU_EN_0, CIU_EN_1);
+		    __func__, enx, CIU_EN_0, CIU_EN_1);
 		return 0;
 	}
 	if (intx < CIU_INT_0 || intx > CIU_INT_1) {
 		printf("%s: invalid intx value %d, should be %d or %d\n",
-		    __FUNCTION__, enx, CIU_INT_0, CIU_INT_1);
+		    __func__, enx, CIU_INT_0, CIU_INT_1);
 		return 0;
 	}
 	if (ciu_ip < CIU_MIPS_IP2 || ciu_ip > CIU_MIPS_IP3) {
 		printf("%s: invalid ciu_ip value %d, should be %d or %d\n",
-		    __FUNCTION__, ciu_ip, CIU_MIPS_IP2, CIU_MIPS_IP3);
+		    __func__, ciu_ip, CIU_MIPS_IP2, CIU_MIPS_IP3);
 		return 0;
 	}
 
@@ -517,7 +384,7 @@ ciu_clear_int_summary(int core_num, int intx, int enx, uint64_t write_bits)
 	    core_num, intx, enx, write_bits);
 #endif
 
-	cpu_status_bits = octeon_disable_interrupts();
+	cpu_status_bits = intr_disable();
 
 	ciu_intr_sum_reg_addr = ciu_get_intr_sum_reg_addr(core_num, intx, enx);
 
@@ -535,7 +402,7 @@ ciu_clear_int_summary(int core_num, int intx, int enx, uint64_t write_bits)
         printf(" Readback: 0x%llX\n\n   ", (uint64_t) oct_read64(ciu_intr_sum_reg_addr));
 #endif
     
-	octeon_set_interrupts(cpu_status_bits);
+	intr_restore(cpu_status_bits);
 }
 
 /*
@@ -550,7 +417,7 @@ ciu_disable_intr(int core_num, int intx, int enx)
 	if (core_num == CIU_THIS_CORE)
         	core_num = octeon_get_core_num();
 
-	cpu_status_bits = octeon_disable_interrupts();
+	cpu_status_bits = intr_disable();
     
 	ciu_intr_reg_addr = ciu_get_intr_en_reg_addr(core_num, intx, enx);
 
@@ -559,7 +426,7 @@ ciu_disable_intr(int core_num, int intx, int enx)
 	oct_write64(ciu_intr_reg_addr, 0LL);
 	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);	/* Bus Barrier */
 
-	octeon_set_interrupts(cpu_status_bits);
+	intr_restore(cpu_status_bits);
 }
 
 void
@@ -580,7 +447,7 @@ ciu_dump_interrutps_enabled(int core_num, int intx, int enx, int ciu_ip)
 #endif
 
         if (!ciu_intr_reg_addr) {
-            printf("Bad call to %s\n", __FUNCTION__);
+            printf("Bad call to %s\n", __func__);
             while(1);
             return;
         }
@@ -612,7 +479,7 @@ void ciu_enable_interrupts(int core_num, int intx, int enx,
 	    core_num, intx, enx, ciu_ip, set_these_interrupt_bits);
 #endif
 
-	cpu_status_bits = octeon_disable_interrupts();
+	cpu_status_bits = intr_disable();
 
 #ifndef OCTEON_SMP_1
 	ciu_intr_reg_addr = ciu_get_intr_en_reg_addr(core_num, intx, enx);
@@ -621,7 +488,7 @@ void ciu_enable_interrupts(int core_num, int intx, int enx,
 #endif
 
         if (!ciu_intr_reg_addr) {
-		printf("Bad call to %s\n", __FUNCTION__);
+		printf("Bad call to %s\n", __func__);
 		while(1);
 		return;	/* XXX */
         }
@@ -634,7 +501,7 @@ void ciu_enable_interrupts(int core_num, int intx, int enx,
 #endif
 	ciu_intr_bits |=  set_these_interrupt_bits;
 	oct_write64(ciu_intr_reg_addr, ciu_intr_bits);
-#ifdef OCTEON_SMP
+#ifdef SMP
 	mips_wbflush();
 #endif
 	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);	/* Bus Barrier */
@@ -644,7 +511,7 @@ void ciu_enable_interrupts(int core_num, int intx, int enx,
 	    (uint64_t)oct_read64(ciu_intr_reg_addr));
 #endif
 
-	octeon_set_interrupts(cpu_status_bits);
+	intr_restore(cpu_status_bits);
 }
 
 unsigned long
@@ -659,12 +526,8 @@ octeon_memory_init(void)
 	uint32_t realmem_bytes;
 
 	if (octeon_board_real()) {
-		printf("octeon_dram == %jx\n", (intmax_t)octeon_dram);
-		printf("reduced to ram: %u MB", (uint32_t)octeon_dram >> 20);
-
 		realmem_bytes = (octeon_dram - PAGE_SIZE);
 		realmem_bytes &= ~(PAGE_SIZE - 1);
-		printf("Real memory bytes is %x\n", realmem_bytes);
 	} else {
 		/* Simulator we limit to 96 meg */
 		realmem_bytes = (96 << 20);
@@ -678,8 +541,6 @@ octeon_memory_init(void)
 			phys_avail[1] = realmem_bytes;
 		realmem_bytes -= OCTEON_DRAM_FIRST_256_END;
 		realmem_bytes &= ~(PAGE_SIZE - 1);
-		printf("phys_avail[0] = %#lx phys_avail[1] = %#lx\n",
-		       (long)phys_avail[0], (long)phys_avail[1]);
 	} else {
 		/* Simulator gets 96Meg period. */
 		phys_avail[1] = (96 << 20);
@@ -705,23 +566,14 @@ octeon_memory_init(void)
 		realmem_bytes &= ~(PAGE_SIZE - 1);
 		/* Now map the rest of the memory */
 		phys_avail[2] = 0x20000000;
-		printf("realmem_bytes is now at %x\n", realmem_bytes);
 		phys_avail[3] = ((uint32_t) 0x20000000 + realmem_bytes);
-		printf("Next block of memory goes from %#lx to %#lx\n",
-		    (long)phys_avail[2], (long)phys_avail[3]);
 		physmem += btoc(phys_avail[3] - phys_avail[2]);
-	} else {
-		printf("realmem_bytes is %d\n", realmem_bytes);
 	}
 	realmem = physmem;
 
 	printf("Total DRAM Size %#X\n", (uint32_t) octeon_dram);
 	printf("Bank 0 = %#08lX   ->  %#08lX\n", (long)phys_avail[0], (long)phys_avail[1]);
 	printf("Bank 1 = %#08lX   ->  %#08lX\n", (long)phys_avail[2], (long)phys_avail[3]);
-	printf("physmem: %#lx\n", physmem);
-
-	Maxmem = physmem;
-
 }
 
 void
@@ -760,7 +612,15 @@ platform_start(__register_t a0, __register_t a1, __register_t a2 __unused,
 		kdb_enter(KDB_WHY_BOOTFLAGS, "Boot flags requested debugger");
 #endif
 	platform_counter_freq = octeon_get_clock_rate();
-	mips_timer_init_params(platform_counter_freq, 1);
+	mips_timer_init_params(platform_counter_freq, 0);
+
+#ifdef SMP
+	/*
+	 * Clear any pending IPIs and enable the IPI interrupt.
+	 */
+	oct_write64(OCTEON_CIU_MBOX_CLRX(0), 0xffffffff);
+	ciu_enable_interrupts(0, CIU_INT_1, CIU_EN_0, OCTEON_CIU_ENABLE_MBOX_INTR, CIU_MIPS_IP3);
+#endif
 }
 
 /* impSTART: This stuff should move back into the Cavium SDK */
@@ -982,7 +842,7 @@ octeon_boot_params_init(register_t ptr)
 
         printf("Boot Descriptor Ver: %u -> %u/%u",
                octeon_bd_ver, octeon_cvmx_bd_ver/100, octeon_cvmx_bd_ver%100);
-        printf("  CPU clock: %uMHz\n", octeon_cpu_clock/1000000);
+        printf("  CPU clock: %uMHz  Core Mask: %#x\n", octeon_cpu_clock/1000000, octeon_core_mask);
         printf("  Dram: %u MB", (uint32_t)(octeon_dram >> 20));
         printf("  Board Type: %u  Revision: %u/%u\n",
                octeon_board_type, octeon_board_rev_major, octeon_board_rev_minor);
diff --git a/sys/mips/cavium/octeon_mp.c b/sys/mips/cavium/octeon_mp.c
new file mode 100644
index 00000000000..85cdc45dadd
--- /dev/null
+++ b/sys/mips/cavium/octeon_mp.c
@@ -0,0 +1,102 @@
+/*-
+ * Copyright (c) 2004-2010 Juli Mallett <jmallett@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/conf.h>
+#include <sys/kernel.h>
+#include <sys/systm.h>
+
+#include <machine/hwfunc.h>
+#include <machine/smp.h>
+
+#include <mips/cavium/octeon_pcmap_regs.h>
+
+unsigned octeon_ap_boot = ~0;
+
+void
+platform_ipi_send(int cpuid)
+{
+	oct_write64(OCTEON_CIU_MBOX_SETX(cpuid), 1);
+	mips_wbflush();
+}
+
+void
+platform_ipi_clear(void)
+{
+	uint64_t action;
+
+	action = oct_read64(OCTEON_CIU_MBOX_CLRX(PCPU_GET(cpuid)));
+	KASSERT(action == 1, ("unexpected IPIs: %#jx", (uintmax_t)action));
+	oct_write64(OCTEON_CIU_MBOX_CLRX(PCPU_GET(cpuid)), action);
+}
+
+int
+platform_ipi_intrnum(void)
+{
+	return (1);
+}
+
+void
+platform_init_ap(int cpuid)
+{
+	/*
+	 * Set the exception base.
+	 */
+	mips_wr_prid1(0x80000000 | cpuid);
+
+	/*
+	 * Set up interrupts, clear IPIs and unmask the IPI interrupt.
+	 */
+	octeon_ciu_reset();
+
+	oct_write64(OCTEON_CIU_MBOX_CLRX(cpuid), 0xffffffff);
+	ciu_enable_interrupts(cpuid, CIU_INT_1, CIU_EN_0, OCTEON_CIU_ENABLE_MBOX_INTR, CIU_MIPS_IP3);
+
+	mips_wbflush();
+}
+
+int
+platform_num_processors(void)
+{
+	return (fls(octeon_core_mask));
+}
+
+int
+platform_start_ap(int cpuid)
+{
+	if (atomic_cmpset_32(&octeon_ap_boot, ~0, cpuid) == 0)
+		return (-1);
+	for (;;) {
+		DELAY(1000);
+		if (atomic_cmpset_32(&octeon_ap_boot, 0, ~0) != 0)
+			return (0);
+		printf("Waiting for cpu%d to start\n", cpuid);
+	}
+}
diff --git a/sys/mips/cavium/octeon_pcmap_regs.h b/sys/mips/cavium/octeon_pcmap_regs.h
index 16dff5b35c9..0ee1a73ea8a 100644
--- a/sys/mips/cavium/octeon_pcmap_regs.h
+++ b/sys/mips/cavium/octeon_pcmap_regs.h
@@ -54,14 +54,6 @@
 
 #ifndef LOCORE
 
-/* XXXimp: From Cavium's include/pcpu.h, need to port that over */
-#ifndef OCTEON_SMP
-#define OCTEON_CORE_ID 0
-#else
-extern struct pcpu *cpuid_to_pcpu[];
-#define OCTEON_CORE_ID (mips_rd_coreid())
-#endif
-
 /*
  * Utility inlines & macros
  */
@@ -324,62 +316,6 @@ static inline void oct_write32 (uint64_t csr_addr, uint32_t val32)
 #define OCTEON_SCRATCH_2   32
 
 
-static inline uint64_t oct_mf_chord (void)
-{
-    uint64_t dest;
-
-    __asm __volatile (	".set push\n"
-                        ".set noreorder\n"
-                        ".set noat\n"
-                        ".set mips64\n"
-			"dmfc2 $1, 0x400\n"
-                        "move %0, $1\n"
-        		".set pop\n"
- 			: "=r" (dest) :  : "$1");
-    return dest;
-}
-
-
-#define MIPS64_DMFCz(cop,regnum,cp0reg,select)  \
-        .word   (0x40200000 | (cop << 25) | (regnum << 16) | (cp0reg << 11) | select)
-
-
-#define mips64_getcpz_xstr(s) mips64_getcpz_str(s)
-#define mips64_getcpz_str(s) #s
-
-#define mips64_dgetcpz(cop,cpzreg,sel,val_ptr) \
-    ({ __asm __volatile( \
-            ".set push\n" \
-            ".set mips3\n" \
-            ".set noreorder\n" \
-            ".set noat\n" \
-            mips64_getcpz_xstr(MIPS64_DMFCz(cop,1,cpzreg,sel)) "\n" \
-            "nop\n" \
-            "nop\n" \
-            "nop\n" \
-            "nop\n" \
-            "sd $1,0(%0)\n" \
-            ".set pop" \
-            : /* no outputs */ : "r" (val_ptr) : "$1"); \
-    })
-
-
-#define mips64_dgetcp2(cp2reg,sel,retval_ptr) \
-    mips64_dgetcpz(2,cp2reg,sel,retval_ptr)
-
-
-#define OCTEON_MF_CHORD(dest)  mips64_dgetcp2(0x400, 0, &dest)
-
-
-
-#define OCTEON_RDHWR(result, regstr) \
-	__asm __volatile (		\
-        		".set mips3\n"		\
-			"rdhwr %0,$" OCTEON_TMP_STR(regstr) "\n"	\
-        		".set mips\n"		\
-			 : "=d" (result));
-
-#define CVMX_MF_CHORD(dest)         OCTEON_RDHWR(dest, 30)
 
 #define OCTEON_CHORD_HEX(dest_ptr)  \
     ({ __asm __volatile( \
@@ -397,15 +333,6 @@ static inline uint64_t oct_mf_chord (void)
             : /* no outputs */ : "r" (dest_ptr) : "$2"); \
     })
 
-
-
-#define OCTEON_MF_CHORD_BAD(dest)	\
-         __asm __volatile (		\
-        		".set mips3\n"		\
-			"dmfc2 %0, 0x400\n"	\
-        		".set mips0\n"		\
- 			: "=&r" (dest) : )
-
 static inline uint64_t oct_scratch_read64 (uint64_t address)
 {
     return(*((volatile uint64_t *)(OCTEON_SCRATCH_BASE + address)));
@@ -417,17 +344,6 @@ static inline void oct_scratch_write64 (uint64_t address, uint64_t value)
 }
 
 
-#define OCTEON_READ_CSR32(addr, val) \
-	addr_ptr = addr; \
-	oct_read_32_ptr(&addr_ptr, &val);
-
-#define OCTEON_WRITE_CSR32(addr, val, val_dummy) \
-	addr_ptr = addr; \
-	oct_write_32_ptr(&addr_ptr, &val); \
-	oct_read64(OCTEON_MIO_BOOT_BIST_STAT);
-
-
-
 /*
  * Octeon Address Space Definitions
  */
@@ -791,12 +707,6 @@ extern void octeon_led_write_hexchar(int char_position, char hexval);
 extern void octeon_led_write_hex(uint32_t wl);
 extern void octeon_led_write_string(const char *str);
 extern void octeon_reset(void);
-extern void octeon_uart_write_byte(int uart_index, uint8_t ch);
-extern void octeon_uart_write_string(int uart_index, const char *str);
-extern void octeon_uart_write_hex(uint32_t wl);
-extern void octeon_uart_write_hex2(uint32_t wl, uint32_t wh);
-extern void octeon_wait_uart_flush(int uart_index, uint8_t ch);
-extern void octeon_uart_write_byte0(uint8_t ch);
 extern void octeon_led_write_char0(char val);
 extern void octeon_led_run_wheel(int *pos, int led_position);
 extern void octeon_debug_symbol(void);
diff --git a/sys/mips/mips/locore.S b/sys/mips/mips/locore.S
index 1574d7542a1..697b81884f5 100644
--- a/sys/mips/mips/locore.S
+++ b/sys/mips/mips/locore.S
@@ -162,6 +162,18 @@ VECTOR(_locore, unknown)
 	sw	a2, _C_LABEL(fenvp)
 #endif
 
+#if defined(TARGET_OCTEON) && defined(SMP)
+	.set push
+	.set mips32r2
+	rdhwr	t2, $0
+	beqz	t2, 1f
+	nop
+	j	octeon_ap_wait
+	nop
+	.set pop
+1:
+#endif
+
 	/*
 	 * Initialize stack and call machine startup.
 	 */
@@ -178,10 +190,10 @@ VECTOR(_locore, unknown)
 	nop
 
 	PTR_LA	sp, _C_LABEL(thread0)
-	lw      a0, TD_PCB(sp)
-	li	t0, ~7
+	PTR_L	a0, TD_PCB(sp)
+	REG_LI	t0, ~7
 	and	a0, a0, t0
-	subu    sp, a0, CALLFRAME_SIZ
+	PTR_SUBU	sp, a0, CALLFRAME_SIZ
 
 	jal	_C_LABEL(mi_startup)		# mi_startup(frame)
 	sw	zero, CALLFRAME_SIZ - 8(sp)	# Zero out old fp for debugger
diff --git a/sys/mips/mips/mp_machdep.c b/sys/mips/mips/mp_machdep.c
index 52d05fad711..f0c142f0c3a 100644
--- a/sys/mips/mips/mp_machdep.c
+++ b/sys/mips/mips/mp_machdep.c
@@ -157,6 +157,8 @@ start_ap(int cpuid)
 	cpus = mp_naps;
 	dpcpu = (void *)kmem_alloc(kernel_map, DPCPU_SIZE);
 
+	mips_sync();
+
 	if (platform_start_ap(cpuid) != 0)
 		return (-1);			/* could not start AP */
 
@@ -246,6 +248,8 @@ smp_init_secondary(u_int32_t cpuid)
 	mips_dcache_wbinv_all();
 	mips_icache_sync_all();
 
+	mips_sync();
+
 	MachSetPID(0);
 
 	pcpu_init(PCPU_ADDR(cpuid), cpuid, sizeof(struct pcpu));
diff --git a/sys/mips/mips/mpboot.S b/sys/mips/mips/mpboot.S
index ce2982d4005..e1c578a7961 100644
--- a/sys/mips/mips/mpboot.S
+++ b/sys/mips/mips/mpboot.S
@@ -36,8 +36,21 @@
 	.set	noat
 	.set	noreorder
 
+#ifdef TARGET_OCTEON
+#define CLEAR_STATUS \
+	mfc0    a0, COP_0_STATUS_REG    ;\
+	li      a2, (MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX) ; \
+	or      a0, a0, a2	        ; \
+	li      a2, ~(MIPS_SR_INT_IE | MIPS_SR_EXL | SR_KSU_USER | MIPS_SR_BEV)   ; \
+	and     a0, a0, a2              ; \
+        mtc0    a0, COP_0_STATUS_REG    
+#else
+#define CLEAR_STATUS \
+	mtc0	zero, COP_0_STATUS_REG
+#endif
+
 GLOBAL(mpentry)
-	mtc0	zero, COP_0_STATUS_REG	/* disable interrupts */
+	CLEAR_STATUS			/* disable interrupts */
 
 	mtc0	zero, COP_0_CAUSE_REG	/* clear soft interrupts */
 

From 2ab78e3ca5e86ce64742020d09b4dcecd067c23c Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 07:20:01 +0000
Subject: [PATCH 101/532] o) Add NPDEPG, like NPTEPG but for PDEs. o) Remove
 NBPG, PGOFSET and PGSHIFT.  Use the standard names. o) Remove some unused
 macros and move things from param.h to vmparam.h that    belong in the
 latter.  (Actually, all of the kernel segment values, virtual    addresses,
 etc., belong in one place, but this is a step in the right    direction.)

---
 sys/mips/include/param.h   | 40 +++++++++++++-------------------------
 sys/mips/include/pte.h     |  2 +-
 sys/mips/include/vmparam.h | 31 +++++++++++------------------
 sys/mips/mips/exception.S  |  6 +++---
 sys/mips/mips/genassym.c   |  3 +--
 sys/mips/mips/locore.S     |  2 +-
 sys/mips/mips/mpboot.S     |  2 +-
 sys/mips/mips/pmap.c       | 18 ++++++++---------
 sys/mips/mips/support.S    |  2 +-
 sys/mips/mips/tlb.S        |  8 ++++----
 sys/mips/mips/trap.c       | 12 ++++++------
 sys/mips/mips/vm_machdep.c |  4 ++--
 12 files changed, 53 insertions(+), 77 deletions(-)

diff --git a/sys/mips/include/param.h b/sys/mips/include/param.h
index 0f380852609..0edb8b93d55 100644
--- a/sys/mips/include/param.h
+++ b/sys/mips/include/param.h
@@ -100,28 +100,14 @@
 #define	CACHE_LINE_SHIFT	6
 #define	CACHE_LINE_SIZE		(1 << CACHE_LINE_SHIFT)
 
-#define	NBPG		4096		/* bytes/page */
-#define	PGOFSET		(NBPG-1)	/* byte offset into page */
-#define	PGSHIFT		12		/* LOG2(NBPG) */
-
 #define	PAGE_SHIFT	12		/* LOG2(PAGE_SIZE) */
 #define	PAGE_SIZE	(1<<PAGE_SHIFT) /* bytes/page */
 #define	PAGE_MASK	(PAGE_SIZE-1)
 #define	NPTEPG		(PAGE_SIZE/(sizeof (pt_entry_t)))
-
-#define	NBSEG		0x400000	/* bytes/segment */
-#define	SEGOFSET	(NBSEG-1)	/* byte offset into segment */
-#define	SEGSHIFT	22		/* LOG2(NBSEG) */
+#define	NPDEPG		(PAGE_SIZE/(sizeof (pd_entry_t)))
 
 #define	MAXPAGESIZES	1		/* maximum number of supported page sizes */
 
-/* XXXimp: This has moved to vmparam.h */
-/* Also, this differs from the mips2 definition, but likely is better */
-/* since this means the kernel won't chew up TLBs when it is executing */
-/* code */
-#define	KERNBASE	0x80000000	/* start of kernel virtual */
-#define	BTOPKERNBASE	((u_long)KERNBASE >> PGSHIFT)
-
 #define	BLKDEV_IOSIZE	2048		/* xxx: Why is this 1/2 page? */
 #define	MAXDUMPPGS	1		/* xxx: why is this only one? */
 
@@ -137,8 +123,8 @@
 #define	UPAGES			2
 
 /* pages ("clicks") (4096 bytes) to disk blocks */
-#define	ctod(x)		((x) << (PGSHIFT - DEV_BSHIFT))
-#define	dtoc(x)		((x) >> (PGSHIFT - DEV_BSHIFT))
+#define	ctod(x)		((x) << (PAGE_SHIFT - DEV_BSHIFT))
+#define	dtoc(x)		((x) >> (PAGE_SHIFT - DEV_BSHIFT))
 
 /*
  * Map a ``block device block'' to a file system block.
@@ -149,18 +135,18 @@
 #define	bdbtofsb(bn)	((bn) / (BLKDEV_IOSIZE/DEV_BSIZE))
 
 /*
- * Conversion macros
+ * Mach derived conversion macros
  */
-#define	mips_round_page(x)	((((unsigned long)(x)) + NBPG - 1) & ~(NBPG-1))
-#define	mips_trunc_page(x)	((unsigned long)(x) & ~(NBPG-1))
-#define	mips_btop(x)		((unsigned long)(x) >> PGSHIFT)
-#define	mips_ptob(x)		((unsigned long)(x) << PGSHIFT)
-#define	round_page		mips_round_page
-#define	trunc_page		mips_trunc_page
-#define	atop(x)			((unsigned long)(x) >> PAGE_SHIFT)
-#define	ptoa(x)			((unsigned long)(x) << PAGE_SHIFT)
+#define round_page(x)		(((unsigned long)(x) + PAGE_MASK) & ~PAGE_MASK)
+#define trunc_page(x)		((unsigned long)(x) & ~PAGE_MASK)
 
-#define	pgtok(x)		((x) * (PAGE_SIZE / 1024))
+#define atop(x)			((unsigned long)(x) >> PAGE_SHIFT)
+#define ptoa(x)			((unsigned long)(x) << PAGE_SHIFT)
+
+#define mips_btop(x)		((unsigned long)(x) >> PAGE_SHIFT)
+#define mips_ptob(x)		((unsigned long)(x) << PAGE_SHIFT)
+
+#define	pgtok(x)		((unsigned long)(x) * (PAGE_SIZE / 1024))
 
 #ifndef _KERNEL
 #define	DELAY(n)	{ register int N = (n); while (--N > 0); }
diff --git a/sys/mips/include/pte.h b/sys/mips/include/pte.h
index db26cbb0929..31b08449711 100644
--- a/sys/mips/include/pte.h
+++ b/sys/mips/include/pte.h
@@ -126,7 +126,7 @@ typedef pt_entry_t *pd_entry_t;
 #define	pfn_to_vad(x)	(((x) & PTE_FRAME) << PTE_SHIFT)
 
 /* User virtual to pte offset in page table */
-#define	vad_to_pte_offset(adr)	(((adr) >> PGSHIFT) & (NPTEPG -1))
+#define	vad_to_pte_offset(adr)	(((adr) >> PAGE_SHIFT) & (NPTEPG -1))
 
 #define	mips_pg_v(entry)	((entry) & PTE_V)
 #define	mips_pg_wired(entry)	((entry) & PTE_WIRED)
diff --git a/sys/mips/include/vmparam.h b/sys/mips/include/vmparam.h
index d4bc7fa07b0..bed5e36398e 100644
--- a/sys/mips/include/vmparam.h
+++ b/sys/mips/include/vmparam.h
@@ -97,13 +97,19 @@
 
 /* user/kernel map constants */
 #define	VM_MIN_ADDRESS		((vm_offset_t)0x00000000)
+#define	VM_MAX_ADDRESS		((vm_offset_t)(intptr_t)(int32_t)0xffffffff)
+
+#define	VM_MINUSER_ADDRESS	((vm_offset_t)0x00000000)
 #define	VM_MAXUSER_ADDRESS	((vm_offset_t)0x80000000)
 #define	VM_MAX_MMAP_ADDR	VM_MAXUSER_ADDRESS
-#define	VM_MAX_ADDRESS		((vm_offset_t)0x80000000)
 
 #define	VM_MIN_KERNEL_ADDRESS		((vm_offset_t)0xC0000000)
-#define	VM_KERNEL_WIRED_ADDR_END	(VM_MIN_KERNEL_ADDRESS)
-#define	VM_MAX_KERNEL_ADDRESS	((vm_offset_t)0xFFFFC000)
+#define	VM_MAX_KERNEL_ADDRESS		((vm_offset_t)0xFFFFC000)
+#if 0
+#define	KERNBASE		(VM_MIN_KERNEL_ADDRESS)
+#else
+#define	KERNBASE		((vm_offset_t)(intptr_t)(int32_t)0x80000000)
+#endif
 
 /*
  * Disable superpage reservations. (not sure if this is right
@@ -175,23 +181,8 @@
  */
 #define	VM_NFREEORDER		9
 
-/*
- * XXXMIPS: This values need to be changed!!!
- */ 
-#if 0
-#define VM_MIN_ADDRESS		((vm_offset_t)0x0000000000010000)
-#define VM_MAXUSER_ADDRESS	((vm_offset_t)MIPS_KSEG0_START-1)
-#define VM_MAX_ADDRESS		((vm_offset_t)0x0000000100000000)
-#define VM_MIN_KERNEL_ADDRESS	((vm_offset_t)MIPS_KSEG3_START)
-#define VM_MAX_KERNEL_ADDRESS	((vm_offset_t)MIPS_KSEG3_END)
-#define	KERNBASE		(VM_MIN_KERNEL_ADDRESS)
-
-/* virtual sizes (bytes) for various kernel submaps */
-#define	VM_KMEM_SIZE		(16*1024*1024)		/* XXX ??? */
-#endif
-
-#define NBSEG		0x400000	/* bytes/segment */
-#define SEGOFSET	(NBSEG-1)	/* byte offset into segment */
 #define SEGSHIFT	22		/* LOG2(NBSEG) */
+#define NBSEG		(1 << SEGSHIFT)	/* bytes/segment */
+#define SEGOFSET	(NBSEG-1)	/* byte offset into segment */
 
 #endif /* !_MACHINE_VMPARAM_H_ */
diff --git a/sys/mips/mips/exception.S b/sys/mips/mips/exception.S
index 161e9475eff..8c32b582c3c 100644
--- a/sys/mips/mips/exception.S
+++ b/sys/mips/mips/exception.S
@@ -158,7 +158,7 @@ MipsDoTLBMiss:
 	lw	k1, 0(k1)			#08: k1=seg entry
 	mfc0	k0, COP_0_BAD_VADDR		#09: k0=bad address (again)
 	beq	k1, zero, 2f			#0a: ==0 -- no page table
-	srl	k0, PGSHIFT - 2			#0b: k0=VPN (aka va>>10)
+	srl	k0, PAGE_SHIFT - 2			#0b: k0=VPN (aka va>>10)
 
 	andi	k0, k0, ((NPTEPG/2) - 1) << 3	#0c: k0=page tab offset
 #xxx mips64 unsafe?
@@ -860,7 +860,7 @@ NLEAF(MipsTLBInvalidException)
 	nop
 
 	mfc0	k0, COP_0_BAD_VADDR
-	srl	k0, PGSHIFT - 2
+	srl	k0, PAGE_SHIFT - 2
 	andi	k0, 0xffc
 	addu	k1, k1, k0
 
@@ -944,7 +944,7 @@ tlb_insert_random:
 	sll	k1, k1, PAGE_SHIFT + 1
 
 	PTR_LA	k0, _C_LABEL(pcpu_space)
-	addiu	k0, (NBPG * 2)
+	addiu	k0, (PAGE_SIZE * 2)
 	addu	k0, k0, k1
 
 	/*
diff --git a/sys/mips/mips/genassym.c b/sys/mips/mips/genassym.c
index f59ef6854ba..6cec2b79435 100644
--- a/sys/mips/mips/genassym.c
+++ b/sys/mips/mips/genassym.c
@@ -90,8 +90,7 @@ ASSYM(VM_MAXUSER_ADDRESS, VM_MAXUSER_ADDRESS);
 ASSYM(SIGF_UC, offsetof(struct sigframe, sf_uc));
 ASSYM(SIGFPE, SIGFPE);
 ASSYM(PAGE_SHIFT, PAGE_SHIFT);
-ASSYM(PGSHIFT, PGSHIFT);
-ASSYM(NBPG, NBPG);
+ASSYM(PAGE_SIZE, PAGE_SIZE);
 ASSYM(SEGSHIFT, SEGSHIFT);
 ASSYM(NPTEPG, NPTEPG);
 ASSYM(TDF_NEEDRESCHED, TDF_NEEDRESCHED);
diff --git a/sys/mips/mips/locore.S b/sys/mips/mips/locore.S
index 697b81884f5..d626add4215 100644
--- a/sys/mips/mips/locore.S
+++ b/sys/mips/mips/locore.S
@@ -178,7 +178,7 @@ VECTOR(_locore, unknown)
 	 * Initialize stack and call machine startup.
 	 */
 	PTR_LA	sp, _C_LABEL(pcpu_space)
-	addiu	sp, (NBPG * 2) - CALLFRAME_SIZ
+	addiu	sp, (PAGE_SIZE * 2) - CALLFRAME_SIZ
 
 	sw	zero, CALLFRAME_SIZ - 4(sp)	# Zero out old ra for debugger
 	sw	zero, CALLFRAME_SIZ - 8(sp)	# Zero out old fp for debugger
diff --git a/sys/mips/mips/mpboot.S b/sys/mips/mips/mpboot.S
index e1c578a7961..631099c5964 100644
--- a/sys/mips/mips/mpboot.S
+++ b/sys/mips/mips/mpboot.S
@@ -66,7 +66,7 @@ GLOBAL(mpentry)
 	 * Initialize stack and call machine startup
 	 */
 	PTR_LA	sp, _C_LABEL(pcpu_space)
-	addiu	sp, (NBPG * 2) - CALLFRAME_SIZ
+	addiu	sp, (PAGE_SIZE * 2) - CALLFRAME_SIZ
 	sll	t0, s0, PAGE_SHIFT + 1
 	addu	sp, sp, t0
 
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index f1056fdcb46..80418a79f2a 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -609,7 +609,7 @@ pmap_invalidate_page_action(void *arg)
 		pmap->pm_asid[PCPU_GET(cpuid)].gen = 0;
 		return;
 	}
-	va = pmap_va_asid(pmap, (va & ~PGOFSET));
+	va = pmap_va_asid(pmap, (va & ~PAGE_MASK));
 	mips_TBIS(va);
 }
 
@@ -765,7 +765,7 @@ pmap_kremove(vm_offset_t va)
 	/*
 	 * Write back all caches from the page being destroyed
 	 */
-	mips_dcache_wbinv_range_index(va, NBPG);
+	mips_dcache_wbinv_range_index(va, PAGE_SIZE);
 
 	pte = pmap_pte(kernel_pmap, va);
 	*pte = PTE_G;
@@ -1516,7 +1516,7 @@ pmap_remove_page(struct pmap *pmap, vm_offset_t va)
 	/*
 	 * Write back all caches from the page being destroyed
 	 */
-	mips_dcache_wbinv_range_index(va, NBPG);
+	mips_dcache_wbinv_range_index(va, PAGE_SIZE);
 
 	/*
 	 * get a local va for mappings for this pmap.
@@ -1603,7 +1603,7 @@ pmap_remove_all(vm_page_t m)
 		 * the page being destroyed
 	 	 */
 		if (m->md.pv_list_count == 1) 
-			mips_dcache_wbinv_range_index(pv->pv_va, NBPG);
+			mips_dcache_wbinv_range_index(pv->pv_va, PAGE_SIZE);
 
 		pv->pv_pmap->pm_stats.resident_count--;
 
@@ -1902,8 +1902,8 @@ validate:
 	 */
 	if (!is_kernel_pmap(pmap) && (pmap == &curproc->p_vmspace->vm_pmap) &&
 	    (prot & VM_PROT_EXECUTE)) {
-		mips_icache_sync_range(va, NBPG);
-		mips_dcache_wbinv_range(va, NBPG);
+		mips_icache_sync_range(va, PAGE_SIZE);
+		mips_dcache_wbinv_range(va, PAGE_SIZE);
 	}
 	vm_page_unlock_queues();
 	PMAP_UNLOCK(pmap);
@@ -2032,8 +2032,8 @@ pmap_enter_quick_locked(pmap_t pmap, vm_offset_t va, vm_page_t m,
 		 * unresolvable TLB miss may occur. */
 		if (pmap == &curproc->p_vmspace->vm_pmap) {
 			va &= ~PAGE_MASK;
-			mips_icache_sync_range(va, NBPG);
-			mips_dcache_wbinv_range(va, NBPG);
+			mips_icache_sync_range(va, PAGE_SIZE);
+			mips_dcache_wbinv_range(va, PAGE_SIZE);
 		}
 	}
 	return (mpte);
@@ -3078,7 +3078,7 @@ pmap_flush_pvcache(vm_page_t m)
 	if (m != NULL) {
 		for (pv = TAILQ_FIRST(&m->md.pv_list); pv;
 	    	    pv = TAILQ_NEXT(pv, pv_list)) {
-			mips_dcache_wbinv_range_index(pv->pv_va, NBPG);
+			mips_dcache_wbinv_range_index(pv->pv_va, PAGE_SIZE);
 		}
 	}
 }
diff --git a/sys/mips/mips/support.S b/sys/mips/mips/support.S
index 2aed3e613fd..d0a06b6d68b 100644
--- a/sys/mips/mips/support.S
+++ b/sys/mips/mips/support.S
@@ -210,7 +210,7 @@ END(fillw)
  * mem_zero_page(addr);
  */
 LEAF(mem_zero_page)
-	li	v0, NBPG
+	li	v0, PAGE_SIZE
 1:
 	subu	v0, 8
 	sd	zero, 0(a0)
diff --git a/sys/mips/mips/tlb.S b/sys/mips/mips/tlb.S
index 46a15f81ca4..f6e7934ed3e 100644
--- a/sys/mips/mips/tlb.S
+++ b/sys/mips/mips/tlb.S
@@ -240,7 +240,7 @@ LEAF(Mips_TLBFlush)
 	# MIPS_KSEG0_START + 2 * i * PAGE_SIZE;
 	# One bogus value for every TLB entry might cause MCHECK exception
 	#
-	sll	t3, t1, PGSHIFT + 1
+	sll	t3, t1, PAGE_SHIFT + 1
 	li	v0, MIPS_KSEG0_START		# invalid address
 	addu	v0, t3
 /*
@@ -299,7 +299,7 @@ LEAF(Mips_TLBFlushAddr)
 	# address calculated by following expression:
 	# MIPS_KSEG0_START + 2 * i * PAGE_SIZE;
 	# One bogus value for every TLB entry might cause MCHECK exception
-	sll	v0, PGSHIFT + 1
+	sll	v0, PAGE_SHIFT + 1
 	addu	t1, v0
 	_MTC0	t1, COP_0_TLB_HI		# Mark entry high as invalid
 
@@ -482,7 +482,7 @@ LEAF(mips_TBIAP)
 	# MIPS_KSEG0_START + 2 * i * PAGE_SIZE;
 	# One bogus value for every TLB entry might cause MCHECK exception
 	#
-	sll	t3, t1, PGSHIFT + 1
+	sll	t3, t1, PAGE_SHIFT + 1
 	li	v0, MIPS_KSEG0_START		# invalid address
 	addu	v0, t3
 
@@ -507,7 +507,7 @@ LEAF(mips_TBIAP)
 	tlbwi					# invalidate the TLB entry
 2:
 	addu	t1, t1, 1
-	addu	v0, 1 << (PGSHIFT + 1)
+	addu	v0, 1 << (PAGE_SHIFT + 1)
 	bne	t1, t2, 1b
 	nop
 
diff --git a/sys/mips/mips/trap.c b/sys/mips/mips/trap.c
index de4bd01bf29..1e28188330b 100644
--- a/sys/mips/mips/trap.c
+++ b/sys/mips/mips/trap.c
@@ -378,12 +378,12 @@ trap(struct trapframe *trapframe)
 			if (!(pte = pmap_segmap(kernel_pmap,
 			    trapframe->badvaddr)))
 				panic("trap: ktlbmod: invalid segmap");
-			pte += (trapframe->badvaddr >> PGSHIFT) & (NPTEPG - 1);
+			pte += (trapframe->badvaddr >> PAGE_SHIFT) & (NPTEPG - 1);
 			entry = *pte;
 #ifdef SMP
 			/* It is possible that some other CPU changed m-bit */
 			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit())) {
-				trapframe->badvaddr &= ~PGOFSET;
+				trapframe->badvaddr &= ~PAGE_MASK;
 				pmap_update_page(kernel_pmap,
 				    trapframe->badvaddr, entry);
 				PMAP_UNLOCK(kernel_pmap);
@@ -401,7 +401,7 @@ trap(struct trapframe *trapframe)
 			}
 			entry |= mips_pg_m_bit();
 			*pte = entry;
-			trapframe->badvaddr &= ~PGOFSET;
+			trapframe->badvaddr &= ~PAGE_MASK;
 			pmap_update_page(kernel_pmap, trapframe->badvaddr, entry);
 			pa = mips_tlbpfn_to_paddr(entry);
 			if (!page_is_managed(pa))
@@ -421,12 +421,12 @@ trap(struct trapframe *trapframe)
 			PMAP_LOCK(pmap);
 			if (!(pte = pmap_segmap(pmap, trapframe->badvaddr)))
 				panic("trap: utlbmod: invalid segmap");
-			pte += (trapframe->badvaddr >> PGSHIFT) & (NPTEPG - 1);
+			pte += (trapframe->badvaddr >> PAGE_SHIFT) & (NPTEPG - 1);
 			entry = *pte;
 #ifdef SMP
 			/* It is possible that some other CPU changed m-bit */
 			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit())) {
-				trapframe->badvaddr = (trapframe->badvaddr & ~PGOFSET);
+				trapframe->badvaddr = (trapframe->badvaddr & ~PAGE_MASK);
 				pmap_update_page(pmap, trapframe->badvaddr, entry);
 				PMAP_UNLOCK(pmap);
 				goto out;
@@ -445,7 +445,7 @@ trap(struct trapframe *trapframe)
 			}
 			entry |= mips_pg_m_bit();
 			*pte = entry;
-			trapframe->badvaddr = (trapframe->badvaddr & ~PGOFSET);
+			trapframe->badvaddr = (trapframe->badvaddr & ~PAGE_MASK);
 			pmap_update_page(pmap, trapframe->badvaddr, entry);
 			trapframe->badvaddr |= (pmap->pm_asid[PCPU_GET(cpuid)].asid << VMTLB_PID_SHIFT);
 			pa = mips_tlbpfn_to_paddr(entry);
diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index 3b7a1a1b18c..10321dac4bf 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -219,7 +219,7 @@ cpu_thread_swapin(struct thread *td)
 	 */
 	if (!(pte = pmap_segmap(kernel_pmap, td->td_md.md_realstack)))
 		panic("cpu_thread_swapin: invalid segmap");
-	pte += ((vm_offset_t)td->td_md.md_realstack >> PGSHIFT) & (NPTEPG - 1);
+	pte += ((vm_offset_t)td->td_md.md_realstack >> PAGE_SHIFT) & (NPTEPG - 1);
 
 	for (i = 0; i < KSTACK_PAGES - 1; i++) {
 		td->td_md.md_upte[i] = *pte & ~(PTE_RO|PTE_WIRED);
@@ -249,7 +249,7 @@ cpu_thread_alloc(struct thread *td)
 
 	if (!(pte = pmap_segmap(kernel_pmap, td->td_md.md_realstack)))
 		panic("cpu_thread_alloc: invalid segmap");
-	pte += ((vm_offset_t)td->td_md.md_realstack >> PGSHIFT) & (NPTEPG - 1);
+	pte += ((vm_offset_t)td->td_md.md_realstack >> PAGE_SHIFT) & (NPTEPG - 1);
 
 	for (i = 0; i < KSTACK_PAGES - 1; i++) {
 		td->td_md.md_upte[i] = *pte & ~(PTE_RO|PTE_WIRED);

From a27e66e8f5d73b57c16fa07d19072d353c944d95 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 17 Apr 2010 09:42:07 +0000
Subject: [PATCH 102/532] o) Make pcb_onfault a pointer rather than an obscure
 integer value. o) Mask off PAGE_MASK bits in pmap_update_page, etc., rather
 than modifying the    badvaddr in trapframe.  Some nearby interfaces already
 did this. o) Make PTEs "unsigned int" for now, not "unsigned long" -- we are
 only ready    for them to be 32-bit on 64-bit platforms. o) Rather than using
 pmap_segmap and calculating the offset into the page table    by hand in
 trap.c, use pmap_pte(). o) Remove unused quad_syscall variable in trap.c. o)
 Log things for illegal instructions like we do for bad page faults. o)
 Various cast cleanups related to how to print registers. o) When logging page
 faults, show the page table information not just for the    program counter,
 but for the fault address. o) Modify support.S to use ABI-neutral macros for
 operating on pointers. o) Consistently use CALLFRAME_SIZ rather than
 STAND_FRAME_SIZE, etc. o) Remove unused insque/remque functions. o) Remove
 some coprocessor 0 accessor functions implemented in assembly that    are
 unused and have inline assembly counterparts.

---
 sys/mips/include/cpu.h     |   4 -
 sys/mips/include/pcb.h     |   2 +-
 sys/mips/include/pte.h     |   2 +-
 sys/mips/mips/pmap.c       |   8 +-
 sys/mips/mips/support.S    | 670 +++++++++++++++----------------------
 sys/mips/mips/trap.c       | 319 ++++++++++--------
 sys/mips/mips/vm_machdep.c |   2 +-
 7 files changed, 459 insertions(+), 548 deletions(-)

diff --git a/sys/mips/include/cpu.h b/sys/mips/include/cpu.h
index 5c5e5e5e92b..83b6a85fe31 100644
--- a/sys/mips/include/cpu.h
+++ b/sys/mips/include/cpu.h
@@ -458,13 +458,9 @@ extern union cpuprid fpu_id;
 struct tlb;
 struct user;
 
-u_int32_t mips_cp0_config1_read(void);
 int Mips_ConfigCache(void);
 void Mips_SetWIRED(int);
 void Mips_SetPID(int);
-u_int Mips_GetCOUNT(void);
-void Mips_SetCOMPARE(u_int);
-u_int Mips_GetCOMPARE(void);
 
 void Mips_SyncCache(void);
 void Mips_SyncDCache(vm_offset_t, int);
diff --git a/sys/mips/include/pcb.h b/sys/mips/include/pcb.h
index f95ef4d8d4a..e0982e35f24 100644
--- a/sys/mips/include/pcb.h
+++ b/sys/mips/include/pcb.h
@@ -51,7 +51,7 @@ struct pcb
 {
 	struct trapframe pcb_regs;	/* saved CPU and registers */
 	__register_t pcb_context[14];	/* kernel context for resume */
-	int	pcb_onfault;		/* for copyin/copyout faults */
+	void *pcb_onfault;		/* for copyin/copyout faults */
 	register_t pcb_tpc;
 };
 
diff --git a/sys/mips/include/pte.h b/sys/mips/include/pte.h
index 31b08449711..e3b46ca80d5 100644
--- a/sys/mips/include/pte.h
+++ b/sys/mips/include/pte.h
@@ -83,7 +83,7 @@ struct tlb {
 	int	tlb_lo1;
 };
 
-typedef unsigned long pt_entry_t;
+typedef unsigned int pt_entry_t;
 typedef pt_entry_t *pd_entry_t;
 
 #define	PDESIZE		sizeof(pd_entry_t)	/* for assembly files */
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 80418a79f2a..6bb38d47e3b 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -660,7 +660,7 @@ pmap_update_page_action(void *arg)
 		pmap->pm_asid[PCPU_GET(cpuid)].gen = 0;
 		return;
 	}
-	va = pmap_va_asid(pmap, va);
+	va = pmap_va_asid(pmap, (va & ~PAGE_MASK));
 	MachTLBUpdate(va, pte);
 }
 
@@ -669,6 +669,8 @@ pmap_TLB_update_kernel(vm_offset_t va, pt_entry_t pte)
 {
 	u_int32_t pid;
 
+	va &= ~PAGE_MASK;
+
 	MachTLBGetPID(pid);
 	va = va | (pid << VMTLB_PID_SHIFT);
 
@@ -1885,7 +1887,7 @@ validate:
 			if (origpte & PTE_M) {
 				KASSERT((origpte & PTE_RW),
 				    ("pmap_enter: modified page not writable:"
-				    " va: %p, pte: 0x%lx", (void *)va, origpte));
+				    " va: %p, pte: 0x%x", (void *)va, origpte));
 				if (page_is_managed(opa))
 					vm_page_dirty(om);
 			}
@@ -2381,7 +2383,7 @@ pmap_remove_pages(pmap_t pmap)
 		m = PHYS_TO_VM_PAGE(mips_tlbpfn_to_paddr(tpte));
 
 		KASSERT(m < &vm_page_array[vm_page_array_size],
-		    ("pmap_remove_pages: bad tpte %lx", tpte));
+		    ("pmap_remove_pages: bad tpte %x", tpte));
 
 		pv->pv_pmap->pm_stats.resident_count--;
 
diff --git a/sys/mips/mips/support.S b/sys/mips/mips/support.S
index d0a06b6d68b..9cbc9982293 100644
--- a/sys/mips/mips/support.S
+++ b/sys/mips/mips/support.S
@@ -103,47 +103,22 @@
  * Primitives
  */
 
-/*
- * This table is indexed by u.u_pcb.pcb_onfault in trap().
- * The reason for using this table rather than storing an address in
- * u.u_pcb.pcb_onfault is simply to make the code faster.
- */
-	.globl	onfault_table
-	.data
-	.align	3
-onfault_table:
-	.word	0			# invalid index number
-#define	BADERR		1
-	.word	baderr
-#define	COPYERR		2
-	.word	copyerr
-#define	FSWBERR		3
-	.word	fswberr
-#define	FSWINTRBERR	4
-	.word	fswintrberr
-#if defined(DDB) || defined(DEBUG)
-#define	DDBERR	5
-	.word	ddberr
-#else
-	.word	0
-#endif
-
 	.text
 
 /*
  * See if access to addr with a len type instruction causes a machine check.
- * len is length of access (1=byte, 2=short, 4=long)
+ * len is length of access (1=byte, 2=short, 4=int)
  *
  * badaddr(addr, len)
  *	char *addr;
  *	int len;
  */
 LEAF(badaddr)
-	li	v0, BADERR
+	PTR_LA	v0, baderr
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
+	PTR_L	v1, PC_CURPCB(v1)
 	bne	a1, 1, 2f
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	b	5f
 	lbu	v0, (a0)
 2:
@@ -154,7 +129,7 @@ LEAF(badaddr)
 4:
 	lw	v0, (a0)
 5:
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero		# made it w/o errors
 baderr:
@@ -169,24 +144,24 @@ END(badaddr)
  * string is too long, return ENAMETOOLONG; else return 0.
  */
 LEAF(copystr)
-	move	t0, a2
-	beq	a2, zero, 4f
+	move		t0, a2
+	beq		a2, zero, 4f
 1:
-	lbu	v0, 0(a0)
-	subu	a2, a2, 1
-	beq	v0, zero, 2f
-	sb	v0, 0(a1)		# each byte until NIL
-	addu	a0, a0, 1
-	bne	a2, zero, 1b		# less than maxlen
-	addu	a1, a1, 1
+	lbu		v0, 0(a0)
+	PTR_SUBU	a2, a2, 1
+	beq		v0, zero, 2f
+	sb		v0, 0(a1)		# each byte until NIL
+	PTR_ADDU	a0, a0, 1
+	bne		a2, zero, 1b		# less than maxlen
+	PTR_ADDU	a1, a1, 1
 4:
-	li	v0, ENAMETOOLONG	# run out of space
+	li		v0, ENAMETOOLONG	# run out of space
 2:
-	beq	a3, zero, 3f		# return num. of copied bytes
-	subu	a2, t0, a2		# if the 4th arg was non-NULL
-	sw	a2, 0(a3)
+	beq		a3, zero, 3f		# return num. of copied bytes
+	PTR_SUBU	a2, t0, a2		# if the 4th arg was non-NULL
+	PTR_S		a2, 0(a3)
 3:
-	j	ra			# v0 is 0 or ENAMETOOLONG
+	j		ra			# v0 is 0 or ENAMETOOLONG
 	nop
 END(copystr)
 
@@ -196,12 +171,12 @@ END(copystr)
  */
 LEAF(fillw)
 1:
-	addiu	a2, a2, -1
-	sh	a0, 0(a1)
-	bne	a2,zero, 1b
-	addiu	a1, a1, 2
+	PTR_ADDU	a2, a2, -1
+	sh		a0, 0(a1)
+	bne		a2,zero, 1b
+	PTR_ADDU	a1, a1, 2
 
-	jr	ra
+	jr		ra
 	nop
 END(fillw)
 
@@ -210,13 +185,13 @@ END(fillw)
  * mem_zero_page(addr);
  */
 LEAF(mem_zero_page)
-	li	v0, PAGE_SIZE
+	li		v0, PAGE_SIZE
 1:
-	subu	v0, 8
-	sd	zero, 0(a0)
-	bne	zero, v0, 1b
-	addu	a0, 8
-	jr	ra
+	PTR_SUBU	v0, 8
+	sd		zero, 0(a0)
+	bne		zero, v0, 1b
+	PTR_ADDU	a0, 8
+	jr		ra
 	nop
 END(mem_zero_page)
 
@@ -228,56 +203,56 @@ END(mem_zero_page)
  *			a2 = count
  */
 LEAF(insb)
-	beq	a2, zero, 2f
-	addu	a2, a1
+	beq		a2, zero, 2f
+	PTR_ADDU	a2, a1
 1:
-	lbu	v0, 0(a0)
-	addiu	a1, 1
-	bne	a1, a2, 1b
-	sb	v0, -1(a1)
+	lbu		v0, 0(a0)
+	PTR_ADDU	a1, 1
+	bne		a1, a2, 1b
+	sb		v0, -1(a1)
 2:
-	jr	ra
+	jr		ra
 	nop
 END(insb)
 
 LEAF(insw)
-	beq	a2, zero, 2f
-	addu	a2, a2
-	addu	a2, a1
+	beq		a2, zero, 2f
+	PTR_ADDU	a2, a2
+	PTR_ADDU	a2, a1
 1:
-	lhu	v0, 0(a0)
-	addiu	a1, 2
-	bne	a1, a2, 1b
-	sh	v0, -2(a1)
+	lhu		v0, 0(a0)
+	PTR_ADDU	a1, 2
+	bne		a1, a2, 1b
+	sh		v0, -2(a1)
 2:
-	jr	ra
+	jr		ra
 	nop
 END(insw)
 
 LEAF(insl)
-	beq	a2, zero, 2f
-	sll	a2, 2
-	addu	a2, a1
+	beq		a2, zero, 2f
+	sll		a2, 2
+	PTR_ADDU	a2, a1
 1:
-	lw	v0, 0(a0)
-	addiu	a1, 4
-	bne	a1, a2, 1b
-	sw	v0, -4(a1)
+	lw		v0, 0(a0)
+	PTR_ADDU	a1, 4
+	bne		a1, a2, 1b
+	sw		v0, -4(a1)
 2:
-	jr	ra
+	jr		ra
 	nop
 END(insl)
 
 LEAF(outsb)
-	beq	a2, zero, 2f
-	addu	a2, a1
+	beq		a2, zero, 2f
+	PTR_ADDU	a2, a1
 1:
-	lbu	v0, 0(a1)
-	addiu	a1, 1
-	bne	a1, a2, 1b
-	sb	v0, 0(a0)
+	lbu		v0, 0(a1)
+	PTR_ADDU	a1, 1
+	bne		a1, a2, 1b
+	sb		v0, 0(a0)
 2:
-	jr	ra
+	jr		ra
 	nop
 END(outsb)
 
@@ -343,22 +318,22 @@ END(outsl)
  *		u_int maxlength;
  *		u_int *lencopied;
  */
-NON_LEAF(copyinstr, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
-	sw	ra, STAND_RA_OFFSET(sp)
+NON_LEAF(copyinstr, CALLFRAME_SIZ, ra)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
+	PTR_LA	v0, copyerr
 	blt	a0, zero, _C_LABEL(copyerr)  # make sure address is in user space
-	li	v0, COPYERR
+	REG_S	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
+	PTR_L	v1, PC_CURPCB(v1)
 	jal	_C_LABEL(copystr)
-	sw	v0, U_PCB_ONFAULT(v1)
-	lw	ra, STAND_RA_OFFSET(sp)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
+	REG_L	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
-	addu	sp, sp, STAND_FRAME_SIZE
+	PTR_ADDU	sp, sp, CALLFRAME_SIZ
 END(copyinstr)
 
 /*
@@ -371,22 +346,22 @@ END(copyinstr)
  *		u_int maxlength;
  *		u_int *lencopied;
  */
-NON_LEAF(copyoutstr, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
-	sw	ra, STAND_RA_OFFSET(sp)
+NON_LEAF(copyoutstr, CALLFRAME_SIZ, ra)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
+	PTR_LA	v0, copyerr
 	blt	a1, zero, _C_LABEL(copyerr)  # make sure address is in user space
-	li	v0, COPYERR
+	REG_S	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
+	PTR_L	v1, PC_CURPCB(v1)
 	jal	_C_LABEL(copystr)
-	sw	v0, U_PCB_ONFAULT(v1)
-	lw	ra, STAND_RA_OFFSET(sp)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
+	REG_L	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
-	addu	sp, sp, STAND_FRAME_SIZE
+	PTR_ADDU	sp, sp, CALLFRAME_SIZ
 END(copyoutstr)
 
 /*
@@ -396,21 +371,21 @@ END(copyoutstr)
  *		caddr_t *to;	(kernel destination address)
  *		unsigned len;
  */
-NON_LEAF(copyin, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
-	sw	ra, STAND_RA_OFFSET(sp)
+NON_LEAF(copyin, CALLFRAME_SIZ, ra)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
+	PTR_LA	v0, copyerr
 	blt	a0, zero, _C_LABEL(copyerr)  # make sure address is in user space
-	li	v0, COPYERR
+	REG_S	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
+	PTR_L	v1, PC_CURPCB(v1)
 	jal	_C_LABEL(bcopy)
-	sw	v0, U_PCB_ONFAULT(v1)
-	lw	ra, STAND_RA_OFFSET(sp)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
+	REG_L	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)	 	# bcopy modified v1, so reload
-	sw	zero, U_PCB_ONFAULT(v1)
-	addu	sp, sp, STAND_FRAME_SIZE
+	PTR_L	v1, PC_CURPCB(v1)	 	# bcopy modified v1, so reload
+	PTR_S	zero, U_PCB_ONFAULT(v1)
+	PTR_ADDU	sp, sp, CALLFRAME_SIZ
 	j	ra
 	move	v0, zero
 END(copyin)
@@ -422,31 +397,28 @@ END(copyin)
  *		caddr_t *to;	(user destination address)
  *		unsigned len;
  */
-NON_LEAF(copyout, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
-	sw	ra, STAND_RA_OFFSET(sp)
+NON_LEAF(copyout, CALLFRAME_SIZ, ra)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
+	PTR_LA	v0, copyerr
 	blt	a1, zero, _C_LABEL(copyerr) # make sure address is in user space
-	li	v0, COPYERR
+	REG_S	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
+	PTR_L	v1, PC_CURPCB(v1)
 	jal	_C_LABEL(bcopy)
-	sw	v0, U_PCB_ONFAULT(v1)
-	lw	ra, STAND_RA_OFFSET(sp)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
+	REG_L	ra, CALLFRAME_RA(sp)
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)		# bcopy modified v1, so reload
-	sw	zero, U_PCB_ONFAULT(v1)
-	addu	sp, sp, STAND_FRAME_SIZE
+	PTR_L	v1, PC_CURPCB(v1)	 	# bcopy modified v1, so reload
+	PTR_S	zero, U_PCB_ONFAULT(v1)
+	PTR_ADDU	sp, sp, CALLFRAME_SIZ
 	j	ra
 	move	v0, zero
 END(copyout)
 
 LEAF(copyerr)
-	lw	ra, STAND_RA_OFFSET(sp)
-	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	zero, U_PCB_ONFAULT(v1)
-	addu	sp, sp, STAND_FRAME_SIZE
+	REG_L	ra, CALLFRAME_RA(sp)
+	PTR_ADDU	sp, sp, CALLFRAME_SIZ
 	j	ra
 	li	v0, EFAULT			# return error
 END(copyerr)
@@ -460,51 +432,55 @@ END(copyerr)
 LEAF(fuword)
 ALEAF(fuword32)
 ALEAF(fuiword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	lw	v0, 0(a0)		# fetch word
 	j	ra
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 END(fuword)
 
 LEAF(fusword)
 ALEAF(fuisword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	lhu	v0, 0(a0)		# fetch short
 	j	ra
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 END(fusword)
 
 LEAF(fubyte)
 ALEAF(fuibyte)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	lbu	v0, 0(a0)		# fetch byte
 	j	ra
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 END(fubyte)
 
 LEAF(suword32)
 #ifndef __mips_n64
 XLEAF(suword)
 #endif
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sw	a1, 0(a0)		# store word
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero
 END(suword32)
@@ -512,13 +488,14 @@ END(suword32)
 #ifdef __mips_n64
 LEAF(suword64)
 XLEAF(suword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sd	a1, 0(a0)		# store word
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero
 END(suword64)
@@ -537,11 +514,12 @@ LEAF(casuword32)
 #ifndef __mips_n64
 XLEAF(casuword)
 #endif
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 1:
 	move	t0, a2
 	ll	v0, 0(a0)
@@ -555,7 +533,7 @@ XLEAF(casuword)
 2:
 	li	v0, -1
 3:
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	jr	ra
 	nop
 END(casuword32)
@@ -563,11 +541,12 @@ END(casuword32)
 #ifdef __mips_n64
 LEAF(casuword64)
 XLEAF(casuword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 1:
 	move	t0, a2
 	lld	v0, 0(a0)
@@ -581,7 +560,7 @@ XLEAF(casuword)
 2:
 	li	v0, -1
 3:
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	jr	ra
 	nop
 END(casuword64)
@@ -593,13 +572,14 @@ END(casuword64)
  * Have to flush instruction cache afterwards.
  */
 LEAF(suiword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sw	a1, 0(a0)		# store word
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	_C_LABEL(Mips_SyncICache)  # FlushICache sets v0 = 0. (Ugly)
 	li	a1, 4			# size of word
 END(suiword)
@@ -610,26 +590,28 @@ END(suiword)
  */
 LEAF(susword)
 ALEAF(suisword)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sh	a1, 0(a0)		# store short
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero
 END(susword)
 
 LEAF(subyte)
 ALEAF(suibyte)
+	PTR_LA	v0, fswberr
 	blt	a0, zero, fswberr	# make sure address is in user space
-	li	v0, FSWBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sb	a1, 0(a0)		# store byte
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero
 END(subyte)
@@ -645,24 +627,26 @@ END(fswberr)
  * The important thing is to prevent sleep() and switch().
  */
 LEAF(fuswintr)
+	PTR_LA	v0, fswintrberr
 	blt	a0, zero, fswintrberr	# make sure address is in user space
-	li	v0, FSWINTRBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	lhu	v0, 0(a0)		# fetch short
 	j	ra
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 END(fuswintr)
 
 LEAF(suswintr)
+	PTR_LA	v0, fswintrberr
 	blt	a0, zero, fswintrberr	# make sure address is in user space
-	li	v0, FSWINTRBERR
+	nop
 	GET_CPU_PCPU(v1)
-	lw	v1, PC_CURPCB(v1)
-	sw	v0, U_PCB_ONFAULT(v1)
+	PTR_L	v1, PC_CURPCB(v1)
+	PTR_S	v0, U_PCB_ONFAULT(v1)
 	sh	a1, 0(a0)		# store short
-	sw	zero, U_PCB_ONFAULT(v1)
+	PTR_S	zero, U_PCB_ONFAULT(v1)
 	j	ra
 	move	v0, zero
 END(suswintr)
@@ -672,111 +656,6 @@ LEAF(fswintrberr)
 	li	v0, -1
 END(fswintrberr)
 
-/*
- * Insert 'p' after 'q'.
- *	_insque(p, q)
- *		caddr_t p, q;
- */
-LEAF(_insque)
-	lw	v0, 0(a1)		# v0 = q->next
-	sw	a1, 4(a0)		# p->prev = q
-	sw	v0, 0(a0)		# p->next = q->next
-	sw	a0, 4(v0)		# q->next->prev = p
-	j	ra
-	sw	a0, 0(a1)		# q->next = p
-END(_insque)
-
-/*
- * Remove item 'p' from queue.
- *	_remque(p)
- *		caddr_t p;
- */
-LEAF(_remque)
-	lw	v0, 0(a0)		# v0 = p->next
-	lw	v1, 4(a0)		# v1 = p->prev
-	nop
-	sw	v0, 0(v1)		# p->prev->next = p->next
-	j	ra
-	sw	v1, 4(v0)		# p->next->prev = p->prev
-END(_remque)
-
-/*--------------------------------------------------------------------------
- *
- * Mips_GetCOUNT --
- *
- *	Mips_GetCOUNT()
- *
- * Results:
- *	Returns the current COUNT reg.
- *
- * Side effects:
- *	None.
- *
- *--------------------------------------------------------------------------
- */
-LEAF(Mips_GetCOUNT)
-	mfc0	v0, COP_0_COUNT
-	nop	#???
-	nop	#???
-	j	ra
-	nop
-END(Mips_GetCOUNT)
-
-/*--------------------------------------------------------------------------
- *
- * Mips_SetCOMPARE --
- *
- *	Mips_SetCOMPARE()
- *
- * Results:
- *	Sets a new value to the COMPARE register.
- *
- * Side effects:
- *	The COMPARE equal interrupt is acknowledged.
- *
- *--------------------------------------------------------------------------
- */
-LEAF(Mips_SetCOMPARE)
-	mtc0	a0, COP_0_COMPARE
-	j	ra
-	nop
-END(Mips_SetCOMPARE)
-
-LEAF(Mips_GetCOMPARE)
-	mfc0	v0, COP_0_COMPARE
-	j	ra
-	nop
-END(Mips_GetCOMPARE)
-
-/*
- * u_int32_t mips_cp0_status_read(void)
- *
- *	Return the current value of the CP0 Status register.
- */
-LEAF(mips_cp0_status_read)
-	mfc0	v0, COP_0_STATUS_REG
-	j	ra
-	nop
-END(mips_cp0_status_read)
-
-/*
- * void mips_cp0_status_write(u_int32_t)
- *
- *	Set the value of the CP0 Status register.
- *
- *	Note: This is almost certainly not the way you want to write a
- *	"permanent" value to to the CP0 Status register, since it gets
- *	saved in trap frames and restores.
- */
-LEAF(mips_cp0_status_write)
-	mtc0	a0, COP_0_STATUS_REG
-	nop
-	nop
-	j	ra
-	nop
-END(mips_cp0_status_write)
-
-
 /*
  * memcpy(to, from, len)
  * {ov}bcopy(from, to, len)
@@ -789,7 +668,7 @@ LEAF(memcpy)
 ALEAF(bcopy)
 ALEAF(ovbcopy)
 	.set	noreorder
-	addu	t0, a0, a2		# t0 = end of s1 region
+	PTR_ADDU	t0, a0, a2		# t0 = end of s1 region
 	sltu	t1, a1, t0
 	sltu	t2, a0, a1
 	and	t1, t1, t2		# t1 = true if from < to < (from+len)
@@ -797,11 +676,11 @@ ALEAF(ovbcopy)
 	slt	t2, a2, 12		# check for small copy
 
 	ble	a2, zero, 2f
-	addu	t1, a1, a2		# t1 = end of to region
+	PTR_ADDU	t1, a1, a2		# t1 = end of to region
 1:
 	lb	v1, -1(t0)		# copy bytes backwards,
-	subu	t0, t0, 1		#   doesnt happen often so do slow way
-	subu	t1, t1, 1
+	PTR_SUBU	t0, t0, 1		#   doesnt happen often so do slow way
+	PTR_SUBU	t1, t1, 1
 	bne	t0, a0, 1b
 	sb	v1, 0(t1)
 2:
@@ -811,59 +690,59 @@ forward:
 	bne	t2, zero, smallcpy	# do a small bcopy
 	xor	v1, a0, a1		# compare low two bits of addresses
 	and	v1, v1, 3
-	subu	a3, zero, a1		# compute # bytes to word align address
+	PTR_SUBU	a3, zero, a1		# compute # bytes to word align address
 	beq	v1, zero, aligned	# addresses can be word aligned
 	and	a3, a3, 3
 
 	beq	a3, zero, 1f
-	subu	a2, a2, a3		# subtract from remaining count
+	PTR_SUBU	a2, a2, a3		# subtract from remaining count
 	LWHI	v1, 0(a0)		# get next 4 bytes (unaligned)
 	LWLO	v1, 3(a0)
-	addu	a0, a0, a3
+	PTR_ADDU	a0, a0, a3
 	SWHI	v1, 0(a1)		# store 1, 2, or 3 bytes to align a1
-	addu	a1, a1, a3
+	PTR_ADDU	a1, a1, a3
 1:
 	and	v1, a2, 3		# compute number of words left
-	subu	a3, a2, v1
+	PTR_SUBU	a3, a2, v1
 	move	a2, v1
-	addu	a3, a3, a0		# compute ending address
+	PTR_ADDU	a3, a3, a0		# compute ending address
 2:
 	LWHI	v1, 0(a0)		# copy words a0 unaligned, a1 aligned
 	LWLO	v1, 3(a0)
-	addu	a0, a0, 4
+	PTR_ADDU	a0, a0, 4
 	sw	v1, 0(a1)
-	addu	a1, a1, 4
+	PTR_ADDU	a1, a1, 4
 	bne	a0, a3, 2b
 	nop				# We have to do this mmu-bug.
 	b	smallcpy
 	nop
 aligned:
 	beq	a3, zero, 1f
-	subu	a2, a2, a3		# subtract from remaining count
+	PTR_SUBU	a2, a2, a3		# subtract from remaining count
 	LWHI	v1, 0(a0)		# copy 1, 2, or 3 bytes to align
-	addu	a0, a0, a3
+	PTR_ADDU	a0, a0, a3
 	SWHI	v1, 0(a1)
-	addu	a1, a1, a3
+	PTR_ADDU	a1, a1, a3
 1:
 	and	v1, a2, 3		# compute number of whole words left
-	subu	a3, a2, v1
+	PTR_SUBU	a3, a2, v1
 	move	a2, v1
-	addu	a3, a3, a0		# compute ending address
+	PTR_ADDU	a3, a3, a0		# compute ending address
 2:
 	lw	v1, 0(a0)		# copy words
-	addu	a0, a0, 4
+	PTR_ADDU	a0, a0, 4
 	sw	v1, 0(a1)
 	bne	a0, a3, 2b
-	addu	a1, a1, 4
+	PTR_ADDU	a1, a1, 4
 smallcpy:
 	ble	a2, zero, 2f
-	addu	a3, a2, a0		# compute ending address
+	PTR_ADDU	a3, a2, a0		# compute ending address
 1:
 	lbu	v1, 0(a0)		# copy bytes
-	addu	a0, a0, 1
+	PTR_ADDU	a0, a0, 1
 	sb	v1, 0(a1)
 	bne	a0, a3, 1b
-	addu	a1, a1, 1	   # MMU BUG ? can not do -1(a1) at 0x80000000!!
+	PTR_ADDU	a1, a1, 1	   # MMU BUG ? can not do -1(a1) at 0x80000000!!
 2:
 	j	ra
 	nop
@@ -883,19 +762,19 @@ LEAF(memset)
 	sll	t2, t1, 16		# shift that left 16
 	or	t1, t2, t1		# or together
 
-	subu	t0, zero, a0		# compute # bytes to word align address
+	PTR_SUBU	t0, zero, a0		# compute # bytes to word align address
 	and	t0, t0, 3
 	beq	t0, zero, 1f		# skip if word aligned
-	subu	a2, a2, t0		# subtract from remaining count
+	PTR_SUBU	a2, a2, t0		# subtract from remaining count
 	SWHI	t1, 0(a0)		# store 1, 2, or 3 bytes to align
-	addu	a0, a0, t0
+	PTR_ADDU	a0, a0, t0
 1:
 	and	v1, a2, 3		# compute number of whole words left
-	subu	t0, a2, v1
-	subu	a2, a2, t0
-	addu	t0, t0, a0		# compute ending address
+	PTR_SUBU	t0, a2, v1
+	PTR_SUBU	a2, a2, t0
+	PTR_ADDU	t0, t0, a0		# compute ending address
 2:
-	addu	a0, a0, 4		# clear words
+	PTR_ADDU	a0, a0, 4		# clear words
 #ifdef MIPS3_5900
 	nop
 	nop
@@ -907,9 +786,9 @@ LEAF(memset)
 
 memsetsmallclr:
 	ble	a2, zero, 2f
-	addu	t0, a2, a0		# compute ending address
+	PTR_ADDU	t0, a2, a0		# compute ending address
 1:
-	addu	a0, a0, 1		# clear bytes
+	PTR_ADDU	a0, a0, 1		# clear bytes
 #ifdef MIPS3_5900
 	nop
 	nop
@@ -931,26 +810,26 @@ LEAF(bzero)
 ALEAF(blkclr)
 	.set	noreorder
 	blt	a1, 12, smallclr	# small amount to clear?
-	subu	a3, zero, a0		# compute # bytes to word align address
+	PTR_SUBU	a3, zero, a0		# compute # bytes to word align address
 	and	a3, a3, 3
 	beq	a3, zero, 1f		# skip if word aligned
-	subu	a1, a1, a3		# subtract from remaining count
+	PTR_SUBU	a1, a1, a3		# subtract from remaining count
 	SWHI	zero, 0(a0)		# clear 1, 2, or 3 bytes to align
-	addu	a0, a0, a3
+	PTR_ADDU	a0, a0, a3
 1:
 	and	v0, a1, 3		# compute number of words left
-	subu	a3, a1, v0
+	PTR_SUBU	a3, a1, v0
 	move	a1, v0
-	addu	a3, a3, a0		# compute ending address
+	PTR_ADDU	a3, a3, a0		# compute ending address
 2:
-	addu	a0, a0, 4		# clear words
+	PTR_ADDU	a0, a0, 4		# clear words
 	bne	a0, a3, 2b		#  unrolling loop does not help
 	sw	zero, -4(a0)		#  since we are limited by memory speed
 smallclr:
 	ble	a1, zero, 2f
-	addu	a3, a1, a0		# compute ending address
+	PTR_ADDU	a3, a1, a0		# compute ending address
 1:
-	addu	a0, a0, 1		# clear bytes
+	PTR_ADDU	a0, a0, 1		# clear bytes
 	bne	a0, a3, 1b
 	sb	zero, -1(a0)
 2:
@@ -967,66 +846,66 @@ LEAF(bcmp)
 	blt	a2, 16, smallcmp	# is it worth any trouble?
 	xor	v0, a0, a1		# compare low two bits of addresses
 	and	v0, v0, 3
-	subu	a3, zero, a1		# compute # bytes to word align address
+	PTR_SUBU	a3, zero, a1		# compute # bytes to word align address
 	bne	v0, zero, unalignedcmp	# not possible to align addresses
 	and	a3, a3, 3
 
 	beq	a3, zero, 1f
-	subu	a2, a2, a3		# subtract from remaining count
+	PTR_SUBU	a2, a2, a3		# subtract from remaining count
 	move	v0, v1			# init v0,v1 so unmodified bytes match
 	LWHI	v0, 0(a0)		# read 1, 2, or 3 bytes
 	LWHI	v1, 0(a1)
-	addu	a1, a1, a3
+	PTR_ADDU	a1, a1, a3
 	bne	v0, v1, nomatch
-	addu	a0, a0, a3
+	PTR_ADDU	a0, a0, a3
 1:
 	and	a3, a2, ~3		# compute number of whole words left
-	subu	a2, a2, a3		#   which has to be >= (16-3) & ~3
-	addu	a3, a3, a0		# compute ending address
+	PTR_SUBU	a2, a2, a3		#   which has to be >= (16-3) & ~3
+	PTR_ADDU	a3, a3, a0		# compute ending address
 2:
 	lw	v0, 0(a0)		# compare words
 	lw	v1, 0(a1)
-	addu	a0, a0, 4
+	PTR_ADDU	a0, a0, 4
 	bne	v0, v1, nomatch
-	addu	a1, a1, 4
+	PTR_ADDU	a1, a1, 4
 	bne	a0, a3, 2b
 	nop
 	b	smallcmp		# finish remainder
 	nop
 unalignedcmp:
 	beq	a3, zero, 2f
-	subu	a2, a2, a3		# subtract from remaining count
-	addu	a3, a3, a0		# compute ending address
+	PTR_SUBU	a2, a2, a3		# subtract from remaining count
+	PTR_ADDU	a3, a3, a0		# compute ending address
 1:
 	lbu	v0, 0(a0)		# compare bytes until a1 word aligned
 	lbu	v1, 0(a1)
-	addu	a0, a0, 1
+	PTR_ADDU	a0, a0, 1
 	bne	v0, v1, nomatch
-	addu	a1, a1, 1
+	PTR_ADDU	a1, a1, 1
 	bne	a0, a3, 1b
 	nop
 2:
 	and	a3, a2, ~3		# compute number of whole words left
-	subu	a2, a2, a3		#   which has to be >= (16-3) & ~3
-	addu	a3, a3, a0		# compute ending address
+	PTR_SUBU	a2, a2, a3		#   which has to be >= (16-3) & ~3
+	PTR_ADDU	a3, a3, a0		# compute ending address
 3:
 	LWHI	v0, 0(a0)		# compare words a0 unaligned, a1 aligned
 	LWLO	v0, 3(a0)
 	lw	v1, 0(a1)
-	addu	a0, a0, 4
+	PTR_ADDU	a0, a0, 4
 	bne	v0, v1, nomatch
-	addu	a1, a1, 4
+	PTR_ADDU	a1, a1, 4
 	bne	a0, a3, 3b
 	nop
 smallcmp:
 	ble	a2, zero, match
-	addu	a3, a2, a0		# compute ending address
+	PTR_ADDU	a3, a2, a0		# compute ending address
 1:
 	lbu	v0, 0(a0)
 	lbu	v1, 0(a1)
-	addu	a0, a0, 1
+	PTR_ADDU	a0, a0, 1
 	bne	v0, v1, nomatch
-	addu	a1, a1, 1
+	PTR_ADDU	a1, a1, 1
 	bne	a0, a3, 1b
 	nop
 match:
@@ -1367,9 +1246,6 @@ END(atomic_subtract_8)
  */
 
 	.set	noreorder		# Noreorder is default style!
-#ifndef _MIPS_ARCH_XLR
-	.set	mips3
-#endif
 
 #if !defined(__mips_n64) && !defined(__mips_n32)	
 	/*
@@ -1426,22 +1302,22 @@ END(atomic_load_64)
 #if defined(DDB) || defined(DEBUG)
 
 LEAF(kdbpeek)
-	li	v1, DDBERR
+	PTR_LA	v1, ddberr
 	and	v0, a0, 3			# unaligned ?
 	GET_CPU_PCPU(t1)
-	lw	t1, PC_CURPCB(t1)
+	PTR_L	t1, PC_CURPCB(t1)
 	bne	v0, zero, 1f
-	sw	v1, U_PCB_ONFAULT(t1)
+	PTR_S	v1, U_PCB_ONFAULT(t1)
 
 	lw	v0, (a0)
 	jr	ra
-	sw	zero, U_PCB_ONFAULT(t1)
+	PTR_S	zero, U_PCB_ONFAULT(t1)
 
 1:
 	LWHI	v0, 0(a0)
 	LWLO	v0, 3(a0)
 	jr	ra
-	sw	zero, U_PCB_ONFAULT(t1)
+	PTR_S	zero, U_PCB_ONFAULT(t1)
 END(kdbpeek)
 
 ddberr:
@@ -1450,44 +1326,31 @@ ddberr:
 
 #if defined(DDB)
 LEAF(kdbpoke)
-	li	v1, DDBERR
+	PTR_LA	v1, ddberr
 	and	v0, a0, 3			# unaligned ?
 	GET_CPU_PCPU(t1)
-	lw	t1, PC_CURPCB(t1)
+	PTR_L	t1, PC_CURPCB(t1)
 	bne	v0, zero, 1f
-	sw	v1, U_PCB_ONFAULT(t1)
+	PTR_S	v1, U_PCB_ONFAULT(t1)
 
 	sw	a1, (a0)
 	jr	ra
-	sw	zero, U_PCB_ONFAULT(t1)
+	PTR_S	zero, U_PCB_ONFAULT(t1)
 
 1:
 	SWHI	a1, 0(a0)
 	SWLO	a1, 3(a0)
 	jr	ra
-	sw	zero, U_PCB_ONFAULT(t1)
+	PTR_S	zero, U_PCB_ONFAULT(t1)
 END(kdbpoke)
 
 	.data
 	.globl	esym
 esym:	.word	0
 
-#ifndef _MIPS_ARCH_XLR
-	.set	mips2
-#endif
 #endif /* DDB */
 #endif /* DDB || DEBUG */
 
-#ifndef MIPS_ISAIII
-#define	STORE	sw	/* 32 bit mode regsave instruction */
-#define	LOAD	lw	/* 32 bit mode regload instruction */
-#define	RSIZE	4	/* 32 bit mode register size */
-#else
-#define	STORE	sd	/* 64 bit mode regsave instruction */
-#define	LOAD	ld	/* 64 bit mode regload instruction */
-#define	RSIZE	8	/* 64 bit mode register size */
-#endif
-
 #define	ITLBNOPFIX	nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;
 
 	.text
@@ -1499,35 +1362,35 @@ LEAF(breakpoint)
 
 LEAF(setjmp)
 	mfc0	v0, COP_0_STATUS_REG	# Later the "real" spl value!
-	STORE	s0, (RSIZE * PREG_S0)(a0)
-	STORE	s1, (RSIZE * PREG_S1)(a0)
-	STORE	s2, (RSIZE * PREG_S2)(a0)
-	STORE	s3, (RSIZE * PREG_S3)(a0)
-	STORE	s4, (RSIZE * PREG_S4)(a0)
-	STORE	s5, (RSIZE * PREG_S5)(a0)
-	STORE	s6, (RSIZE * PREG_S6)(a0)
-	STORE	s7, (RSIZE * PREG_S7)(a0)
-	STORE	s8, (RSIZE * PREG_S8)(a0)
-	STORE	sp, (RSIZE * PREG_SP)(a0)
-	STORE	ra, (RSIZE * PREG_RA)(a0)
-	STORE	v0, (RSIZE * PREG_SR)(a0)
+	REG_S	s0, (SZREG * PREG_S0)(a0)
+	REG_S	s1, (SZREG * PREG_S1)(a0)
+	REG_S	s2, (SZREG * PREG_S2)(a0)
+	REG_S	s3, (SZREG * PREG_S3)(a0)
+	REG_S	s4, (SZREG * PREG_S4)(a0)
+	REG_S	s5, (SZREG * PREG_S5)(a0)
+	REG_S	s6, (SZREG * PREG_S6)(a0)
+	REG_S	s7, (SZREG * PREG_S7)(a0)
+	REG_S	s8, (SZREG * PREG_S8)(a0)
+	REG_S	sp, (SZREG * PREG_SP)(a0)
+	REG_S	ra, (SZREG * PREG_RA)(a0)
+	REG_S	v0, (SZREG * PREG_SR)(a0)
 	jr	ra
 	li	v0, 0			# setjmp return
 END(setjmp)
 
 LEAF(longjmp)
-	LOAD	v0, (RSIZE * PREG_SR)(a0)
-	LOAD	ra, (RSIZE * PREG_RA)(a0)
-	LOAD	s0, (RSIZE * PREG_S0)(a0)
-	LOAD	s1, (RSIZE * PREG_S1)(a0)
-	LOAD	s2, (RSIZE * PREG_S2)(a0)
-	LOAD	s3, (RSIZE * PREG_S3)(a0)
-	LOAD	s4, (RSIZE * PREG_S4)(a0)
-	LOAD	s5, (RSIZE * PREG_S5)(a0)
-	LOAD	s6, (RSIZE * PREG_S6)(a0)
-	LOAD	s7, (RSIZE * PREG_S7)(a0)
-	LOAD	s8, (RSIZE * PREG_S8)(a0)
-	LOAD	sp, (RSIZE * PREG_SP)(a0)
+	REG_L	v0, (SZREG * PREG_SR)(a0)
+	REG_L	ra, (SZREG * PREG_RA)(a0)
+	REG_L	s0, (SZREG * PREG_S0)(a0)
+	REG_L	s1, (SZREG * PREG_S1)(a0)
+	REG_L	s2, (SZREG * PREG_S2)(a0)
+	REG_L	s3, (SZREG * PREG_S3)(a0)
+	REG_L	s4, (SZREG * PREG_S4)(a0)
+	REG_L	s5, (SZREG * PREG_S5)(a0)
+	REG_L	s6, (SZREG * PREG_S6)(a0)
+	REG_L	s7, (SZREG * PREG_S7)(a0)
+	REG_L	s8, (SZREG * PREG_S8)(a0)
+	REG_L	sp, (SZREG * PREG_SP)(a0)
 	mtc0	v0, COP_0_STATUS_REG	# Later the "real" spl value!
 	ITLBNOPFIX
 	jr	ra
@@ -1538,7 +1401,6 @@ LEAF(fusufault)
 	GET_CPU_PCPU(t0)
 	lw	t0, PC_CURTHREAD(t0)
 	lw	t0, TD_PCB(t0)
-	sw	zero, U_PCB_ONFAULT(t0)
 	li	v0, -1
 	j	ra
 END(fusufault)
@@ -1547,8 +1409,7 @@ END(fusufault)
        a pointer that is in user space.	 It will be used as the basic primitive
        for a kernel supported user space lock implementation. */
 LEAF(casuptr)
-
-	li	t0, VM_MAXUSER_ADDRESS /* verify address validity */
+	PTR_LI	t0, VM_MAXUSER_ADDRESS /* verify address validity */
 	blt	a0, t0, fusufault		/* trap faults */
 	nop
 
@@ -1556,8 +1417,8 @@ LEAF(casuptr)
 	lw	t1, PC_CURTHREAD(t1)
 	lw	t1, TD_PCB(t1)
 
-	lw	t2, fusufault
-	sw	t2, U_PCB_ONFAULT(t1)
+	PTR_LA	t2, fusufault
+	PTR_S	t2, U_PCB_ONFAULT(t1)
 1:
 	ll	v0, 0(a0)		/* try to load the old value */
 	beq	v0, a1, 2f		/* compare */
@@ -1565,7 +1426,7 @@ LEAF(casuptr)
 	sc	t0, 0(a0)		/* write if address still locked */
 	beq	t0, zero, 1b			/* if it failed, spin */
 2:
-	sw	zero, U_PCB_ONFAULT(t1) /* clean up */
+	PTR_S	zero, U_PCB_ONFAULT(t1) /* clean up */
 	j	ra
 END(casuptr)
 
@@ -1593,7 +1454,7 @@ END(octeon_get_shadow)
  * octeon_set_control(addr, uint32_t val)
  */
 LEAF(octeon_set_control)
-	.set mips64r2
+	.set push
 	or      t1, a1, zero
 /*	dmfc0   a1, 9, 7*/
 	.word 0x40254807
@@ -1603,20 +1464,21 @@ LEAF(octeon_set_control)
 	.word 0x40a54807
 	jr	ra
 	nop
-        .set	mips0
+	.set pop
 END(octeon_set_control)
 
 /*
  * octeon_get_control(addr)
  */
 LEAF(octeon_get_control)
+	.set push
 	.set mips64r2
 /*	dmfc0   a1, 9, 7 */
 	.word 0x40254807
 	sd	a1, 0(a0)
 	jr	ra
 	nop
-        .set	mips0
+	.set pop
 END(octeon_get_control)
 #endif
 
diff --git a/sys/mips/mips/trap.c b/sys/mips/mips/trap.c
index 1e28188330b..17a6be5e7fd 100644
--- a/sys/mips/mips/trap.c
+++ b/sys/mips/mips/trap.c
@@ -99,8 +99,7 @@ __FBSDID("$FreeBSD$");
 int trap_debug = 1;
 #endif
 
-extern unsigned onfault_table[];
-
+static void log_illegal_instruction(const char *, struct trapframe *);
 static void log_bad_page_fault(char *, struct trapframe *, int);
 static void log_frame_dump(struct trapframe *frame);
 static void get_mapping_info(vm_offset_t, pd_entry_t **, pt_entry_t **);
@@ -226,8 +225,8 @@ void stacktrace(struct trapframe *);
 void logstacktrace(struct trapframe *);
 #endif
 
-#define	KERNLAND(x) ((int)(x) < 0)
-#define	DELAYBRANCH(x) ((int)(x) < 0)
+#define	KERNLAND(x)	((vm_offset_t)(x) >= VM_MIN_KERNEL_ADDRESS && (vm_offset_t)(x) < VM_MAX_KERNEL_ADDRESS)
+#define	DELAYBRANCH(x)	((int)(x) < 0)
 
 /*
  * MIPS load/store access type
@@ -263,6 +262,7 @@ SYSCTL_INT(_vm, OID_AUTO, allow_unaligned_acc, CTLFLAG_RW,
 static int emulate_unaligned_access(struct trapframe *frame);
 
 extern char *syscallnames[];
+extern void fswintrberr(void); /* XXX */
 
 /*
  * Handle an exception.
@@ -281,13 +281,12 @@ trap(struct trapframe *trapframe)
 	struct proc *p = curproc;
 	vm_prot_t ftype;
 	pt_entry_t *pte;
-	unsigned int entry;
 	pmap_t pmap;
-	int quad_syscall = 0;
 	int access_type;
 	ksiginfo_t ksi;
 	char *msg = NULL;
-	register_t addr = 0;
+	intptr_t addr = 0;
+	register_t pc;
 
 	trapdebug_enter(trapframe, 0);
 	
@@ -330,9 +329,9 @@ trap(struct trapframe *trapframe)
 		printf("cpuid = %d\n", PCPU_GET(cpuid));
 #endif
 		MachTLBGetPID(pid);
-		printf("badaddr = 0x%0x, pc = 0x%0x, ra = 0x%0x, sp = 0x%0x, sr = 0x%x, pid = %d, ASID = 0x%x\n",
-		    trapframe->badvaddr, trapframe->pc, trapframe->ra,
-		    trapframe->sp, trapframe->sr,
+		printf("badaddr = %#jx, pc = %#jx, ra = %#jx, sp = %#jx, sr = %jx, pid = %d, ASID = %u\n",
+		    (intmax_t)trapframe->badvaddr, (intmax_t)trapframe->pc, (intmax_t)trapframe->ra,
+		    (intmax_t)trapframe->sp, (intmax_t)trapframe->sr,
 		    (curproc ? curproc->p_pid : -1), pid);
 
 		switch (type & ~T_USER) {
@@ -354,7 +353,7 @@ trap(struct trapframe *trapframe)
 		    ((type & ~T_USER) != T_SYSCALL)) {
 			if (++count == 3) {
 				trap_frame_dump(trapframe);
-				panic("too many faults at %x\n", last_badvaddr);
+				panic("too many faults at %p\n", (void *)last_badvaddr);
 			}
 		} else {
 			last_badvaddr = this_badvaddr;
@@ -375,35 +374,30 @@ trap(struct trapframe *trapframe)
 			vm_offset_t pa;
 
 			PMAP_LOCK(kernel_pmap);
-			if (!(pte = pmap_segmap(kernel_pmap,
-			    trapframe->badvaddr)))
-				panic("trap: ktlbmod: invalid segmap");
-			pte += (trapframe->badvaddr >> PAGE_SHIFT) & (NPTEPG - 1);
-			entry = *pte;
+			pte = pmap_pte(kernel_pmap, trapframe->badvaddr);
+			if (pte == NULL)
+				panic("trap: ktlbmod: can't find PTE");
 #ifdef SMP
 			/* It is possible that some other CPU changed m-bit */
-			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit())) {
-				trapframe->badvaddr &= ~PAGE_MASK;
+			if (!mips_pg_v(*pte) || (*pte & mips_pg_m_bit())) {
 				pmap_update_page(kernel_pmap,
-				    trapframe->badvaddr, entry);
+				    trapframe->badvaddr, *pte);
 				PMAP_UNLOCK(kernel_pmap);
 				return (trapframe->pc);
 			}
 #else
-			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit()))
+			if (!mips_pg_v(*pte) || (*pte & mips_pg_m_bit()))
 				panic("trap: ktlbmod: invalid pte");
 #endif
-			if (entry & mips_pg_ro_bit()) {
+			if (*pte & mips_pg_ro_bit()) {
 				/* write to read only page in the kernel */
 				ftype = VM_PROT_WRITE;
 				PMAP_UNLOCK(kernel_pmap);
 				goto kernel_fault;
 			}
-			entry |= mips_pg_m_bit();
-			*pte = entry;
-			trapframe->badvaddr &= ~PAGE_MASK;
-			pmap_update_page(kernel_pmap, trapframe->badvaddr, entry);
-			pa = mips_tlbpfn_to_paddr(entry);
+			*pte |= mips_pg_m_bit();
+			pmap_update_page(kernel_pmap, trapframe->badvaddr, *pte);
+			pa = mips_tlbpfn_to_paddr(*pte);
 			if (!page_is_managed(pa))
 				panic("trap: ktlbmod: unmanaged page");
 			pmap_set_modified(pa);
@@ -419,36 +413,30 @@ trap(struct trapframe *trapframe)
 			pmap = &p->p_vmspace->vm_pmap;
 
 			PMAP_LOCK(pmap);
-			if (!(pte = pmap_segmap(pmap, trapframe->badvaddr)))
-				panic("trap: utlbmod: invalid segmap");
-			pte += (trapframe->badvaddr >> PAGE_SHIFT) & (NPTEPG - 1);
-			entry = *pte;
+			pte = pmap_pte(pmap, trapframe->badvaddr);
+			if (pte == NULL)
+				panic("trap: utlbmod: can't find PTE");
 #ifdef SMP
 			/* It is possible that some other CPU changed m-bit */
-			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit())) {
-				trapframe->badvaddr = (trapframe->badvaddr & ~PAGE_MASK);
-				pmap_update_page(pmap, trapframe->badvaddr, entry);
+			if (!mips_pg_v(*pte) || (*pte & mips_pg_m_bit())) {
+				pmap_update_page(pmap, trapframe->badvaddr, *pte);
 				PMAP_UNLOCK(pmap);
 				goto out;
 			}
 #else
-			if (!mips_pg_v(entry) || (entry & mips_pg_m_bit())) {
+			if (!mips_pg_v(*pte) || (*pte & mips_pg_m_bit()))
 				panic("trap: utlbmod: invalid pte");
-			}
 #endif
 
-			if (entry & mips_pg_ro_bit()) {
+			if (*pte & mips_pg_ro_bit()) {
 				/* write to read only page */
 				ftype = VM_PROT_WRITE;
 				PMAP_UNLOCK(pmap);
 				goto dofault;
 			}
-			entry |= mips_pg_m_bit();
-			*pte = entry;
-			trapframe->badvaddr = (trapframe->badvaddr & ~PAGE_MASK);
-			pmap_update_page(pmap, trapframe->badvaddr, entry);
-			trapframe->badvaddr |= (pmap->pm_asid[PCPU_GET(cpuid)].asid << VMTLB_PID_SHIFT);
-			pa = mips_tlbpfn_to_paddr(entry);
+			*pte |= mips_pg_m_bit();
+			pmap_update_page(pmap, trapframe->badvaddr, *pte);
+			pa = mips_tlbpfn_to_paddr(*pte);
 			if (!page_is_managed(pa))
 				panic("trap: utlbmod: unmanaged page");
 			pmap_set_modified(pa);
@@ -473,22 +461,29 @@ trap(struct trapframe *trapframe)
 			rv = vm_fault(kernel_map, va, ftype, VM_FAULT_NORMAL);
 			if (rv == KERN_SUCCESS)
 				return (trapframe->pc);
-			if ((i = td->td_pcb->pcb_onfault) != 0) {
-				td->td_pcb->pcb_onfault = 0;
-				return (onfault_table[i]);
+			if (td->td_pcb->pcb_onfault != NULL) {
+				pc = (register_t)(intptr_t)td->td_pcb->pcb_onfault;
+				td->td_pcb->pcb_onfault = NULL;
+				return (pc);
 			}
 			goto err;
 		}
-		/*
+
+                /*
 		 * It is an error for the kernel to access user space except
 		 * through the copyin/copyout routines.
 		 */
-		if ((i = td->td_pcb->pcb_onfault) == 0)
+		if (td->td_pcb->pcb_onfault == NULL)
 			goto err;
+
 		/* check for fuswintr() or suswintr() getting a page fault */
-		if (i == 4) {
-			return (onfault_table[i]);
+		/* XXX There must be a nicer way to do this.  */
+		if (td->td_pcb->pcb_onfault == fswintrberr) {
+			pc = (register_t)(intptr_t)td->td_pcb->pcb_onfault;
+			td->td_pcb->pcb_onfault = NULL;
+			return (pc);
 		}
+
 		goto dofault;
 
 	case T_TLB_LD_MISS + T_USER:
@@ -507,7 +502,7 @@ dofault:
 			vm = p->p_vmspace;
 			map = &vm->vm_map;
 			va = trunc_page((vm_offset_t)trapframe->badvaddr);
-			if ((vm_offset_t)trapframe->badvaddr >= VM_MIN_KERNEL_ADDRESS) {
+			if (KERNLAND(trapframe->badvaddr)) {
 				/*
 				 * Don't allow user-mode faults in kernel
 				 * address space.
@@ -529,9 +524,9 @@ dofault:
 			--p->p_lock;
 			PROC_UNLOCK(p);
 #ifdef VMFAULT_TRACE
-			printf("vm_fault(%p (pmap %p), %x (%x), %x, %d) -> %x at pc %x\n",
-			    map, &vm->vm_pmap, va, trapframe->badvaddr, ftype, VM_FAULT_NORMAL,
-			    rv, trapframe->pc);
+			printf("vm_fault(%p (pmap %p), %p (%p), %x, %d) -> %x at pc %p\n",
+			    map, &vm->vm_pmap, (void *)va, (void *)(intptr_t)trapframe->badvaddr,
+			    ftype, VM_FAULT_NORMAL, rv, (void *)(intptr_t)trapframe->pc);
 #endif
 
 			if (rv == KERN_SUCCESS) {
@@ -542,9 +537,10 @@ dofault:
 			}
 	nogo:
 			if (!usermode) {
-				if ((i = td->td_pcb->pcb_onfault) != 0) {
-					td->td_pcb->pcb_onfault = 0;
-					return (onfault_table[i]);
+				if (td->td_pcb->pcb_onfault != NULL) {
+					pc = (register_t)(intptr_t)td->td_pcb->pcb_onfault;
+					td->td_pcb->pcb_onfault = NULL;
+					return (pc);
 				}
 				goto err;
 			}
@@ -606,6 +602,8 @@ dofault:
 			int nargs, nsaved;
 			register_t args[8];
 
+			bzero(args, sizeof args);
+
 			/*
 			 * note: PCPU_LAZY_INC() can only be used if we can
 			 * afford occassional inaccuracy in the count.
@@ -654,7 +652,6 @@ dofault:
 				args[0] = locr0->a2;
 				args[1] = locr0->a3;
 				nsaved = 2;
-				quad_syscall = 1;
 				break;
 
 			default:
@@ -679,7 +676,7 @@ dofault:
 			nargs = callp->sy_narg;
 
 			if (nargs > nsaved) {
-				i = copyin((caddr_t)(locr0->sp +
+				i = copyin((caddr_t)(intptr_t)(locr0->sp +
 				    4 * sizeof(register_t)), (caddr_t)&args[nsaved],
 				    (u_int)(nargs - nsaved) * sizeof(register_t));
 				if (i) {
@@ -770,7 +767,8 @@ dofault:
 
 	case T_BREAK + T_USER:
 		{
-			uintptr_t va, instr;
+			intptr_t va;
+			uint32_t instr;
 
 			/* compute address of break instruction */
 			va = trapframe->pc;
@@ -803,7 +801,7 @@ dofault:
 	case T_IWATCH + T_USER:
 	case T_DWATCH + T_USER:
 		{
-			uintptr_t va;
+			intptr_t va;
 
 			/* compute address of trapped instruction */
 			va = trapframe->pc;
@@ -817,7 +815,8 @@ dofault:
 
 	case T_TRAP + T_USER:
 		{
-			uintptr_t va, instr;
+			intptr_t va;
+			uint32_t instr;
 			struct trapframe *locr0 = td->td_frame;
 
 			/* compute address of trap instruction */
@@ -839,6 +838,7 @@ dofault:
 		}
 
 	case T_RES_INST + T_USER:
+		log_illegal_instruction("RES_INST", trapframe);
 		i = SIGILL;
 		addr = trapframe->pc;
 		break;
@@ -853,11 +853,13 @@ dofault:
 #if !defined(CPU_HAVEFPU)
 		/* FP (COP1) instruction */
 		if ((trapframe->cause & CR_COP_ERR) == 0x10000000) {
+			log_illegal_instruction("COP1_UNUSABLE", trapframe);
 			i = SIGILL;
 			break;
 		}
 #endif
 		if ((trapframe->cause & CR_COP_ERR) != 0x10000000) {
+			log_illegal_instruction("COPn_UNUSABLE", trapframe);
 			i = SIGILL;	/* only FPU instructions allowed */
 			break;
 		}
@@ -872,13 +874,13 @@ dofault:
 #if !defined(SMP) && (defined(DDB) || defined(DEBUG))
 		trapDump("fpintr");
 #else
-		printf("FPU Trap: PC %x CR %x SR %x\n",
-		    trapframe->pc, trapframe->cause, trapframe->sr);
+		printf("FPU Trap: PC %#jx CR %x SR %x\n",
+		    (intmax_t)trapframe->pc, (unsigned)trapframe->cause, (unsigned)trapframe->sr);
 		goto err;
 #endif
 
 	case T_FPE + T_USER:
-		MachFPTrap(trapframe->sr, trapframe->cause, trapframe->pc);
+		MipsFPTrap(trapframe->sr, trapframe->cause, trapframe->pc);
 		goto out;
 
 	case T_OVFLOW + T_USER:
@@ -889,8 +891,8 @@ dofault:
 	case T_ADDR_ERR_LD:	/* misaligned access */
 	case T_ADDR_ERR_ST:	/* misaligned access */
 #ifdef TRAP_DEBUG
-		printf("+++ ADDR_ERR: type = %d, badvaddr = %x\n", type,
-		    trapframe->badvaddr);
+		printf("+++ ADDR_ERR: type = %d, badvaddr = %#jx\n", type,
+		    (intmax_t)trapframe->badvaddr);
 #endif
 		/* Only allow emulation on a user address */
 		if (allow_unaligned_acc &&
@@ -922,10 +924,12 @@ dofault:
 		/* FALLTHROUGH */
 
 	case T_BUS_ERR_LD_ST:	/* BERR asserted to cpu */
-		if ((i = td->td_pcb->pcb_onfault) != 0) {
-			td->td_pcb->pcb_onfault = 0;
-			return (onfault_table[i]);
+		if (td->td_pcb->pcb_onfault != NULL) {
+			pc = (register_t)(intptr_t)td->td_pcb->pcb_onfault;
+			td->td_pcb->pcb_onfault = NULL;
+			return (pc);
 		}
+
 		/* FALLTHROUGH */
 
 	default:
@@ -947,9 +951,9 @@ err:
 			printf("kernel mode)\n");
 
 #ifdef TRAP_DEBUG
-		printf("badvaddr = %x, pc = %x, ra = %x, sr = 0x%x\n",
-		       trapframe->badvaddr, trapframe->pc, trapframe->ra,
-		       trapframe->sr);
+		printf("badvaddr = %#jx, pc = %#jx, ra = %#jx, sr = %#jxx\n",
+		       (intmax_t)trapframe->badvaddr, (intmax_t)trapframe->pc, (intmax_t)trapframe->ra,
+		       (intmax_t)trapframe->sr);
 #endif
 
 #ifdef KDB
@@ -997,11 +1001,11 @@ trapDump(char *msg)
 		if (trp->cause == 0)
 			break;
 
-		printf("%s: ADR %x PC %x CR %x SR %x\n",
+		printf("%s: ADR %jx PC %jx CR %jx SR %jx\n",
 		    trap_type[(trp->cause & CR_EXC_CODE) >> CR_EXC_CODE_SHIFT],
-		    trp->vadr, trp->pc, trp->cause, trp->status);
+		    (intmax_t)trp->vadr, (intmax_t)trp->pc, (intmax_t)trp->cause, (intmax_t)trp->status);
 
-		printf("   RA %x SP %x code %d\n", trp->ra, trp->sp, trp->code);
+		printf("   RA %jx SP %jx code %d\n", (intmax_t)trp->ra, (intmax_t)trp->sp, (int)trp->code);
 	}
 	intr_restore(s);
 }
@@ -1165,39 +1169,39 @@ static void
 log_frame_dump(struct trapframe *frame)
 {
 	log(LOG_ERR, "Trapframe Register Dump:\n");
-	log(LOG_ERR, "\tzero: %p\tat: %p\tv0: %p\tv1: %p\n",
-	    (void *)0, (void *)frame->ast, (void *)frame->v0, (void *)frame->v1);
+	log(LOG_ERR, "\tzero: %#jx\tat: %#jx\tv0: %#jx\tv1: %#jx\n",
+	    (intmax_t)0, (intmax_t)frame->ast, (intmax_t)frame->v0, (intmax_t)frame->v1);
 
-	log(LOG_ERR, "\ta0: %p\ta1: %p\ta2: %p\ta3: %p\n",
-	    (void *)frame->a0, (void *)frame->a1, (void *)frame->a2, (void *)frame->a3);
+	log(LOG_ERR, "\ta0: %#jx\ta1: %#jx\ta2: %#jx\ta3: %#jx\n",
+	    (intmax_t)frame->a0, (intmax_t)frame->a1, (intmax_t)frame->a2, (intmax_t)frame->a3);
 
-	log(LOG_ERR, "\tt0: %p\tt1: %p\tt2: %p\tt3: %p\n",
-	    (void *)frame->t0, (void *)frame->t1, (void *)frame->t2, (void *)frame->t3);
+	log(LOG_ERR, "\tt0: %#jx\tt1: %#jx\tt2: %#jx\tt3: %#jx\n",
+	    (intmax_t)frame->t0, (intmax_t)frame->t1, (intmax_t)frame->t2, (intmax_t)frame->t3);
 
-	log(LOG_ERR, "\tt4: %p\tt5: %p\tt6: %p\tt7: %p\n",
-	    (void *)frame->t4, (void *)frame->t5, (void *)frame->t6, (void *)frame->t7);
+	log(LOG_ERR, "\tt4: %#jx\tt5: %#jx\tt6: %#jx\tt7: %#jx\n",
+	    (intmax_t)frame->t4, (intmax_t)frame->t5, (intmax_t)frame->t6, (intmax_t)frame->t7);
 
-	log(LOG_ERR, "\tt8: %p\tt9: %p\ts0: %p\ts1: %p\n",
-	    (void *)frame->t8, (void *)frame->t9, (void *)frame->s0, (void *)frame->s1);
+	log(LOG_ERR, "\tt8: %#jx\tt9: %#jx\ts0: %#jx\ts1: %#jx\n",
+	    (intmax_t)frame->t8, (intmax_t)frame->t9, (intmax_t)frame->s0, (intmax_t)frame->s1);
 
-	log(LOG_ERR, "\ts2: %p\ts3: %p\ts4: %p\ts5: %p\n",
-	    (void *)frame->s2, (void *)frame->s3, (void *)frame->s4, (void *)frame->s5);
+	log(LOG_ERR, "\ts2: %#jx\ts3: %#jx\ts4: %#jx\ts5: %#jx\n",
+	    (intmax_t)frame->s2, (intmax_t)frame->s3, (intmax_t)frame->s4, (intmax_t)frame->s5);
 
-	log(LOG_ERR, "\ts6: %p\ts7: %p\tk0: %p\tk1: %p\n",
-	    (void *)frame->s6, (void *)frame->s7, (void *)frame->k0, (void *)frame->k1);
+	log(LOG_ERR, "\ts6: %#jx\ts7: %#jx\tk0: %#jx\tk1: %#jx\n",
+	    (intmax_t)frame->s6, (intmax_t)frame->s7, (intmax_t)frame->k0, (intmax_t)frame->k1);
 
-	log(LOG_ERR, "\tgp: %p\tsp: %p\ts8: %p\tra: %p\n",
-	    (void *)frame->gp, (void *)frame->sp, (void *)frame->s8, (void *)frame->ra);
+	log(LOG_ERR, "\tgp: %#jx\tsp: %#jx\ts8: %#jx\tra: %#jx\n",
+	    (intmax_t)frame->gp, (intmax_t)frame->sp, (intmax_t)frame->s8, (intmax_t)frame->ra);
 
-	log(LOG_ERR, "\tsr: %p\tmullo: %p\tmulhi: %p\tbadvaddr: %p\n",
-	    (void *)frame->sr, (void *)frame->mullo, (void *)frame->mulhi, (void *)frame->badvaddr);
+	log(LOG_ERR, "\tsr: %#jx\tmullo: %#jx\tmulhi: %#jx\tbadvaddr: %#jx\n",
+	    (intmax_t)frame->sr, (intmax_t)frame->mullo, (intmax_t)frame->mulhi, (intmax_t)frame->badvaddr);
 
 #ifdef IC_REG
-	log(LOG_ERR, "\tcause: %p\tpc: %p\tic: %p\n",
-	    (void *)frame->cause, (void *)frame->pc, (void *)frame->ic);
+	log(LOG_ERR, "\tcause: %#jx\tpc: %#jx\tic: %#jx\n",
+	    (intmax_t)frame->cause, (intmax_t)frame->pc, (intmax_t)frame->ic);
 #else
-	log(LOG_ERR, "\tcause: %p\tpc: %p\n",
-	    (void *)frame->cause, (void *)frame->pc);
+	log(LOG_ERR, "\tcause: %#jx\tpc: %#jx\n",
+	    (intmax_t)frame->cause, (intmax_t)frame->pc);
 #endif
 }
 
@@ -1206,39 +1210,39 @@ static void
 trap_frame_dump(struct trapframe *frame)
 {
 	printf("Trapframe Register Dump:\n");
-	printf("\tzero: %p\tat: %p\tv0: %p\tv1: %p\n",
-	    (void *)0, (void *)frame->ast, (void *)frame->v0, (void *)frame->v1);
+	printf("\tzero: %#jx\tat: %#jx\tv0: %#jx\tv1: %#jx\n",
+	    (intmax_t)0, (intmax_t)frame->ast, (intmax_t)frame->v0, (intmax_t)frame->v1);
 
-	printf("\ta0: %p\ta1: %p\ta2: %p\ta3: %p\n",
-	    (void *)frame->a0, (void *)frame->a1, (void *)frame->a2, (void *)frame->a3);
+	printf("\ta0: %#jx\ta1: %#jx\ta2: %#jx\ta3: %#jx\n",
+	    (intmax_t)frame->a0, (intmax_t)frame->a1, (intmax_t)frame->a2, (intmax_t)frame->a3);
 
-	printf("\tt0: %p\tt1: %p\tt2: %p\tt3: %p\n",
-	    (void *)frame->t0, (void *)frame->t1, (void *)frame->t2, (void *)frame->t3);
+	printf("\tt0: %#jx\tt1: %#jx\tt2: %#jx\tt3: %#jx\n",
+	    (intmax_t)frame->t0, (intmax_t)frame->t1, (intmax_t)frame->t2, (intmax_t)frame->t3);
 
-	printf("\tt4: %p\tt5: %p\tt6: %p\tt7: %p\n",
-	    (void *)frame->t4, (void *)frame->t5, (void *)frame->t6, (void *)frame->t7);
+	printf("\tt4: %#jx\tt5: %#jx\tt6: %#jx\tt7: %#jx\n",
+	    (intmax_t)frame->t4, (intmax_t)frame->t5, (intmax_t)frame->t6, (intmax_t)frame->t7);
 
-	printf("\tt8: %p\tt9: %p\ts0: %p\ts1: %p\n",
-	    (void *)frame->t8, (void *)frame->t9, (void *)frame->s0, (void *)frame->s1);
+	printf("\tt8: %#jx\tt9: %#jx\ts0: %#jx\ts1: %#jx\n",
+	    (intmax_t)frame->t8, (intmax_t)frame->t9, (intmax_t)frame->s0, (intmax_t)frame->s1);
 
-	printf("\ts2: %p\ts3: %p\ts4: %p\ts5: %p\n",
-	    (void *)frame->s2, (void *)frame->s3, (void *)frame->s4, (void *)frame->s5);
+	printf("\ts2: %#jx\ts3: %#jx\ts4: %#jx\ts5: %#jx\n",
+	    (intmax_t)frame->s2, (intmax_t)frame->s3, (intmax_t)frame->s4, (intmax_t)frame->s5);
 
-	printf("\ts6: %p\ts7: %p\tk0: %p\tk1: %p\n",
-	    (void *)frame->s6, (void *)frame->s7, (void *)frame->k0, (void *)frame->k1);
+	printf("\ts6: %#jx\ts7: %#jx\tk0: %#jx\tk1: %#jx\n",
+	    (intmax_t)frame->s6, (intmax_t)frame->s7, (intmax_t)frame->k0, (intmax_t)frame->k1);
 
-	printf("\tgp: %p\tsp: %p\ts8: %p\tra: %p\n",
-	    (void *)frame->gp, (void *)frame->sp, (void *)frame->s8, (void *)frame->ra);
+	printf("\tgp: %#jx\tsp: %#jx\ts8: %#jx\tra: %#jx\n",
+	    (intmax_t)frame->gp, (intmax_t)frame->sp, (intmax_t)frame->s8, (intmax_t)frame->ra);
 
-	printf("\tsr: %p\tmullo: %p\tmulhi: %p\tbadvaddr: %p\n",
-	    (void *)frame->sr, (void *)frame->mullo, (void *)frame->mulhi, (void *)frame->badvaddr);
+	printf("\tsr: %#jx\tmullo: %#jx\tmulhi: %#jx\tbadvaddr: %#jx\n",
+	    (intmax_t)frame->sr, (intmax_t)frame->mullo, (intmax_t)frame->mulhi, (intmax_t)frame->badvaddr);
 
 #ifdef IC_REG
-	printf("\tcause: %p\tpc: %p\tic: %p\n",
-	    (void *)frame->cause, (void *)frame->pc, (void *)frame->ic);
+	printf("\tcause: %#jx\tpc: %#jx\tic: %#jx\n",
+	    (intmax_t)frame->cause, (intmax_t)frame->pc, (intmax_t)frame->ic);
 #else
-	printf("\tcause: %p\tpc: %p\n",
-	    (void *)frame->cause, (void *)frame->pc);
+	printf("\tcause: %#jx\tpc: %#jx\n",
+	    (intmax_t)frame->cause, (intmax_t)frame->pc);
 #endif
 }
 
@@ -1252,7 +1256,7 @@ get_mapping_info(vm_offset_t va, pd_entry_t **pdepp, pt_entry_t **ptepp)
 	pd_entry_t *pdep;
 	struct proc *p = curproc;
 
-	pdep = (&(p->p_vmspace->vm_pmap.pm_segtab[va >> SEGSHIFT]));
+	pdep = (&(p->p_vmspace->vm_pmap.pm_segtab[(va >> SEGSHIFT) & (NPDEPG - 1)]));
 	if (*pdep)
 		ptep = pmap_pte(&p->p_vmspace->vm_pmap, va);
 	else
@@ -1262,6 +1266,50 @@ get_mapping_info(vm_offset_t va, pd_entry_t **pdepp, pt_entry_t **ptepp)
 	*ptepp = ptep;
 }
 
+static void
+log_illegal_instruction(const char *msg, struct trapframe *frame)
+{
+	pt_entry_t *ptep;
+	pd_entry_t *pdep;
+	unsigned int *addr;
+	struct proc *p = curproc;
+	register_t pc;
+
+#ifdef SMP
+	printf("cpuid = %d\n", PCPU_GET(cpuid));
+#endif
+	pc = frame->pc + (DELAYBRANCH(frame->cause) ? 4 : 0);
+	log(LOG_ERR, "%s: pid %d (%s), uid %d: pc %#jx ra %#jx\n",
+	    msg, p->p_pid, p->p_comm,
+	    p->p_ucred ? p->p_ucred->cr_uid : -1,
+	    (intmax_t)pc,
+	    (intmax_t)frame->ra);
+
+	/* log registers in trap frame */
+	log_frame_dump(frame);
+
+	get_mapping_info((vm_offset_t)pc, &pdep, &ptep);
+
+	/*
+	 * Dump a few words around faulting instruction, if the addres is
+	 * valid.
+	 */
+	if (!(pc & 3) &&
+	    useracc((caddr_t)(intptr_t)pc, sizeof(int) * 4, VM_PROT_READ)) {
+		/* dump page table entry for faulting instruction */
+		log(LOG_ERR, "Page table info for pc address %#jx: pde = %p, pte = %#x\n",
+		    (intmax_t)pc, (void *)(intptr_t)*pdep, ptep ? *ptep : 0);
+
+		addr = (unsigned int *)(intptr_t)pc;
+		log(LOG_ERR, "Dumping 4 words starting at pc address %p: \n",
+		    addr);
+		log(LOG_ERR, "%08x %08x %08x %08x\n",
+		    addr[0], addr[1], addr[2], addr[3]);
+	} else {
+		log(LOG_ERR, "pc address %#jx is inaccessible, pde = %p, pte = %#x\n",
+		    (intmax_t)pc, (void *)(intptr_t)*pdep, ptep ? *ptep : 0);
+	}
+}
 
 static void
 log_bad_page_fault(char *msg, struct trapframe *frame, int trap_type)
@@ -1293,12 +1341,12 @@ log_bad_page_fault(char *msg, struct trapframe *frame, int trap_type)
 	}
 
 	pc = frame->pc + (DELAYBRANCH(frame->cause) ? 4 : 0);
-	log(LOG_ERR, "%s: pid %d (%s), uid %d: pc %p got a %s fault at %p\n",
+	log(LOG_ERR, "%s: pid %d (%s), uid %d: pc %#jx got a %s fault at %#jx\n",
 	    msg, p->p_pid, p->p_comm,
 	    p->p_ucred ? p->p_ucred->cr_uid : -1,
-	    (void *)pc,
+	    (intmax_t)pc,
 	    read_or_write,
-	    (void *)frame->badvaddr);
+	    (intmax_t)frame->badvaddr);
 
 	/* log registers in trap frame */
 	log_frame_dump(frame);
@@ -1311,21 +1359,24 @@ log_bad_page_fault(char *msg, struct trapframe *frame, int trap_type)
 	 */
 	if (!(pc & 3) && (pc != frame->badvaddr) &&
 	    (trap_type != T_BUS_ERR_IFETCH) &&
-	    useracc((caddr_t)pc, sizeof(int) * 4, VM_PROT_READ)) {
+	    useracc((caddr_t)(intptr_t)pc, sizeof(int) * 4, VM_PROT_READ)) {
 		/* dump page table entry for faulting instruction */
-		log(LOG_ERR, "Page table info for pc address %p: pde = %p, pte = 0x%lx\n",
-		    (void *)pc, *pdep, ptep ? *ptep : 0);
+		log(LOG_ERR, "Page table info for pc address %#jx: pde = %p, pte = %#x\n",
+		    (intmax_t)pc, (void *)(intptr_t)*pdep, ptep ? *ptep : 0);
 
-		addr = (unsigned int *)pc;
+		addr = (unsigned int *)(intptr_t)pc;
 		log(LOG_ERR, "Dumping 4 words starting at pc address %p: \n",
 		    addr);
 		log(LOG_ERR, "%08x %08x %08x %08x\n",
 		    addr[0], addr[1], addr[2], addr[3]);
 	} else {
-		log(LOG_ERR, "pc address %p is inaccessible, pde = 0x%p, pte = 0x%lx\n",
-		    (void *)pc, *pdep, ptep ? *ptep : 0);
+		log(LOG_ERR, "pc address %#jx is inaccessible, pde = %p, pte = %#x\n",
+		    (intmax_t)pc, (void *)(intptr_t)*pdep, ptep ? *ptep : 0);
 	}
-	/*	panic("Bad trap");*/
+
+	get_mapping_info((vm_offset_t)frame->badvaddr, &pdep, &ptep);
+	log(LOG_ERR, "Page table info for bad address %#jx: pde = %p, pte = %#x\n",
+	    (intmax_t)frame->badvaddr, (void *)(intptr_t)*pdep, ptep ? *ptep : 0);
 }
 
 
@@ -1336,7 +1387,7 @@ static int
 mips_unaligned_load_store(struct trapframe *frame, register_t addr, register_t pc)
 {
 	register_t *reg = (register_t *) frame;
-	u_int32_t inst = *((u_int32_t *) pc);
+	u_int32_t inst = *((u_int32_t *)(intptr_t)pc);
 	u_int32_t value_msb, value;
 	int access_type = 0;
 
@@ -1432,9 +1483,9 @@ emulate_unaligned_access(struct trapframe *frame)
 			else
 				frame->pc += 4;
 
-			log(LOG_INFO, "Unaligned %s: pc=%p, badvaddr=%p\n",
-			    access_name[access_type - 1], (void *)pc,
-			    (void *)frame->badvaddr);
+			log(LOG_INFO, "Unaligned %s: pc=%#jx, badvaddr=%#jx\n",
+			    access_name[access_type - 1], (intmax_t)pc,
+			    (intmax_t)frame->badvaddr);
 		}
 	}
 	return access_type;
diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index 10321dac4bf..cb8a8d90a1a 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -647,7 +647,7 @@ DB_SHOW_COMMAND(pcb, ddb_dump_pcb)
 	DB_PRINT_REG_ARRAY(pcb, pcb_context, PCB_REG_GP);
 	DB_PRINT_REG_ARRAY(pcb, pcb_context, PCB_REG_PC);
 
-	db_printf("PCB onfault = %d\n", pcb->pcb_onfault);
+	db_printf("PCB onfault = %p\n", pcb->pcb_onfault);
 	db_printf("md_saved_intr = 0x%0lx\n", (long)td->td_md.md_saved_intr);
 	db_printf("md_spinlock_count = %d\n", td->td_md.md_spinlock_count);
 

From 25eb6a7d87e4085a8a205dffee2a3730883617b2 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Sat, 17 Apr 2010 11:40:39 +0000
Subject: [PATCH 103/532] Add another ICH7M chipset that works.

MFC after:	1 week
---
 sys/dev/ahci/ahci.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/dev/ahci/ahci.c b/sys/dev/ahci/ahci.c
index 5d399683945..a35fbacdab6 100644
--- a/sys/dev/ahci/ahci.c
+++ b/sys/dev/ahci/ahci.c
@@ -126,6 +126,7 @@ static struct {
 	{0x26838086, 0x00, "Intel ESB2",	0},
 	{0x27c18086, 0x00, "Intel ICH7",	0},
 	{0x27c38086, 0x00, "Intel ICH7",	0},
+	{0x27c48086, 0x00, "Intel ICH7M",	0},
 	{0x27c58086, 0x00, "Intel ICH7M",	0},
 	{0x27c68086, 0x00, "Intel ICH7M",	0},
 	{0x28218086, 0x00, "Intel ICH8",	0},

From 37f144eb5dcee9f71e4d683031bb555f827dea80 Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Sat, 17 Apr 2010 12:22:44 +0000
Subject: [PATCH 104/532] Fix a bug where SACKs are not sent when they should.
 Move some protection code to INVARIANTS. Cleanups.

MFC after: 3 days.
---
 sys/netinet/sctp_indata.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)

diff --git a/sys/netinet/sctp_indata.c b/sys/netinet/sctp_indata.c
index 53dcf8f2a6f..a7e6c087fb4 100644
--- a/sys/netinet/sctp_indata.c
+++ b/sys/netinet/sctp_indata.c
@@ -1466,7 +1466,7 @@ sctp_process_a_data_chunk(struct sctp_tcb *stcb, struct sctp_association *asoc,
 		asoc->send_sack = 1;
 	}
 	protocol_id = ch->dp.protocol_id;
-	ordered = ((ch->ch.chunk_flags & SCTP_DATA_UNORDERED) == 0);
+	ordered = ((chunk_flags & SCTP_DATA_UNORDERED) == 0);
 	if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_MAP_LOGGING_ENABLE) {
 		sctp_log_map(tsn, asoc->cumulative_tsn, asoc->highest_tsn_inside_map, SCTP_MAP_TSN_ENTERS);
 	}
@@ -2300,7 +2300,12 @@ sctp_slide_mapping_arrays(struct sctp_tcb *stcb)
 	if ((asoc->cumulative_tsn == highest_tsn) && (at >= 8)) {
 		/* The complete array was completed by a single FR */
 		/* highest becomes the cum-ack */
-		int clr, i;
+		int clr;
+
+#ifdef INVARIANTS
+		unsigned int i;
+
+#endif
 
 		/* clear the array */
 		clr = ((at + 7) >> 3);
@@ -2309,12 +2314,14 @@ sctp_slide_mapping_arrays(struct sctp_tcb *stcb)
 		}
 		memset(asoc->mapping_array, 0, clr);
 		memset(asoc->nr_mapping_array, 0, clr);
+#ifdef INVARIANTS
 		for (i = 0; i < asoc->mapping_array_size; i++) {
 			if ((asoc->mapping_array[i]) || (asoc->nr_mapping_array[i])) {
 				printf("Error Mapping array's not clean at clear\n");
 				sctp_print_mapping_array(asoc);
 			}
 		}
+#endif
 		asoc->mapping_array_base_tsn = asoc->cumulative_tsn + 1;
 		asoc->highest_tsn_inside_nr_map = asoc->highest_tsn_inside_map = asoc->cumulative_tsn;
 	} else if (at >= 8) {
@@ -2371,10 +2378,8 @@ sctp_slide_mapping_arrays(struct sctp_tcb *stcb)
 			int ii;
 
 			for (ii = 0; ii < distance; ii++) {
-				asoc->mapping_array[ii] =
-				    asoc->mapping_array[slide_from + ii];
-				asoc->nr_mapping_array[ii] =
-				    asoc->nr_mapping_array[slide_from + ii];
+				asoc->mapping_array[ii] = asoc->mapping_array[slide_from + ii];
+				asoc->nr_mapping_array[ii] = asoc->nr_mapping_array[slide_from + ii];
 
 			}
 			for (ii = distance; ii < asoc->mapping_array_size; ii++) {
@@ -2808,8 +2813,8 @@ sctp_process_data(struct mbuf **mm, int iphlen, int *offset, int length,
 		stcb->asoc.send_sack = 1;
 	}
 	/* Start a sack timer or QUEUE a SACK for sending */
-	if ((stcb->asoc.cumulative_tsn == stcb->asoc.highest_tsn_inside_map) &&
-	    (stcb->asoc.mapping_array[0] != 0xff)) {
+	if ((stcb->asoc.cumulative_tsn == stcb->asoc.highest_tsn_inside_nr_map) &&
+	    (stcb->asoc.nr_mapping_array[0] != 0xff)) {
 		if ((stcb->asoc.data_pkts_seen >= stcb->asoc.sack_freq) ||
 		    (stcb->asoc.delayed_ack == 0) ||
 		    (stcb->asoc.numduptsns) ||

From 8eac1f947715c47f877e577515a3289aba705513 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 17 Apr 2010 14:35:46 +0000
Subject: [PATCH 105/532] sh: On startup of the shell, use PWD from the
 environment if it is valid. Unset PWD if it is incorrect and no value for it
 can be determined. This preserves the logical current directory across shell
 invocations.

Example (assuming /home is a symlink):
$ cd
$ pwd
/home/foo
$ sh
$ pwd
/home/foo

Formerly the second pwd would show the physical path (symlinks resolved).
---
 bin/sh/cd.c                               | 27 ++++++++++++++++++-----
 bin/sh/cd.h                               |  2 +-
 bin/sh/main.c                             |  5 +----
 tools/regression/bin/sh/parameters/pwd2.0 | 24 ++++++++++++++++++++
 4 files changed, 48 insertions(+), 10 deletions(-)
 create mode 100644 tools/regression/bin/sh/parameters/pwd2.0

diff --git a/bin/sh/cd.c b/bin/sh/cd.c
index 40801ba2c97..aee06e8fb00 100644
--- a/bin/sh/cd.c
+++ b/bin/sh/cd.c
@@ -70,6 +70,7 @@ STATIC int docd(char *, int, int);
 STATIC char *getcomponent(void);
 STATIC char *findcwd(char *);
 STATIC void updatepwd(char *);
+STATIC char *getpwd(void);
 STATIC char *getpwd2(void);
 
 STATIC char *curdir = NULL;	/* current working directory */
@@ -351,7 +352,7 @@ pwdcmd(int argc, char **argv)
 /*
  * Get the current directory and cache the result in curdir.
  */
-char *
+STATIC char *
 getpwd(void)
 {
 	char *p;
@@ -374,7 +375,6 @@ getpwd(void)
 STATIC char *
 getpwd2(void)
 {
-	struct stat stdot, stpwd;
 	char *pwd;
 	int i;
 
@@ -387,12 +387,29 @@ getpwd2(void)
 			break;
 	}
 
-	pwd = getenv("PWD");
+	return NULL;
+}
+
+/*
+ * Initialize PWD in a new shell.
+ * If the shell is interactive, we need to warn if this fails.
+ */
+void
+pwd_init(int warn)
+{
+	char *pwd;
+	struct stat stdot, stpwd;
+
+	pwd = lookupvar("PWD");
 	if (pwd && *pwd == '/' && stat(".", &stdot) != -1 &&
 	    stat(pwd, &stpwd) != -1 &&
 	    stdot.st_dev == stpwd.st_dev &&
 	    stdot.st_ino == stpwd.st_ino) {
-		return pwd;
+		if (curdir)
+			ckfree(curdir);
+		curdir = savestr(pwd);
 	}
-	return NULL;
+	if (getpwd() == NULL && warn)
+		out2fmt_flush("sh: cannot determine working directory\n");
+	setvar("PWD", curdir, VEXPORT);
 }
diff --git a/bin/sh/cd.h b/bin/sh/cd.h
index 0a2d48999e7..f88ce26ad4c 100644
--- a/bin/sh/cd.h
+++ b/bin/sh/cd.h
@@ -29,6 +29,6 @@
  * $FreeBSD$
  */
 
-char	*getpwd(void);
+void	 pwd_init(int);
 int	 cdcmd (int, char **);
 int	 pwdcmd(int, char **);
diff --git a/bin/sh/main.c b/bin/sh/main.c
index 50a88132ab3..ecbb12dc640 100644
--- a/bin/sh/main.c
+++ b/bin/sh/main.c
@@ -153,10 +153,7 @@ main(int argc, char *argv[])
 	init();
 	setstackmark(&smark);
 	procargs(argc, argv);
-	if (getpwd() == NULL && iflag)
-		out2fmt_flush("sh: cannot determine working directory\n");
-	if (getpwd() != NULL)
-		setvar ("PWD", getpwd(), VEXPORT);
+	pwd_init(iflag);
 	if (iflag)
 		chkmail(1);
 	if (argv[0] && argv[0][0] == '-') {
diff --git a/tools/regression/bin/sh/parameters/pwd2.0 b/tools/regression/bin/sh/parameters/pwd2.0
new file mode 100644
index 00000000000..29b5531aaa8
--- /dev/null
+++ b/tools/regression/bin/sh/parameters/pwd2.0
@@ -0,0 +1,24 @@
+# $FreeBSD$
+# Check that PWD is exported and accepted from the environment.
+set -e
+
+T=$(mktemp -d ${TMPDIR:-/tmp}/sh-test.XXXXXX)
+trap 'rm -rf $T' 0
+cd -P $T
+TP=$(pwd)
+mkdir test1
+ln -s test1 link
+cd link
+[ "$PWD" = "$TP/link" ]
+[ "$(pwd)" = "$TP/link" ]
+[ "$(pwd -P)" = "$TP/test1" ]
+[ "$(sh -c pwd)" = "$TP/link" ]
+[ "$(sh -c pwd\ -P)" = "$TP/test1" ]
+cd ..
+[ "$(pwd)" = "$TP" ]
+cd -P link
+[ "$PWD" = "$TP/test1" ]
+[ "$(pwd)" = "$TP/test1" ]
+[ "$(pwd -P)" = "$TP/test1" ]
+[ "$(sh -c pwd)" = "$TP/test1" ]
+[ "$(sh -c pwd\ -P)" = "$TP/test1" ]

From 702f62fb48f1cd993cfdfb7d84e17e3874f11635 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 17 Apr 2010 15:52:50 +0000
Subject: [PATCH 106/532] getcwd(3): Clarify that EACCES may or may not be
 checked.

POSIX permits but does not require checking access on the current and parent
directories.

Because various programs do not like it if getcwd(3) fails, it seems best
to avoid checking access as much as possible. There are various reports in
GNATS about this (search for getcwd).

Our getcwd(3) implementation first queries the kernel for the pathname
directly, which does not check any permissions but sometimes fails, and then
falls back to reading all parent directories for the names.

PR:		standards/44425
MFC after:	2 weeks
---
 lib/libc/gen/getcwd.3 | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/lib/libc/gen/getcwd.3 b/lib/libc/gen/getcwd.3
index 12a2bd664af..88291c3e2ee 100644
--- a/lib/libc/gen/getcwd.3
+++ b/lib/libc/gen/getcwd.3
@@ -28,7 +28,7 @@
 .\"     @(#)getcwd.3	8.2 (Berkeley) 12/11/93
 .\" $FreeBSD$
 .\"
-.Dd November 24, 1997
+.Dd April 17, 2010
 .Dt GETCWD 3
 .Os
 .Sh NAME
@@ -108,8 +108,6 @@ The
 function
 will fail if:
 .Bl -tag -width Er
-.It Bq Er EACCES
-Read or search permission was denied for a component of the pathname.
 .It Bq Er EINVAL
 The
 .Fa size
@@ -124,6 +122,16 @@ The
 argument is greater than zero but smaller than the length of the pathname
 plus 1.
 .El
+.Pp
+The
+.Fn getcwd
+function
+may fail if:
+.Bl -tag -width Er
+.It Bq Er EACCES
+Read or search permission was denied for a component of the pathname.
+This is only checked in limited cases, depending on implementation details.
+.El
 .Sh SEE ALSO
 .Xr chdir 2 ,
 .Xr fchdir 2 ,

From 0b6ace474376f863982bec0c04b1498ff74b0c1d Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sat, 17 Apr 2010 17:02:17 +0000
Subject: [PATCH 107/532] Setting PG_REFERENCED on the requested page in
 swap_pager_getpages() is either redundant or harmful, depending on the
 caller.  For example, when called by vm_fault(), it is redundant.  However,
 when called by vm_thread_swapin(), it is harmful.  Specifically, if the
 thread is later swapped out, having PG_REFERENCED set on its stack pages
 leads the page daemon to reactivate these stack pages and delay their
 reclamation.

Reviewed by:	kib
MFC after:	3 weeks
---
 sys/vm/swap_pager.c | 6 +-----
 1 file changed, 1 insertion(+), 5 deletions(-)

diff --git a/sys/vm/swap_pager.c b/sys/vm/swap_pager.c
index ef64b31ae09..f1d89d5e030 100644
--- a/sys/vm/swap_pager.c
+++ b/sys/vm/swap_pager.c
@@ -1101,8 +1101,7 @@ swap_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 	 * happen.  Note that blk, iblk & jblk can be SWAPBLK_NONE, but the 
 	 * loops are set up such that the case(s) are handled implicitly.
 	 *
-	 * The swp_*() calls must be made at splvm().  vm_page_free() does
-	 * not need to be, but it will go a little faster if it is.
+	 * The swp_*() calls must be made with the object locked.
 	 */
 	blk = swp_pager_meta_ctl(mreq->object, mreq->pindex, 0);
 
@@ -1212,9 +1211,6 @@ swap_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 	VM_OBJECT_LOCK(object);
 	while ((mreq->oflags & VPO_SWAPINPROG) != 0) {
 		mreq->oflags |= VPO_WANTED;
-		vm_page_lock_queues();
-		vm_page_flag_set(mreq, PG_REFERENCED);
-		vm_page_unlock_queues();
 		PCPU_INC(cnt.v_intrans);
 		if (msleep(mreq, VM_OBJECT_MTX(object), PSWP, "swread", hz*20)) {
 			printf(

From d02ae91a4c13ee1cd599ccd96566b84259ec15c2 Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:13:52 +0000
Subject: [PATCH 108/532] Fix comment about ipw_assoc and remove some
 whitespaces; no functional changes.

Approved by:	rpaulo (mentor)
---
 sys/dev/ipw/if_ipw.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index 0b217583f24..dc2a61f009d 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -888,10 +888,10 @@ ipw_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg)
 			/*
 			 * XXX when joining an ibss network we are called
 			 * with a SCAN -> RUN transition on scan complete.
-			 * Use that to call ipw_auth_and_assoc.  On completing
-			 * the join we are then called again with an
-			 * AUTH -> RUN transition and we want to do nothing.
-			 * This is all totally bogus and needs to be redone.
+			 * Use that to call ipw_assoc.  On completing the
+			 * join we are then called again with an AUTH -> RUN
+			 * transition and we want to do nothing.  This is
+			 * all totally bogus and needs to be redone.
 			 */
 			if (ostate == IEEE80211_S_SCAN)
 				ipw_assoc(ic, vap);
@@ -909,7 +909,7 @@ ipw_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg)
 
 	case IEEE80211_S_ASSOC:
 		/*
-		 * If we are not transitioning from AUTH the resend the
+		 * If we are not transitioning from AUTH then resend the
 		 * association request.
 		 */
 		if (ostate != IEEE80211_S_AUTH)
@@ -1070,7 +1070,7 @@ ipw_rx_newstate_intr(struct ipw_softc *sc, struct ipw_soft_buf *sbuf)
 
 	case IPW_STATE_DISABLED:
 		/* XXX? is this right? */
-		sc->flags &= ~(IPW_FLAG_HACK | IPW_FLAG_SCANNING | 
+		sc->flags &= ~(IPW_FLAG_HACK | IPW_FLAG_SCANNING |
 		    IPW_FLAG_ASSOCIATING | IPW_FLAG_ASSOCIATED);
 		DPRINTFN(2, ("Firmware disabled (%s flags 0x%x)\n",
 			IEEESTATE(vap), sc->flags));

From 892d2e9d9bb81d56487590e7780a4229df8bbaef Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:14:49 +0000
Subject: [PATCH 109/532] Pass correct RSSI to ieee80211_input*().

Approved by:	rpaulo (mentor)
---
 sys/dev/ipw/if_ipw.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index dc2a61f009d..3a8d85d7ca6 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -1234,10 +1234,10 @@ ipw_rx_data_intr(struct ipw_softc *sc, struct ipw_status *status,
 	IPW_UNLOCK(sc);
 	ni = ieee80211_find_rxnode(ic, mtod(m, struct ieee80211_frame_min *));
 	if (ni != NULL) {
-		(void) ieee80211_input(ni, m, rssi, nf);
+		(void) ieee80211_input(ni, m, rssi - nf, nf);
 		ieee80211_free_node(ni);
 	} else
-		(void) ieee80211_input_all(ic, m, rssi, nf);
+		(void) ieee80211_input_all(ic, m, rssi - nf, nf);
 	IPW_LOCK(sc);
 
 	bus_dmamap_sync(sc->rbd_dmat, sc->rbd_map, BUS_DMASYNC_PREWRITE);

From 557c25b48fe48116f27951dc263cd4fbf133b388 Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:16:14 +0000
Subject: [PATCH 110/532] - Make ipw usable again by moving directly into ASSOC
 state. - No need to manually switch to RUN state, assoc response takes care  
 of that.

Approved by:	rpaulo (mentor)
---
 sys/dev/ipw/if_ipw.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index 3a8d85d7ca6..2e62949a4bb 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -904,6 +904,13 @@ ipw_newstate(struct ieee80211vap *vap, enum ieee80211_state nstate, int arg)
 		break;
 
 	case IEEE80211_S_AUTH:
+		/*
+		 * Move to ASSOC state after the ipw_assoc() call.  Firmware
+		 * takes care of authentication, after the call we'll receive
+		 * only an assoc response which would otherwise be discared
+		 * if we are still in AUTH state.
+		 */
+		nstate = IEEE80211_S_ASSOC;
 		ipw_assoc(ic, vap);
 		break;
 
@@ -1021,7 +1028,6 @@ ipw_rx_newstate_intr(struct ipw_softc *sc, struct ipw_soft_buf *sbuf)
 		}
 		sc->flags &= ~IPW_FLAG_ASSOCIATING;
 		sc->flags |= IPW_FLAG_ASSOCIATED;
-		ieee80211_new_state(vap, IEEE80211_S_RUN, -1);
 		break;
 
 	case IPW_STATE_SCANNING:

From 6e4c30c9cdebd9bd441d57a8884babcd40e2df79 Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:17:25 +0000
Subject: [PATCH 111/532] Use iv_appie_wpa, with this commit WPA works again.

Approved by:	rpaulo (mentor)
---
 sys/dev/ipw/if_ipw.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index 2e62949a4bb..9a17e6173c8 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -2266,8 +2266,8 @@ ipw_assoc(struct ieee80211com *ic, struct ieee80211vap *vap)
 	if (error != 0)
 		goto done;
 
-	if (vap->iv_appie_assocreq != NULL) {
-		struct ieee80211_appie *ie = vap->iv_appie_assocreq;
+	if (vap->iv_appie_wpa != NULL) {
+		struct ieee80211_appie *ie = vap->iv_appie_wpa;
 		error = ipw_setwpaie(sc, ie->ie_data, ie->ie_len);
 		if (error != 0)
 			goto done;

From 82ff07c047997e086c6ad79595d97b31dc555813 Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:18:46 +0000
Subject: [PATCH 112/532] Remove IPW_LOCK_DECL and fix various LORs.

Approved by:	rpaulo (mentor)
---
 sys/dev/ipw/if_ipw.c    | 27 ++++++++++++++-------------
 sys/dev/ipw/if_ipwvar.h | 11 ++---------
 2 files changed, 16 insertions(+), 22 deletions(-)

diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index 9a17e6173c8..3315d654b1c 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -1040,8 +1040,10 @@ ipw_rx_newstate_intr(struct ipw_softc *sc, struct ipw_soft_buf *sbuf)
 		 * we checked the 802.11 layer state.
 		 */
 		if (sc->flags & IPW_FLAG_ASSOCIATED) {
+			IPW_UNLOCK(sc);
 			/* XXX probably need to issue disassoc to fw */
 			ieee80211_beacon_miss(ic);
+			IPW_LOCK(sc);
 		}
 		break;
 
@@ -1060,7 +1062,9 @@ ipw_rx_newstate_intr(struct ipw_softc *sc, struct ipw_soft_buf *sbuf)
 			break;
 		}
 		if (sc->flags & IPW_FLAG_SCANNING) {
+			IPW_UNLOCK(sc);
 			ieee80211_scan_done(vap);
+			IPW_LOCK(sc);
 			sc->flags &= ~IPW_FLAG_SCANNING;
 			sc->sc_scan_timer = 0;
 		}
@@ -1070,8 +1074,11 @@ ipw_rx_newstate_intr(struct ipw_softc *sc, struct ipw_soft_buf *sbuf)
 		DPRINTFN(2, ("Association lost (%s flags 0x%x)\n",
 			IEEESTATE(vap), sc->flags));
 		sc->flags &= ~(IPW_FLAG_ASSOCIATING | IPW_FLAG_ASSOCIATED);
-		if (vap->iv_state == IEEE80211_S_RUN)
+		if (vap->iv_state == IEEE80211_S_RUN) {
+			IPW_UNLOCK(sc);
 			ieee80211_new_state(vap, IEEE80211_S_SCAN, -1);
+			IPW_LOCK(sc);
+		}
 		break;
 
 	case IPW_STATE_DISABLED:
@@ -1170,7 +1177,6 @@ ipw_rx_data_intr(struct ipw_softc *sc, struct ipw_status *status,
 	bus_addr_t physaddr;
 	int error;
 	int8_t rssi, nf;
-	IPW_LOCK_DECL;
 
 	DPRINTFN(5, ("received frame len=%u, rssi=%u\n", le32toh(status->len),
 	    status->rssi));
@@ -1384,8 +1390,11 @@ ipw_fatal_error_intr(struct ipw_softc *sc)
 	struct ieee80211vap *vap = TAILQ_FIRST(&ic->ic_vaps);
 
 	device_printf(sc->sc_dev, "firmware error\n");
-	if (vap != NULL)
+	if (vap != NULL) {
+		IPW_UNLOCK(sc);
 		ieee80211_cancel_scan(vap);
+		IPW_LOCK(sc);
+	}
 	ieee80211_runtask(ic, &sc->sc_init_task);
 }
 
@@ -1394,7 +1403,6 @@ ipw_intr(void *arg)
 {
 	struct ipw_softc *sc = arg;
 	uint32_t r;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 
@@ -1724,7 +1732,6 @@ static void
 ipw_start(struct ifnet *ifp)
 {
 	struct ipw_softc *sc = ifp->if_softc;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	ipw_start_locked(ifp);
@@ -1781,7 +1788,9 @@ ipw_watchdog(void *arg)
 			DPRINTFN(3, ("Scan timeout\n"));
 			/* End the scan */
 			if (sc->flags & IPW_FLAG_SCANNING) {
+				IPW_UNLOCK(sc);
 				ieee80211_scan_done(TAILQ_FIRST(&ic->ic_vaps));
+				IPW_LOCK(sc);
 				sc->flags &= ~IPW_FLAG_SCANNING;
 			}
 		}
@@ -1797,7 +1806,6 @@ ipw_ioctl(struct ifnet *ifp, u_long cmd, caddr_t data)
 	struct ieee80211com *ic = ifp->if_l2com;
 	struct ifreq *ifr = (struct ifreq *) data;
 	int error = 0, startall = 0;
-	IPW_LOCK_DECL;
 
 	switch (cmd) {
 	case SIOCSIFFLAGS:
@@ -2207,7 +2215,6 @@ ipw_assoc(struct ieee80211com *ic, struct ieee80211vap *vap)
 	struct ipw_security security;
 	uint32_t data;
 	int error;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	error = ipw_disable(sc);
@@ -2297,7 +2304,6 @@ ipw_disassoc(struct ieee80211com *ic, struct ieee80211vap *vap)
 	struct ifnet *ifp = vap->iv_ic->ic_ifp;
 	struct ieee80211_node *ni = vap->iv_bss;
 	struct ipw_softc *sc = ifp->if_softc;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	DPRINTF(("Disassociate from %6D\n", ni->ni_bssid, ":"));
@@ -2333,7 +2339,6 @@ ipw_init(void *priv)
 	struct ipw_softc *sc = priv;
 	struct ifnet *ifp = sc->sc_ifp;
 	struct ieee80211com *ic = ifp->if_l2com;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	ipw_init_locked(sc);
@@ -2540,7 +2545,6 @@ static void
 ipw_stop(void *priv)
 {
 	struct ipw_softc *sc = priv;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	ipw_stop_locked(sc);
@@ -2667,7 +2671,6 @@ ipw_scan_start(struct ieee80211com *ic)
 {
 	struct ifnet *ifp = ic->ic_ifp;
 	struct ipw_softc *sc = ifp->if_softc;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	ipw_scan(sc);
@@ -2679,7 +2682,6 @@ ipw_set_channel(struct ieee80211com *ic)
 {
 	struct ifnet *ifp = ic->ic_ifp;
 	struct ipw_softc *sc = ifp->if_softc;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	if (ic->ic_opmode == IEEE80211_M_MONITOR) {
@@ -2707,7 +2709,6 @@ ipw_scan_end(struct ieee80211com *ic)
 {
 	struct ifnet *ifp = ic->ic_ifp;
 	struct ipw_softc *sc = ifp->if_softc;
-	IPW_LOCK_DECL;
 
 	IPW_LOCK(sc);
 	sc->flags &= ~IPW_FLAG_SCANNING;
diff --git a/sys/dev/ipw/if_ipwvar.h b/sys/dev/ipw/if_ipwvar.h
index 8d9e049ccfb..89702d0c634 100644
--- a/sys/dev/ipw/if_ipwvar.h
+++ b/sys/dev/ipw/if_ipwvar.h
@@ -164,13 +164,6 @@ struct ipw_softc {
  * NB.: This models the only instance of async locking in ipw_init_locked
  *	and must be kept in sync.
  */
-#define	IPW_LOCK_DECL	int	__waslocked = 0
-#define IPW_LOCK(sc)	do {				\
-	if (!(__waslocked = mtx_owned(&(sc)->sc_mtx)))	\
-		mtx_lock(&sc->sc_mtx);			\
-} while (0)
-#define IPW_UNLOCK(sc)	do {				\
-	if (!__waslocked)				\
-		mtx_unlock(&sc->sc_mtx);		\
-} while (0)
+#define IPW_LOCK(sc)		mtx_lock(&sc->sc_mtx);
+#define IPW_UNLOCK(sc)		mtx_unlock(&sc->sc_mtx);
 #define IPW_LOCK_ASSERT(sc)	mtx_assert(&(sc)->sc_mtx, MA_OWNED)

From aefea7f519b1c2aee4f74acd1ecaf1216956fdf7 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:35:07 +0000
Subject: [PATCH 113/532] In vm_object_backing_scan(), setting PG_REFERENCED on
 a page before sleeping on that page is nonsensical.  Doing so reduces the
 likelihood that the page daemon will reclaim the page before the thread
 waiting in vm_object_backing_scan() is reawakened.  However, it does not
 guarantee that the page is not reclaimed, so vm_object_backing_scan()
 restarts after reawakening.  More importantly, this muddles the meaning of
 PG_REFERENCED.  There is no reason to believe that the caller of
 vm_object_backing_scan() is going to use (i.e., access) the contents of the
 page.  There is especially no reason to believe that an access is more likely
 because vm_object_backing_scan() had to sleep on the page.

Discussed with:	kib
MFC after:	3 weeks
---
 sys/vm/vm_object.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 399cb10a1f9..7e84e58e35f 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1553,9 +1553,6 @@ vm_object_backing_scan(vm_object_t object, int op)
 				}
 			} else if (op & OBSC_COLLAPSE_WAIT) {
 				if ((p->oflags & VPO_BUSY) || p->busy) {
-					vm_page_lock_queues();
-					vm_page_flag_set(p, PG_REFERENCED);
-					vm_page_unlock_queues();
 					VM_OBJECT_UNLOCK(object);
 					p->oflags |= VPO_WANTED;
 					msleep(p, VM_OBJECT_MTX(backing_object),

From 1a5873681676b938759aee294c9a98c8845a593a Mon Sep 17 00:00:00 2001
From: Doug Barton <dougb@FreeBSD.org>
Date: Sat, 17 Apr 2010 18:48:18 +0000
Subject: [PATCH 114/532] In case a user wants to configure only an IPv6
 link-local address add an example that shows how to do it.

---
 etc/defaults/rc.conf     | 1 +
 share/man/man5/rc.conf.5 | 7 +++++++
 2 files changed, 8 insertions(+)

diff --git a/etc/defaults/rc.conf b/etc/defaults/rc.conf
index a5da7c2417e..48233bdb1cf 100644
--- a/etc/defaults/rc.conf
+++ b/etc/defaults/rc.conf
@@ -211,6 +211,7 @@ ifconfig_lo0="inet 127.0.0.1"	# default loopback device configuration.
 #ifconfig_lo0_alias0="inet 127.0.0.254 netmask 0xffffffff" # Sample alias entry.
 #ifconfig_ed0_ipx="ipx 0x00010010"	# Sample IPX address family entry.
 #ifconfig_ed0_ipv6="RTADV" 	# Sample IPv6 entry for RA/rtsol(8)
+#ifconfig_ed0_ipv6="inet6 auto_linklocal" 	# To configure only link-local
 #ifconfig_ed0_ipv6="inet6 2001:db8:1::1 prefixlen 64" # Sample IPv6 addr entry
 #ifconfig_ed0_alias0="inet6 2001:db8:2::1 prefixlen 64" # Sample IPv6 alias
 #ifconfig_fxp0_name="net0"	# Change interface name from fxp0 to net0.
diff --git a/share/man/man5/rc.conf.5 b/share/man/man5/rc.conf.5
index c0b358d4499..e93205be539 100644
--- a/share/man/man5/rc.conf.5
+++ b/share/man/man5/rc.conf.5
@@ -1303,6 +1303,13 @@ interface:
 ifconfig_ed0_ipv6="RTADV"
 .Ed
 .Pp
+To configure only a link-local address on the
+.Dq Li ed0
+interface:
+.Bd -literal
+ifconfig_ed0_ipv6="inet6 auto_linklocal"
+.Ed
+.Pp
 To disable RA the
 .Dq Li NORTADV
 option is available, although not required if manual

From b11b56b55b0fef3215817f1d4b99661809e9afea Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sat, 17 Apr 2010 21:14:37 +0000
Subject: [PATCH 115/532] In vm_object_madvise() setting PG_REFERENCED on a
 page before sleeping on that page only makes sense if the advice is
 MADV_WILLNEED.  In that case, the intention is to activate the page, so
 discouraging the page daemon from reclaiming the page makes sense.  In
 contrast, in the other cases, MADV_DONTNEED and MADV_FREE, it makes no sense
 whatsoever to discourage the page daemon from reclaiming the page by setting
 PG_REFERENCED.

Wrap a nearby line.

Discussed with:	kib
MFC after:	3 weeks
---
 sys/vm/vm_object.c | 11 +++++++++--
 1 file changed, 9 insertions(+), 2 deletions(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 7e84e58e35f..1f9ba6d4956 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1205,12 +1205,19 @@ shadowlookup:
 			goto unlock_tobject;
 		}
 		if ((m->oflags & VPO_BUSY) || m->busy) {
-			vm_page_flag_set(m, PG_REFERENCED);
+			if (advise == MADV_WILLNEED)
+				/*
+				 * Reference the page before unlocking and
+				 * sleeping so that the page daemon is less
+				 * likely to reclaim it. 
+				 */
+				vm_page_flag_set(m, PG_REFERENCED);
 			vm_page_unlock_queues();
 			if (object != tobject)
 				VM_OBJECT_UNLOCK(object);
 			m->oflags |= VPO_WANTED;
-			msleep(m, VM_OBJECT_MTX(tobject), PDROP | PVM, "madvpo", 0);
+			msleep(m, VM_OBJECT_MTX(tobject), PDROP | PVM, "madvpo",
+			    0);
 			VM_OBJECT_LOCK(object);
   			goto relookup;
 		}

From 9eb448a7e634f3edde2ca41327e4a7a0dcac564e Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Sat, 17 Apr 2010 21:31:42 +0000
Subject: [PATCH 116/532] Use ubthidhci_enable="NO" to avoid the bootup
 warning.

Submitted by: 	Jilles Tjoelker <jilles@stack.nl>
MFC after:	3 days
---
 etc/defaults/rc.conf | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/etc/defaults/rc.conf b/etc/defaults/rc.conf
index 48233bdb1cf..ab50920dccb 100644
--- a/etc/defaults/rc.conf
+++ b/etc/defaults/rc.conf
@@ -437,7 +437,7 @@ rfcomm_pppd_server_two_channel="3"	# Override local channel for 'two'
 #rfcomm_pppd_server_two_register_sp="NO"	# Override SP and DUN register
 #rfcomm_pppd_server_two_register_dun="NO"	# for 'two'
 
-#ubthidhci_enable="YES"		# Switch an USB BT controller present on
+ubthidhci_enable="NO"		# Switch an USB BT controller present on
 #ubthidhci_busnum="3"		# bus 3 and addr 2 from HID mode to HCI mode.
 #ubthidhci_addr="2"		# Check usbconfig list to find the correct
 				# numbers for your system.

From 5d66b54e27eb01b2bbea010f559c4c985c67886a Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 17 Apr 2010 22:39:53 +0000
Subject: [PATCH 117/532] ln: Refuse deleting a directory entry by hardlinking
 it to itself.

Two pathnames refer to the same directory entry iff the directories match
and the final components' names match.

Example: (assuming file1 is an existing file)
  ln -f file1 file1
This now fails while leaving file1 intact. It used to delete file1 and then
complain it cannot be linked because it is gone.

With -i, this error is detected before the question is asked.

MFC after:	2 weeks
---
 bin/ln/ln.c | 61 +++++++++++++++++++++++++++++++++++++++++++++++++++--
 1 file changed, 59 insertions(+), 2 deletions(-)

diff --git a/bin/ln/ln.c b/bin/ln/ln.c
index e946646775e..d77939291f7 100644
--- a/bin/ln/ln.c
+++ b/bin/ln/ln.c
@@ -172,6 +172,52 @@ main(int argc, char *argv[])
 	exit(exitval);
 }
 
+/*
+ * Two pathnames refer to the same directory entry if the directories match
+ * and the final components' names match.
+ */
+static int
+samedirent(const char *path1, const char *path2)
+{
+	const char *file1, *file2;
+	char pathbuf[PATH_MAX];
+	struct stat sb1, sb2;
+
+	if (strcmp(path1, path2) == 0)
+		return 1;
+	file1 = strrchr(path1, '/');
+	if (file1 != NULL)
+		file1++;
+	else
+		file1 = path1;
+	file2 = strrchr(path2, '/');
+	if (file2 != NULL)
+		file2++;
+	else
+		file2 = path2;
+	if (strcmp(file1, file2) != 0)
+		return 0;
+	if (file1 - path1 >= PATH_MAX || file2 - path2 >= PATH_MAX)
+		return 0;
+	if (file1 == path1)
+		memcpy(pathbuf, ".", 2);
+	else {
+		memcpy(pathbuf, path1, file1 - path1);
+		pathbuf[file1 - path1] = '\0';
+	}
+	if (stat(pathbuf, &sb1) != 0)
+		return 0;
+	if (file2 == path2)
+		memcpy(pathbuf, ".", 2);
+	else {
+		memcpy(pathbuf, path2, file2 - path2);
+		pathbuf[file2 - path2] = '\0';
+	}
+	if (stat(pathbuf, &sb2) != 0)
+		return 0;
+	return sb1.st_dev == sb2.st_dev && sb1.st_ino == sb2.st_ino;
+}
+
 int
 linkit(const char *source, const char *target, int isdir)
 {
@@ -215,7 +261,6 @@ linkit(const char *source, const char *target, int isdir)
 		target = path;
 	}
 
-	exists = !lstat(target, &sb);
 	/*
 	 * If the link source doesn't exist, and a symbolic link was
 	 * requested, and -w was specified, give a warning.
@@ -242,8 +287,20 @@ linkit(const char *source, const char *target, int isdir)
 				warn("warning: %s", source);
 		}
 	}
+
 	/*
-	 * If the file exists, then unlink it forcibly if -f was specified
+	 * If the file exists, first check it is not the same directory entry.
+	 */
+	exists = !lstat(target, &sb);
+	if (exists) {
+		if (!sflag && samedirent(source, target)) {
+			warnx("%s and %s are the same directory entry",
+			    source, target);
+			return (1);
+		}
+	}
+	/*
+	 * Then unlink it forcibly if -f was specified
 	 * and interactively if -i was specified.
 	 */
 	if (fflag && exists) {

From 5195ca23073524c7c25453d289c451040c15af41 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:20:33 +0000
Subject: [PATCH 118/532] Set ARC_L2_WRITING on L2ARC header creation.

Obtained from:	OpenSolaris
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
index ca8ffb1ea13..e7043c99160 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
@@ -4562,6 +4562,7 @@ l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
 			hdrl2->b_dev = dev;
 			hdrl2->b_daddr = dev->l2ad_hand;
 
+			ab->b_flags |= ARC_L2_WRITING;
 			ab->b_l2hdr = hdrl2;
 			list_insert_head(dev->l2ad_buflist, ab);
 			buf_data = ab->b_buf->b_data;

From 57a81a8bbc37dadcc96dd6ed12c802e7ef99a7c8 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:21:52 +0000
Subject: [PATCH 119/532] Remove racy assertion.

Obtained from:	OpenSolaris
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
index e7043c99160..4cd1cff2ecf 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
@@ -2731,7 +2731,6 @@ arc_read(zio_t *pio, spa_t *spa, blkptr_t *bp, arc_buf_t *pbuf,
     uint32_t *arc_flags, const zbookmark_t *zb)
 {
 	int err;
-	arc_buf_hdr_t *hdr = pbuf->b_hdr;
 
 	ASSERT(!refcount_is_zero(&pbuf->b_hdr->b_refcnt));
 	ASSERT3U((char *)bp - (char *)pbuf->b_data, <, pbuf->b_hdr->b_size);
@@ -2739,8 +2738,6 @@ arc_read(zio_t *pio, spa_t *spa, blkptr_t *bp, arc_buf_t *pbuf,
 
 	err = arc_read_nolock(pio, spa, bp, done, private, priority,
 	    zio_flags, arc_flags, zb);
-
-	ASSERT3P(hdr, ==, pbuf->b_hdr);
 	rw_exit(&pbuf->b_lock);
 	return (err);
 }

From ad3cb80827341a867218b67720a64ca41f5135bb Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:25:40 +0000
Subject: [PATCH 120/532] Extend locks scope to match OpenSolaris.

---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
index 4cd1cff2ecf..7d218a44dc6 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
@@ -1106,8 +1106,6 @@ add_reference(arc_buf_hdr_t *ab, kmutex_t *hash_lock, void *tag)
 		mutex_enter(lock);
 		ASSERT(list_link_active(&ab->b_arc_node));
 		list_remove(list, ab);
-		mutex_exit(lock);
-
 		if (GHOST_STATE(ab->b_state)) {
 			ASSERT3U(ab->b_datacnt, ==, 0);
 			ASSERT3P(ab->b_buf, ==, NULL);
@@ -1116,6 +1114,7 @@ add_reference(arc_buf_hdr_t *ab, kmutex_t *hash_lock, void *tag)
 		ASSERT(delta > 0);
 		ASSERT3U(*size, >=, delta);
 		atomic_add_64(size, -delta);
+		mutex_exit(lock);
 		/* remove the prefetch flag if we get a reference */
 		if (ab->b_flags & ARC_PREFETCH)
 			ab->b_flags &= ~ARC_PREFETCH;
@@ -1138,15 +1137,13 @@ remove_reference(arc_buf_hdr_t *ab, kmutex_t *hash_lock, void *tag)
 		kmutex_t *lock;
 
 		get_buf_info(ab, state, &list, &lock);
-
 		ASSERT(!MUTEX_HELD(lock));
 		mutex_enter(lock);
 		ASSERT(!list_link_active(&ab->b_arc_node));
 		list_insert_head(list, ab);
-		mutex_exit(lock);
-
 		ASSERT(ab->b_datacnt > 0);
 		atomic_add_64(size, ab->b_size * ab->b_datacnt);
+		mutex_exit(lock);
 	}
 	return (cnt);
 }

From eb998be67d2bd8509b9c37c66c9a18f81253deaf Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:27:07 +0000
Subject: [PATCH 121/532] Add missing list and lock destruction.

---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
index 7d218a44dc6..740a0a57f3e 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
@@ -3850,12 +3850,14 @@ arc_fini(void)
 		list_destroy(&arc_mru_ghost->arcs_lists[i]);
 		list_destroy(&arc_mfu->arcs_lists[i]);
 		list_destroy(&arc_mfu_ghost->arcs_lists[i]);
+		list_destroy(&arc_l2c_only->arcs_lists[i]);
 
 		mutex_destroy(&arc_anon->arcs_locks[i].arcs_lock);
 		mutex_destroy(&arc_mru->arcs_locks[i].arcs_lock);
 		mutex_destroy(&arc_mru_ghost->arcs_locks[i].arcs_lock);
 		mutex_destroy(&arc_mfu->arcs_locks[i].arcs_lock);
 		mutex_destroy(&arc_mfu_ghost->arcs_locks[i].arcs_lock);
+		mutex_destroy(&arc_l2c_only->arcs_locks[i].arcs_lock);
 	}
 	
 	mutex_destroy(&zfs_write_limit_lock);

From 224329fb6b7aa08f0384ee0ec7c1a6fa6dd9828c Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:36:53 +0000
Subject: [PATCH 122/532] Style fixes.

---
 .../opensolaris/uts/common/fs/zfs/arc.c       | 120 ++++++++----------
 1 file changed, 54 insertions(+), 66 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
index 740a0a57f3e..ceb4c87f2f6 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/arc.c
@@ -131,7 +131,6 @@
 #include <sys/kstat.h>
 #include <sys/sdt.h>
 
-#include <sys/ktr.h>
 #include <vm/vm_pageout.h>
 
 static kmutex_t		arc_reclaim_thr_lock;
@@ -238,7 +237,7 @@ struct arcs_lock {
  */
 #define ARC_BUFC_NUMDATALISTS		16
 #define ARC_BUFC_NUMMETADATALISTS	16
-#define ARC_BUFC_NUMLISTS	(ARC_BUFC_NUMMETADATALISTS+ARC_BUFC_NUMDATALISTS)
+#define ARC_BUFC_NUMLISTS	(ARC_BUFC_NUMMETADATALISTS + ARC_BUFC_NUMDATALISTS)
 
 typedef struct arc_state {
 	uint64_t arcs_lsize[ARC_BUFC_NUMTYPES];	/* amount of evictable data */
@@ -247,7 +246,7 @@ typedef struct arc_state {
 	struct arcs_lock arcs_locks[ARC_BUFC_NUMLISTS] __aligned(CACHE_LINE_SIZE);
 } arc_state_t;
 
-#define ARCS_LOCK(s, i) &((s)->arcs_locks[(i)].arcs_lock)
+#define ARCS_LOCK(s, i)	(&((s)->arcs_locks[(i)].arcs_lock))
 
 /* The 6 states: */
 static arc_state_t ARC_anon;
@@ -307,9 +306,9 @@ typedef struct arc_stats {
 	kstat_named_t arcstat_l2_hdr_size;
 	kstat_named_t arcstat_memory_throttle_count;
 	kstat_named_t arcstat_l2_write_trylock_fail;
-	kstat_named_t arcstat_l2_write_in_l2;
 	kstat_named_t arcstat_l2_write_passed_headroom;
 	kstat_named_t arcstat_l2_write_spa_mismatch;
+	kstat_named_t arcstat_l2_write_in_l2;
 	kstat_named_t arcstat_l2_write_hdr_io_in_progress;
 	kstat_named_t arcstat_l2_write_not_cacheable;
 	kstat_named_t arcstat_l2_write_full;
@@ -370,19 +369,19 @@ static arc_stats_t arc_stats = {
 	{ "l2_size",			KSTAT_DATA_UINT64 },
 	{ "l2_hdr_size",		KSTAT_DATA_UINT64 },
 	{ "memory_throttle_count",	KSTAT_DATA_UINT64 },
-	{ "l2_write_trylock_fail", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_in_l2", 		KSTAT_DATA_UINT64 },
-	{ "l2_write_passed_headroom", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_spa_mismatch", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_io_in_progress", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_not_cacheable", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_full", 		KSTAT_DATA_UINT64 },
-	{ "l2_write_buffer_iter", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_pios", 		KSTAT_DATA_UINT64 },
-	{ "l2_write_bytes_written", 		KSTAT_DATA_UINT64 },
-	{ "l2_write_buffer_bytes_scanned", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_buffer_list_iter", 	KSTAT_DATA_UINT64 },
-	{ "l2_write_buffer_list_null_iter", 	KSTAT_DATA_UINT64 }
+	{ "l2_write_trylock_fail",	KSTAT_DATA_UINT64 },
+	{ "l2_write_passed_headroom",	KSTAT_DATA_UINT64 },
+	{ "l2_write_spa_mismatch",	KSTAT_DATA_UINT64 },
+	{ "l2_write_in_l2",		KSTAT_DATA_UINT64 },
+	{ "l2_write_io_in_progress",	KSTAT_DATA_UINT64 },
+	{ "l2_write_not_cacheable",	KSTAT_DATA_UINT64 },
+	{ "l2_write_full",		KSTAT_DATA_UINT64 },
+	{ "l2_write_buffer_iter",	KSTAT_DATA_UINT64 },
+	{ "l2_write_pios",		KSTAT_DATA_UINT64 },
+	{ "l2_write_bytes_written",	KSTAT_DATA_UINT64 },
+	{ "l2_write_buffer_bytes_scanned", KSTAT_DATA_UINT64 },
+	{ "l2_write_buffer_list_iter",	KSTAT_DATA_UINT64 },
+	{ "l2_write_buffer_list_null_iter", KSTAT_DATA_UINT64 }
 };
 
 #define	ARCSTAT(stat)	(arc_stats.stat.value.ui64)
@@ -390,7 +389,7 @@ static arc_stats_t arc_stats = {
 #define	ARCSTAT_INCR(stat, val) \
 	atomic_add_64(&arc_stats.stat.value.ui64, (val));
 
-#define	ARCSTAT_BUMP(stat) 	ARCSTAT_INCR(stat, 1)
+#define	ARCSTAT_BUMP(stat)	ARCSTAT_INCR(stat, 1)
 #define	ARCSTAT_BUMPDOWN(stat)	ARCSTAT_INCR(stat, -1)
 
 #define	ARCSTAT_MAX(stat, val) {					\
@@ -424,7 +423,7 @@ static arc_stats_t arc_stats = {
 	}
 
 kstat_t			*arc_ksp;
-static arc_state_t 	*arc_anon;
+static arc_state_t	*arc_anon;
 static arc_state_t	*arc_mru;
 static arc_state_t	*arc_mru_ghost;
 static arc_state_t	*arc_mfu;
@@ -1076,10 +1075,10 @@ get_buf_info(arc_buf_hdr_t *ab, arc_state_t *state, list_t **list, kmutex_t **lo
 {
 	uint64_t buf_hashid = buf_hash(ab->b_spa, &ab->b_dva, ab->b_birth);
 
-	if (ab->b_type == ARC_BUFC_METADATA) 
-		buf_hashid &= (ARC_BUFC_NUMMETADATALISTS-1);
+	if (ab->b_type == ARC_BUFC_METADATA)
+		buf_hashid &= (ARC_BUFC_NUMMETADATALISTS - 1);
 	else {
-		buf_hashid &= (ARC_BUFC_NUMDATALISTS-1);
+		buf_hashid &= (ARC_BUFC_NUMDATALISTS - 1);
 		buf_hashid += ARC_BUFC_NUMMETADATALISTS;
 	}
 
@@ -1096,10 +1095,10 @@ add_reference(arc_buf_hdr_t *ab, kmutex_t *hash_lock, void *tag)
 
 	if ((refcount_add(&ab->b_refcnt, tag) == 1) &&
 	    (ab->b_state != arc_anon)) {
-		list_t *list;
-		kmutex_t *lock;
 		uint64_t delta = ab->b_size * ab->b_datacnt;
 		uint64_t *size = &ab->b_state->arcs_lsize[ab->b_type];
+		list_t *list;
+		kmutex_t *lock;
 
 		get_buf_info(ab, ab->b_state, &list, &lock);
 		ASSERT(!MUTEX_HELD(lock));
@@ -1179,7 +1178,6 @@ arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *ab, kmutex_t *hash_lock)
 
 			get_buf_info(ab, old_state, &list, &lock);
 			use_mutex = !MUTEX_HELD(lock);
-
 			if (use_mutex)
 				mutex_enter(lock);
 
@@ -1202,13 +1200,11 @@ arc_change_state(arc_state_t *new_state, arc_buf_hdr_t *ab, kmutex_t *hash_lock)
 				mutex_exit(lock);
 		}
 		if (new_state != arc_anon) {
-			int use_mutex; 
+			int use_mutex;
 			uint64_t *size = &new_state->arcs_lsize[ab->b_type];
 
 			get_buf_info(ab, new_state, &list, &lock);
 			use_mutex = !MUTEX_HELD(lock);
-			
-			
 			if (use_mutex)
 				mutex_enter(lock);
 
@@ -1626,7 +1622,7 @@ arc_evict(arc_state_t *state, spa_t *spa, int64_t bytes, boolean_t recycle,
 	ASSERT(state == arc_mru || state == arc_mfu);
 
 	evicted_state = (state == arc_mru) ? arc_mru_ghost : arc_mfu_ghost;
-	
+
 	if (type == ARC_BUFC_METADATA) {
 		offset = 0;
 		list_count = ARC_BUFC_NUMMETADATALISTS;
@@ -1635,7 +1631,6 @@ arc_evict(arc_state_t *state, spa_t *spa, int64_t bytes, boolean_t recycle,
 		idx = evict_metadata_offset;
 	} else {
 		offset = ARC_BUFC_NUMMETADATALISTS;
-
 		list_start = &state->arcs_lists[offset];
 		evicted_list_start = &evicted_state->arcs_lists[offset];
 		list_count = ARC_BUFC_NUMDATALISTS;
@@ -1643,12 +1638,12 @@ arc_evict(arc_state_t *state, spa_t *spa, int64_t bytes, boolean_t recycle,
 	}
 	bytes_remaining = evicted_state->arcs_lsize[type];
 	count = 0;
-	
+
 evict_start:
 	list = &list_start[idx];
 	evicted_list = &evicted_list_start[idx];
 	lock = ARCS_LOCK(state, (offset + idx));
-	evicted_lock = ARCS_LOCK(evicted_state, (offset + idx)); 
+	evicted_lock = ARCS_LOCK(evicted_state, (offset + idx));
 
 	mutex_enter(lock);
 	mutex_enter(evicted_lock);
@@ -1718,7 +1713,7 @@ evict_start:
 			if (bytes_remaining > 0) {
 				mutex_exit(evicted_lock);
 				mutex_exit(lock);
-				idx  = ((idx + 1)&(list_count-1));
+				idx  = ((idx + 1) & (list_count - 1));
 				count++;
 				goto evict_start;
 			}
@@ -1729,8 +1724,8 @@ evict_start:
 
 	mutex_exit(evicted_lock);
 	mutex_exit(lock);
-	
-	idx  = ((idx + 1)&(list_count-1));
+
+	idx  = ((idx + 1) & (list_count - 1));
 	count++;
 
 	if (bytes_evicted < bytes) {
@@ -1740,11 +1735,11 @@ evict_start:
 			dprintf("only evicted %lld bytes from %x",
 			    (longlong_t)bytes_evicted, state);
 	}
-	if (type == ARC_BUFC_METADATA) 
+	if (type == ARC_BUFC_METADATA)
 		evict_metadata_offset = idx;
 	else
 		evict_data_offset = idx;
-		
+
 	if (skipped)
 		ARCSTAT_INCR(arcstat_evict_skip, skipped);
 
@@ -1801,7 +1796,7 @@ arc_evict_ghost(arc_state_t *state, spa_t *spa, int64_t bytes)
 	list_start = &state->arcs_lists[ARC_BUFC_NUMMETADATALISTS];
 	list_count = ARC_BUFC_NUMDATALISTS;
 	offset = ARC_BUFC_NUMMETADATALISTS;
-	
+
 evict_start:
 	list = &list_start[idx];
 	lock = ARCS_LOCK(state, idx + offset);
@@ -1848,12 +1843,12 @@ evict_start:
 		}
 	}
 	mutex_exit(lock);
-	idx  = ((idx + 1)&(ARC_BUFC_NUMDATALISTS-1));
+	idx  = ((idx + 1) & (ARC_BUFC_NUMDATALISTS - 1));
 	count++;
-	
+
 	if (count < list_count)
 		goto evict_start;
-	
+
 	evict_offset = idx;
 	if ((uintptr_t)list > (uintptr_t)&state->arcs_lists[ARC_BUFC_NUMMETADATALISTS] &&
 	    (bytes < 0 || bytes_deleted < bytes)) {
@@ -1942,7 +1937,7 @@ arc_do_user_evicts(void)
 	/*
 	 * Move list over to avoid LOR
 	 */
-restart:	
+restart:
 	mutex_enter(&arc_eviction_mtx);
 	tmp_arc_eviction_list = arc_eviction_list;
 	arc_eviction_list = NULL;
@@ -2053,7 +2048,7 @@ arc_reclaim_needed(void)
 		return (0);
 
 	/*
-	 * If pages are needed or we're within 2048 pages 
+	 * If pages are needed or we're within 2048 pages
 	 * of needing to page need to reclaim
 	 */
 	if (vm_pages_needed || (vm_paging_target() > -2048))
@@ -2611,10 +2606,7 @@ arc_read_done(zio_t *zio)
 	hdr->b_flags &= ~ARC_L2_EVICTED;
 	if (l2arc_noprefetch && (hdr->b_flags & ARC_PREFETCH))
 		hdr->b_flags &= ~ARC_L2CACHE;
-#if 0	
-	else if ((hdr->b_flags & ARC_PREFETCH) == 0)
-		hdr->b_flags |= ARC_L2CACHE;
-#endif	
+
 	/* byteswap if necessary */
 	callback_list = hdr->b_acb;
 	ASSERT(callback_list != NULL);
@@ -2951,7 +2943,7 @@ top:
 				 * released by l2arc_read_done().
 				 */
 				rzio = zio_read_phys(pio, vd, addr, size,
-				    buf->b_data, ZIO_CHECKSUM_OFF, 
+				    buf->b_data, ZIO_CHECKSUM_OFF,
 				    l2arc_read_done, cb, priority, zio_flags |
 				    ZIO_FLAG_DONT_CACHE | ZIO_FLAG_CANFAIL |
 				    ZIO_FLAG_DONT_PROPAGATE |
@@ -3048,7 +3040,7 @@ arc_buf_evict(arc_buf_t *buf)
 	arc_buf_t **bufp;
 	list_t *list, *evicted_list;
 	kmutex_t *lock, *evicted_lock;
-	
+
 	rw_enter(&buf->b_lock, RW_WRITER);
 	hdr = buf->b_hdr;
 	if (hdr == NULL) {
@@ -3723,7 +3715,6 @@ arc_init(void)
 	arc_size = 0;
 
 	for (i = 0; i < ARC_BUFC_NUMLISTS; i++) {
-		
 		mutex_init(&arc_anon->arcs_locks[i].arcs_lock,
 		    NULL, MUTEX_DEFAULT, NULL);
 		mutex_init(&arc_mru->arcs_locks[i].arcs_lock,
@@ -3736,7 +3727,7 @@ arc_init(void)
 		    NULL, MUTEX_DEFAULT, NULL);
 		mutex_init(&arc_l2c_only->arcs_locks[i].arcs_lock,
 		    NULL, MUTEX_DEFAULT, NULL);
-	
+
 		list_create(&arc_mru->arcs_lists[i],
 		    sizeof (arc_buf_hdr_t), offsetof(arc_buf_hdr_t, b_arc_node));
 		list_create(&arc_mru_ghost->arcs_lists[i],
@@ -3786,7 +3777,7 @@ arc_init(void)
 #ifdef _KERNEL
 	if (TUNABLE_INT_FETCH("vfs.zfs.prefetch_disable", &zfs_prefetch_disable))
 		prefetch_tunable_set = 1;
-	
+
 #ifdef __i386__
 	if (prefetch_tunable_set == 0) {
 		printf("ZFS NOTICE: Prefetch is disabled by default on i386 "
@@ -3795,7 +3786,7 @@ arc_init(void)
 		    "to /boot/loader.conf.\n");
 		zfs_prefetch_disable=1;
 	}
-#else	
+#else
 	if ((((uint64_t)physmem * PAGESIZE) < (1ULL << 32)) &&
 	    prefetch_tunable_set == 0) {
 		printf("ZFS NOTICE: Prefetch is disabled by default if less "
@@ -3804,7 +3795,7 @@ arc_init(void)
 		    "to /boot/loader.conf.\n");
 		zfs_prefetch_disable=1;
 	}
-#endif	
+#endif
 	/* Warn about ZFS memory and address space requirements. */
 	if (((uint64_t)physmem * PAGESIZE) < (256 + 128 + 64) * (1 << 20)) {
 		printf("ZFS WARNING: Recommended minimum RAM size is 512MB; "
@@ -3824,7 +3815,7 @@ void
 arc_fini(void)
 {
 	int i;
-	
+
 	mutex_enter(&arc_reclaim_thr_lock);
 	arc_thread_exit = 1;
 	cv_signal(&arc_reclaim_thr_cv);
@@ -3859,7 +3850,7 @@ arc_fini(void)
 		mutex_destroy(&arc_mfu_ghost->arcs_locks[i].arcs_lock);
 		mutex_destroy(&arc_l2c_only->arcs_locks[i].arcs_lock);
 	}
-	
+
 	mutex_destroy(&zfs_write_limit_lock);
 
 	buf_fini();
@@ -4255,18 +4246,18 @@ l2arc_list_locked(int list_num, kmutex_t **lock)
 {
 	list_t *list;
 	int idx;
-	
-	ASSERT(list_num >= 0 && list_num < 2*ARC_BUFC_NUMLISTS);
+
+	ASSERT(list_num >= 0 && list_num < 2 * ARC_BUFC_NUMLISTS);
 
 	if (list_num < ARC_BUFC_NUMMETADATALISTS) {
 		idx = list_num;
 		list = &arc_mfu->arcs_lists[idx];
 		*lock = ARCS_LOCK(arc_mfu, idx);
-	} else if (list_num < ARC_BUFC_NUMMETADATALISTS*2) {
+	} else if (list_num < ARC_BUFC_NUMMETADATALISTS * 2) {
 		idx = list_num - ARC_BUFC_NUMMETADATALISTS;
 		list = &arc_mru->arcs_lists[idx];
 		*lock = ARCS_LOCK(arc_mru, idx);
-	} else if (list_num < (ARC_BUFC_NUMMETADATALISTS*2 +
+	} else if (list_num < (ARC_BUFC_NUMMETADATALISTS * 2 +
 		ARC_BUFC_NUMDATALISTS)) {
 		idx = list_num - ARC_BUFC_NUMMETADATALISTS;
 		list = &arc_mfu->arcs_lists[idx];
@@ -4277,8 +4268,6 @@ l2arc_list_locked(int list_num, kmutex_t **lock)
 		*lock = ARCS_LOCK(arc_mru, idx);
 	}
 
-	CTR3(KTR_SPARE2, "list=%p list_num=%d idx=%d",
-	    list, list_num, idx);
 	ASSERT(!(MUTEX_HELD(*lock)));
 	mutex_enter(*lock);
 	return (list);
@@ -4448,7 +4437,7 @@ l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
 	 * Copy buffers for L2ARC writing.
 	 */
 	mutex_enter(&l2arc_buflist_mtx);
-	for (try = 0; try < 2*ARC_BUFC_NUMLISTS; try++) {
+	for (try = 0; try < 2 * ARC_BUFC_NUMLISTS; try++) {
 		list = l2arc_list_locked(try, &list_lock);
 		passed_sz = 0;
 		ARCSTAT_BUMP(arcstat_l2_write_buffer_list_iter);
@@ -4464,9 +4453,8 @@ l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
 			ab = list_head(list);
 		else
 			ab = list_tail(list);
-		if (ab == NULL) {
+		if (ab == NULL)
 			ARCSTAT_BUMP(arcstat_l2_write_buffer_list_null_iter);
-		}
 
 		for (; ab; ab = ab_prev) {
 			if (arc_warm == B_FALSE)
@@ -4474,7 +4462,7 @@ l2arc_write_buffers(spa_t *spa, l2arc_dev_t *dev, uint64_t target_sz)
 			else
 				ab_prev = list_prev(list, ab);
 			ARCSTAT_INCR(arcstat_l2_write_buffer_bytes_scanned, ab->b_size);
-		
+
 			hash_lock = HDR_LOCK(ab);
 			have_lock = MUTEX_HELD(hash_lock);
 			if (!have_lock && !mutex_tryenter(hash_lock)) {

From bd7226a57219b0e1d6d68c2cab87dace5f40ee1b Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sun, 18 Apr 2010 12:43:33 +0000
Subject: [PATCH 123/532] Restore previous order.

---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h
index 3ca7249305a..f3e00877a8e 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/arc.h
@@ -55,8 +55,8 @@ struct arc_buf {
 };
 
 typedef enum arc_buf_contents {
-	ARC_BUFC_METADATA,			/* buffer contains metadata */
 	ARC_BUFC_DATA,				/* buffer contains data */
+	ARC_BUFC_METADATA,			/* buffer contains metadata */
 	ARC_BUFC_NUMTYPES
 } arc_buf_contents_t;
 /*

From 4b9dd5d53737effedae23defa9bda6f0c74320a5 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 18 Apr 2010 17:50:09 +0000
Subject: [PATCH 124/532] There is no justification for vm_object_split()
 setting PG_REFERENCED on a page that it is going to sleep on.  Eliminate it.

MFC after:	3 weeks
---
 sys/vm/vm_object.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 1f9ba6d4956..7e4dbc85350 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1422,7 +1422,6 @@ retry:
 		 * not be changed by this operation.
 		 */
 		if ((m->oflags & VPO_BUSY) || m->busy) {
-			vm_page_flag_set(m, PG_REFERENCED);
 			vm_page_unlock_queues();
 			VM_OBJECT_UNLOCK(new_object);
 			m->oflags |= VPO_WANTED;

From 8d0331993042102b34533c1e0e91fe1f5f8783c9 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sun, 18 Apr 2010 18:23:11 +0000
Subject: [PATCH 125/532] Revert r206649.

Simplify the presented declaration of struct sigaction, noting the
caveat in the text. Real layout of the structure and exposed
implementation namespace only obfuscates the usage.

Submitted by:	bde
MFC after:	3 days
---
 lib/libc/sys/sigaction.2 | 31 +++++++++++++------------------
 1 file changed, 13 insertions(+), 18 deletions(-)

diff --git a/lib/libc/sys/sigaction.2 b/lib/libc/sys/sigaction.2
index 51b6c47c78f..3b4f2f0196d 100644
--- a/lib/libc/sys/sigaction.2
+++ b/lib/libc/sys/sigaction.2
@@ -28,7 +28,7 @@
 .\"	From: @(#)sigaction.2	8.2 (Berkeley) 4/3/94
 .\" $FreeBSD$
 .\"
-.Dd April 13, 2010
+.Dd April 18, 2010
 .Dt SIGACTION 2
 .Os
 .Sh NAME
@@ -40,16 +40,11 @@
 .In signal.h
 .Bd -literal
 struct  sigaction {
-        union {
-                void    (*__sa_handler)(int);
-                void    (*__sa_sigaction)(int, siginfo_t *, void *);
-        } __sigaction_u;                /* signal handler */
+        void    (*sa_handler)(int);
+        void    (*sa_sigaction)(int, siginfo_t *, void *);
         int     sa_flags;               /* see signal options below */
         sigset_t sa_mask;               /* signal mask to apply */
 };
-
-#define	sa_handler	__sigaction_u.__sa_handler
-#define	sa_sigaction	__sigaction_u.__sa_sigaction
 .Ed
 .Ft int
 .Fo sigaction
@@ -148,6 +143,16 @@ If
 is non-zero, the previous handling information for the signal
 is returned to the user.
 .Pp
+The above declaration of
+.Vt "struct sigaction"
+is not literal.
+It is provided only to list the accessible members.
+See
+.In sys/signal.h
+for the actual definition.
+In particular, the storage occupied by sa_handler and sa_sigaction overlaps,
+and an application can not use both simultaneously.
+.Pp
 Once a signal handler is installed, it normally remains installed
 until another
 .Fn sigaction
@@ -496,16 +501,6 @@ or
 .Dv SIG_IGN
 this way.
 .Pp
-If preprocessing symbol
-.Va _POSIX_C_SOURCE
-with the value >= 199309 is not defined, the following declaration for
-the handler shall be used:
-.Bl -tag -offset indent -width short
-.It Tn POSIX Dv SA_SIGINFO :
-.Ft void
-.Fn handler int "struct __sigaction *" "void *" ;
-.El
-.Pp
 If the
 .Dv SA_SIGINFO
 flag is not set, the handler function should match

From 685b0bb81486270480e6cdf7748769d28eb7e7c3 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Sun, 18 Apr 2010 18:43:36 +0000
Subject: [PATCH 126/532] Delete svn:executable prop.

---
 sys/mips/rmi/dev/sec/desc.h | 0
 1 file changed, 0 insertions(+), 0 deletions(-)
 mode change 100755 => 100644 sys/mips/rmi/dev/sec/desc.h

diff --git a/sys/mips/rmi/dev/sec/desc.h b/sys/mips/rmi/dev/sec/desc.h
old mode 100755
new mode 100644

From b28889a2fc4a6f08ffc0c3b2991310c270df854b Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 18 Apr 2010 21:29:28 +0000
Subject: [PATCH 127/532] Remove a nonsensical test from vm_pageout_clean().  A
 page can't be in the inactive queue and have a non-zero wire count.

Reviewed by:	kib
MFC after:	3 weeks
---
 sys/vm/vm_pageout.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index dc2b3b77d56..d5ede5a7503 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -350,7 +350,6 @@ more:
 		vm_page_test_dirty(p);
 		if (p->dirty == 0 ||
 		    p->queue != PQ_INACTIVE ||
-		    p->wire_count != 0 ||	/* may be held by buf cache */
 		    p->hold_count != 0) {	/* may be undergoing I/O */
 			ib = 0;
 			break;
@@ -378,7 +377,6 @@ more:
 		vm_page_test_dirty(p);
 		if (p->dirty == 0 ||
 		    p->queue != PQ_INACTIVE ||
-		    p->wire_count != 0 ||	/* may be held by buf cache */
 		    p->hold_count != 0) {	/* may be undergoing I/O */
 			break;
 		}

From 348c81a032e288731dbe9891650ff8cbac6dbe9e Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 18 Apr 2010 22:13:45 +0000
Subject: [PATCH 128/532] sh: Add testcases for double-quotes within quoted
 ${var+-...} (non-POSIX).

POSIX leaves things like "${var+"word"}" undefined.
We follow traditional ash behaviour here.
Hence, these testcases also work on stable/8.
---
 .../regression/bin/sh/expansion/plus-minus3.0 | 44 +++++++++++++++++++
 1 file changed, 44 insertions(+)
 create mode 100644 tools/regression/bin/sh/expansion/plus-minus3.0

diff --git a/tools/regression/bin/sh/expansion/plus-minus3.0 b/tools/regression/bin/sh/expansion/plus-minus3.0
new file mode 100644
index 00000000000..49fcdc2f8f5
--- /dev/null
+++ b/tools/regression/bin/sh/expansion/plus-minus3.0
@@ -0,0 +1,44 @@
+# $FreeBSD$
+
+e= q='?' a='*' t=texttext s='ast*que?non' p='/et[c]/' w='a b c' b='{{(#)}}'
+h='##'
+failures=''
+ok=''
+
+testcase() {
+	code="$1"
+	expected="$2"
+	oIFS="$IFS"
+	eval "$code"
+	IFS='|'
+	result="$#|$*"
+	IFS="$oIFS"
+	if [ "x$result" = "x$expected" ]; then
+		ok=x$ok
+	else
+		failures=x$failures
+		echo "For $code, expected $expected actual $result"
+	fi
+}
+
+# We follow original ash behaviour for quoted ${var+-=?} expansions:
+# a double-quote in one switches back to unquoted state.
+# This allows expanding a variable as a single word if it is set
+# and substituting multiple words otherwise.
+# It is also close to the Bourne and Korn shells.
+# POSIX leaves this undefined, and various other shells treat
+# such double-quotes as introducing a second level of quoting
+# which does not do much except quoting close braces.
+
+testcase 'set -- "${p+"/et[c]/"}"'		'1|/etc/'
+testcase 'set -- "${p-"/et[c]/"}"'		'1|/et[c]/'
+testcase 'set -- "${p+"$p"}"'			'1|/etc/'
+testcase 'set -- "${p-"$p"}"'			'1|/et[c]/'
+testcase 'set -- "${p+"""/et[c]/"}"'		'1|/etc/'
+testcase 'set -- "${p-"""/et[c]/"}"'		'1|/et[c]/'
+testcase 'set -- "${p+"""$p"}"'			'1|/etc/'
+testcase 'set -- "${p-"""$p"}"'			'1|/et[c]/'
+testcase 'set -- "${p+"\@"}"'			'1|@'
+testcase 'set -- "${p+"'\''/et[c]/'\''"}"'	'1|/et[c]/'
+
+test "x$failures" = x

From 7ea710b3b174504918463bbe3456c5cf057baafe Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Sun, 18 Apr 2010 22:21:23 +0000
Subject: [PATCH 129/532] Avoid extraneous recovery cycles in the experimental
 NFS client when an NFSv4 server reboots, by doing two things. 1 - Make the
 function that acquires a stateid for I/O operations     block until recovery
 is complete, so that it doesn't acquire     out of date stateids. 2 - Only
 allow a recovery once every 1/2 of a lease duration, since     the NFSv4
 server must provide a recovery grace period of at     least a lease duration.
 This should avoid recoveries caused     by an out of date stateid that was
 acquired for an I/O op.     just before a recovery cycle started.

MFC after:	1 week
---
 sys/fs/nfs/nfsclstate.h        |  1 +
 sys/fs/nfsclient/nfs_clstate.c | 33 +++++++++++++++++++++++++++++----
 2 files changed, 30 insertions(+), 4 deletions(-)

diff --git a/sys/fs/nfs/nfsclstate.h b/sys/fs/nfs/nfsclstate.h
index edd479cf2ce..72d8eebfc39 100644
--- a/sys/fs/nfs/nfsclstate.h
+++ b/sys/fs/nfs/nfsclstate.h
@@ -74,6 +74,7 @@ struct nfsclclient {
 #define	NFSCLFLAGS_EXPIREIT	0x0040
 #define	NFSCLFLAGS_FIRSTDELEG	0x0080
 #define	NFSCLFLAGS_GOTDELEG	0x0100
+#define	NFSCLFLAGS_RECVRINPROG	0x0200
 
 struct nfsclowner {
 	LIST_ENTRY(nfsclowner)	nfsow_list;
diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index 7bccbfcc5d7..6a189635ff7 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -480,6 +480,13 @@ nfscl_getstateid(vnode_t vp, u_int8_t *nfhp, int fhlen, u_int32_t mode,
 		return (EACCES);
 	}
 
+	/*
+	 * Wait for recovery to complete.
+	 */
+	while ((clp->nfsc_flags & NFSCLFLAGS_RECVRINPROG))
+		(void) nfsmsleep(&clp->nfsc_flags, NFSCLSTATEMUTEXPTR,
+		    PZERO, "nfsrecvr", NULL);
+
 	/*
 	 * First, look for a delegation.
 	 */
@@ -1778,6 +1785,7 @@ nfscl_recover(struct nfsclclient *clp, struct ucred *cred, NFSPROC_T *p)
 	 * block when trying to use state.
 	 */
 	NFSLOCKCLSTATE();
+	clp->nfsc_flags |= NFSCLFLAGS_RECVRINPROG;
 	do {
 		igotlock = nfsv4_lock(&clp->nfsc_lock, 1, NULL,
 		    NFSCLSTATEMUTEXPTR);
@@ -1794,9 +1802,10 @@ nfscl_recover(struct nfsclclient *clp, struct ucred *cred, NFSPROC_T *p)
 	     error == NFSERR_STALEDONTRECOVER) && --trycnt > 0);
 	if (error) {
 		nfscl_cleanclient(clp);
-		clp->nfsc_flags &= ~(NFSCLFLAGS_HASCLIENTID |
-		    NFSCLFLAGS_RECOVER);
 		NFSLOCKCLSTATE();
+		clp->nfsc_flags &= ~(NFSCLFLAGS_HASCLIENTID |
+		    NFSCLFLAGS_RECOVER | NFSCLFLAGS_RECVRINPROG);
+		wakeup(&clp->nfsc_flags);
 		nfsv4_unlock(&clp->nfsc_lock, 0);
 		NFSUNLOCKCLSTATE();
 		return;
@@ -2057,6 +2066,8 @@ nfscl_recover(struct nfsclclient *clp, struct ucred *cred, NFSPROC_T *p)
 	}
 
 	NFSLOCKCLSTATE();
+	clp->nfsc_flags &= ~NFSCLFLAGS_RECVRINPROG;
+	wakeup(&clp->nfsc_flags);
 	nfsv4_unlock(&clp->nfsc_lock, 0);
 	NFSUNLOCKCLSTATE();
 	NFSFREECRED(tcred);
@@ -2316,6 +2327,7 @@ nfscl_renewthread(struct nfsclclient *clp, NFSPROC_T *p)
 	struct ucred *cred;
 	u_int32_t clidrev;
 	int error, cbpathdown, islept, igotlock, ret, clearok;
+	uint32_t recover_done_time = 0;
 
 	cred = newnfs_getcred();
 	NFSLOCKCLSTATE();
@@ -2324,8 +2336,21 @@ nfscl_renewthread(struct nfsclclient *clp, NFSPROC_T *p)
 	for(;;) {
 		newnfs_setroot(cred);
 		cbpathdown = 0;
-		if (clp->nfsc_flags & NFSCLFLAGS_RECOVER)
-			nfscl_recover(clp, cred, p);
+		if (clp->nfsc_flags & NFSCLFLAGS_RECOVER) {
+			/*
+			 * Only allow one recover within 1/2 of the lease
+			 * duration (nfsc_renew).
+			 */
+			if (recover_done_time < NFSD_MONOSEC) {
+				recover_done_time = NFSD_MONOSEC +
+				    clp->nfsc_renew;
+				nfscl_recover(clp, cred, p);
+			} else {
+				NFSLOCKCLSTATE();
+				clp->nfsc_flags &= ~NFSCLFLAGS_RECOVER;
+				NFSUNLOCKCLSTATE();
+			}
+		}
 		if (clp->nfsc_expire <= NFSD_MONOSEC &&
 		    (clp->nfsc_flags & NFSCLFLAGS_HASCLIENTID)) {
 			clp->nfsc_expire = NFSD_MONOSEC + clp->nfsc_renew;

From ca596a25f07f94a05515cd150b1d7e6f81c1894d Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sun, 18 Apr 2010 22:32:07 +0000
Subject: [PATCH 130/532] o) Add a VM find-space option,
 VMFS_TLB_ALIGNED_SPACE, which searches the    address space for an address as
 aligned by the new pmap_align_tlb()    function, which is for constraints
 imposed by the TLB. [1] o) Add a kmem_alloc_nofault_space() function, which
 acts like    kmem_alloc_nofault() but allows the caller to specify which
 find-space    option to use. [1] o) Use kmem_alloc_nofault_space() with
 VMFS_TLB_ALIGNED_SPACE to allocate the    kernel stack address on MIPS. [1]
 o) Make pmap_align_tlb() on MIPS align addresses so that they do not start on
    an odd boundary within the TLB, so that they are suitable for insertion as
    wired entries and do not have to share a TLB entry with another mapping,  
  assuming they are appropriately-sized. o) Eliminate md_realstack now that
 the kstack will be appropriately-aligned on    MIPS. o) Increase the number
 of guard pages to 2 so that we retain the proper    alignment of the kstack
 address.

Reviewed by:	[1] alc
X-MFC-after:	Making sure alc has not come up with a better interface.
---
 sys/mips/include/param.h   |  7 ++-----
 sys/mips/include/proc.h    |  3 +--
 sys/mips/mips/exception.S  |  4 ++--
 sys/mips/mips/genassym.c   |  2 +-
 sys/mips/mips/machdep.c    |  5 ++---
 sys/mips/mips/pmap.c       | 15 +++++++++++++++
 sys/mips/mips/swtch.S      |  2 +-
 sys/mips/mips/vm_machdep.c | 26 +++++++-------------------
 sys/vm/pmap.h              |  3 +++
 sys/vm/vm_extern.h         |  1 +
 sys/vm/vm_glue.c           |  9 +++++++++
 sys/vm/vm_kern.c           | 29 +++++++++++++++++++++++++++++
 sys/vm/vm_map.c            | 13 ++++++++++++-
 sys/vm/vm_map.h            |  3 +++
 14 files changed, 88 insertions(+), 34 deletions(-)

diff --git a/sys/mips/include/param.h b/sys/mips/include/param.h
index 0edb8b93d55..06cdeac0483 100644
--- a/sys/mips/include/param.h
+++ b/sys/mips/include/param.h
@@ -113,12 +113,9 @@
 
 /*
  * The kernel stack needs to be aligned on a (PAGE_SIZE * 2) boundary.
- *
- * Although we allocate 3 pages for the kernel stack we end up using
- * only the 2 pages that are aligned on a (PAGE_SIZE * 2) boundary.
  */
-#define	KSTACK_PAGES		3	/* kernel stack*/
-#define	KSTACK_GUARD_PAGES	1	/* pages of kstack guard; 0 disables */
+#define	KSTACK_PAGES		2	/* kernel stack*/
+#define	KSTACK_GUARD_PAGES	2	/* pages of kstack guard; 0 disables */
 
 #define	UPAGES			2
 
diff --git a/sys/mips/include/proc.h b/sys/mips/include/proc.h
index 99dab78e165..0491b11b02e 100644
--- a/sys/mips/include/proc.h
+++ b/sys/mips/include/proc.h
@@ -44,7 +44,7 @@
  */
 struct mdthread {
 	int	md_flags;		/* machine-dependent flags */
-	int	md_upte[KSTACK_PAGES - 1]; /* ptes for mapping u pcb */
+	int	md_upte[KSTACK_PAGES];	/* ptes for mapping u pcb */
 	int	md_ss_addr;		/* single step address for ptrace */
 	int	md_ss_instr;		/* single step instruction for ptrace */
 	register_t	md_saved_intr;
@@ -53,7 +53,6 @@ struct mdthread {
 	int	md_pc_ctrl;		/* performance counter control */
 	int	md_pc_count;		/* performance counter */
 	int	md_pc_spill;		/* performance counter spill */
-	vm_offset_t	md_realstack;
 	void	*md_tls;
 };
 
diff --git a/sys/mips/mips/exception.S b/sys/mips/mips/exception.S
index 8c32b582c3c..aa775ea5d83 100644
--- a/sys/mips/mips/exception.S
+++ b/sys/mips/mips/exception.S
@@ -928,7 +928,7 @@ tlb_insert_random:
 	 */
 	GET_CPU_PCPU(k1)
 	lw	k0, PC_CURTHREAD(k1)
-	lw	k0, TD_REALKSTACK(k0)
+	lw	k0, TD_KSTACK(k0)
 	sltu	k0, k0, sp
 	bnez	k0, _C_LABEL(MipsKernGenException)
 	nop
@@ -975,7 +975,7 @@ tlb_insert_random:
 	 */
 	GET_CPU_PCPU(k1)
 	lw	k0, PC_CURTHREAD(k1)
-	sw	zero, TD_REALKSTACK(k0)
+	sw	zero, TD_KSTACK(k0)
 
 	move	a1, a0
 	PANIC("kernel stack overflow - trapframe at %p")
diff --git a/sys/mips/mips/genassym.c b/sys/mips/mips/genassym.c
index 6cec2b79435..b10a5e2327b 100644
--- a/sys/mips/mips/genassym.c
+++ b/sys/mips/mips/genassym.c
@@ -65,7 +65,7 @@ __FBSDID("$FreeBSD$");
 
 ASSYM(TD_PCB, offsetof(struct thread, td_pcb));
 ASSYM(TD_UPTE, offsetof(struct thread, td_md.md_upte));
-ASSYM(TD_REALKSTACK, offsetof(struct thread, td_md.md_realstack));
+ASSYM(TD_KSTACK, offsetof(struct thread, td_kstack));
 ASSYM(TD_FLAGS, offsetof(struct thread, td_flags));
 ASSYM(TD_LOCK, offsetof(struct thread, td_lock));
 ASSYM(TD_FRAME, offsetof(struct thread, td_frame));
diff --git a/sys/mips/mips/machdep.c b/sys/mips/mips/machdep.c
index 6ca4ec9409a..b3f36e53fe6 100644
--- a/sys/mips/mips/machdep.c
+++ b/sys/mips/mips/machdep.c
@@ -298,14 +298,13 @@ mips_proc0_init(void)
 		(long)kstack0));
 	thread0.td_kstack = kstack0;
 	thread0.td_kstack_pages = KSTACK_PAGES;
-	thread0.td_md.md_realstack = roundup2(thread0.td_kstack, PAGE_SIZE * 2);
 	/* 
 	 * Do not use cpu_thread_alloc to initialize these fields 
 	 * thread0 is the only thread that has kstack located in KSEG0 
 	 * while cpu_thread_alloc handles kstack allocated in KSEG2.
 	 */
-	thread0.td_pcb = (struct pcb *)(thread0.td_md.md_realstack +
-	    (thread0.td_kstack_pages - 1) * PAGE_SIZE) - 1;
+	thread0.td_pcb = (struct pcb *)(thread0.td_kstack +
+	    thread0.td_kstack_pages * PAGE_SIZE) - 1;
 	thread0.td_frame = &thread0.td_pcb->pcb_regs;
 
 	/* Steal memory for the dynamic per-cpu area. */
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 6bb38d47e3b..e3b7daf2ba7 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -2813,6 +2813,21 @@ pmap_align_superpage(vm_object_t object, vm_ooffset_t offset,
 		*addr = ((*addr + SEGOFSET) & ~SEGOFSET) + superpage_offset;
 }
 
+/*
+ * 	Increase the starting virtual address of the given mapping so
+ * 	that it is aligned to not be the second page in a TLB entry.
+ * 	This routine assumes that the length is appropriately-sized so
+ * 	that the allocation does not share a TLB entry at all if required.
+ */
+void
+pmap_align_tlb(vm_offset_t *addr)
+{
+	if ((*addr & PAGE_SIZE) == 0)
+		return;
+	*addr += PAGE_SIZE;
+	return;
+}
+
 int pmap_pid_dump(int pid);
 
 int
diff --git a/sys/mips/mips/swtch.S b/sys/mips/mips/swtch.S
index 4fc3cbd6cb9..2919d21e5f3 100644
--- a/sys/mips/mips/swtch.S
+++ b/sys/mips/mips/swtch.S
@@ -339,7 +339,7 @@ blocked_loop:
 	sw	a1, PC_CURTHREAD(a3)
 	lw	a2, TD_PCB(a1)
 	sw	a2, PC_CURPCB(a3)
-	lw	v0, TD_REALKSTACK(a1)
+	lw	v0, TD_KSTACK(a1)
 	li	s0, MIPS_KSEG2_START		# If Uarea addr is below kseg2,
 	bltu	v0, s0, sw2			# no need to insert in TLB.
 	lw	a1, TD_UPTE+0(s7)		# t0 = first u. pte
diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index cb8a8d90a1a..d385fddd9c4 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -217,13 +217,9 @@ cpu_thread_swapin(struct thread *td)
 	 * part of the thread struct so cpu_switch() can quickly map in
 	 * the pcb struct and kernel stack.
 	 */
-	if (!(pte = pmap_segmap(kernel_pmap, td->td_md.md_realstack)))
-		panic("cpu_thread_swapin: invalid segmap");
-	pte += ((vm_offset_t)td->td_md.md_realstack >> PAGE_SHIFT) & (NPTEPG - 1);
-
-	for (i = 0; i < KSTACK_PAGES - 1; i++) {
+	for (i = 0; i < KSTACK_PAGES; i++) {
+		pte = pmap_pte(kernel_pmap, td->td_kstack + i * PAGE_SIZE);
 		td->td_md.md_upte[i] = *pte & ~(PTE_RO|PTE_WIRED);
-		pte++;
 	}
 }
 
@@ -238,22 +234,14 @@ cpu_thread_alloc(struct thread *td)
 	pt_entry_t *pte;
 	int i;
 
-	if (td->td_kstack & (1 << PAGE_SHIFT))
-		td->td_md.md_realstack = td->td_kstack + PAGE_SIZE;
-	else
-		td->td_md.md_realstack = td->td_kstack;
-
-	td->td_pcb = (struct pcb *)(td->td_md.md_realstack +
-	    (td->td_kstack_pages - 1) * PAGE_SIZE) - 1;
+	KASSERT((td->td_kstack & (1 << PAGE_SHIFT)) == 0, ("kernel stack must be aligned."));
+	td->td_pcb = (struct pcb *)(td->td_kstack +
+	    td->td_kstack_pages * PAGE_SIZE) - 1;
 	td->td_frame = &td->td_pcb->pcb_regs;
 
-	if (!(pte = pmap_segmap(kernel_pmap, td->td_md.md_realstack)))
-		panic("cpu_thread_alloc: invalid segmap");
-	pte += ((vm_offset_t)td->td_md.md_realstack >> PAGE_SHIFT) & (NPTEPG - 1);
-
-	for (i = 0; i < KSTACK_PAGES - 1; i++) {
+	for (i = 0; i < KSTACK_PAGES; i++) {
+		pte = pmap_pte(kernel_pmap, td->td_kstack + i * PAGE_SIZE);
 		td->td_md.md_upte[i] = *pte & ~(PTE_RO|PTE_WIRED);
-		pte++;
 	}
 }
 
diff --git a/sys/vm/pmap.h b/sys/vm/pmap.h
index 02fda073aea..51a3bbe4c7c 100644
--- a/sys/vm/pmap.h
+++ b/sys/vm/pmap.h
@@ -98,6 +98,9 @@ extern vm_offset_t kernel_vm_end;
 
 void		 pmap_align_superpage(vm_object_t, vm_ooffset_t, vm_offset_t *,
 		    vm_size_t);
+#if defined(__mips__)
+void		 pmap_align_tlb(vm_offset_t *);
+#endif
 void		 pmap_change_wiring(pmap_t, vm_offset_t, boolean_t);
 void		 pmap_clear_modify(vm_page_t m);
 void		 pmap_clear_reference(vm_page_t m);
diff --git a/sys/vm/vm_extern.h b/sys/vm/vm_extern.h
index 83c468e0801..ca6d49c18f2 100644
--- a/sys/vm/vm_extern.h
+++ b/sys/vm/vm_extern.h
@@ -47,6 +47,7 @@ vm_offset_t kmem_alloc_contig(vm_map_t map, vm_size_t size, int flags,
     vm_paddr_t low, vm_paddr_t high, unsigned long alignment,
     unsigned long boundary, vm_memattr_t memattr);
 vm_offset_t kmem_alloc_nofault(vm_map_t, vm_size_t);
+vm_offset_t kmem_alloc_nofault_space(vm_map_t, vm_size_t, int);
 vm_offset_t kmem_alloc_wait(vm_map_t, vm_size_t);
 void kmem_free(vm_map_t, vm_offset_t, vm_size_t);
 void kmem_free_wakeup(vm_map_t, vm_offset_t, vm_size_t);
diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index c080ca04c6c..bcdfd343e1a 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -373,8 +373,17 @@ vm_thread_new(struct thread *td, int pages)
 	/*
 	 * Get a kernel virtual address for this thread's kstack.
 	 */
+#if defined(__mips__)
+	/*
+	 * We need to align the kstack's mapped address to fit within
+	 * a single TLB entry.
+	 */
+	ks = kmem_alloc_nofault_space(kernel_map,
+	    (pages + KSTACK_GUARD_PAGES) * PAGE_SIZE, VMFS_TLB_ALIGNED_SPACE);
+#else
 	ks = kmem_alloc_nofault(kernel_map,
 	   (pages + KSTACK_GUARD_PAGES) * PAGE_SIZE);
+#endif
 	if (ks == 0) {
 		printf("vm_thread_new: kstack allocation failed\n");
 		vm_object_deallocate(ksobj);
diff --git a/sys/vm/vm_kern.c b/sys/vm/vm_kern.c
index 90065721da8..739d289b613 100644
--- a/sys/vm/vm_kern.c
+++ b/sys/vm/vm_kern.c
@@ -118,6 +118,35 @@ kmem_alloc_nofault(map, size)
 	return (addr);
 }
 
+/*
+ *	kmem_alloc_nofault_space:
+ *
+ *	Allocate a virtual address range with no underlying object and
+ *	no initial mapping to physical memory within the specified
+ *	address space.  Any mapping from this range to physical memory
+ *	must be explicitly created prior to its use, typically with
+ *	pmap_qenter().  Any attempt to create a mapping on demand
+ *	through vm_fault() will result in a panic. 
+ */
+vm_offset_t
+kmem_alloc_nofault_space(map, size, find_space)
+	vm_map_t map;
+	vm_size_t size;
+	int find_space;
+{
+	vm_offset_t addr;
+	int result;
+
+	size = round_page(size);
+	addr = vm_map_min(map);
+	result = vm_map_find(map, NULL, 0, &addr, size, find_space,
+	    VM_PROT_ALL, VM_PROT_ALL, MAP_NOFAULT);
+	if (result != KERN_SUCCESS) {
+		return (0);
+	}
+	return (addr);
+}
+
 /*
  *	Allocate wired-down memory in the kernel's address map
  *	or a submap.
diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c
index fe0c0f51378..1d22fa6b116 100644
--- a/sys/vm/vm_map.c
+++ b/sys/vm/vm_map.c
@@ -1394,9 +1394,20 @@ vm_map_find(vm_map_t map, vm_object_t object, vm_ooffset_t offset,
 				vm_map_unlock(map);
 				return (KERN_NO_SPACE);
 			}
-			if (find_space == VMFS_ALIGNED_SPACE)
+			switch (find_space) {
+			case VMFS_ALIGNED_SPACE:
 				pmap_align_superpage(object, offset, addr,
 				    length);
+				break;
+#ifdef VMFS_TLB_ALIGNED_SPACE
+			case VMFS_TLB_ALIGNED_SPACE:
+				pmap_align_tlb(addr);
+				break;
+#endif
+			default:
+				break;
+			}
+
 			start = *addr;
 		}
 		result = vm_map_insert(map, object, offset, start, start +
diff --git a/sys/vm/vm_map.h b/sys/vm/vm_map.h
index 5454ce6ed15..d5c5b511c43 100644
--- a/sys/vm/vm_map.h
+++ b/sys/vm/vm_map.h
@@ -326,6 +326,9 @@ long vmspace_wired_count(struct vmspace *vmspace);
 #define	VMFS_NO_SPACE		0	/* don't find; use the given range */
 #define	VMFS_ANY_SPACE		1	/* find a range with any alignment */
 #define	VMFS_ALIGNED_SPACE	2	/* find a superpage-aligned range */
+#if defined(__mips__)
+#define	VMFS_TLB_ALIGNED_SPACE	3	/* find a TLB entry aligned range */
+#endif
 
 /*
  * vm_map_wire and vm_map_unwire option flags

From 7b7d5b6c5829b4840dd666f5e887122f70c10d29 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 19 Apr 2010 00:18:14 +0000
Subject: [PATCH 131/532] vm_thread_swapout() can safely dirty the page before
 rather than after acquiring the page queues lock.

---
 sys/vm/vm_glue.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index bcdfd343e1a..dbd5065d8b0 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -523,8 +523,8 @@ vm_thread_swapout(struct thread *td)
 		m = vm_page_lookup(ksobj, i);
 		if (m == NULL)
 			panic("vm_thread_swapout: kstack already missing?");
-		vm_page_lock_queues();
 		vm_page_dirty(m);
+		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_unlock_queues();
 	}

From 5f3173b5176c1a702f16c6fb1a021e7e7be405c6 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Mon, 19 Apr 2010 06:01:58 +0000
Subject: [PATCH 132/532] o) Fix XKPHYS physical address extraction.  Also
 define cache coherency    attributes for XKPHYS. o) Make coprocessor 0
 accessor function macros for register+selector registers    take the full
 name so that e.g. (as done in this commit), prid selector 1    can be written
 through mips_wr_ebase() rather than mips_wr_prid1(). o) Allow for sign
 extension of 32-bit segment addresses. o) Remove an unused MIPS-I register
 number.

---
 sys/mips/cavium/octeon_machdep.c | 10 --------
 sys/mips/cavium/octeon_mp.c      |  2 +-
 sys/mips/include/cpufunc.h       | 32 ++++++++++++-------------
 sys/mips/include/cpuregs.h       | 40 ++++++++++++++++++++------------
 sys/mips/mips/genassym.c         |  6 +++++
 5 files changed, 48 insertions(+), 42 deletions(-)

diff --git a/sys/mips/cavium/octeon_machdep.c b/sys/mips/cavium/octeon_machdep.c
index c796071a385..d5df4b426ea 100644
--- a/sys/mips/cavium/octeon_machdep.c
+++ b/sys/mips/cavium/octeon_machdep.c
@@ -86,16 +86,6 @@ static void octeon_boot_params_init(register_t ptr);
 static uint64_t ciu_get_intr_sum_reg_addr(int core_num, int intx, int enx);
 static uint64_t ciu_get_intr_en_reg_addr(int core_num, int intx, int enx);
 
-static __inline void
-mips_wr_ebase(u_int32_t a0)
-{
-	__asm __volatile("mtc0 %[a0], $15, 1 ;"
-	    :
-	    :     [a0] "r"(a0));
-
-	mips_barrier();
-}
-
 void
 platform_cpu_init()
 {
diff --git a/sys/mips/cavium/octeon_mp.c b/sys/mips/cavium/octeon_mp.c
index 85cdc45dadd..8ded87ee282 100644
--- a/sys/mips/cavium/octeon_mp.c
+++ b/sys/mips/cavium/octeon_mp.c
@@ -69,7 +69,7 @@ platform_init_ap(int cpuid)
 	/*
 	 * Set the exception base.
 	 */
-	mips_wr_prid1(0x80000000 | cpuid);
+	mips_wr_ebase(0x80000000 | cpuid);
 
 	/*
 	 * Set up interrupts, clear IPIs and unmask the IPI interrupt.
diff --git a/sys/mips/include/cpufunc.h b/sys/mips/include/cpufunc.h
index 9c34e3152d8..6520671137a 100644
--- a/sys/mips/include/cpufunc.h
+++ b/sys/mips/include/cpufunc.h
@@ -166,7 +166,7 @@ mips_wr_ ## n (uint32_t a0)					\
 
 #define	MIPS_RDRW32_COP0_SEL(n,r,s)					\
 static __inline uint32_t					\
-mips_rd_ ## n ## s(void)						\
+mips_rd_ ## n(void)						\
 {								\
 	int v0;							\
 	__asm __volatile ("mfc0 %[v0], $"__XSTRING(r)", "__XSTRING(s)";"	\
@@ -175,7 +175,7 @@ mips_rd_ ## n ## s(void)						\
 	return (v0);						\
 }								\
 static __inline void						\
-mips_wr_ ## n ## s(uint32_t a0)					\
+mips_wr_ ## n(uint32_t a0)					\
 {								\
 	__asm __volatile ("mtc0 %[a0], $"__XSTRING(r)", "__XSTRING(s)";"	\
 			 __XSTRING(COP0_SYNC)";"		\
@@ -201,9 +201,9 @@ static __inline void mips_sync_icache (void)
 
 MIPS_RDRW32_COP0(compare, MIPS_COP_0_COMPARE);
 MIPS_RDRW32_COP0(config, MIPS_COP_0_CONFIG);
-MIPS_RDRW32_COP0_SEL(config, MIPS_COP_0_CONFIG, 1);
-MIPS_RDRW32_COP0_SEL(config, MIPS_COP_0_CONFIG, 2);
-MIPS_RDRW32_COP0_SEL(config, MIPS_COP_0_CONFIG, 3);
+MIPS_RDRW32_COP0_SEL(config1, MIPS_COP_0_CONFIG, 1);
+MIPS_RDRW32_COP0_SEL(config2, MIPS_COP_0_CONFIG, 2);
+MIPS_RDRW32_COP0_SEL(config3, MIPS_COP_0_CONFIG, 3);
 MIPS_RDRW32_COP0(count, MIPS_COP_0_COUNT);
 MIPS_RDRW32_COP0(index, MIPS_COP_0_TLB_INDEX);
 MIPS_RDRW32_COP0(wired, MIPS_COP_0_TLB_WIRED);
@@ -219,20 +219,20 @@ MIPS_RDRW32_COP0(pagemask, MIPS_COP_0_TLB_PG_MASK);
 #endif
 MIPS_RDRW32_COP0(prid, MIPS_COP_0_PRID);
 /* XXX 64-bit?  */
-MIPS_RDRW32_COP0_SEL(prid, MIPS_COP_0_PRID, 1);
+MIPS_RDRW32_COP0_SEL(ebase, MIPS_COP_0_PRID, 1);
 MIPS_RDRW32_COP0(watchlo, MIPS_COP_0_WATCH_LO);
-MIPS_RDRW32_COP0_SEL(watchlo, MIPS_COP_0_WATCH_LO, 1);
-MIPS_RDRW32_COP0_SEL(watchlo, MIPS_COP_0_WATCH_LO, 2);
-MIPS_RDRW32_COP0_SEL(watchlo, MIPS_COP_0_WATCH_LO, 3);
+MIPS_RDRW32_COP0_SEL(watchlo1, MIPS_COP_0_WATCH_LO, 1);
+MIPS_RDRW32_COP0_SEL(watchlo2, MIPS_COP_0_WATCH_LO, 2);
+MIPS_RDRW32_COP0_SEL(watchlo3, MIPS_COP_0_WATCH_LO, 3);
 MIPS_RDRW32_COP0(watchhi, MIPS_COP_0_WATCH_HI);
-MIPS_RDRW32_COP0_SEL(watchhi, MIPS_COP_0_WATCH_HI, 1);
-MIPS_RDRW32_COP0_SEL(watchhi, MIPS_COP_0_WATCH_HI, 2);
-MIPS_RDRW32_COP0_SEL(watchhi, MIPS_COP_0_WATCH_HI, 3);
+MIPS_RDRW32_COP0_SEL(watchhi1, MIPS_COP_0_WATCH_HI, 1);
+MIPS_RDRW32_COP0_SEL(watchhi2, MIPS_COP_0_WATCH_HI, 2);
+MIPS_RDRW32_COP0_SEL(watchhi3, MIPS_COP_0_WATCH_HI, 3);
 
-MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 0);
-MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 1);
-MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 2);
-MIPS_RDRW32_COP0_SEL(perfcnt, MIPS_COP_0_PERFCNT, 3);
+MIPS_RDRW32_COP0_SEL(perfcnt0, MIPS_COP_0_PERFCNT, 0);
+MIPS_RDRW32_COP0_SEL(perfcnt1, MIPS_COP_0_PERFCNT, 1);
+MIPS_RDRW32_COP0_SEL(perfcnt2, MIPS_COP_0_PERFCNT, 2);
+MIPS_RDRW32_COP0_SEL(perfcnt3, MIPS_COP_0_PERFCNT, 3);
 
 #undef	MIPS_RDRW32_COP0
 
diff --git a/sys/mips/include/cpuregs.h b/sys/mips/include/cpuregs.h
index 74e789c1416..3f4ffd208ba 100644
--- a/sys/mips/include/cpuregs.h
+++ b/sys/mips/include/cpuregs.h
@@ -78,21 +78,36 @@
  * Caching of mapped addresses is controlled by bits in the TLB entry.
  */
 
-#define	MIPS_KUSEG_START		0x0
-#define	MIPS_KSEG0_START		0x80000000
-#define	MIPS_KSEG0_END			0x9fffffff
-#define	MIPS_KSEG1_START		0xa0000000
-#define	MIPS_KSEG1_END			0xbfffffff
-#define	MIPS_KSSEG_START		0xc0000000
-#define	MIPS_KSSEG_END			0xdfffffff
+#if !defined(_LOCORE)
+#define	MIPS_KUSEG_START		0x00000000
+#define	MIPS_KSEG0_START		((intptr_t)(int32_t)0x80000000)
+#define	MIPS_KSEG0_END			((intptr_t)(int32_t)0x9fffffff)
+#define	MIPS_KSEG1_START		((intptr_t)(int32_t)0xa0000000)
+#define	MIPS_KSEG1_END			((intptr_t)(int32_t)0xbfffffff)
+#define	MIPS_KSSEG_START		((intptr_t)(int32_t)0xc0000000)
+#define	MIPS_KSSEG_END			((intptr_t)(int32_t)0xdfffffff)
+#define	MIPS_KSEG3_START		((intptr_t)(int32_t)0xe0000000)
+#define	MIPS_KSEG3_END			((intptr_t)(int32_t)0xffffffff)
+
 #define MIPS_KSEG2_START		MIPS_KSSEG_START
 #define MIPS_KSEG2_END			MIPS_KSSEG_END
-#define	MIPS_KSEG3_START		0xe0000000
-#define	MIPS_KSEG3_END			0xffffffff
+#endif
+
+#define	MIPS_XKPHYS_START		0x8000000000000000
+#define	MIPS_XKPHYS_END			0xbfffffffffffffff
+
+#define	MIPS_XKPHYS_CCA_UC		0x02	/* Uncached.  */
+#define	MIPS_XKPHYS_CCA_CNC		0x03	/* Cacheable non-coherent.  */
 
 #define	MIPS_PHYS_TO_XKPHYS(cca,x) \
 	((0x2ULL << 62) | ((unsigned long long)(cca) << 59) | (x))
-#define	MIPS_XKPHYS_TO_PHYS(x)	((x) & 0x0effffffffffffffULL)
+#define	MIPS_XKPHYS_TO_PHYS(x)	((x) & 0x07ffffffffffffffULL)
+
+#define	MIPS_XUSEG_START		0x0000000000000000
+#define	MIPS_XUSEG_END			0x0000010000000000
+
+#define	MIPS_XKSEG_START		0xc000000000000000
+#define	MIPS_XKSEG_END			0xc00000ff80000000
 
 /* CPU dependent mtc0 hazard hook */
 #ifdef TARGET_OCTEON
@@ -471,7 +486,6 @@
  *				     (3=32bit, 6=64bit, i=impl dep)
  *  0	MIPS_COP_0_TLB_INDEX	3333 TLB Index.
  *  1	MIPS_COP_0_TLB_RANDOM	3333 TLB Random.
- *  2	MIPS_COP_0_TLB_LOW	3... r3k TLB entry low.
  *  2	MIPS_COP_0_TLB_LO0	.636 r4k TLB entry low.
  *  3	MIPS_COP_0_TLB_LO1	.636 r4k TLB entry low, extended.
  *  4	MIPS_COP_0_TLB_CONTEXT	3636 TLB Context.
@@ -531,10 +545,6 @@
 #define	MIPS_COP_0_EXC_PC	_(14)
 #define	MIPS_COP_0_PRID		_(15)
 
-
-/* MIPS-I */
-#define	MIPS_COP_0_TLB_LOW	_(2)
-
 /* MIPS-III */
 #define	MIPS_COP_0_TLB_LO0	_(2)
 #define	MIPS_COP_0_TLB_LO1	_(3)
diff --git a/sys/mips/mips/genassym.c b/sys/mips/mips/genassym.c
index b10a5e2327b..53f962b69f3 100644
--- a/sys/mips/mips/genassym.c
+++ b/sys/mips/mips/genassym.c
@@ -54,6 +54,7 @@ __FBSDID("$FreeBSD$");
 #include <vm/pmap.h>
 #include <vm/vm_map.h>
 #include <sys/proc.h>
+#include <machine/cpuregs.h>
 #include <machine/pcb.h>
 #include <machine/sigframe.h>
 #include <machine/proc.h>
@@ -91,9 +92,14 @@ ASSYM(SIGF_UC, offsetof(struct sigframe, sf_uc));
 ASSYM(SIGFPE, SIGFPE);
 ASSYM(PAGE_SHIFT, PAGE_SHIFT);
 ASSYM(PAGE_SIZE, PAGE_SIZE);
+ASSYM(PAGE_MASK, PAGE_MASK);
 ASSYM(SEGSHIFT, SEGSHIFT);
 ASSYM(NPTEPG, NPTEPG);
 ASSYM(TDF_NEEDRESCHED, TDF_NEEDRESCHED);
 ASSYM(TDF_ASTPENDING, TDF_ASTPENDING);
 ASSYM(PCPU_SIZE, sizeof(struct pcpu));
 ASSYM(MAXCOMLEN, MAXCOMLEN);
+
+ASSYM(MIPS_KSEG0_START, MIPS_KSEG0_START);
+ASSYM(MIPS_KSEG1_START, MIPS_KSEG1_START);
+ASSYM(MIPS_KSEG2_START, MIPS_KSEG2_START);

From 745bba1ab0e26d708f044f6f3fdce5364e8ab4f9 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Mon, 19 Apr 2010 07:34:26 +0000
Subject: [PATCH 133/532] o) Eliminate the "stand" frame and its use.  Use
 CALLFRAME_* everywhere. o) Use <machine/asm.h> macros for register-width,
 etc., rather than doing it    by hand in a few more assembly files. o) Reduce
 diffs between various bits of TLB refill code in exception.S and    between
 interrupt processing code. o) Use PTR_* to operate on registers that are
 pointers (e.g. sp). o) Add and use a macro, CLEAR_PTE_SWBITS rather than
 using the    mysteriously-named WIRED_SHIFT to select bits to truncate when
 loading PTEs. o) Don't doubly disable interrupts by moving zero to the status
 register,    especially since that has the nasty side-effect of taking us out
 of 64-bit    mode. o) Use CLEAR_STATUS to disable interrupts the first time.
 o) Keep SR_PX set as well as SR_[KSU]X when doing exception processing.  This
    is the bit that determines whether 64-bit operations are allowed. o) Don't
 enable interrupts until configure_final(), like most other ports.

---
 sys/mips/include/regnum.h  |   4 -
 sys/mips/mips/autoconf.c   |   1 +
 sys/mips/mips/exception.S  | 431 ++++++++++++++++---------------------
 sys/mips/mips/fp.S         |  10 +-
 sys/mips/mips/machdep.c    |   1 -
 sys/mips/mips/swtch.S      | 131 ++++-------
 sys/mips/mips/vm_machdep.c |   6 +-
 7 files changed, 236 insertions(+), 348 deletions(-)

diff --git a/sys/mips/include/regnum.h b/sys/mips/include/regnum.h
index baa60bd8b86..1c22bb96961 100644
--- a/sys/mips/include/regnum.h
+++ b/sys/mips/include/regnum.h
@@ -42,10 +42,6 @@
 #ifndef _MACHINE_REGNUM_H_
 #define	_MACHINE_REGNUM_H_
 
-#define	STAND_ARG_SIZE		16
-#define	STAND_FRAME_SIZE	24
-#define	STAND_RA_OFFSET		20
-
 /* This must match the numbers
  * in pcb.h and is used by
  * swtch.S
diff --git a/sys/mips/mips/autoconf.c b/sys/mips/mips/autoconf.c
index 99fd541927f..e16c05c147d 100644
--- a/sys/mips/mips/autoconf.c
+++ b/sys/mips/mips/autoconf.c
@@ -102,6 +102,7 @@ static void
 configure_final(dummy)
 	void *dummy;
 {
+	intr_enable();
 
 	cninit_finish(); 
 
diff --git a/sys/mips/mips/exception.S b/sys/mips/mips/exception.S
index aa775ea5d83..aa2a0a6b5e9 100644
--- a/sys/mips/mips/exception.S
+++ b/sys/mips/mips/exception.S
@@ -66,37 +66,14 @@
 
 #include "assym.s"
 
-#if defined(ISA_MIPS32)
-#undef WITH_64BIT_CP0
-#elif defined(ISA_MIPS64)
-#define WITH_64BIT_CP0
-#elif defined(ISA_MIPS3)
-#define WITH_64BIT_CP0
-#else
-#error "Please write the code for this ISA"
-#endif
+/*
+ * Clear the software-managed bits in a PTE in register pr.
+ */
+#define	CLEAR_PTE_SWBITS(pr)							\
+	sll		pr, 2 ;							\
+	srl		pr, 2			# keep bottom 30 bits
 
-#ifdef WITH_64BIT_CP0
-#define _SLL	dsll
-#define	_SRL	dsrl
-#define	_MFC0	dmfc0
-#define	_MTC0	dmtc0
-#define WIRED_SHIFT 34
-#else
-#define _SLL	sll
-#define	_SRL	srl
-#define	_MFC0	mfc0
-#define	_MTC0	mtc0
-#define WIRED_SHIFT 2
-#endif
 	.set	noreorder			# Noreorder is default style!
-#if defined(ISA_MIPS32)
-	.set	mips32
-#elif defined(ISA_MIPS64)
-	.set	mips64
-#elif defined(ISA_MIPS3)
-	.set	mips3
-#endif
 
 /*
  * Reasonable limit
@@ -125,12 +102,12 @@
  *
  *
  */
-
-	.set	noat
 VECTOR(MipsTLBMiss, unknown)
-	j	_C_LABEL(MipsDoTLBMiss)
-	mfc0	k0, COP_0_BAD_VADDR		# get the fault address
-	nop
+	.set push
+	.set noat
+	j	MipsDoTLBMiss
+	MFC0	k0, COP_0_BAD_VADDR		# get the fault address
+	.set pop
 VECTOR_END(MipsTLBMiss)
 
 /*
@@ -145,42 +122,40 @@ VECTOR_END(MipsTLBMiss)
  * let the processor trap to load the correct value after service.
  *----------------------------------------------------------------------------
  */
+ 	.set push
+	.set noat
 MipsDoTLBMiss:
-						#k0 already has BadVA
-	bltz	k0, 1f				#02: k0<0 -> 1f (kernel fault)
-	srl	k0, k0, SEGSHIFT - 2		#03: k0=seg offset (almost)
+	bltz		k0, 1f				#02: k0<0 -> 1f (kernel fault)
+	PTR_SRL		k0, k0, SEGSHIFT - 2		#03: k0=seg offset (almost)
+
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_SEGBASE(k1)
-	beqz	k1, 2f			      #05: make sure segbase is not null
-	andi	k0, k0, 0x7fc			#06: k0=seg offset (mask 0x3)
-#xxx mips64 unsafe?
-	addu	k1, k0, k1			#07: k1=seg entry address
-	lw	k1, 0(k1)			#08: k1=seg entry
-	mfc0	k0, COP_0_BAD_VADDR		#09: k0=bad address (again)
-	beq	k1, zero, 2f			#0a: ==0 -- no page table
-	srl	k0, PAGE_SHIFT - 2			#0b: k0=VPN (aka va>>10)
+	PTR_L		k1, PC_SEGBASE(k1)
+	beqz		k1, 2f				#05: make sure segbase is not null
+	andi		k0, k0, 0xffc			#06: k0=seg offset (mask 0x3)
+	PTR_ADDU	k1, k0, k1			#07: k1=seg entry address
 
-	andi	k0, k0, ((NPTEPG/2) - 1) << 3	#0c: k0=page tab offset
-#xxx mips64 unsafe?
-	addu	k1, k1, k0			#0d: k1=pte address
-	lw	k0, 0(k1)			#0e: k0=lo0 pte
-	lw	k1, 4(k1)			#0f: k1=lo1 pte
-	_SLL	k0, k0, WIRED_SHIFT		#10: keep bottom 30 bits
-	_SRL	k0, k0, WIRED_SHIFT		#11: keep bottom 30 bits
-	_MTC0	k0, COP_0_TLB_LO0		#12: lo0 is loaded
-	_SLL	k1, k1, WIRED_SHIFT		#13: keep bottom 30 bits
-	_SRL	k1, k1, WIRED_SHIFT		#14: keep bottom 30 bits
-	_MTC0	k1, COP_0_TLB_LO1		#15: lo1 is loaded
+	PTR_L		k1, 0(k1)			#08: k1=seg entry
+	MFC0		k0, COP_0_BAD_VADDR		#09: k0=bad address (again)
+	beq		k1, zero, 2f			#0a: ==0 -- no page table
+	srl		k0, PAGE_SHIFT - 2		#0b: k0=VPN (aka va>>10)
+	andi		k0, k0, 0xff8			#0c: k0=page tab offset
+	PTR_ADDU	k1, k1, k0			#0d: k1=pte address
+	lw		k0, 0(k1)			#0e: k0=lo0 pte
+	lw		k1, 4(k1)			#0f: k1=lo0 pte
+	CLEAR_PTE_SWBITS(k0)
+	MTC0		k0, COP_0_TLB_LO0		#12: lo0 is loaded
+	COP0_SYNC
+	CLEAR_PTE_SWBITS(k1)
+	MTC0		k1, COP_0_TLB_LO1		#15: lo1 is loaded
+	COP0_SYNC
+	tlbwr						#1a: write to tlb
 	HAZARD_DELAY
-	tlbwr					#1a: write to tlb
-	HAZARD_DELAY
-	eret					#1f: retUrn from exception
-1:	j _C_LABEL(MipsTLBMissException)	#20: kernel exception
-	nop					#21: branch delay slot
-2:	j	SlowFault			#22: no page table present
-	nop					#23: branch delay slot
-
-	.set	at
+	eret						#1f: retUrn from exception
+1:	j		MipsTLBMissException		#20: kernel exception
+	nop						#21: branch delay slot
+2:	j		SlowFault			#22: no page table present
+	nop						#23: branch delay slot
+	.set pop
 
 /*
  * This code is copied to the general exception vector address to
@@ -207,7 +182,7 @@ VECTOR(MipsException, unknown)
 						#  the cause is already
 						#  shifted left by 2 bits so
 						#  we dont have to shift.
-	lw	k0, 0(k0)			# Get the function address
+	PTR_L	k0, 0(k0)			# Get the function address
 	nop
 	j	k0				# Jump to the function.
 	nop
@@ -244,20 +219,9 @@ SlowFault:
  *
  *----------------------------------------------------------------------------
  */
-#if defined(ISA_MIPS32)
-#define	STORE	sw	/* 32 bit mode regsave instruction */
-#define	LOAD	lw	/* 32 bit mode regload instruction */
-#define	RSIZE	4	/* 32 bit mode register size */
-#elif defined(ISA_MIPS64)
-#define	STORE	sd	/* 64 bit mode regsave instruction */
-#define	LOAD	ld	/* 64 bit mode regload instruction */
-#define	RSIZE	8	/* 64 bit mode register size */
-#else
-#error "Please write code for this isa."	
-#endif
 
 #define	SAVE_REG(reg, offs, base) \
-	STORE	reg, STAND_ARG_SIZE + (RSIZE * offs) (base)
+	REG_S	reg, CALLFRAME_SIZ + (SZREG * offs) (base)
 
 #ifdef TARGET_OCTEON
 #define CLEAR_STATUS \
@@ -274,7 +238,7 @@ SlowFault:
 	and     a0, a0, a2              ; \
 	mtc0	a0, COP_0_STATUS_REG
 #endif
-	
+
 /*
  * Save CPU and CP0 register state.
  *
@@ -317,8 +281,8 @@ SlowFault:
 	mfhi	v1			;\
 	mfc0	a0, COP_0_STATUS_REG	;\
 	mfc0	a1, COP_0_CAUSE_REG	;\
-	mfc0	a2, COP_0_BAD_VADDR	;\
-	mfc0	a3, COP_0_EXC_PC	;\
+	MFC0	a2, COP_0_BAD_VADDR	;\
+	MFC0	a3, COP_0_EXC_PC	;\
 	SAVE_REG(v0, MULLO, sp)		;\
 	SAVE_REG(v1, MULHI, sp)		;\
 	SAVE_REG(a0, SR, sp)		;\
@@ -332,20 +296,20 @@ SlowFault:
 	PTR_ADDU v0, sp, KERN_EXC_FRAME_SIZE ;\
 	SAVE_REG(v0, SP, sp)		;\
 	CLEAR_STATUS			;\
-	PTR_ADDU a0, sp, STAND_ARG_SIZE	;\
+	PTR_ADDU a0, sp, CALLFRAME_SIZ	;\
 	ITLBNOPFIX
 
 #define	RESTORE_REG(reg, offs, base) \
-	LOAD	reg, STAND_ARG_SIZE + (RSIZE * offs) (base)
+	REG_L	reg, CALLFRAME_SIZ + (SZREG * offs) (base)
 
 #define	RESTORE_CPU \
-	mtc0	zero,COP_0_STATUS_REG	;\
+	CLEAR_STATUS			;\
 	RESTORE_REG(k0, SR, sp)		;\
 	RESTORE_REG(t0, MULLO, sp)	;\
 	RESTORE_REG(t1, MULHI, sp)	;\
 	mtlo	t0			;\
 	mthi	t1			;\
-	_MTC0	v0, COP_0_EXC_PC	;\
+	MTC0	v0, COP_0_EXC_PC	;\
 	.set noat		        ;\
 	RESTORE_REG(AT, AST, sp)	;\
 	RESTORE_REG(v0, V0, sp)		;\
@@ -384,13 +348,13 @@ SlowFault:
  * the status register and the multiply lo and high registers.
  * In addition, we set this up for linkage conventions.
  */
-#define	KERN_REG_SIZE		(NUMSAVEREGS * RSIZE)
-#define	KERN_EXC_FRAME_SIZE	(STAND_FRAME_SIZE + KERN_REG_SIZE + 16)
+#define	KERN_REG_SIZE		(NUMSAVEREGS * SZREG)
+#define	KERN_EXC_FRAME_SIZE	(CALLFRAME_SIZ + KERN_REG_SIZE + 16)
 
 NNON_LEAF(MipsKernGenException, KERN_EXC_FRAME_SIZE, ra)
 	.set	noat
-	subu	sp, sp, KERN_EXC_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - KERN_EXC_FRAME_SIZE)
+	PTR_SUBU	sp, sp, KERN_EXC_FRAME_SIZE
+	.mask	0x80000000, (CALLFRAME_RA - KERN_EXC_FRAME_SIZE)
 /*
  *  Save CPU state, building 'frame'.
  */
@@ -401,7 +365,7 @@ NNON_LEAF(MipsKernGenException, KERN_EXC_FRAME_SIZE, ra)
 	PTR_LA	gp, _C_LABEL(_gp)
 	PTR_LA	k0, _C_LABEL(trap)
 	jalr	k0
-	sw	a3, STAND_RA_OFFSET + KERN_REG_SIZE(sp)		# for debugging
+	REG_S	a3, CALLFRAME_RA + KERN_REG_SIZE(sp)		# for debugging
 
 	/*
 	 * Update interrupt mask in saved status register
@@ -410,7 +374,6 @@ NNON_LEAF(MipsKernGenException, KERN_EXC_FRAME_SIZE, ra)
 	 * in trap handler
 	 */
 	mfc0	a0, COP_0_STATUS_REG
-	mtc0	zero, COP_0_STATUS_REG
 	and	a0, a0, SR_INT_MASK
 	RESTORE_REG(a1, SR, sp)
 	and	a1, a1, ~SR_INT_MASK
@@ -424,10 +387,10 @@ END(MipsKernGenException)
 
 
 #define	SAVE_U_PCB_REG(reg, offs, base) \
-	STORE	reg, U_PCB_REGS + (RSIZE * offs) (base)
+	REG_S	reg, U_PCB_REGS + (SZREG * offs) (base)
 
 #define	RESTORE_U_PCB_REG(reg, offs, base) \
-	LOAD	reg, U_PCB_REGS + (RSIZE * offs) (base)
+	REG_L	reg, U_PCB_REGS + (SZREG * offs) (base)
 
 /*----------------------------------------------------------------------------
  *
@@ -443,14 +406,14 @@ END(MipsKernGenException)
  *
  *----------------------------------------------------------------------------
  */
-NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
+NNON_LEAF(MipsUserGenException, CALLFRAME_SIZ, ra)
 	.set	noat
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
 /*
  * Save all of the registers except for the kernel temporaries in u.u_pcb.
  */
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_CURPCB(k1)
+	PTR_L	k1, PC_CURPCB(k1)
 	SAVE_U_PCB_REG(AT, AST, k1)
 	.set	at
 	SAVE_U_PCB_REG(v0, V0, k1)
@@ -476,17 +439,17 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
 	SAVE_U_PCB_REG(s2, S2, k1)
 	SAVE_U_PCB_REG(s3, S3, k1)
 	SAVE_U_PCB_REG(s4, S4, k1)
-	mfc0	a2, COP_0_BAD_VADDR		# Third arg is the fault addr
+	MFC0	a2, COP_0_BAD_VADDR		# Third arg is the fault addr
 	SAVE_U_PCB_REG(s5, S5, k1)
 	SAVE_U_PCB_REG(s6, S6, k1)
 	SAVE_U_PCB_REG(s7, S7, k1)
 	SAVE_U_PCB_REG(t8, T8, k1)
-	mfc0	a3, COP_0_EXC_PC		# Fourth arg is the pc.
+	MFC0	a3, COP_0_EXC_PC		# Fourth arg is the pc.
 	SAVE_U_PCB_REG(t9, T9, k1)
 	SAVE_U_PCB_REG(gp, GP, k1)
 	SAVE_U_PCB_REG(sp, SP, k1)
 	SAVE_U_PCB_REG(s8, S8, k1)
-	subu	sp, k1, STAND_FRAME_SIZE	 # switch to kernel SP
+	PTR_SUBU	sp, k1, CALLFRAME_SIZ	 # switch to kernel SP
 	SAVE_U_PCB_REG(ra, RA, k1)
 	SAVE_U_PCB_REG(v0, MULLO, k1)
 	SAVE_U_PCB_REG(v1, MULHI, k1)
@@ -494,12 +457,12 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
 	SAVE_U_PCB_REG(a1, CAUSE, k1)
 	SAVE_U_PCB_REG(a2, BADVADDR, k1)
 	SAVE_U_PCB_REG(a3, PC, k1)
-	sw	a3, STAND_RA_OFFSET(sp)	# for debugging
+	REG_S	a3, CALLFRAME_RA(sp)	# for debugging
 	PTR_LA	gp, _C_LABEL(_gp)	# switch to kernel GP
 # Turn off fpu and enter kernel mode
 	and	t0, a0, ~(SR_COP_1_BIT | SR_EXL | SR_KSU_MASK | SR_INT_ENAB)
 #ifdef TARGET_OCTEON
-	or      t0, t0, (MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX)
+	or      t0, t0, (MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX | MIPS32_SR_PX)
 #endif	
 	mtc0	t0, COP_0_STATUS_REG
 	PTR_ADDU a0, k1, U_PCB_REGS
@@ -518,10 +481,7 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
  */
 	DO_AST
 
-	mfc0	t0, COP_0_STATUS_REG	# disable int
-	and	t0, t0, ~(MIPS_SR_INT_IE)
-	mtc0	t0, COP_0_STATUS_REG
-	ITLBNOPFIX
+	CLEAR_STATUS
 
 /*
  * The use of k1 for storing the PCB pointer must be done only
@@ -529,7 +489,7 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
  * by the interrupt code.
  */
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_CURPCB(k1)
+	PTR_L	k1, PC_CURPCB(k1)
 
 	/*
 	 * Update interrupt mask in saved status register
@@ -549,7 +509,7 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
 	mthi	t1
 	RESTORE_U_PCB_REG(a0, PC, k1)
 	RESTORE_U_PCB_REG(v0, V0, k1)
-        _MTC0	a0, COP_0_EXC_PC	# set return address
+        MTC0	a0, COP_0_EXC_PC	# set return address
 	RESTORE_U_PCB_REG(v1, V1, k1)
 	RESTORE_U_PCB_REG(a0, A0, k1)
 	RESTORE_U_PCB_REG(a1, A1, k1)
@@ -578,9 +538,6 @@ NNON_LEAF(MipsUserGenException, STAND_FRAME_SIZE, ra)
 	RESTORE_U_PCB_REG(k0, SR, k1)
 	RESTORE_U_PCB_REG(s8, S8, k1)
 	RESTORE_U_PCB_REG(ra, RA, k1)
-#ifdef TARGET_OCTEON
-	and k0, k0, ~(MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX)
-#endif
 	.set noat
 	RESTORE_U_PCB_REG(AT, AST, k1)
 
@@ -610,27 +567,25 @@ END(MipsUserGenException)
 
 NNON_LEAF(MipsKernIntr, KERN_EXC_FRAME_SIZE, ra)
 	.set	noat
-	subu	sp, sp, KERN_EXC_FRAME_SIZE
-	.mask	0x80000000, (STAND_RA_OFFSET - KERN_EXC_FRAME_SIZE)
+	PTR_SUBU	sp, sp, KERN_EXC_FRAME_SIZE
+	.mask	0x80000000, (CALLFRAME_RA - KERN_EXC_FRAME_SIZE)
 /*
- * Save the relevant kernel registers onto the stack.
+ *  Save CPU state, building 'frame'.
  */
 	SAVE_CPU
-        
 /*
- * Call the interrupt handler.
+ *  Call the interrupt handler. a0 points at the saved frame.
  */
 	PTR_LA	gp, _C_LABEL(_gp)
-	PTR_ADDU a0, sp, STAND_ARG_SIZE
 	PTR_LA	k0, _C_LABEL(cpu_intr)
 	jalr	k0
-	sw	a3, STAND_RA_OFFSET + KERN_REG_SIZE(sp)
-	/* Why no AST processing here? */
+	REG_S	a3, CALLFRAME_RA + KERN_REG_SIZE(sp)		# for debugging
 
 	/*
 	 * Update interrupt mask in saved status register
 	 * Some of interrupts could be disabled by
-	 * intr filters
+	 * intr filters if interrupts are enabled later
+	 * in trap handler
 	 */
 	mfc0	a0, COP_0_STATUS_REG
 	and	a0, a0, SR_INT_MASK
@@ -638,12 +593,8 @@ NNON_LEAF(MipsKernIntr, KERN_EXC_FRAME_SIZE, ra)
 	and	a1, a1, ~SR_INT_MASK
 	or	a1, a1, a0
 	SAVE_REG(a1, SR, sp)
-
-/*
- * Restore registers and return from the interrupt.
- */
-	lw	v0, STAND_RA_OFFSET + KERN_REG_SIZE(sp)
-	RESTORE_CPU
+	REG_L	v0, CALLFRAME_RA + KERN_REG_SIZE(sp)
+	RESTORE_CPU			# v0 contains the return address.
 	sync
 	eret
 	.set	at
@@ -668,15 +619,15 @@ END(MipsKernIntr)
  *
  *----------------------------------------------------------------------------
  */
-NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
+NNON_LEAF(MipsUserIntr, CALLFRAME_SIZ, ra)
 	.set	noat
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
 /*
  * Save the relevant user registers into the u.u_pcb struct.
  * We don't need to save s0 - s8 because the compiler does it for us.
  */
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_CURPCB(k1)
+	PTR_L	k1, PC_CURPCB(k1)
 	SAVE_U_PCB_REG(AT, AST, k1)
 	.set	at
 	SAVE_U_PCB_REG(v0, V0, k1)
@@ -715,19 +666,19 @@ NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
 	mfhi	v1
 	mfc0	a0, COP_0_STATUS_REG
 	mfc0	a1, COP_0_CAUSE_REG
-	mfc0	a3, COP_0_EXC_PC
+	MFC0	a3, COP_0_EXC_PC
 	SAVE_U_PCB_REG(v0, MULLO, k1)
 	SAVE_U_PCB_REG(v1, MULHI, k1)
 	SAVE_U_PCB_REG(a0, SR, k1)
 	SAVE_U_PCB_REG(a1, CAUSE, k1)
 	SAVE_U_PCB_REG(a3, PC, k1)	# PC in a3, note used later!
-	subu	sp, k1, STAND_FRAME_SIZE  # switch to kernel SP
+	PTR_SUBU	sp, k1, CALLFRAME_SIZ  # switch to kernel SP
 	PTR_LA	gp, _C_LABEL(_gp)	# switch to kernel GP
 
 # Turn off fpu, disable interrupts, set kernel mode kernel mode, clear exception level.
 	and	t0, a0, ~(SR_COP_1_BIT | SR_EXL | SR_INT_ENAB | SR_KSU_MASK)
 #ifdef TARGET_OCTEON
-	or      t0, t0, (MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX)
+	or      t0, t0, (MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX | MIPS32_SR_PX)
 #endif	
 	mtc0	t0, COP_0_STATUS_REG
 	ITLBNOPFIX
@@ -737,7 +688,7 @@ NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
  */
 	PTR_LA	k0, _C_LABEL(cpu_intr)
 	jalr	k0
-	sw	a3, STAND_RA_OFFSET(sp)	# for debugging
+	REG_S	a3, CALLFRAME_RA(sp)	# for debugging
 
 /*
  * Enable interrupts before doing ast().
@@ -759,13 +710,10 @@ NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
 /*
  * Restore user registers and return. 
  */
-	mfc0	t0, COP_0_STATUS_REG	# disable int
-	and	t0, t0, ~(MIPS_SR_INT_IE)
-	mtc0	t0, COP_0_STATUS_REG
-	ITLBNOPFIX
+ 	CLEAR_STATUS
 
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_CURPCB(k1)
+	PTR_L	k1, PC_CURPCB(k1)
 
 	/*
 	 * Update interrupt mask in saved status register
@@ -793,7 +741,7 @@ NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
 	RESTORE_U_PCB_REG(t2, PC, k1)
 	mtlo	t0
 	mthi	t1
-	_MTC0	t2, COP_0_EXC_PC	# set return address
+	MTC0	t2, COP_0_EXC_PC	# set return address
 	RESTORE_U_PCB_REG(v0, V0, k1)
 	RESTORE_U_PCB_REG(v1, V1, k1)
 	RESTORE_U_PCB_REG(a0, A0, k1)
@@ -814,9 +762,6 @@ NNON_LEAF(MipsUserIntr, STAND_FRAME_SIZE, ra)
 	RESTORE_U_PCB_REG(k0, SR, k1)
 	RESTORE_U_PCB_REG(sp, SP, k1)
 	RESTORE_U_PCB_REG(ra, RA, k1)
-#ifdef TARGET_OCTEON
-	and k0, k0, ~(MIPS_SR_KX | MIPS_SR_SX | MIPS_SR_UX)
-#endif
 	.set	noat
 	RESTORE_U_PCB_REG(AT, AST, k1)
 
@@ -832,78 +777,78 @@ NLEAF(MipsTLBInvalidException)
 	.set noat
 	.set noreorder
 
-	mfc0	k0, COP_0_BAD_VADDR
-	li	k1, VM_MAXUSER_ADDRESS
-	sltu	k1, k0, k1
-	bnez	k1, 1f
+	MFC0		k0, COP_0_BAD_VADDR
+	PTR_LI		k1, VM_MAXUSER_ADDRESS
+	sltu		k1, k0, k1
+	bnez		k1, 1f
 	nop
 
-	/* badvaddr = kernel address */
-	lui	k1, %hi(_C_LABEL(kernel_segmap))
-	b	2f
-	lw	k1, %lo(_C_LABEL(kernel_segmap))(k1)
+	/* Kernel address.  */
+	lui		k1, %hi(kernel_segmap)		# k1=hi of segbase
+	b		2f
+	PTR_L		k1, %lo(kernel_segmap)(k1)	# k1=segment tab base
 
-1:
-	/* badvaddr = user address */
+1:	/* User address.  */
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_SEGBASE(k1)
+	PTR_L		k1, PC_SEGBASE(k1)
 
-2:
-	beqz	k1, 3f			/* invalid page directory pointer */
+2:	/* Validate page directory pointer.  */
+	beqz		k1, 3f
 	nop
 
-	srl	k0, SEGSHIFT - 2
-	andi	k0, 0xffc
-	addu	k1, k1, k0
-	lw	k1, 0(k1)
-	beqz	k1, 3f			/* invalid page table page pointer */
+	PTR_SRL		k0, SEGSHIFT - 2		# k0=seg offset (almost)
+	beq		k1, zero, MipsKernGenException	# ==0 -- no seg tab
+	andi		k0, k0, 0xffc			# k0=seg offset (mask 0x3)
+	PTR_ADDU	k1, k0, k1			# k1=seg entry address
+	PTR_L		k1, 0(k1)			# k1=seg entry
+
+	/* Validate page table pointer.  */
+	beqz		k1, 3f
 	nop
 
-	mfc0	k0, COP_0_BAD_VADDR
-	srl	k0, PAGE_SHIFT - 2
-	andi	k0, 0xffc
-	addu	k1, k1, k0
+	MFC0		k0, COP_0_BAD_VADDR		# k0=bad address (again)
+	PTR_SRL		k0, PAGE_SHIFT - 2		# k0=VPN
+	andi		k0, k0, 0xffc			# k0=page tab offset
+	PTR_ADDU	k1, k1, k0			# k1=pte address
+	lw		k0, 0(k1)			# k0=this PTE
 
-	lw	k0, 0(k1)
-	andi	k0, PTE_V
-	beqz	k0, 3f			/* invalid page table entry */
+	/* Validate page table entry.  */
+	andi		k0, PTE_V
+	beqz		k0, 3f
 	nop
 
-	andi	k0, k1, 4
-	bnez	k0, odd_page
+	/* Check whether this is an even or odd entry.  */
+	andi		k0, k1, 4
+	bnez		k0, odd_page
 	nop
 
-even_page:
-	lw	k0, 0(k1)
-	_SLL	k0, k0, WIRED_SHIFT
-	_SRL	k0, k0, WIRED_SHIFT
-	_MTC0	k0, COP_0_TLB_LO0
+	lw		k0, 0(k1)
+	lw		k1, 4(k1)
+	CLEAR_PTE_SWBITS(k0)
+	MTC0		k0, COP_0_TLB_LO0
+	COP0_SYNC
+	CLEAR_PTE_SWBITS(k1)
+	MTC0		k1, COP_0_TLB_LO1
+	COP0_SYNC
 
-	lw	k0, 4(k1)
-	_SLL	k0, k0, WIRED_SHIFT
-	_SRL	k0, k0, WIRED_SHIFT
-	_MTC0	k0, COP_0_TLB_LO1
-
-	b	tlb_insert_entry
+	b		tlb_insert_entry
 	nop
 
 odd_page:
-	lw	k0, 0(k1)
-	_SLL	k0, k0, WIRED_SHIFT
-	_SRL	k0, k0, WIRED_SHIFT
-	_MTC0	k0, COP_0_TLB_LO1
-
-	lw	k0, -4(k1)
-	_SLL	k0, k0, WIRED_SHIFT
-	_SRL	k0, k0, WIRED_SHIFT
-	_MTC0	k0, COP_0_TLB_LO0
+	lw		k0, -4(k1)
+	lw		k1, 0(k1)
+	CLEAR_PTE_SWBITS(k0)
+	MTC0		k0, COP_0_TLB_LO0
+	COP0_SYNC
+	CLEAR_PTE_SWBITS(k1)
+	MTC0		k1, COP_0_TLB_LO1
+	COP0_SYNC
 
 tlb_insert_entry:
 	tlbp
 	HAZARD_DELAY
-	mfc0	k0, COP_0_TLB_INDEX
-	HAZARD_DELAY
-	bltz	k0, tlb_insert_random
+	mfc0		k0, COP_0_TLB_INDEX
+	bltz		k0, tlb_insert_random
 	nop
 	tlbwi
 	eret
@@ -927,8 +872,8 @@ tlb_insert_random:
 	 * Check for kernel stack overflow.
 	 */
 	GET_CPU_PCPU(k1)
-	lw	k0, PC_CURTHREAD(k1)
-	lw	k0, TD_KSTACK(k0)
+	PTR_L	k0, PC_CURTHREAD(k1)
+	PTR_L	k0, TD_KSTACK(k0)
 	sltu	k0, k0, sp
 	bnez	k0, _C_LABEL(MipsKernGenException)
 	nop
@@ -944,8 +889,8 @@ tlb_insert_random:
 	sll	k1, k1, PAGE_SHIFT + 1
 
 	PTR_LA	k0, _C_LABEL(pcpu_space)
-	addiu	k0, (PAGE_SIZE * 2)
-	addu	k0, k0, k1
+	PTR_ADDU	k0, PAGE_SIZE * 2
+	PTR_ADDU	k0, k0, k1
 
 	/*
 	 * Stash the original value of 'sp' so we can update trapframe later.
@@ -954,12 +899,12 @@ tlb_insert_random:
 	move	k1, sp
 
 	move	sp, k0
-	subu	sp, sp, KERN_EXC_FRAME_SIZE
+	PTR_SUBU	sp, sp, KERN_EXC_FRAME_SIZE
 
 	move	k0, ra
 	move	ra, zero
-	sw	ra, CALLFRAME_RA(sp)	/* stop the ddb backtrace right here */
-	sw	zero, CALLFRAME_SP(sp)
+	REG_S	ra, CALLFRAME_RA(sp)	/* stop the ddb backtrace right here */
+	REG_S	zero, CALLFRAME_SP(sp)
 	move	ra, k0
 
 	SAVE_CPU
@@ -974,8 +919,8 @@ tlb_insert_random:
 	 * Squelch any more overflow checks by setting the stack base to 0.
 	 */
 	GET_CPU_PCPU(k1)
-	lw	k0, PC_CURTHREAD(k1)
-	sw	zero, TD_KSTACK(k0)
+	PTR_L	k0, PC_CURTHREAD(k1)
+	PTR_S	zero, TD_KSTACK(k0)
 
 	move	a1, a0
 	PANIC("kernel stack overflow - trapframe at %p")
@@ -1008,34 +953,30 @@ END(MipsTLBInvalidException)
  */
 NLEAF(MipsTLBMissException)
 	.set	noat
-	mfc0	k0, COP_0_BAD_VADDR		# k0=bad address
-	li	k1, (VM_MAX_KERNEL_ADDRESS)	# check fault address against
-	sltu	k1, k1, k0			# upper bound of kernel_segmap
-	bnez	k1, _C_LABEL(MipsKernGenException)  # out of bound
-	lui	k1, %hi(_C_LABEL(kernel_segmap))  # k1=hi of segbase
-	srl	k0, 20				# k0=seg offset (almost)
-	lw	k1, %lo(_C_LABEL(kernel_segmap))(k1)  # k1=segment tab base
-	beq	k1, zero, _C_LABEL(MipsKernGenException)  # ==0 -- no seg tab
-	andi	k0, k0, 0xffc			# k0=seg offset (mask 0x3)
-#xxx mips64 unsafe
-	addu	k1, k0, k1			# k1=seg entry address
-	lw	k1, 0(k1)			# k1=seg entry
-	mfc0	k0, COP_0_BAD_VADDR		# k0=bad address (again)
-	beq	k1, zero, _C_LABEL(MipsKernGenException)  # ==0 -- no page table
-	srl	k0, 10				# k0=VPN (aka va>>10)
-	andi	k0, k0, 0xff8			# k0=page tab offset
-#xxx mips64 unsafe
-	addu	k1, k1, k0			# k1=pte address
-	lw	k0, 0(k1)			# k0=lo0 pte
-	lw	k1, 4(k1)			# k1=lo1 pte
-	_SLL	k0, WIRED_SHIFT			# chop bits [31..30]
-	_SRL	k0, WIRED_SHIFT			# chop bits [31..30]
-	_MTC0	k0, COP_0_TLB_LO0		# lo0 is loaded
-	_SLL	k1, WIRED_SHIFT			# chop bits [31..30]
-	_SRL	k1, WIRED_SHIFT			# chop bits [31..30]
-	_MTC0	k1, COP_0_TLB_LO1		# lo1 is loaded
-
-	HAZARD_DELAY
+	MFC0		k0, COP_0_BAD_VADDR		# k0=bad address
+	PTR_LI		k1, VM_MAX_KERNEL_ADDRESS	# check fault address against
+	sltu		k1, k1, k0			# upper bound of kernel_segmap
+	bnez		k1, MipsKernGenException	# out of bound
+	lui		k1, %hi(kernel_segmap)		# k1=hi of segbase
+	PTR_SRL		k0, SEGSHIFT - 2		# k0=seg offset (almost)
+	PTR_L		k1, %lo(kernel_segmap)(k1)	# k1=segment tab base
+	beq		k1, zero, MipsKernGenException	# ==0 -- no seg tab
+	andi		k0, k0, 0xffc			# k0=seg offset (mask 0x3)
+	PTR_ADDU	k1, k0, k1			# k1=seg entry address
+	PTR_L		k1, 0(k1)			# k1=seg entry
+	MFC0		k0, COP_0_BAD_VADDR		# k0=bad address (again)
+	beq		k1, zero, MipsKernGenException	# ==0 -- no page table
+	PTR_SRL		k0, PAGE_SHIFT - 2		# k0=VPN
+	andi		k0, k0, 0xff8			# k0=page tab offset
+	PTR_ADDU	k1, k1, k0			# k1=pte address
+	lw		k0, 0(k1)			# k0=lo0 pte
+	lw		k1, 4(k1)			# k1=lo1 pte
+	CLEAR_PTE_SWBITS(k0)
+	MTC0		k0, COP_0_TLB_LO0		# lo0 is loaded
+	COP0_SYNC
+	CLEAR_PTE_SWBITS(k1)
+	MTC0		k1, COP_0_TLB_LO1		# lo1 is loaded
+	COP0_SYNC
 	tlbwr					# write to tlb
 	HAZARD_DELAY
 	eret					# return from exception
@@ -1061,11 +1002,11 @@ END(MipsTLBMissException)
  *
  *----------------------------------------------------------------------------
  */
-NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
+NON_LEAF(MipsFPTrap, CALLFRAME_SIZ, ra)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
 	mfc0	t0, COP_0_STATUS_REG
-	sw	ra, STAND_RA_OFFSET(sp)
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
+	REG_S	ra, CALLFRAME_RA(sp)
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
 
 	or	t1, t0, SR_COP_1_BIT
 	mtc0	t1, COP_0_STATUS_REG
@@ -1086,10 +1027,10 @@ NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
  * The instruction is in the branch delay slot so the branch will have to
  * be emulated to get the resulting PC.
  */
-	sw	a2, STAND_FRAME_SIZE + 8(sp)
+	PTR_S	a2, CALLFRAME_SIZ + 8(sp)
 	GET_CPU_PCPU(a0)
 #mips64 unsafe?
-	lw	a0, PC_CURPCB(a0)
+	PTR_L	a0, PC_CURPCB(a0)
 	PTR_ADDU a0, a0, U_PCB_REGS		# first arg is ptr to CPU registers
 	move	a1, a2				# second arg is instruction PC
 	move	a2, t1				# third arg is floating point CSR
@@ -1100,7 +1041,7 @@ NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
  * Now load the floating-point instruction in the branch delay slot
  * to be emulated.
  */
-	lw	a2, STAND_FRAME_SIZE + 8(sp)	# restore EXC pc
+	PTR_L	a2, CALLFRAME_SIZ + 8(sp)	# restore EXC pc
 	b	2f
 	lw	a0, 4(a2)			# a0 = coproc instruction
 /*
@@ -1110,10 +1051,10 @@ NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
 1:
 	lw	a0, 0(a2)			# a0 = coproc instruction
 #xxx mips64 unsafe?
-	addu	v0, a2, 4			# v0 = next pc
+	PTR_ADDU	v0, a2, 4			# v0 = next pc
 2:
 	GET_CPU_PCPU(t2)
-	lw	t2, PC_CURPCB(t2)
+	PTR_L	t2, PC_CURPCB(t2)
 	SAVE_U_PCB_REG(v0, PC, t2)		# save new pc
 /*
  * Check to see if the instruction to be emulated is a floating-point
@@ -1127,7 +1068,7 @@ NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
  */
 3:
 	GET_CPU_PCPU(a0)
-	lw	a0, PC_CURTHREAD(a0)		# get current thread
+	PTR_L	a0, PC_CURTHREAD(a0)		# get current thread
 	cfc1	a2, FPC_CSR			# code = FP execptions
 	ctc1	zero, FPC_CSR			# Clear exceptions
 	PTR_LA	t3, _C_LABEL(trapsignal)
@@ -1149,12 +1090,12 @@ NON_LEAF(MipsFPTrap, STAND_FRAME_SIZE, ra)
  */
 FPReturn:
 	mfc0	t0, COP_0_STATUS_REG
-	lw	ra, STAND_RA_OFFSET(sp)
+	PTR_L	ra, CALLFRAME_RA(sp)
 	and	t0, t0, ~SR_COP_1_BIT
 	mtc0	t0, COP_0_STATUS_REG
 	ITLBNOPFIX
 	j	ra
-	PTR_ADDU sp, sp, STAND_FRAME_SIZE
+	PTR_ADDU sp, sp, CALLFRAME_SIZ
 END(MipsFPTrap)
 
 /*
@@ -1182,7 +1123,7 @@ VECTOR(MipsCache, unknown)
 	PTR_LA	k0, _C_LABEL(MipsCacheException)
 	li	k1, MIPS_PHYS_MASK
 	and	k0, k1
-	li	k1, MIPS_KSEG1_START
+	PTR_LI	k1, MIPS_KSEG1_START
 	or	k0, k1
 	j	k0
 	nop
@@ -1200,16 +1141,16 @@ NESTED_NOPROFILE(MipsCacheException, KERN_EXC_FRAME_SIZE, ra)
 	.mask	0x80000000, -4
 	PTR_LA	k0, _C_LABEL(panic)		# return to panic
 	PTR_LA	a0, 9f				# panicstr
-	_MFC0	a1, COP_0_ERROR_PC
+	MFC0	a1, COP_0_ERROR_PC
 	mfc0	a2, COP_0_CACHE_ERR		# 3rd arg cache error
 
-	_MTC0	k0, COP_0_ERROR_PC		# set return address
+	MTC0	k0, COP_0_ERROR_PC		# set return address
 
 	mfc0	k0, COP_0_STATUS_REG		# restore status
 	li	k1, SR_DIAG_DE			# ignore further errors
 	or	k0, k1
 	mtc0	k0, COP_0_STATUS_REG		# restore status
-	HAZARD_DELAY
+	COP0_SYNC
 
 	eret
 
diff --git a/sys/mips/mips/fp.S b/sys/mips/mips/fp.S
index afe9f03d1e0..ce1702eb225 100644
--- a/sys/mips/mips/fp.S
+++ b/sys/mips/mips/fp.S
@@ -94,9 +94,9 @@
  *
  *----------------------------------------------------------------------------
  */
-NON_LEAF(MipsEmulateFP, STAND_FRAME_SIZE, ra)
-	subu	sp, sp, STAND_FRAME_SIZE
-	sw	ra, STAND_RA_OFFSET(sp)
+NON_LEAF(MipsEmulateFP, CALLFRAME_SIZ, ra)
+	subu	sp, sp, CALLFRAME_SIZ
+	sw	ra, CALLFRAME_RA(sp)
 /*
  * Decode the FMT field (bits 24-21) and FUNCTION field (bits 5-0).
  */
@@ -2247,8 +2247,8 @@ result_fs_d:					# result is FS
 	jal	set_fd_d			# save result (in t0,t1,t2,t3)
 
 done:
-	lw	ra, STAND_RA_OFFSET(sp)
-	addu	sp, sp, STAND_FRAME_SIZE
+	lw	ra, CALLFRAME_RA(sp)
+	addu	sp, sp, CALLFRAME_SIZ
 	j	ra
 END(MipsEmulateFP)
 
diff --git a/sys/mips/mips/machdep.c b/sys/mips/mips/machdep.c
index b3f36e53fe6..a9c484c806e 100644
--- a/sys/mips/mips/machdep.c
+++ b/sys/mips/mips/machdep.c
@@ -369,7 +369,6 @@ mips_vector_init(void)
 	 * when handler is installed for it
 	 */
 	set_intr_mask(ALL_INT_MASK);
-	intr_enable();
 
 	/* Clear BEV in SR so we start handling our own exceptions */
 	mips_wr_status(mips_rd_status() & ~SR_BOOT_EXC_VEC);
diff --git a/sys/mips/mips/swtch.S b/sys/mips/mips/swtch.S
index 2919d21e5f3..f287476a3b3 100644
--- a/sys/mips/mips/swtch.S
+++ b/sys/mips/mips/swtch.S
@@ -65,53 +65,7 @@
 
 #include "assym.s"
 
-#if defined(ISA_MIPS32)
-#undef WITH_64BIT_CP0
-#elif defined(ISA_MIPS64)
-#define WITH_64BIT_CP0
-#elif defined(ISA_MIPS3)
-#define WITH_64BIT_CP0
-#else
-#error "Please write the code for this ISA"
-#endif
-
-#ifdef WITH_64BIT_CP0
-#define _SLL	dsll
-#define	_SRL	dsrl
-#define	_MFC0	dmfc0
-#define	_MTC0	dmtc0
-#define WIRED_SHIFT 34
-#else
-#define _SLL	sll
-#define	_SRL	srl
-#define	_MFC0	mfc0
-#define	_MTC0	mtc0
-#define WIRED_SHIFT 2
-#endif
 	.set	noreorder			# Noreorder is default style!
-#if defined(ISA_MIPS32)
-	.set	mips32
-#elif defined(ISA_MIPS64)
-	.set	mips64
-#elif defined(ISA_MIPS3)
-	.set	mips3
-#endif
-
-#if defined(ISA_MIPS32)
-#define	STORE		sw		/* 32 bit mode regsave instruction */
-#define	LOAD		lw		/* 32 bit mode regload instruction */
-#define	RSIZE		4		/* 32 bit mode register size */
-#define	STORE_FP	swc1		/* 32 bit mode fp regsave instruction */
-#define	LOAD_FP		lwc1		/* 32 bit mode fp regload instruction */
-#define	FP_RSIZE	4		/* 32 bit mode fp register size */
-#else
-#define	STORE		sd		/* 64 bit mode regsave instruction */
-#define	LOAD		ld		/* 64 bit mode regload instruction */
-#define	RSIZE		8		/* 64 bit mode register size */
-#define	STORE_FP	sdc1		/* 64 bit mode fp regsave instruction */
-#define	LOAD_FP		ldc1		/* 64 bit mode fp regload instruction */
-#define	FP_RSIZE	8		/* 64 bit mode fp register size */
-#endif
 
 /*
  * FREEBSD_DEVELOPERS_FIXME
@@ -125,28 +79,28 @@
 #endif
 
 #define	SAVE_U_PCB_REG(reg, offs, base) \
-	STORE	reg, U_PCB_REGS + (RSIZE * offs) (base)
+	REG_S	reg, U_PCB_REGS + (SZREG * offs) (base)
 
 #define	RESTORE_U_PCB_REG(reg, offs, base) \
-	LOAD	reg, U_PCB_REGS + (RSIZE * offs) (base)
+	REG_L	reg, U_PCB_REGS + (SZREG * offs) (base)
 
 #define	SAVE_U_PCB_FPREG(reg, offs, base) \
-	STORE_FP reg, U_PCB_FPREGS + (FP_RSIZE * offs) (base)
+	FP_S	reg, U_PCB_FPREGS + (SZFPREG * offs) (base)
 
 #define	RESTORE_U_PCB_FPREG(reg, offs, base) \
-	LOAD_FP	reg, U_PCB_FPREGS + (FP_RSIZE * offs) (base)
+	FP_L	reg, U_PCB_FPREGS + (SZFPREG * offs) (base)
 
 #define	SAVE_U_PCB_FPSR(reg, offs, base) \
-	STORE	reg, U_PCB_FPREGS + (FP_RSIZE * offs) (base)
+	REG_S	reg, U_PCB_FPREGS + (SZFPREG * offs) (base)
 
 #define	RESTORE_U_PCB_FPSR(reg, offs, base) \
-	LOAD	reg, U_PCB_FPREGS + (FP_RSIZE * offs) (base)
+	REG_L	reg, U_PCB_FPREGS + (SZFPREG * offs) (base)
 
 #define	SAVE_U_PCB_CONTEXT(reg, offs, base) \
-	STORE	reg, U_PCB_CONTEXT + (RSIZE * offs) (base)
+	REG_S	reg, U_PCB_CONTEXT + (SZREG * offs) (base)
 
 #define	RESTORE_U_PCB_CONTEXT(reg, offs, base) \
-	LOAD	reg, U_PCB_CONTEXT + (RSIZE * offs) (base)
+	REG_L	reg, U_PCB_CONTEXT + (SZREG * offs) (base)
 
 #define	ITLBNOPFIX	nop;nop;nop;nop;nop;nop;nop;nop;nop;nop;
 
@@ -172,7 +126,7 @@ LEAF(fork_trampoline)
  */
 	.set	noat
 	GET_CPU_PCPU(k1)
-	lw	k1, PC_CURPCB(k1)
+	PTR_L	k1, PC_CURPCB(k1)
 
 	RESTORE_U_PCB_REG(t0, MULLO, k1)
 	RESTORE_U_PCB_REG(t1, MULHI, k1)
@@ -181,7 +135,7 @@ LEAF(fork_trampoline)
 	RESTORE_U_PCB_REG(a0, PC, k1)
 	RESTORE_U_PCB_REG(AT, AST, k1)
 	RESTORE_U_PCB_REG(v0, V0, k1)
-	_MTC0	a0, COP_0_EXC_PC	# set return address
+	MTC0	a0, COP_0_EXC_PC	# set return address
 
 	RESTORE_U_PCB_REG(v1, V1, k1)
 	RESTORE_U_PCB_REG(a0, A0, k1)
@@ -265,7 +219,7 @@ END(savectx)
 
 KSEG0TEXT_START;
 
-NON_LEAF(mips_cpu_throw, STAND_FRAME_SIZE, ra)
+NON_LEAF(mips_cpu_throw, CALLFRAME_SIZ, ra)
 	mfc0	t0, COP_0_STATUS_REG		# t0 = saved status register
 	nop
 	nop
@@ -285,7 +239,7 @@ END(mips_cpu_throw)
  *	a2 - mtx
  * Find the highest priority process and resume it.
  */
-NON_LEAF(cpu_switch, STAND_FRAME_SIZE, ra)
+NON_LEAF(cpu_switch, CALLFRAME_SIZ, ra)
 	mfc0	t0, COP_0_STATUS_REG		# t0 = saved status register
 	nop
 	nop
@@ -294,11 +248,11 @@ NON_LEAF(cpu_switch, STAND_FRAME_SIZE, ra)
 	ITLBNOPFIX
 	beqz	a0, mips_sw1
 	move	a3, a0
-	lw	a0, TD_PCB(a0)		# load PCB addr of curproc
+	PTR_L	a0, TD_PCB(a0)		# load PCB addr of curproc
 	SAVE_U_PCB_CONTEXT(sp, PREG_SP, a0)		# save old sp
-	subu	sp, sp, STAND_FRAME_SIZE
-	sw	ra, STAND_RA_OFFSET(sp)
-	.mask	0x80000000, (STAND_RA_OFFSET - STAND_FRAME_SIZE)
+	PTR_SUBU	sp, sp, CALLFRAME_SIZ
+	REG_S	ra, CALLFRAME_RA(sp)
+	.mask	0x80000000, (CALLFRAME_RA - CALLFRAME_SIZ)
 	SAVE_U_PCB_CONTEXT(s0, PREG_S0, a0)		# do a 'savectx()'
 	SAVE_U_PCB_CONTEXT(s1, PREG_S1, a0)
 	SAVE_U_PCB_CONTEXT(s2, PREG_S2, a0)
@@ -321,13 +275,13 @@ getpc:
 	 * to be saved with the other registers do so here.
 	 */
 
-	sw	a2, TD_LOCK(a3)			# Switchout td_lock 
+	PTR_S	a2, TD_LOCK(a3)			# Switchout td_lock 
 
 mips_sw1:
 #if defined(SMP) && defined(SCHED_ULE)
 	PTR_LA	t0, _C_LABEL(blocked_lock)
 blocked_loop:
-	lw	t1, TD_LOCK(a1)
+	PTR_L	t1, TD_LOCK(a1)
 	beq	t0, t1, blocked_loop
 	nop
 #endif
@@ -336,20 +290,16 @@ blocked_loop:
  * Switch to new context.
  */
 	GET_CPU_PCPU(a3)
-	sw	a1, PC_CURTHREAD(a3)
-	lw	a2, TD_PCB(a1)
-	sw	a2, PC_CURPCB(a3)
-	lw	v0, TD_KSTACK(a1)
-	li	s0, MIPS_KSEG2_START		# If Uarea addr is below kseg2,
+	PTR_S	a1, PC_CURTHREAD(a3)
+	PTR_L	a2, TD_PCB(a1)
+	PTR_S	a2, PC_CURPCB(a3)
+	PTR_L	v0, TD_KSTACK(a1)
+#if !defined(__mips_n64)
+	PTR_LI	s0, MIPS_KSEG2_START		# If Uarea addr is below kseg2,
 	bltu	v0, s0, sw2			# no need to insert in TLB.
-	lw	a1, TD_UPTE+0(s7)		# t0 = first u. pte
-	lw	a2, TD_UPTE+4(s7)		# t1 = 2nd u. pte
-	and	s0, v0, PTE_ODDPG
-	beq	s0, zero, entry0
-	nop
-
-	PANIC_KSEG0("USPACE sat on odd page boundary", t1)
-
+#endif
+	lw	a1, TD_UPTE + 0(s7)		# a1 = u. pte #0
+	lw	a2, TD_UPTE + 4(s7)		# a2 = u. pte #1
 /*
  * Wiredown the USPACE of newproc in TLB entry#0.  Check whether target
  * USPACE is already in another place of TLB before that, and if so
@@ -357,31 +307,32 @@ blocked_loop:
  * NOTE: This is hard coded to UPAGES == 2.
  * Also, there should be no TLB faults at this point.
  */
-entry0:
-	mtc0	v0, COP_0_TLB_HI		# VPN = va
+	MTC0	v0, COP_0_TLB_HI		# VPN = va
 	HAZARD_DELAY
 	tlbp					# probe VPN
 	HAZARD_DELAY
 	mfc0	s0, COP_0_TLB_INDEX
-	nop
-pgm:
+	HAZARD_DELAY
+
+	PTR_LI	t1, MIPS_KSEG0_START		# invalidate tlb entry
 	bltz	s0, entry0set
-	li	t1, MIPS_KSEG0_START		# invalidate tlb entry
+	nop
 	sll	s0, PAGE_SHIFT + 1
 	addu	t1, s0
-	mtc0	t1, COP_0_TLB_HI
+	MTC0	t1, COP_0_TLB_HI
 	mtc0	zero, COP_0_TLB_LO0
 	mtc0	zero, COP_0_TLB_LO1
 	HAZARD_DELAY
 	tlbwi
 	HAZARD_DELAY
-	mtc0	v0, COP_0_TLB_HI		# set VPN again
+	MTC0	v0, COP_0_TLB_HI		# set VPN again
+
 entry0set:
 /* SMP!! - Works only for  unshared TLB case - i.e. no v-cpus */
 	mtc0	zero, COP_0_TLB_INDEX		# TLB entry #0
-#	or	a1, PG_G
+	HAZARD_DELAY
 	mtc0	a1, COP_0_TLB_LO0		# upte[0]
-#	or	a2, PG_G
+	HAZARD_DELAY
 	mtc0	a2, COP_0_TLB_LO1		# upte[1]
 	HAZARD_DELAY
 	tlbwi					# set TLB entry #0
@@ -396,7 +347,7 @@ sw2:
 /*
  * Restore registers and return.
  */
-	lw	a0, TD_PCB(s7)
+	PTR_L	a0, TD_PCB(s7)
 	RESTORE_U_PCB_CONTEXT(gp, PREG_GP, a0)
 	RESTORE_U_PCB_CONTEXT(v0, PREG_SR, a0)	# restore kernel context
 	RESTORE_U_PCB_CONTEXT(ra, PREG_RA, a0)
@@ -457,7 +408,7 @@ LEAF(MipsSwitchFPState)
  * First read out the status register to make sure that all FP operations
  * have completed.
  */
-	lw	a0, TD_PCB(a0)			# get pointer to pcb for proc
+	PTR_L	a0, TD_PCB(a0)			# get pointer to pcb for proc
 	cfc1	t0, FPC_CSR			# stall til FP done
 	cfc1	t0, FPC_CSR			# now get status
 	li	t3, ~SR_COP_1_BIT
@@ -567,13 +518,13 @@ END(MipsSwitchFPState)
  *----------------------------------------------------------------------------
  */
 LEAF(MipsSaveCurFPState)
-	lw	a0, TD_PCB(a0)			# get pointer to pcb for thread
+	PTR_L	a0, TD_PCB(a0)			# get pointer to pcb for thread
 	mfc0	t1, COP_0_STATUS_REG		# Disable interrupts and
 	li	t0, SR_COP_1_BIT		#  enable the coprocessor
 	mtc0	t0, COP_0_STATUS_REG
 	ITLBNOPFIX
 	GET_CPU_PCPU(a1)
-	sw	zero, PC_FPCURTHREAD(a1)	# indicate state has been saved
+	PTR_S	zero, PC_FPCURTHREAD(a1)	# indicate state has been saved
 /*
  * First read out the status register to make sure that all FP operations
  * have completed.
diff --git a/sys/mips/mips/vm_machdep.c b/sys/mips/mips/vm_machdep.c
index d385fddd9c4..20243038d2b 100644
--- a/sys/mips/mips/vm_machdep.c
+++ b/sys/mips/mips/vm_machdep.c
@@ -144,7 +144,7 @@ cpu_fork(register struct thread *td1,register struct proc *p2,
 	pcb2->pcb_context[PCB_REG_RA] = (register_t)fork_trampoline;
 	/* Make sp 64-bit aligned */
 	pcb2->pcb_context[PCB_REG_SP] = (register_t)(((vm_offset_t)td2->td_pcb &
-	    ~(sizeof(__int64_t) - 1)) - STAND_FRAME_SIZE);
+	    ~(sizeof(__int64_t) - 1)) - CALLFRAME_SIZ);
 	pcb2->pcb_context[PCB_REG_S0] = (register_t)fork_return;
 	pcb2->pcb_context[PCB_REG_S1] = (register_t)td2;
 	pcb2->pcb_context[PCB_REG_S2] = (register_t)td2->td_frame;
@@ -339,7 +339,7 @@ cpu_set_upcall(struct thread *td, struct thread *td0)
 	pcb2->pcb_context[PCB_REG_RA] = (register_t)fork_trampoline;
 	/* Make sp 64-bit aligned */
 	pcb2->pcb_context[PCB_REG_SP] = (register_t)(((vm_offset_t)td->td_pcb &
-	    ~(sizeof(__int64_t) - 1)) - STAND_FRAME_SIZE);
+	    ~(sizeof(__int64_t) - 1)) - CALLFRAME_SIZ);
 	pcb2->pcb_context[PCB_REG_S0] = (register_t)fork_return;
 	pcb2->pcb_context[PCB_REG_S1] = (register_t)td;
 	pcb2->pcb_context[PCB_REG_S2] = (register_t)td->td_frame;
@@ -386,7 +386,7 @@ cpu_set_upcall_kse(struct thread *td, void (*entry)(void *), void *arg,
 	* in ``See MIPS Run'' by D. Sweetman, p. 269
 	* align stack */
 	sp = ((register_t)(stack->ss_sp + stack->ss_size) & ~0x7) -
-	    STAND_FRAME_SIZE;
+	    CALLFRAME_SIZ;
 
 	/*
 	 * Set the trap frame to point at the beginning of the uts

From 540247e8c1f94ae64703b57e86022df5d63f3ded Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Mon, 19 Apr 2010 07:51:57 +0000
Subject: [PATCH 134/532] Remove unused file.

---
 sys/conf/files.mips     |   1 -
 sys/mips/mips/copystr.S | 171 ----------------------------------------
 2 files changed, 172 deletions(-)
 delete mode 100644 sys/mips/mips/copystr.S

diff --git a/sys/conf/files.mips b/sys/conf/files.mips
index 60ab0c2102d..f88949ce9d8 100644
--- a/sys/conf/files.mips
+++ b/sys/conf/files.mips
@@ -50,7 +50,6 @@ mips/mips/bus_space_generic.c 	standard
 mips/mips/busdma_machdep.c 	standard
 mips/mips/cache.c		standard
 mips/mips/cache_mipsNN.c	standard
-#mips/mips/copystr.S		standard
 mips/mips/db_disasm.c		optional	ddb
 mips/mips/db_interface.c	optional	ddb
 mips/mips/db_trace.c		optional	ddb
diff --git a/sys/mips/mips/copystr.S b/sys/mips/mips/copystr.S
deleted file mode 100644
index 35e79057a53..00000000000
--- a/sys/mips/mips/copystr.S
+++ /dev/null
@@ -1,171 +0,0 @@
-/*      $NetBSD: copy.S,v 1.5 2007/10/17 19:55:37 garbled Exp $ */
-
-/*-
- * Copyright (c) 1992, 1993
- *      The Regents of the University of California.  All rights reserved.
- * 
- * This code is derived from software contributed to Berkeley by
- * Digital Equipment Corporation and Ralph Campbell.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 3. Neither the name of the University nor the names of its contributors
- *    may be used to endorse or promote products derived from this software
- *    without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- * Copyright (C) 1989 Digital Equipment Corporation.
- * Permission to use, copy, modify, and distribute this software and
- * its documentation for any purpose and without fee is hereby granted,
- * provided that the above copyright notice appears in all copies.
- * Digital Equipment Corporation makes no representations about the
- * suitability of this software for any purpose.  It is provided "as is"
- * without express or implied warranty.
- * 
- * from: Header: /sprite/src/kernel/mach/ds3100.md/RCS/loMem.s,
- *      v 1.1 89/07/11 17:55:04 nelson Exp  SPRITE (DECWRL)
- * from: Header: /sprite/src/kernel/mach/ds3100.md/RCS/machAsm.s,
- *      v 9.2 90/01/29 18:00:39 shirriff Exp  SPRITE (DECWRL)
- * from: Header: /sprite/src/kernel/vm/ds3100.md/vmPmaxAsm.s,
- *      v 1.1 89/07/10 14:27:41 nelson Exp  SPRITE (DECWRL)
- * 
- *      @(#)locore.s    8.5 (Berkeley) 1/4/94
- */
-	
-#include "assym.s"
-#include <machine/asm.h>
-#include <machine/asmacros.h>
-__FBSDID("$FreeBSD$");
-
-#include <sys/errno.h>
-
-/*
- * copystr(9)
- * <v0>int copystr(<a0>const void *src, <a1>void *dst, <a2>size_t len,
- *                 <a3>size_t *done)
- */
-ENTRY(copystr)
-	.set noreorder
-	.set noat
-	move	v0, zero
-	beqz	a2, 2f
-	move	t1, zero
-1:	subu	a2, 1			/*XXX mips64 unsafe -- long */
-	lbu	t0, 0(a0)
-	PTR_ADDU a0, 1
-	sb	t0, 0(a1)
-	PTR_ADDU a1, 1
-	beqz	t0, 3f /* NULL  - end of string*/
-	addu	t1, 1			/*XXX mips64 unsafe -- long */
-	bnez	a2, 1b
-	nop
-2:	/* ENAMETOOLONG */
-	li	v0, ENAMETOOLONG
-3:	/* done != NULL -> how many bytes were copied */
-	beqz	a3, 4f
-	nop
-	sw	t1, 0(a3)		/*XXX mips64 unsafe -- long */
-4:	jr	ra
-	nop
-	.set reorder
-	.set at
-END(copystr)
-
-/*
- * int copyinstr(void *uaddr, void *kaddr, size_t maxlen, size_t *lencopied)
- * Copy a NIL-terminated string, at most maxlen characters long, from the
- * user's address space.  Return the number of characters copied (including
- * the NIL) in *lencopied.  If the string is too long, return ENAMETOOLONG;
- * else return 0 or EFAULT.
- */
-LEAF(copyinstr)
-	.set noreorder
-	.set noat
-	lw	t2, pcpup
-	lw	v1, PC_CURPCB(t2)
-	PTR_LA	v0, _C_LABEL(copystrerr)
-	blt	a0, zero, _C_LABEL(copystrerr)
-	sw	v0, PCB_ONFAULT(v1)
-	move	t0, a2
-	beq	a2, zero, 4f
-1:
-	lbu	v0, 0(a0)
-	subu	a2, a2, 1		/*xxx mips64 unsafe -- long */
-	beq	v0, zero, 2f
-	sb	v0, 0(a1)
-	PTR_ADDU a0, a0, 1
-	bne	a2, zero, 1b
-	PTR_ADDU a1, a1, 1
-4:
-	li	v0, ENAMETOOLONG
-2:
-	beq	a3, zero, 3f
-	subu	a2, t0, a2		/*xxx mips64 unsafe -- long */
-	sw	a2, 0(a3)		/*xxx mips64 unsafe -- long */
-3:
-	j	ra				# v0 is 0 or ENAMETOOLONG
-	sw	zero, PCB_ONFAULT(v1)
-	.set reorder
-	.set at
-END(copyinstr)
-
-/*
- * int copyoutstr(void *uaddr, void *kaddr, size_t maxlen, size_t *lencopied);
- * Copy a NIL-terminated string, at most maxlen characters long, into the
- * user's address space.  Return the number of characters copied (including
- * the NIL) in *lencopied.  If the string is too long, return ENAMETOOLONG;
- * else return 0 or EFAULT.
- */
-LEAF(copyoutstr)
-	.set noreorder
-	.set noat
-	lw	t2, pcpup
-	lw	v1, PC_CURPCB(t2)
-	PTR_LA	v0, _C_LABEL(copystrerr)
-	blt	a1, zero, _C_LABEL(copystrerr)
-	sw	v0, PCB_ONFAULT(v1)
-	move	t0, a2
-	beq	a2, zero, 4f
-1:
-	lbu	v0, 0(a0)
-	subu	a2, a2, 1		/*xxx mips64 unsafe -- long */
-	beq	v0, zero, 2f
-	sb	v0, 0(a1)
-	PTR_ADDU a0, a0, 1
-	bne	a2, zero, 1b
-	PTR_ADDU a1, a1, 1
-4:
-	li	v0, ENAMETOOLONG
-2:
-	beq	a3, zero, 3f
-	subu	a2, t0, a2		/*xxx mips64 unsafe -- long */
-	sw	a2, 0(a3)		/*xxx mips64 unsafe -- long */
-3:
-	j	ra				# v0 is 0 or ENAMETOOLONG
-	sw	zero, PCB_ONFAULT(v1)
-	.set reorder
-	.set at
-END(copyoutstr)
-
-LEAF(copystrerr)
-	sw	zero, PCB_ONFAULT(v1)
-	j	ra
-	li	v0, EFAULT			# return EFAULT
-END(copystrerr)

From fe7d6b6e6936d303fc397ddf187c083c46007bf7 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Mon, 19 Apr 2010 09:03:34 +0000
Subject: [PATCH 135/532] Fix MALTA64 build.

---
 sys/mips/malta/gt_pci.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/mips/malta/gt_pci.c b/sys/mips/malta/gt_pci.c
index 2f0eadaba4e..237e74227e0 100644
--- a/sys/mips/malta/gt_pci.c
+++ b/sys/mips/malta/gt_pci.c
@@ -109,8 +109,8 @@ struct gt_pci_softc {
 	struct rman		sc_mem_rman;
 	struct rman		sc_io_rman;
 	struct rman		sc_irq_rman;
-	uint32_t		sc_mem;
-	uint32_t		sc_io;
+	unsigned long		sc_mem;
+	bus_space_handle_t	sc_io;
 
 	struct resource		*sc_irq;
 	struct intr_event	*sc_eventstab[ICU_LEN];

From 0e568ab25c4d3bae23b3452fffe9e4263fac9a8d Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Mon, 19 Apr 2010 09:03:36 +0000
Subject: [PATCH 136/532] Partially MFp4 #176265 by pjd@:

 - Properly initialize and destroy system_taskq.
 - Add a dummy implementation of taskq_create_proc().

Note: We do not currently use system_taskq in ZFS so this is mostly a
no-op at this time.  Proper system_taskq initialization is required
by newer ZFS code.

Ok'ed by:	pjd
MFC after:	2 weeks
---
 .../compat/opensolaris/kern/opensolaris_taskq.c | 17 ++++++++++++-----
 1 file changed, 12 insertions(+), 5 deletions(-)

diff --git a/sys/cddl/compat/opensolaris/kern/opensolaris_taskq.c b/sys/cddl/compat/opensolaris/kern/opensolaris_taskq.c
index 584be24100e..1e3b1eff391 100644
--- a/sys/cddl/compat/opensolaris/kern/opensolaris_taskq.c
+++ b/sys/cddl/compat/opensolaris/kern/opensolaris_taskq.c
@@ -52,9 +52,9 @@ static void
 system_taskq_init(void *arg)
 {
 
-	system_taskq = (taskq_t *)taskqueue_thread;
 	taskq_zone = uma_zcreate("taskq_zone", sizeof(struct ostask),
 	    NULL, NULL, NULL, NULL, 0, 0);
+	system_taskq = taskq_create("system_taskq", mp_ncpus, 0, 0, 0, 0);
 }
 SYSINIT(system_taskq_init, SI_SUB_CONFIGURE, SI_ORDER_ANY, system_taskq_init, NULL);
 
@@ -62,6 +62,7 @@ static void
 system_taskq_fini(void *arg)
 {
 
+	taskq_destroy(system_taskq);
 	uma_zdestroy(taskq_zone);
 }
 SYSUNINIT(system_taskq_fini, SI_SUB_CONFIGURE, SI_ORDER_ANY, system_taskq_fini, NULL);
@@ -72,10 +73,8 @@ taskq_create(const char *name, int nthreads, pri_t pri, int minalloc __unused,
 {
 	taskq_t *tq;
 
-	if ((flags & TASKQ_THREADS_CPU_PCT) != 0) {
-		/* TODO: Calculate number od threads. */
-		printf("%s: TASKQ_THREADS_CPU_PCT\n", __func__);
-	}
+	if ((flags & TASKQ_THREADS_CPU_PCT) != 0)
+		nthreads = MAX((mp_ncpus * nthreads) / 100, 1);
 
 	tq = kmem_alloc(sizeof(*tq), KM_SLEEP);
 	tq->tq_queue = taskqueue_create(name, M_WAITOK, taskqueue_thread_enqueue,
@@ -85,6 +84,14 @@ taskq_create(const char *name, int nthreads, pri_t pri, int minalloc __unused,
 	return ((taskq_t *)tq);
 }
 
+taskq_t *
+taskq_create_proc(const char *name, int nthreads, pri_t pri, int minalloc,
+    int maxalloc, proc_t *proc __unused, uint_t flags)
+{
+
+	return (taskq_create(name, nthreads, pri, minalloc, maxalloc, flags));
+}
+
 void
 taskq_destroy(taskq_t *tq)
 {

From 0ff2943ecb402e2acaaec69ebf3707f4f8d0cee0 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Mon, 19 Apr 2010 14:07:33 +0000
Subject: [PATCH 137/532] Revert r206755. It causes some laptops to stop
 booting.

---
 sys/dev/ahci/ahci.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/ahci/ahci.c b/sys/dev/ahci/ahci.c
index a35fbacdab6..5d399683945 100644
--- a/sys/dev/ahci/ahci.c
+++ b/sys/dev/ahci/ahci.c
@@ -126,7 +126,6 @@ static struct {
 	{0x26838086, 0x00, "Intel ESB2",	0},
 	{0x27c18086, 0x00, "Intel ICH7",	0},
 	{0x27c38086, 0x00, "Intel ICH7",	0},
-	{0x27c48086, 0x00, "Intel ICH7M",	0},
 	{0x27c58086, 0x00, "Intel ICH7M",	0},
 	{0x27c68086, 0x00, "Intel ICH7M",	0},
 	{0x28218086, 0x00, "Intel ICH8",	0},

From 307b49efefe0161f8b128f1786ea2169a66aac71 Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Mon, 19 Apr 2010 14:15:58 +0000
Subject: [PATCH 138/532] Get delayed SACK working again.

MFC after: 3 days.
---
 sys/netinet/sctp_indata.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sys/netinet/sctp_indata.c b/sys/netinet/sctp_indata.c
index a7e6c087fb4..97a9733ab84 100644
--- a/sys/netinet/sctp_indata.c
+++ b/sys/netinet/sctp_indata.c
@@ -2123,6 +2123,10 @@ failed_pdapi_express_del:
 		}
 	}
 finish_express_del:
+	if (tsn == (asoc->cumulative_tsn + 1)) {
+		/* Update cum-ack */
+		asoc->cumulative_tsn = tsn;
+	}
 	if (last_chunk) {
 		*m = NULL;
 	}

From 3560b8af1fdb83b0b6c864144eaf0343077bf0cc Mon Sep 17 00:00:00 2001
From: Nathan Whitehorn <nwhitehorn@FreeBSD.org>
Date: Mon, 19 Apr 2010 14:34:44 +0000
Subject: [PATCH 139/532] Fix brokenness in top on big-endian 32-bit systems
 introduced when changing format_k2 to take a long long. Because itoa is
 defined as a K&R C function, without prototyping its arguments, format_k2
 passed a 64-bit value, but itoa() received only the first word, showing '0'
 in all memory fields.

---
 contrib/top/utils.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/contrib/top/utils.c b/contrib/top/utils.c
index a38b2178fcc..43072b1eccb 100644
--- a/contrib/top/utils.c
+++ b/contrib/top/utils.c
@@ -499,7 +499,7 @@ unsigned long long amt;
 	}
     }
 
-    p = strecpy(p, itoa(amt));
+    p = strecpy(p, itoa((int)amt));
     *p++ = tag;
     *p = '\0';
 

From 37133ba7027d430eccfe8b95c22c55c60ad8e4a1 Mon Sep 17 00:00:00 2001
From: Luigi Rizzo <luigi@FreeBSD.org>
Date: Mon, 19 Apr 2010 15:11:45 +0000
Subject: [PATCH 140/532] Slightly different handling of printf/snprintf for
 unaligned uint64_t, which should improve readability, and also to ease the
 port to platforms that do not support %llu

MFC after:	3 days
---
 sbin/ipfw/dummynet.c |  8 +++---
 sbin/ipfw/ipfw2.c    | 66 +++++++++++++++++++++++---------------------
 sbin/ipfw/ipfw2.h    |  2 +-
 3 files changed, 40 insertions(+), 36 deletions(-)

diff --git a/sbin/ipfw/dummynet.c b/sbin/ipfw/dummynet.c
index eb6547a046c..19d52a44b14 100644
--- a/sbin/ipfw/dummynet.c
+++ b/sbin/ipfw/dummynet.c
@@ -203,9 +203,9 @@ list_flow(struct dn_flow *ni)
 		    inet_ntop(AF_INET6, &(id->dst_ip6), buff, sizeof(buff)),
 		    id->dst_port);
 	}
-	printf("%4llu %8llu %2u %4u %3u\n",
-	    align_uint64(&ni->tot_pkts),
-	    align_uint64(&ni->tot_bytes),
+	pr_u64(&ni->tot_pkts, 4);
+	pr_u64(&ni->tot_bytes, 8);
+	printf("%2u %4u %3u\n",
 	    ni->length, ni->len_bytes, ni->drops);
 }
 
@@ -290,8 +290,8 @@ static void
 list_pipes(struct dn_id *oid, struct dn_id *end)
 {
     char buf[160];	/* pending buffer */
-    buf[0] = '\0';
 
+    buf[0] = '\0';
     for (; oid != end; oid = O_NEXT(oid, oid->len)) {
 	if (oid->len < sizeof(*oid))
 		errx(1, "invalid oid len %d\n", oid->len);
diff --git a/sbin/ipfw/ipfw2.c b/sbin/ipfw/ipfw2.c
index 41f6e3ab3e8..b0f76830c40 100644
--- a/sbin/ipfw/ipfw2.c
+++ b/sbin/ipfw/ipfw2.c
@@ -314,22 +314,27 @@ static struct _s_x rule_options[] = {
 	{ NULL, 0 }	/* terminator */
 };
 
-/*  
- * The following is used to generate a printable argument for
- * 64-bit numbers, irrespective of platform alignment and bit size.
- * Because all the printf in this program use %llu as a format,
- * we just return an unsigned long long, which is larger than
- * we need in certain cases, but saves the hassle of using
- * PRIu64 as a format specifier.
- * We don't care about inlining, this is not performance critical code.
+/*
+ * Helper routine to print a possibly unaligned uint64_t on
+ * various platform. If width > 0, print the value with
+ * the desired width, followed by a space;
+ * otherwise, return the required width.
  */
-unsigned long long
-align_uint64(const uint64_t *pll)
+int
+pr_u64(uint64_t *pd, int width)
 {
-	uint64_t ret;
+#ifdef TCC
+#define U64_FMT "I64"
+#else
+#define U64_FMT "llu"
+#endif
+	uint64_t d;
 
-	bcopy (pll, &ret, sizeof(ret));
-	return ret;
+	bcopy (pd, &d, sizeof(d));
+	return (width > 0) ?
+		printf("%*" U64_FMT " ", width, d) :
+		snprintf(NULL, 0, "%" U64_FMT, d) ;
+#undef U64_FMT
 }
 
 void *
@@ -973,9 +978,10 @@ show_ipfw(struct ip_fw *rule, int pcwidth, int bcwidth)
 	}
 	printf("%05u ", rule->rulenum);
 
-	if (pcwidth>0 || bcwidth>0)
-		printf("%*llu %*llu ", pcwidth, align_uint64(&rule->pcnt),
-		    bcwidth, align_uint64(&rule->bcnt));
+	if (pcwidth > 0 || bcwidth > 0) {
+		pr_u64(&rule->pcnt, pcwidth);
+		pr_u64(&rule->bcnt, bcwidth);
+	}
 
 	if (co.do_time == 2)
 		printf("%10u ", rule->timestamp);
@@ -1577,10 +1583,12 @@ show_dyn_ipfw(ipfw_dyn_rule *d, int pcwidth, int bcwidth)
 	}
 	bcopy(&d->rule, &rulenum, sizeof(rulenum));
 	printf("%05d", rulenum);
-	if (pcwidth>0 || bcwidth>0)
-	    printf(" %*llu %*llu (%ds)", pcwidth,
-		align_uint64(&d->pcnt), bcwidth,
-		align_uint64(&d->bcnt), d->expire);
+	if (pcwidth > 0 || bcwidth > 0) {
+		printf(" ");
+		pr_u64(&d->pcnt, pcwidth);
+		pr_u64(&d->bcnt, bcwidth);
+		printf("(%ds)", d->expire);
+	}
 	switch (d->dyn_type) {
 	case O_LIMIT_PARENT:
 		printf(" PARENT %d", d->count);
@@ -1645,9 +1653,9 @@ ipfw_sets_handler(char *av[])
 				free(data);
 			nalloc = nalloc * 2 + 200;
 			nbytes = nalloc;
-		data = safe_calloc(1, nbytes);
-		if (do_cmd(IP_FW_GET, data, (uintptr_t)&nbytes) < 0)
-			err(EX_OSERR, "getsockopt(IP_FW_GET)");
+			data = safe_calloc(1, nbytes);
+			if (do_cmd(IP_FW_GET, data, (uintptr_t)&nbytes) < 0)
+				err(EX_OSERR, "getsockopt(IP_FW_GET)");
 		}
 
 		bcopy(&((struct ip_fw *)data)->next_rule,
@@ -1837,14 +1845,12 @@ ipfw_list(int ac, char *av[], int show_counters)
 				continue;
 
 			/* packet counter */
-			width = snprintf(NULL, 0, "%llu",
-			    align_uint64(&r->pcnt));
+			width = pr_u64(&r->pcnt, 0);
 			if (width > pcwidth)
 				pcwidth = width;
 
 			/* byte counter */
-			width = snprintf(NULL, 0, "%llu",
-			    align_uint64(&r->bcnt));
+			width = pr_u64(&r->bcnt, 0);
 			if (width > bcwidth)
 				bcwidth = width;
 		}
@@ -1858,13 +1864,11 @@ ipfw_list(int ac, char *av[], int show_counters)
 				if (set != co.use_set - 1)
 					continue;
 			}
-			width = snprintf(NULL, 0, "%llu",
-			    align_uint64(&d->pcnt));
+			width = pr_u64(&d->pcnt, 0);
 			if (width > pcwidth)
 				pcwidth = width;
 
-			width = snprintf(NULL, 0, "%llu",
-			    align_uint64(&d->bcnt));
+			width = pr_u64(&d->bcnt, 0);
 			if (width > bcwidth)
 				bcwidth = width;
 		}
diff --git a/sbin/ipfw/ipfw2.h b/sbin/ipfw/ipfw2.h
index d1729841e97..8566cdef021 100644
--- a/sbin/ipfw/ipfw2.h
+++ b/sbin/ipfw/ipfw2.h
@@ -207,7 +207,7 @@ enum tokens {
 #define NEED(_p, msg)      {if (!_p) errx(EX_USAGE, msg);}
 #define NEED1(msg)      {if (!(*av)) errx(EX_USAGE, msg);}
 
-unsigned long long align_uint64(const uint64_t *pll);
+int pr_u64(uint64_t *pd, int width);
 
 /* memory allocation support */
 void *safe_calloc(size_t number, size_t size);

From 3579cf4c4f96448eccadce9da9431aa253c60da5 Mon Sep 17 00:00:00 2001
From: "Kenneth D. Merry" <ken@FreeBSD.org>
Date: Mon, 19 Apr 2010 15:15:36 +0000
Subject: [PATCH 141/532] Don't clear other flags (e.g. CSUM_TCP) when setting
 CSUM_TSO.  This was causing TSO to break for the Xen netfront driver.

Reviewed by:	gibbs, rwatson
MFC after:	7 days
---
 sys/netinet/tcp_output.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/netinet/tcp_output.c b/sys/netinet/tcp_output.c
index ebe5e36ee3f..f9d1b6371e8 100644
--- a/sys/netinet/tcp_output.c
+++ b/sys/netinet/tcp_output.c
@@ -1048,7 +1048,7 @@ send:
 	 * XXX: Fixme: This is currently not the case for IPv6.
 	 */
 	if (tso) {
-		m->m_pkthdr.csum_flags = CSUM_TSO;
+		m->m_pkthdr.csum_flags |= CSUM_TSO;
 		m->m_pkthdr.tso_segsz = tp->t_maxopd - optlen;
 	}
 

From 6ba1ccc0f29eb64b529034ca224ce2c1e8f74fd8 Mon Sep 17 00:00:00 2001
From: Luigi Rizzo <luigi@FreeBSD.org>
Date: Mon, 19 Apr 2010 16:17:30 +0000
Subject: [PATCH 142/532] whitespace fixes (trailing whitespace, bad
 indentation after a merge, etc.)

---
 sys/netinet/ip_dummynet.h        | 54 ++++++++++++++++----------------
 sys/netinet/ipfw/dn_sched.h      |  4 +--
 sys/netinet/ipfw/dn_sched_rr.c   |  6 ++--
 sys/netinet/ipfw/dn_sched_wf2q.c |  8 ++---
 sys/netinet/ipfw/ip_dn_private.h |  2 +-
 sys/netinet/ipfw/ip_dummynet.c   | 36 +++++++++++----------
 sys/netinet/ipfw/ip_fw_pfil.c    |  8 ++---
 sys/netinet/ipfw/ip_fw_private.h |  2 +-
 8 files changed, 61 insertions(+), 59 deletions(-)

diff --git a/sys/netinet/ip_dummynet.h b/sys/netinet/ip_dummynet.h
index 0bbc32638c9..dc2c3412a08 100644
--- a/sys/netinet/ip_dummynet.h
+++ b/sys/netinet/ip_dummynet.h
@@ -87,14 +87,14 @@ enum {
 	DN_SYSCTL_SET,
 
 	DN_LAST,
-} ;
+};
 
 enum { /* subtype for schedulers, flowset and the like */
 	DN_SCHED_UNKNOWN = 0,
 	DN_SCHED_FIFO = 1,
 	DN_SCHED_WF2QP = 2,
 	/* others are in individual modules */
-} ;
+};
 
 enum {	/* user flags */
 	DN_HAVE_MASK	= 0x0001,	/* fs or sched has a mask */
@@ -113,16 +113,16 @@ enum {	/* user flags */
 struct dn_link {
 	struct dn_id oid;
 
-    /*
+	/*
 	 * Userland sets bw and delay in bits/s and milliseconds.
 	 * The kernel converts this back and forth to bits/tick and ticks.
 	 * XXX what about burst ?
-     */
+	 */
 	int32_t		link_nr;
 	int		bandwidth;	/* bit/s or bits/tick.   */
 	int		delay;		/* ms and ticks */
 	uint64_t	burst;		/* scaled. bits*Hz  XXX */
-} ;
+};
 
 /*
  * A flowset, which is a template for flows. Contains parameters
@@ -132,13 +132,13 @@ struct dn_link {
  */
 struct dn_fs {
 	struct dn_id oid;
-	uint32_t fs_nr;	/* the flowset number */
-	uint32_t flags;	/* userland flags */
-    int qsize ;			/* queue size in slots or bytes */
-	int32_t plr;	/* PLR, pkt loss rate (2^31-1 means 100%) */
+	uint32_t fs_nr;		/* the flowset number */
+	uint32_t flags;		/* userland flags */
+	int qsize;		/* queue size in slots or bytes */
+	int32_t plr;		/* PLR, pkt loss rate (2^31-1 means 100%) */
 	uint32_t buckets;	/* buckets used for the queue hash table */
 
-    struct ipfw_flow_id flow_mask ;
+	struct ipfw_flow_id flow_mask;
 	uint32_t sched_nr;	/* the scheduler we attach to */
 	/* generic scheduler parameters. Leave them at -1 if unset.
 	 * Now we use 0: weight, 1: lmax, 2: priority
@@ -149,14 +149,14 @@ struct dn_fs {
 	 * weight and probabilities are in the range 0..1 represented
 	 * in fixed point arithmetic with SCALE_RED decimal bits.
 	 */
-#define SCALE_RED               16
-#define SCALE(x)                ( (x) << SCALE_RED )
-#define SCALE_VAL(x)            ( (x) >> SCALE_RED )
-#define SCALE_MUL(x,y)          ( ( (x) * (y) ) >> SCALE_RED )
-    int w_q ;			/* queue weight (scaled) */
-    int max_th ;		/* maximum threshold for queue (scaled) */
-    int min_th ;		/* minimum threshold for queue (scaled) */
-    int max_p ;			/* maximum value for p_b (scaled) */
+#define SCALE_RED	16
+#define SCALE(x)	( (x) << SCALE_RED )
+#define SCALE_VAL(x)	( (x) >> SCALE_RED )
+#define SCALE_MUL(x,y)	( ( (x) * (y) ) >> SCALE_RED )
+	int w_q ;		/* queue weight (scaled) */
+	int max_th ;		/* maximum threshold for queue (scaled) */
+	int min_th ;		/* minimum threshold for queue (scaled) */
+	int max_p ;		/* maximum value for p_b (scaled) */
 
 };
 
@@ -177,10 +177,10 @@ struct dn_flow {
 };
 
 
-    /*
+/*
  * Scheduler template, mostly indicating the name, number,
  * sched_mask and buckets.
-     */
+ */
 struct dn_sch {
 	struct dn_id	oid;
 	uint32_t	sched_nr; /* N, scheduler number */
@@ -199,14 +199,14 @@ struct dn_sch {
 #define	ED_MAX_SAMPLES_NO	1024
 struct dn_profile {
 	struct dn_id	oid;
-    /* fields to simulate a delay profile */
+	/* fields to simulate a delay profile */
 #define ED_MAX_NAME_LEN		32
-    char name[ED_MAX_NAME_LEN];
-	int		link_nr;
-    int loss_level;
-	int		bandwidth;	// XXX use link bandwidth?
-	int		samples_no;	/* actual length of samples[] */
-	int samples[ED_MAX_SAMPLES_NO]; /* may be shorter */
+	char	name[ED_MAX_NAME_LEN];
+	int	link_nr;
+	int	loss_level;
+	int	bandwidth;			// XXX use link bandwidth?
+	int	samples_no;			/* actual len of samples[] */
+	int	samples[ED_MAX_SAMPLES_NO];	/* may be shorter */
 };
 
 
diff --git a/sys/netinet/ipfw/dn_sched.h b/sys/netinet/ipfw/dn_sched.h
index fe54b0205cf..b6bf24e466a 100644
--- a/sys/netinet/ipfw/dn_sched.h
+++ b/sys/netinet/ipfw/dn_sched.h
@@ -140,9 +140,9 @@ struct dn_alg {
 
 /* MSVC does not support initializers so we need this ugly macro */
 #ifdef _WIN32
-#define _SI(fld)        
+#define _SI(fld)
 #else
-#define _SI(fld)        fld
+#define _SI(fld)	fld
 #endif
 
 /*
diff --git a/sys/netinet/ipfw/dn_sched_rr.c b/sys/netinet/ipfw/dn_sched_rr.c
index fc7be001b30..1bbd80057c3 100644
--- a/sys/netinet/ipfw/dn_sched_rr.c
+++ b/sys/netinet/ipfw/dn_sched_rr.c
@@ -94,7 +94,7 @@ rr_remove_head(struct rr_si *si)
 	if (si->head == NULL)
 		return; /* empty queue */
 	si->head->status = 0;
-	
+
 	if (si->head == si->tail) {
 		si->head = si->tail = NULL;
 		return;
@@ -141,7 +141,7 @@ next_pointer(struct rr_si *si)
 	si->tail = si->tail->qnext;
 }
 
-static int 
+static int
 rr_enqueue(struct dn_sch_inst *_si, struct dn_queue *q, struct mbuf *m)
 {
 	struct rr_si *si;
@@ -154,7 +154,7 @@ rr_enqueue(struct dn_sch_inst *_si, struct dn_queue *q, struct mbuf *m)
 			return 0;
 	}
 
-	/* If reach this point, queue q was idle */ 
+	/* If reach this point, queue q was idle */
 	si = (struct rr_si *)(_si + 1);
 	rrq = (struct rr_queue *)q;
 
diff --git a/sys/netinet/ipfw/dn_sched_wf2q.c b/sys/netinet/ipfw/dn_sched_wf2q.c
index 1fbc1202e40..55a49550b7f 100644
--- a/sys/netinet/ipfw/dn_sched_wf2q.c
+++ b/sys/netinet/ipfw/dn_sched_wf2q.c
@@ -125,7 +125,7 @@ idle_check(struct wf2qp_si *si, int n, int force)
     }
 }
 
-static int 
+static int
 wf2qp_enqueue(struct dn_sch_inst *_si, struct dn_queue *q, struct mbuf *m)
 {
     struct dn_fsk *fs = q->fs;
@@ -140,7 +140,7 @@ wf2qp_enqueue(struct dn_sch_inst *_si, struct dn_queue *q, struct mbuf *m)
 	    return 0;
     }
 
-    /* If reach this point, queue q was idle */ 
+    /* If reach this point, queue q was idle */
     alg_fq = (struct wf2qp_queue *)q;
 
     if (DN_KEY_LT(alg_fq->F, alg_fq->S)) {
@@ -318,7 +318,7 @@ wf2qp_free_queue(struct dn_queue *q)
 {
 	struct wf2qp_queue *alg_fq = (struct wf2qp_queue *)q;
 	struct wf2qp_si *si = (struct wf2qp_si *)(q->_si + 1);
-    
+   
 	if (alg_fq->S >= alg_fq->F + 1)
 		return 0;	/* nothing to do, not in any heap */
 	si->wsum -= q->fs->fs.par[0];
@@ -361,7 +361,7 @@ static struct dn_alg wf2qp_desc = {
 	_SI( .destroy = )  NULL,
 	_SI( .new_sched = ) wf2qp_new_sched,
 	_SI( .free_sched = ) wf2qp_free_sched,
-    
+
 	_SI( .new_fsk = ) wf2qp_new_fsk,
 	_SI( .free_fsk = )  NULL,
 
diff --git a/sys/netinet/ipfw/ip_dn_private.h b/sys/netinet/ipfw/ip_dn_private.h
index f1a7f3fa509..03b43dba55d 100644
--- a/sys/netinet/ipfw/ip_dn_private.h
+++ b/sys/netinet/ipfw/ip_dn_private.h
@@ -149,7 +149,7 @@ struct dn_parms {
 	int drain_sch;
 	uint32_t expire;
 	uint32_t expire_cycle;	/* tick count */
-	
+
 	int init_done;
 
 	/* if the upper half is busy doing something long,
diff --git a/sys/netinet/ipfw/ip_dummynet.c b/sys/netinet/ipfw/ip_dummynet.c
index 0a6d16dc018..01714aa6634 100644
--- a/sys/netinet/ipfw/ip_dummynet.c
+++ b/sys/netinet/ipfw/ip_dummynet.c
@@ -1547,28 +1547,28 @@ config_profile(struct dn_profile *pf, struct dn_id *arg)
 	/* XXX other sanity checks */
 	DN_BH_WLOCK();
 	for (; i < 2*DN_MAX_ID; i += DN_MAX_ID) {
-	s = locate_scheduler(i);
+		s = locate_scheduler(i);
 
-	if (s == NULL) {
+		if (s == NULL) {
 			err = EINVAL;
 			break;
-	}
-	dn_cfg.id++;
-	/*
-	 * If we had a profile and the new one does not fit,
-	 * or it is deleted, then we need to free memory.
-	 */
-	if (s->profile && (pf->samples_no == 0 ||
-			s->profile->oid.len < pf->oid.len)) {
-		free(s->profile, M_DUMMYNET);
-		s->profile = NULL;
-	}
+		}
+		dn_cfg.id++;
+		/*
+		 * If we had a profile and the new one does not fit,
+		 * or it is deleted, then we need to free memory.
+		 */
+		if (s->profile && (pf->samples_no == 0 ||
+		    s->profile->oid.len < pf->oid.len)) {
+			free(s->profile, M_DUMMYNET);
+			s->profile = NULL;
+		}
 		if (pf->samples_no == 0)
 			continue;
-	/*
+		/*
 		 * new profile, possibly allocate memory
-	 * and copy data.
-	 */
+		 * and copy data.
+		 */
 		if (s->profile == NULL)
 			s->profile = malloc(pf->oid.len,
 			    M_DUMMYNET, M_NOWAIT | M_ZERO);
@@ -1642,7 +1642,8 @@ do_config(void *p, int l)
 		default:
 			D("cmd %d not implemented", o->type);
 			break;
-#ifdef EMULATE_SYSCTL		
+
+#ifdef EMULATE_SYSCTL
 		/* sysctl emulation.
 		 * if we recognize the command, jump to the correct
 		 * handler and return
@@ -1651,6 +1652,7 @@ do_config(void *p, int l)
 			err = kesysctl_emu_set(p, l);
 			return err;
 #endif
+
 		case DN_CMD_CONFIG: /* simply a header */
 			break;
 
diff --git a/sys/netinet/ipfw/ip_fw_pfil.c b/sys/netinet/ipfw/ip_fw_pfil.c
index b4e31d4c8c8..e87a4c973fe 100644
--- a/sys/netinet/ipfw/ip_fw_pfil.c
+++ b/sys/netinet/ipfw/ip_fw_pfil.c
@@ -147,8 +147,8 @@ again:
 	switch (ipfw) {
 	case IP_FW_PASS:
 		/* next_hop may be set by ipfw_chk */
-                if (args.next_hop == NULL)
-                        break; /* pass */
+		if (args.next_hop == NULL)
+			break; /* pass */
 #ifndef IPFIREWALL_FORWARD
 		ret = EACCES;
 #else
@@ -347,14 +347,14 @@ ipfw_attach_hooks(int arg)
 
 	if (arg == 0) /* detach */
 		ipfw_hook(0, AF_INET);
-        else if (V_fw_enable && ipfw_hook(1, AF_INET) != 0) {
+	else if (V_fw_enable && ipfw_hook(1, AF_INET) != 0) {
                 error = ENOENT; /* see ip_fw_pfil.c::ipfw_hook() */
                 printf("ipfw_hook() error\n");
         }
 #ifdef INET6
 	if (arg == 0) /* detach */
 		ipfw_hook(0, AF_INET6);
-        else if (V_fw6_enable && ipfw_hook(1, AF_INET6) != 0) {
+	else if (V_fw6_enable && ipfw_hook(1, AF_INET6) != 0) {
                 error = ENOENT;
                 printf("ipfw6_hook() error\n");
         }
diff --git a/sys/netinet/ipfw/ip_fw_private.h b/sys/netinet/ipfw/ip_fw_private.h
index c29ae0ad90b..ac55433750a 100644
--- a/sys/netinet/ipfw/ip_fw_private.h
+++ b/sys/netinet/ipfw/ip_fw_private.h
@@ -214,7 +214,7 @@ struct ip_fw_chain {
 	struct ip_fw	*default_rule;
 	int		n_rules;	/* number of static rules */
 	int		static_len;	/* total len of static rules */
-	struct ip_fw    **map;	/* array of rule ptrs to ease lookup */
+	struct ip_fw	**map;		/* array of rule ptrs to ease lookup */
 	LIST_HEAD(nat_list, cfg_nat) nat;       /* list of nat entries */
 	struct radix_node_head *tables[IPFW_TABLES_MAX];
 #if defined( __linux__ ) || defined( _WIN32 )

From d55ebfbd4e1ccc00cad998a5e9d24de4b98b448e Mon Sep 17 00:00:00 2001
From: Luigi Rizzo <luigi@FreeBSD.org>
Date: Mon, 19 Apr 2010 16:35:47 +0000
Subject: [PATCH 143/532] fix 64-bit build

Reported by:	Robert Noland
---
 sbin/ipfw/ipfw2.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sbin/ipfw/ipfw2.c b/sbin/ipfw/ipfw2.c
index b0f76830c40..f313b51b968 100644
--- a/sbin/ipfw/ipfw2.c
+++ b/sbin/ipfw/ipfw2.c
@@ -328,9 +328,11 @@ pr_u64(uint64_t *pd, int width)
 #else
 #define U64_FMT "llu"
 #endif
-	uint64_t d;
+	uint64_t u;
+	unsigned long long d;
 
-	bcopy (pd, &d, sizeof(d));
+	bcopy (pd, &u, sizeof(u));
+	d = u;
 	return (width > 0) ?
 		printf("%*" U64_FMT " ", width, d) :
 		snprintf(NULL, 0, "%" U64_FMT, d) ;

From 4ad53b2511c4a4421c361c2fe6622e4865d4695e Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Mon, 19 Apr 2010 17:16:23 +0000
Subject: [PATCH 144/532] Dump the AR_PHY_TURBO register on the AR5416. This
 register holds 11n configurations.

Sponsored by:	iXsystems, inc.
---
 tools/tools/ath/common/dumpregs_5416.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/tools/tools/ath/common/dumpregs_5416.c b/tools/tools/ath/common/dumpregs_5416.c
index a074ff7feae..0bc66f35471 100644
--- a/tools/tools/ath/common/dumpregs_5416.c
+++ b/tools/tools/ath/common/dumpregs_5416.c
@@ -394,6 +394,7 @@ static struct dumpreg ar5416regs[] = {
     DEFBASIC(AR_SLP_CNT,	"SLPCNT"),
     DEFBASIC(AR_SLP_MIB_CTRL,	"SLPMIB"),
     DEFBASIC(AR_EXTRCCNT,	"EXTRCCNT"),
+    DEFBASIC(AR_PHY_TURBO,	"PHYTURBO"),
 
     DEFVOID(AR_PHY_ADC_SERIAL_CTL,	"PHY_ADC_SERIAL_CTL"),
 

From 002d1d1c385f0c9316948f2ed64973fbf496e270 Mon Sep 17 00:00:00 2001
From: Jaakko Heinonen <jh@FreeBSD.org>
Date: Mon, 19 Apr 2010 20:07:35 +0000
Subject: [PATCH 145/532] Fix ddb(4) "show geom addr" command when INVARIANTS
 is enabled. Don't assert that the topology lock is held when g_valid_obj() is
 called from debugger.

MFC after:	1 week
---
 sys/geom/geom_subr.c | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

diff --git a/sys/geom/geom_subr.c b/sys/geom/geom_subr.c
index 9bef0e3638f..772277e02a7 100644
--- a/sys/geom/geom_subr.c
+++ b/sys/geom/geom_subr.c
@@ -59,6 +59,10 @@ __FBSDID("$FreeBSD$");
 #include <ddb/ddb.h>
 #endif
 
+#ifdef KDB
+#include <sys/kdb.h>
+#endif
+
 struct class_list_head g_classes = LIST_HEAD_INITIALIZER(g_classes);
 static struct g_tailq_head geoms = TAILQ_HEAD_INITIALIZER(geoms);
 char *g_wait_event, *g_wait_up, *g_wait_down, *g_wait_sim;
@@ -1011,12 +1015,11 @@ g_getattr__(const char *attr, struct g_consumer *cp, void *var, int len)
 
 #if defined(DIAGNOSTIC) || defined(DDB)
 /*
- * This function walks (topologically unsafely) the mesh and return a
- * non-zero integer if it finds the argument pointer is an object.
- * The return value indicates which type of object it is belived to be.
- * If topology is not locked, this function is potentially dangerous,
- * but since it is for debugging purposes and can be useful for instance
- * from DDB, we do not assert topology lock is held.
+ * This function walks the mesh and returns a non-zero integer if it
+ * finds the argument pointer is an object. The return value indicates
+ * which type of object it is believed to be. If topology is not locked,
+ * this function is potentially dangerous, but we don't assert that the
+ * topology lock is held when called from debugger.
  */
 int
 g_valid_obj(void const *ptr)
@@ -1026,7 +1029,10 @@ g_valid_obj(void const *ptr)
 	struct g_consumer *cp;
 	struct g_provider *pp;
 
-	g_topology_assert();
+#ifdef KDB
+	if (kdb_active == 0)
+#endif
+		g_topology_assert();
 
 	LIST_FOREACH(mp, &g_classes, class) {
 		if (ptr == mp)

From 206727f1fa03d968343c3cd1cb49ba1618ae620e Mon Sep 17 00:00:00 2001
From: Edwin Groothuis <edwin@FreeBSD.org>
Date: Mon, 19 Apr 2010 20:54:00 +0000
Subject: [PATCH 146/532] Vendor import of tzdata2010i:

- Marocca will have DST this year between May and August
- Historical data for Taiwan
- No more DST for the Argentinian province/state San Luis this year.

Obtained from:	ftp://elsie.nci.nih.gov/pub/
---
 africa       | 40 +++++++++++++++++++++++++++++++++++---
 asia         | 55 +++++++++++++++++++++++++++++++++++++++++++++++-----
 southamerica | 27 +++++++++++++++++++++++---
 3 files changed, 111 insertions(+), 11 deletions(-)

diff --git a/africa b/africa
index ad89bb743c1..8cae7a682b0 100644
--- a/africa
+++ b/africa
@@ -1,5 +1,5 @@
 # <pre>
-# @(#)africa	8.23
+# @(#)africa	8.26
 # This file is in the public domain, so clarified as of
 # 2009-05-17 by Arthur David Olson.
 
@@ -680,6 +680,21 @@ Zone	Indian/Mayotte	3:00:56 -	LMT	1911 Jul	# Mamoutzou
 # http://www.worldtimezone.com/dst_news/dst_news_morocco03.html
 # </a>
 
+# From Steffen Thorsen (2010-04-13):
+# Several news media in Morocco report that the Ministry of Modernization
+# of Public Sectors has announced that Morocco will have DST from
+# 2010-05-02 to 2010-08-08.
+#
+# Example:
+# <a href="http://www.lavieeco.com/actualites/4099-le-maroc-passera-a-l-heure-d-ete-gmt1-le-2-mai.html">
+# http://www.lavieeco.com/actualites/4099-le-maroc-passera-a-l-heure-d-ete-gmt1-le-2-mai.html
+# </a>
+# (French)
+# Our page:
+# <a href="http://www.timeanddate.com/news/time/morocco-starts-dst-2010.html">
+# http://www.timeanddate.com/news/time/morocco-starts-dst-2010.html
+# </a>
+
 # RULE	NAME	FROM	TO	TYPE	IN	ON	AT	SAVE	LETTER/S
 
 Rule	Morocco	1939	only	-	Sep	12	 0:00	1:00	S
@@ -701,6 +716,8 @@ Rule	Morocco	2008	only	-	Jun	 1	 0:00	1:00	S
 Rule	Morocco	2008	only	-	Sep	 1	 0:00	0	-
 Rule	Morocco	2009	only	-	Jun	 1	 0:00	1:00	S
 Rule	Morocco	2009	only	-	Aug	 21	 0:00	0	-
+Rule	Morocco	2010	only	-	May	 2	 0:00	1:00	S
+Rule	Morocco	2010	only	-	Aug	 8	 0:00	0	-
 # Zone	NAME		GMTOFF	RULES	FORMAT	[UNTIL]
 Zone Africa/Casablanca	-0:30:20 -	LMT	1913 Oct 26
 			 0:00	Morocco	WE%sT	1984 Mar 16
@@ -942,6 +959,24 @@ Zone	Africa/Lome	0:04:52 -	LMT	1893
 # Therefore, the standard time will be kept unchanged the whole year long."
 # So foregoing DST seems to be an exception (albeit one that may be repeated in the  future).
 
+# From Alexander Krivenyshev (2010-03-27):
+# According to some news reports Tunis confirmed not to use DST in 2010
+#
+# (translation):
+# "The Tunisian government has decided to abandon DST, which was scheduled on
+# Sunday...
+# Tunisian authorities had suspended the DST for the first time last year also
+# coincided with the month of Ramadan..."
+#
+# (in Arabic)
+# <a href="http://www.moheet.com/show_news.aspx?nid=358861&pg=1">
+# http://www.moheet.com/show_news.aspx?nid=358861&pg=1
+# <a href="http://www.almadenahnews.com/newss/news.php?c=118&id=38036">
+# http://www.almadenahnews.com/newss/news.php?c=118&id=38036
+# or
+# <a href="http://www.worldtimezone.com/dst_news/dst_news_tunis02.html">
+# http://www.worldtimezone.com/dst_news/dst_news_tunis02.html
+
 # Rule	NAME	FROM	TO	TYPE	IN	ON	AT	SAVE	LETTER/S
 Rule	Tunisia	1939	only	-	Apr	15	23:00s	1:00	S
 Rule	Tunisia	1939	only	-	Nov	18	23:00s	0	-
@@ -968,8 +1003,7 @@ Rule	Tunisia	2005	only	-	May	 1	 0:00s	1:00	S
 Rule	Tunisia	2005	only	-	Sep	30	 1:00s	0	-
 Rule	Tunisia	2006	2008	-	Mar	lastSun	 2:00s	1:00	S
 Rule	Tunisia	2006	2008	-	Oct	lastSun	 2:00s	0	-
-Rule	Tunisia	2010	max	-	Mar	lastSun	 2:00s	1:00	S
-Rule	Tunisia	2010	max	-	Oct	lastSun	 2:00s	0	-
+
 # Shanks & Pottenger give 0:09:20 for Paris Mean Time; go with Howse's
 # more precise 0:09:21.
 # Shanks & Pottenger say the 1911 switch was on Mar 9; go with Howse's Mar 11.
diff --git a/asia b/asia
index 8cafc5973e2..78ff2ffedcf 100644
--- a/asia
+++ b/asia
@@ -1,4 +1,4 @@
-# @(#)asia	8.58
+# @(#)asia	8.60
 # This file is in the public domain, so clarified as of
 # 2009-05-17 by Arthur David Olson.
 
@@ -566,6 +566,28 @@ Zone	Asia/Hong_Kong	7:36:36 -	LMT	1904 Oct 30
 # was still controlled by Japan.  This is hard to believe, but we don't
 # have any other information.
 
+# From smallufo (2010-04-03):
+# According to Taiwan's CWB,
+# <a href="http://www.cwb.gov.tw/V6/astronomy/cdata/summert.htm">
+# http://www.cwb.gov.tw/V6/astronomy/cdata/summert.htm
+# </a>
+# Taipei has DST in 1979 between July 1st and Sep 30.
+
+# From Arthur David Olson (2010-04-07):
+# Here's Google's translation of the table at the bottom of the "summert.htm" page:
+# Decade 	                                                    Name                      Start and end date
+# Republic of China 34 years to 40 years (AD 1945-1951 years) Summer Time               May 1 to September 30 
+# 41 years of the Republic of China (AD 1952)                 Daylight Saving Time      March 1 to October 31 
+# Republic of China 42 years to 43 years (AD 1953-1954 years) Daylight Saving Time      April 1 to October 31 
+# In the 44 years to 45 years (AD 1955-1956 years)            Daylight Saving Time      April 1 to September 30 
+# Republic of China 46 years to 48 years (AD 1957-1959)       Summer Time               April 1 to September 30 
+# Republic of China 49 years to 50 years (AD 1960-1961)       Summer Time               June 1 to September 30 
+# Republic of China 51 years to 62 years (AD 1962-1973 years) Stop Summer Time 
+# Republic of China 63 years to 64 years (1974-1975 AD)       Daylight Saving Time      April 1 to September 30 
+# Republic of China 65 years to 67 years (1976-1978 AD)       Stop Daylight Saving Time 
+# Republic of China 68 years (AD 1979)                        Daylight Saving Time      July 1 to September 30 
+# Republic of China since 69 years (AD 1980)                  Stop Daylight Saving Time
+
 # Rule	NAME	FROM	TO	TYPE	IN	ON	AT	SAVE	LETTER/S
 Rule	Taiwan	1945	1951	-	May	1	0:00	1:00	D
 Rule	Taiwan	1945	1951	-	Oct	1	0:00	0	S
@@ -576,8 +598,9 @@ Rule	Taiwan	1955	1961	-	Oct	1	0:00	0	S
 Rule	Taiwan	1960	1961	-	Jun	1	0:00	1:00	D
 Rule	Taiwan	1974	1975	-	Apr	1	0:00	1:00	D
 Rule	Taiwan	1974	1975	-	Oct	1	0:00	0	S
-Rule	Taiwan	1980	only	-	Jun	30	0:00	1:00	D
-Rule	Taiwan	1980	only	-	Sep	30	0:00	0	S
+Rule	Taiwan	1979	only	-	Jun	30	0:00	1:00	D
+Rule	Taiwan	1979	only	-	Sep	30	0:00	0	S
+
 # Zone	NAME		GMTOFF	RULES	FORMAT	[UNTIL]
 Zone	Asia/Taipei	8:06:00 -	LMT	1896 # or Taibei or T'ai-pei
 			8:00	Taiwan	C%sT
@@ -1912,13 +1935,35 @@ Zone	Asia/Muscat	3:54:20 -	LMT	1920
 # [T]he German Consulate General in Karachi reported me today that Pakistan
 # will go back to standard time on 1st of November.
 
+# From Steffen Thorsen (2010-03-26):
+# Steffen Thorsen wrote:
+# > On Thursday (2010-03-25) it was announced that DST would start in
+# > Pakistan on 2010-04-01.
+# >
+# > Then today, the president said that they might have to revert the
+# > decision if it is not supported by the parliament. So at the time
+# > being, it seems unclear if DST will be actually observed or not - but
+# > April 1 could be a more likely date than April 15.
+# Now, it seems that the decision to not observe DST in final:
+#
+# "Govt Withdraws Plan To Advance Clocks"
+# <a href="http://www.apakistannews.com/govt-withdraws-plan-to-advance-clocks-172041">
+# http://www.apakistannews.com/govt-withdraws-plan-to-advance-clocks-172041
+# </a>
+#
+# "People laud PM's announcement to end DST"
+# <a href="http://www.app.com.pk/en_/index.php?option=com_content&task=view&id=99374&Itemid=2">
+# http://www.app.com.pk/en_/index.php?option=com_content&task=view&id=99374&Itemid=2
+# </a>
+
 # Rule	NAME	FROM	TO	TYPE	IN	ON	AT	SAVE	LETTER/S
 Rule Pakistan	2002	only	-	Apr	Sun>=2	0:01	1:00	S
 Rule Pakistan	2002	only	-	Oct	Sun>=2	0:01	0	-
 Rule Pakistan	2008	only	-	Jun	1	0:00	1:00	S
 Rule Pakistan	2008	only	-	Nov	1	0:00	0	-
-Rule Pakistan	2009	max	-	Apr	15	0:00	1:00	S
-Rule Pakistan	2009	max	-	Nov	1	0:00	0	-
+Rule Pakistan	2009	only	-	Apr	15	0:00	1:00	S
+Rule Pakistan	2009	only	-	Nov	1	0:00	0	-
+
 # Zone	NAME		GMTOFF	RULES	FORMAT	[UNTIL]
 Zone	Asia/Karachi	4:28:12 -	LMT	1907
 			5:30	-	IST	1942 Sep
diff --git a/southamerica b/southamerica
index 8db45ccda9a..7355022929d 100644
--- a/southamerica
+++ b/southamerica
@@ -1,5 +1,5 @@
 # <pre>
-# @(#)southamerica	8.43
+# @(#)southamerica	8.44
 # This file is in the public domain, so clarified as of
 # 2009-05-17 by Arthur David Olson.
 
@@ -437,6 +437,27 @@ Rule	Arg	2008	only	-	Oct	Sun>=15	0:00	1:00	S
 # of the country calls it "ART".
 # ...
 
+# From Alexander Krivenyshev (2010-04-09):
+# According to news reports from El Diario de la Republica Province San
+# Luis, Argentina (standard time UTC-04) will keep Daylight Saving Time
+# after April 11, 2010--will continue to have same time as rest of
+# Argentina (UTC-3) (no DST).
+#
+# Confirmaron la pr&oacute;rroga del huso horario de verano (Spanish)
+# <a href="http://www.eldiariodelarepublica.com/index.php?option=com_content&task=view&id=29383&Itemid=9">
+# http://www.eldiariodelarepublica.com/index.php?option=com_content&task=view&id=29383&Itemid=9
+# </a>
+# or (some English translation):
+# <a href="http://www.worldtimezone.com/dst_news/dst_news_argentina08.html">
+# http://www.worldtimezone.com/dst_news/dst_news_argentina08.html
+# </a>
+
+# From Mariano Absatz (2010-04-12):
+# yes...I can confirm this...and given that San Luis keeps calling
+# UTC-03:00 "summer time", we should't just let San Luis go back to "Arg"
+# rules...San Luis is still using "Western ARgentina Time" and it got
+# stuck on Summer daylight savings time even though the summer is over.
+
 # Zone	NAME		GMTOFF	RULES	FORMAT	[UNTIL]
 #
 # Buenos Aires (BA), Capital Federal (CF),
@@ -570,8 +591,8 @@ Zone America/Argentina/Mendoza -4:35:16 - LMT	1894 Oct 31
 #
 # San Luis (SL)
 
-Rule	SanLuis	2008	max	-	Mar	Sun>=8	0:00	0	-
-Rule	SanLuis	2007	max	-	Oct	Sun>=8	0:00	1:00	S
+Rule	SanLuis	2008	2009	-	Mar	Sun>=8	0:00	0	-
+Rule	SanLuis	2007	2009	-	Oct	Sun>=8	0:00	1:00	S
 
 Zone America/Argentina/San_Luis -4:25:24 - LMT	1894 Oct 31
 			-4:16:48 -	CMT	1920 May

From 6da6d0a9e3f608feb5ca67cf633670359157549b Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Mon, 19 Apr 2010 22:10:40 +0000
Subject: [PATCH 147/532] With r206844, CSUM_TCP is also set for CSUM_TSO case.
 Modify drivers to take into account for the change. Basically CSUM_TSO should
 be checked before checking CSUM_TCP.

---
 sys/dev/age/if_age.c |  28 ++++++------
 sys/dev/alc/if_alc.c |  44 +++++++++---------
 sys/dev/ale/if_ale.c |  19 ++++----
 sys/dev/fxp/if_fxp.c | 106 +++++++++++++++++++++----------------------
 sys/dev/msk/if_msk.c |  48 ++++++++++----------
 sys/dev/nfe/if_nfe.c |  13 +++---
 6 files changed, 125 insertions(+), 133 deletions(-)

diff --git a/sys/dev/age/if_age.c b/sys/dev/age/if_age.c
index 3c5a1079774..15ce0eff851 100644
--- a/sys/dev/age/if_age.c
+++ b/sys/dev/age/if_age.c
@@ -1629,22 +1629,8 @@ age_encap(struct age_softc *sc, struct mbuf **m_head)
 	}
 
 	m = *m_head;
-	/* Configure Tx IP/TCP/UDP checksum offload. */
-	if ((m->m_pkthdr.csum_flags & AGE_CSUM_FEATURES) != 0) {
-		cflags |= AGE_TD_CSUM;
-		if ((m->m_pkthdr.csum_flags & CSUM_TCP) != 0)
-			cflags |= AGE_TD_TCPCSUM;
-		if ((m->m_pkthdr.csum_flags & CSUM_UDP) != 0)
-			cflags |= AGE_TD_UDPCSUM;
-		/* Set checksum start offset. */
-		cflags |= (poff << AGE_TD_CSUM_PLOADOFFSET_SHIFT);
-		/* Set checksum insertion position of TCP/UDP. */
-		cflags |= ((poff + m->m_pkthdr.csum_data) <<
-		    AGE_TD_CSUM_XSUMOFFSET_SHIFT);
-	}
-
-	/* Configure TSO. */
 	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+		/* Configure TSO. */
 		if (poff + (tcp->th_off << 2) == m->m_pkthdr.len) {
 			/* Not TSO but IP/TCP checksum offload. */
 			cflags |= AGE_TD_IPCSUM | AGE_TD_TCPCSUM;
@@ -1660,6 +1646,18 @@ age_encap(struct age_softc *sc, struct mbuf **m_head)
 		/* Set IP/TCP header size. */
 		cflags |= ip->ip_hl << AGE_TD_IPHDR_LEN_SHIFT;
 		cflags |= tcp->th_off << AGE_TD_TSO_TCPHDR_LEN_SHIFT;
+	} else if ((m->m_pkthdr.csum_flags & AGE_CSUM_FEATURES) != 0) {
+		/* Configure Tx IP/TCP/UDP checksum offload. */
+		cflags |= AGE_TD_CSUM;
+		if ((m->m_pkthdr.csum_flags & CSUM_TCP) != 0)
+			cflags |= AGE_TD_TCPCSUM;
+		if ((m->m_pkthdr.csum_flags & CSUM_UDP) != 0)
+			cflags |= AGE_TD_UDPCSUM;
+		/* Set checksum start offset. */
+		cflags |= (poff << AGE_TD_CSUM_PLOADOFFSET_SHIFT);
+		/* Set checksum insertion position of TCP/UDP. */
+		cflags |= ((poff + m->m_pkthdr.csum_data) <<
+		    AGE_TD_CSUM_XSUMOFFSET_SHIFT);
 	}
 
 	/* Configure VLAN hardware tag insertion. */
diff --git a/sys/dev/alc/if_alc.c b/sys/dev/alc/if_alc.c
index e95b04453cd..c685b84e863 100644
--- a/sys/dev/alc/if_alc.c
+++ b/sys/dev/alc/if_alc.c
@@ -1908,28 +1908,7 @@ alc_encap(struct alc_softc *sc, struct mbuf **m_head)
 		vtag = (vtag << TD_VLAN_SHIFT) & TD_VLAN_MASK;
 		cflags |= TD_INS_VLAN_TAG;
 	}
-	/* Configure Tx checksum offload. */
-	if ((m->m_pkthdr.csum_flags & ALC_CSUM_FEATURES) != 0) {
-#ifdef ALC_USE_CUSTOM_CSUM
-		cflags |= TD_CUSTOM_CSUM;
-		/* Set checksum start offset. */
-		cflags |= ((poff >> 1) << TD_PLOAD_OFFSET_SHIFT) &
-		    TD_PLOAD_OFFSET_MASK;
-		/* Set checksum insertion position of TCP/UDP. */
-		cflags |= (((poff + m->m_pkthdr.csum_data) >> 1) <<
-		    TD_CUSTOM_CSUM_OFFSET_SHIFT) & TD_CUSTOM_CSUM_OFFSET_MASK;
-#else
-		if ((m->m_pkthdr.csum_flags & CSUM_IP) != 0)
-			cflags |= TD_IPCSUM;
-		if ((m->m_pkthdr.csum_flags & CSUM_TCP) != 0)
-			cflags |= TD_TCPCSUM;
-		if ((m->m_pkthdr.csum_flags & CSUM_UDP) != 0)
-			cflags |= TD_UDPCSUM;
-		/* Set TCP/UDP header offset. */
-		cflags |= (poff << TD_L4HDR_OFFSET_SHIFT) &
-		    TD_L4HDR_OFFSET_MASK;
-#endif
-	} else if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
 		/* Request TSO and set MSS. */
 		cflags |= TD_TSO | TD_TSO_DESCV1;
 		cflags |= ((uint32_t)m->m_pkthdr.tso_segsz << TD_MSS_SHIFT) &
@@ -1961,6 +1940,27 @@ alc_encap(struct alc_softc *sc, struct mbuf **m_head)
 		}
 		/* Handle remaining fragments. */
 		idx = 1;
+	} else if ((m->m_pkthdr.csum_flags & ALC_CSUM_FEATURES) != 0) {
+		/* Configure Tx checksum offload. */
+#ifdef ALC_USE_CUSTOM_CSUM
+		cflags |= TD_CUSTOM_CSUM;
+		/* Set checksum start offset. */
+		cflags |= ((poff >> 1) << TD_PLOAD_OFFSET_SHIFT) &
+		    TD_PLOAD_OFFSET_MASK;
+		/* Set checksum insertion position of TCP/UDP. */
+		cflags |= (((poff + m->m_pkthdr.csum_data) >> 1) <<
+		    TD_CUSTOM_CSUM_OFFSET_SHIFT) & TD_CUSTOM_CSUM_OFFSET_MASK;
+#else
+		if ((m->m_pkthdr.csum_flags & CSUM_IP) != 0)
+			cflags |= TD_IPCSUM;
+		if ((m->m_pkthdr.csum_flags & CSUM_TCP) != 0)
+			cflags |= TD_TCPCSUM;
+		if ((m->m_pkthdr.csum_flags & CSUM_UDP) != 0)
+			cflags |= TD_UDPCSUM;
+		/* Set TCP/UDP header offset. */
+		cflags |= (poff << TD_L4HDR_OFFSET_SHIFT) &
+		    TD_L4HDR_OFFSET_MASK;
+#endif
 	}
 	for (; idx < nsegs; idx++) {
 		desc = &sc->alc_rdata.alc_tx_ring[prod];
diff --git a/sys/dev/ale/if_ale.c b/sys/dev/ale/if_ale.c
index 76f1b74f2af..b3e71872d5b 100644
--- a/sys/dev/ale/if_ale.c
+++ b/sys/dev/ale/if_ale.c
@@ -1737,8 +1737,14 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 	bus_dmamap_sync(sc->ale_cdata.ale_tx_tag, map, BUS_DMASYNC_PREWRITE);
 
 	m = *m_head;
-	/* Configure Tx checksum offload. */
-	if ((m->m_pkthdr.csum_flags & ALE_CSUM_FEATURES) != 0) {
+	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+		/* Request TSO and set MSS. */
+		cflags |= ALE_TD_TSO;
+		cflags |= ((uint32_t)m->m_pkthdr.tso_segsz << ALE_TD_MSS_SHIFT);
+		/* Set IP/TCP header size. */
+		cflags |= ip->ip_hl << ALE_TD_IPHDR_LEN_SHIFT;
+		cflags |= tcp->th_off << ALE_TD_TCPHDR_LEN_SHIFT;
+	} else if ((m->m_pkthdr.csum_flags & ALE_CSUM_FEATURES) != 0) {
 		/*
 		 * AR81xx supports Tx custom checksum offload feature
 		 * that offloads single 16bit checksum computation.
@@ -1769,15 +1775,6 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 		    ALE_TD_CSUM_XSUMOFFSET_SHIFT);
 	}
 
-	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
-		/* Request TSO and set MSS. */
-		cflags |= ALE_TD_TSO;
-		cflags |= ((uint32_t)m->m_pkthdr.tso_segsz << ALE_TD_MSS_SHIFT);
-		/* Set IP/TCP header size. */
-		cflags |= ip->ip_hl << ALE_TD_IPHDR_LEN_SHIFT;
-		cflags |= tcp->th_off << ALE_TD_TCPHDR_LEN_SHIFT;
-	}
-
 	/* Configure VLAN hardware tag insertion. */
 	if ((m->m_flags & M_VLANTAG) != 0) {
 		vtag = ALE_TX_VLAN_TAG(m->m_pkthdr.ether_vtag);
diff --git a/sys/dev/fxp/if_fxp.c b/sys/dev/fxp/if_fxp.c
index a8d961eb0fb..a480e7cc6a4 100644
--- a/sys/dev/fxp/if_fxp.c
+++ b/sys/dev/fxp/if_fxp.c
@@ -1417,60 +1417,6 @@ fxp_encap(struct fxp_softc *sc, struct mbuf **m_head)
 		    FXP_IPCB_HARDWAREPARSING_ENABLE;
 
 	m = *m_head;
-	/*
-	 * Deal with TCP/IP checksum offload. Note that
-	 * in order for TCP checksum offload to work,
-	 * the pseudo header checksum must have already
-	 * been computed and stored in the checksum field
-	 * in the TCP header. The stack should have
-	 * already done this for us.
-	 */
-	if (m->m_pkthdr.csum_flags & FXP_CSUM_FEATURES) {
-		txp->tx_cb->ipcb_ip_schedule = FXP_IPCB_TCPUDP_CHECKSUM_ENABLE;
-		if (m->m_pkthdr.csum_flags & CSUM_TCP)
-			txp->tx_cb->ipcb_ip_schedule |= FXP_IPCB_TCP_PACKET;
-
-#ifdef FXP_IP_CSUM_WAR
-		/*
-		 * XXX The 82550 chip appears to have trouble
-		 * dealing with IP header checksums in very small
-		 * datagrams, namely fragments from 1 to 3 bytes
-		 * in size. For example, say you want to transmit
-		 * a UDP packet of 1473 bytes. The packet will be
-		 * fragmented over two IP datagrams, the latter
-		 * containing only one byte of data. The 82550 will
-		 * botch the header checksum on the 1-byte fragment.
-		 * As long as the datagram contains 4 or more bytes
-		 * of data, you're ok.
-		 *
-                 * The following code attempts to work around this
-		 * problem: if the datagram is less than 38 bytes
-		 * in size (14 bytes ether header, 20 bytes IP header,
-		 * plus 4 bytes of data), we punt and compute the IP
-		 * header checksum by hand. This workaround doesn't
-		 * work very well, however, since it can be fooled
-		 * by things like VLAN tags and IP options that make
-		 * the header sizes/offsets vary.
-		 */
-
-		if (m->m_pkthdr.csum_flags & CSUM_IP) {
-			if (m->m_pkthdr.len < 38) {
-				struct ip *ip;
-				m->m_data += ETHER_HDR_LEN;
-				ip = mtod(m, struct ip *);
-				ip->ip_sum = in_cksum(m, ip->ip_hl << 2);
-				m->m_data -= ETHER_HDR_LEN;
-				m->m_pkthdr.csum_flags &= ~CSUM_IP;
-			} else {
-				txp->tx_cb->ipcb_ip_activation_high =
-				    FXP_IPCB_HARDWAREPARSING_ENABLE;
-				txp->tx_cb->ipcb_ip_schedule |=
-				    FXP_IPCB_IP_CHECKSUM_ENABLE;
-			}
-		}
-#endif
-	}
-
 	if (m->m_pkthdr.csum_flags & CSUM_TSO) {
 		/*
 		 * 82550/82551 requires ethernet/IP/TCP headers must be
@@ -1539,6 +1485,58 @@ fxp_encap(struct fxp_softc *sc, struct mbuf **m_head)
 		tcp_payload = m->m_pkthdr.len - ip_off - (ip->ip_hl << 2);
 		tcp_payload -= tcp->th_off << 2;
 		*m_head = m;
+	} else if (m->m_pkthdr.csum_flags & FXP_CSUM_FEATURES) {
+		/*
+		 * Deal with TCP/IP checksum offload. Note that
+		 * in order for TCP checksum offload to work,
+		 * the pseudo header checksum must have already
+		 * been computed and stored in the checksum field
+		 * in the TCP header. The stack should have
+		 * already done this for us.
+		 */
+		txp->tx_cb->ipcb_ip_schedule = FXP_IPCB_TCPUDP_CHECKSUM_ENABLE;
+		if (m->m_pkthdr.csum_flags & CSUM_TCP)
+			txp->tx_cb->ipcb_ip_schedule |= FXP_IPCB_TCP_PACKET;
+
+#ifdef FXP_IP_CSUM_WAR
+		/*
+		 * XXX The 82550 chip appears to have trouble
+		 * dealing with IP header checksums in very small
+		 * datagrams, namely fragments from 1 to 3 bytes
+		 * in size. For example, say you want to transmit
+		 * a UDP packet of 1473 bytes. The packet will be
+		 * fragmented over two IP datagrams, the latter
+		 * containing only one byte of data. The 82550 will
+		 * botch the header checksum on the 1-byte fragment.
+		 * As long as the datagram contains 4 or more bytes
+		 * of data, you're ok.
+		 *
+                 * The following code attempts to work around this
+		 * problem: if the datagram is less than 38 bytes
+		 * in size (14 bytes ether header, 20 bytes IP header,
+		 * plus 4 bytes of data), we punt and compute the IP
+		 * header checksum by hand. This workaround doesn't
+		 * work very well, however, since it can be fooled
+		 * by things like VLAN tags and IP options that make
+		 * the header sizes/offsets vary.
+		 */
+
+		if (m->m_pkthdr.csum_flags & CSUM_IP) {
+			if (m->m_pkthdr.len < 38) {
+				struct ip *ip;
+				m->m_data += ETHER_HDR_LEN;
+				ip = mtod(m, struct ip *);
+				ip->ip_sum = in_cksum(m, ip->ip_hl << 2);
+				m->m_data -= ETHER_HDR_LEN;
+				m->m_pkthdr.csum_flags &= ~CSUM_IP;
+			} else {
+				txp->tx_cb->ipcb_ip_activation_high =
+				    FXP_IPCB_HARDWAREPARSING_ENABLE;
+				txp->tx_cb->ipcb_ip_schedule |=
+				    FXP_IPCB_IP_CHECKSUM_ENABLE;
+			}
+		}
+#endif
 	}
 
 	error = bus_dmamap_load_mbuf_sg(sc->fxp_txmtag, txp->tx_map, *m_head,
diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index 61b131bf451..c083e1e10f6 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -2605,23 +2605,32 @@ msk_encap(struct msk_if_softc *sc_if, struct mbuf **m_head)
 		ip = (struct ip *)(mtod(m, char *) + offset);
 		offset += (ip->ip_hl << 2);
 		tcp_offset = offset;
-		/*
-		 * It seems that Yukon II has Tx checksum offload bug for
-		 * small TCP packets that's less than 60 bytes in size
-		 * (e.g. TCP window probe packet, pure ACK packet).
-		 * Common work around like padding with zeros to make the
-		 * frame minimum ethernet frame size didn't work at all.
-		 * Instead of disabling checksum offload completely we
-		 * resort to S/W checksum routine when we encounter short
-		 * TCP frames.
-		 * Short UDP packets appear to be handled correctly by
-		 * Yukon II. Also I assume this bug does not happen on
-		 * controllers that use newer descriptor format or
-		 * automatic Tx checksum calaulcation.
-		 */
-		if ((sc_if->msk_flags & MSK_FLAG_AUTOTX_CSUM) == 0 &&
+		if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+			m = m_pullup(m, offset + sizeof(struct tcphdr));
+			if (m == NULL) {
+				*m_head = NULL;
+				return (ENOBUFS);
+			}
+			tcp = (struct tcphdr *)(mtod(m, char *) + offset);
+			offset += (tcp->th_off << 2);
+		} else if ((sc_if->msk_flags & MSK_FLAG_AUTOTX_CSUM) == 0 &&
 		    (m->m_pkthdr.len < MSK_MIN_FRAMELEN) &&
 		    (m->m_pkthdr.csum_flags & CSUM_TCP) != 0) {
+			/*
+			 * It seems that Yukon II has Tx checksum offload bug
+			 * for small TCP packets that's less than 60 bytes in
+			 * size (e.g. TCP window probe packet, pure ACK packet).
+			 * Common work around like padding with zeros to make
+			 * the frame minimum ethernet frame size didn't work at
+			 * all.
+			 * Instead of disabling checksum offload completely we
+			 * resort to S/W checksum routine when we encounter
+			 * short TCP frames.
+			 * Short UDP packets appear to be handled correctly by
+			 * Yukon II. Also I assume this bug does not happen on
+			 * controllers that use newer descriptor format or
+			 * automatic Tx checksum calaulcation.
+			 */
 			m = m_pullup(m, offset + sizeof(struct tcphdr));
 			if (m == NULL) {
 				*m_head = NULL;
@@ -2632,15 +2641,6 @@ msk_encap(struct msk_if_softc *sc_if, struct mbuf **m_head)
 			    m->m_pkthdr.len, offset);
 			m->m_pkthdr.csum_flags &= ~CSUM_TCP;
 		}
-		if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
-			m = m_pullup(m, offset + sizeof(struct tcphdr));
-			if (m == NULL) {
-				*m_head = NULL;
-				return (ENOBUFS);
-			}
-			tcp = (struct tcphdr *)(mtod(m, char *) + offset);
-			offset += (tcp->th_off << 2);
-		}
 		*m_head = m;
 	}
 
diff --git a/sys/dev/nfe/if_nfe.c b/sys/dev/nfe/if_nfe.c
index 5d0bfd268f6..7d77754b90c 100644
--- a/sys/dev/nfe/if_nfe.c
+++ b/sys/dev/nfe/if_nfe.c
@@ -2366,7 +2366,12 @@ nfe_encap(struct nfe_softc *sc, struct mbuf **m_head)
 	m = *m_head;
 	cflags = flags = 0;
 	tso_segsz = 0;
-	if ((m->m_pkthdr.csum_flags & NFE_CSUM_FEATURES) != 0) {
+	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+		tso_segsz = (uint32_t)m->m_pkthdr.tso_segsz <<
+		    NFE_TX_TSO_SHIFT;
+		cflags &= ~(NFE_TX_IP_CSUM | NFE_TX_TCP_UDP_CSUM);
+		cflags |= NFE_TX_TSO;
+	} else if ((m->m_pkthdr.csum_flags & NFE_CSUM_FEATURES) != 0) {
 		if ((m->m_pkthdr.csum_flags & CSUM_IP) != 0)
 			cflags |= NFE_TX_IP_CSUM;
 		if ((m->m_pkthdr.csum_flags & CSUM_TCP) != 0)
@@ -2374,12 +2379,6 @@ nfe_encap(struct nfe_softc *sc, struct mbuf **m_head)
 		if ((m->m_pkthdr.csum_flags & CSUM_UDP) != 0)
 			cflags |= NFE_TX_TCP_UDP_CSUM;
 	}
-	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
-		tso_segsz = (uint32_t)m->m_pkthdr.tso_segsz <<
-		    NFE_TX_TSO_SHIFT;
-		cflags &= ~(NFE_TX_IP_CSUM | NFE_TX_TCP_UDP_CSUM);
-		cflags |= NFE_TX_TSO;
-	}
 
 	for (i = 0; i < nsegs; i++) {
 		if (sc->nfe_flags & NFE_40BIT_ADDR) {

From 248bb9379fd3db5e637005b70c851d2bf08ca042 Mon Sep 17 00:00:00 2001
From: Attilio Rao <attilio@FreeBSD.org>
Date: Mon, 19 Apr 2010 23:27:54 +0000
Subject: [PATCH 148/532] Fix a deadlock in the shutdown code: When performing
 a smp_rendezvous() or more likely, on amd64 and i386, a smp_tlb_shootdown()
 the caller will end up with the smp_ipi_mtx spinlock held, busy-waiting for
 other CPUs to acknowledge the operation. As long as CPUs are suspended (via
 cpu_reset()) between the active mask read and IPI sending there can be a
 deadlock where the caller will wait forever for a dead CPU to acknowledge the
 operation. Please note that on CPU0 that is going to be someway heavier
 because of the spinlocks being disabled earlier than quitting the machine.

Fix this bug by calling cpu_reset() with the smp_ipi_mtx held.
Note that it is very likely that a saner offline/online CPUs mechanism
will help heavilly in fixing similar cases as it is likely more bugs
of this type may arise in the future.

Reported by:	rwatson
Discussed with:	jhb
Tested by:	rnoland, Giovanni Trematerra
		<giovanni dot trematerra at gmail dot com>
MFC:		2 weeks

Special deciation to:	anyone who made possible to have 16-ways machines
			in Netperf
---
 sys/kern/kern_shutdown.c | 21 +++++++++++++--------
 1 file changed, 13 insertions(+), 8 deletions(-)

diff --git a/sys/kern/kern_shutdown.c b/sys/kern/kern_shutdown.c
index 9b28af40c28..4b27d45f6c0 100644
--- a/sys/kern/kern_shutdown.c
+++ b/sys/kern/kern_shutdown.c
@@ -62,7 +62,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/reboot.h>
 #include <sys/resourcevar.h>
 #include <sys/sched.h>
-#include <sys/smp.h>		/* smp_active */
+#include <sys/smp.h>
 #include <sys/sysctl.h>
 #include <sys/sysproto.h>
 
@@ -485,15 +485,20 @@ static void
 shutdown_reset(void *junk, int howto)
 {
 
-	/*
-	 * Disable interrupts on CPU0 in order to avoid fast handlers
-	 * to preempt the stopping process and to deadlock against other
-	 * CPUs.
-	 */
-	spinlock_enter();
-
 	printf("Rebooting...\n");
 	DELAY(1000000);	/* wait 1 sec for printf's to complete and be read */
+
+	/*
+	 * Acquiring smp_ipi_mtx here has a double effect:
+	 * - it disables interrupts avoiding CPU0 preemption
+	 *   by fast handlers (thus deadlocking  against other CPUs)
+	 * - it avoids deadlocks against smp_rendezvous() or, more 
+	 *   generally, threads busy-waiting, with this spinlock held,
+	 *   and waiting for responses by threads on other CPUs
+	 *   (ie. smp_tlb_shootdown()).
+	 */
+	mtx_lock_spin(&smp_ipi_mtx);
+
 	/* cpu_boot(howto); */ /* doesn't do anything at the moment */
 	cpu_reset();
 	/* NOTREACHED */ /* assuming reset worked */

From 95335fd84431aace91d15fe4a5b8aa9c94afa1b5 Mon Sep 17 00:00:00 2001
From: Attilio Rao <attilio@FreeBSD.org>
Date: Mon, 19 Apr 2010 23:40:46 +0000
Subject: [PATCH 149/532] getblk lockmgr is mostly used as a msleep() and may
 lead too easilly to false positives. Whitelist it.

Reported by:	Erik Cederstrand <erik at cederstrand dot dk>
---
 sys/kern/kern_clock.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/kern/kern_clock.c b/sys/kern/kern_clock.c
index 98e276bb34a..da05cc16356 100644
--- a/sys/kern/kern_clock.c
+++ b/sys/kern/kern_clock.c
@@ -163,6 +163,7 @@ SYSCTL_PROC(_kern, OID_AUTO, cp_times, CTLTYPE_LONG|CTLFLAG_RD|CTLFLAG_MPSAFE,
 
 #ifdef DEADLKRES
 static const char *blessed[] = {
+	"getblk",
 	"so_snd_sx",
 	"so_rcv_sx",
 	NULL

From a318bc273dca5d2fe11b487a42c8a116e12bbc04 Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Tue, 20 Apr 2010 01:02:39 +0000
Subject: [PATCH 150/532] For the experimental NFS client doing an NFSv4 mount,
 set the NFSCLFLAGS_RECVRINPROG while doing recovery from an expired lease in
 a manner similar to r206818 for server reboot recovery. This will prevent the
 function that acquires stateids for I/O operations from acquiring out of date
 stateids during recovery. Also, fix up mutex locking on the nfsc_flags field.

MFC after:	1 week
---
 sys/fs/nfsclient/nfs_clstate.c | 9 ++++++---
 1 file changed, 6 insertions(+), 3 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index 6a189635ff7..78b596640cf 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -2111,6 +2111,7 @@ nfscl_hasexpired(struct nfsclclient *clp, u_int32_t clidrev, NFSPROC_T *p)
 		NFSUNLOCKCLSTATE();
 		return (0);
 	}
+	clp->nfsc_flags |= NFSCLFLAGS_RECVRINPROG;
 	NFSUNLOCKCLSTATE();
 
 	nmp = clp->nfsc_nmp;
@@ -2127,6 +2128,7 @@ nfscl_hasexpired(struct nfsclclient *clp, u_int32_t clidrev, NFSPROC_T *p)
 		 * Clear out any state.
 		 */
 		nfscl_cleanclient(clp);
+		NFSLOCKCLSTATE();
 		clp->nfsc_flags &= ~(NFSCLFLAGS_HASCLIENTID |
 		    NFSCLFLAGS_RECOVER);
 	} else {
@@ -2140,14 +2142,15 @@ nfscl_hasexpired(struct nfsclclient *clp, u_int32_t clidrev, NFSPROC_T *p)
 		 * Expire the state for the client.
 		 */
 		nfscl_expireclient(clp, nmp, cred, p);
+		NFSLOCKCLSTATE();
 		clp->nfsc_flags |= NFSCLFLAGS_HASCLIENTID;
 		clp->nfsc_flags &= ~NFSCLFLAGS_RECOVER;
 	}
-	NFSFREECRED(cred);
-	clp->nfsc_flags &= ~NFSCLFLAGS_EXPIREIT;
-	NFSLOCKCLSTATE();
+	clp->nfsc_flags &= ~(NFSCLFLAGS_EXPIREIT | NFSCLFLAGS_RECVRINPROG);
+	wakeup(&clp->nfsc_flags);
 	nfsv4_unlock(&clp->nfsc_lock, 0);
 	NFSUNLOCKCLSTATE();
+	NFSFREECRED(cred);
 	return (error);
 }
 

From 0d526fa82972822897b544107dcd0d56a50f8d3c Mon Sep 17 00:00:00 2001
From: Nathan Whitehorn <nwhitehorn@FreeBSD.org>
Date: Tue, 20 Apr 2010 01:12:23 +0000
Subject: [PATCH 151/532] Add gpart and glabel to the release CD mfsroot. Even
 if sysinstall cannot partition disks on powerpc, this will allow the user to.

PR:		powerpc/93203
Obtained from:	ia64
MFC after:	1 week
---
 release/powerpc/boot_crunch.conf | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/release/powerpc/boot_crunch.conf b/release/powerpc/boot_crunch.conf
index 9f7a3010b11..4868519c34e 100644
--- a/release/powerpc/boot_crunch.conf
+++ b/release/powerpc/boot_crunch.conf
@@ -15,6 +15,7 @@ srcdirs /usr/src/sbin
 progs camcontrol
 progs dhclient
 progs fsck_ffs
+progs geom
 progs ifconfig
 progs mount_msdosfs
 progs mount_nfs
@@ -25,6 +26,8 @@ progs rtsol
 progs tunefs
 ln fsck_ffs fsck_4.2bsd
 ln fsck_ffs fsck_ufs
+ln geom glabel
+ln geom gpart
 
 srcdirs /usr/src/usr.bin
 progs cpio
@@ -43,4 +46,4 @@ progs usbconfig
 
 libs -ll -ledit -lutil -lmd -lcrypt -lftpio -lz -lnetgraph
 libs -ldialog -lncurses -ldisk -lcam -lkiconv -lsbuf -lufs
-libs -lbsdxml -larchive -lbz2 -lusb -ljail
+libs -lgeom -lbsdxml -larchive -lbz2 -lusb -ljail

From 5d4a7b794595092d7946546d8fee39cf63b4e393 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Tue, 20 Apr 2010 04:16:39 +0000
Subject: [PATCH 152/532] Eliminate an unnecessary call to pmap_remove_all(). 
 If a page belongs to an object whose reference count is zero, then that page
 cannot possibly be mapped.

---
 sys/vm/vm_pageout.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index d5ede5a7503..735beee4b8b 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -1122,7 +1122,8 @@ unlock_and_continue:
 			    m->act_count == 0) {
 				page_shortage--;
 				if (object->ref_count == 0) {
-					pmap_remove_all(m);
+					KASSERT(!pmap_page_is_mapped(m),
+				    ("vm_pageout_scan: page %p is mapped", m));
 					if (m->dirty == 0)
 						vm_page_cache(m);
 					else

From a8734ffdb3cb5e4e2a29628843cb9a8f4b76991a Mon Sep 17 00:00:00 2001
From: Maxim Konovalov <maxim@FreeBSD.org>
Date: Tue, 20 Apr 2010 06:10:05 +0000
Subject: [PATCH 153/532] o Add do-not-fragment option support to ping6(8).

PR:		bin/145759
Submitted by:	pluknet
MFC after:	1 month
---
 sbin/ping6/ping6.8 |  8 +++++---
 sbin/ping6/ping6.c | 14 +++++++++++---
 2 files changed, 16 insertions(+), 6 deletions(-)

diff --git a/sbin/ping6/ping6.8 b/sbin/ping6/ping6.8
index 3897a90c473..3298deaa5a2 100644
--- a/sbin/ping6/ping6.8
+++ b/sbin/ping6/ping6.8
@@ -29,7 +29,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd August 27, 2008
+.Dd April 20, 2010
 .Dt PING6 8
 .Os
 .Sh NAME
@@ -40,9 +40,9 @@ packets to network hosts
 .Sh SYNOPSIS
 .Nm
 .\" without ipsec, or new ipsec
-.Op Fl dfHmnNoqrRtvwW
+.Op Fl DdfHmnNoqrRtvwW
 .\" old ipsec
-.\" .Op Fl AdEfmnNqRtvwW
+.\" .Op Fl ADdEfmnNqRtvwW
 .Bk -words
 .Op Fl a Ar addrtype
 .Ek
@@ -141,6 +141,8 @@ Stop after sending
 .Ar count
 .Tn ECHO_RESPONSE
 packets.
+.It Fl D
+Disable IPv6 fragmentation.
 .It Fl d
 Set the
 .Dv SO_DEBUG
diff --git a/sbin/ping6/ping6.c b/sbin/ping6/ping6.c
index f7dba273650..69a98b23500 100644
--- a/sbin/ping6/ping6.c
+++ b/sbin/ping6/ping6.c
@@ -191,6 +191,7 @@ struct tv32 {
 #define F_ONCE		0x200000
 #define F_AUDIBLE	0x400000
 #define F_MISSED	0x800000
+#define F_DONTFRAG	0x1000000
 #define F_NOUSERDATA	(F_NODEADDR | F_FQDN | F_FQDNOLD | F_SUPTYPES)
 u_int options;
 
@@ -349,7 +350,7 @@ main(argc, argv)
 #endif /*IPSEC_POLICY_IPSEC*/
 #endif
 	while ((ch = getopt(argc, argv,
-	    "a:b:c:dfHg:h:I:i:l:mnNop:qrRS:s:tvwW" ADDOPTS)) != -1) {
+	    "a:b:c:DdfHg:h:I:i:l:mnNop:qrRS:s:tvwW" ADDOPTS)) != -1) {
 #undef ADDOPTS
 		switch (ch) {
 		case 'a':
@@ -415,6 +416,9 @@ main(argc, argv)
 				errx(1,
 				    "illegal number of packets -- %s", optarg);
 			break;
+		case 'D':
+			options |= F_DONTFRAG;
+			break;
 		case 'd':
 			options |= F_SO_DEBUG;
 			break;
@@ -742,7 +746,11 @@ main(argc, argv)
 	for (i = 0; i < sizeof(nonce); i += sizeof(u_int32_t))
 		*((u_int32_t *)&nonce[i]) = arc4random();
 #endif
-
+	optval = 1;
+	if (options & F_DONTFRAG)
+		if (setsockopt(s, IPPROTO_IPV6, IPV6_DONTFRAG,
+		    &optval, sizeof(optval)) == -1)
+			err(1, "IPV6_DONTFRAG");
 	hold = 1;
 
 	if (options & F_SO_DEBUG)
@@ -2780,7 +2788,7 @@ usage()
 	    "A"
 #endif
 	    "usage: ping6 [-"
-	    "d"
+	    "Dd"
 #if defined(IPSEC) && !defined(IPSEC_POLICY_IPSEC)
 	    "E"
 #endif

From 553aff12d462be96f31ded5685c391c9c402d66a Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Tue, 20 Apr 2010 08:50:19 +0000
Subject: [PATCH 154/532] Really print the nr_mapping array when it should be
 printed.`

MFC after: 3 days.
---
 sys/netinet/sctputil.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/netinet/sctputil.c b/sys/netinet/sctputil.c
index 0f2805fb613..26122ce4abf 100644
--- a/sys/netinet/sctputil.c
+++ b/sys/netinet/sctputil.c
@@ -1215,7 +1215,7 @@ sctp_print_mapping_array(struct sctp_association *asoc)
 	}
 	printf("Non renegable mapping array (last %d entries are zero):\n", asoc->mapping_array_size - limit);
 	for (i = 0; i < limit; i++) {
-		printf("%2.2x%c", asoc->mapping_array[i], ((i + 1) % 16) ? ' ' : '\n');
+		printf("%2.2x%c", asoc->nr_mapping_array[i], ((i + 1) % 16) ? ' ' : '\n');
 	}
 	if (limit % 16)
 		printf("\n");

From ee94f0a272922e3e5a9fa907e5e73497991b6fcf Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Tue, 20 Apr 2010 08:51:21 +0000
Subject: [PATCH 155/532] Update highest_tsn variables when sliding mapping
 arrays.

---
 sys/netinet/sctp_indata.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/sys/netinet/sctp_indata.c b/sys/netinet/sctp_indata.c
index 97a9733ab84..c1430a53f40 100644
--- a/sys/netinet/sctp_indata.c
+++ b/sys/netinet/sctp_indata.c
@@ -2390,6 +2390,12 @@ sctp_slide_mapping_arrays(struct sctp_tcb *stcb)
 				asoc->mapping_array[ii] = 0;
 				asoc->nr_mapping_array[ii] = 0;
 			}
+			if (asoc->highest_tsn_inside_map + 1 == asoc->mapping_array_base_tsn) {
+				asoc->highest_tsn_inside_map += (slide_from << 3);
+			}
+			if (asoc->highest_tsn_inside_nr_map + 1 == asoc->mapping_array_base_tsn) {
+				asoc->highest_tsn_inside_nr_map += (slide_from << 3);
+			}
 			asoc->mapping_array_base_tsn += (slide_from << 3);
 			if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_MAP_LOGGING_ENABLE) {
 				sctp_log_map(asoc->mapping_array_base_tsn,

From 9d79ec20fb18158492642c73fb8a17b531308a7d Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 20 Apr 2010 10:16:44 +0000
Subject: [PATCH 156/532] Slightly modernize realpath(3).

SUSv4 requires that implementation returns EINVAL if supplied path is NULL,
and ENOENT if path is empty string [1].
Bring prototype in conformance with SUSv4, adding restrict keywords.
Allow the resolved path buffer pointer be NULL, in which case realpath(3)
allocates storage with malloc().

PR:	kern/121897 [1]
MFC after:	2 weeks
---
 include/stdlib.h           |  2 +-
 lib/libc/stdlib/realpath.3 | 19 +++++++++++---
 lib/libc/stdlib/realpath.c | 54 ++++++++++++++++++++++++++++++++++----
 3 files changed, 66 insertions(+), 9 deletions(-)

diff --git a/include/stdlib.h b/include/stdlib.h
index b8cd1a389f8..b833203c65b 100644
--- a/include/stdlib.h
+++ b/include/stdlib.h
@@ -201,7 +201,7 @@ int	 posix_openpt(int);
 char	*ptsname(int);
 int	 putenv(char *);
 long	 random(void);
-char	*realpath(const char *, char resolved_path[]);
+char	*realpath(const char * __restrict, char * __restrict);
 unsigned short
 	*seed48(unsigned short[3]);
 #ifndef _SETKEY_DECLARED
diff --git a/lib/libc/stdlib/realpath.3 b/lib/libc/stdlib/realpath.3
index 4205a3d6bbc..371ee3b2804 100644
--- a/lib/libc/stdlib/realpath.3
+++ b/lib/libc/stdlib/realpath.3
@@ -31,7 +31,7 @@
 .\"     @(#)realpath.3	8.2 (Berkeley) 2/16/94
 .\" $FreeBSD$
 .\"
-.Dd February 16, 1994
+.Dd April 19, 2010
 .Dt REALPATH 3
 .Os
 .Sh NAME
@@ -43,7 +43,7 @@
 .In sys/param.h
 .In stdlib.h
 .Ft "char *"
-.Fn realpath "const char *pathname" "char resolved_path[PATH_MAX]"
+.Fn realpath "const char *pathname" "char *resolved_path"
 .Sh DESCRIPTION
 The
 .Fn realpath
@@ -64,7 +64,8 @@ argument
 .Em must
 refer to a buffer capable of storing at least
 .Dv PATH_MAX
-characters.
+characters, or be
+.Dv NULL .
 .Pp
 The
 .Fn realpath
@@ -82,6 +83,13 @@ The
 function returns
 .Fa resolved_path
 on success.
+If the function was supplied
+.Dv NULL
+as
+.Fa resolved_path ,
+and operation did not cause errors, the returned value is
+a null-terminated string in a buffer allocated by a call to
+.Fn malloc 3 .
 If an error occurs,
 .Fn realpath
 returns
@@ -89,6 +97,11 @@ returns
 and
 .Fa resolved_path
 contains the pathname which caused the problem.
+If
+.Fa resolved_path
+was
+.Dv NULL ,
+then information of the failed pathname component is lost.
 .Sh ERRORS
 The function
 .Fn realpath
diff --git a/lib/libc/stdlib/realpath.c b/lib/libc/stdlib/realpath.c
index 3082f5f4b95..31c93e67741 100644
--- a/lib/libc/stdlib/realpath.c
+++ b/lib/libc/stdlib/realpath.c
@@ -43,23 +43,37 @@ __FBSDID("$FreeBSD$");
 #include "un-namespace.h"
 
 /*
- * char *realpath(const char *path, char resolved[PATH_MAX]);
- *
  * Find the real name of path, by removing all ".", ".." and symlink
  * components.  Returns (resolved) on success, or (NULL) on failure,
  * in which case the path which caused trouble is left in (resolved).
  */
 char *
-realpath(const char *path, char resolved[PATH_MAX])
+realpath(const char * __restrict path, char * __restrict resolved)
 {
 	struct stat sb;
 	char *p, *q, *s;
 	size_t left_len, resolved_len;
 	unsigned symlinks;
-	int serrno, slen;
+	int serrno, slen, m;
 	char left[PATH_MAX], next_token[PATH_MAX], symlink[PATH_MAX];
 
+	if (path == NULL) {
+		errno = EINVAL;
+		return (NULL);
+	}
+	if (path[0] == '\0') {
+		errno = ENOENT;
+		return (NULL);
+	}
 	serrno = errno;
+	if (resolved == NULL) {
+		resolved = malloc(PATH_MAX);
+		if (resolved == NULL)
+			return (NULL);
+		m = 1;
+	} else
+		m = 0;
+
 	symlinks = 0;
 	if (path[0] == '/') {
 		resolved[0] = '/';
@@ -71,12 +85,19 @@ realpath(const char *path, char resolved[PATH_MAX])
 	} else {
 		if (getcwd(resolved, PATH_MAX) == NULL) {
 			strlcpy(resolved, ".", PATH_MAX);
+			if (m) {
+				serrno = errno;
+				free(resolved);
+				errno = serrno;
+			}
 			return (NULL);
 		}
 		resolved_len = strlen(resolved);
 		left_len = strlcpy(left, path, sizeof(left));
 	}
 	if (left_len >= sizeof(left) || resolved_len >= PATH_MAX) {
+		if (m)
+			free(resolved);
 		errno = ENAMETOOLONG;
 		return (NULL);
 	}
@@ -92,6 +113,8 @@ realpath(const char *path, char resolved[PATH_MAX])
 		p = strchr(left, '/');
 		s = p ? p : left + left_len;
 		if (s - left >= sizeof(next_token)) {
+			if (m)
+				free(resolved);
 			errno = ENAMETOOLONG;
 			return (NULL);
 		}
@@ -102,6 +125,8 @@ realpath(const char *path, char resolved[PATH_MAX])
 			memmove(left, s + 1, left_len + 1);
 		if (resolved[resolved_len - 1] != '/') {
 			if (resolved_len + 1 >= PATH_MAX) {
+				if (m)
+					free(resolved);
 				errno = ENAMETOOLONG;
 				return (NULL);
 			}
@@ -133,6 +158,8 @@ realpath(const char *path, char resolved[PATH_MAX])
 		 */
 		resolved_len = strlcat(resolved, next_token, PATH_MAX);
 		if (resolved_len >= PATH_MAX) {
+			if (m)
+				free(resolved);
 			errno = ENAMETOOLONG;
 			return (NULL);
 		}
@@ -141,16 +168,29 @@ realpath(const char *path, char resolved[PATH_MAX])
 				errno = serrno;
 				return (resolved);
 			}
+			if (m) {
+				serrno = errno;
+				free(resolved);
+				errno = serrno;
+			}
 			return (NULL);
 		}
 		if (S_ISLNK(sb.st_mode)) {
 			if (symlinks++ > MAXSYMLINKS) {
+				if (m)
+					free(resolved);
 				errno = ELOOP;
 				return (NULL);
 			}
 			slen = readlink(resolved, symlink, sizeof(symlink) - 1);
-			if (slen < 0)
+			if (slen < 0) {
+				if (m) {
+					serrno = errno;
+					free(resolved);
+					errno = serrno;
+				}
 				return (NULL);
+			}
 			symlink[slen] = '\0';
 			if (symlink[0] == '/') {
 				resolved[1] = 0;
@@ -171,6 +211,8 @@ realpath(const char *path, char resolved[PATH_MAX])
 			if (p != NULL) {
 				if (symlink[slen - 1] != '/') {
 					if (slen + 1 >= sizeof(symlink)) {
+						if (m)
+							free(resolved);
 						errno = ENAMETOOLONG;
 						return (NULL);
 					}
@@ -179,6 +221,8 @@ realpath(const char *path, char resolved[PATH_MAX])
 				}
 				left_len = strlcat(symlink, left, sizeof(left));
 				if (left_len >= sizeof(left)) {
+					if (m)
+						free(resolved);
 					errno = ENAMETOOLONG;
 					return (NULL);
 				}

From 5673e3cb08f0b3d813ce50e658fbc822e200904a Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 20 Apr 2010 10:19:27 +0000
Subject: [PATCH 157/532] The cache_enter(9) function shall not be called for
 doomed dvp. Assert this.

In the reported panic, vdestroy() fired the assertion "vp has namecache
for ..", because pseudofs may end up doing cache_enter() with reclaimed
dvp, after dotdot lookup temporary unlocked dvp.
Similar problem exists in ufs_lookup() for "." lookup, when vnode
lock needs to be upgraded.

Verify that dvp is not reclaimed before calling cache_enter().

Reported and tested by:	pho
Reviewed by:	kan
MFC after:	2 weeks
---
 sys/fs/pseudofs/pseudofs_vnops.c | 2 +-
 sys/kern/vfs_cache.c             | 2 ++
 sys/ufs/ufs/ufs_lookup.c         | 8 ++++++++
 3 files changed, 11 insertions(+), 1 deletion(-)

diff --git a/sys/fs/pseudofs/pseudofs_vnops.c b/sys/fs/pseudofs/pseudofs_vnops.c
index 5854378336c..f8343a97401 100644
--- a/sys/fs/pseudofs/pseudofs_vnops.c
+++ b/sys/fs/pseudofs/pseudofs_vnops.c
@@ -542,7 +542,7 @@ pfs_lookup(struct vop_cachedlookup_args *va)
 
 	if (cnp->cn_flags & ISDOTDOT)
 		vn_lock(vn, LK_EXCLUSIVE|LK_RETRY);
-	if (cnp->cn_flags & MAKEENTRY)
+	if (cnp->cn_flags & MAKEENTRY && !(vn->v_iflag & VI_DOOMED))
 		cache_enter(vn, *vpp, cnp);
 	PFS_RETURN (0);
  failed:
diff --git a/sys/kern/vfs_cache.c b/sys/kern/vfs_cache.c
index 533ec97875d..a13a721850d 100644
--- a/sys/kern/vfs_cache.c
+++ b/sys/kern/vfs_cache.c
@@ -611,6 +611,8 @@ cache_enter(dvp, vp, cnp)
 	CTR3(KTR_VFS, "cache_enter(%p, %p, %s)", dvp, vp, cnp->cn_nameptr);
 	VNASSERT(vp == NULL || (vp->v_iflag & VI_DOOMED) == 0, vp,
 	    ("cache_enter: Adding a doomed vnode"));
+	VNASSERT(dvp == NULL || (dvp->v_iflag & VI_DOOMED) == 0, dvp,
+	    ("cache_enter: Doomed vnode used as src"));
 
 	if (!doingcache)
 		return;
diff --git a/sys/ufs/ufs/ufs_lookup.c b/sys/ufs/ufs/ufs_lookup.c
index b0247e77d0d..ab71cf6df55 100644
--- a/sys/ufs/ufs/ufs_lookup.c
+++ b/sys/ufs/ufs/ufs_lookup.c
@@ -704,6 +704,14 @@ found:
 				vn_lock(vdp, LK_UPGRADE | LK_RETRY);
 			else /* if (ltype == LK_SHARED) */
 				vn_lock(vdp, LK_DOWNGRADE | LK_RETRY);
+			/*
+			 * Relock for the "." case may left us with
+			 * reclaimed vnode.
+			 */
+			if (vdp->v_iflag & VI_DOOMED) {
+				vrele(vdp);
+				return (ENOENT);
+			}
 		}
 		*vpp = vdp;
 	} else {

From 153bfab76a878d55c9fe0958d331cf98f9b8a055 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Tue, 20 Apr 2010 10:42:08 +0000
Subject: [PATCH 158/532] Remove svn:executable prop.

---
 sys/mips/rmi/debug.h                  | 0
 sys/mips/rmi/msgring.h                | 0
 sys/mips/rmi/shared_structs.h         | 0
 sys/mips/rmi/shared_structs_func.h    | 0
 sys/mips/rmi/shared_structs_offsets.h | 0
 5 files changed, 0 insertions(+), 0 deletions(-)
 mode change 100755 => 100644 sys/mips/rmi/debug.h
 mode change 100755 => 100644 sys/mips/rmi/msgring.h
 mode change 100755 => 100644 sys/mips/rmi/shared_structs.h
 mode change 100755 => 100644 sys/mips/rmi/shared_structs_func.h
 mode change 100755 => 100644 sys/mips/rmi/shared_structs_offsets.h

diff --git a/sys/mips/rmi/debug.h b/sys/mips/rmi/debug.h
old mode 100755
new mode 100644
diff --git a/sys/mips/rmi/msgring.h b/sys/mips/rmi/msgring.h
old mode 100755
new mode 100644
diff --git a/sys/mips/rmi/shared_structs.h b/sys/mips/rmi/shared_structs.h
old mode 100755
new mode 100644
diff --git a/sys/mips/rmi/shared_structs_func.h b/sys/mips/rmi/shared_structs_func.h
old mode 100755
new mode 100644
diff --git a/sys/mips/rmi/shared_structs_offsets.h b/sys/mips/rmi/shared_structs_offsets.h
old mode 100755
new mode 100644

From 0a2d5fea5921acbf8945307784cf6e4b7924a95a Mon Sep 17 00:00:00 2001
From: Attilio Rao <attilio@FreeBSD.org>
Date: Tue, 20 Apr 2010 12:22:06 +0000
Subject: [PATCH 159/532] Fix compilation in the !SMP case. Keep the interrupts
 disabled in order to avoid preemption problems.

Reported by:	tinderbox, b.f. <bf1783 at googlemail dot com>
MFC:		2 weeks
X-MFC:		r206878
---
 sys/kern/kern_shutdown.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/sys/kern/kern_shutdown.c b/sys/kern/kern_shutdown.c
index 4b27d45f6c0..bd3c7aa478c 100644
--- a/sys/kern/kern_shutdown.c
+++ b/sys/kern/kern_shutdown.c
@@ -496,8 +496,14 @@ shutdown_reset(void *junk, int howto)
 	 *   generally, threads busy-waiting, with this spinlock held,
 	 *   and waiting for responses by threads on other CPUs
 	 *   (ie. smp_tlb_shootdown()).
+	 *
+	 * For the !SMP case it just needs to handle the former problem.
 	 */
+#ifdef SMP
 	mtx_lock_spin(&smp_ipi_mtx);
+#else
+	spinlock_enter();
+#endif
 
 	/* cpu_boot(howto); */ /* doesn't do anything at the moment */
 	cpu_reset();

From 655c8a600bb02846e05f137609abb62134a13a26 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 20 Apr 2010 14:22:29 +0000
Subject: [PATCH 160/532] Free() is not allowed to modify errno, remove safety
 brackets around it [1]. Add small optimization, do not copy a string to the
 buffer that is to be freed immediately after.

Noted by:	jh [1]
Reviewed by:	jh
MFC after:	2 weeks
---
 lib/libc/stdlib/realpath.c | 18 +++++-------------
 1 file changed, 5 insertions(+), 13 deletions(-)

diff --git a/lib/libc/stdlib/realpath.c b/lib/libc/stdlib/realpath.c
index 31c93e67741..e75ee4ac823 100644
--- a/lib/libc/stdlib/realpath.c
+++ b/lib/libc/stdlib/realpath.c
@@ -84,12 +84,10 @@ realpath(const char * __restrict path, char * __restrict resolved)
 		left_len = strlcpy(left, path + 1, sizeof(left));
 	} else {
 		if (getcwd(resolved, PATH_MAX) == NULL) {
-			strlcpy(resolved, ".", PATH_MAX);
-			if (m) {
-				serrno = errno;
+			if (m)
 				free(resolved);
-				errno = serrno;
-			}
+			else
+				strlcpy(resolved, ".", PATH_MAX);
 			return (NULL);
 		}
 		resolved_len = strlen(resolved);
@@ -168,11 +166,8 @@ realpath(const char * __restrict path, char * __restrict resolved)
 				errno = serrno;
 				return (resolved);
 			}
-			if (m) {
-				serrno = errno;
+			if (m)
 				free(resolved);
-				errno = serrno;
-			}
 			return (NULL);
 		}
 		if (S_ISLNK(sb.st_mode)) {
@@ -184,11 +179,8 @@ realpath(const char * __restrict path, char * __restrict resolved)
 			}
 			slen = readlink(resolved, symlink, sizeof(symlink) - 1);
 			if (slen < 0) {
-				if (m) {
-					serrno = errno;
+				if (m)
 					free(resolved);
-					errno = serrno;
-				}
 				return (NULL);
 			}
 			symlink[slen] = '\0';

From b527a2a4dc2a83b4da9a2b2fa6cdc6db4733da12 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Tue, 20 Apr 2010 16:30:17 +0000
Subject: [PATCH 161/532] The amd64 version of the cyclic dtrace module is a
 verbatim copy of the i386 version, so instead having a copy of the same file,
 use Makefile foo to include the i386 version on amd64.

---
 sys/cddl/dev/cyclic/amd64/cyclic_machdep.c | 133 ---------------------
 sys/modules/cyclic/Makefile                |   2 +-
 2 files changed, 1 insertion(+), 134 deletions(-)
 delete mode 100644 sys/cddl/dev/cyclic/amd64/cyclic_machdep.c

diff --git a/sys/cddl/dev/cyclic/amd64/cyclic_machdep.c b/sys/cddl/dev/cyclic/amd64/cyclic_machdep.c
deleted file mode 100644
index 9a15b2cdad0..00000000000
--- a/sys/cddl/dev/cyclic/amd64/cyclic_machdep.c
+++ /dev/null
@@ -1,133 +0,0 @@
-/*-
- * Copyright 2007 John Birrell <jb@FreeBSD.org>
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- * 
- * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- *
- * $FreeBSD$
- *
- */
-
-static void enable(cyb_arg_t);
-static void disable(cyb_arg_t);
-static void reprogram(cyb_arg_t, hrtime_t);
-static void xcall(cyb_arg_t, cpu_t *, cyc_func_t, void *);
-
-static cyc_backend_t	be	= {
-	NULL,		/* cyb_configure */
-	NULL,		/* cyb_unconfigure */
-	enable,
-	disable,
-	reprogram,
-	xcall,
-	NULL		/* cyb_arg_t cyb_arg */
-};
-
-static void
-cyclic_ap_start(void *dummy)
-{
-	/* Initialise the rest of the CPUs. */
-	cyclic_mp_init();
-}
-
-SYSINIT(cyclic_ap_start, SI_SUB_SMP, SI_ORDER_ANY, cyclic_ap_start, NULL);
-
-/*
- *  Machine dependent cyclic subsystem initialisation.
- */
-static void
-cyclic_machdep_init(void)
-{
-	/* Register the cyclic backend. */
-	cyclic_init(&be);
-}
-
-static void
-cyclic_machdep_uninit(void)
-{
-	int i;
-
-	for (i = 0; i <= mp_maxid; i++)
-		/* Reset the cyclic clock callback hook. */
-		lapic_cyclic_clock_func[i] = NULL;
-
-	/* De-register the cyclic backend. */
-	cyclic_uninit();
-}
-
-static hrtime_t exp_due[MAXCPU];
-
-/*
- * This function is the one registered by the machine dependent
- * initialiser as the callback for high speed timer events.
- */
-static void
-cyclic_clock(struct trapframe *frame)
-{
-	cpu_t *c = &solaris_cpu[curcpu];
-
-	if (c->cpu_cyclic != NULL && gethrtime() >= exp_due[curcpu]) {
-		if (TRAPF_USERMODE(frame)) {
-			c->cpu_profile_pc = 0;
-			c->cpu_profile_upc = TRAPF_PC(frame);
-		} else {
-			c->cpu_profile_pc = TRAPF_PC(frame);
-			c->cpu_profile_upc = 0;
-		}
-
-		c->cpu_intr_actv = 1;
-
-		/* Fire any timers that are due. */
-		cyclic_fire(c);
-
-		c->cpu_intr_actv = 0;
-	}
-}
-
-static void enable(cyb_arg_t arg)
-{
-	/* Register the cyclic clock callback function. */
-	lapic_cyclic_clock_func[curcpu] = cyclic_clock;
-}
-
-static void disable(cyb_arg_t arg)
-{
-	/* Reset the cyclic clock callback function. */
-	lapic_cyclic_clock_func[curcpu] = NULL;
-}
-
-static void reprogram(cyb_arg_t arg, hrtime_t exp)
-{
-	exp_due[curcpu] = exp;
-}
-
-static void xcall(cyb_arg_t arg, cpu_t *c, cyc_func_t func, void *param)
-{
-	/*
-	 * If the target CPU is the current one, just call the
-	 * function. This covers the non-SMP case.
-	 */
-	if (c == &solaris_cpu[curcpu])
-		(*func)(param);
-	else
-		smp_rendezvous_cpus((cpumask_t) (1 << c->cpuid), NULL,
-		    func, smp_no_rendevous_barrier, param);
-}
diff --git a/sys/modules/cyclic/Makefile b/sys/modules/cyclic/Makefile
index db99488e138..371dac68054 100644
--- a/sys/modules/cyclic/Makefile
+++ b/sys/modules/cyclic/Makefile
@@ -10,7 +10,7 @@ SRCS+=		vnode_if.h
 CFLAGS+=	-I${.CURDIR}/../../cddl/compat/opensolaris		\
 		-I${.CURDIR}/../../cddl/contrib/opensolaris/uts/common	\
 		-I${.CURDIR}/../..					\
-		-I${.CURDIR}/../../cddl/dev/cyclic/${MACHINE_ARCH}
+		-I${.CURDIR}/../../cddl/dev/cyclic/${MACHINE_ARCH:S/amd64/i386/}
 
 CFLAGS+=	-DDEBUG=1
 

From ff569d8436052b01c102a791ff60caf4a05b0319 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Tue, 20 Apr 2010 17:03:30 +0000
Subject: [PATCH 162/532] Rename the cyclic global variable
 lapic_cyclic_clock_func to just cyclic_clock_func. This will make more sense
 when we start developing non x86 cyclic version.

---
 sys/amd64/amd64/local_apic.c              | 6 +++---
 sys/cddl/dev/cyclic/i386/cyclic_machdep.c | 6 +++---
 sys/i386/i386/local_apic.c                | 6 +++---
 sys/sys/dtrace_bsd.h                      | 2 +-
 4 files changed, 10 insertions(+), 10 deletions(-)

diff --git a/sys/amd64/amd64/local_apic.c b/sys/amd64/amd64/local_apic.c
index 8edc971c95c..c8f60f06675 100644
--- a/sys/amd64/amd64/local_apic.c
+++ b/sys/amd64/amd64/local_apic.c
@@ -70,7 +70,7 @@ __FBSDID("$FreeBSD$");
 
 #ifdef KDTRACE_HOOKS
 #include <sys/dtrace_bsd.h>
-cyclic_clock_func_t	lapic_cyclic_clock_func[MAXCPU];
+cyclic_clock_func_t	cyclic_clock_func[MAXCPU];
 #endif
 
 /* Sanity checks on IDT vectors. */
@@ -778,8 +778,8 @@ lapic_handle_timer(struct trapframe *frame)
 	 * timers.
 	 */
 	int cpu = PCPU_GET(cpuid);
-	if (lapic_cyclic_clock_func[cpu] != NULL)
-		(*lapic_cyclic_clock_func[cpu])(frame);
+	if (cyclic_clock_func[cpu] != NULL)
+		(*cyclic_clock_func[cpu])(frame);
 #endif
 
 	/* Fire hardclock at hz. */
diff --git a/sys/cddl/dev/cyclic/i386/cyclic_machdep.c b/sys/cddl/dev/cyclic/i386/cyclic_machdep.c
index fa40db9329c..0b6ab5936e2 100644
--- a/sys/cddl/dev/cyclic/i386/cyclic_machdep.c
+++ b/sys/cddl/dev/cyclic/i386/cyclic_machdep.c
@@ -67,7 +67,7 @@ cyclic_machdep_uninit(void)
 
 	for (i = 0; i <= mp_maxid; i++)
 		/* Reset the cyclic clock callback hook. */
-		lapic_cyclic_clock_func[i] = NULL;
+		cyclic_clock_func[i] = NULL;
 
 	/* De-register the cyclic backend. */
 	cyclic_uninit();
@@ -105,13 +105,13 @@ cyclic_clock(struct trapframe *frame)
 static void enable(cyb_arg_t arg)
 {
 	/* Register the cyclic clock callback function. */
-	lapic_cyclic_clock_func[curcpu] = cyclic_clock;
+	cyclic_clock_func[curcpu] = cyclic_clock;
 }
 
 static void disable(cyb_arg_t arg)
 {
 	/* Reset the cyclic clock callback function. */
-	lapic_cyclic_clock_func[curcpu] = NULL;
+	cyclic_clock_func[curcpu] = NULL;
 }
 
 static void reprogram(cyb_arg_t arg, hrtime_t exp)
diff --git a/sys/i386/i386/local_apic.c b/sys/i386/i386/local_apic.c
index 302535f27f9..d894d3271d0 100644
--- a/sys/i386/i386/local_apic.c
+++ b/sys/i386/i386/local_apic.c
@@ -71,7 +71,7 @@ __FBSDID("$FreeBSD$");
 
 #ifdef KDTRACE_HOOKS
 #include <sys/dtrace_bsd.h>
-cyclic_clock_func_t	lapic_cyclic_clock_func[MAXCPU];
+cyclic_clock_func_t	cyclic_clock_func[MAXCPU];
 #endif
 
 /* Sanity checks on IDT vectors. */
@@ -779,8 +779,8 @@ lapic_handle_timer(struct trapframe *frame)
 	 * timers.
 	 */
 	int cpu = PCPU_GET(cpuid);
-	if (lapic_cyclic_clock_func[cpu] != NULL)
-		(*lapic_cyclic_clock_func[cpu])(frame);
+	if (cyclic_clock_func[cpu] != NULL)
+		(*cyclic_clock_func[cpu])(frame);
 #endif
 
 	/* Fire hardclock at hz. */
diff --git a/sys/sys/dtrace_bsd.h b/sys/sys/dtrace_bsd.h
index f3232841fa0..a14a1a160f3 100644
--- a/sys/sys/dtrace_bsd.h
+++ b/sys/sys/dtrace_bsd.h
@@ -50,7 +50,7 @@ typedef	void (*cyclic_clock_func_t)(struct trapframe *);
  *
  * Defining them here avoids a proliferation of header files.
  */
-extern cyclic_clock_func_t     lapic_cyclic_clock_func[];
+extern cyclic_clock_func_t     cyclic_clock_func[];
 
 /*
  * The dtrace module handles traps that occur during a DTrace probe.

From 0017771c681d770dcec4aec1efb48438b82fd814 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Tue, 20 Apr 2010 17:22:20 +0000
Subject: [PATCH 163/532] Add the necessary hooks for dtrace cyclic module.

---
 sys/mips/mips/tick.c | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/sys/mips/mips/tick.c b/sys/mips/mips/tick.c
index d2678b1f637..9a9108c0b83 100644
--- a/sys/mips/mips/tick.c
+++ b/sys/mips/mips/tick.c
@@ -295,7 +295,16 @@ clock_intr(void *arg)
 	 */
 	if (delta > cycles_per_hz)
 		delta = cycles_per_hz;
-
+#if KDTRACE_HOOKS
+	/*
+	 * If the DTrace hooks are configured and a callback function
+	 * has been registered, then call it to process the high speed
+	 * timers.
+	 */
+	int cpu = PCPU_GET(cpuid);
+	if (cyclic_clock_func[cpu] != NULL)
+		(*cyclic_clock_func[cpu])(tf);
+#endif
 	/* Fire hardclock at hz. */
 	cpu_ticks->hard_ticks += delta;
 	if (cpu_ticks->hard_ticks >= cycles_per_hz) {

From 0e954830e516457c9d94b6a20ca4e54d3055e8d3 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 20 Apr 2010 17:57:43 +0000
Subject: [PATCH 164/532] Make this file more C++ friendly.

---
 sys/sys/thr.h | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/sys/sys/thr.h b/sys/sys/thr.h
index ae4a65d368f..7ccc8724ffb 100644
--- a/sys/sys/thr.h
+++ b/sys/sys/thr.h
@@ -30,6 +30,7 @@
 #ifndef _SYS_THR_H_
 #define	_SYS_THR_H_
 
+#include <sys/cdefs.h>
 #include <sys/_types.h>
 #include <sys/sched.h>
 
@@ -68,6 +69,7 @@ typedef __pid_t		pid_t;
 #define _PID_T_DECLARED
 #endif
 
+__BEGIN_DECLS
 int thr_create(ucontext_t *ctx, long *id, int flags);
 int thr_new(struct thr_param *param, int param_size);
 int thr_self(long *id);
@@ -77,7 +79,7 @@ int thr_kill2(pid_t pid, long id, int sig);
 int thr_suspend(const struct timespec *timeout);
 int thr_wake(long id);
 int thr_set_name(long id, const char *name);
-
+__END_DECLS
 #endif /* !_KERNEL */
 
 #endif /* ! _SYS_THR_H_ */

From 60e3ce611c7b983ce0d18862cfc112beb4bc74b7 Mon Sep 17 00:00:00 2001
From: Ana Kukec <anchie@FreeBSD.org>
Date: Tue, 20 Apr 2010 18:16:56 +0000
Subject: [PATCH 165/532] Add myself and list bz@ as my mentor.

Approved by:	bz (mentor)
---
 share/misc/committers-src.dot | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/share/misc/committers-src.dot b/share/misc/committers-src.dot
index a52aa272a54..f7e4a326e66 100644
--- a/share/misc/committers-src.dot
+++ b/share/misc/committers-src.dot
@@ -55,6 +55,7 @@ node [color=lightblue2, style=filled, bgcolor=black];
 ache [label="Andrey Chernov\nache@FreeBSD.org\n1993/10/31"]
 akiyama [label="Shunsuke Akiyama\nakiyama@FreeBSD.org\n2000/06/19"]
 ambrisko [label="Doug Ambrisko\nambrisko@FreeBSD.org\n2001/12/19"]
+anchie [label="Ana Kukec\anchie@FreeBSD.org\n2010/04/14"]
 andre [label="Andre Oppermann\nandre@FreeBSD.org\n2003/11/12"]
 anholt [label="Eric Anholt\nanholt@FreeBSD.org\n2002/04/22"]
 antoine [label="Antoine Brodin\nantoine@FreeBSD.org\n2008/02/03"]
@@ -238,6 +239,7 @@ brian -> joe
 brooks -> bushman
 brooks -> jamie
 
+bz -> anchie
 bz -> jamie
 bz -> syrinx
 

From 20c14533dda078fcc83f6ad7f1221933f1fdd87b Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Tue, 20 Apr 2010 18:42:51 +0000
Subject: [PATCH 166/532] Vendor import of zlib 1.2.5.

---
 ChangeLog  | 28 +++++++++++++++++++++++++++-
 README     |  4 ++--
 crc32.c    |  2 +-
 deflate.c  |  2 +-
 deflate.h  | 24 ++++++++++++------------
 gzguts.h   | 12 ++++++++----
 gzlib.c    |  6 +++---
 inffast.c  |  4 ++--
 inffast.h  |  4 ++--
 inftrees.c |  6 +++---
 inftrees.h |  4 ++--
 trees.c    | 17 +++++++++--------
 trees.h    |  4 ++--
 zconf.h    | 14 ++++++++++----
 zlib.3     |  4 ++--
 zlib.h     | 10 +++++-----
 zutil.c    | 24 ++++++++++++------------
 zutil.h    | 22 ++++++++++++++--------
 18 files changed, 117 insertions(+), 74 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 898c197f0cf..f310bb0fcdb 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,6 +1,32 @@
 
                 ChangeLog file for zlib
 
+Changes in 1.2.5 (19 Apr 2010)
+- Disable visibility attribute in win32/Makefile.gcc [Bar-Lev]
+- Default to libdir as sharedlibdir in configure [Nieder]
+- Update copyright dates on modified source files
+- Update trees.c to be able to generate modified trees.h
+- Exit configure for MinGW, suggesting win32/Makefile.gcc
+
+Changes in 1.2.4.5 (18 Apr 2010)
+- Set sharedlibdir in configure [Torok]
+- Set LDFLAGS in Makefile.in [Bar-Lev]
+- Avoid mkdir objs race condition in Makefile.in [Bowler]
+- Add ZLIB_INTERNAL in front of internal inter-module functions and arrays
+- Define ZLIB_INTERNAL to hide internal functions and arrays for GNU C
+- Don't use hidden attribute when it is a warning generator (e.g. Solaris)
+
+Changes in 1.2.4.4 (18 Apr 2010)
+- Fix CROSS_PREFIX executable testing, CHOST extract, mingw* [Torok]
+- Undefine _LARGEFILE64_SOURCE in zconf.h if it is zero, but not if empty
+- Try to use bash or ksh regardless of functionality of /bin/sh
+- Fix configure incompatibility with NetBSD sh
+- Remove attempt to run under bash or ksh since have better NetBSD fix
+- Fix win32/Makefile.gcc for MinGW [Bar-Lev]
+- Add diagnostic messages when using CROSS_PREFIX in configure
+- Added --sharedlibdir option to configure [Weigelt]
+- Use hidden visibility attribute when available [Frysinger]
+
 Changes in 1.2.4.3 (10 Apr 2010)
 - Only use CROSS_PREFIX in configure for ar and ranlib if they exist
 - Use CROSS_PREFIX for nm [Bar-Lev]
@@ -151,7 +177,7 @@ Changes in 1.2.3.6 (17 Jan 2010)
 - Correct email address in configure for system options
 - Update make_vms.com and add make_vms.com to contrib/minizip [Zinser]
 - Update zlib.map [Brown]
-- Fix Makefile.in for Solaris 10 make of example64 and minizip64 [Tšršk]
+- Fix Makefile.in for Solaris 10 make of example64 and minizip64 [Torok]
 - Apply various fixes to CMakeLists.txt [Lowman]
 - Add checks on len in gzread() and gzwrite()
 - Add error message for no more room for gzungetc()
diff --git a/README b/README
index 68ff992369c..d4219bf889f 100644
--- a/README
+++ b/README
@@ -1,6 +1,6 @@
 ZLIB DATA COMPRESSION LIBRARY
 
-zlib 1.2.4.3 is a general purpose data compression library.  All the code is
+zlib 1.2.5 is a general purpose data compression library.  All the code is
 thread safe.  The data format used by the zlib library is described by RFCs
 (Request for Comments) 1950 to 1952 in the files
 http://www.ietf.org/rfc/rfc1950.txt (zlib format), rfc1951.txt (deflate format)
@@ -31,7 +31,7 @@ Mark Nelson <markn@ieee.org> wrote an article about zlib for the Jan.  1997
 issue of Dr.  Dobb's Journal; a copy of the article is available at
 http://marknelson.us/1997/01/01/zlib-engine/ .
 
-The changes made in version 1.2.4.3 are documented in the file ChangeLog.
+The changes made in version 1.2.5 are documented in the file ChangeLog.
 
 Unsupported third party contributions are provided in directory contrib/ .
 
diff --git a/crc32.c b/crc32.c
index 6b81fb4a1ca..91be372d224 100644
--- a/crc32.c
+++ b/crc32.c
@@ -1,5 +1,5 @@
 /* crc32.c -- compute the CRC-32 of a data stream
- * Copyright (C) 1995-2006 Mark Adler
+ * Copyright (C) 1995-2006, 2010 Mark Adler
  * For conditions of distribution and use, see copyright notice in zlib.h
  *
  * Thanks to Rodney Brown <rbrown64@csc.com.au> for his contribution of faster
diff --git a/deflate.c b/deflate.c
index efb26bf5d91..5c4022f3d47 100644
--- a/deflate.c
+++ b/deflate.c
@@ -52,7 +52,7 @@
 #include "deflate.h"
 
 const char deflate_copyright[] =
-   " deflate 1.2.4.3 Copyright 1995-2010 Jean-loup Gailly and Mark Adler ";
+   " deflate 1.2.5 Copyright 1995-2010 Jean-loup Gailly and Mark Adler ";
 /*
   If you use the zlib library in a product, an acknowledgment is welcome
   in the documentation of your product. If for some reason you cannot
diff --git a/deflate.h b/deflate.h
index f53deba852b..cbf0d1ea5d9 100644
--- a/deflate.h
+++ b/deflate.h
@@ -1,5 +1,5 @@
 /* deflate.h -- internal compression state
- * Copyright (C) 1995-2009 Jean-loup Gailly
+ * Copyright (C) 1995-2010 Jean-loup Gailly
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
@@ -290,13 +290,13 @@ typedef struct internal_state {
    memory checker errors from longest match routines */
 
         /* in trees.c */
-void _tr_init         OF((deflate_state *s));
-int  _tr_tally        OF((deflate_state *s, unsigned dist, unsigned lc));
-void _tr_flush_block  OF((deflate_state *s, charf *buf, ulg stored_len,
-                          int last));
-void _tr_align        OF((deflate_state *s));
-void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
-                          int last));
+void ZLIB_INTERNAL _tr_init OF((deflate_state *s));
+int ZLIB_INTERNAL _tr_tally OF((deflate_state *s, unsigned dist, unsigned lc));
+void ZLIB_INTERNAL _tr_flush_block OF((deflate_state *s, charf *buf,
+                        ulg stored_len, int last));
+void ZLIB_INTERNAL _tr_align OF((deflate_state *s));
+void ZLIB_INTERNAL _tr_stored_block OF((deflate_state *s, charf *buf,
+                        ulg stored_len, int last));
 
 #define d_code(dist) \
    ((dist) < 256 ? _dist_code[dist] : _dist_code[256+((dist)>>7)])
@@ -309,11 +309,11 @@ void _tr_stored_block OF((deflate_state *s, charf *buf, ulg stored_len,
 /* Inline versions of _tr_tally for speed: */
 
 #if defined(GEN_TREES_H) || !defined(STDC)
-  extern uch _length_code[];
-  extern uch _dist_code[];
+  extern uch ZLIB_INTERNAL _length_code[];
+  extern uch ZLIB_INTERNAL _dist_code[];
 #else
-  extern const uch _length_code[];
-  extern const uch _dist_code[];
+  extern const uch ZLIB_INTERNAL _length_code[];
+  extern const uch ZLIB_INTERNAL _dist_code[];
 #endif
 
 # define _tr_tally_lit(s, c, flush) \
diff --git a/gzguts.h b/gzguts.h
index b0a4cbdc336..0f8fb79f87d 100644
--- a/gzguts.h
+++ b/gzguts.h
@@ -12,7 +12,11 @@
 #  endif
 #endif
 
-#define ZLIB_INTERNAL
+#if ((__GNUC__-0) * 10 + __GNUC_MINOR__-0 >= 33) && !defined(NO_VIZ)
+#  define ZLIB_INTERNAL __attribute__((visibility ("hidden")))
+#else
+#  define ZLIB_INTERNAL
+#endif
 
 #include <stdio.h>
 #include "zlib.h"
@@ -112,9 +116,9 @@ typedef struct {
 typedef gz_state FAR *gz_statep;
 
 /* shared functions */
-ZEXTERN void ZEXPORT gz_error OF((gz_statep, int, const char *));
+void ZLIB_INTERNAL gz_error OF((gz_statep, int, const char *));
 #if defined UNDER_CE
-ZEXTERN char ZEXPORT *gz_strwinerror OF((DWORD error));
+char ZLIB_INTERNAL *gz_strwinerror OF((DWORD error));
 #endif
 
 /* GT_OFF(x), where x is an unsigned value, is true if x > maximum z_off64_t
@@ -123,6 +127,6 @@ ZEXTERN char ZEXPORT *gz_strwinerror OF((DWORD error));
 #ifdef INT_MAX
 #  define GT_OFF(x) (sizeof(int) == sizeof(z_off64_t) && (x) > INT_MAX)
 #else
-ZEXTERN unsigned ZEXPORT gz_intmax OF((void));
+unsigned ZLIB_INTERNAL gz_intmax OF((void));
 #  define GT_OFF(x) (sizeof(int) == sizeof(z_off64_t) && (x) > gz_intmax())
 #endif
diff --git a/gzlib.c b/gzlib.c
index 18390026905..603e60ed544 100644
--- a/gzlib.c
+++ b/gzlib.c
@@ -26,7 +26,7 @@ local gzFile gz_open OF((const char *, int, const char *));
 
    The gz_strwinerror function does not change the current setting of
    GetLastError. */
-char ZEXPORT *gz_strwinerror (error)
+char ZLIB_INTERNAL *gz_strwinerror (error)
      DWORD error;
 {
     static char buf[1024];
@@ -482,7 +482,7 @@ void ZEXPORT gzclearerr(file)
    memory).  Simply save the error message as a static string.  If there is an
    allocation failure constructing the error message, then convert the error to
    out of memory. */
-void ZEXPORT gz_error(state, err, msg)
+void ZLIB_INTERNAL gz_error(state, err, msg)
     gz_statep state;
     int err;
     const char *msg;
@@ -522,7 +522,7 @@ void ZEXPORT gz_error(state, err, msg)
    available) -- we need to do this to cover cases where 2's complement not
    used, since C standard permits 1's complement and sign-bit representations,
    otherwise we could just use ((unsigned)-1) >> 1 */
-unsigned ZEXPORT gz_intmax()
+unsigned ZLIB_INTERNAL gz_intmax()
 {
     unsigned p, q;
 
diff --git a/inffast.c b/inffast.c
index 0a0761f3de0..2f1d60b43b8 100644
--- a/inffast.c
+++ b/inffast.c
@@ -1,5 +1,5 @@
 /* inffast.c -- fast decoding
- * Copyright (C) 1995-2008 Mark Adler
+ * Copyright (C) 1995-2008, 2010 Mark Adler
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
@@ -64,7 +64,7 @@
       requires strm->avail_out >= 258 for each loop to avoid checking for
       output space.
  */
-void inflate_fast(strm, start)
+void ZLIB_INTERNAL inflate_fast(strm, start)
 z_streamp strm;
 unsigned start;         /* inflate()'s starting value for strm->avail_out */
 {
diff --git a/inffast.h b/inffast.h
index 1e88d2d97b5..e5c1aa4ca8c 100644
--- a/inffast.h
+++ b/inffast.h
@@ -1,5 +1,5 @@
 /* inffast.h -- header to use inffast.c
- * Copyright (C) 1995-2003 Mark Adler
+ * Copyright (C) 1995-2003, 2010 Mark Adler
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
@@ -8,4 +8,4 @@
    subject to change. Applications should only use zlib.h.
  */
 
-void inflate_fast OF((z_streamp strm, unsigned start));
+void ZLIB_INTERNAL inflate_fast OF((z_streamp strm, unsigned start));
diff --git a/inftrees.c b/inftrees.c
index 430b1741585..11e9c52accb 100644
--- a/inftrees.c
+++ b/inftrees.c
@@ -9,7 +9,7 @@
 #define MAXBITS 15
 
 const char inflate_copyright[] =
-   " inflate 1.2.4.3 Copyright 1995-2010 Mark Adler ";
+   " inflate 1.2.5 Copyright 1995-2010 Mark Adler ";
 /*
   If you use the zlib library in a product, an acknowledgment is welcome
   in the documentation of your product. If for some reason you cannot
@@ -29,7 +29,7 @@ const char inflate_copyright[] =
    table index bits.  It will differ if the request is greater than the
    longest code or if it is less than the shortest code.
  */
-int inflate_table(type, lens, codes, table, bits, work)
+int ZLIB_INTERNAL inflate_table(type, lens, codes, table, bits, work)
 codetype type;
 unsigned short FAR *lens;
 unsigned codes;
@@ -62,7 +62,7 @@ unsigned short FAR *work;
         35, 43, 51, 59, 67, 83, 99, 115, 131, 163, 195, 227, 258, 0, 0};
     static const unsigned short lext[31] = { /* Length codes 257..285 extra */
         16, 16, 16, 16, 16, 16, 16, 16, 17, 17, 17, 17, 18, 18, 18, 18,
-        19, 19, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 16, 195, 66};
+        19, 19, 19, 19, 20, 20, 20, 20, 21, 21, 21, 21, 16, 73, 195};
     static const unsigned short dbase[32] = { /* Distance codes 0..29 base */
         1, 2, 3, 4, 5, 7, 9, 13, 17, 25, 33, 49, 65, 97, 129, 193,
         257, 385, 513, 769, 1025, 1537, 2049, 3073, 4097, 6145,
diff --git a/inftrees.h b/inftrees.h
index 67461da0348..baa53a0b1a1 100644
--- a/inftrees.h
+++ b/inftrees.h
@@ -1,5 +1,5 @@
 /* inftrees.h -- header to use inftrees.c
- * Copyright (C) 1995-2005 Mark Adler
+ * Copyright (C) 1995-2005, 2010 Mark Adler
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
@@ -57,6 +57,6 @@ typedef enum {
     DISTS
 } codetype;
 
-extern int inflate_table OF((codetype type, unsigned short FAR *lens,
+int ZLIB_INTERNAL inflate_table OF((codetype type, unsigned short FAR *lens,
                              unsigned codes, code FAR * FAR *table,
                              unsigned FAR *bits, unsigned short FAR *work));
diff --git a/trees.c b/trees.c
index 1a6e997ac0a..56e9bb1c115 100644
--- a/trees.c
+++ b/trees.c
@@ -1,5 +1,5 @@
 /* trees.c -- output deflated data using Huffman coding
- * Copyright (C) 1995-2009 Jean-loup Gailly
+ * Copyright (C) 1995-2010 Jean-loup Gailly
  * detect_data_type() function provided freely by Cosmin Truta, 2006
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
@@ -351,13 +351,14 @@ void gen_trees_header()
                 static_dtree[i].Len, SEPARATOR(i, D_CODES-1, 5));
     }
 
-    fprintf(header, "const uch _dist_code[DIST_CODE_LEN] = {\n");
+    fprintf(header, "const uch ZLIB_INTERNAL _dist_code[DIST_CODE_LEN] = {\n");
     for (i = 0; i < DIST_CODE_LEN; i++) {
         fprintf(header, "%2u%s", _dist_code[i],
                 SEPARATOR(i, DIST_CODE_LEN-1, 20));
     }
 
-    fprintf(header, "const uch _length_code[MAX_MATCH-MIN_MATCH+1]= {\n");
+    fprintf(header,
+        "const uch ZLIB_INTERNAL _length_code[MAX_MATCH-MIN_MATCH+1]= {\n");
     for (i = 0; i < MAX_MATCH-MIN_MATCH+1; i++) {
         fprintf(header, "%2u%s", _length_code[i],
                 SEPARATOR(i, MAX_MATCH-MIN_MATCH, 20));
@@ -382,7 +383,7 @@ void gen_trees_header()
 /* ===========================================================================
  * Initialize the tree data structures for a new zlib stream.
  */
-void _tr_init(s)
+void ZLIB_INTERNAL _tr_init(s)
     deflate_state *s;
 {
     tr_static_init();
@@ -867,7 +868,7 @@ local void send_all_trees(s, lcodes, dcodes, blcodes)
 /* ===========================================================================
  * Send a stored block
  */
-void _tr_stored_block(s, buf, stored_len, last)
+void ZLIB_INTERNAL _tr_stored_block(s, buf, stored_len, last)
     deflate_state *s;
     charf *buf;       /* input block */
     ulg stored_len;   /* length of input block */
@@ -892,7 +893,7 @@ void _tr_stored_block(s, buf, stored_len, last)
  * To simplify the code, we assume the worst case of last real code encoded
  * on one bit only.
  */
-void _tr_align(s)
+void ZLIB_INTERNAL _tr_align(s)
     deflate_state *s;
 {
     send_bits(s, STATIC_TREES<<1, 3);
@@ -921,7 +922,7 @@ void _tr_align(s)
  * Determine the best encoding for the current block: dynamic trees, static
  * trees or store, and output the encoded block to the zip file.
  */
-void _tr_flush_block(s, buf, stored_len, last)
+void ZLIB_INTERNAL _tr_flush_block(s, buf, stored_len, last)
     deflate_state *s;
     charf *buf;       /* input block, or NULL if too old */
     ulg stored_len;   /* length of input block */
@@ -1022,7 +1023,7 @@ void _tr_flush_block(s, buf, stored_len, last)
  * Save the match info and tally the frequency counts. Return true if
  * the current block must be flushed.
  */
-int _tr_tally (s, dist, lc)
+int ZLIB_INTERNAL _tr_tally (s, dist, lc)
     deflate_state *s;
     unsigned dist;  /* distance of matched string */
     unsigned lc;    /* match length-MIN_MATCH or unmatched char (if dist==0) */
diff --git a/trees.h b/trees.h
index 72facf900f7..d35639d82a2 100644
--- a/trees.h
+++ b/trees.h
@@ -70,7 +70,7 @@ local const ct_data static_dtree[D_CODES] = {
 {{19},{ 5}}, {{11},{ 5}}, {{27},{ 5}}, {{ 7},{ 5}}, {{23},{ 5}}
 };
 
-const uch _dist_code[DIST_CODE_LEN] = {
+const uch ZLIB_INTERNAL _dist_code[DIST_CODE_LEN] = {
  0,  1,  2,  3,  4,  4,  5,  5,  6,  6,  6,  6,  7,  7,  7,  7,  8,  8,  8,  8,
  8,  8,  8,  8,  9,  9,  9,  9,  9,  9,  9,  9, 10, 10, 10, 10, 10, 10, 10, 10,
 10, 10, 10, 10, 10, 10, 10, 10, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11, 11,
@@ -99,7 +99,7 @@ const uch _dist_code[DIST_CODE_LEN] = {
 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29, 29
 };
 
-const uch _length_code[MAX_MATCH-MIN_MATCH+1]= {
+const uch ZLIB_INTERNAL _length_code[MAX_MATCH-MIN_MATCH+1]= {
  0,  1,  2,  3,  4,  5,  6,  7,  8,  8,  9,  9, 10, 10, 11, 11, 12, 12, 12, 12,
 13, 13, 13, 13, 14, 14, 14, 14, 15, 15, 15, 15, 16, 16, 16, 16, 16, 16, 16, 16,
 17, 17, 17, 17, 17, 17, 17, 17, 18, 18, 18, 18, 18, 18, 18, 18, 19, 19, 19, 19,
diff --git a/zconf.h b/zconf.h
index 1988920109f..02ce56c4313 100644
--- a/zconf.h
+++ b/zconf.h
@@ -315,10 +315,6 @@
 #  endif
 #endif
 
-#ifdef HAVE_VISIBILITY_PRAGMA
-#  define ZEXTERN __attribute__((visibility ("default"))) extern
-#endif
-
 #ifndef ZEXTERN
 #  define ZEXTERN extern
 #endif
@@ -368,6 +364,16 @@ typedef uLong FAR uLongf;
 #  include <sys/types.h>    /* for off_t */
 #endif
 
+/* a little trick to accommodate both "#define _LARGEFILE64_SOURCE" and
+ * "#define _LARGEFILE64_SOURCE 1" as requesting 64-bit operations, (even
+ * though the former does not conform to the LFS document), but considering
+ * both "#undef _LARGEFILE64_SOURCE" and "#define _LARGEFILE64_SOURCE 0" as
+ * equivalently requesting no 64-bit operations
+ */
+#if -_LARGEFILE64_SOURCE - -1 == 1
+#  undef _LARGEFILE64_SOURCE
+#endif
+
 #if defined(Z_HAVE_UNISTD_H) || defined(_LARGEFILE64_SOURCE)
 #  include <unistd.h>       /* for SEEK_* and off_t */
 #  ifdef VMS
diff --git a/zlib.3 b/zlib.3
index b5346168d0e..27adc4cd10a 100644
--- a/zlib.3
+++ b/zlib.3
@@ -1,4 +1,4 @@
-.TH ZLIB 3 "10 Apr 2010"
+.TH ZLIB 3 "19 Apr 2010"
 .SH NAME
 zlib \- compression/decompression library
 .SH SYNOPSIS
@@ -125,7 +125,7 @@ before asking for help.
 Send questions and/or comments to zlib@gzip.org,
 or (for the Windows DLL version) to Gilles Vollant (info@winimage.com).
 .SH AUTHORS
-Version 1.2.4.3
+Version 1.2.5
 Copyright (C) 1995-2010 Jean-loup Gailly (jloup@gzip.org)
 and Mark Adler (madler@alumni.caltech.edu).
 .LP
diff --git a/zlib.h b/zlib.h
index 699630c1ddc..bfbba83e8ee 100644
--- a/zlib.h
+++ b/zlib.h
@@ -1,5 +1,5 @@
 /* zlib.h -- interface of the 'zlib' general purpose compression library
-  version 1.2.4.3, April 10th, 2010
+  version 1.2.5, April 19th, 2010
 
   Copyright (C) 1995-2010 Jean-loup Gailly and Mark Adler
 
@@ -37,12 +37,12 @@
 extern "C" {
 #endif
 
-#define ZLIB_VERSION "1.2.4.3"
-#define ZLIB_VERNUM 0x1243
+#define ZLIB_VERSION "1.2.5"
+#define ZLIB_VERNUM 0x1250
 #define ZLIB_VER_MAJOR 1
 #define ZLIB_VER_MINOR 2
-#define ZLIB_VER_REVISION 4
-#define ZLIB_VER_SUBREVISION 3
+#define ZLIB_VER_REVISION 5
+#define ZLIB_VER_SUBREVISION 0
 
 /*
     The 'zlib' compression library provides in-memory compression and
diff --git a/zutil.c b/zutil.c
index 6ba16f68c23..898ed345b0e 100644
--- a/zutil.c
+++ b/zutil.c
@@ -1,5 +1,5 @@
 /* zutil.c -- target dependent utility functions for the compression library
- * Copyright (C) 1995-2005 Jean-loup Gailly.
+ * Copyright (C) 1995-2005, 2010 Jean-loup Gailly.
  * For conditions of distribution and use, see copyright notice in zlib.h
  */
 
@@ -117,9 +117,9 @@ uLong ZEXPORT zlibCompileFlags()
 #  ifndef verbose
 #    define verbose 0
 #  endif
-int z_verbose = verbose;
+int ZLIB_INTERNAL z_verbose = verbose;
 
-void z_error (m)
+void ZLIB_INTERNAL z_error (m)
     char *m;
 {
     fprintf(stderr, "%s\n", m);
@@ -146,7 +146,7 @@ const char * ZEXPORT zError(err)
 
 #ifndef HAVE_MEMCPY
 
-void zmemcpy(dest, source, len)
+void ZLIB_INTERNAL zmemcpy(dest, source, len)
     Bytef* dest;
     const Bytef* source;
     uInt  len;
@@ -157,7 +157,7 @@ void zmemcpy(dest, source, len)
     } while (--len != 0);
 }
 
-int zmemcmp(s1, s2, len)
+int ZLIB_INTERNAL zmemcmp(s1, s2, len)
     const Bytef* s1;
     const Bytef* s2;
     uInt  len;
@@ -170,7 +170,7 @@ int zmemcmp(s1, s2, len)
     return 0;
 }
 
-void zmemzero(dest, len)
+void ZLIB_INTERNAL zmemzero(dest, len)
     Bytef* dest;
     uInt  len;
 {
@@ -213,7 +213,7 @@ local ptr_table table[MAX_PTR];
  * a protected system like OS/2. Use Microsoft C instead.
  */
 
-voidpf zcalloc (voidpf opaque, unsigned items, unsigned size)
+voidpf ZLIB_INTERNAL zcalloc (voidpf opaque, unsigned items, unsigned size)
 {
     voidpf buf = opaque; /* just to make some compilers happy */
     ulg bsize = (ulg)items*size;
@@ -237,7 +237,7 @@ voidpf zcalloc (voidpf opaque, unsigned items, unsigned size)
     return buf;
 }
 
-void  zcfree (voidpf opaque, voidpf ptr)
+void ZLIB_INTERNAL zcfree (voidpf opaque, voidpf ptr)
 {
     int n;
     if (*(ush*)&ptr != 0) { /* object < 64K */
@@ -272,13 +272,13 @@ void  zcfree (voidpf opaque, voidpf ptr)
 #  define _hfree   hfree
 #endif
 
-voidpf zcalloc (voidpf opaque, uInt items, uInt size)
+voidpf ZLIB_INTERNAL zcalloc (voidpf opaque, uInt items, uInt size)
 {
     if (opaque) opaque = 0; /* to make compiler happy */
     return _halloc((long)items, size);
 }
 
-void  zcfree (voidpf opaque, voidpf ptr)
+void ZLIB_INTERNAL zcfree (voidpf opaque, voidpf ptr)
 {
     if (opaque) opaque = 0; /* to make compiler happy */
     _hfree(ptr);
@@ -297,7 +297,7 @@ extern voidp  calloc OF((uInt items, uInt size));
 extern void   free   OF((voidpf ptr));
 #endif
 
-voidpf zcalloc (opaque, items, size)
+voidpf ZLIB_INTERNAL zcalloc (opaque, items, size)
     voidpf opaque;
     unsigned items;
     unsigned size;
@@ -307,7 +307,7 @@ voidpf zcalloc (opaque, items, size)
                               (voidpf)calloc(items, size);
 }
 
-void  zcfree (opaque, ptr)
+void ZLIB_INTERNAL zcfree (opaque, ptr)
     voidpf opaque;
     voidpf ptr;
 {
diff --git a/zutil.h b/zutil.h
index a250088640a..258fa88799a 100644
--- a/zutil.h
+++ b/zutil.h
@@ -13,7 +13,12 @@
 #ifndef ZUTIL_H
 #define ZUTIL_H
 
-#define ZLIB_INTERNAL
+#if ((__GNUC__-0) * 10 + __GNUC_MINOR__-0 >= 33) && !defined(NO_VIZ)
+#  define ZLIB_INTERNAL __attribute__((visibility ("hidden")))
+#else
+#  define ZLIB_INTERNAL
+#endif
+
 #include "zlib.h"
 
 #ifdef STDC
@@ -231,16 +236,16 @@ extern const char * const z_errmsg[10]; /* indexed by 2-zlib_error */
 #    define zmemzero(dest, len) memset(dest, 0, len)
 #  endif
 #else
-   extern void zmemcpy  OF((Bytef* dest, const Bytef* source, uInt len));
-   extern int  zmemcmp  OF((const Bytef* s1, const Bytef* s2, uInt len));
-   extern void zmemzero OF((Bytef* dest, uInt len));
+   void ZLIB_INTERNAL zmemcpy OF((Bytef* dest, const Bytef* source, uInt len));
+   int ZLIB_INTERNAL zmemcmp OF((const Bytef* s1, const Bytef* s2, uInt len));
+   void ZLIB_INTERNAL zmemzero OF((Bytef* dest, uInt len));
 #endif
 
 /* Diagnostic functions */
 #ifdef DEBUG
 #  include <stdio.h>
-   extern int z_verbose;
-   extern void z_error    OF((char *m));
+   extern int ZLIB_INTERNAL z_verbose;
+   extern void ZLIB_INTERNAL z_error OF((char *m));
 #  define Assert(cond,msg) {if(!(cond)) z_error(msg);}
 #  define Trace(x) {if (z_verbose>=0) fprintf x ;}
 #  define Tracev(x) {if (z_verbose>0) fprintf x ;}
@@ -257,8 +262,9 @@ extern const char * const z_errmsg[10]; /* indexed by 2-zlib_error */
 #endif
 
 
-voidpf zcalloc OF((voidpf opaque, unsigned items, unsigned size));
-void   zcfree  OF((voidpf opaque, voidpf ptr));
+voidpf ZLIB_INTERNAL zcalloc OF((voidpf opaque, unsigned items,
+                        unsigned size));
+void ZLIB_INTERNAL zcfree  OF((voidpf opaque, voidpf ptr));
 
 #define ZALLOC(strm, items, size) \
            (*((strm)->zalloc))((strm)->opaque, (items), (size))

From 62592d917310c364b915d2c4be86bc1e0da1c1a1 Mon Sep 17 00:00:00 2001
From: Rebecca Cran <brucec@FreeBSD.org>
Date: Tue, 20 Apr 2010 19:30:12 +0000
Subject: [PATCH 167/532] It's not necessary to reset the chip every time an
 input overflow event occurs. In addition, the delay when programming the
 short cable fix should be 100us, not 100ms.

PR:	kern/64556
Submitted by:	Thomas Hurst <tom at hur.st>
Approved by:	rrs (mentor)
MFC after:	1 week
---
 sys/dev/sis/if_sis.c | 15 +++------------
 1 file changed, 3 insertions(+), 12 deletions(-)

diff --git a/sys/dev/sis/if_sis.c b/sys/dev/sis/if_sis.c
index aac46f7bda0..19833423c82 100644
--- a/sys/dev/sis/if_sis.c
+++ b/sys/dev/sis/if_sis.c
@@ -1483,15 +1483,6 @@ sis_rxeof(struct sis_softc *sc)
 	return (rx_npkts);
 }
 
-static void
-sis_rxeoc(struct sis_softc *sc)
-{
-
-	SIS_LOCK_ASSERT(sc);
-	sis_rxeof(sc);
-	sis_initl(sc);
-}
-
 /*
  * A frame was downloaded to the chip. It's safe for us to clean up
  * the list buffers.
@@ -1614,7 +1605,7 @@ sis_poll(struct ifnet *ifp, enum poll_cmd cmd, int count)
 		status = CSR_READ_4(sc, SIS_ISR);
 
 		if (status & (SIS_ISR_RX_ERR|SIS_ISR_RX_OFLOW))
-			sis_rxeoc(sc);
+			ifp->if_ierrors++;
 
 		if (status & (SIS_ISR_RX_IDLE))
 			SIS_SETBIT(sc, SIS_CSR, SIS_CSR_RX_ENABLE);
@@ -1672,7 +1663,7 @@ sis_intr(void *arg)
 			sis_rxeof(sc);
 
 		if (status & SIS_ISR_RX_OFLOW)
-			sis_rxeoc(sc);
+			ifp->if_ierrors++;
 
 		if (status & (SIS_ISR_RX_IDLE))
 			SIS_SETBIT(sc, SIS_CSR, SIS_CSR_RX_ENABLE);
@@ -2017,7 +2008,7 @@ sis_initl(struct sis_softc *sc)
 		CSR_WRITE_4(sc, NS_PHY_PAGE, 0x0001);
 		reg = CSR_READ_4(sc, NS_PHY_DSPCFG) & 0xfff;
 		CSR_WRITE_4(sc, NS_PHY_DSPCFG, reg | 0x1000);
-		DELAY(100000);
+		DELAY(100);
 		reg = CSR_READ_4(sc, NS_PHY_TDATA) & 0xff;
 		if ((reg & 0x0080) == 0 || (reg > 0xd8 && reg <= 0xff)) {
 			device_printf(sc->sis_dev,

From bd34cda61a6e3ac2aaa9a18f4853498a41e65018 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 20 Apr 2010 20:36:38 +0000
Subject: [PATCH 168/532] Bump minor version of config to reflect the new
 option remapping feature.  The kernel makefiles have specifically not been
 bumped because nothing uses this new feature and doing so forces everybody to
 recompile for no good reason.  This chnage will be MFC'd where the kernel
 version numbers for amd64 and ia64 will be bumped, since those are the only
 two that have use the option remapping feature.  Once merged, this will give
 a better error message to folks that are using buildkernel without buildworld
 or kernel-toolchain to update their kernels.

MFC after:	3 days
---
 usr.sbin/config/configvers.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.sbin/config/configvers.h b/usr.sbin/config/configvers.h
index 28451585ec2..73310bd6f41 100644
--- a/usr.sbin/config/configvers.h
+++ b/usr.sbin/config/configvers.h
@@ -49,5 +49,5 @@
  *
  * $FreeBSD$
  */
-#define	CONFIGVERS	600007
+#define	CONFIGVERS	600008
 #define	MAJOR_VERS(x)	((x) / 100000)

From e8b93e425769587779f6ee962c61dc095ea8cdc0 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 20 Apr 2010 20:39:42 +0000
Subject: [PATCH 169/532] Make sure that we free the passed in data message if
 we don't actually insert it onto the queue.  Also, fix a mtx leak if someone
 turns off devctl while we're processing a messages.

MFC after:	5 days
---
 sys/kern/subr_bus.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/sys/kern/subr_bus.c b/sys/kern/subr_bus.c
index 6e939c0ed77..9d3292aac45 100644
--- a/sys/kern/subr_bus.c
+++ b/sys/kern/subr_bus.c
@@ -545,15 +545,16 @@ devctl_queue_data(char *data)
 	struct proc *p;
 
 	if (strlen(data) == 0)
-		return;
+		goto out;
 	if (devctl_queue_length == 0)
-		return;
+		goto out;
 	n1 = malloc(sizeof(*n1), M_BUS, M_NOWAIT);
 	if (n1 == NULL)
-		return;
+		goto out;
 	n1->dei_data = data;
 	mtx_lock(&devsoftc.mtx);
 	if (devctl_queue_length == 0) {
+		mtx_unlock(&devsoftc.mtx);
 		free(n1->dei_data, M_BUS);
 		free(n1, M_BUS);
 		return;
@@ -577,6 +578,14 @@ devctl_queue_data(char *data)
 		psignal(p, SIGIO);
 		PROC_UNLOCK(p);
 	}
+	return;
+out:
+	/*
+	 * We have to free data on all error paths since the caller
+	 * assumes it will be free'd when this item is dequeued.
+	 */
+	free(data, M_BUS);
+	return;
 }
 
 /**

From 2669f95457b9136a7c9fa78e879cc53f68799078 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Tue, 20 Apr 2010 20:47:58 +0000
Subject: [PATCH 170/532] Initial import of TestFloat 2a.

Obtained from:	http://www.jhauser.us/arithmetic/TestFloat.html
---
 processors/386-gcc.h                    |   80 +
 processors/SPARC-gcc.h                  |   80 +
 testfloat/386-Win32-gcc/Makefile        |   64 +
 testfloat/386-Win32-gcc/milieu.h        |   51 +
 testfloat/386-Win32-gcc/systflags.S     |   41 +
 testfloat/386-Win32-gcc/systfloat.S     |  332 ++
 testfloat/386-Win32-gcc/systfloat.h     |  181 ++
 testfloat/386-Win32-gcc/systmodes.S     |   82 +
 testfloat/README.txt                    |   50 +
 testfloat/SPARC-Solaris-gcc/Makefile    |   64 +
 testfloat/SPARC-Solaris-gcc/milieu.h    |   51 +
 testfloat/SPARC-Solaris-gcc/systflags.c |   42 +
 testfloat/SPARC-Solaris-gcc/systfloat.S |  941 ++++++
 testfloat/SPARC-Solaris-gcc/systfloat.h |  205 ++
 testfloat/SPARC-Solaris-gcc/systmodes.c |   51 +
 testfloat/fail.c                        |   46 +
 testfloat/fail.h                        |   29 +
 testfloat/random.c                      |   65 +
 testfloat/random.h                      |   32 +
 testfloat/slowfloat-32.c                | 1183 ++++++++
 testfloat/slowfloat-64.c                | 2109 +++++++++++++
 testfloat/slowfloat.c                   |   35 +
 testfloat/slowfloat.h                   |  167 +
 testfloat/systemBugs.txt                |  323 ++
 testfloat/systflags.h                   |   33 +
 testfloat/systfloat.c                   |  553 ++++
 testfloat/systfloat.h                   |  233 ++
 testfloat/systmodes.h                   |   42 +
 testfloat/templates/Makefile            |   67 +
 testfloat/templates/milieu.h            |   62 +
 testfloat/templates/systflags.c         |   41 +
 testfloat/templates/systmodes.c         |   58 +
 testfloat/testCases.c                   | 3679 +++++++++++++++++++++++
 testfloat/testCases.h                   |   69 +
 testfloat/testFunction.c                | 1149 +++++++
 testfloat/testFunction.h                |  135 +
 testfloat/testLoops.c                   | 2713 +++++++++++++++++
 testfloat/testLoops.h                   |  143 +
 testfloat/testfloat-history.txt         |   57 +
 testfloat/testfloat-source.txt          |  444 +++
 testfloat/testfloat.c                   |  295 ++
 testfloat/testfloat.txt                 |  771 +++++
 testfloat/testsoftfloat.c               | 1040 +++++++
 testfloat/writeHex.c                    |  183 ++
 testfloat/writeHex.h                    |   42 +
 45 files changed, 18113 insertions(+)
 create mode 100644 processors/386-gcc.h
 create mode 100644 processors/SPARC-gcc.h
 create mode 100644 testfloat/386-Win32-gcc/Makefile
 create mode 100644 testfloat/386-Win32-gcc/milieu.h
 create mode 100644 testfloat/386-Win32-gcc/systflags.S
 create mode 100644 testfloat/386-Win32-gcc/systfloat.S
 create mode 100644 testfloat/386-Win32-gcc/systfloat.h
 create mode 100644 testfloat/386-Win32-gcc/systmodes.S
 create mode 100644 testfloat/README.txt
 create mode 100644 testfloat/SPARC-Solaris-gcc/Makefile
 create mode 100644 testfloat/SPARC-Solaris-gcc/milieu.h
 create mode 100644 testfloat/SPARC-Solaris-gcc/systflags.c
 create mode 100644 testfloat/SPARC-Solaris-gcc/systfloat.S
 create mode 100644 testfloat/SPARC-Solaris-gcc/systfloat.h
 create mode 100644 testfloat/SPARC-Solaris-gcc/systmodes.c
 create mode 100644 testfloat/fail.c
 create mode 100644 testfloat/fail.h
 create mode 100644 testfloat/random.c
 create mode 100644 testfloat/random.h
 create mode 100644 testfloat/slowfloat-32.c
 create mode 100644 testfloat/slowfloat-64.c
 create mode 100644 testfloat/slowfloat.c
 create mode 100644 testfloat/slowfloat.h
 create mode 100644 testfloat/systemBugs.txt
 create mode 100644 testfloat/systflags.h
 create mode 100644 testfloat/systfloat.c
 create mode 100644 testfloat/systfloat.h
 create mode 100644 testfloat/systmodes.h
 create mode 100644 testfloat/templates/Makefile
 create mode 100644 testfloat/templates/milieu.h
 create mode 100644 testfloat/templates/systflags.c
 create mode 100644 testfloat/templates/systmodes.c
 create mode 100644 testfloat/testCases.c
 create mode 100644 testfloat/testCases.h
 create mode 100644 testfloat/testFunction.c
 create mode 100644 testfloat/testFunction.h
 create mode 100644 testfloat/testLoops.c
 create mode 100644 testfloat/testLoops.h
 create mode 100644 testfloat/testfloat-history.txt
 create mode 100644 testfloat/testfloat-source.txt
 create mode 100644 testfloat/testfloat.c
 create mode 100644 testfloat/testfloat.txt
 create mode 100644 testfloat/testsoftfloat.c
 create mode 100644 testfloat/writeHex.c
 create mode 100644 testfloat/writeHex.h

diff --git a/processors/386-gcc.h b/processors/386-gcc.h
new file mode 100644
index 00000000000..1fe8bb9ca03
--- /dev/null
+++ b/processors/386-gcc.h
@@ -0,0 +1,80 @@
+
+/*
+-------------------------------------------------------------------------------
+One of the macros `BIGENDIAN' or `LITTLEENDIAN' must be defined.
+-------------------------------------------------------------------------------
+*/
+#define LITTLEENDIAN
+
+/*
+-------------------------------------------------------------------------------
+The macro `BITS64' can be defined to indicate that 64-bit integer types are
+supported by the compiler.
+-------------------------------------------------------------------------------
+*/
+#define BITS64
+
+/*
+-------------------------------------------------------------------------------
+Each of the following `typedef's defines the most convenient type that holds
+integers of at least as many bits as specified.  For example, `uint8' should
+be the most convenient type that can hold unsigned integers of as many as
+8 bits.  The `flag' type must be able to hold either a 0 or 1.  For most
+implementations of C, `flag', `uint8', and `int8' should all be `typedef'ed
+to the same as `int'.
+-------------------------------------------------------------------------------
+*/
+typedef char flag;
+typedef unsigned char uint8;
+typedef signed char int8;
+typedef int uint16;
+typedef int int16;
+typedef unsigned int uint32;
+typedef signed int int32;
+#ifdef BITS64
+typedef unsigned long long int uint64;
+typedef signed long long int int64;
+#endif
+
+/*
+-------------------------------------------------------------------------------
+Each of the following `typedef's defines a type that holds integers
+of _exactly_ the number of bits specified.  For instance, for most
+implementation of C, `bits16' and `sbits16' should be `typedef'ed to
+`unsigned short int' and `signed short int' (or `short int'), respectively.
+-------------------------------------------------------------------------------
+*/
+typedef unsigned char bits8;
+typedef signed char sbits8;
+typedef unsigned short int bits16;
+typedef signed short int sbits16;
+typedef unsigned int bits32;
+typedef signed int sbits32;
+#ifdef BITS64
+typedef unsigned long long int bits64;
+typedef signed long long int sbits64;
+#endif
+
+#ifdef BITS64
+/*
+-------------------------------------------------------------------------------
+The `LIT64' macro takes as its argument a textual integer literal and
+if necessary ``marks'' the literal as having a 64-bit integer type.
+For example, the GNU C Compiler (`gcc') requires that 64-bit literals be
+appended with the letters `LL' standing for `long long', which is `gcc's
+name for the 64-bit integer type.  Some compilers may allow `LIT64' to be
+defined as the identity macro:  `#define LIT64( a ) a'.
+-------------------------------------------------------------------------------
+*/
+#define LIT64( a ) a##LL
+#endif
+
+/*
+-------------------------------------------------------------------------------
+The macro `INLINE' can be used before functions that should be inlined.  If
+a compiler does not support explicit inlining, this macro should be defined
+to be `static'.
+-------------------------------------------------------------------------------
+*/
+#define INLINE extern inline
+
diff --git a/processors/SPARC-gcc.h b/processors/SPARC-gcc.h
new file mode 100644
index 00000000000..f926615f99e
--- /dev/null
+++ b/processors/SPARC-gcc.h
@@ -0,0 +1,80 @@
+
+/*
+-------------------------------------------------------------------------------
+One of the macros `BIGENDIAN' or `LITTLEENDIAN' must be defined.
+-------------------------------------------------------------------------------
+*/
+#define BIGENDIAN
+
+/*
+-------------------------------------------------------------------------------
+The macro `BITS64' can be defined to indicate that 64-bit integer types are
+supported by the compiler.
+-------------------------------------------------------------------------------
+*/
+#define BITS64
+
+/*
+-------------------------------------------------------------------------------
+Each of the following `typedef's defines the most convenient type that holds
+integers of at least as many bits as specified.  For example, `uint8' should
+be the most convenient type that can hold unsigned integers of as many as
+8 bits.  The `flag' type must be able to hold either a 0 or 1.  For most
+implementations of C, `flag', `uint8', and `int8' should all be `typedef'ed
+to the same as `int'.
+-------------------------------------------------------------------------------
+*/
+typedef int flag;
+typedef int uint8;
+typedef int int8;
+typedef int uint16;
+typedef int int16;
+typedef unsigned int uint32;
+typedef signed int int32;
+#ifdef BITS64
+typedef unsigned long long int uint64;
+typedef signed long long int int64;
+#endif
+
+/*
+-------------------------------------------------------------------------------
+Each of the following `typedef's defines a type that holds integers
+of _exactly_ the number of bits specified.  For instance, for most
+implementation of C, `bits16' and `sbits16' should be `typedef'ed to
+`unsigned short int' and `signed short int' (or `short int'), respectively.
+-------------------------------------------------------------------------------
+*/
+typedef unsigned char bits8;
+typedef signed char sbits8;
+typedef unsigned short int bits16;
+typedef signed short int sbits16;
+typedef unsigned int bits32;
+typedef signed int sbits32;
+#ifdef BITS64
+typedef unsigned long long int bits64;
+typedef signed long long int sbits64;
+#endif
+
+#ifdef BITS64
+/*
+-------------------------------------------------------------------------------
+The `LIT64' macro takes as its argument a textual integer literal and
+if necessary ``marks'' the literal as having a 64-bit integer type.
+For example, the GNU C Compiler (`gcc') requires that 64-bit literals be
+appended with the letters `LL' standing for `long long', which is `gcc's
+name for the 64-bit integer type.  Some compilers may allow `LIT64' to be
+defined as the identity macro:  `#define LIT64( a ) a'.
+-------------------------------------------------------------------------------
+*/
+#define LIT64( a ) a##LL
+#endif
+
+/*
+-------------------------------------------------------------------------------
+The macro `INLINE' can be used before functions that should be inlined.  If
+a compiler does not support explicit inlining, this macro should be defined
+to be `static'.
+-------------------------------------------------------------------------------
+*/
+#define INLINE extern inline
+
diff --git a/testfloat/386-Win32-gcc/Makefile b/testfloat/386-Win32-gcc/Makefile
new file mode 100644
index 00000000000..bba8b9e5b1e
--- /dev/null
+++ b/testfloat/386-Win32-gcc/Makefile
@@ -0,0 +1,64 @@
+
+PROCESSOR_H = ../../processors/386-gcc.h
+SOFTFLOAT_VERSION = bits64
+TARGET = 386-Win32-gcc
+SOFTFLOAT_DIR = ../../softfloat/$(SOFTFLOAT_VERSION)/$(TARGET)
+
+OBJ = .o
+EXE = .exe
+INCLUDES = -I. -I.. -I$(SOFTFLOAT_DIR)
+COMPILE_ASM = gcc -c -o $@
+COMPILE_C = gcc -c -o $@ $(INCLUDES) -I- -O2
+COMPILE_SLOWFLOAT_C = gcc -c -o $@ $(INCLUDES) -I- -O3
+LINK = gcc -o $@
+
+SOFTFLOAT_H = $(SOFTFLOAT_DIR)/softfloat.h
+SOFTFLOAT_OBJ = $(SOFTFLOAT_DIR)/softfloat$(OBJ)
+
+ALL: testsoftfloat$(EXE) testfloat$(EXE)
+
+milieu.h: $(PROCESSOR_H)
+	touch milieu.h
+
+fail$(OBJ): milieu.h ../fail.h
+	$(COMPILE_C) ../fail.c
+
+random$(OBJ): milieu.h ../random.h
+	$(COMPILE_C) ../random.c
+
+testCases$(OBJ): milieu.h ../fail.h ../random.h $(SOFTFLOAT_H) ../testCases.h ../testCases.c
+	$(COMPILE_C) ../testCases.c
+
+writeHex$(OBJ): milieu.h $(SOFTFLOAT_H) ../writeHex.h ../writeHex.c
+	$(COMPILE_C) ../writeHex.c
+
+testLoops$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../writeHex.h ../testLoops.h ../testLoops.c
+	$(COMPILE_C) ../testLoops.c
+
+slowfloat$(OBJ): milieu.h $(SOFTFLOAT_H) ../slowfloat.h ../slowfloat-32.c ../slowfloat-64.c ../slowfloat.c
+	$(COMPILE_SLOWFLOAT_C) ../slowfloat.c
+
+testsoftfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../slowfloat.h ../testsoftfloat.c
+	$(COMPILE_C) ../testsoftfloat.c
+
+testsoftfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+
+systmodes$(OBJ): systmodes.S
+	$(COMPILE_ASM) systmodes.S
+
+systflags$(OBJ): systflags.S
+	$(COMPILE_ASM) systflags.S
+
+systfloat$(OBJ): systfloat.S
+	$(COMPILE_ASM) systfloat.S
+
+testFunction$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systmodes.h ../systflags.h systfloat.h ../testFunction.h ../testFunction.c
+	$(COMPILE_C) ../testFunction.c
+
+testfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systflags.h ../testFunction.h ../testfloat.c
+	$(COMPILE_C) ../testfloat.c
+
+testfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+
diff --git a/testfloat/386-Win32-gcc/milieu.h b/testfloat/386-Win32-gcc/milieu.h
new file mode 100644
index 00000000000..7f795dd9396
--- /dev/null
+++ b/testfloat/386-Win32-gcc/milieu.h
@@ -0,0 +1,51 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+Include common integer types and flags.
+-------------------------------------------------------------------------------
+*/
+#include "../../processors/386-gcc.h"
+
+/*
+-------------------------------------------------------------------------------
+If the `BITS64' macro is defined by the processor header file but the
+version of SoftFloat being tested is the 32-bit one (`bits32'), the `BITS64'
+macro must be undefined here.
+-------------------------------------------------------------------------------
+#undef BITS64
+*/
+
+/*
+-------------------------------------------------------------------------------
+Symbolic Boolean literals.
+-------------------------------------------------------------------------------
+*/
+enum {
+    FALSE = 0,
+    TRUE  = 1
+};
+
diff --git a/testfloat/386-Win32-gcc/systflags.S b/testfloat/386-Win32-gcc/systflags.S
new file mode 100644
index 00000000000..b338e54be82
--- /dev/null
+++ b/testfloat/386-Win32-gcc/systflags.S
@@ -0,0 +1,41 @@
+
+/*
+===============================================================================
+
+This GNU assembler source file is part of TestFloat, Release 2a, a package
+of programs for testing the correctness of floating-point arithmetic
+complying to the IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+	.text
+
+/*
+-------------------------------------------------------------------------------
+Clears the system's IEC/IEEE floating-point exception flags.  Returns the
+previous value of the flags.
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_float_flags_clear
+_syst_float_flags_clear:
+	fnstsw %ax
+	fnclex
+	andl $61,%eax
+	ret
+
diff --git a/testfloat/386-Win32-gcc/systfloat.S b/testfloat/386-Win32-gcc/systfloat.S
new file mode 100644
index 00000000000..01e8983e494
--- /dev/null
+++ b/testfloat/386-Win32-gcc/systfloat.S
@@ -0,0 +1,332 @@
+
+/*
+===============================================================================
+
+This GNU assembler source file is part of TestFloat, Release 2a, a package
+of programs for testing the correctness of floating-point arithmetic
+complying to the IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+	.text
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_int32_to_floatx80
+_syst_int32_to_floatx80:
+	fildl 8(%esp)
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_int64_to_floatx80
+_syst_int64_to_floatx80:
+	fildq 8(%esp)
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_float32_to_floatx80
+_syst_float32_to_floatx80:
+	flds 8(%esp)
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_float64_to_floatx80
+_syst_float64_to_floatx80:
+	fldl 8(%esp)
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_to_int32
+_syst_floatx80_to_int32:
+	fldt 4(%esp)
+	subl $4,%esp
+	fistpl (%esp)
+	movl (%esp),%eax
+	addl $4,%esp
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_to_int64
+_syst_floatx80_to_int64:
+	fldt 4(%esp)
+	subl $8,%esp
+	fistpq (%esp)
+	movl (%esp),%eax
+	movl 4(%esp),%edx
+	addl $8,%esp
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_to_float32
+_syst_floatx80_to_float32:
+	fldt 4(%esp)
+	subl $4,%esp
+	fstps (%esp)
+	movl (%esp),%eax
+	addl $4,%esp
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_to_float64
+_syst_floatx80_to_float64:
+	fldt 4(%esp)
+	subl $8,%esp
+	fstpl (%esp)
+	movl 4(%esp),%edx
+	movl (%esp),%eax
+	addl $8,%esp
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_round_to_int
+_syst_floatx80_round_to_int:
+	fldt 8(%esp)
+	frndint
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_add
+_syst_floatx80_add:
+	fldt 8(%esp)
+	fldt 20(%esp)
+	faddp
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_sub
+_syst_floatx80_sub:
+	fldt 8(%esp)
+	fldt 20(%esp)
+	fsubrp
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_mul
+_syst_floatx80_mul:
+	fldt 8(%esp)
+	fldt 20(%esp)
+	fmulp
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_div
+_syst_floatx80_div:
+	fldt 8(%esp)
+	fldt 20(%esp)
+	fdivrp
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_rem
+_syst_floatx80_rem:
+	fldt 20(%esp)
+	fldt 8(%esp)
+floatx80_rem_loop:
+	fprem1
+	fnstsw %ax
+	btw $10,%ax
+	jc floatx80_rem_loop
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	fstp %st(0)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_sqrt
+_syst_floatx80_sqrt:
+	fldt 8(%esp)
+	fsqrt
+	movl 4(%esp),%eax
+	fstpt (%eax)
+	ret $4
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_eq
+_syst_floatx80_eq:
+	fldt 16(%esp)
+	fldt 4(%esp)
+	fucompp
+	fnstsw %ax
+	andw $17664,%ax
+	cmpw $16384,%ax
+	seteb %al
+	movzb %al,%eax
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_le
+_syst_floatx80_le:
+	fldt 4(%esp)
+	fldt 16(%esp)
+	fcompp
+	fnstsw %ax
+	notl %eax
+	shrl $8,%eax
+	andl $1,%eax
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_lt
+_syst_floatx80_lt:
+	fldt 4(%esp)
+	fldt 16(%esp)
+	fcompp
+	fnstsw %ax
+	andw $17664,%ax
+	setzb %al
+	movzb %al,%eax
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_eq_signaling
+_syst_floatx80_eq_signaling:
+	fldt 16(%esp)
+	fldt 4(%esp)
+	fcompp
+	fnstsw %ax
+	andw $17664,%ax
+	cmpw $16384,%ax
+	seteb %al
+	movzb %al,%eax
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_le_quiet
+_syst_floatx80_le_quiet:
+	fldt 4(%esp)
+	fldt 16(%esp)
+	fucompp
+	fnstsw %ax
+	notl %eax
+	shrl $8,%eax
+	andl $1,%eax
+	ret
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 2
+.globl _syst_floatx80_lt_quiet
+_syst_floatx80_lt_quiet:
+	fldt 4(%esp)
+	fldt 16(%esp)
+	fucompp
+	fnstsw %ax
+	andw $17664,%ax
+	setzb %al
+	movzb %al,%eax
+	ret
+
diff --git a/testfloat/386-Win32-gcc/systfloat.h b/testfloat/386-Win32-gcc/systfloat.h
new file mode 100644
index 00000000000..167b9160203
--- /dev/null
+++ b/testfloat/386-Win32-gcc/systfloat.h
@@ -0,0 +1,181 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+The following macros are defined to indicate that the corresponding
+functions exist.
+-------------------------------------------------------------------------------
+*/
+#define SYST_INT32_TO_FLOATX80
+#define SYST_INT64_TO_FLOATX80
+#define SYST_FLOAT32_TO_FLOATX80
+#define SYST_FLOAT64_TO_FLOATX80
+#define SYST_FLOATX80_TO_INT32
+#define SYST_FLOATX80_TO_INT64
+#define SYST_FLOATX80_TO_FLOAT32
+#define SYST_FLOATX80_TO_FLOAT64
+#define SYST_FLOATX80_ROUND_TO_INT
+#define SYST_FLOATX80_ADD
+#define SYST_FLOATX80_SUB
+#define SYST_FLOATX80_MUL
+#define SYST_FLOATX80_DIV
+#define SYST_FLOATX80_REM
+#define SYST_FLOATX80_SQRT
+#define SYST_FLOATX80_EQ
+#define SYST_FLOATX80_LE
+#define SYST_FLOATX80_LT
+#define SYST_FLOATX80_EQ_SIGNALING
+#define SYST_FLOATX80_LE_QUIET
+#define SYST_FLOATX80_LT_QUIET
+
+/*
+-------------------------------------------------------------------------------
+System function declarations.  (Some of these functions may not exist.)
+-------------------------------------------------------------------------------
+*/
+float32 syst_int32_to_float32( int32 );
+float64 syst_int32_to_float64( int32 );
+#ifdef FLOATX80
+floatx80 syst_int32_to_floatx80( int32 );
+#endif
+#ifdef FLOAT128
+float128 syst_int32_to_float128( int32 );
+#endif
+#ifdef BITS64
+float32 syst_int64_to_float32( int64 );
+float64 syst_int64_to_float64( int64 );
+#ifdef FLOATX80
+floatx80 syst_int64_to_floatx80( int64 );
+#endif
+#ifdef FLOAT128
+float128 syst_int64_to_float128( int64 );
+#endif
+#endif
+int32 syst_float32_to_int32( float32 );
+int32 syst_float32_to_int32_round_to_zero( float32 );
+#ifdef BITS64
+int64 syst_float32_to_int64( float32 );
+int64 syst_float32_to_int64_round_to_zero( float32 );
+#endif
+float64 syst_float32_to_float64( float32 );
+#ifdef FLOATX80
+floatx80 syst_float32_to_floatx80( float32 );
+#endif
+#ifdef FLOAT128
+float128 syst_float32_to_float128( float32 );
+#endif
+float32 syst_float32_round_to_int( float32 );
+float32 syst_float32_add( float32, float32 );
+float32 syst_float32_sub( float32, float32 );
+float32 syst_float32_mul( float32, float32 );
+float32 syst_float32_div( float32, float32 );
+float32 syst_float32_rem( float32, float32 );
+float32 syst_float32_sqrt( float32 );
+flag syst_float32_eq( float32, float32 );
+flag syst_float32_le( float32, float32 );
+flag syst_float32_lt( float32, float32 );
+flag syst_float32_eq_signaling( float32, float32 );
+flag syst_float32_le_quiet( float32, float32 );
+flag syst_float32_lt_quiet( float32, float32 );
+int32 syst_float64_to_int32( float64 );
+int32 syst_float64_to_int32_round_to_zero( float64 );
+#ifdef BITS64
+int64 syst_float64_to_int64( float64 );
+int64 syst_float64_to_int64_round_to_zero( float64 );
+#endif
+float32 syst_float64_to_float32( float64 );
+#ifdef FLOATX80
+floatx80 syst_float64_to_floatx80( float64 );
+#endif
+#ifdef FLOAT128
+float128 syst_float64_to_float128( float64 );
+#endif
+float64 syst_float64_round_to_int( float64 );
+float64 syst_float64_add( float64, float64 );
+float64 syst_float64_sub( float64, float64 );
+float64 syst_float64_mul( float64, float64 );
+float64 syst_float64_div( float64, float64 );
+float64 syst_float64_rem( float64, float64 );
+float64 syst_float64_sqrt( float64 );
+flag syst_float64_eq( float64, float64 );
+flag syst_float64_le( float64, float64 );
+flag syst_float64_lt( float64, float64 );
+flag syst_float64_eq_signaling( float64, float64 );
+flag syst_float64_le_quiet( float64, float64 );
+flag syst_float64_lt_quiet( float64, float64 );
+#ifdef FLOATX80
+int32 syst_floatx80_to_int32( floatx80 );
+int32 syst_floatx80_to_int32_round_to_zero( floatx80 );
+#ifdef BITS64
+int64 syst_floatx80_to_int64( floatx80 );
+int64 syst_floatx80_to_int64_round_to_zero( floatx80 );
+#endif
+float32 syst_floatx80_to_float32( floatx80 );
+float64 syst_floatx80_to_float64( floatx80 );
+#ifdef FLOAT128
+float128 syst_floatx80_to_float128( floatx80 );
+#endif
+floatx80 syst_floatx80_round_to_int( floatx80 );
+floatx80 syst_floatx80_add( floatx80, floatx80 );
+floatx80 syst_floatx80_sub( floatx80, floatx80 );
+floatx80 syst_floatx80_mul( floatx80, floatx80 );
+floatx80 syst_floatx80_div( floatx80, floatx80 );
+floatx80 syst_floatx80_rem( floatx80, floatx80 );
+floatx80 syst_floatx80_sqrt( floatx80 );
+flag syst_floatx80_eq( floatx80, floatx80 );
+flag syst_floatx80_le( floatx80, floatx80 );
+flag syst_floatx80_lt( floatx80, floatx80 );
+flag syst_floatx80_eq_signaling( floatx80, floatx80 );
+flag syst_floatx80_le_quiet( floatx80, floatx80 );
+flag syst_floatx80_lt_quiet( floatx80, floatx80 );
+#endif
+#ifdef FLOAT128
+int32 syst_float128_to_int32( float128 );
+int32 syst_float128_to_int32_round_to_zero( float128 );
+#ifdef BITS64
+int64 syst_float128_to_int64( float128 );
+int64 syst_float128_to_int64_round_to_zero( float128 );
+#endif
+float32 syst_float128_to_float32( float128 );
+float64 syst_float128_to_float64( float128 );
+#ifdef FLOATX80
+floatx80 syst_float128_to_floatx80( float128 );
+#endif
+float128 syst_float128_round_to_int( float128 );
+float128 syst_float128_add( float128, float128 );
+float128 syst_float128_sub( float128, float128 );
+float128 syst_float128_mul( float128, float128 );
+float128 syst_float128_div( float128, float128 );
+float128 syst_float128_rem( float128, float128 );
+float128 syst_float128_sqrt( float128 );
+flag syst_float128_eq( float128, float128 );
+flag syst_float128_le( float128, float128 );
+flag syst_float128_lt( float128, float128 );
+flag syst_float128_eq_signaling( float128, float128 );
+flag syst_float128_le_quiet( float128, float128 );
+flag syst_float128_lt_quiet( float128, float128 );
+#endif
+
diff --git a/testfloat/386-Win32-gcc/systmodes.S b/testfloat/386-Win32-gcc/systmodes.S
new file mode 100644
index 00000000000..a8f3c8bc889
--- /dev/null
+++ b/testfloat/386-Win32-gcc/systmodes.S
@@ -0,0 +1,82 @@
+
+/*
+===============================================================================
+
+This GNU assembler source file is part of TestFloat, Release 2a, a package
+of programs for testing the correctness of floating-point arithmetic
+complying to the IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+	.text
+
+/*
+-------------------------------------------------------------------------------
+Sets the system's IEC/IEEE floating-point rounding mode.  Also disables all
+system exception traps.
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global _syst_float_set_rounding_mode
+_syst_float_set_rounding_mode:
+	movb 4(%esp),%al
+	andb $3,%al
+	shlw $10,%ax
+	orw $63,%ax
+	subl $2,%esp
+	fnstcw 0(%esp)
+	andw $768,0(%esp)
+	orw %ax,0(%esp)
+	fldcw 0(%esp)
+	addl $2,%esp
+	ret
+
+/*
+-------------------------------------------------------------------------------
+Sets the rounding precision of subsequent extended double-precision
+operations.  The `precision' argument should be one of 0, 32, 64, or 80.
+If `precision' is 32, the rounding precision is set equivalent to single
+precision; else if `precision' is 64, the rounding precision is set
+equivalent to double precision; else the rounding precision is set to full
+extended double precision.
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global _syst_float_set_rounding_precision
+_syst_float_set_rounding_precision:
+	movb 4(%esp),%al
+	movb $0,%ah
+	cmpb $32,%al
+	je setRoundingPrecision
+	movb $2,%ah
+	cmpb $64,%al
+	je setRoundingPrecision
+	movb $3,%ah
+	cmpb $80,%al
+	je setRoundingPrecision
+	movb $0,%ah
+setRoundingPrecision:
+	movb $0,%al
+	subl $2,%esp
+	fnstcw 0(%esp)
+	andw $64767,0(%esp)
+	orw %ax,0(%esp)
+	fldcw 0(%esp)
+	addl $2,%esp
+	ret
+
diff --git a/testfloat/README.txt b/testfloat/README.txt
new file mode 100644
index 00000000000..fe99fd95d64
--- /dev/null
+++ b/testfloat/README.txt
@@ -0,0 +1,50 @@
+
+Package Overview for TestFloat Release 2a
+
+John R. Hauser
+1998 December 16
+
+
+TestFloat is a program for testing that a floating-point implementation
+conforms to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
+TestFloat is distributed in the form of C source code.  The TestFloat
+package actually provides two related programs:
+
+-- The `testfloat' program tests a system's floating-point for conformance
+   to the IEC/IEEE Standard.  This program uses the SoftFloat software
+   floating-point implementation as a basis for comparison.
+
+-- The `testsoftfloat' program tests SoftFloat itself for conformance to
+   the IEC/IEEE Standard.  These tests are performed by comparing against a
+   separate, slower software floating-point that is included in the TestFloat
+   package.
+
+TestFloat depends on SoftFloat, but SoftFloat is not included in the
+TestFloat package.  SoftFloat can be obtained through the Web page `http://
+HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/SoftFloat.html'.
+
+TestFloat is documented in three text files:
+
+   testfloat.txt          Documentation for using the TestFloat programs
+                              (both `testfloat' and `testsoftfloat').
+   testfloat-source.txt   Documentation for porting and compiling TestFloat.
+   testfloat-history.txt  History of major changes to TestFloat.
+
+The following file is also provided:
+
+   systemBugs.txt         Information about processor bugs found using
+                              TestFloat.
+
+Other files in the package comprise the source code for TestFloat.
+
+Please be aware that some work is involved in porting this software to other
+targets.  It is not just a matter of getting `make' to complete without
+error messages.  I would have written the code that way if I could, but
+there are fundamental differences between systems that I can't make go away.
+You should not attempt to compile the TestFloat sources without first
+reading `testfloat-source.txt'.
+
+At the time of this writing, the most up-to-date information about
+TestFloat and the latest release can be found at the Web page `http://
+HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
diff --git a/testfloat/SPARC-Solaris-gcc/Makefile b/testfloat/SPARC-Solaris-gcc/Makefile
new file mode 100644
index 00000000000..ee53aa19d33
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/Makefile
@@ -0,0 +1,64 @@
+
+PROCESSOR_H = ../../processors/SPARC-gcc.h
+SOFTFLOAT_VERSION = bits64
+TARGET = SPARC-Solaris-gcc
+SOFTFLOAT_DIR = ../../softfloat/$(SOFTFLOAT_VERSION)/$(TARGET)
+
+OBJ = .o
+EXE =
+INCLUDES = -I. -I.. -I$(SOFTFLOAT_DIR)
+COMPILE_ASM = gcc -c -o $@
+COMPILE_C = gcc -c -o $@ $(INCLUDES) -I- -O2
+COMPILE_SLOWFLOAT_C = gcc -c -o $@ $(INCLUDES) -I- -O3
+LINK = gcc -o $@
+
+SOFTFLOAT_H = $(SOFTFLOAT_DIR)/softfloat.h
+SOFTFLOAT_OBJ = $(SOFTFLOAT_DIR)/softfloat$(OBJ)
+
+ALL: testsoftfloat$(EXE) testfloat$(EXE)
+
+milieu.h: $(PROCESSOR_H)
+	touch milieu.h
+
+fail$(OBJ): milieu.h ../fail.h
+	$(COMPILE_C) ../fail.c
+
+random$(OBJ): milieu.h ../random.h
+	$(COMPILE_C) ../random.c
+
+testCases$(OBJ): milieu.h ../fail.h ../random.h $(SOFTFLOAT_H) ../testCases.h ../testCases.c
+	$(COMPILE_C) ../testCases.c
+
+writeHex$(OBJ): milieu.h $(SOFTFLOAT_H) ../writeHex.h ../writeHex.c
+	$(COMPILE_C) ../writeHex.c
+
+testLoops$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../writeHex.h ../testLoops.h ../testLoops.c
+	$(COMPILE_C) ../testLoops.c
+
+slowfloat$(OBJ): milieu.h $(SOFTFLOAT_H) ../slowfloat.h ../slowfloat-32.c ../slowfloat-64.c ../slowfloat.c
+	$(COMPILE_SLOWFLOAT_C) ../slowfloat.c
+
+testsoftfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../slowfloat.h ../testsoftfloat.c
+	$(COMPILE_C) ../testsoftfloat.c
+
+testsoftfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+
+systmodes$(OBJ): milieu.h ../systmodes.h systmodes.c
+	$(COMPILE_C) systmodes.c
+
+systflags$(OBJ): milieu.h ../systflags.h systflags.c
+	$(COMPILE_C) systflags.c
+
+systfloat$(OBJ): systfloat.S
+	$(COMPILE_ASM) systfloat.S
+
+testFunction$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systmodes.h ../systflags.h systfloat.h ../testFunction.h ../testFunction.c
+	$(COMPILE_C) ../testFunction.c
+
+testfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systflags.h ../testFunction.h ../testfloat.c
+	$(COMPILE_C) ../testfloat.c
+
+testfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+
diff --git a/testfloat/SPARC-Solaris-gcc/milieu.h b/testfloat/SPARC-Solaris-gcc/milieu.h
new file mode 100644
index 00000000000..4aa7b13f0ea
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/milieu.h
@@ -0,0 +1,51 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+Include common integer types and flags.
+-------------------------------------------------------------------------------
+*/
+#include "../../processors/SPARC-gcc.h"
+
+/*
+-------------------------------------------------------------------------------
+If the `BITS64' macro is defined by the processor header file but the
+version of SoftFloat being used/tested is the 32-bit one (`bits32'), the
+`BITS64' macro must be undefined here.
+-------------------------------------------------------------------------------
+#undef BITS64
+*/
+
+/*
+-------------------------------------------------------------------------------
+Symbolic Boolean literals.
+-------------------------------------------------------------------------------
+*/
+enum {
+    FALSE = 0,
+    TRUE  = 1
+};
+
diff --git a/testfloat/SPARC-Solaris-gcc/systflags.c b/testfloat/SPARC-Solaris-gcc/systflags.c
new file mode 100644
index 00000000000..d0e29a1a4b2
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/systflags.c
@@ -0,0 +1,42 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <ieeefp.h>
+#include "milieu.h"
+#include "systflags.h"
+
+/*
+-------------------------------------------------------------------------------
+Clears the system's IEC/IEEE floating-point exception flags.  Returns the
+previous value of the flags.
+-------------------------------------------------------------------------------
+*/
+int8 syst_float_flags_clear( void )
+{
+
+    return fpsetsticky( 0 );
+
+}
+
diff --git a/testfloat/SPARC-Solaris-gcc/systfloat.S b/testfloat/SPARC-Solaris-gcc/systfloat.S
new file mode 100644
index 00000000000..1c0264d5bf6
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/systfloat.S
@@ -0,0 +1,941 @@
+
+/*
+===============================================================================
+
+This GNU assembler source file is part of TestFloat, Release 2a, a package
+of programs for testing the correctness of floating-point arithmetic
+complying to the IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+	.text
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_int32_to_float32
+syst_int32_to_float32:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fitos %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_int32_to_float64
+syst_int32_to_float64:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fitod %f0,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_int32_to_float128
+syst_int32_to_float128:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fitoq %f0,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_to_int32_round_to_zero
+syst_float32_to_int32_round_to_zero:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fstoi %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_to_float64
+syst_float32_to_float64:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fstod %f0,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_to_float128
+syst_float32_to_float128:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fstoq %f0,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_add
+syst_float32_add:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fadds %f0,%f1,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_sub
+syst_float32_sub:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fsubs %f0,%f1,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_mul
+syst_float32_mul:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fmuls %f0,%f1,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_div
+syst_float32_div:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fdivs %f0,%f1,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_sqrt
+syst_float32_sqrt:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	ld [%sp+96],%f0
+	fsqrts %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_eq
+syst_float32_eq:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmps %f0,%f1
+	mov 0,%i0
+	fbe,a float32EqExit
+	mov 1,%i0
+float32EqExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_le
+syst_float32_le:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmpes %f0,%f1
+	mov 0,%i0
+	fble,a float32LeExit
+	mov 1,%i0
+float32LeExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_lt
+syst_float32_lt:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmpes %f0,%f1
+	mov 0,%i0
+	fbl,a float32LtExit
+	mov 1,%i0
+float32LtExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_eq_signaling
+syst_float32_eq_signaling:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmpes %f0,%f1
+	mov 0,%i0
+	fbe,a float32EqSignalingExit
+	mov 1,%i0
+float32EqSignalingExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_le_quiet
+syst_float32_le_quiet:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmps %f0,%f1
+	mov 0,%i0
+	fble,a float32LeQuietExit
+	mov 1,%i0
+float32LeQuietExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float32_lt_quiet
+syst_float32_lt_quiet:
+	save %sp,-128,%sp
+
+	st %i0,[%sp+96]
+	st %i1,[%sp+100]
+	ld [%sp+96],%f0
+	ld [%sp+100],%f1
+	fcmps %f0,%f1
+	mov 0,%i0
+	fbl,a float32LtQuietExit
+	mov 1,%i0
+float32LtQuietExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_to_int32_round_to_zero
+syst_float64_to_int32_round_to_zero:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	ldd [%sp+96],%f0
+	fdtoi %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_to_float32
+syst_float64_to_float32:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	ldd [%sp+96],%f0
+	fdtos %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_to_float128
+syst_float64_to_float128:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	ldd [%sp+96],%f0
+	fdtoq %f0,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_add
+syst_float64_add:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	faddd %f0,%f2,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_sub
+syst_float64_sub:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fsubd %f0,%f2,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_mul
+syst_float64_mul:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fmuld %f0,%f2,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_div
+syst_float64_div:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fdivd %f0,%f2,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_sqrt
+syst_float64_sqrt:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	ldd [%sp+96],%f0
+	fsqrtd %f0,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_eq
+syst_float64_eq:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmpd %f0,%f2
+	mov 0,%i0
+	fbe,a float64EqExit
+	mov 1,%i0
+float64EqExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_le
+syst_float64_le:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmped %f0,%f2
+	mov 0,%i0
+	fble,a float64LeExit
+	mov 1,%i0
+float64LeExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_lt
+syst_float64_lt:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmped %f0,%f2
+	mov 0,%i0
+	fbl,a float64LtExit
+	mov 1,%i0
+float64LtExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_eq_signaling
+syst_float64_eq_signaling:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmped %f0,%f2
+	mov 0,%i0
+	fbe,a float64EqSignalingExit
+	mov 1,%i0
+float64EqSignalingExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_le_quiet
+syst_float64_le_quiet:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmpd %f0,%f2
+	mov 0,%i0
+	fble,a float64LeQuietExit
+	mov 1,%i0
+float64LeQuietExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float64_lt_quiet
+syst_float64_lt_quiet:
+	save %sp,-128,%sp
+
+	std %i0,[%sp+96]
+	std %i2,[%sp+104]
+	ldd [%sp+96],%f0
+	ldd [%sp+104],%f2
+	fcmpd %f0,%f2
+	mov 0,%i0
+	fbl,a float64LtQuietExit
+	mov 1,%i0
+float64LtQuietExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_to_int32_round_to_zero
+syst_float128_to_int32_round_to_zero:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	fqtoi %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_to_float32
+syst_float128_to_float32:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	fqtos %f0,%f0
+	st %f0,[%sp+96]
+	ld [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_to_float64
+syst_float128_to_float64:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	fqtod %f0,%f0
+	std %f0,[%sp+96]
+	ldd [%sp+96],%i0
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_add
+syst_float128_add:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	faddq %f0,%f4,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_sub
+syst_float128_sub:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fsubq %f0,%f4,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_mul
+syst_float128_mul:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fmulq %f0,%f4,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_div
+syst_float128_div:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fdivq %f0,%f4,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_sqrt
+syst_float128_sqrt:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	fsqrtq %f0,%f0
+	ld [%sp+192],%o0
+	std %f0,[%o0]
+	std %f2,[%o0+8]
+
+	jmp %i7+12
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_eq
+syst_float128_eq:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpq %f0,%f4
+	mov 0,%i0
+	fbe,a float128EqExit
+	mov 1,%i0
+float128EqExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_le
+syst_float128_le:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpeq %f0,%f4
+	mov 0,%i0
+	fble,a float128Le
+	mov 1,%i0
+float128Le:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_lt
+syst_float128_lt:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpeq %f0,%f4
+	mov 0,%i0
+	fbl,a float128Lt
+	mov 1,%i0
+float128Lt:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_eq_signaling
+syst_float128_eq_signaling:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpeq %f0,%f4
+	mov 0,%i0
+	fbe,a float128EqSignalingExit
+	mov 1,%i0
+float128EqSignalingExit:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_le_quiet
+syst_float128_le_quiet:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpq %f0,%f4
+	mov 0,%i0
+	fble,a float128LeQuiet
+	mov 1,%i0
+float128LeQuiet:
+
+	ret
+	restore
+
+/*
+-------------------------------------------------------------------------------
+-------------------------------------------------------------------------------
+*/
+	.align 4
+	.global syst_float128_lt_quiet
+syst_float128_lt_quiet:
+	save %sp,-128,%sp
+
+	ldd [%i0],%f0
+	ldd [%i0+8],%f2
+	ldd [%i1],%f4
+	ldd [%i1+8],%f6
+	fcmpq %f0,%f4
+	mov 0,%i0
+	fbl,a float128LtQuiet
+	mov 1,%i0
+float128LtQuiet:
+
+	ret
+	restore
+
diff --git a/testfloat/SPARC-Solaris-gcc/systfloat.h b/testfloat/SPARC-Solaris-gcc/systfloat.h
new file mode 100644
index 00000000000..956e470f4fb
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/systfloat.h
@@ -0,0 +1,205 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+The following macros are defined to indicate that the corresponding
+functions exist.
+-------------------------------------------------------------------------------
+*/
+#define SYST_INT32_TO_FLOAT32
+#define SYST_INT32_TO_FLOAT64
+#define SYST_INT32_TO_FLOAT128
+#define SYST_FLOAT32_TO_INT32_ROUND_TO_ZERO
+#define SYST_FLOAT32_TO_FLOAT64
+#define SYST_FLOAT32_TO_FLOAT128
+#define SYST_FLOAT32_ADD
+#define SYST_FLOAT32_SUB
+#define SYST_FLOAT32_MUL
+#define SYST_FLOAT32_DIV
+#define SYST_FLOAT32_SQRT
+#define SYST_FLOAT32_EQ
+#define SYST_FLOAT32_LE
+#define SYST_FLOAT32_LT
+#define SYST_FLOAT32_EQ_SIGNALING
+#define SYST_FLOAT32_LE_QUIET
+#define SYST_FLOAT32_LT_QUIET
+#define SYST_FLOAT64_TO_INT32_ROUND_TO_ZERO
+#define SYST_FLOAT64_TO_FLOAT32
+#define SYST_FLOAT64_TO_FLOAT128
+#define SYST_FLOAT64_ADD
+#define SYST_FLOAT64_SUB
+#define SYST_FLOAT64_MUL
+#define SYST_FLOAT64_DIV
+#define SYST_FLOAT64_SQRT
+#define SYST_FLOAT64_EQ
+#define SYST_FLOAT64_LE
+#define SYST_FLOAT64_LT
+#define SYST_FLOAT64_EQ_SIGNALING
+#define SYST_FLOAT64_LE_QUIET
+#define SYST_FLOAT64_LT_QUIET
+#define SYST_FLOAT128_TO_INT32_ROUND_TO_ZERO
+#define SYST_FLOAT128_TO_FLOAT32
+#define SYST_FLOAT128_TO_FLOAT64
+#define SYST_FLOAT128_ADD
+#define SYST_FLOAT128_SUB
+#define SYST_FLOAT128_MUL
+#define SYST_FLOAT128_DIV
+#define SYST_FLOAT128_SQRT
+#define SYST_FLOAT128_EQ
+#define SYST_FLOAT128_LE
+#define SYST_FLOAT128_LT
+#define SYST_FLOAT128_EQ_SIGNALING
+#define SYST_FLOAT128_LE_QUIET
+#define SYST_FLOAT128_LT_QUIET
+
+/*
+-------------------------------------------------------------------------------
+System function declarations.  (Some of these functions may not exist.)
+-------------------------------------------------------------------------------
+*/
+float32 syst_int32_to_float32( int32 );
+float64 syst_int32_to_float64( int32 );
+#ifdef FLOATX80
+floatx80 syst_int32_to_floatx80( int32 );
+#endif
+#ifdef FLOAT128
+float128 syst_int32_to_float128( int32 );
+#endif
+#ifdef BITS64
+float32 syst_int64_to_float32( int64 );
+float64 syst_int64_to_float64( int64 );
+#ifdef FLOATX80
+floatx80 syst_int64_to_floatx80( int64 );
+#endif
+#ifdef FLOAT128
+float128 syst_int64_to_float128( int64 );
+#endif
+#endif
+int32 syst_float32_to_int32( float32 );
+int32 syst_float32_to_int32_round_to_zero( float32 );
+#ifdef BITS64
+int64 syst_float32_to_int64( float32 );
+int64 syst_float32_to_int64_round_to_zero( float32 );
+#endif
+float64 syst_float32_to_float64( float32 );
+#ifdef FLOATX80
+floatx80 syst_float32_to_floatx80( float32 );
+#endif
+#ifdef FLOAT128
+float128 syst_float32_to_float128( float32 );
+#endif
+float32 syst_float32_round_to_int( float32 );
+float32 syst_float32_add( float32, float32 );
+float32 syst_float32_sub( float32, float32 );
+float32 syst_float32_mul( float32, float32 );
+float32 syst_float32_div( float32, float32 );
+float32 syst_float32_rem( float32, float32 );
+float32 syst_float32_sqrt( float32 );
+flag syst_float32_eq( float32, float32 );
+flag syst_float32_le( float32, float32 );
+flag syst_float32_lt( float32, float32 );
+flag syst_float32_eq_signaling( float32, float32 );
+flag syst_float32_le_quiet( float32, float32 );
+flag syst_float32_lt_quiet( float32, float32 );
+int32 syst_float64_to_int32( float64 );
+int32 syst_float64_to_int32_round_to_zero( float64 );
+#ifdef BITS64
+int64 syst_float64_to_int64( float64 );
+int64 syst_float64_to_int64_round_to_zero( float64 );
+#endif
+float32 syst_float64_to_float32( float64 );
+#ifdef FLOATX80
+floatx80 syst_float64_to_floatx80( float64 );
+#endif
+#ifdef FLOAT128
+float128 syst_float64_to_float128( float64 );
+#endif
+float64 syst_float64_round_to_int( float64 );
+float64 syst_float64_add( float64, float64 );
+float64 syst_float64_sub( float64, float64 );
+float64 syst_float64_mul( float64, float64 );
+float64 syst_float64_div( float64, float64 );
+float64 syst_float64_rem( float64, float64 );
+float64 syst_float64_sqrt( float64 );
+flag syst_float64_eq( float64, float64 );
+flag syst_float64_le( float64, float64 );
+flag syst_float64_lt( float64, float64 );
+flag syst_float64_eq_signaling( float64, float64 );
+flag syst_float64_le_quiet( float64, float64 );
+flag syst_float64_lt_quiet( float64, float64 );
+#ifdef FLOATX80
+int32 syst_floatx80_to_int32( floatx80 );
+int32 syst_floatx80_to_int32_round_to_zero( floatx80 );
+#ifdef BITS64
+int64 syst_floatx80_to_int64( floatx80 );
+int64 syst_floatx80_to_int64_round_to_zero( floatx80 );
+#endif
+float32 syst_floatx80_to_float32( floatx80 );
+float64 syst_floatx80_to_float64( floatx80 );
+#ifdef FLOAT128
+float128 syst_floatx80_to_float128( floatx80 );
+#endif
+floatx80 syst_floatx80_round_to_int( floatx80 );
+floatx80 syst_floatx80_add( floatx80, floatx80 );
+floatx80 syst_floatx80_sub( floatx80, floatx80 );
+floatx80 syst_floatx80_mul( floatx80, floatx80 );
+floatx80 syst_floatx80_div( floatx80, floatx80 );
+floatx80 syst_floatx80_rem( floatx80, floatx80 );
+floatx80 syst_floatx80_sqrt( floatx80 );
+flag syst_floatx80_eq( floatx80, floatx80 );
+flag syst_floatx80_le( floatx80, floatx80 );
+flag syst_floatx80_lt( floatx80, floatx80 );
+flag syst_floatx80_eq_signaling( floatx80, floatx80 );
+flag syst_floatx80_le_quiet( floatx80, floatx80 );
+flag syst_floatx80_lt_quiet( floatx80, floatx80 );
+#endif
+#ifdef FLOAT128
+int32 syst_float128_to_int32( float128 );
+int32 syst_float128_to_int32_round_to_zero( float128 );
+#ifdef BITS64
+int64 syst_float128_to_int64( float128 );
+int64 syst_float128_to_int64_round_to_zero( float128 );
+#endif
+float32 syst_float128_to_float32( float128 );
+float64 syst_float128_to_float64( float128 );
+#ifdef FLOATX80
+floatx80 syst_float128_to_floatx80( float128 );
+#endif
+float128 syst_float128_round_to_int( float128 );
+float128 syst_float128_add( float128, float128 );
+float128 syst_float128_sub( float128, float128 );
+float128 syst_float128_mul( float128, float128 );
+float128 syst_float128_div( float128, float128 );
+float128 syst_float128_rem( float128, float128 );
+float128 syst_float128_sqrt( float128 );
+flag syst_float128_eq( float128, float128 );
+flag syst_float128_le( float128, float128 );
+flag syst_float128_lt( float128, float128 );
+flag syst_float128_eq_signaling( float128, float128 );
+flag syst_float128_le_quiet( float128, float128 );
+flag syst_float128_lt_quiet( float128, float128 );
+#endif
+
diff --git a/testfloat/SPARC-Solaris-gcc/systmodes.c b/testfloat/SPARC-Solaris-gcc/systmodes.c
new file mode 100644
index 00000000000..3981cd79a6a
--- /dev/null
+++ b/testfloat/SPARC-Solaris-gcc/systmodes.c
@@ -0,0 +1,51 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <ieeefp.h>
+#include "milieu.h"
+#include "systmodes.h"
+
+/*
+-------------------------------------------------------------------------------
+Sets the system's IEC/IEEE floating-point rounding mode.
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_mode( int8 roundingMode )
+{
+
+    (void) fpsetround( roundingMode );
+
+}
+
+/*
+-------------------------------------------------------------------------------
+Does nothing.
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_precision( int8 precision )
+{
+
+}
+
diff --git a/testfloat/fail.c b/testfloat/fail.c
new file mode 100644
index 00000000000..30bbea6bd7d
--- /dev/null
+++ b/testfloat/fail.c
@@ -0,0 +1,46 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdlib.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include "milieu.h"
+#include "fail.h"
+
+char *fail_programName = "";
+
+void fail( const char *message, ... )
+{
+    va_list varArgs;
+
+    fprintf( stderr, "%s: ", fail_programName );
+    va_start( varArgs, message );
+    vfprintf( stderr, message, varArgs );
+    va_end( varArgs );
+    fputs( ".\n", stderr );
+    exit( EXIT_FAILURE );
+
+}
+
diff --git a/testfloat/fail.h b/testfloat/fail.h
new file mode 100644
index 00000000000..9c338da260a
--- /dev/null
+++ b/testfloat/fail.h
@@ -0,0 +1,29 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+extern char *fail_programName;
+
+void fail( const char *, ... );
+
diff --git a/testfloat/random.c b/testfloat/random.c
new file mode 100644
index 00000000000..53453665760
--- /dev/null
+++ b/testfloat/random.c
@@ -0,0 +1,65 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdlib.h>
+#include "milieu.h"
+#include "random.h"
+
+uint8 randomUint8( void )
+{
+
+    return (bits8) ( rand()>>4 );
+
+}
+
+uint16 randomUint16( void )
+{
+
+    return ( ( rand() & 0x0FF0 )<<4 ) | ( ( rand()>>4 ) & 0xFF );
+
+}
+
+uint32 randomUint32( void )
+{
+
+    return
+          ( ( (uint32) ( rand() & 0x0FF0 ) )<<20 )
+        | ( ( (uint32) ( rand() & 0x0FF0 ) )<<12 )
+        | ( ( rand() & 0x0FF0 )<<4 )
+        | ( ( rand()>>4 ) & 0xFF );
+
+}
+
+#ifdef BITS64
+
+uint64 randomUint64( void )
+{
+
+    return ( ( (uint64) randomUint32() )<<32 ) | randomUint32();
+
+}
+
+#endif
+
diff --git a/testfloat/random.h b/testfloat/random.h
new file mode 100644
index 00000000000..7375499ea96
--- /dev/null
+++ b/testfloat/random.h
@@ -0,0 +1,32 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+uint8 randomUint8( void );
+uint16 randomUint16( void );
+uint32 randomUint32( void );
+#ifdef BITS64
+uint64 randomUint64( void );
+#endif
+
diff --git a/testfloat/slowfloat-32.c b/testfloat/slowfloat-32.c
new file mode 100644
index 00000000000..549654b05e6
--- /dev/null
+++ b/testfloat/slowfloat-32.c
@@ -0,0 +1,1183 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+int8 slow_float_rounding_mode;
+int8 slow_float_exception_flags;
+int8 slow_float_detect_tininess;
+
+typedef struct {
+    bits32 a0, a1;
+} bits64X;
+
+typedef struct {
+    flag isNaN;
+    flag isInf;
+    flag isZero;
+    flag sign;
+    int16 exp;
+    bits64X sig;
+} floatX;
+
+static const floatX floatXNaN = { TRUE, FALSE, FALSE, FALSE, 0, { 0, 0 } };
+static const floatX floatXPositiveZero =
+    { FALSE, FALSE, TRUE, FALSE, 0, { 0, 0 } };
+static const floatX floatXNegativeZero =
+    { FALSE, FALSE, TRUE, TRUE, 0, { 0, 0 } };
+
+static bits64X shortShift64Left( bits64X a, int8 shiftCount )
+{
+    int8 negShiftCount;
+
+    negShiftCount = ( - shiftCount & 31 );
+    a.a0 = ( a.a0<<shiftCount ) | ( a.a1>>negShiftCount );
+    a.a1 <<= shiftCount;
+    return a;
+
+}
+
+static bits64X shortShift64RightJamming( bits64X a, int8 shiftCount )
+{
+    int8 negShiftCount;
+    bits32 extra;
+
+    negShiftCount = ( - shiftCount & 31 );
+    extra = a.a1<<negShiftCount;
+    a.a1 = ( a.a0<<negShiftCount ) | ( a.a1>>shiftCount ) | ( extra != 0 );
+    a.a0 >>= shiftCount;
+    return a;
+
+}
+
+static bits64X neg64( bits64X a )
+{
+
+    if ( a.a1 == 0 ) {
+        a.a0 = - a.a0;
+    }
+    else {
+        a.a1 = - a.a1;
+        a.a0 = ~ a.a0;
+    }
+    return a;
+
+}
+
+static bits64X add64( bits64X a, bits64X b )
+{
+
+    a.a1 += b.a1;
+    a.a0 += b.a0 + ( a.a1 < b.a1 );
+    return a;
+
+}
+
+static flag eq64( bits64X a, bits64X b )
+{
+
+    return ( a.a0 == b.a0 ) && ( a.a1 == b.a1 );
+
+}
+
+static flag le64( bits64X a, bits64X b )
+{
+
+    return ( a.a0 < b.a0 ) || ( ( a.a0 == b.a0 ) && ( a.a1 <= b.a1 ) );
+
+}
+
+static flag lt64( bits64X a, bits64X b )
+{
+
+    return ( a.a0 < b.a0 ) || ( ( a.a0 == b.a0 ) && ( a.a1 < b.a1 ) );
+
+}
+
+static floatX roundFloatXTo24( flag isTiny, floatX zx )
+{
+
+    if ( zx.sig.a1 ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( zx.sig.a1 < 0x80000000 ) goto noIncrement;
+            if ( ( zx.sig.a1 == 0x80000000 ) && ! ( zx.sig.a0 & 1 ) ) {
+                goto noIncrement;
+            }
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        ++zx.sig.a0;
+        if ( zx.sig.a0 == 0x01000000 ) {
+            zx.sig.a0 = 0x00800000;
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    zx.sig.a1 = 0;
+    return zx;
+
+}
+
+static floatX roundFloatXTo53( flag isTiny, floatX zx )
+{
+    int8 roundBits;
+
+    roundBits = zx.sig.a1 & 7;
+    zx.sig.a1 -= roundBits;
+    if ( roundBits ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( roundBits < 4 ) goto noIncrement;
+            if ( ( roundBits == 4 ) && ! ( zx.sig.a1 & 8 ) ) goto noIncrement;
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        zx.sig.a1 += 8;
+        zx.sig.a0 += ( zx.sig.a1 == 0 );
+        if ( zx.sig.a0 == 0x01000000 ) {
+            zx.sig.a0 = 0x00800000;
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    return zx;
+
+}
+
+static floatX int32ToFloatX( int32 a )
+{
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.sign = ( a < 0 );
+    ax.sig.a1 = ax.sign ? - a : a;
+    ax.sig.a0 = 0;
+    if ( a == 0 ) {
+        ax.isZero = TRUE;
+        return ax;
+    }
+    ax.isZero = FALSE;
+    ax.sig = shortShift64Left( ax.sig, 23 );
+    ax.exp = 32;
+    while ( ax.sig.a0 < 0x00800000 ) {
+        ax.sig = shortShift64Left( ax.sig, 1 );
+        --ax.exp;
+    }
+    return ax;
+
+}
+
+static int32 floatXToInt32( floatX ax )
+{
+    int8 savedExceptionFlags;
+    int16 shiftCount;
+    int32 z;
+
+    if ( ax.isInf || ax.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+        return ( ax.isInf & ax.sign ) ? 0x80000000 : 0x7FFFFFFF;
+    }
+    if ( ax.isZero ) return 0;
+    savedExceptionFlags = slow_float_exception_flags;
+    shiftCount = 52 - ax.exp;
+    if ( 56 < shiftCount ) {
+        ax.sig.a1 = 1;
+        ax.sig.a0 = 0;
+    }
+    else {
+        while ( 0 < shiftCount ) {
+            ax.sig = shortShift64RightJamming( ax.sig, 1 );
+            --shiftCount;
+        }
+    }
+    ax = roundFloatXTo53( FALSE, ax );
+    ax.sig = shortShift64RightJamming( ax.sig, 3 );
+    z = ax.sig.a1;
+    if ( ax.sign ) z = - z;
+    if (    ( shiftCount < 0 )
+         || ax.sig.a0
+         || ( ( z != 0 ) && ( ( ax.sign ^ ( z < 0 ) ) != 0 ) )
+       ) {
+        slow_float_exception_flags = savedExceptionFlags | float_flag_invalid;
+        return ax.sign ? 0x80000000 : 0x7FFFFFFF;
+    }
+    return z;
+
+}
+
+static floatX float32ToFloatX( float32 a )
+{
+    int16 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+    ax.sign = ( ( a & 0x80000000 ) != 0 );
+    expField = ( a>>23 ) & 0xFF;
+    ax.sig.a1 = 0;
+    ax.sig.a0 = a & 0x007FFFFF;
+    if ( expField == 0 ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x7F;
+            do {
+                ax.sig.a0 <<= 1;
+                --expField;
+            } while ( ax.sig.a0 < 0x00800000 );
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0xFF ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.sig.a0 |= 0x00800000;
+        ax.exp = expField - 0x7F;
+    }
+    return ax;
+
+}
+
+static float32 floatXToFloat32( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int16 expField;
+    float32 z;
+
+    if ( zx.isZero ) return zx.sign ? 0x80000000 : 0;
+    if ( zx.isInf ) return zx.sign ? 0xFF800000 : 0x7F800000;
+    if ( zx.isNaN ) return 0xFFFFFFFF;
+    while ( 0x01000000 <= zx.sig.a0 ) {
+        zx.sig = shortShift64RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < 0x00800000 ) {
+        zx.sig = shortShift64Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x7F <= 0 );
+    zx = roundFloatXTo24( isTiny, zx );
+    expField = zx.exp + 0x7F;
+    if ( 0xFF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z = 0xFF800000;
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z = 0xFF7FFFFF;
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z = 0x7F800000;
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z = 0x7F7FFFFF;
+                break;
+            }
+        }
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x7F;
+        if ( expField < -27 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift64RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        zx = roundFloatXTo24( isTiny, zx );
+        expField = ( 0x00800000 <= zx.sig.a0 ) ? 1 : 0;
+    }
+    z = expField;
+    z <<= 23;
+    if ( zx.sign ) z |= 0x80000000;
+    z |= zx.sig.a0 & 0x007FFFFF;
+    return z;
+
+}
+
+static floatX float64ToFloatX( float64 a )
+{
+    int16 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+#ifdef BITS64
+    ax.sign = ( ( a & LIT64( 0x8000000000000000 ) ) != 0 );
+    expField = ( a>>52 ) & 0x7FF;
+    ax.sig.a1 = a;
+    ax.sig.a0 = ( a>>32 ) & 0x000FFFFF;
+#else
+    ax.sign = ( ( a.high & 0x80000000 ) != 0 );
+    expField = ( a.high>>( 52 - 32 ) ) & 0x7FF;
+    ax.sig.a1 = a.low;
+    ax.sig.a0 = a.high & 0x000FFFFF;
+#endif
+    if ( expField == 0 ) {
+        if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x3FF;
+            do {
+                ax.sig = shortShift64Left( ax.sig, 1 );
+                --expField;
+            } while ( ax.sig.a0 < 0x00100000 );
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0x7FF ) {
+        if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.exp = expField - 0x3FF;
+        ax.sig.a0 |= 0x00100000;
+    }
+    ax.sig = shortShift64Left( ax.sig, 3 );
+    return ax;
+
+}
+
+static float64 floatXToFloat64( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int16 expField;
+    float64 z;
+
+#ifdef BITS64
+    if ( zx.isZero ) return zx.sign ? LIT64( 0x8000000000000000 ) : 0;
+    if ( zx.isInf ) {
+        return
+              zx.sign ? LIT64( 0xFFF0000000000000 )
+            : LIT64( 0x7FF0000000000000 );
+    }
+    if ( zx.isNaN ) return LIT64( 0xFFFFFFFFFFFFFFFF );
+#else
+    if ( zx.isZero ) {
+        z.low = 0;
+        z.high = zx.sign ? 0x80000000 : 0;
+        return z;
+    }
+    if ( zx.isInf ) {
+        z.low = 0;
+        z.high = zx.sign ? 0xFFF00000 : 0x7FF00000;
+        return z;
+    }
+    if ( zx.isNaN ) {
+        z.high = z.low = 0xFFFFFFFF;
+        return z;
+    }
+#endif
+    while ( 0x01000000 <= zx.sig.a0 ) {
+        zx.sig = shortShift64RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < 0x00800000 ) {
+        zx.sig = shortShift64Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x3FF <= 0 );
+    zx = roundFloatXTo53( isTiny, zx );
+    expField = zx.exp + 0x3FF;
+    if ( 0x7FF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+#ifdef BITS64
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z = LIT64( 0xFFF0000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z = LIT64( 0xFFEFFFFFFFFFFFFF );
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z = LIT64( 0x7FF0000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z = LIT64( 0x7FEFFFFFFFFFFFFF );
+                break;
+            }
+        }
+#else
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z.low = 0;
+                z.high = 0xFFF00000;
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z.low = 0xFFFFFFFF;
+                z.high = 0xFFEFFFFF;
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z.low = 0;
+                z.high = 0x7FF00000;
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z.low = 0xFFFFFFFF;
+                z.high = 0x7FEFFFFF;
+                break;
+            }
+        }
+#endif
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x3FF;
+        if ( expField < -56 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift64RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        zx = roundFloatXTo53( isTiny, zx );
+        expField = ( 0x00800000 <= zx.sig.a0 ) ? 1 : 0;
+    }
+    zx.sig = shortShift64RightJamming( zx.sig, 3 );
+#ifdef BITS64
+    z = expField;
+    z <<= 52;
+    if ( zx.sign ) z |= LIT64( 0x8000000000000000 );
+    z |= ( ( (bits64) ( zx.sig.a0 & 0x000FFFFF ) )<<32 ) | zx.sig.a1;
+#else
+    z.low = zx.sig.a1;
+    z.high = expField;
+    z.high <<= 52 - 32;
+    if ( zx.sign ) z.high |= 0x80000000;
+    z.high |= zx.sig.a0 & 0x000FFFFF;
+#endif
+    return z;
+
+}
+
+static floatX floatXInvalid( void )
+{
+
+    slow_float_exception_flags |= float_flag_invalid;
+    return floatXNaN;
+
+}
+
+static floatX floatXRoundToInt( floatX ax )
+{
+    int16 shiftCount, i;
+
+    if ( ax.isNaN || ax.isInf ) return ax;
+    shiftCount = 52 - ax.exp;
+    if ( shiftCount <= 0 ) return ax;
+    if ( 55 < shiftCount ) {
+        ax.exp = 52;
+        ax.sig.a1 = ! ax.isZero;
+        ax.sig.a0 = 0;
+    }
+    else {
+        while ( 0 < shiftCount ) {
+            ax.sig = shortShift64RightJamming( ax.sig, 1 );
+            ++ax.exp;
+            --shiftCount;
+        }
+    }
+    ax = roundFloatXTo53( FALSE, ax );
+    if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) ax.isZero = TRUE;
+    return ax;
+
+}
+
+static floatX floatXAdd( floatX ax, floatX bx )
+{
+    int16 expDiff;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf && bx.isInf ) {
+        if ( ax.sign == bx.sign ) return ax;
+        return floatXInvalid();
+    }
+    if ( ax.isInf ) return ax;
+    if ( bx.isInf ) return bx;
+    if ( ax.isZero && bx.isZero ) {
+        if ( ax.sign == bx.sign ) return ax;
+        goto completeCancellation;
+    }
+    if (    ( ax.sign != bx.sign )
+         && ( ax.exp == bx.exp )
+         && eq64( ax.sig, bx.sig )
+       ) {
+ completeCancellation:
+        return
+              ( slow_float_rounding_mode == float_round_down ) ?
+                  floatXNegativeZero
+            : floatXPositiveZero;
+    }
+    if ( ax.isZero ) return bx;
+    if ( bx.isZero ) return ax;
+    expDiff = ax.exp - bx.exp;
+    if ( expDiff < 0 ) {
+        zx = ax;
+        zx.exp = bx.exp;
+        if ( expDiff < -56 ) {
+            zx.sig.a1 = 1;
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expDiff < 0 ) {
+                zx.sig = shortShift64RightJamming( zx.sig, 1 );
+                ++expDiff;
+            }
+        }
+        if ( ax.sign != bx.sign ) zx.sig = neg64( zx.sig );
+        zx.sign = bx.sign;
+        zx.sig = add64( zx.sig, bx.sig );
+    }
+    else {
+        zx = bx;
+        zx.exp = ax.exp;
+        if ( 56 < expDiff ) {
+            zx.sig.a1 = 1;
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( 0 < expDiff ) {
+                zx.sig = shortShift64RightJamming( zx.sig, 1 );
+                --expDiff;
+            }
+        }
+        if ( ax.sign != bx.sign ) zx.sig = neg64( zx.sig );
+        zx.sign = ax.sign;
+        zx.sig = add64( zx.sig, ax.sig );
+    }
+    if ( zx.sig.a0 & 0x80000000 ) {
+        zx.sig = neg64( zx.sig );
+        zx.sign = ! zx.sign;
+    }
+    return zx;
+
+}
+
+static floatX floatXMul( floatX ax, floatX bx )
+{
+    int8 bitNum;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf ) {
+        if ( bx.isZero ) return floatXInvalid();
+        if ( bx.sign ) ax.sign = ! ax.sign;
+        return ax;
+    }
+    if ( bx.isInf ) {
+        if ( ax.isZero ) return floatXInvalid();
+        if ( ax.sign ) bx.sign = ! bx.sign;
+        return bx;
+    }
+    zx = ax;
+    zx.sign ^= bx.sign;
+    if ( ax.isZero || bx.isZero ) {
+        return zx.sign ? floatXNegativeZero : floatXPositiveZero;
+    }
+    zx.exp += bx.exp + 1;
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    for ( bitNum = 0; bitNum < 55; ++bitNum ) {
+        if ( bx.sig.a1 & 2 ) zx.sig = add64( zx.sig, ax.sig );
+        bx.sig = shortShift64RightJamming( bx.sig, 1 );
+        zx.sig = shortShift64RightJamming( zx.sig, 1 );
+    }
+    return zx;
+
+}
+
+static floatX floatXDiv( floatX ax, floatX bx )
+{
+    bits64X negBSig;
+    int8 bitNum;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf ) {
+        if ( bx.isInf ) return floatXInvalid();
+        if ( bx.sign ) ax.sign = ! ax.sign;
+        return ax;
+    }
+    if ( bx.isZero ) {
+        if ( ax.isZero ) return floatXInvalid();
+        slow_float_exception_flags |= float_flag_divbyzero;
+        if ( ax.sign ) bx.sign = ! bx.sign;
+        bx.isZero = FALSE;
+        bx.isInf = TRUE;
+        return bx;
+    }
+    zx = ax;
+    zx.sign ^= bx.sign;
+    if ( ax.isZero || bx.isInf ) {
+        return zx.sign ? floatXNegativeZero : floatXPositiveZero;
+    }
+    zx.exp -= bx.exp + 1;
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    negBSig = neg64( bx.sig );
+    for ( bitNum = 0; bitNum < 56; ++bitNum ) {
+        if ( le64( bx.sig, ax.sig ) ) {
+            zx.sig.a1 |= 1;
+            ax.sig = add64( ax.sig, negBSig );
+        }
+        ax.sig = shortShift64Left( ax.sig, 1 );
+        zx.sig = shortShift64Left( zx.sig, 1 );
+    }
+    if ( ax.sig.a0 || ax.sig.a1 ) zx.sig.a1 |= 1;
+    return zx;
+
+}
+
+static floatX floatXRem( floatX ax, floatX bx )
+{
+    bits64X negBSig;
+    flag lastQuotientBit;
+    bits64X savedASig;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf || bx.isZero ) return floatXInvalid();
+    if ( ax.isZero || bx.isInf ) return ax;
+    --bx.exp;
+    if ( ax.exp < bx.exp ) return ax;
+    bx.sig = shortShift64Left( bx.sig, 1 );
+    negBSig = neg64( bx.sig );
+    while ( bx.exp < ax.exp ) {
+        if ( le64( bx.sig, ax.sig ) ) ax.sig = add64( ax.sig, negBSig );
+        ax.sig = shortShift64Left( ax.sig, 1 );
+        --ax.exp;
+    }
+    lastQuotientBit = le64( bx.sig, ax.sig );
+    if ( lastQuotientBit ) ax.sig = add64( ax.sig, negBSig );
+    savedASig = ax.sig;
+    ax.sig = neg64( add64( ax.sig, negBSig ) );
+    if ( lt64( ax.sig, savedASig ) ) {
+        ax.sign = ! ax.sign;
+    }
+    else if ( lt64( savedASig, ax.sig ) ) {
+        ax.sig = savedASig;
+    }
+    else {
+        if ( lastQuotientBit ) {
+            ax.sign = ! ax.sign;
+        }
+        else {
+            ax.sig = savedASig;
+        }
+    }
+    if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) ax.isZero = TRUE;
+    return ax;
+
+}
+
+static floatX floatXSqrt( floatX ax )
+{
+    int8 bitNum;
+    bits64X bitSig, savedASig;
+    floatX zx;
+
+    if ( ax.isNaN || ax.isZero ) return ax;
+    if ( ax.sign ) return floatXInvalid();
+    if ( ax.isInf ) return ax;
+    zx = ax;
+    zx.exp >>= 1;
+    if ( ( ax.exp & 1 ) == 0 ) ax.sig = shortShift64RightJamming( ax.sig, 1 );
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    bitSig.a1 = 0;
+    bitSig.a0 = 0x00800000;
+    for ( bitNum = 0; bitNum < 56; ++bitNum ) {
+        savedASig = ax.sig;
+        ax.sig = add64( ax.sig, neg64( zx.sig ) );
+        ax.sig = shortShift64Left( ax.sig, 1 );
+        ax.sig = add64( ax.sig, neg64( bitSig ) );
+        if ( ax.sig.a0 & 0x80000000 ) {
+            ax.sig = shortShift64Left( savedASig, 1 );
+        }
+        else {
+            zx.sig.a1 |= bitSig.a1;
+            zx.sig.a0 |= bitSig.a0;
+        }
+        bitSig = shortShift64RightJamming( bitSig, 1 );
+    }
+    if ( ax.sig.a0 || ax.sig.a1 ) zx.sig.a1 |= 1;
+    return zx;
+
+}
+
+static flag floatXEq( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return TRUE;
+    if ( ax.sign != bx.sign ) return FALSE;
+    if ( ax.isInf || bx.isInf ) return ax.isInf && bx.isInf;
+    return ( ax.exp == bx.exp ) && eq64( ax.sig, bx.sig );
+
+}
+
+static flag floatXLe( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return TRUE;
+    if ( ax.sign != bx.sign ) return ax.sign;
+    if ( ax.sign ) {
+        if ( ax.isInf || bx.isZero ) return TRUE;
+        if ( bx.isInf || ax.isZero ) return FALSE;
+        if ( bx.exp < ax.exp ) return TRUE;
+        if ( ax.exp < bx.exp ) return FALSE;
+        return le64( bx.sig, ax.sig );
+    }
+    else {
+        if ( bx.isInf || ax.isZero ) return TRUE;
+        if ( ax.isInf || bx.isZero ) return FALSE;
+        if ( ax.exp < bx.exp ) return TRUE;
+        if ( bx.exp < ax.exp ) return FALSE;
+        return le64( ax.sig, bx.sig );
+    }
+
+}
+
+static flag floatXLt( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return FALSE;
+    if ( ax.sign != bx.sign ) return ax.sign;
+    if ( ax.isInf && bx.isInf ) return FALSE;
+    if ( ax.sign ) {
+        if ( ax.isInf || bx.isZero ) return TRUE;
+        if ( bx.isInf || ax.isZero ) return FALSE;
+        if ( bx.exp < ax.exp ) return TRUE;
+        if ( ax.exp < bx.exp ) return FALSE;
+        return lt64( bx.sig, ax.sig );
+    }
+    else {
+        if ( bx.isInf || ax.isZero ) return TRUE;
+        if ( ax.isInf || bx.isZero ) return FALSE;
+        if ( ax.exp < bx.exp ) return TRUE;
+        if ( bx.exp < ax.exp ) return FALSE;
+        return lt64( ax.sig, bx.sig );
+    }
+
+}
+
+float32 slow_int32_to_float32( int32 a )
+{
+
+    return floatXToFloat32( int32ToFloatX( a ) );
+
+}
+
+float64 slow_int32_to_float64( int32 a )
+{
+
+    return floatXToFloat64( int32ToFloatX( a ) );
+
+}
+
+int32 slow_float32_to_int32( float32 a )
+{
+
+    return floatXToInt32( float32ToFloatX( a ) );
+
+}
+
+int32 slow_float32_to_int32_round_to_zero( float32 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( float32ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float64 slow_float32_to_float64( float32 a )
+{
+
+    return floatXToFloat64( float32ToFloatX( a ) );
+
+}
+
+float32 slow_float32_round_to_int( float32 a )
+{
+
+    return floatXToFloat32( floatXRoundToInt( float32ToFloatX( a ) ) );
+
+}
+
+float32 slow_float32_add( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXAdd( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_sub( float32 a, float32 b )
+{
+
+    b ^= 0x80000000;
+    return
+        floatXToFloat32(
+            floatXAdd( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_mul( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXMul( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_div( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXDiv( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_rem( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXRem( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_sqrt( float32 a )
+{
+
+    return floatXToFloat32( floatXSqrt( float32ToFloatX( a ) ) );
+
+}
+
+flag slow_float32_eq( float32 a, float32 b )
+{
+
+    return floatXEq( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+flag slow_float32_le( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_float32_lt( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_float32_eq_signaling( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_float32_le_quiet( float32 a, float32 b )
+{
+
+    return floatXLe( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+flag slow_float32_lt_quiet( float32 a, float32 b )
+{
+
+    return floatXLt( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+int32 slow_float64_to_int32( float64 a )
+{
+
+    return floatXToInt32( float64ToFloatX( a ) );
+
+}
+
+int32 slow_float64_to_int32_round_to_zero( float64 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( float64ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float32 slow_float64_to_float32( float64 a )
+{
+
+    return floatXToFloat32( float64ToFloatX( a ) );
+
+}
+
+float64 slow_float64_round_to_int( float64 a )
+{
+
+    return floatXToFloat64( floatXRoundToInt( float64ToFloatX( a ) ) );
+
+}
+
+float64 slow_float64_add( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXAdd( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_sub( float64 a, float64 b )
+{
+
+#ifdef BITS64
+    b ^= LIT64( 0x8000000000000000 );
+#else
+    b.high ^= 0x80000000;
+#endif
+    return
+        floatXToFloat64(
+            floatXAdd( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_mul( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXMul( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_div( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXDiv( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_rem( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXRem( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_sqrt( float64 a )
+{
+
+    return floatXToFloat64( floatXSqrt( float64ToFloatX( a ) ) );
+
+}
+
+flag slow_float64_eq( float64 a, float64 b )
+{
+
+    return floatXEq( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
+flag slow_float64_le( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_float64_lt( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_float64_eq_signaling( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_float64_le_quiet( float64 a, float64 b )
+{
+
+    return floatXLe( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
+flag slow_float64_lt_quiet( float64 a, float64 b )
+{
+
+    return floatXLt( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
diff --git a/testfloat/slowfloat-64.c b/testfloat/slowfloat-64.c
new file mode 100644
index 00000000000..27e56e1152a
--- /dev/null
+++ b/testfloat/slowfloat-64.c
@@ -0,0 +1,2109 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+int8 slow_float_rounding_mode;
+int8 slow_float_exception_flags;
+int8 slow_float_detect_tininess;
+#ifdef FLOATX80
+int8 slow_floatx80_rounding_precision;
+#endif
+
+typedef struct {
+    bits64 a0, a1;
+} bits128X;
+
+typedef struct {
+    flag isNaN;
+    flag isInf;
+    flag isZero;
+    flag sign;
+    int32 exp;
+    bits128X sig;
+} floatX;
+
+static const floatX floatXNaN = { TRUE, FALSE, FALSE, FALSE, 0, { 0, 0 } };
+static const floatX floatXPositiveZero =
+    { FALSE, FALSE, TRUE, FALSE, 0, { 0, 0 } };
+static const floatX floatXNegativeZero =
+    { FALSE, FALSE, TRUE, TRUE, 0, { 0, 0 } };
+
+static bits128X shortShift128Left( bits128X a, int8 shiftCount )
+{
+    int8 negShiftCount;
+
+    negShiftCount = ( - shiftCount & 63 );
+    a.a0 = ( a.a0<<shiftCount ) | ( a.a1>>negShiftCount );
+    a.a1 <<= shiftCount;
+    return a;
+
+}
+
+static bits128X shortShift128RightJamming( bits128X a, int8 shiftCount )
+{
+    int8 negShiftCount;
+    bits64 extra;
+
+    negShiftCount = ( - shiftCount & 63 );
+    extra = a.a1<<negShiftCount;
+    a.a1 = ( a.a0<<negShiftCount ) | ( a.a1>>shiftCount ) | ( extra != 0 );
+    a.a0 >>= shiftCount;
+    return a;
+
+}
+
+static bits128X neg128( bits128X a )
+{
+
+    if ( a.a1 == 0 ) {
+        a.a0 = - a.a0;
+    }
+    else {
+        a.a1 = - a.a1;
+        a.a0 = ~ a.a0;
+    }
+    return a;
+
+}
+
+static bits128X add128( bits128X a, bits128X b )
+{
+
+    a.a1 += b.a1;
+    a.a0 += b.a0 + ( a.a1 < b.a1 );
+    return a;
+
+}
+
+static flag eq128( bits128X a, bits128X b )
+{
+
+    return ( a.a0 == b.a0 ) && ( a.a1 == b.a1 );
+
+}
+
+static flag le128( bits128X a, bits128X b )
+{
+
+    return ( a.a0 < b.a0 ) || ( ( a.a0 == b.a0 ) && ( a.a1 <= b.a1 ) );
+
+}
+
+static flag lt128( bits128X a, bits128X b )
+{
+
+    return ( a.a0 < b.a0 ) || ( ( a.a0 == b.a0 ) && ( a.a1 < b.a1 ) );
+
+}
+
+static floatX roundFloatXTo24( flag isTiny, floatX zx )
+{
+    bits32 roundBits;
+
+    zx.sig.a0 |= ( zx.sig.a1 != 0 );
+    zx.sig.a1 = 0;
+    roundBits = zx.sig.a0 & 0xFFFFFFFF;
+    zx.sig.a0 -= roundBits;
+    if ( roundBits ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( roundBits < 0x80000000 ) goto noIncrement;
+            if (    ( roundBits == 0x80000000 )
+                 && ! ( zx.sig.a0 & LIT64( 0x100000000 ) ) ) {
+                goto noIncrement;
+            }
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        zx.sig.a0 += LIT64( 0x100000000 );
+        if ( zx.sig.a0 == LIT64( 0x0100000000000000 ) ) {
+            zx.sig.a0 = LIT64( 0x0080000000000000 );
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    return zx;
+
+}
+
+static floatX roundFloatXTo53( flag isTiny, floatX zx )
+{
+    int8 roundBits;
+
+    zx.sig.a0 |= ( zx.sig.a1 != 0 );
+    zx.sig.a1 = 0;
+    roundBits = zx.sig.a0 & 7;
+    zx.sig.a0 -= roundBits;
+    if ( roundBits ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( roundBits < 4 ) goto noIncrement;
+            if ( ( roundBits == 4 ) && ! ( zx.sig.a0 & 8 ) ) goto noIncrement;
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        zx.sig.a0 += 8;
+        if ( zx.sig.a0 == LIT64( 0x0100000000000000 ) ) {
+            zx.sig.a0 = LIT64( 0x0080000000000000 );
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    return zx;
+
+}
+
+static floatX roundFloatXTo64( flag isTiny, floatX zx )
+{
+    int64 roundBits;
+
+    roundBits = zx.sig.a1 & LIT64( 0x00FFFFFFFFFFFFFF );
+    zx.sig.a1 -= roundBits;
+    if ( roundBits ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( roundBits < LIT64( 0x0080000000000000 ) ) goto noIncrement;
+            if (    ( roundBits == LIT64( 0x0080000000000000 ) )
+                 && ! ( zx.sig.a1 & LIT64( 0x0100000000000000 ) ) ) {
+                goto noIncrement;
+            }
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        zx.sig.a1 += LIT64( 0x0100000000000000 );
+        zx.sig.a0 += ( zx.sig.a1 == 0 );
+        if ( zx.sig.a0 == LIT64( 0x0100000000000000 ) ) {
+            zx.sig.a0 = LIT64( 0x0080000000000000 );
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    return zx;
+
+}
+
+static floatX roundFloatXTo113( flag isTiny, floatX zx )
+{
+    int8 roundBits;
+
+    roundBits = zx.sig.a1 & 0x7F;
+    zx.sig.a1 -= roundBits;
+    if ( roundBits ) {
+        slow_float_exception_flags |= float_flag_inexact;
+        if ( isTiny ) slow_float_exception_flags |= float_flag_underflow;
+        switch ( slow_float_rounding_mode ) {
+         case float_round_nearest_even:
+            if ( roundBits < 0x40 ) goto noIncrement;
+            if (    ( roundBits == 0x40 )
+                 && ! ( zx.sig.a1 & 0x80 ) ) goto noIncrement;
+            break;
+         case float_round_to_zero:
+            goto noIncrement;
+         case float_round_down:
+            if ( ! zx.sign ) goto noIncrement;
+            break;
+         case float_round_up:
+            if ( zx.sign ) goto noIncrement;
+            break;
+        }
+        zx.sig.a1 += 0x80;
+        zx.sig.a0 += ( zx.sig.a1 == 0 );
+        if ( zx.sig.a0 == LIT64( 0x0100000000000000 ) ) {
+            zx.sig.a0 = LIT64( 0x0080000000000000 );
+            ++zx.exp;
+        }
+    }
+ noIncrement:
+    return zx;
+
+}
+
+static floatX int32ToFloatX( int32 a )
+{
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.sign = ( a < 0 );
+    ax.sig.a1 = 0;
+    ax.sig.a0 = ax.sign ? - (bits64) a : a;
+    if ( a == 0 ) {
+        ax.isZero = TRUE;
+        return ax;
+    }
+    ax.isZero = FALSE;
+    ax.sig.a0 <<= 24;
+    ax.exp = 31;
+    while ( ax.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        ax.sig.a0 <<= 1;
+        --ax.exp;
+    }
+    return ax;
+
+}
+
+static int32 floatXToInt32( floatX ax )
+{
+    int8 savedExceptionFlags;
+    int32 shiftCount;
+    int32 z;
+
+    if ( ax.isInf || ax.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+        return ( ax.isInf & ax.sign ) ? (sbits32) 0x80000000 : 0x7FFFFFFF;
+    }
+    if ( ax.isZero ) return 0;
+    savedExceptionFlags = slow_float_exception_flags;
+    shiftCount = 52 - ax.exp;
+    if ( 56 < shiftCount ) {
+        ax.sig.a1 = 1;
+        ax.sig.a0 = 0;
+    }
+    else {
+        while ( 0 < shiftCount ) {
+            ax.sig = shortShift128RightJamming( ax.sig, 1 );
+            --shiftCount;
+        }
+    }
+    ax = roundFloatXTo53( FALSE, ax );
+    ax.sig = shortShift128RightJamming( ax.sig, 3 );
+    z = ax.sig.a0;
+    if ( ax.sign ) z = - z;
+    if (    ( shiftCount < 0 )
+         || ( ax.sig.a0>>32 )
+         || ( ( z != 0 ) && ( ( ax.sign ^ ( z < 0 ) ) != 0 ) )
+       ) {
+        slow_float_exception_flags = savedExceptionFlags | float_flag_invalid;
+        return ax.sign ? (sbits32) 0x80000000 : 0x7FFFFFFF;
+    }
+    return z;
+
+}
+
+static floatX int64ToFloatX( int64 a )
+{
+    uint64 absA;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.sign = ( a < 0 );
+    ax.sig.a1 = ax.sign ? - a : a;
+    ax.sig.a0 = 0;
+    if ( a == 0 ) {
+        ax.isZero = TRUE;
+        return ax;
+    }
+    ax.isZero = FALSE;
+    ax.sig = shortShift128Left( ax.sig, 56 );
+    ax.exp = 63;
+    while ( ax.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        ax.sig = shortShift128Left( ax.sig, 1 );
+        --ax.exp;
+    }
+    return ax;
+
+}
+
+static int64 floatXToInt64( floatX ax )
+{
+    int8 savedExceptionFlags;
+    int32 shiftCount;
+    int64 z;
+
+    if ( ax.isInf || ax.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+        return
+              ( ax.isInf & ax.sign ) ? (sbits64) LIT64( 0x8000000000000000 )
+            : LIT64( 0x7FFFFFFFFFFFFFFF );
+    }
+    if ( ax.isZero ) return 0;
+    savedExceptionFlags = slow_float_exception_flags;
+    shiftCount = 112 - ax.exp;
+    if ( 116 < shiftCount ) {
+        ax.sig.a1 = 1;
+        ax.sig.a0 = 0;
+    }
+    else {
+        while ( 0 < shiftCount ) {
+            ax.sig = shortShift128RightJamming( ax.sig, 1 );
+            --shiftCount;
+        }
+    }
+    ax = roundFloatXTo113( FALSE, ax );
+    ax.sig = shortShift128RightJamming( ax.sig, 7 );
+    z = ax.sig.a1;
+    if ( ax.sign ) z = - z;
+    if (    ( shiftCount < 0 )
+         || ax.sig.a0
+         || ( ( z != 0 ) && ( ( ax.sign ^ ( z < 0 ) ) != 0 ) )
+       ) {
+        slow_float_exception_flags = savedExceptionFlags | float_flag_invalid;
+        return
+              ax.sign ? (sbits64) LIT64( 0x8000000000000000 )
+            : LIT64( 0x7FFFFFFFFFFFFFFF );
+    }
+    return z;
+
+}
+
+static floatX float32ToFloatX( float32 a )
+{
+    int16 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+    ax.sign = ( ( a & 0x80000000 ) != 0 );
+    expField = ( a>>23 ) & 0xFF;
+    ax.sig.a1 = 0;
+    ax.sig.a0 = a & 0x007FFFFF;
+    ax.sig.a0 <<= 32;
+    if ( expField == 0 ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x7F;
+            do {
+                ax.sig.a0 <<= 1;
+                --expField;
+            } while ( ax.sig.a0 < LIT64( 0x0080000000000000 ) );
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0xFF ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.sig.a0 |= LIT64( 0x0080000000000000 );
+        ax.exp = expField - 0x7F;
+    }
+    return ax;
+
+}
+
+static float32 floatXToFloat32( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int32 expField;
+    float32 z;
+
+    if ( zx.isZero ) return zx.sign ? 0x80000000 : 0;
+    if ( zx.isInf ) return zx.sign ? 0xFF800000 : 0x7F800000;
+    if ( zx.isNaN ) return 0xFFFFFFFF;
+    while ( LIT64( 0x0100000000000000 ) <= zx.sig.a0 ) {
+        zx.sig = shortShift128RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        zx.sig = shortShift128Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x7F <= 0 );
+    zx = roundFloatXTo24( isTiny, zx );
+    expField = zx.exp + 0x7F;
+    if ( 0xFF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z = 0xFF800000;
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z = 0xFF7FFFFF;
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z = 0x7F800000;
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z = 0x7F7FFFFF;
+                break;
+            }
+        }
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x7F;
+        if ( expField < -27 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        zx = roundFloatXTo24( isTiny, zx );
+        expField = ( LIT64( 0x0080000000000000 ) <= zx.sig.a0 ) ? 1 : 0;
+    }
+    z = expField;
+    z <<= 23;
+    if ( zx.sign ) z |= 0x80000000;
+    z |= ( zx.sig.a0>>32 ) & 0x007FFFFF;
+    return z;
+
+}
+
+static floatX float64ToFloatX( float64 a )
+{
+    int16 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+    ax.sign = ( ( a & LIT64( 0x8000000000000000 ) ) != 0 );
+    expField = ( a>>52 ) & 0x7FF;
+    ax.sig.a1 = 0;
+    ax.sig.a0 = a & LIT64( 0x000FFFFFFFFFFFFF );
+    if ( expField == 0 ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x3FF;
+            do {
+                ax.sig.a0 <<= 1;
+                --expField;
+            } while ( ax.sig.a0 < LIT64( 0x0010000000000000 ) );
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0x7FF ) {
+        if ( ax.sig.a0 == 0 ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.exp = expField - 0x3FF;
+        ax.sig.a0 |= LIT64( 0x0010000000000000 );
+    }
+    ax.sig.a0 <<= 3;
+    return ax;
+
+}
+
+static float64 floatXToFloat64( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int32 expField;
+    float64 z;
+
+    if ( zx.isZero ) return zx.sign ? LIT64( 0x8000000000000000 ) : 0;
+    if ( zx.isInf ) {
+        return
+              zx.sign ? LIT64( 0xFFF0000000000000 )
+            : LIT64( 0x7FF0000000000000 );
+    }
+    if ( zx.isNaN ) return LIT64( 0xFFFFFFFFFFFFFFFF );
+    while ( LIT64( 0x0100000000000000 ) <= zx.sig.a0 ) {
+        zx.sig = shortShift128RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        zx.sig = shortShift128Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x3FF <= 0 );
+    zx = roundFloatXTo53( isTiny, zx );
+    expField = zx.exp + 0x3FF;
+    if ( 0x7FF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z = LIT64( 0xFFF0000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z = LIT64( 0xFFEFFFFFFFFFFFFF );
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z = LIT64( 0x7FF0000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z = LIT64( 0x7FEFFFFFFFFFFFFF );
+                break;
+            }
+        }
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x3FF;
+        if ( expField < -56 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        zx = roundFloatXTo53( isTiny, zx );
+        expField = ( LIT64( 0x0080000000000000 ) <= zx.sig.a0 ) ? 1 : 0;
+    }
+    zx.sig.a0 >>= 3;
+    z = expField;
+    z <<= 52;
+    if ( zx.sign ) z |= LIT64( 0x8000000000000000 );
+    z |= zx.sig.a0 & LIT64( 0x000FFFFFFFFFFFFF );
+    return z;
+
+}
+
+#ifdef FLOATX80
+
+static floatX floatx80ToFloatX( floatx80 a )
+{
+    int32 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+    ax.sign = ( ( a.high & 0x8000 ) != 0 );
+    expField = a.high & 0x7FFF;
+    ax.sig.a1 = a.low;
+    ax.sig.a0 = 0;
+    if ( expField == 0 ) {
+        if ( ax.sig.a1 == 0 ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x3FFF;
+            while ( ax.sig.a1 < LIT64( 0x8000000000000000 ) ) {
+                ax.sig.a1 <<= 1;
+                --expField;
+            }
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0x7FFF ) {
+        if ( ( ax.sig.a1 & LIT64( 0x7FFFFFFFFFFFFFFF ) ) == 0 ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.exp = expField - 0x3FFF;
+    }
+    ax.sig = shortShift128Left( ax.sig, 56 );
+    return ax;
+
+}
+
+static floatx80 floatXToFloatx80( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int32 expField;
+    floatx80 z;
+
+    if ( zx.isZero ) {
+        z.low = 0;
+        z.high = zx.sign ? 0x8000 : 0;
+        return z;
+    }
+    if ( zx.isInf ) {
+        z.low = LIT64( 0x8000000000000000 );
+        z.high = zx.sign ? 0xFFFF : 0x7FFF;
+        return z;
+    }
+    if ( zx.isNaN ) {
+        z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+        z.high = 0xFFFF;
+        return z;
+    }
+    while ( LIT64( 0x0100000000000000 ) <= zx.sig.a0 ) {
+        zx.sig = shortShift128RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        zx.sig = shortShift128Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x3FFF <= 0 );
+    switch ( slow_floatx80_rounding_precision ) {
+     case 32:
+        zx = roundFloatXTo24( isTiny, zx );
+        break;
+     case 64:
+        zx = roundFloatXTo53( isTiny, zx );
+        break;
+     default:
+        zx = roundFloatXTo64( isTiny, zx );
+        break;
+    }
+    expField = zx.exp + 0x3FFF;
+    if ( 0x7FFF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z.low = LIT64( 0x8000000000000000 );
+                z.high = 0xFFFF;
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                switch ( slow_floatx80_rounding_precision ) {
+                 case 32:
+                    z.low = LIT64( 0xFFFFFF0000000000 );
+                    break;
+                 case 64:
+                    z.low = LIT64( 0xFFFFFFFFFFFFF800 );
+                    break;
+                 default:
+                    z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+                    break;
+                }
+                z.high = 0xFFFE;
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z.low = LIT64( 0x8000000000000000 );
+                z.high = 0x7FFF;
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                switch ( slow_floatx80_rounding_precision ) {
+                 case 32:
+                    z.low = LIT64( 0xFFFFFF0000000000 );
+                    break;
+                 case 64:
+                    z.low = LIT64( 0xFFFFFFFFFFFFF800 );
+                    break;
+                 default:
+                    z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+                    break;
+                }
+                z.high = 0x7FFE;
+                break;
+            }
+        }
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x3FFF;
+        if ( expField < -70 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        switch ( slow_floatx80_rounding_precision ) {
+         case 32:
+            zx = roundFloatXTo24( isTiny, zx );
+            break;
+         case 64:
+            zx = roundFloatXTo53( isTiny, zx );
+            break;
+         default:
+            zx = roundFloatXTo64( isTiny, zx );
+            break;
+        }
+        expField = ( LIT64( 0x0080000000000000 ) <= zx.sig.a0 ) ? 1 : 0;
+    }
+    zx.sig = shortShift128RightJamming( zx.sig, 56 );
+    z.low = zx.sig.a1;
+    z.high = expField;
+    if ( zx.sign ) z.high |= 0x8000;
+    return z;
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+static floatX float128ToFloatX( float128 a )
+{
+    int32 expField;
+    floatX ax;
+
+    ax.isNaN = FALSE;
+    ax.isInf = FALSE;
+    ax.isZero = FALSE;
+    ax.sign = ( ( a.high & LIT64( 0x8000000000000000 ) ) != 0 );
+    expField = ( a.high>>48 ) & 0x7FFF;
+    ax.sig.a1 = a.low;
+    ax.sig.a0 = a.high & LIT64( 0x0000FFFFFFFFFFFF );
+    if ( expField == 0 ) {
+        if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) {
+            ax.isZero = TRUE;
+        }
+        else {
+            expField = 1 - 0x3FFF;
+            do {
+                ax.sig = shortShift128Left( ax.sig, 1 );
+                --expField;
+            } while ( ax.sig.a0 < LIT64( 0x0001000000000000 ) );
+            ax.exp = expField;
+        }
+    }
+    else if ( expField == 0x7FFF ) {
+        if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) {
+            ax.isInf = TRUE;
+        }
+        else {
+            ax.isNaN = TRUE;
+        }
+    }
+    else {
+        ax.exp = expField - 0x3FFF;
+        ax.sig.a0 |= LIT64( 0x0001000000000000 );
+    }
+    ax.sig = shortShift128Left( ax.sig, 7 );
+    return ax;
+
+}
+
+static float128 floatXToFloat128( floatX zx )
+{
+    floatX savedZ;
+    flag isTiny;
+    int32 expField;
+    float128 z;
+
+    if ( zx.isZero ) {
+        z.low = 0;
+        z.high = zx.sign ? LIT64( 0x8000000000000000 ) : 0;
+        return z;
+    }
+    if ( zx.isInf ) {
+        z.low = 0;
+        z.high =
+              zx.sign ? LIT64( 0xFFFF000000000000 )
+            : LIT64( 0x7FFF000000000000 );
+        return z;
+    }
+    if ( zx.isNaN ) {
+        z.high = z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+        return z;
+    }
+    while ( LIT64( 0x0100000000000000 ) <= zx.sig.a0 ) {
+        zx.sig = shortShift128RightJamming( zx.sig, 1 );
+        ++zx.exp;
+    }
+    while ( zx.sig.a0 < LIT64( 0x0080000000000000 ) ) {
+        zx.sig = shortShift128Left( zx.sig, 1 );
+        --zx.exp;
+    }
+    savedZ = zx;
+    isTiny =
+           ( slow_float_detect_tininess == float_tininess_before_rounding )
+        && ( zx.exp + 0x3FFF <= 0 );
+    zx = roundFloatXTo113( isTiny, zx );
+    expField = zx.exp + 0x3FFF;
+    if ( 0x7FFF <= expField ) {
+        slow_float_exception_flags |=
+            float_flag_overflow | float_flag_inexact;
+        if ( zx.sign ) {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_down:
+                z.low = 0;
+                z.high = LIT64( 0xFFFF000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_up:
+                z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+                z.high = LIT64( 0xFFFEFFFFFFFFFFFF );
+                break;
+            }
+        }
+        else {
+            switch ( slow_float_rounding_mode ) {
+             case float_round_nearest_even:
+             case float_round_up:
+                z.low = 0;
+                z.high = LIT64( 0x7FFF000000000000 );
+                break;
+             case float_round_to_zero:
+             case float_round_down:
+                z.low = LIT64( 0xFFFFFFFFFFFFFFFF );
+                z.high = LIT64( 0x7FFEFFFFFFFFFFFF );
+                break;
+            }
+        }
+        return z;
+    }
+    if ( expField <= 0 ) {
+        isTiny = TRUE;
+        zx = savedZ;
+        expField = zx.exp + 0x3FFF;
+        if ( expField < -120 ) {
+            zx.sig.a1 = ( zx.sig.a0 != 0 ) || ( zx.sig.a1 != 0 );
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expField <= 0 ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                ++expField;
+            }
+        }
+        zx = roundFloatXTo113( isTiny, zx );
+        expField = ( LIT64( 0x0080000000000000 ) <= zx.sig.a0 ) ? 1 : 0;
+    }
+    zx.sig = shortShift128RightJamming( zx.sig, 7 );
+    z.low = zx.sig.a1;
+    z.high = expField;
+    z.high <<= 48;
+    if ( zx.sign ) z.high |= LIT64( 0x8000000000000000 );
+    z.high |= zx.sig.a0 & LIT64( 0x0000FFFFFFFFFFFF );
+    return z;
+
+}
+
+#endif
+
+static floatX floatXInvalid( void )
+{
+
+    slow_float_exception_flags |= float_flag_invalid;
+    return floatXNaN;
+
+}
+
+static floatX floatXRoundToInt( floatX ax )
+{
+    int32 shiftCount, i;
+
+    if ( ax.isNaN || ax.isInf ) return ax;
+    shiftCount = 112 - ax.exp;
+    if ( shiftCount <= 0 ) return ax;
+    if ( 119 < shiftCount ) {
+        ax.exp = 112;
+        ax.sig.a1 = ! ax.isZero;
+        ax.sig.a0 = 0;
+    }
+    else {
+        while ( 0 < shiftCount ) {
+            ax.sig = shortShift128RightJamming( ax.sig, 1 );
+            ++ax.exp;
+            --shiftCount;
+        }
+    }
+    ax = roundFloatXTo113( FALSE, ax );
+    if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) ax.isZero = TRUE;
+    return ax;
+
+}
+
+static floatX floatXAdd( floatX ax, floatX bx )
+{
+    int32 expDiff;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf && bx.isInf ) {
+        if ( ax.sign == bx.sign ) return ax;
+        return floatXInvalid();
+    }
+    if ( ax.isInf ) return ax;
+    if ( bx.isInf ) return bx;
+    if ( ax.isZero && bx.isZero ) {
+        if ( ax.sign == bx.sign ) return ax;
+        goto completeCancellation;
+    }
+    if (    ( ax.sign != bx.sign )
+         && ( ax.exp == bx.exp )
+         && eq128( ax.sig, bx.sig )
+       ) {
+ completeCancellation:
+        return
+              ( slow_float_rounding_mode == float_round_down ) ?
+                  floatXNegativeZero
+            : floatXPositiveZero;
+    }
+    if ( ax.isZero ) return bx;
+    if ( bx.isZero ) return ax;
+    expDiff = ax.exp - bx.exp;
+    if ( expDiff < 0 ) {
+        zx = ax;
+        zx.exp = bx.exp;
+        if ( expDiff < -120 ) {
+            zx.sig.a1 = 1;
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( expDiff < 0 ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                ++expDiff;
+            }
+        }
+        if ( ax.sign != bx.sign ) zx.sig = neg128( zx.sig );
+        zx.sign = bx.sign;
+        zx.sig = add128( zx.sig, bx.sig );
+    }
+    else {
+        zx = bx;
+        zx.exp = ax.exp;
+        if ( 120 < expDiff ) {
+            zx.sig.a1 = 1;
+            zx.sig.a0 = 0;
+        }
+        else {
+            while ( 0 < expDiff ) {
+                zx.sig = shortShift128RightJamming( zx.sig, 1 );
+                --expDiff;
+            }
+        }
+        if ( ax.sign != bx.sign ) zx.sig = neg128( zx.sig );
+        zx.sign = ax.sign;
+        zx.sig = add128( zx.sig, ax.sig );
+    }
+    if ( zx.sig.a0 & LIT64( 0x8000000000000000 ) ) {
+        zx.sig = neg128( zx.sig );
+        zx.sign = ! zx.sign;
+    }
+    return zx;
+
+}
+
+static floatX floatXMul( floatX ax, floatX bx )
+{
+    int8 bitNum;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf ) {
+        if ( bx.isZero ) return floatXInvalid();
+        if ( bx.sign ) ax.sign = ! ax.sign;
+        return ax;
+    }
+    if ( bx.isInf ) {
+        if ( ax.isZero ) return floatXInvalid();
+        if ( ax.sign ) bx.sign = ! bx.sign;
+        return bx;
+    }
+    zx = ax;
+    zx.sign ^= bx.sign;
+    if ( ax.isZero || bx.isZero ) {
+        return zx.sign ? floatXNegativeZero : floatXPositiveZero;
+    }
+    zx.exp += bx.exp + 1;
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    for ( bitNum = 0; bitNum < 119; ++bitNum ) {
+        if ( bx.sig.a1 & 2 ) zx.sig = add128( zx.sig, ax.sig );
+        bx.sig = shortShift128RightJamming( bx.sig, 1 );
+        zx.sig = shortShift128RightJamming( zx.sig, 1 );
+    }
+    return zx;
+
+}
+
+static floatX floatXDiv( floatX ax, floatX bx )
+{
+    bits128X negBSig;
+    int8 bitNum;
+    floatX zx;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf ) {
+        if ( bx.isInf ) return floatXInvalid();
+        if ( bx.sign ) ax.sign = ! ax.sign;
+        return ax;
+    }
+    if ( bx.isZero ) {
+        if ( ax.isZero ) return floatXInvalid();
+        slow_float_exception_flags |= float_flag_divbyzero;
+        if ( ax.sign ) bx.sign = ! bx.sign;
+        bx.isZero = FALSE;
+        bx.isInf = TRUE;
+        return bx;
+    }
+    zx = ax;
+    zx.sign ^= bx.sign;
+    if ( ax.isZero || bx.isInf ) {
+        return zx.sign ? floatXNegativeZero : floatXPositiveZero;
+    }
+    zx.exp -= bx.exp + 1;
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    negBSig = neg128( bx.sig );
+    for ( bitNum = 0; bitNum < 120; ++bitNum ) {
+        if ( le128( bx.sig, ax.sig ) ) {
+            zx.sig.a1 |= 1;
+            ax.sig = add128( ax.sig, negBSig );
+        }
+        ax.sig = shortShift128Left( ax.sig, 1 );
+        zx.sig = shortShift128Left( zx.sig, 1 );
+    }
+    if ( ax.sig.a0 || ax.sig.a1 ) zx.sig.a1 |= 1;
+    return zx;
+
+}
+
+static floatX floatXRem( floatX ax, floatX bx )
+{
+    bits128X negBSig;
+    flag lastQuotientBit;
+    bits128X savedASig;
+
+    if ( ax.isNaN ) return ax;
+    if ( bx.isNaN ) return bx;
+    if ( ax.isInf || bx.isZero ) return floatXInvalid();
+    if ( ax.isZero || bx.isInf ) return ax;
+    --bx.exp;
+    if ( ax.exp < bx.exp ) return ax;
+    bx.sig = shortShift128Left( bx.sig, 1 );
+    negBSig = neg128( bx.sig );
+    while ( bx.exp < ax.exp ) {
+        if ( le128( bx.sig, ax.sig ) ) ax.sig = add128( ax.sig, negBSig );
+        ax.sig = shortShift128Left( ax.sig, 1 );
+        --ax.exp;
+    }
+    lastQuotientBit = le128( bx.sig, ax.sig );
+    if ( lastQuotientBit ) ax.sig = add128( ax.sig, negBSig );
+    savedASig = ax.sig;
+    ax.sig = neg128( add128( ax.sig, negBSig ) );
+    if ( lt128( ax.sig, savedASig ) ) {
+        ax.sign = ! ax.sign;
+    }
+    else if ( lt128( savedASig, ax.sig ) ) {
+        ax.sig = savedASig;
+    }
+    else {
+        if ( lastQuotientBit ) {
+            ax.sign = ! ax.sign;
+        }
+        else {
+            ax.sig = savedASig;
+        }
+    }
+    if ( ( ax.sig.a0 == 0 ) && ( ax.sig.a1 == 0 ) ) ax.isZero = TRUE;
+    return ax;
+
+}
+
+static floatX floatXSqrt( floatX ax )
+{
+    int8 bitNum;
+    bits128X bitSig, savedASig;
+    floatX zx;
+
+    if ( ax.isNaN || ax.isZero ) return ax;
+    if ( ax.sign ) return floatXInvalid();
+    if ( ax.isInf ) return ax;
+    zx = ax;
+    zx.exp >>= 1;
+    if ( ( ax.exp & 1 ) == 0 ) ax.sig = shortShift128RightJamming( ax.sig, 1 );
+    zx.sig.a1 = 0;
+    zx.sig.a0 = 0;
+    bitSig.a1 = 0;
+    bitSig.a0 = LIT64( 0x0080000000000000 );
+    for ( bitNum = 0; bitNum < 120; ++bitNum ) {
+        savedASig = ax.sig;
+        ax.sig = add128( ax.sig, neg128( zx.sig ) );
+        ax.sig = shortShift128Left( ax.sig, 1 );
+        ax.sig = add128( ax.sig, neg128( bitSig ) );
+        if ( ax.sig.a0 & LIT64( 0x8000000000000000 ) ) {
+            ax.sig = shortShift128Left( savedASig, 1 );
+        }
+        else {
+            zx.sig.a1 |= bitSig.a1;
+            zx.sig.a0 |= bitSig.a0;
+        }
+        bitSig = shortShift128RightJamming( bitSig, 1 );
+    }
+    if ( ax.sig.a0 || ax.sig.a1 ) zx.sig.a1 |= 1;
+    return zx;
+
+}
+
+static flag floatXEq( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return TRUE;
+    if ( ax.sign != bx.sign ) return FALSE;
+    if ( ax.isInf || bx.isInf ) return ax.isInf && bx.isInf;
+    return ( ax.exp == bx.exp ) && eq128( ax.sig, bx.sig );
+
+}
+
+static flag floatXLe( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return TRUE;
+    if ( ax.sign != bx.sign ) return ax.sign;
+    if ( ax.sign ) {
+        if ( ax.isInf || bx.isZero ) return TRUE;
+        if ( bx.isInf || ax.isZero ) return FALSE;
+        if ( bx.exp < ax.exp ) return TRUE;
+        if ( ax.exp < bx.exp ) return FALSE;
+        return le128( bx.sig, ax.sig );
+    }
+    else {
+        if ( bx.isInf || ax.isZero ) return TRUE;
+        if ( ax.isInf || bx.isZero ) return FALSE;
+        if ( ax.exp < bx.exp ) return TRUE;
+        if ( bx.exp < ax.exp ) return FALSE;
+        return le128( ax.sig, bx.sig );
+    }
+
+}
+
+static flag floatXLt( floatX ax, floatX bx )
+{
+
+    if ( ax.isNaN || bx.isNaN ) return FALSE;
+    if ( ax.isZero && bx.isZero ) return FALSE;
+    if ( ax.sign != bx.sign ) return ax.sign;
+    if ( ax.isInf && bx.isInf ) return FALSE;
+    if ( ax.sign ) {
+        if ( ax.isInf || bx.isZero ) return TRUE;
+        if ( bx.isInf || ax.isZero ) return FALSE;
+        if ( bx.exp < ax.exp ) return TRUE;
+        if ( ax.exp < bx.exp ) return FALSE;
+        return lt128( bx.sig, ax.sig );
+    }
+    else {
+        if ( bx.isInf || ax.isZero ) return TRUE;
+        if ( ax.isInf || bx.isZero ) return FALSE;
+        if ( ax.exp < bx.exp ) return TRUE;
+        if ( bx.exp < ax.exp ) return FALSE;
+        return lt128( ax.sig, bx.sig );
+    }
+
+}
+
+float32 slow_int32_to_float32( int32 a )
+{
+
+    return floatXToFloat32( int32ToFloatX( a ) );
+
+}
+
+float64 slow_int32_to_float64( int32 a )
+{
+
+    return floatXToFloat64( int32ToFloatX( a ) );
+
+}
+
+#ifdef FLOATX80
+
+floatx80 slow_int32_to_floatx80( int32 a )
+{
+
+    return floatXToFloatx80( int32ToFloatX( a ) );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+float128 slow_int32_to_float128( int32 a )
+{
+
+    return floatXToFloat128( int32ToFloatX( a ) );
+
+}
+
+#endif
+
+float32 slow_int64_to_float32( int64 a )
+{
+
+    return floatXToFloat32( int64ToFloatX( a ) );
+
+}
+
+float64 slow_int64_to_float64( int64 a )
+{
+
+    return floatXToFloat64( int64ToFloatX( a ) );
+
+}
+
+#ifdef FLOATX80
+
+floatx80 slow_int64_to_floatx80( int64 a )
+{
+
+    return floatXToFloatx80( int64ToFloatX( a ) );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+float128 slow_int64_to_float128( int64 a )
+{
+
+    return floatXToFloat128( int64ToFloatX( a ) );
+
+}
+
+#endif
+
+int32 slow_float32_to_int32( float32 a )
+{
+
+    return floatXToInt32( float32ToFloatX( a ) );
+
+}
+
+int32 slow_float32_to_int32_round_to_zero( float32 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( float32ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+int64 slow_float32_to_int64( float32 a )
+{
+
+    return floatXToInt64( float32ToFloatX( a ) );
+
+}
+
+int64 slow_float32_to_int64_round_to_zero( float32 a )
+{
+    int8 savedRoundingMode;
+    int64 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt64( float32ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float64 slow_float32_to_float64( float32 a )
+{
+
+    return floatXToFloat64( float32ToFloatX( a ) );
+
+}
+
+#ifdef FLOATX80
+
+floatx80 slow_float32_to_floatx80( float32 a )
+{
+
+    return floatXToFloatx80( float32ToFloatX( a ) );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+float128 slow_float32_to_float128( float32 a )
+{
+
+    return floatXToFloat128( float32ToFloatX( a ) );
+
+}
+
+#endif
+
+float32 slow_float32_round_to_int( float32 a )
+{
+
+    return floatXToFloat32( floatXRoundToInt( float32ToFloatX( a ) ) );
+
+}
+
+float32 slow_float32_add( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXAdd( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_sub( float32 a, float32 b )
+{
+
+    b ^= 0x80000000;
+    return
+        floatXToFloat32(
+            floatXAdd( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_mul( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXMul( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_div( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXDiv( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_rem( float32 a, float32 b )
+{
+
+    return
+        floatXToFloat32(
+            floatXRem( float32ToFloatX( a ), float32ToFloatX( b ) ) );
+
+}
+
+float32 slow_float32_sqrt( float32 a )
+{
+
+    return floatXToFloat32( floatXSqrt( float32ToFloatX( a ) ) );
+
+}
+
+flag slow_float32_eq( float32 a, float32 b )
+{
+
+    return floatXEq( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+flag slow_float32_le( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_float32_lt( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_float32_eq_signaling( float32 a, float32 b )
+{
+    floatX ax, bx;
+
+    ax = float32ToFloatX( a );
+    bx = float32ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_float32_le_quiet( float32 a, float32 b )
+{
+
+    return floatXLe( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+flag slow_float32_lt_quiet( float32 a, float32 b )
+{
+
+    return floatXLt( float32ToFloatX( a ), float32ToFloatX( b ) );
+
+}
+
+int32 slow_float64_to_int32( float64 a )
+{
+
+    return floatXToInt32( float64ToFloatX( a ) );
+
+}
+
+int32 slow_float64_to_int32_round_to_zero( float64 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( float64ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+int64 slow_float64_to_int64( float64 a )
+{
+
+    return floatXToInt64( float64ToFloatX( a ) );
+
+}
+
+int64 slow_float64_to_int64_round_to_zero( float64 a )
+{
+    int8 savedRoundingMode;
+    int64 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt64( float64ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float32 slow_float64_to_float32( float64 a )
+{
+
+    return floatXToFloat32( float64ToFloatX( a ) );
+
+}
+
+#ifdef FLOATX80
+
+floatx80 slow_float64_to_floatx80( float64 a )
+{
+
+    return floatXToFloatx80( float64ToFloatX( a ) );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+float128 slow_float64_to_float128( float64 a )
+{
+
+    return floatXToFloat128( float64ToFloatX( a ) );
+
+}
+
+#endif
+
+float64 slow_float64_round_to_int( float64 a )
+{
+
+    return floatXToFloat64( floatXRoundToInt( float64ToFloatX( a ) ) );
+
+}
+
+float64 slow_float64_add( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXAdd( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_sub( float64 a, float64 b )
+{
+
+    b ^= LIT64( 0x8000000000000000 );
+    return
+        floatXToFloat64(
+            floatXAdd( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_mul( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXMul( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_div( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXDiv( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_rem( float64 a, float64 b )
+{
+
+    return
+        floatXToFloat64(
+            floatXRem( float64ToFloatX( a ), float64ToFloatX( b ) ) );
+
+}
+
+float64 slow_float64_sqrt( float64 a )
+{
+
+    return floatXToFloat64( floatXSqrt( float64ToFloatX( a ) ) );
+
+}
+
+flag slow_float64_eq( float64 a, float64 b )
+{
+
+    return floatXEq( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
+flag slow_float64_le( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_float64_lt( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_float64_eq_signaling( float64 a, float64 b )
+{
+    floatX ax, bx;
+
+    ax = float64ToFloatX( a );
+    bx = float64ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_float64_le_quiet( float64 a, float64 b )
+{
+
+    return floatXLe( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
+flag slow_float64_lt_quiet( float64 a, float64 b )
+{
+
+    return floatXLt( float64ToFloatX( a ), float64ToFloatX( b ) );
+
+}
+
+#ifdef FLOATX80
+
+int32 slow_floatx80_to_int32( floatx80 a )
+{
+
+    return floatXToInt32( floatx80ToFloatX( a ) );
+
+}
+
+int32 slow_floatx80_to_int32_round_to_zero( floatx80 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( floatx80ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+int64 slow_floatx80_to_int64( floatx80 a )
+{
+
+    return floatXToInt64( floatx80ToFloatX( a ) );
+
+}
+
+int64 slow_floatx80_to_int64_round_to_zero( floatx80 a )
+{
+    int8 savedRoundingMode;
+    int64 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt64( floatx80ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float32 slow_floatx80_to_float32( floatx80 a )
+{
+
+    return floatXToFloat32( floatx80ToFloatX( a ) );
+
+}
+
+float64 slow_floatx80_to_float64( floatx80 a )
+{
+
+    return floatXToFloat64( floatx80ToFloatX( a ) );
+
+}
+
+#ifdef FLOAT128
+
+float128 slow_floatx80_to_float128( floatx80 a )
+{
+
+    return floatXToFloat128( floatx80ToFloatX( a ) );
+
+}
+
+#endif
+
+floatx80 slow_floatx80_round_to_int( floatx80 a )
+{
+
+    return floatXToFloatx80( floatXRoundToInt( floatx80ToFloatX( a ) ) );
+
+}
+
+floatx80 slow_floatx80_add( floatx80 a, floatx80 b )
+{
+
+    return
+        floatXToFloatx80(
+            floatXAdd( floatx80ToFloatX( a ), floatx80ToFloatX( b ) ) );
+
+}
+
+floatx80 slow_floatx80_sub( floatx80 a, floatx80 b )
+{
+
+    b.high ^= 0x8000;
+    return
+        floatXToFloatx80(
+            floatXAdd( floatx80ToFloatX( a ), floatx80ToFloatX( b ) ) );
+
+}
+
+floatx80 slow_floatx80_mul( floatx80 a, floatx80 b )
+{
+
+    return
+        floatXToFloatx80(
+            floatXMul( floatx80ToFloatX( a ), floatx80ToFloatX( b ) ) );
+
+}
+
+floatx80 slow_floatx80_div( floatx80 a, floatx80 b )
+{
+
+    return
+        floatXToFloatx80(
+            floatXDiv( floatx80ToFloatX( a ), floatx80ToFloatX( b ) ) );
+
+}
+
+floatx80 slow_floatx80_rem( floatx80 a, floatx80 b )
+{
+
+    return
+        floatXToFloatx80(
+            floatXRem( floatx80ToFloatX( a ), floatx80ToFloatX( b ) ) );
+
+}
+
+floatx80 slow_floatx80_sqrt( floatx80 a )
+{
+
+    return floatXToFloatx80( floatXSqrt( floatx80ToFloatX( a ) ) );
+
+}
+
+flag slow_floatx80_eq( floatx80 a, floatx80 b )
+{
+
+    return floatXEq( floatx80ToFloatX( a ), floatx80ToFloatX( b ) );
+
+}
+
+flag slow_floatx80_le( floatx80 a, floatx80 b )
+{
+    floatX ax, bx;
+
+    ax = floatx80ToFloatX( a );
+    bx = floatx80ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_floatx80_lt( floatx80 a, floatx80 b )
+{
+    floatX ax, bx;
+
+    ax = floatx80ToFloatX( a );
+    bx = floatx80ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_floatx80_eq_signaling( floatx80 a, floatx80 b )
+{
+    floatX ax, bx;
+
+    ax = floatx80ToFloatX( a );
+    bx = floatx80ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_floatx80_le_quiet( floatx80 a, floatx80 b )
+{
+
+    return floatXLe( floatx80ToFloatX( a ), floatx80ToFloatX( b ) );
+
+}
+
+flag slow_floatx80_lt_quiet( floatx80 a, floatx80 b )
+{
+
+    return floatXLt( floatx80ToFloatX( a ), floatx80ToFloatX( b ) );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+int32 slow_float128_to_int32( float128 a )
+{
+
+    return floatXToInt32( float128ToFloatX( a ) );
+
+}
+
+int32 slow_float128_to_int32_round_to_zero( float128 a )
+{
+    int8 savedRoundingMode;
+    int32 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt32( float128ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+int64 slow_float128_to_int64( float128 a )
+{
+
+    return floatXToInt64( float128ToFloatX( a ) );
+
+}
+
+int64 slow_float128_to_int64_round_to_zero( float128 a )
+{
+    int8 savedRoundingMode;
+    int64 z;
+
+    savedRoundingMode = slow_float_rounding_mode;
+    slow_float_rounding_mode = float_round_to_zero;
+    z = floatXToInt64( float128ToFloatX( a ) );
+    slow_float_rounding_mode = savedRoundingMode;
+    return z;
+
+}
+
+float32 slow_float128_to_float32( float128 a )
+{
+
+    return floatXToFloat32( float128ToFloatX( a ) );
+
+}
+
+float64 slow_float128_to_float64( float128 a )
+{
+
+    return floatXToFloat64( float128ToFloatX( a ) );
+
+}
+
+#ifdef FLOATX80
+
+floatx80 slow_float128_to_floatx80( float128 a )
+{
+
+    return floatXToFloatx80( float128ToFloatX( a ) );
+
+}
+
+#endif
+
+float128 slow_float128_round_to_int( float128 a )
+{
+
+    return floatXToFloat128( floatXRoundToInt( float128ToFloatX( a ) ) );
+
+}
+
+float128 slow_float128_add( float128 a, float128 b )
+{
+
+    return
+        floatXToFloat128(
+            floatXAdd( float128ToFloatX( a ), float128ToFloatX( b ) ) );
+
+}
+
+float128 slow_float128_sub( float128 a, float128 b )
+{
+
+    b.high ^= LIT64( 0x8000000000000000 );
+    return
+        floatXToFloat128(
+            floatXAdd( float128ToFloatX( a ), float128ToFloatX( b ) ) );
+
+}
+
+float128 slow_float128_mul( float128 a, float128 b )
+{
+
+    return
+        floatXToFloat128(
+            floatXMul( float128ToFloatX( a ), float128ToFloatX( b ) ) );
+
+}
+
+float128 slow_float128_div( float128 a, float128 b )
+{
+
+    return
+        floatXToFloat128(
+            floatXDiv( float128ToFloatX( a ), float128ToFloatX( b ) ) );
+
+}
+
+float128 slow_float128_rem( float128 a, float128 b )
+{
+
+    return
+        floatXToFloat128(
+            floatXRem( float128ToFloatX( a ), float128ToFloatX( b ) ) );
+
+}
+
+float128 slow_float128_sqrt( float128 a )
+{
+
+    return floatXToFloat128( floatXSqrt( float128ToFloatX( a ) ) );
+
+}
+
+flag slow_float128_eq( float128 a, float128 b )
+{
+
+    return floatXEq( float128ToFloatX( a ), float128ToFloatX( b ) );
+
+}
+
+flag slow_float128_le( float128 a, float128 b )
+{
+    floatX ax, bx;
+
+    ax = float128ToFloatX( a );
+    bx = float128ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLe( ax, bx );
+
+}
+
+flag slow_float128_lt( float128 a, float128 b )
+{
+    floatX ax, bx;
+
+    ax = float128ToFloatX( a );
+    bx = float128ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXLt( ax, bx );
+
+}
+
+flag slow_float128_eq_signaling( float128 a, float128 b )
+{
+    floatX ax, bx;
+
+    ax = float128ToFloatX( a );
+    bx = float128ToFloatX( b );
+    if ( ax.isNaN || bx.isNaN ) {
+        slow_float_exception_flags |= float_flag_invalid;
+    }
+    return floatXEq( ax, bx );
+
+}
+
+flag slow_float128_le_quiet( float128 a, float128 b )
+{
+
+    return floatXLe( float128ToFloatX( a ), float128ToFloatX( b ) );
+
+}
+
+flag slow_float128_lt_quiet( float128 a, float128 b )
+{
+
+    return floatXLt( float128ToFloatX( a ), float128ToFloatX( b ) );
+
+}
+
+#endif
+
diff --git a/testfloat/slowfloat.c b/testfloat/slowfloat.c
new file mode 100644
index 00000000000..ea69f82908e
--- /dev/null
+++ b/testfloat/slowfloat.c
@@ -0,0 +1,35 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include "milieu.h"
+#include "softfloat.h"
+#include "slowfloat.h"
+
+#ifdef BITS64
+#include "slowfloat-64.c"
+#else
+#include "slowfloat-32.c"
+#endif
+
diff --git a/testfloat/slowfloat.h b/testfloat/slowfloat.h
new file mode 100644
index 00000000000..45c6c6be00a
--- /dev/null
+++ b/testfloat/slowfloat.h
@@ -0,0 +1,167 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+extern int8 slow_float_rounding_mode;
+extern int8 slow_float_exception_flags;
+extern int8 slow_float_detect_tininess;
+#ifdef FLOATX80
+extern int8 slow_floatx80_rounding_precision;
+#endif
+
+float32 slow_int32_to_float32( int32 );
+float64 slow_int32_to_float64( int32 );
+#ifdef FLOATX80
+floatx80 slow_int32_to_floatx80( int32 );
+#endif
+#ifdef FLOAT128
+float128 slow_int32_to_float128( int32 );
+#endif
+#ifdef BITS64
+float32 slow_int64_to_float32( int64 );
+float64 slow_int64_to_float64( int64 );
+#ifdef FLOATX80
+floatx80 slow_int64_to_floatx80( int64 );
+#endif
+#ifdef FLOAT128
+float128 slow_int64_to_float128( int64 );
+#endif
+#endif
+
+int32 slow_float32_to_int32( float32 );
+int32 slow_float32_to_int32_round_to_zero( float32 );
+#ifdef BITS64
+int64 slow_float32_to_int64( float32 );
+int64 slow_float32_to_int64_round_to_zero( float32 );
+#endif
+float64 slow_float32_to_float64( float32 );
+#ifdef FLOATX80
+floatx80 slow_float32_to_floatx80( float32 );
+#endif
+#ifdef FLOAT128
+float128 slow_float32_to_float128( float32 );
+#endif
+
+float32 slow_float32_round_to_int( float32 );
+float32 slow_float32_add( float32, float32 );
+float32 slow_float32_sub( float32, float32 );
+float32 slow_float32_mul( float32, float32 );
+float32 slow_float32_div( float32, float32 );
+float32 slow_float32_rem( float32, float32 );
+float32 slow_float32_sqrt( float32 );
+flag slow_float32_eq( float32, float32 );
+flag slow_float32_le( float32, float32 );
+flag slow_float32_lt( float32, float32 );
+flag slow_float32_eq_signaling( float32, float32 );
+flag slow_float32_le_quiet( float32, float32 );
+flag slow_float32_lt_quiet( float32, float32 );
+
+int32 slow_float64_to_int32( float64 );
+int32 slow_float64_to_int32_round_to_zero( float64 );
+#ifdef BITS64
+int64 slow_float64_to_int64( float64 );
+int64 slow_float64_to_int64_round_to_zero( float64 );
+#endif
+float32 slow_float64_to_float32( float64 );
+#ifdef FLOATX80
+floatx80 slow_float64_to_floatx80( float64 );
+#endif
+#ifdef FLOAT128
+float128 slow_float64_to_float128( float64 );
+#endif
+
+float64 slow_float64_round_to_int( float64 );
+float64 slow_float64_add( float64, float64 );
+float64 slow_float64_sub( float64, float64 );
+float64 slow_float64_mul( float64, float64 );
+float64 slow_float64_div( float64, float64 );
+float64 slow_float64_rem( float64, float64 );
+float64 slow_float64_sqrt( float64 );
+flag slow_float64_eq( float64, float64 );
+flag slow_float64_le( float64, float64 );
+flag slow_float64_lt( float64, float64 );
+flag slow_float64_eq_signaling( float64, float64 );
+flag slow_float64_le_quiet( float64, float64 );
+flag slow_float64_lt_quiet( float64, float64 );
+
+#ifdef FLOATX80
+
+int32 slow_floatx80_to_int32( floatx80 );
+int32 slow_floatx80_to_int32_round_to_zero( floatx80 );
+#ifdef BITS64
+int64 slow_floatx80_to_int64( floatx80 );
+int64 slow_floatx80_to_int64_round_to_zero( floatx80 );
+#endif
+float32 slow_floatx80_to_float32( floatx80 );
+float64 slow_floatx80_to_float64( floatx80 );
+#ifdef FLOAT128
+float128 slow_floatx80_to_float128( floatx80 );
+#endif
+
+floatx80 slow_floatx80_round_to_int( floatx80 );
+floatx80 slow_floatx80_add( floatx80, floatx80 );
+floatx80 slow_floatx80_sub( floatx80, floatx80 );
+floatx80 slow_floatx80_mul( floatx80, floatx80 );
+floatx80 slow_floatx80_div( floatx80, floatx80 );
+floatx80 slow_floatx80_rem( floatx80, floatx80 );
+floatx80 slow_floatx80_sqrt( floatx80 );
+flag slow_floatx80_eq( floatx80, floatx80 );
+flag slow_floatx80_le( floatx80, floatx80 );
+flag slow_floatx80_lt( floatx80, floatx80 );
+flag slow_floatx80_eq_signaling( floatx80, floatx80 );
+flag slow_floatx80_le_quiet( floatx80, floatx80 );
+flag slow_floatx80_lt_quiet( floatx80, floatx80 );
+
+#endif
+
+#ifdef FLOAT128
+
+int32 slow_float128_to_int32( float128 );
+int32 slow_float128_to_int32_round_to_zero( float128 );
+#ifdef BITS64
+int64 slow_float128_to_int64( float128 );
+int64 slow_float128_to_int64_round_to_zero( float128 );
+#endif
+float32 slow_float128_to_float32( float128 );
+float64 slow_float128_to_float64( float128 );
+#ifdef FLOATX80
+floatx80 slow_float128_to_floatx80( float128 );
+#endif
+
+float128 slow_float128_round_to_int( float128 );
+float128 slow_float128_add( float128, float128 );
+float128 slow_float128_sub( float128, float128 );
+float128 slow_float128_mul( float128, float128 );
+float128 slow_float128_div( float128, float128 );
+float128 slow_float128_rem( float128, float128 );
+float128 slow_float128_sqrt( float128 );
+flag slow_float128_eq( float128, float128 );
+flag slow_float128_le( float128, float128 );
+flag slow_float128_lt( float128, float128 );
+flag slow_float128_eq_signaling( float128, float128 );
+flag slow_float128_le_quiet( float128, float128 );
+flag slow_float128_lt_quiet( float128, float128 );
+
+#endif
+
diff --git a/testfloat/systemBugs.txt b/testfloat/systemBugs.txt
new file mode 100644
index 00000000000..a0d171a31bb
--- /dev/null
+++ b/testfloat/systemBugs.txt
@@ -0,0 +1,323 @@
+
+Known Floating-point Bugs Detected by TestFloat
+
+John R. Hauser
+1997 December 15
+
+
+-------------------------------------------------------------------------------
+Introduction
+
+Several popular systems have bugs that TestFloat is very likely to run
+across.  The ones I know of are documented here.  First off, TestFloat finds
+no errors in the following processors/machines:
+
+    AMD 486 DX4's
+    Sun UltraSPARC 1's and 2's
+
+On the other hand, bugs are found in these processors/machines:
+
+    Older Intel Pentiums (with the divide bug)
+    Intel Pentium Pros
+    Sun SPARCstation 1's and IPX's
+    Sun SPARCstation 10's
+    HP Precision Architecture processors, with HP-UX prior to version 10.10
+
+For some reason, most of the bugs found involve conversions from floating-
+point to integer formats.
+
+The bugs are shown as actual TestFloat error lines, along with a brief
+explanation.  The error lines given are not necesarily exhaustive and were
+not necessarily output in the order shown.
+
+This document does not pretend to be an authoritative bug listing for all
+commercial processors.  The vast majority of processors are absent from this
+list because I have never run TestFloat on such machines and I thus have no
+knowledge of what bugs TestFloat might find in them.
+
+The latest version of this file can be found at the Web page `http://
+http.cs.berkeley.edu/~jhauser/arithmetic/testfloat.html'.
+
+
+-------------------------------------------------------------------------------
+Older Intel Pentiums (with the divide bug)
+
+The following conversion problems are found on Pentiums that also suffer
+from the infamous floating-point divide bug.  These bugs have been fixed on
+newer Pentiums.  (TestFloat does not find the divide bug.)
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+floatx80_to_int32
+
+-- A few small fractions are treated as though they were zero.
+
+	Errors found in floatx80_to_int32, rounding nearest_even:
+	3FFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.C000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	Errors found in floatx80_to_int32, rounding to_zero:
+	3FFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.C000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFC.C000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	Errors found in floatx80_to_int32, rounding down:
+	3FFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	3FFC.C000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFB.8000000000000000  soft: FFFFFFFF ....x  syst: 00000000 .....
+	BFFC.8000000000000000  soft: FFFFFFFF ....x  syst: 00000000 .....
+	BFFC.C000000000000000  soft: FFFFFFFF ....x  syst: 00000000 .....
+	Errors found in floatx80_to_int32, rounding up:
+	3FFB.8000000000000000  soft: 00000001 ....x  syst: 00000000 .....
+	3FFC.8000000000000000  soft: 00000001 ....x  syst: 00000000 .....
+	3FFC.C000000000000000  soft: 00000001 ....x  syst: 00000000 .....
+	BFFB.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+	BFFC.8000000000000000  soft: 00000000 ....x  syst: 00000000 .....
+
+   3FFB.8000000000000000 is the fraction 1/16; 3FFC.8000000000000000 is 1/8;
+   and 3FFC.C000000000000000 is 3/16.  Both positive and negative inputs are
+   affected.
+
+-- Some (all?) positive floating-point values between 2^32 - 1/2
+   (401E.FFFFFFFF00000000) and 2^32 (401F.0000000000000000) are rounded to
+   zero when the rounding mode is nearest/even or up.
+
+	Errors found in floatx80_to_int32, rounding nearest_even:
+	401E.FFFFFFFF80000000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFC00001FE  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFF8000000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFEC00000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFF002000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFC00000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFE00000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFD7FFE  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFFFFFE  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFFFFFF  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	Errors found in floatx80_to_int32, rounding up:
+	401E.FFFFFFFF00800000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFF80000000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFEFFFC000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFC000000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFE7FFFFF  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFF00000  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFE0800  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFF7FFB  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFFFFFE  soft: 7FFFFFFF v....  syst: 00000000 ....x
+	401E.FFFFFFFFFFFFFFFF  soft: 7FFFFFFF v....  syst: 00000000 ....x
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Intel Pentium Pros
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+floatx80_to_int32
+
+-- The inexact flag is sometimes raised instead of the invalid flag for
+   floating-point inputs under -(2^32) (C01F.0000000000000000).  This bug is
+   sporadic.  It appears to be deterministic but dependent on the sequence
+   of operations executed.
+
+	Errors found in floatx80_to_int32, rounding nearest_even:
+	C01F.C000000000000002  soft: 80000000 v....  syst: 80000000 ....x
+	C021.F00000000000003F  soft: 80000000 v....  syst: 80000000 ....x
+	Errors found in floatx80_to_int32, rounding to_zero:
+	C021.F00000000000003F  soft: 80000000 v....  syst: 80000000 ....x
+	Errors found in floatx80_to_int32, rounding up:
+	C01F.C000000000000007  soft: 80000000 v....  syst: 80000000 ....x
+	C01F.C000000000001000  soft: 80000000 v....  syst: 80000000 ....x
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Sun SPARCstation 1's and IPX's
+
+Some older SPARCstations appear confused about whether underflow tininess is
+detected before or after rounding.  For conversions from double precision
+to single precision, tininess is detected after rounding, while for all
+quadruple-precision operations it is detected before rounding.  Single- and
+double-precision multipies go both ways:
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+float32_mul, float64_mul
+
+-- For multiplies, underflow tininess is detected _before_ rounding if one
+   of the inputs is subnormal, and _after_ rounding otherwise.  If tininess
+   is assumed to be detected before rounding, the following errors are
+   generated:
+
+	Errors found in float32_mul, rounding nearest_even:
+	001.000001  07E.7FFFFE  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000001  87E.7FFFFE  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000002  07E.7FFFFC  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000002  87E.7FFFFC  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000004  07E.7FFFF8  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	Errors found in float32_mul, rounding down:
+	001.000001  87E.7FFFFE  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000002  87E.7FFFFC  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000004  87E.7FFFF8  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000008  87E.7FFFF0  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	001.000010  87E.7FFFE0  soft: 801.000000 ...ux  syst: 801.000000 ....x
+	Errors found in float32_mul, rounding up:
+	001.000001  07E.7FFFFE  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000002  07E.7FFFFC  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000004  07E.7FFFF8  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000008  07E.7FFFF0  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	001.000010  07E.7FFFE0  soft: 001.000000 ...ux  syst: 001.000000 ....x
+	Errors found in float64_mul, rounding nearest_even:
+	001.0000000000001  3FE.FFFFFFFFFFFFE
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000001  BFE.FFFFFFFFFFFFE
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000002  3FE.FFFFFFFFFFFFC
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000002  BFE.FFFFFFFFFFFFC
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000004  3FE.FFFFFFFFFFFF8
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	Errors found in float64_mul, rounding down:
+	001.0000000000001  BFE.FFFFFFFFFFFFE
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000002  BFE.FFFFFFFFFFFFC
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000004  BFE.FFFFFFFFFFFF8
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000008  BFE.FFFFFFFFFFFF0
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	001.0000000000010  BFE.FFFFFFFFFFFE0
+		soft: 801.0000000000000 ...ux  syst: 801.0000000000000 ....x
+	Errors found in float64_mul, rounding up:
+	001.0000000000001  3FE.FFFFFFFFFFFFE
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000002  3FE.FFFFFFFFFFFFC
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000004  3FE.FFFFFFFFFFFF8
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000008  3FE.FFFFFFFFFFFF0
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+	001.0000000000010  3FE.FFFFFFFFFFFE0
+		soft: 001.0000000000000 ...ux  syst: 001.0000000000000 ....x
+
+   If we assume tininess should be detected after rounding, we get the
+   following errors:
+
+	Errors found in float32_mul, rounding nearest_even:
+	000.7FFC00  07F.000400  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFC00  87F.000400  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFE00  07F.000200  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFE00  87F.000200  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFF00  07F.000100  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	Errors found in float32_mul, rounding down:
+	000.7FFC00  87F.000400  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFE00  87F.000200  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFF00  87F.000100  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFF80  87F.000080  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	000.7FFFC0  87F.000040  soft: 801.000000 ....x  syst: 801.000000 ...ux
+	Errors found in float32_mul, rounding up:
+	000.7FFC00  07F.000400  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFE00  07F.000200  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFF00  07F.000100  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFF80  07F.000080  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	000.7FFFC0  07F.000040  soft: 001.000000 ....x  syst: 001.000000 ...ux
+	Errors found in float64_mul, rounding nearest_even:
+	000.FFFFFFE000000  3FF.0000002000000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFE000000  BFF.0000002000000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFF000000  3FF.0000001000000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFF000000  BFF.0000001000000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFF800000  3FF.0000000800000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	Errors found in float64_mul, rounding down:
+	000.FFFFFFE000000  BFF.0000002000000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFF000000  BFF.0000001000000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFF800000  BFF.0000000800000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFFC00000  BFF.0000000400000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	000.FFFFFFFE00000  BFF.0000000200000
+		soft: 801.0000000000000 ....x  syst: 801.0000000000000 ...ux
+	Errors found in float64_mul, rounding up:
+	000.FFFFFFE000000  3FF.0000002000000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFF000000  3FF.0000001000000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFF800000  3FF.0000000800000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFFC00000  3FF.0000000400000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+	000.FFFFFFFE00000  3FF.0000000200000
+		soft: 001.0000000000000 ....x  syst: 001.0000000000000 ...ux
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Sun SPARCstation 10's
+
+Like other SPARCstations, some SPARCstation 10's are inconsistent regarding
+underflow tininess, detecting it after rounding for single- and double-
+precision operations and before rounding for quadruple-precision operations.
+The following bug has also been observed.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+float32_to_int32_round_to_zero, float64_to_int32_round_to_zero
+
+-- Single- and double-precision NaNs are converted to the integer zero.
+   (The invalid exception flag is raised correctly.)
+
+	Errors found in float32_to_int32_round_to_zero:
+	8FF.5D36AC  soft: 7FFFFFFF v....  syst: 00000000 v....
+	0FF.7FFFC0  soft: 7FFFFFFF v....  syst: 00000000 v....
+	8FF.7C0000  soft: 7FFFFFFF v....  syst: 00000000 v....
+	0FF.2AB7ED  soft: 7FFFFFFF v....  syst: 00000000 v....
+	0FF.03FFFF  soft: 7FFFFFFF v....  syst: 00000000 v....
+	Errors found in float64_to_int32_round_to_zero:
+	7FF.45AD84DB2524A  soft: 7FFFFFFF v....  syst: 00000000 v....
+	7FF.CFEE063EE0512  soft: 7FFFFFFF v....  syst: 00000000 v....
+	7FF.89FF03AB7DBA2  soft: 7FFFFFFF v....  syst: 00000000 v....
+	7FF.FFFFFFFFFF800  soft: 7FFFFFFF v....  syst: 00000000 v....
+	FFF.68A6410E91BF6  soft: 7FFFFFFF v....  syst: 00000000 v....
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+HP Precision Architecture processors, with HP-UX prior to version 10.10
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+float32_to_int32_round_to_zero, float64_to_int32_round_to_zero
+
+-- When the floating-point value is too large, the overflow and inexact
+   exception flags are raised instead of the invalid flag.
+
+	Errors found in float32_to_int32_round_to_zero:
+	89E.000007  soft: 80000000 v....  syst: 80000000 ..o.x
+	0A2.000020  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	8FA.7C0000  soft: 80000000 v....  syst: 80000000 ..o.x
+	Errors found in float64_to_int32_round_to_zero:
+	7FD.0448700002F1C  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	DAA.F000000000000  soft: 80000000 v....  syst: 80000000 ..o.x
+	41E.063DA00005E65  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	47E.FFFF800000000  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	51F.0000000000004  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	DDA.0000001FFFFFF  soft: 80000000 v....  syst: 80000000 ..o.x
+	D70.00000000003FF  soft: 80000000 v....  syst: 80000000 ..o.x
+	C7E.0000100000000  soft: 80000000 v....  syst: 80000000 ..o.x
+	47E.000000000007F  soft: 7FFFFFFF v....  syst: 7FFFFFFF ..o.x
+	D57.000000000FFFF  soft: 80000000 v....  syst: 80000000 ..o.x
+
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
diff --git a/testfloat/systflags.h b/testfloat/systflags.h
new file mode 100644
index 00000000000..23e6b236472
--- /dev/null
+++ b/testfloat/systflags.h
@@ -0,0 +1,33 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+Target-specific function for clearing the system's IEC/IEEE floating-point
+exception flags.  The previous value of the flags is returned.
+-------------------------------------------------------------------------------
+*/
+int8 syst_float_flags_clear( void );
+
diff --git a/testfloat/systfloat.c b/testfloat/systfloat.c
new file mode 100644
index 00000000000..08548c4981e
--- /dev/null
+++ b/testfloat/systfloat.c
@@ -0,0 +1,553 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <math.h>
+#include "milieu.h"
+#include "softfloat.h"
+#include "systfloat.h"
+
+float32 syst_int32_to_float32( int32 a )
+{
+    float32 z;
+
+    *( (float *) &z ) = a;
+    return z;
+
+}
+
+float64 syst_int32_to_float64( int32 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = a;
+    return z;
+
+}
+
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+
+floatx80 syst_int32_to_floatx80( int32 a )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) = a;
+    return z;
+
+}
+
+#endif
+
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+
+float128 syst_int32_to_float128( int32 a )
+{
+    float128 z;
+
+    *( (long double *) &z ) = a;
+    return z;
+
+}
+
+#endif
+
+#ifdef BITS64
+
+float32 syst_int64_to_float32( int64 a )
+{
+    float32 z;
+
+    *( (float *) &z ) = a;
+    return z;
+
+}
+
+float64 syst_int64_to_float64( int64 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = a;
+    return z;
+
+}
+
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+
+floatx80 syst_int64_to_floatx80( int64 a )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) = a;
+    return z;
+
+}
+
+#endif
+
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+
+float128 syst_int64_to_float128( int64 a )
+{
+    float128 z;
+
+    *( (long double *) &z ) = a;
+    return z;
+
+}
+
+#endif
+
+#endif
+
+int32 syst_float32_to_int32_round_to_zero( float32 a )
+{
+
+    return *( (float *) &a );
+
+}
+
+#ifdef BITS64
+
+int64 syst_float32_to_int64_round_to_zero( float32 a )
+{
+
+    return *( (float *) &a );
+
+}
+
+#endif
+
+float64 syst_float32_to_float64( float32 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (float *) &a );
+    return z;
+
+}
+
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+
+floatx80 syst_float32_to_floatx80( float32 a )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) = *( (float *) &a );
+    return z;
+
+}
+
+#endif
+
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+
+float128 syst_float32_to_float128( float32 a )
+{
+    float128 z;
+
+    *( (long double *) &z ) = *( (float *) &a );
+    return z;
+
+}
+
+#endif
+
+float32 syst_float32_add( float32 a, float32 b )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (float *) &a ) + *( (float *) &b );
+    return z;
+
+}
+
+float32 syst_float32_sub( float32 a, float32 b )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (float *) &a ) - *( (float *) &b );
+    return z;
+
+}
+
+float32 syst_float32_mul( float32 a, float32 b )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (float *) &a ) * *( (float *) &b );
+    return z;
+
+}
+
+float32 syst_float32_div( float32 a, float32 b )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (float *) &a ) / *( (float *) &b );
+    return z;
+
+}
+
+flag syst_float32_eq( float32 a, float32 b )
+{
+
+    return ( *( (float *) &a ) == *( (float *) &b ) );
+
+}
+
+flag syst_float32_le( float32 a, float32 b )
+{
+
+    return ( *( (float *) &a ) <= *( (float *) &b ) );
+
+}
+
+flag syst_float32_lt( float32 a, float32 b )
+{
+
+    return ( *( (float *) &a ) < *( (float *) &b ) );
+
+}
+
+int32 syst_float64_to_int32_round_to_zero( float64 a )
+{
+
+    return *( (double *) &a );
+
+}
+
+#ifdef BITS64
+
+int64 syst_float64_to_int64_round_to_zero( float64 a )
+{
+
+    return *( (double *) &a );
+
+}
+
+#endif
+
+float32 syst_float64_to_float32( float64 a )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (double *) &a );
+    return z;
+
+}
+
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+
+floatx80 syst_float64_to_floatx80( float64 a )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) = *( (double *) &a );
+    return z;
+
+}
+
+#endif
+
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+
+float128 syst_float64_to_float128( float64 a )
+{
+    float128 z;
+
+    *( (long double *) &z ) = *( (double *) &a );
+    return z;
+
+}
+
+#endif
+
+float64 syst_float64_add( float64 a, float64 b )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (double *) &a ) + *( (double *) &b );
+    return z;
+
+}
+
+float64 syst_float64_sub( float64 a, float64 b )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (double *) &a ) - *( (double *) &b );
+    return z;
+
+}
+
+float64 syst_float64_mul( float64 a, float64 b )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (double *) &a ) * *( (double *) &b );
+    return z;
+
+}
+
+float64 syst_float64_div( float64 a, float64 b )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (double *) &a ) / *( (double *) &b );
+    return z;
+
+}
+
+float64 syst_float64_sqrt( float64 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = sqrt( *( (double *) &a ) );
+    return z;
+
+}
+
+flag syst_float64_eq( float64 a, float64 b )
+{
+
+    return ( *( (double *) &a ) == *( (double *) &b ) );
+
+}
+
+flag syst_float64_le( float64 a, float64 b )
+{
+
+    return ( *( (double *) &a ) <= *( (double *) &b ) );
+
+}
+
+flag syst_float64_lt( float64 a, float64 b )
+{
+
+    return ( *( (double *) &a ) < *( (double *) &b ) );
+
+}
+
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+
+int32 syst_floatx80_to_int32_round_to_zero( floatx80 a )
+{
+
+    return *( (long double *) &a );
+
+}
+
+#ifdef BITS64
+
+int64 syst_floatx80_to_int64_round_to_zero( floatx80 a )
+{
+
+    return *( (long double *) &a );
+
+}
+
+#endif
+
+float32 syst_floatx80_to_float32( floatx80 a )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (long double *) &a );
+    return z;
+
+}
+
+float64 syst_floatx80_to_float64( floatx80 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (long double *) &a );
+    return z;
+
+}
+
+floatx80 syst_floatx80_add( floatx80 a, floatx80 b )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) + *( (long double *) &b );
+    return z;
+
+}
+
+floatx80 syst_floatx80_sub( floatx80 a, floatx80 b )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) - *( (long double *) &b );
+    return z;
+
+}
+
+floatx80 syst_floatx80_mul( floatx80 a, floatx80 b )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) * *( (long double *) &b );
+    return z;
+
+}
+
+floatx80 syst_floatx80_div( floatx80 a, floatx80 b )
+{
+    floatx80 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) / *( (long double *) &b );
+    return z;
+
+}
+
+flag syst_floatx80_eq( floatx80 a, floatx80 b )
+{
+
+    return ( *( (long double *) &a ) == *( (long double *) &b ) );
+
+}
+
+flag syst_floatx80_le( floatx80 a, floatx80 b )
+{
+
+    return ( *( (long double *) &a ) <= *( (long double *) &b ) );
+
+}
+
+flag syst_floatx80_lt( floatx80 a, floatx80 b )
+{
+
+    return ( *( (long double *) &a ) < *( (long double *) &b ) );
+
+}
+
+#endif
+
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+
+int32 syst_float128_to_int32_round_to_zero( float128 a )
+{
+
+    return *( (long double *) &a );
+
+}
+
+#ifdef BITS64
+
+int64 syst_float128_to_int64_round_to_zero( float128 a )
+{
+
+    return *( (long double *) &a );
+
+}
+
+#endif
+
+float32 syst_float128_to_float32( float128 a )
+{
+    float32 z;
+
+    *( (float *) &z ) = *( (long double *) &a );
+    return z;
+
+}
+
+float64 syst_float128_to_float64( float128 a )
+{
+    float64 z;
+
+    *( (double *) &z ) = *( (long double *) &a );
+    return z;
+
+}
+
+float128 syst_float128_add( float128 a, float128 b )
+{
+    float128 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) + *( (long double *) &b );
+    return z;
+
+}
+
+float128 syst_float128_sub( float128 a, float128 b )
+{
+    float128 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) - *( (long double *) &b );
+    return z;
+
+}
+
+float128 syst_float128_mul( float128 a, float128 b )
+{
+    float128 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) * *( (long double *) &b );
+    return z;
+
+}
+
+float128 syst_float128_div( float128 a, float128 b )
+{
+    float128 z;
+
+    *( (long double *) &z ) =
+        *( (long double *) &a ) / *( (long double *) &b );
+    return z;
+
+}
+
+flag syst_float128_eq( float128 a, float128 b )
+{
+
+    return ( *( (long double *) &a ) == *( (long double *) &b ) );
+
+}
+
+flag syst_float128_le( float128 a, float128 b )
+{
+
+    return ( *( (long double *) &a ) <= *( (long double *) &b ) );
+
+}
+
+flag syst_float128_lt( float128 a, float128 b )
+{
+
+    return ( *( (long double *) &a ) < *( (long double *) &b ) );
+
+}
+
+#endif
+
diff --git a/testfloat/systfloat.h b/testfloat/systfloat.h
new file mode 100644
index 00000000000..88f0dc40498
--- /dev/null
+++ b/testfloat/systfloat.h
@@ -0,0 +1,233 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+The following macros are defined to indicate that the corresponding
+functions exist.
+-------------------------------------------------------------------------------
+*/
+#define SYST_INT32_TO_FLOAT32
+#define SYST_INT32_TO_FLOAT64
+#ifdef BITS64
+#define SYST_INT64_TO_FLOAT32
+#define SYST_INT64_TO_FLOAT64
+#endif
+#define SYST_FLOAT32_TO_INT32_ROUND_TO_ZERO
+#ifdef BITS64
+#define SYST_FLOAT32_TO_INT64_ROUND_TO_ZERO
+#endif
+#define SYST_FLOAT32_TO_FLOAT64
+#define SYST_FLOAT32_ADD
+#define SYST_FLOAT32_SUB
+#define SYST_FLOAT32_MUL
+#define SYST_FLOAT32_DIV
+#define SYST_FLOAT32_EQ
+#define SYST_FLOAT32_LE
+#define SYST_FLOAT32_LT
+#define SYST_FLOAT64_TO_INT32_ROUND_TO_ZERO
+#ifdef BITS64
+#define SYST_FLOAT64_TO_INT64_ROUND_TO_ZERO
+#endif
+#define SYST_FLOAT64_TO_FLOAT32
+#define SYST_FLOAT64_ADD
+#define SYST_FLOAT64_SUB
+#define SYST_FLOAT64_MUL
+#define SYST_FLOAT64_DIV
+#define SYST_FLOAT64_SQRT
+#define SYST_FLOAT64_EQ
+#define SYST_FLOAT64_LE
+#define SYST_FLOAT64_LT
+#if defined( FLOATX80 ) && defined( LONG_DOUBLE_IS_FLOATX80 )
+#define SYST_INT32_TO_FLOATX80
+#ifdef BITS64
+#define SYST_INT64_TO_FLOATX80
+#endif
+#define SYST_FLOAT32_TO_FLOATX80
+#define SYST_FLOAT64_TO_FLOATX80
+#define SYST_FLOATX80_TO_INT32_ROUND_TO_ZERO
+#ifdef BITS64
+#define SYST_FLOATX80_TO_INT64_ROUND_TO_ZERO
+#endif
+#define SYST_FLOATX80_TO_FLOAT32
+#define SYST_FLOATX80_TO_FLOAT64
+#define SYST_FLOATX80_ADD
+#define SYST_FLOATX80_SUB
+#define SYST_FLOATX80_MUL
+#define SYST_FLOATX80_DIV
+#define SYST_FLOATX80_EQ
+#define SYST_FLOATX80_LE
+#define SYST_FLOATX80_LT
+#endif
+#if defined( FLOAT128 ) && defined( LONG_DOUBLE_IS_FLOAT128 )
+#define SYST_INT32_TO_FLOAT128
+#ifdef BITS64
+#define SYST_INT64_TO_FLOAT128
+#endif
+#define SYST_FLOAT32_TO_FLOAT128
+#define SYST_FLOAT64_TO_FLOAT128
+#define SYST_FLOAT128_TO_INT32_ROUND_TO_ZERO
+#ifdef BITS64
+#define SYST_FLOAT128_TO_INT64_ROUND_TO_ZERO
+#endif
+#define SYST_FLOAT128_TO_FLOAT32
+#define SYST_FLOAT128_TO_FLOAT64
+#define SYST_FLOAT128_ADD
+#define SYST_FLOAT128_SUB
+#define SYST_FLOAT128_MUL
+#define SYST_FLOAT128_DIV
+#define SYST_FLOAT128_EQ
+#define SYST_FLOAT128_LE
+#define SYST_FLOAT128_LT
+#endif
+
+/*
+-------------------------------------------------------------------------------
+System function declarations.  (Some of these functions may not exist.)
+-------------------------------------------------------------------------------
+*/
+float32 syst_int32_to_float32( int32 );
+float64 syst_int32_to_float64( int32 );
+#ifdef FLOATX80
+floatx80 syst_int32_to_floatx80( int32 );
+#endif
+#ifdef FLOAT128
+float128 syst_int32_to_float128( int32 );
+#endif
+#ifdef BITS64
+float32 syst_int64_to_float32( int64 );
+float64 syst_int64_to_float64( int64 );
+#ifdef FLOATX80
+floatx80 syst_int64_to_floatx80( int64 );
+#endif
+#ifdef FLOAT128
+float128 syst_int64_to_float128( int64 );
+#endif
+#endif
+int32 syst_float32_to_int32( float32 );
+int32 syst_float32_to_int32_round_to_zero( float32 );
+#ifdef BITS64
+int64 syst_float32_to_int64( float32 );
+int64 syst_float32_to_int64_round_to_zero( float32 );
+#endif
+float64 syst_float32_to_float64( float32 );
+#ifdef FLOATX80
+floatx80 syst_float32_to_floatx80( float32 );
+#endif
+#ifdef FLOAT128
+float128 syst_float32_to_float128( float32 );
+#endif
+float32 syst_float32_round_to_int( float32 );
+float32 syst_float32_add( float32, float32 );
+float32 syst_float32_sub( float32, float32 );
+float32 syst_float32_mul( float32, float32 );
+float32 syst_float32_div( float32, float32 );
+float32 syst_float32_rem( float32, float32 );
+float32 syst_float32_sqrt( float32 );
+flag syst_float32_eq( float32, float32 );
+flag syst_float32_le( float32, float32 );
+flag syst_float32_lt( float32, float32 );
+flag syst_float32_eq_signaling( float32, float32 );
+flag syst_float32_le_quiet( float32, float32 );
+flag syst_float32_lt_quiet( float32, float32 );
+int32 syst_float64_to_int32( float64 );
+int32 syst_float64_to_int32_round_to_zero( float64 );
+#ifdef BITS64
+int64 syst_float64_to_int64( float64 );
+int64 syst_float64_to_int64_round_to_zero( float64 );
+#endif
+float32 syst_float64_to_float32( float64 );
+#ifdef FLOATX80
+floatx80 syst_float64_to_floatx80( float64 );
+#endif
+#ifdef FLOAT128
+float128 syst_float64_to_float128( float64 );
+#endif
+float64 syst_float64_round_to_int( float64 );
+float64 syst_float64_add( float64, float64 );
+float64 syst_float64_sub( float64, float64 );
+float64 syst_float64_mul( float64, float64 );
+float64 syst_float64_div( float64, float64 );
+float64 syst_float64_rem( float64, float64 );
+float64 syst_float64_sqrt( float64 );
+flag syst_float64_eq( float64, float64 );
+flag syst_float64_le( float64, float64 );
+flag syst_float64_lt( float64, float64 );
+flag syst_float64_eq_signaling( float64, float64 );
+flag syst_float64_le_quiet( float64, float64 );
+flag syst_float64_lt_quiet( float64, float64 );
+#ifdef FLOATX80
+int32 syst_floatx80_to_int32( floatx80 );
+int32 syst_floatx80_to_int32_round_to_zero( floatx80 );
+#ifdef BITS64
+int64 syst_floatx80_to_int64( floatx80 );
+int64 syst_floatx80_to_int64_round_to_zero( floatx80 );
+#endif
+float32 syst_floatx80_to_float32( floatx80 );
+float64 syst_floatx80_to_float64( floatx80 );
+#ifdef FLOAT128
+float128 syst_floatx80_to_float128( floatx80 );
+#endif
+floatx80 syst_floatx80_round_to_int( floatx80 );
+floatx80 syst_floatx80_add( floatx80, floatx80 );
+floatx80 syst_floatx80_sub( floatx80, floatx80 );
+floatx80 syst_floatx80_mul( floatx80, floatx80 );
+floatx80 syst_floatx80_div( floatx80, floatx80 );
+floatx80 syst_floatx80_rem( floatx80, floatx80 );
+floatx80 syst_floatx80_sqrt( floatx80 );
+flag syst_floatx80_eq( floatx80, floatx80 );
+flag syst_floatx80_le( floatx80, floatx80 );
+flag syst_floatx80_lt( floatx80, floatx80 );
+flag syst_floatx80_eq_signaling( floatx80, floatx80 );
+flag syst_floatx80_le_quiet( floatx80, floatx80 );
+flag syst_floatx80_lt_quiet( floatx80, floatx80 );
+#endif
+#ifdef FLOAT128
+int32 syst_float128_to_int32( float128 );
+int32 syst_float128_to_int32_round_to_zero( float128 );
+#ifdef BITS64
+int64 syst_float128_to_int64( float128 );
+int64 syst_float128_to_int64_round_to_zero( float128 );
+#endif
+float32 syst_float128_to_float32( float128 );
+float64 syst_float128_to_float64( float128 );
+#ifdef FLOATX80
+floatx80 syst_float128_to_floatx80( float128 );
+#endif
+float128 syst_float128_round_to_int( float128 );
+float128 syst_float128_add( float128, float128 );
+float128 syst_float128_sub( float128, float128 );
+float128 syst_float128_mul( float128, float128 );
+float128 syst_float128_div( float128, float128 );
+float128 syst_float128_rem( float128, float128 );
+float128 syst_float128_sqrt( float128 );
+flag syst_float128_eq( float128, float128 );
+flag syst_float128_le( float128, float128 );
+flag syst_float128_lt( float128, float128 );
+flag syst_float128_eq_signaling( float128, float128 );
+flag syst_float128_le_quiet( float128, float128 );
+flag syst_float128_lt_quiet( float128, float128 );
+#endif
+
diff --git a/testfloat/systmodes.h b/testfloat/systmodes.h
new file mode 100644
index 00000000000..b2befa4ad59
--- /dev/null
+++ b/testfloat/systmodes.h
@@ -0,0 +1,42 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+Target-specific function for setting the system's IEC/IEEE floating-point
+rounding mode.  Other system modes are also initialized as necessary (for
+example, exception trapping may be disabled).
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_mode( int8 );
+
+/*
+-------------------------------------------------------------------------------
+Target-specific function for setting the IEC/IEEE rounding precision of
+subsequent extended double-precision operations performed by the system.
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_precision( int8 );
+
diff --git a/testfloat/templates/Makefile b/testfloat/templates/Makefile
new file mode 100644
index 00000000000..f5f3cde713c
--- /dev/null
+++ b/testfloat/templates/Makefile
@@ -0,0 +1,67 @@
+
+PROCESSOR_H = ../../processors/!!!processor.h
+SOFTFLOAT_VERSION = bits64
+TARGET = !!!target
+SOFTFLOAT_DIR = ../../softfloat/$(SOFTFLOAT_VERSION)/$(TARGET)
+
+OBJ = .o
+EXE =
+INCLUDES = -I. -I.. -I$(SOFTFLOAT_DIR)
+COMPILE_C = gcc -c -o $@ $(INCLUDES) -I- -O2
+COMPILE_SLOWFLOAT_C = gcc -c -o $@ $(INCLUDES) -I- -O3
+LINK = gcc -o $@
+
+SOFTFLOAT_H = $(SOFTFLOAT_DIR)/softfloat.h
+SOFTFLOAT_OBJ = $(SOFTFLOAT_DIR)/softfloat$(OBJ)
+
+ALL: testsoftfloat$(EXE) testfloat$(EXE)
+
+systmodes$(OBJ): milieu.h ../systmodes.h systmodes.c
+	$(COMPILE_C) systmodes.c
+
+systflags$(OBJ): milieu.h ../systflags.h systflags.c
+	$(COMPILE_C) systflags.c
+
+systfloat$(OBJ): milieu.h $(SOFTFLOAT_H) ../systfloat.h ../systfloat.c
+	$(COMPILE_C) ../systfloat.c
+
+#------------------------------------------------------------------------------
+# Probably O.K. below here.
+#------------------------------------------------------------------------------
+
+milieu.h: $(PROCESSOR_H)
+	touch milieu.h
+
+fail$(OBJ): milieu.h ../fail.h
+	$(COMPILE_C) ../fail.c
+
+random$(OBJ): milieu.h ../random.h
+	$(COMPILE_C) ../random.c
+
+testCases$(OBJ): milieu.h ../fail.h ../random.h $(SOFTFLOAT_H) ../testCases.h ../testCases.c
+	$(COMPILE_C) ../testCases.c
+
+writeHex$(OBJ): milieu.h $(SOFTFLOAT_H) ../writeHex.h ../writeHex.c
+	$(COMPILE_C) ../writeHex.c
+
+testLoops$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../writeHex.h ../testLoops.h ../testLoops.c
+	$(COMPILE_C) ../testLoops.c
+
+slowfloat$(OBJ): milieu.h $(SOFTFLOAT_H) ../slowfloat.h ../slowfloat-32.c ../slowfloat-64.c ../slowfloat.c
+	$(COMPILE_SLOWFLOAT_C) ../slowfloat.c
+
+testsoftfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../slowfloat.h ../testsoftfloat.c
+	$(COMPILE_C) ../testsoftfloat.c
+
+testsoftfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) slowfloat$(OBJ) testsoftfloat$(OBJ)
+
+testFunction$(OBJ): milieu.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systmodes.h ../systflags.h systfloat.h ../testFunction.h ../testFunction.c
+	$(COMPILE_C) ../testFunction.c
+
+testfloat$(OBJ): milieu.h ../fail.h $(SOFTFLOAT_H) ../testCases.h ../testLoops.h ../systflags.h ../testFunction.h testfloat.c
+	$(COMPILE_C) testfloat.c
+
+testfloat$(EXE): fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+	$(LINK) fail$(OBJ) random$(OBJ) $(SOFTFLOAT_OBJ) testCases$(OBJ) writeHex$(OBJ) testLoops$(OBJ) systmodes$(OBJ) systflags$(OBJ) systfloat$(OBJ) testFunction$(OBJ) testfloat$(OBJ)
+
diff --git a/testfloat/templates/milieu.h b/testfloat/templates/milieu.h
new file mode 100644
index 00000000000..56d3ac49a9a
--- /dev/null
+++ b/testfloat/templates/milieu.h
@@ -0,0 +1,62 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+/*
+-------------------------------------------------------------------------------
+Include common integer types and flags.
+-------------------------------------------------------------------------------
+*/
+#include "../../processors/!!!processor.h"
+
+/*
+-------------------------------------------------------------------------------
+If the `BITS64' macro is defined by the processor header file but the
+version of SoftFloat being used/tested is the 32-bit one (`bits32'), the
+`BITS64' macro must be undefined here.
+-------------------------------------------------------------------------------
+#undef BITS64
+*/
+
+/*
+-------------------------------------------------------------------------------
+The macro `LONG_DOUBLE_IS_FLOATX80' can be defined to indicate that the
+C compiler supports the type `long double' as an extended double-precision
+format.  Alternatively, the macro `LONG_DOUBLE_IS_FLOAT128' can be defined
+to indicate that `long double' is a quadruple-precision format.  If neither
+of these macros is defined, `long double' will be ignored.
+-------------------------------------------------------------------------------
+#define LONG_DOUBLE_IS_FLOATX80
+*/
+
+/*
+-------------------------------------------------------------------------------
+Symbolic Boolean literals.
+-------------------------------------------------------------------------------
+*/
+enum {
+    FALSE = 0,
+    TRUE  = 1
+};
+
diff --git a/testfloat/templates/systflags.c b/testfloat/templates/systflags.c
new file mode 100644
index 00000000000..3eb86c1ade8
--- /dev/null
+++ b/testfloat/templates/systflags.c
@@ -0,0 +1,41 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include "milieu.h"
+#include "systflags.h"
+
+/*
+-------------------------------------------------------------------------------
+Clears the system's IEC/IEEE floating-point exception flags.  Returns the
+previous value of the flags.
+-------------------------------------------------------------------------------
+*/
+int8 syst_float_flags_clear( void )
+{
+
+    !!!code
+
+}
+
diff --git a/testfloat/templates/systmodes.c b/testfloat/templates/systmodes.c
new file mode 100644
index 00000000000..49450ab258b
--- /dev/null
+++ b/testfloat/templates/systmodes.c
@@ -0,0 +1,58 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include "milieu.h"
+#include "systmodes.h"
+
+/*
+-------------------------------------------------------------------------------
+Sets the system's IEC/IEEE floating-point rounding mode.  Also disables all
+system exception traps.
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_mode( int8 roundingMode )
+{
+
+    !!!code
+
+}
+
+/*
+-------------------------------------------------------------------------------
+Sets the rounding precision of subsequent extended double-precision
+operations.  The `precision' argument should be one of 0, 32, 64, or 80.
+If `precision' is 32, the rounding precision is set equivalent to single
+precision; else if `precision' is 64, the rounding precision is set
+equivalent to double precision; else the rounding precision is set to full
+extended double precision.
+-------------------------------------------------------------------------------
+*/
+void syst_float_set_rounding_precision( int8 precision )
+{
+
+    !!!code (possibly empty)
+
+}
+
diff --git a/testfloat/testCases.c b/testfloat/testCases.c
new file mode 100644
index 00000000000..44676306af3
--- /dev/null
+++ b/testfloat/testCases.c
@@ -0,0 +1,3679 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include "milieu.h"
+#include "fail.h"
+#include "random.h"
+#include "softfloat.h"
+#include "testCases.h"
+
+typedef struct {
+    int16 expNum, term1Num, term2Num;
+    flag done;
+} sequenceT;
+
+enum {
+    int32NumP1 = 124
+};
+
+static const uint32 int32P1[ int32NumP1 ] = {
+    0x00000000,
+    0x00000001,
+    0x00000002,
+    0x00000004,
+    0x00000008,
+    0x00000010,
+    0x00000020,
+    0x00000040,
+    0x00000080,
+    0x00000100,
+    0x00000200,
+    0x00000400,
+    0x00000800,
+    0x00001000,
+    0x00002000,
+    0x00004000,
+    0x00008000,
+    0x00010000,
+    0x00020000,
+    0x00040000,
+    0x00080000,
+    0x00100000,
+    0x00200000,
+    0x00400000,
+    0x00800000,
+    0x01000000,
+    0x02000000,
+    0x04000000,
+    0x08000000,
+    0x10000000,
+    0x20000000,
+    0x40000000,
+    0x80000000,
+    0xC0000000,
+    0xE0000000,
+    0xF0000000,
+    0xF8000000,
+    0xFC000000,
+    0xFE000000,
+    0xFF000000,
+    0xFF800000,
+    0xFFC00000,
+    0xFFE00000,
+    0xFFF00000,
+    0xFFF80000,
+    0xFFFC0000,
+    0xFFFE0000,
+    0xFFFF0000,
+    0xFFFF8000,
+    0xFFFFC000,
+    0xFFFFE000,
+    0xFFFFF000,
+    0xFFFFF800,
+    0xFFFFFC00,
+    0xFFFFFE00,
+    0xFFFFFF00,
+    0xFFFFFF80,
+    0xFFFFFFC0,
+    0xFFFFFFE0,
+    0xFFFFFFF0,
+    0xFFFFFFF8,
+    0xFFFFFFFC,
+    0xFFFFFFFE,
+    0xFFFFFFFF,
+    0xFFFFFFFD,
+    0xFFFFFFFB,
+    0xFFFFFFF7,
+    0xFFFFFFEF,
+    0xFFFFFFDF,
+    0xFFFFFFBF,
+    0xFFFFFF7F,
+    0xFFFFFEFF,
+    0xFFFFFDFF,
+    0xFFFFFBFF,
+    0xFFFFF7FF,
+    0xFFFFEFFF,
+    0xFFFFDFFF,
+    0xFFFFBFFF,
+    0xFFFF7FFF,
+    0xFFFEFFFF,
+    0xFFFDFFFF,
+    0xFFFBFFFF,
+    0xFFF7FFFF,
+    0xFFEFFFFF,
+    0xFFDFFFFF,
+    0xFFBFFFFF,
+    0xFF7FFFFF,
+    0xFEFFFFFF,
+    0xFDFFFFFF,
+    0xFBFFFFFF,
+    0xF7FFFFFF,
+    0xEFFFFFFF,
+    0xDFFFFFFF,
+    0xBFFFFFFF,
+    0x7FFFFFFF,
+    0x3FFFFFFF,
+    0x1FFFFFFF,
+    0x0FFFFFFF,
+    0x07FFFFFF,
+    0x03FFFFFF,
+    0x01FFFFFF,
+    0x00FFFFFF,
+    0x007FFFFF,
+    0x003FFFFF,
+    0x001FFFFF,
+    0x000FFFFF,
+    0x0007FFFF,
+    0x0003FFFF,
+    0x0001FFFF,
+    0x0000FFFF,
+    0x00007FFF,
+    0x00003FFF,
+    0x00001FFF,
+    0x00000FFF,
+    0x000007FF,
+    0x000003FF,
+    0x000001FF,
+    0x000000FF,
+    0x0000007F,
+    0x0000003F,
+    0x0000001F,
+    0x0000000F,
+    0x00000007,
+    0x00000003
+};
+
+static int32 int32NextP1( sequenceT *sequencePtr )
+{
+    uint8 termNum;
+    int32 z;
+
+    termNum = sequencePtr->term1Num;
+    z = int32P1[ termNum ];
+    ++termNum;
+    if ( int32NumP1 <= termNum ) {
+        termNum = 0;
+        sequencePtr->done = TRUE;
+    }
+    sequencePtr->term1Num = termNum;
+    return (sbits32) z;
+
+}
+
+static const int32 int32NumP2 = ( int32NumP1 * int32NumP1 + int32NumP1 ) / 2;
+
+static int32 int32NextP2( sequenceT *sequencePtr )
+{
+    uint8 term1Num, term2Num;
+    int32 z;
+
+    term2Num = sequencePtr->term2Num;
+    term1Num = sequencePtr->term1Num;
+    z = int32P1[ term1Num ] + int32P1[ term2Num ];
+    ++term2Num;
+    if ( int32NumP1 <= term2Num ) {
+        ++term1Num;
+        if ( int32NumP1 <= term1Num ) {
+            term1Num = 0;
+            sequencePtr->done = TRUE;
+        }
+        term2Num = term1Num;
+        sequencePtr->term1Num = term1Num;
+    }
+    sequencePtr->term2Num = term2Num;
+    return (sbits32) z;
+
+}
+
+static int32 int32RandomP3( void )
+{
+
+    return
+        (sbits32) (
+              int32P1[ randomUint8() % int32NumP1 ]
+            + int32P1[ randomUint8() % int32NumP1 ]
+            + int32P1[ randomUint8() % int32NumP1 ]
+        );
+
+}
+
+enum {
+    int32NumPInfWeightMasks = 29
+};
+
+static const uint32 int32PInfWeightMasks[ int32NumPInfWeightMasks ] = {
+    0xFFFFFFFF,
+    0x7FFFFFFF,
+    0x3FFFFFFF,
+    0x1FFFFFFF,
+    0x0FFFFFFF,
+    0x07FFFFFF,
+    0x03FFFFFF,
+    0x01FFFFFF,
+    0x00FFFFFF,
+    0x007FFFFF,
+    0x003FFFFF,
+    0x001FFFFF,
+    0x000FFFFF,
+    0x0007FFFF,
+    0x0003FFFF,
+    0x0001FFFF,
+    0x0000FFFF,
+    0x00007FFF,
+    0x00003FFF,
+    0x00001FFF,
+    0x00000FFF,
+    0x000007FF,
+    0x000003FF,
+    0x000001FF,
+    0x000000FF,
+    0x0000007F,
+    0x0000003F,
+    0x0000001F,
+    0x0000000F
+};
+
+static const uint32 int32PInfWeightOffsets[ int32NumPInfWeightMasks ] = {
+    0x00000000,
+    0xC0000000,
+    0xE0000000,
+    0xF0000000,
+    0xF8000000,
+    0xFC000000,
+    0xFE000000,
+    0xFF000000,
+    0xFF800000,
+    0xFFC00000,
+    0xFFE00000,
+    0xFFF00000,
+    0xFFF80000,
+    0xFFFC0000,
+    0xFFFE0000,
+    0xFFFF0000,
+    0xFFFF8000,
+    0xFFFFC000,
+    0xFFFFE000,
+    0xFFFFF000,
+    0xFFFFF800,
+    0xFFFFFC00,
+    0xFFFFFE00,
+    0xFFFFFF00,
+    0xFFFFFF80,
+    0xFFFFFFC0,
+    0xFFFFFFE0,
+    0xFFFFFFF0,
+    0xFFFFFFF8
+};
+
+static int32 int32RandomPInf( void )
+{
+    int8 weightMaskNum;
+
+    weightMaskNum = randomUint8() % int32NumPInfWeightMasks;
+    return
+        (sbits32) (
+              ( randomUint32() & int32PInfWeightMasks[ weightMaskNum ] )
+            + int32PInfWeightOffsets[ weightMaskNum ]
+        );
+
+}
+
+#ifdef BITS64
+
+enum {
+    int64NumP1 = 252
+};
+
+static const uint64 int64P1[ int64NumP1 ] = {
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0x0000000000000001 ),
+    LIT64( 0x0000000000000002 ),
+    LIT64( 0x0000000000000004 ),
+    LIT64( 0x0000000000000008 ),
+    LIT64( 0x0000000000000010 ),
+    LIT64( 0x0000000000000020 ),
+    LIT64( 0x0000000000000040 ),
+    LIT64( 0x0000000000000080 ),
+    LIT64( 0x0000000000000100 ),
+    LIT64( 0x0000000000000200 ),
+    LIT64( 0x0000000000000400 ),
+    LIT64( 0x0000000000000800 ),
+    LIT64( 0x0000000000001000 ),
+    LIT64( 0x0000000000002000 ),
+    LIT64( 0x0000000000004000 ),
+    LIT64( 0x0000000000008000 ),
+    LIT64( 0x0000000000010000 ),
+    LIT64( 0x0000000000020000 ),
+    LIT64( 0x0000000000040000 ),
+    LIT64( 0x0000000000080000 ),
+    LIT64( 0x0000000000100000 ),
+    LIT64( 0x0000000000200000 ),
+    LIT64( 0x0000000000400000 ),
+    LIT64( 0x0000000000800000 ),
+    LIT64( 0x0000000001000000 ),
+    LIT64( 0x0000000002000000 ),
+    LIT64( 0x0000000004000000 ),
+    LIT64( 0x0000000008000000 ),
+    LIT64( 0x0000000010000000 ),
+    LIT64( 0x0000000020000000 ),
+    LIT64( 0x0000000040000000 ),
+    LIT64( 0x0000000080000000 ),
+    LIT64( 0x0000000100000000 ),
+    LIT64( 0x0000000200000000 ),
+    LIT64( 0x0000000400000000 ),
+    LIT64( 0x0000000800000000 ),
+    LIT64( 0x0000001000000000 ),
+    LIT64( 0x0000002000000000 ),
+    LIT64( 0x0000004000000000 ),
+    LIT64( 0x0000008000000000 ),
+    LIT64( 0x0000010000000000 ),
+    LIT64( 0x0000020000000000 ),
+    LIT64( 0x0000040000000000 ),
+    LIT64( 0x0000080000000000 ),
+    LIT64( 0x0000100000000000 ),
+    LIT64( 0x0000200000000000 ),
+    LIT64( 0x0000400000000000 ),
+    LIT64( 0x0000800000000000 ),
+    LIT64( 0x0001000000000000 ),
+    LIT64( 0x0002000000000000 ),
+    LIT64( 0x0004000000000000 ),
+    LIT64( 0x0008000000000000 ),
+    LIT64( 0x0010000000000000 ),
+    LIT64( 0x0020000000000000 ),
+    LIT64( 0x0040000000000000 ),
+    LIT64( 0x0080000000000000 ),
+    LIT64( 0x0100000000000000 ),
+    LIT64( 0x0200000000000000 ),
+    LIT64( 0x0400000000000000 ),
+    LIT64( 0x0800000000000000 ),
+    LIT64( 0x1000000000000000 ),
+    LIT64( 0x2000000000000000 ),
+    LIT64( 0x4000000000000000 ),
+    LIT64( 0x8000000000000000 ),
+    LIT64( 0xC000000000000000 ),
+    LIT64( 0xE000000000000000 ),
+    LIT64( 0xF000000000000000 ),
+    LIT64( 0xF800000000000000 ),
+    LIT64( 0xFC00000000000000 ),
+    LIT64( 0xFE00000000000000 ),
+    LIT64( 0xFF00000000000000 ),
+    LIT64( 0xFF80000000000000 ),
+    LIT64( 0xFFC0000000000000 ),
+    LIT64( 0xFFE0000000000000 ),
+    LIT64( 0xFFF0000000000000 ),
+    LIT64( 0xFFF8000000000000 ),
+    LIT64( 0xFFFC000000000000 ),
+    LIT64( 0xFFFE000000000000 ),
+    LIT64( 0xFFFF000000000000 ),
+    LIT64( 0xFFFF800000000000 ),
+    LIT64( 0xFFFFC00000000000 ),
+    LIT64( 0xFFFFE00000000000 ),
+    LIT64( 0xFFFFF00000000000 ),
+    LIT64( 0xFFFFF80000000000 ),
+    LIT64( 0xFFFFFC0000000000 ),
+    LIT64( 0xFFFFFE0000000000 ),
+    LIT64( 0xFFFFFF0000000000 ),
+    LIT64( 0xFFFFFF8000000000 ),
+    LIT64( 0xFFFFFFC000000000 ),
+    LIT64( 0xFFFFFFE000000000 ),
+    LIT64( 0xFFFFFFF000000000 ),
+    LIT64( 0xFFFFFFF800000000 ),
+    LIT64( 0xFFFFFFFC00000000 ),
+    LIT64( 0xFFFFFFFE00000000 ),
+    LIT64( 0xFFFFFFFF00000000 ),
+    LIT64( 0xFFFFFFFF80000000 ),
+    LIT64( 0xFFFFFFFFC0000000 ),
+    LIT64( 0xFFFFFFFFE0000000 ),
+    LIT64( 0xFFFFFFFFF0000000 ),
+    LIT64( 0xFFFFFFFFF8000000 ),
+    LIT64( 0xFFFFFFFFFC000000 ),
+    LIT64( 0xFFFFFFFFFE000000 ),
+    LIT64( 0xFFFFFFFFFF000000 ),
+    LIT64( 0xFFFFFFFFFF800000 ),
+    LIT64( 0xFFFFFFFFFFC00000 ),
+    LIT64( 0xFFFFFFFFFFE00000 ),
+    LIT64( 0xFFFFFFFFFFF00000 ),
+    LIT64( 0xFFFFFFFFFFF80000 ),
+    LIT64( 0xFFFFFFFFFFFC0000 ),
+    LIT64( 0xFFFFFFFFFFFE0000 ),
+    LIT64( 0xFFFFFFFFFFFF0000 ),
+    LIT64( 0xFFFFFFFFFFFF8000 ),
+    LIT64( 0xFFFFFFFFFFFFC000 ),
+    LIT64( 0xFFFFFFFFFFFFE000 ),
+    LIT64( 0xFFFFFFFFFFFFF000 ),
+    LIT64( 0xFFFFFFFFFFFFF800 ),
+    LIT64( 0xFFFFFFFFFFFFFC00 ),
+    LIT64( 0xFFFFFFFFFFFFFE00 ),
+    LIT64( 0xFFFFFFFFFFFFFF00 ),
+    LIT64( 0xFFFFFFFFFFFFFF80 ),
+    LIT64( 0xFFFFFFFFFFFFFFC0 ),
+    LIT64( 0xFFFFFFFFFFFFFFE0 ),
+    LIT64( 0xFFFFFFFFFFFFFFF0 ),
+    LIT64( 0xFFFFFFFFFFFFFFF8 ),
+    LIT64( 0xFFFFFFFFFFFFFFFC ),
+    LIT64( 0xFFFFFFFFFFFFFFFE ),
+    LIT64( 0xFFFFFFFFFFFFFFFF ),
+    LIT64( 0xFFFFFFFFFFFFFFFD ),
+    LIT64( 0xFFFFFFFFFFFFFFFB ),
+    LIT64( 0xFFFFFFFFFFFFFFF7 ),
+    LIT64( 0xFFFFFFFFFFFFFFEF ),
+    LIT64( 0xFFFFFFFFFFFFFFDF ),
+    LIT64( 0xFFFFFFFFFFFFFFBF ),
+    LIT64( 0xFFFFFFFFFFFFFF7F ),
+    LIT64( 0xFFFFFFFFFFFFFEFF ),
+    LIT64( 0xFFFFFFFFFFFFFDFF ),
+    LIT64( 0xFFFFFFFFFFFFFBFF ),
+    LIT64( 0xFFFFFFFFFFFFF7FF ),
+    LIT64( 0xFFFFFFFFFFFFEFFF ),
+    LIT64( 0xFFFFFFFFFFFFDFFF ),
+    LIT64( 0xFFFFFFFFFFFFBFFF ),
+    LIT64( 0xFFFFFFFFFFFF7FFF ),
+    LIT64( 0xFFFFFFFFFFFEFFFF ),
+    LIT64( 0xFFFFFFFFFFFDFFFF ),
+    LIT64( 0xFFFFFFFFFFFBFFFF ),
+    LIT64( 0xFFFFFFFFFFF7FFFF ),
+    LIT64( 0xFFFFFFFFFFEFFFFF ),
+    LIT64( 0xFFFFFFFFFFDFFFFF ),
+    LIT64( 0xFFFFFFFFFFBFFFFF ),
+    LIT64( 0xFFFFFFFFFF7FFFFF ),
+    LIT64( 0xFFFFFFFFFEFFFFFF ),
+    LIT64( 0xFFFFFFFFFDFFFFFF ),
+    LIT64( 0xFFFFFFFFFBFFFFFF ),
+    LIT64( 0xFFFFFFFFF7FFFFFF ),
+    LIT64( 0xFFFFFFFFEFFFFFFF ),
+    LIT64( 0xFFFFFFFFDFFFFFFF ),
+    LIT64( 0xFFFFFFFFBFFFFFFF ),
+    LIT64( 0xFFFFFFFF7FFFFFFF ),
+    LIT64( 0xFFFFFFFEFFFFFFFF ),
+    LIT64( 0xFFFFFFFDFFFFFFFF ),
+    LIT64( 0xFFFFFFFBFFFFFFFF ),
+    LIT64( 0xFFFFFFF7FFFFFFFF ),
+    LIT64( 0xFFFFFFEFFFFFFFFF ),
+    LIT64( 0xFFFFFFDFFFFFFFFF ),
+    LIT64( 0xFFFFFFBFFFFFFFFF ),
+    LIT64( 0xFFFFFF7FFFFFFFFF ),
+    LIT64( 0xFFFFFEFFFFFFFFFF ),
+    LIT64( 0xFFFFFDFFFFFFFFFF ),
+    LIT64( 0xFFFFFBFFFFFFFFFF ),
+    LIT64( 0xFFFFF7FFFFFFFFFF ),
+    LIT64( 0xFFFFEFFFFFFFFFFF ),
+    LIT64( 0xFFFFDFFFFFFFFFFF ),
+    LIT64( 0xFFFFBFFFFFFFFFFF ),
+    LIT64( 0xFFFF7FFFFFFFFFFF ),
+    LIT64( 0xFFFEFFFFFFFFFFFF ),
+    LIT64( 0xFFFDFFFFFFFFFFFF ),
+    LIT64( 0xFFFBFFFFFFFFFFFF ),
+    LIT64( 0xFFF7FFFFFFFFFFFF ),
+    LIT64( 0xFFEFFFFFFFFFFFFF ),
+    LIT64( 0xFFDFFFFFFFFFFFFF ),
+    LIT64( 0xFFBFFFFFFFFFFFFF ),
+    LIT64( 0xFF7FFFFFFFFFFFFF ),
+    LIT64( 0xFEFFFFFFFFFFFFFF ),
+    LIT64( 0xFDFFFFFFFFFFFFFF ),
+    LIT64( 0xFBFFFFFFFFFFFFFF ),
+    LIT64( 0xF7FFFFFFFFFFFFFF ),
+    LIT64( 0xEFFFFFFFFFFFFFFF ),
+    LIT64( 0xDFFFFFFFFFFFFFFF ),
+    LIT64( 0xBFFFFFFFFFFFFFFF ),
+    LIT64( 0x7FFFFFFFFFFFFFFF ),
+    LIT64( 0x3FFFFFFFFFFFFFFF ),
+    LIT64( 0x1FFFFFFFFFFFFFFF ),
+    LIT64( 0x0FFFFFFFFFFFFFFF ),
+    LIT64( 0x07FFFFFFFFFFFFFF ),
+    LIT64( 0x03FFFFFFFFFFFFFF ),
+    LIT64( 0x01FFFFFFFFFFFFFF ),
+    LIT64( 0x00FFFFFFFFFFFFFF ),
+    LIT64( 0x007FFFFFFFFFFFFF ),
+    LIT64( 0x003FFFFFFFFFFFFF ),
+    LIT64( 0x001FFFFFFFFFFFFF ),
+    LIT64( 0x000FFFFFFFFFFFFF ),
+    LIT64( 0x0007FFFFFFFFFFFF ),
+    LIT64( 0x0003FFFFFFFFFFFF ),
+    LIT64( 0x0001FFFFFFFFFFFF ),
+    LIT64( 0x0000FFFFFFFFFFFF ),
+    LIT64( 0x00007FFFFFFFFFFF ),
+    LIT64( 0x00003FFFFFFFFFFF ),
+    LIT64( 0x00001FFFFFFFFFFF ),
+    LIT64( 0x00000FFFFFFFFFFF ),
+    LIT64( 0x000007FFFFFFFFFF ),
+    LIT64( 0x000003FFFFFFFFFF ),
+    LIT64( 0x000001FFFFFFFFFF ),
+    LIT64( 0x000000FFFFFFFFFF ),
+    LIT64( 0x0000007FFFFFFFFF ),
+    LIT64( 0x0000003FFFFFFFFF ),
+    LIT64( 0x0000001FFFFFFFFF ),
+    LIT64( 0x0000000FFFFFFFFF ),
+    LIT64( 0x00000007FFFFFFFF ),
+    LIT64( 0x00000003FFFFFFFF ),
+    LIT64( 0x00000001FFFFFFFF ),
+    LIT64( 0x00000000FFFFFFFF ),
+    LIT64( 0x000000007FFFFFFF ),
+    LIT64( 0x000000003FFFFFFF ),
+    LIT64( 0x000000001FFFFFFF ),
+    LIT64( 0x000000000FFFFFFF ),
+    LIT64( 0x0000000007FFFFFF ),
+    LIT64( 0x0000000003FFFFFF ),
+    LIT64( 0x0000000001FFFFFF ),
+    LIT64( 0x0000000000FFFFFF ),
+    LIT64( 0x00000000007FFFFF ),
+    LIT64( 0x00000000003FFFFF ),
+    LIT64( 0x00000000001FFFFF ),
+    LIT64( 0x00000000000FFFFF ),
+    LIT64( 0x000000000007FFFF ),
+    LIT64( 0x000000000003FFFF ),
+    LIT64( 0x000000000001FFFF ),
+    LIT64( 0x000000000000FFFF ),
+    LIT64( 0x0000000000007FFF ),
+    LIT64( 0x0000000000003FFF ),
+    LIT64( 0x0000000000001FFF ),
+    LIT64( 0x0000000000000FFF ),
+    LIT64( 0x00000000000007FF ),
+    LIT64( 0x00000000000003FF ),
+    LIT64( 0x00000000000001FF ),
+    LIT64( 0x00000000000000FF ),
+    LIT64( 0x000000000000007F ),
+    LIT64( 0x000000000000003F ),
+    LIT64( 0x000000000000001F ),
+    LIT64( 0x000000000000000F ),
+    LIT64( 0x0000000000000007 ),
+    LIT64( 0x0000000000000003 )
+};
+
+static int64 int64NextP1( sequenceT *sequencePtr )
+{
+    uint8 termNum;
+    int64 z;
+
+    termNum = sequencePtr->term1Num;
+    z = int64P1[ termNum ];
+    ++termNum;
+    if ( int64NumP1 <= termNum ) {
+        termNum = 0;
+        sequencePtr->done = TRUE;
+    }
+    sequencePtr->term1Num = termNum;
+    return (sbits64) z;
+
+}
+
+static const int64 int64NumP2 = ( int64NumP1 * int64NumP1 + int64NumP1 ) / 2;
+
+static int64 int64NextP2( sequenceT *sequencePtr )
+{
+    uint8 term1Num, term2Num;
+    int64 z;
+
+    term2Num = sequencePtr->term2Num;
+    term1Num = sequencePtr->term1Num;
+    z = int64P1[ term1Num ] + int64P1[ term2Num ];
+    ++term2Num;
+    if ( int64NumP1 <= term2Num ) {
+        ++term1Num;
+        if ( int64NumP1 <= term1Num ) {
+            term1Num = 0;
+            sequencePtr->done = TRUE;
+        }
+        term2Num = term1Num;
+        sequencePtr->term1Num = term1Num;
+    }
+    sequencePtr->term2Num = term2Num;
+    return (sbits64) z;
+
+}
+
+static int64 int64RandomP3( void )
+{
+
+    return
+        (sbits64) (
+              int64P1[ randomUint8() % int64NumP1 ]
+            + int64P1[ randomUint8() % int64NumP1 ]
+            + int64P1[ randomUint8() % int64NumP1 ]
+        );
+
+}
+
+enum {
+    int64NumPInfWeightMasks = 61
+};
+
+static const uint64 int64PInfWeightMasks[ int64NumPInfWeightMasks ] = {
+    LIT64( 0xFFFFFFFFFFFFFFFF ),
+    LIT64( 0x7FFFFFFFFFFFFFFF ),
+    LIT64( 0x3FFFFFFFFFFFFFFF ),
+    LIT64( 0x1FFFFFFFFFFFFFFF ),
+    LIT64( 0x0FFFFFFFFFFFFFFF ),
+    LIT64( 0x07FFFFFFFFFFFFFF ),
+    LIT64( 0x03FFFFFFFFFFFFFF ),
+    LIT64( 0x01FFFFFFFFFFFFFF ),
+    LIT64( 0x00FFFFFFFFFFFFFF ),
+    LIT64( 0x007FFFFFFFFFFFFF ),
+    LIT64( 0x003FFFFFFFFFFFFF ),
+    LIT64( 0x001FFFFFFFFFFFFF ),
+    LIT64( 0x000FFFFFFFFFFFFF ),
+    LIT64( 0x0007FFFFFFFFFFFF ),
+    LIT64( 0x0003FFFFFFFFFFFF ),
+    LIT64( 0x0001FFFFFFFFFFFF ),
+    LIT64( 0x0000FFFFFFFFFFFF ),
+    LIT64( 0x00007FFFFFFFFFFF ),
+    LIT64( 0x00003FFFFFFFFFFF ),
+    LIT64( 0x00001FFFFFFFFFFF ),
+    LIT64( 0x00000FFFFFFFFFFF ),
+    LIT64( 0x000007FFFFFFFFFF ),
+    LIT64( 0x000003FFFFFFFFFF ),
+    LIT64( 0x000001FFFFFFFFFF ),
+    LIT64( 0x000000FFFFFFFFFF ),
+    LIT64( 0x0000007FFFFFFFFF ),
+    LIT64( 0x0000003FFFFFFFFF ),
+    LIT64( 0x0000001FFFFFFFFF ),
+    LIT64( 0x0000000FFFFFFFFF ),
+    LIT64( 0x00000007FFFFFFFF ),
+    LIT64( 0x00000003FFFFFFFF ),
+    LIT64( 0x00000001FFFFFFFF ),
+    LIT64( 0x00000000FFFFFFFF ),
+    LIT64( 0x000000007FFFFFFF ),
+    LIT64( 0x000000003FFFFFFF ),
+    LIT64( 0x000000001FFFFFFF ),
+    LIT64( 0x000000000FFFFFFF ),
+    LIT64( 0x0000000007FFFFFF ),
+    LIT64( 0x0000000003FFFFFF ),
+    LIT64( 0x0000000001FFFFFF ),
+    LIT64( 0x0000000000FFFFFF ),
+    LIT64( 0x00000000007FFFFF ),
+    LIT64( 0x00000000003FFFFF ),
+    LIT64( 0x00000000001FFFFF ),
+    LIT64( 0x00000000000FFFFF ),
+    LIT64( 0x000000000007FFFF ),
+    LIT64( 0x000000000003FFFF ),
+    LIT64( 0x000000000001FFFF ),
+    LIT64( 0x000000000000FFFF ),
+    LIT64( 0x0000000000007FFF ),
+    LIT64( 0x0000000000003FFF ),
+    LIT64( 0x0000000000001FFF ),
+    LIT64( 0x0000000000000FFF ),
+    LIT64( 0x00000000000007FF ),
+    LIT64( 0x00000000000003FF ),
+    LIT64( 0x00000000000001FF ),
+    LIT64( 0x00000000000000FF ),
+    LIT64( 0x000000000000007F ),
+    LIT64( 0x000000000000003F ),
+    LIT64( 0x000000000000001F ),
+    LIT64( 0x000000000000000F )
+};
+
+static const uint64 int64PInfWeightOffsets[ int64NumPInfWeightMasks ] = {
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0xC000000000000000 ),
+    LIT64( 0xE000000000000000 ),
+    LIT64( 0xF000000000000000 ),
+    LIT64( 0xF800000000000000 ),
+    LIT64( 0xFC00000000000000 ),
+    LIT64( 0xFE00000000000000 ),
+    LIT64( 0xFF00000000000000 ),
+    LIT64( 0xFF80000000000000 ),
+    LIT64( 0xFFC0000000000000 ),
+    LIT64( 0xFFE0000000000000 ),
+    LIT64( 0xFFF0000000000000 ),
+    LIT64( 0xFFF8000000000000 ),
+    LIT64( 0xFFFC000000000000 ),
+    LIT64( 0xFFFE000000000000 ),
+    LIT64( 0xFFFF000000000000 ),
+    LIT64( 0xFFFF800000000000 ),
+    LIT64( 0xFFFFC00000000000 ),
+    LIT64( 0xFFFFE00000000000 ),
+    LIT64( 0xFFFFF00000000000 ),
+    LIT64( 0xFFFFF80000000000 ),
+    LIT64( 0xFFFFFC0000000000 ),
+    LIT64( 0xFFFFFE0000000000 ),
+    LIT64( 0xFFFFFF0000000000 ),
+    LIT64( 0xFFFFFF8000000000 ),
+    LIT64( 0xFFFFFFC000000000 ),
+    LIT64( 0xFFFFFFE000000000 ),
+    LIT64( 0xFFFFFFF000000000 ),
+    LIT64( 0xFFFFFFF800000000 ),
+    LIT64( 0xFFFFFFFC00000000 ),
+    LIT64( 0xFFFFFFFE00000000 ),
+    LIT64( 0xFFFFFFFF00000000 ),
+    LIT64( 0xFFFFFFFF80000000 ),
+    LIT64( 0xFFFFFFFFC0000000 ),
+    LIT64( 0xFFFFFFFFE0000000 ),
+    LIT64( 0xFFFFFFFFF0000000 ),
+    LIT64( 0xFFFFFFFFF8000000 ),
+    LIT64( 0xFFFFFFFFFC000000 ),
+    LIT64( 0xFFFFFFFFFE000000 ),
+    LIT64( 0xFFFFFFFFFF000000 ),
+    LIT64( 0xFFFFFFFFFF800000 ),
+    LIT64( 0xFFFFFFFFFFC00000 ),
+    LIT64( 0xFFFFFFFFFFE00000 ),
+    LIT64( 0xFFFFFFFFFFF00000 ),
+    LIT64( 0xFFFFFFFFFFF80000 ),
+    LIT64( 0xFFFFFFFFFFFC0000 ),
+    LIT64( 0xFFFFFFFFFFFE0000 ),
+    LIT64( 0xFFFFFFFFFFFF0000 ),
+    LIT64( 0xFFFFFFFFFFFF8000 ),
+    LIT64( 0xFFFFFFFFFFFFC000 ),
+    LIT64( 0xFFFFFFFFFFFFE000 ),
+    LIT64( 0xFFFFFFFFFFFFF000 ),
+    LIT64( 0xFFFFFFFFFFFFF800 ),
+    LIT64( 0xFFFFFFFFFFFFFC00 ),
+    LIT64( 0xFFFFFFFFFFFFFE00 ),
+    LIT64( 0xFFFFFFFFFFFFFF00 ),
+    LIT64( 0xFFFFFFFFFFFFFF80 ),
+    LIT64( 0xFFFFFFFFFFFFFFC0 ),
+    LIT64( 0xFFFFFFFFFFFFFFE0 ),
+    LIT64( 0xFFFFFFFFFFFFFFF0 ),
+    LIT64( 0xFFFFFFFFFFFFFFF8 )
+};
+
+static int64 int64RandomPInf( void )
+{
+    int8 weightMaskNum;
+
+    weightMaskNum = randomUint8() % int64NumPInfWeightMasks;
+    return
+        (sbits64) (
+              ( randomUint64() & int64PInfWeightMasks[ weightMaskNum ] )
+            + int64PInfWeightOffsets[ weightMaskNum ]
+        );
+
+}
+
+#endif
+
+enum {
+    float32NumQIn  = 22,
+    float32NumQOut = 50,
+    float32NumP1   =  4,
+    float32NumP2   = 88
+};
+
+static const uint32 float32QIn[ float32NumQIn ] = {
+    0x00000000,		/* positive, subnormal		*/
+    0x00800000,		/* positive, -126		*/
+    0x33800000,		/* positive,  -24		*/
+    0x3E800000,		/* positive,   -2		*/
+    0x3F000000,		/* positive,   -1		*/
+    0x3F800000,		/* positive,    0		*/
+    0x40000000,		/* positive,    1		*/
+    0x40800000,		/* positive,    2		*/
+    0x4B800000,		/* positive,   24		*/
+    0x7F000000,		/* positive,  127		*/
+    0x7F800000,		/* positive, infinity or NaN	*/
+    0x80000000,		/* negative, subnormal		*/
+    0x80800000,		/* negative, -126		*/
+    0xB3800000,		/* negative,  -24		*/
+    0xBE800000,		/* negative,   -2		*/
+    0xBF000000,		/* negative,   -1		*/
+    0xBF800000,		/* negative,    0		*/
+    0xC0000000,		/* negative,    1		*/
+    0xC0800000,		/* negative,    2		*/
+    0xCB800000,		/* negative,   24		*/
+    0xFE800000,		/* negative,  126		*/
+    0xFF800000		/* negative, infinity or NaN	*/
+};
+
+static const uint32 float32QOut[ float32NumQOut ] = {
+    0x00000000,		/* positive, subnormal		*/
+    0x00800000,		/* positive, -126		*/
+    0x01000000,		/* positive, -125		*/
+    0x33800000,		/* positive,  -24		*/
+    0x3D800000,		/* positive,   -4		*/
+    0x3E000000,		/* positive,   -3		*/
+    0x3E800000,		/* positive,   -2		*/
+    0x3F000000,		/* positive,   -1		*/
+    0x3F800000,		/* positive,    0		*/
+    0x40000000,		/* positive,    1		*/
+    0x40800000,		/* positive,    2		*/
+    0x41000000,		/* positive,    3		*/
+    0x41800000,		/* positive,    4		*/
+    0x4B800000,		/* positive,   24		*/
+    0x4E000000,		/* positive,   29		*/
+    0x4E800000,		/* positive,   30		*/
+    0x4F000000,		/* positive,   31		*/
+    0x4F800000,		/* positive,   32		*/
+    0x5E000000,		/* positive,   61		*/
+    0x5E800000,		/* positive,   62		*/
+    0x5F000000,		/* positive,   63		*/
+    0x5F800000,		/* positive,   64		*/
+    0x7E800000,		/* positive,  126		*/
+    0x7F000000,		/* positive,  127		*/
+    0x7F800000,		/* positive, infinity or NaN	*/
+    0x80000000,		/* negative, subnormal		*/
+    0x80800000,		/* negative, -126		*/
+    0x81000000,		/* negative, -125		*/
+    0xB3800000,		/* negative,  -24		*/
+    0xBD800000,		/* negative,   -4		*/
+    0xBE000000,		/* negative,   -3		*/
+    0xBE800000,		/* negative,   -2		*/
+    0xBF000000,		/* negative,   -1		*/
+    0xBF800000,		/* negative,    0		*/
+    0xC0000000,		/* negative,    1		*/
+    0xC0800000,		/* negative,    2		*/
+    0xC1000000,		/* negative,    3		*/
+    0xC1800000,		/* negative,    4		*/
+    0xCB800000,		/* negative,   24		*/
+    0xCE000000,		/* negative,   29		*/
+    0xCE800000,		/* negative,   30		*/
+    0xCF000000,		/* negative,   31		*/
+    0xCF800000,		/* negative,   32		*/
+    0xDE000000,		/* negative,   61		*/
+    0xDE800000,		/* negative,   62		*/
+    0xDF000000,		/* negative,   63		*/
+    0xDF800000,		/* negative,   64		*/
+    0xFE800000,		/* negative,  126		*/
+    0xFF000000,		/* negative,  127		*/
+    0xFF800000		/* negative, infinity or NaN	*/
+};
+
+static const uint32 float32P1[ float32NumP1 ] = {
+    0x00000000,
+    0x00000001,
+    0x007FFFFF,
+    0x007FFFFE
+};
+
+static const uint32 float32P2[ float32NumP2 ] = {
+    0x00000000,
+    0x00000001,
+    0x00000002,
+    0x00000004,
+    0x00000008,
+    0x00000010,
+    0x00000020,
+    0x00000040,
+    0x00000080,
+    0x00000100,
+    0x00000200,
+    0x00000400,
+    0x00000800,
+    0x00001000,
+    0x00002000,
+    0x00004000,
+    0x00008000,
+    0x00010000,
+    0x00020000,
+    0x00040000,
+    0x00080000,
+    0x00100000,
+    0x00200000,
+    0x00400000,
+    0x00600000,
+    0x00700000,
+    0x00780000,
+    0x007C0000,
+    0x007E0000,
+    0x007F0000,
+    0x007F8000,
+    0x007FC000,
+    0x007FE000,
+    0x007FF000,
+    0x007FF800,
+    0x007FFC00,
+    0x007FFE00,
+    0x007FFF00,
+    0x007FFF80,
+    0x007FFFC0,
+    0x007FFFE0,
+    0x007FFFF0,
+    0x007FFFF8,
+    0x007FFFFC,
+    0x007FFFFE,
+    0x007FFFFF,
+    0x007FFFFD,
+    0x007FFFFB,
+    0x007FFFF7,
+    0x007FFFEF,
+    0x007FFFDF,
+    0x007FFFBF,
+    0x007FFF7F,
+    0x007FFEFF,
+    0x007FFDFF,
+    0x007FFBFF,
+    0x007FF7FF,
+    0x007FEFFF,
+    0x007FDFFF,
+    0x007FBFFF,
+    0x007F7FFF,
+    0x007EFFFF,
+    0x007DFFFF,
+    0x007BFFFF,
+    0x0077FFFF,
+    0x006FFFFF,
+    0x005FFFFF,
+    0x003FFFFF,
+    0x001FFFFF,
+    0x000FFFFF,
+    0x0007FFFF,
+    0x0003FFFF,
+    0x0001FFFF,
+    0x0000FFFF,
+    0x00007FFF,
+    0x00003FFF,
+    0x00001FFF,
+    0x00000FFF,
+    0x000007FF,
+    0x000003FF,
+    0x000001FF,
+    0x000000FF,
+    0x0000007F,
+    0x0000003F,
+    0x0000001F,
+    0x0000000F,
+    0x00000007,
+    0x00000003
+};
+
+static const uint32 float32NumQInP1 = float32NumQIn * float32NumP1;
+static const uint32 float32NumQOutP1 = float32NumQOut * float32NumP1;
+
+static float32 float32NextQInP1( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float32 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z = float32QIn[ expNum ] | float32P1[ sigNum ];
+    ++sigNum;
+    if ( float32NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float32NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float32 float32NextQOutP1( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float32 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z = float32QOut[ expNum ] | float32P1[ sigNum ];
+    ++sigNum;
+    if ( float32NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float32NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static const uint32 float32NumQInP2 = float32NumQIn * float32NumP2;
+static const uint32 float32NumQOutP2 = float32NumQOut * float32NumP2;
+
+static float32 float32NextQInP2( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float32 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z = float32QIn[ expNum ] | float32P2[ sigNum ];
+    ++sigNum;
+    if ( float32NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float32NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float32 float32NextQOutP2( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float32 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z = float32QOut[ expNum ] | float32P2[ sigNum ];
+    ++sigNum;
+    if ( float32NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float32NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float32 float32RandomQOutP3( void )
+{
+
+    return
+          float32QOut[ randomUint8() % float32NumQOut ]
+        | (   (   float32P2[ randomUint8() % float32NumP2 ]
+                + float32P2[ randomUint8() % float32NumP2 ] )
+            & 0x007FFFFF );
+
+}
+
+static float32 float32RandomQOutPInf( void )
+{
+
+    return
+          float32QOut[ randomUint8() % float32NumQOut ]
+        | ( randomUint32() & 0x007FFFFF );
+
+}
+
+enum {
+    float32NumQInfWeightMasks = 7
+};
+
+static const uint32 float32QInfWeightMasks[ float32NumQInfWeightMasks ] = {
+    0x7F800000,
+    0x7F800000,
+    0x3F800000,
+    0x1F800000,
+    0x0F800000,
+    0x07800000,
+    0x03800000
+};
+
+static const uint32 float32QInfWeightOffsets[ float32NumQInfWeightMasks ] = {
+    0x00000000,
+    0x00000000,
+    0x20000000,
+    0x30000000,
+    0x38000000,
+    0x3C000000,
+    0x3E000000
+};
+
+static float32 float32RandomQInfP3( void )
+{
+    int8 weightMaskNum;
+
+    weightMaskNum = randomUint8() % float32NumQInfWeightMasks;
+    return
+          ( ( (uint32) ( randomUint8() & 1 ) )<<31 )
+        | (   (   ( ( (uint32) ( randomUint16() & 0x1FF ) )<<23 )
+                & float32QInfWeightMasks[ weightMaskNum ] )
+            + float32QInfWeightOffsets[ weightMaskNum ]
+          )
+        | (   (   float32P2[ randomUint8() % float32NumP2 ]
+                + float32P2[ randomUint8() % float32NumP2 ] )
+            & 0x007FFFFF );
+
+}
+
+static float32 float32RandomQInfPInf( void )
+{
+    int8 weightMaskNum;
+
+    weightMaskNum = randomUint8() % float32NumQInfWeightMasks;
+    return
+          ( ( (uint32) ( randomUint8() & 1 ) )<<31 )
+        | (   (   ( ( (uint32) ( randomUint16() & 0x1FF ) )<<23 )
+                & float32QInfWeightMasks[ weightMaskNum ] )
+            + float32QInfWeightOffsets[ weightMaskNum ]
+          )
+        | ( randomUint32() & 0x007FFFFF );
+
+}
+
+static float32 float32Random( void )
+{
+
+    switch ( randomUint8() & 7 ) {
+     case 0:
+     case 1:
+     case 2:
+        return float32RandomQOutP3();
+     case 3:
+        return float32RandomQOutPInf();
+     case 4:
+     case 5:
+     case 6:
+        return float32RandomQInfP3();
+     case 7:
+        return float32RandomQInfPInf();
+    }
+
+}
+
+#ifdef BITS64
+#define SETFLOAT64( z, zHigh, zLow ) z = ( ( (float64) zHigh )<<32 ) | zLow
+#else
+#define SETFLOAT64( z, zHigh, zLow ) z.low = zLow; z.high = zHigh
+#endif
+
+enum {
+    float64NumQIn  =  22,
+    float64NumQOut =  64,
+    float64NumP1   =   4,
+    float64NumP2   = 204
+};
+
+static const uint32 float64QIn[ float64NumQIn ] = {
+    0x00000000,		/* positive, subnormal		*/
+    0x00100000,		/* positive, -1022		*/
+    0x3CA00000,		/* positive,   -53		*/
+    0x3FD00000,		/* positive,    -2		*/
+    0x3FE00000,		/* positive,    -1		*/
+    0x3FF00000,		/* positive,     0		*/
+    0x40000000,		/* positive,     1		*/
+    0x40100000,		/* positive,     2		*/
+    0x43400000,		/* positive,    53		*/
+    0x7FE00000,		/* positive,  1023		*/
+    0x7FF00000,		/* positive, infinity or NaN	*/
+    0x80000000,		/* negative, subnormal		*/
+    0x80100000,		/* negative, -1022		*/
+    0xBCA00000,		/* negative,   -53		*/
+    0xBFD00000,		/* negative,    -2		*/
+    0xBFE00000,		/* negative,    -1		*/
+    0xBFF00000,		/* negative,     0		*/
+    0xC0000000,		/* negative,     1		*/
+    0xC0100000,		/* negative,     2		*/
+    0xC3400000,		/* negative,    53		*/
+    0xFFE00000,		/* negative,  1023		*/
+    0xFFF00000		/* negative, infinity or NaN	*/
+};
+
+static const uint32 float64QOut[ float64NumQOut ] = {
+    0x00000000,		/* positive, subnormal		*/
+    0x00100000,		/* positive, -1022		*/
+    0x00200000,		/* positive, -1021		*/
+    0x37E00000,		/* positive,  -129		*/
+    0x37F00000,		/* positive,  -128		*/
+    0x38000000,		/* positive,  -127		*/
+    0x38100000,		/* positive,  -126		*/
+    0x3CA00000,		/* positive,   -53		*/
+    0x3FB00000,		/* positive,    -4		*/
+    0x3FC00000,		/* positive,    -3		*/
+    0x3FD00000,		/* positive,    -2		*/
+    0x3FE00000,		/* positive,    -1		*/
+    0x3FF00000,		/* positive,     0		*/
+    0x40000000,		/* positive,     1		*/
+    0x40100000,		/* positive,     2		*/
+    0x40200000,		/* positive,     3		*/
+    0x40300000,		/* positive,     4		*/
+    0x41C00000,		/* positive,    29		*/
+    0x41D00000,		/* positive,    30		*/
+    0x41E00000,		/* positive,    31		*/
+    0x41F00000,		/* positive,    32		*/
+    0x43400000,		/* positive,    53		*/
+    0x43C00000,		/* positive,    61		*/
+    0x43D00000,		/* positive,    62		*/
+    0x43E00000,		/* positive,    63		*/
+    0x43F00000,		/* positive,    64		*/
+    0x47E00000,		/* positive,   127		*/
+    0x47F00000,		/* positive,   128		*/
+    0x48000000,		/* positive,   129		*/
+    0x7FD00000,		/* positive,  1022		*/
+    0x7FE00000,		/* positive,  1023		*/
+    0x7FF00000,		/* positive, infinity or NaN	*/
+    0x80000000,		/* negative, subnormal		*/
+    0x80100000,		/* negative, -1022		*/
+    0x80200000,		/* negative, -1021		*/
+    0xB7E00000,		/* negative,  -129		*/
+    0xB7F00000,		/* negative,  -128		*/
+    0xB8000000,		/* negative,  -127		*/
+    0xB8100000,		/* negative,  -126		*/
+    0xBCA00000,		/* negative,   -53		*/
+    0xBFB00000,		/* negative,    -4		*/
+    0xBFC00000,		/* negative,    -3		*/
+    0xBFD00000,		/* negative,    -2		*/
+    0xBFE00000,		/* negative,    -1		*/
+    0xBFF00000,		/* negative,     0		*/
+    0xC0000000,		/* negative,     1		*/
+    0xC0100000,		/* negative,     2		*/
+    0xC0200000,		/* negative,     3		*/
+    0xC0300000,		/* negative,     4		*/
+    0xC1C00000,		/* negative,    29		*/
+    0xC1D00000,		/* negative,    30		*/
+    0xC1E00000,		/* negative,    31		*/
+    0xC1F00000,		/* negative,    32		*/
+    0xC3400000,		/* negative,    53		*/
+    0xC3C00000,		/* negative,    61		*/
+    0xC3D00000,		/* negative,    62		*/
+    0xC3E00000,		/* negative,    63		*/
+    0xC3F00000,		/* negative,    64		*/
+    0xC7E00000,		/* negative,   127		*/
+    0xC7F00000,		/* negative,   128		*/
+    0xC8000000,		/* negative,   129		*/
+    0xFFD00000,		/* negative,  1022		*/
+    0xFFE00000,		/* negative,  1023		*/
+    0xFFF00000		/* negative, infinity or NaN	*/
+};
+
+static const struct { bits32 high, low; } float64P1[ float64NumP1 ] = {
+    { 0x00000000, 0x00000000 },
+    { 0x00000000, 0x00000001 },
+    { 0x000FFFFF, 0xFFFFFFFF },
+    { 0x000FFFFF, 0xFFFFFFFE }
+};
+
+static const struct { bits32 high, low; } float64P2[ float64NumP2 ] = {
+    { 0x00000000, 0x00000000 },
+    { 0x00000000, 0x00000001 },
+    { 0x00000000, 0x00000002 },
+    { 0x00000000, 0x00000004 },
+    { 0x00000000, 0x00000008 },
+    { 0x00000000, 0x00000010 },
+    { 0x00000000, 0x00000020 },
+    { 0x00000000, 0x00000040 },
+    { 0x00000000, 0x00000080 },
+    { 0x00000000, 0x00000100 },
+    { 0x00000000, 0x00000200 },
+    { 0x00000000, 0x00000400 },
+    { 0x00000000, 0x00000800 },
+    { 0x00000000, 0x00001000 },
+    { 0x00000000, 0x00002000 },
+    { 0x00000000, 0x00004000 },
+    { 0x00000000, 0x00008000 },
+    { 0x00000000, 0x00010000 },
+    { 0x00000000, 0x00020000 },
+    { 0x00000000, 0x00040000 },
+    { 0x00000000, 0x00080000 },
+    { 0x00000000, 0x00100000 },
+    { 0x00000000, 0x00200000 },
+    { 0x00000000, 0x00400000 },
+    { 0x00000000, 0x00800000 },
+    { 0x00000000, 0x01000000 },
+    { 0x00000000, 0x02000000 },
+    { 0x00000000, 0x04000000 },
+    { 0x00000000, 0x08000000 },
+    { 0x00000000, 0x10000000 },
+    { 0x00000000, 0x20000000 },
+    { 0x00000000, 0x40000000 },
+    { 0x00000000, 0x80000000 },
+    { 0x00000001, 0x00000000 },
+    { 0x00000002, 0x00000000 },
+    { 0x00000004, 0x00000000 },
+    { 0x00000008, 0x00000000 },
+    { 0x00000010, 0x00000000 },
+    { 0x00000020, 0x00000000 },
+    { 0x00000040, 0x00000000 },
+    { 0x00000080, 0x00000000 },
+    { 0x00000100, 0x00000000 },
+    { 0x00000200, 0x00000000 },
+    { 0x00000400, 0x00000000 },
+    { 0x00000800, 0x00000000 },
+    { 0x00001000, 0x00000000 },
+    { 0x00002000, 0x00000000 },
+    { 0x00004000, 0x00000000 },
+    { 0x00008000, 0x00000000 },
+    { 0x00010000, 0x00000000 },
+    { 0x00020000, 0x00000000 },
+    { 0x00040000, 0x00000000 },
+    { 0x00080000, 0x00000000 },
+    { 0x000C0000, 0x00000000 },
+    { 0x000E0000, 0x00000000 },
+    { 0x000F0000, 0x00000000 },
+    { 0x000F8000, 0x00000000 },
+    { 0x000FC000, 0x00000000 },
+    { 0x000FE000, 0x00000000 },
+    { 0x000FF000, 0x00000000 },
+    { 0x000FF800, 0x00000000 },
+    { 0x000FFC00, 0x00000000 },
+    { 0x000FFE00, 0x00000000 },
+    { 0x000FFF00, 0x00000000 },
+    { 0x000FFF80, 0x00000000 },
+    { 0x000FFFC0, 0x00000000 },
+    { 0x000FFFE0, 0x00000000 },
+    { 0x000FFFF0, 0x00000000 },
+    { 0x000FFFF8, 0x00000000 },
+    { 0x000FFFFC, 0x00000000 },
+    { 0x000FFFFE, 0x00000000 },
+    { 0x000FFFFF, 0x00000000 },
+    { 0x000FFFFF, 0x80000000 },
+    { 0x000FFFFF, 0xC0000000 },
+    { 0x000FFFFF, 0xE0000000 },
+    { 0x000FFFFF, 0xF0000000 },
+    { 0x000FFFFF, 0xF8000000 },
+    { 0x000FFFFF, 0xFC000000 },
+    { 0x000FFFFF, 0xFE000000 },
+    { 0x000FFFFF, 0xFF000000 },
+    { 0x000FFFFF, 0xFF800000 },
+    { 0x000FFFFF, 0xFFC00000 },
+    { 0x000FFFFF, 0xFFE00000 },
+    { 0x000FFFFF, 0xFFF00000 },
+    { 0x000FFFFF, 0xFFF80000 },
+    { 0x000FFFFF, 0xFFFC0000 },
+    { 0x000FFFFF, 0xFFFE0000 },
+    { 0x000FFFFF, 0xFFFF0000 },
+    { 0x000FFFFF, 0xFFFF8000 },
+    { 0x000FFFFF, 0xFFFFC000 },
+    { 0x000FFFFF, 0xFFFFE000 },
+    { 0x000FFFFF, 0xFFFFF000 },
+    { 0x000FFFFF, 0xFFFFF800 },
+    { 0x000FFFFF, 0xFFFFFC00 },
+    { 0x000FFFFF, 0xFFFFFE00 },
+    { 0x000FFFFF, 0xFFFFFF00 },
+    { 0x000FFFFF, 0xFFFFFF80 },
+    { 0x000FFFFF, 0xFFFFFFC0 },
+    { 0x000FFFFF, 0xFFFFFFE0 },
+    { 0x000FFFFF, 0xFFFFFFF0 },
+    { 0x000FFFFF, 0xFFFFFFF8 },
+    { 0x000FFFFF, 0xFFFFFFFC },
+    { 0x000FFFFF, 0xFFFFFFFE },
+    { 0x000FFFFF, 0xFFFFFFFF },
+    { 0x000FFFFF, 0xFFFFFFFD },
+    { 0x000FFFFF, 0xFFFFFFFB },
+    { 0x000FFFFF, 0xFFFFFFF7 },
+    { 0x000FFFFF, 0xFFFFFFEF },
+    { 0x000FFFFF, 0xFFFFFFDF },
+    { 0x000FFFFF, 0xFFFFFFBF },
+    { 0x000FFFFF, 0xFFFFFF7F },
+    { 0x000FFFFF, 0xFFFFFEFF },
+    { 0x000FFFFF, 0xFFFFFDFF },
+    { 0x000FFFFF, 0xFFFFFBFF },
+    { 0x000FFFFF, 0xFFFFF7FF },
+    { 0x000FFFFF, 0xFFFFEFFF },
+    { 0x000FFFFF, 0xFFFFDFFF },
+    { 0x000FFFFF, 0xFFFFBFFF },
+    { 0x000FFFFF, 0xFFFF7FFF },
+    { 0x000FFFFF, 0xFFFEFFFF },
+    { 0x000FFFFF, 0xFFFDFFFF },
+    { 0x000FFFFF, 0xFFFBFFFF },
+    { 0x000FFFFF, 0xFFF7FFFF },
+    { 0x000FFFFF, 0xFFEFFFFF },
+    { 0x000FFFFF, 0xFFDFFFFF },
+    { 0x000FFFFF, 0xFFBFFFFF },
+    { 0x000FFFFF, 0xFF7FFFFF },
+    { 0x000FFFFF, 0xFEFFFFFF },
+    { 0x000FFFFF, 0xFDFFFFFF },
+    { 0x000FFFFF, 0xFBFFFFFF },
+    { 0x000FFFFF, 0xF7FFFFFF },
+    { 0x000FFFFF, 0xEFFFFFFF },
+    { 0x000FFFFF, 0xDFFFFFFF },
+    { 0x000FFFFF, 0xBFFFFFFF },
+    { 0x000FFFFF, 0x7FFFFFFF },
+    { 0x000FFFFE, 0xFFFFFFFF },
+    { 0x000FFFFD, 0xFFFFFFFF },
+    { 0x000FFFFB, 0xFFFFFFFF },
+    { 0x000FFFF7, 0xFFFFFFFF },
+    { 0x000FFFEF, 0xFFFFFFFF },
+    { 0x000FFFDF, 0xFFFFFFFF },
+    { 0x000FFFBF, 0xFFFFFFFF },
+    { 0x000FFF7F, 0xFFFFFFFF },
+    { 0x000FFEFF, 0xFFFFFFFF },
+    { 0x000FFDFF, 0xFFFFFFFF },
+    { 0x000FFBFF, 0xFFFFFFFF },
+    { 0x000FF7FF, 0xFFFFFFFF },
+    { 0x000FEFFF, 0xFFFFFFFF },
+    { 0x000FDFFF, 0xFFFFFFFF },
+    { 0x000FBFFF, 0xFFFFFFFF },
+    { 0x000F7FFF, 0xFFFFFFFF },
+    { 0x000EFFFF, 0xFFFFFFFF },
+    { 0x000DFFFF, 0xFFFFFFFF },
+    { 0x000BFFFF, 0xFFFFFFFF },
+    { 0x0007FFFF, 0xFFFFFFFF },
+    { 0x0003FFFF, 0xFFFFFFFF },
+    { 0x0001FFFF, 0xFFFFFFFF },
+    { 0x0000FFFF, 0xFFFFFFFF },
+    { 0x00007FFF, 0xFFFFFFFF },
+    { 0x00003FFF, 0xFFFFFFFF },
+    { 0x00001FFF, 0xFFFFFFFF },
+    { 0x00000FFF, 0xFFFFFFFF },
+    { 0x000007FF, 0xFFFFFFFF },
+    { 0x000003FF, 0xFFFFFFFF },
+    { 0x000001FF, 0xFFFFFFFF },
+    { 0x000000FF, 0xFFFFFFFF },
+    { 0x0000007F, 0xFFFFFFFF },
+    { 0x0000003F, 0xFFFFFFFF },
+    { 0x0000001F, 0xFFFFFFFF },
+    { 0x0000000F, 0xFFFFFFFF },
+    { 0x00000007, 0xFFFFFFFF },
+    { 0x00000003, 0xFFFFFFFF },
+    { 0x00000001, 0xFFFFFFFF },
+    { 0x00000000, 0xFFFFFFFF },
+    { 0x00000000, 0x7FFFFFFF },
+    { 0x00000000, 0x3FFFFFFF },
+    { 0x00000000, 0x1FFFFFFF },
+    { 0x00000000, 0x0FFFFFFF },
+    { 0x00000000, 0x07FFFFFF },
+    { 0x00000000, 0x03FFFFFF },
+    { 0x00000000, 0x01FFFFFF },
+    { 0x00000000, 0x00FFFFFF },
+    { 0x00000000, 0x007FFFFF },
+    { 0x00000000, 0x003FFFFF },
+    { 0x00000000, 0x001FFFFF },
+    { 0x00000000, 0x000FFFFF },
+    { 0x00000000, 0x0007FFFF },
+    { 0x00000000, 0x0003FFFF },
+    { 0x00000000, 0x0001FFFF },
+    { 0x00000000, 0x0000FFFF },
+    { 0x00000000, 0x00007FFF },
+    { 0x00000000, 0x00003FFF },
+    { 0x00000000, 0x00001FFF },
+    { 0x00000000, 0x00000FFF },
+    { 0x00000000, 0x000007FF },
+    { 0x00000000, 0x000003FF },
+    { 0x00000000, 0x000001FF },
+    { 0x00000000, 0x000000FF },
+    { 0x00000000, 0x0000007F },
+    { 0x00000000, 0x0000003F },
+    { 0x00000000, 0x0000001F },
+    { 0x00000000, 0x0000000F },
+    { 0x00000000, 0x00000007 },
+    { 0x00000000, 0x00000003 }
+};
+
+static const uint32 float64NumQInP1 = float64NumQIn * float64NumP1;
+static const uint32 float64NumQOutP1 = float64NumQOut * float64NumP1;
+
+static float64 float64NextQInP1( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float64 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    SETFLOAT64(
+        z,
+        float64QIn[ expNum ] | float64P1[ sigNum ].high,
+        float64P1[ sigNum ].low
+    );
+    ++sigNum;
+    if ( float64NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float64NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float64 float64NextQOutP1( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float64 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    SETFLOAT64(
+        z,
+        float64QOut[ expNum ] | float64P1[ sigNum ].high,
+        float64P1[ sigNum ].low
+    );
+    ++sigNum;
+    if ( float64NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float64NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static const uint32 float64NumQInP2 = float64NumQIn * float64NumP2;
+static const uint32 float64NumQOutP2 = float64NumQOut * float64NumP2;
+
+static float64 float64NextQInP2( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float64 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    SETFLOAT64(
+        z,
+        float64QIn[ expNum ] | float64P2[ sigNum ].high,
+        float64P2[ sigNum ].low
+    );
+    ++sigNum;
+    if ( float64NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float64NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float64 float64NextQOutP2( sequenceT *sequencePtr )
+{
+    uint8 expNum, sigNum;
+    float64 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    SETFLOAT64(
+        z,
+        float64QOut[ expNum ] | float64P2[ sigNum ].high,
+        float64P2[ sigNum ].low
+    );
+    ++sigNum;
+    if ( float64NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float64NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float64 float64RandomQOutP3( void )
+{
+    int8 sigNum1, sigNum2;
+    uint32 sig1Low, sig2Low, zLow;
+    float64 z;
+
+    sigNum1 = randomUint8() % float64NumP2;
+    sigNum2 = randomUint8() % float64NumP2;
+    sig1Low = float64P2[ sigNum1 ].low;
+    sig2Low = float64P2[ sigNum2 ].low;
+    zLow = sig1Low + sig2Low;
+    SETFLOAT64(
+        z,
+          float64QOut[ randomUint8() % float64NumQOut ]
+        | (   (   float64P2[ sigNum1 ].high
+                + float64P2[ sigNum2 ].high
+                + ( zLow < sig1Low )
+              )
+            & 0x000FFFFF
+          ),
+        zLow
+    );
+    return z;
+
+}
+
+static float64 float64RandomQOutPInf( void )
+{
+    float64 z;
+
+    SETFLOAT64(
+        z,
+          float64QOut[ randomUint8() % float64NumQOut ]
+        | ( randomUint32() & 0x000FFFFF ),
+        randomUint32()
+    );
+    return z;
+
+}
+
+enum {
+    float64NumQInfWeightMasks = 10
+};
+
+static const uint32 float64QInfWeightMasks[ float64NumQInfWeightMasks ] = {
+    0x7FF00000,
+    0x7FF00000,
+    0x3FF00000,
+    0x1FF00000,
+    0x0FF00000,
+    0x07F00000,
+    0x03F00000,
+    0x01F00000,
+    0x00F00000,
+    0x00700000
+};
+
+static const uint32 float64QInfWeightOffsets[ float64NumQInfWeightMasks ] = {
+    0x00000000,
+    0x00000000,
+    0x20000000,
+    0x30000000,
+    0x38000000,
+    0x3C000000,
+    0x3E000000,
+    0x3F000000,
+    0x3F800000,
+    0x3FC00000
+};
+
+static float64 float64RandomQInfP3( void )
+{
+    int8 sigNum1, sigNum2;
+    uint32 sig1Low, sig2Low, zLow;
+    int8 weightMaskNum;
+    float64 z;
+
+    sigNum1 = randomUint8() % float64NumP2;
+    sigNum2 = randomUint8() % float64NumP2;
+    sig1Low = float64P2[ sigNum1 ].low;
+    sig2Low = float64P2[ sigNum2 ].low;
+    zLow = sig1Low + sig2Low;
+    weightMaskNum = randomUint8() % float64NumQInfWeightMasks;
+    SETFLOAT64(
+        z,
+          ( ( (uint32) ( randomUint8() & 1 ) )<<31 )
+        | (   (   ( ( (uint32) ( randomUint16() & 0xFFF ) )<<20 )
+                & float64QInfWeightMasks[ weightMaskNum ] )
+            + float64QInfWeightOffsets[ weightMaskNum ]
+          )
+        | (   (   float64P2[ sigNum1 ].high
+                + float64P2[ sigNum2 ].high
+                + ( zLow < sig1Low )
+              )
+            & 0x000FFFFF
+          ),
+        zLow
+    );
+    return z;
+
+}
+
+static float64 float64RandomQInfPInf( void )
+{
+    int8 weightMaskNum;
+    float64 z;
+
+    weightMaskNum = randomUint8() % float64NumQInfWeightMasks;
+    SETFLOAT64(
+        z,
+          ( ( (uint32) ( randomUint8() & 1 ) )<<31 )
+        | (   (   ( ( (uint32) ( randomUint16() & 0xFFF ) )<<20 )
+                & float64QInfWeightMasks[ weightMaskNum ] )
+            + float64QInfWeightOffsets[ weightMaskNum ]
+          )
+        | ( randomUint32() & 0x000FFFFF ),
+        randomUint32()
+    );
+    return z;
+
+}
+
+static float64 float64Random( void )
+{
+
+    switch ( randomUint8() & 7 ) {
+     case 0:
+     case 1:
+     case 2:
+        return float64RandomQOutP3();
+     case 3:
+        return float64RandomQOutPInf();
+     case 4:
+     case 5:
+     case 6:
+        return float64RandomQInfP3();
+     case 7:
+        return float64RandomQInfPInf();
+    }
+
+}
+
+#ifdef FLOATX80
+
+enum {
+    floatx80NumQIn  =  22,
+    floatx80NumQOut =  76,
+    floatx80NumP1   =   4,
+    floatx80NumP2   = 248
+};
+
+static const uint16 floatx80QIn[ floatx80NumQIn ] = {
+    0x0000,		/* positive, subnormal		*/
+    0x0001,		/* positive, -16382		*/
+    0x3FBF,		/* positive,    -64		*/
+    0x3FFD,		/* positive,     -2		*/
+    0x3FFE,		/* positive,     -1		*/
+    0x3FFF,		/* positive,      0		*/
+    0x4000,		/* positive,      1		*/
+    0x4001,		/* positive,      2		*/
+    0x403F,		/* positive,     64		*/
+    0x7FFE,		/* positive,  16383		*/
+    0x7FFF,		/* positive, infinity or NaN	*/
+    0x8000,		/* negative, subnormal		*/
+    0x8001,		/* negative, -16382		*/
+    0xBFBF,		/* negative,    -64		*/
+    0xBFFD,		/* negative,     -2		*/
+    0xBFFE,		/* negative,     -1		*/
+    0xBFFF,		/* negative,      0		*/
+    0xC000,		/* negative,      1		*/
+    0xC001,		/* negative,      2		*/
+    0xC03F,		/* negative,     64		*/
+    0xFFFE,		/* negative,  16383		*/
+    0xFFFF		/* negative, infinity or NaN	*/
+};
+
+static const uint16 floatx80QOut[ floatx80NumQOut ] = {
+    0x0000,		/* positive, subnormal		*/
+    0x0001,		/* positive, -16382		*/
+    0x0002,		/* positive, -16381		*/
+    0x3BFE,		/* positive,  -1025		*/
+    0x3BFF,		/* positive,  -1024		*/
+    0x3C00,		/* positive,  -1023		*/
+    0x3C01,		/* positive,  -1022		*/
+    0x3F7E,		/* positive,   -129		*/
+    0x3F7F,		/* positive,   -128		*/
+    0x3F80,		/* positive,   -127		*/
+    0x3F81,		/* positive,   -126		*/
+    0x3FBF,		/* positive,    -64		*/
+    0x3FFB,		/* positive,     -4		*/
+    0x3FFC,		/* positive,     -3		*/
+    0x3FFD,		/* positive,     -2		*/
+    0x3FFE,		/* positive,     -1		*/
+    0x3FFF,		/* positive,      0		*/
+    0x4000,		/* positive,      1		*/
+    0x4001,		/* positive,      2		*/
+    0x4002,		/* positive,      3		*/
+    0x4003,		/* positive,      4		*/
+    0x401C,		/* positive,     29		*/
+    0x401D,		/* positive,     30		*/
+    0x401E,		/* positive,     31		*/
+    0x401F,		/* positive,     32		*/
+    0x403C,		/* positive,     61		*/
+    0x403D,		/* positive,     62		*/
+    0x403E,		/* positive,     63		*/
+    0x403F,		/* positive,     64		*/
+    0x407E,		/* positive,    127		*/
+    0x407F,		/* positive,    128		*/
+    0x4080,		/* positive,    129		*/
+    0x43FE,		/* positive,   1023		*/
+    0x43FF,		/* positive,   1024		*/
+    0x4400,		/* positive,   1025		*/
+    0x7FFD,		/* positive,  16382		*/
+    0x7FFE,		/* positive,  16383		*/
+    0x7FFF,		/* positive, infinity or NaN	*/
+    0x8000,		/* negative, subnormal		*/
+    0x8001,		/* negative, -16382		*/
+    0x8002,		/* negative, -16381		*/
+    0xBBFE,		/* negative,  -1025		*/
+    0xBBFF,		/* negative,  -1024		*/
+    0xBC00,		/* negative,  -1023		*/
+    0xBC01,		/* negative,  -1022		*/
+    0xBF7E,		/* negative,   -129		*/
+    0xBF7F,		/* negative,   -128		*/
+    0xBF80,		/* negative,   -127		*/
+    0xBF81,		/* negative,   -126		*/
+    0xBFBF,		/* negative,    -64		*/
+    0xBFFB,		/* negative,     -4		*/
+    0xBFFC,		/* negative,     -3		*/
+    0xBFFD,		/* negative,     -2		*/
+    0xBFFE,		/* negative,     -1		*/
+    0xBFFF,		/* negative,      0		*/
+    0xC000,		/* negative,      1		*/
+    0xC001,		/* negative,      2		*/
+    0xC002,		/* negative,      3		*/
+    0xC003,		/* negative,      4		*/
+    0xC01C,		/* negative,     29		*/
+    0xC01D,		/* negative,     30		*/
+    0xC01E,		/* negative,     31		*/
+    0xC01F,		/* negative,     32		*/
+    0xC03C,		/* negative,     61		*/
+    0xC03D,		/* negative,     62		*/
+    0xC03E,		/* negative,     63		*/
+    0xC03F,		/* negative,     64		*/
+    0xC07E,		/* negative,    127		*/
+    0xC07F,		/* negative,    128		*/
+    0xC080,		/* negative,    129		*/
+    0xC3FE,		/* negative,   1023		*/
+    0xC3FF,		/* negative,   1024		*/
+    0xC400,		/* negative,   1025		*/
+    0xFFFD,		/* negative,  16382		*/
+    0xFFFE,		/* negative,  16383		*/
+    0xFFFF		/* negative, infinity or NaN	*/
+};
+
+static const bits64 floatx80P1[ floatx80NumP1 ] = {
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0x0000000000000001 ),
+    LIT64( 0x7FFFFFFFFFFFFFFF ),
+    LIT64( 0x7FFFFFFFFFFFFFFE )
+};
+
+static const bits64 floatx80P2[ floatx80NumP2 ] = {
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0x0000000000000001 ),
+    LIT64( 0x0000000000000002 ),
+    LIT64( 0x0000000000000004 ),
+    LIT64( 0x0000000000000008 ),
+    LIT64( 0x0000000000000010 ),
+    LIT64( 0x0000000000000020 ),
+    LIT64( 0x0000000000000040 ),
+    LIT64( 0x0000000000000080 ),
+    LIT64( 0x0000000000000100 ),
+    LIT64( 0x0000000000000200 ),
+    LIT64( 0x0000000000000400 ),
+    LIT64( 0x0000000000000800 ),
+    LIT64( 0x0000000000001000 ),
+    LIT64( 0x0000000000002000 ),
+    LIT64( 0x0000000000004000 ),
+    LIT64( 0x0000000000008000 ),
+    LIT64( 0x0000000000010000 ),
+    LIT64( 0x0000000000020000 ),
+    LIT64( 0x0000000000040000 ),
+    LIT64( 0x0000000000080000 ),
+    LIT64( 0x0000000000100000 ),
+    LIT64( 0x0000000000200000 ),
+    LIT64( 0x0000000000400000 ),
+    LIT64( 0x0000000000800000 ),
+    LIT64( 0x0000000001000000 ),
+    LIT64( 0x0000000002000000 ),
+    LIT64( 0x0000000004000000 ),
+    LIT64( 0x0000000008000000 ),
+    LIT64( 0x0000000010000000 ),
+    LIT64( 0x0000000020000000 ),
+    LIT64( 0x0000000040000000 ),
+    LIT64( 0x0000000080000000 ),
+    LIT64( 0x0000000100000000 ),
+    LIT64( 0x0000000200000000 ),
+    LIT64( 0x0000000400000000 ),
+    LIT64( 0x0000000800000000 ),
+    LIT64( 0x0000001000000000 ),
+    LIT64( 0x0000002000000000 ),
+    LIT64( 0x0000004000000000 ),
+    LIT64( 0x0000008000000000 ),
+    LIT64( 0x0000010000000000 ),
+    LIT64( 0x0000020000000000 ),
+    LIT64( 0x0000040000000000 ),
+    LIT64( 0x0000080000000000 ),
+    LIT64( 0x0000100000000000 ),
+    LIT64( 0x0000200000000000 ),
+    LIT64( 0x0000400000000000 ),
+    LIT64( 0x0000800000000000 ),
+    LIT64( 0x0001000000000000 ),
+    LIT64( 0x0002000000000000 ),
+    LIT64( 0x0004000000000000 ),
+    LIT64( 0x0008000000000000 ),
+    LIT64( 0x0010000000000000 ),
+    LIT64( 0x0020000000000000 ),
+    LIT64( 0x0040000000000000 ),
+    LIT64( 0x0080000000000000 ),
+    LIT64( 0x0100000000000000 ),
+    LIT64( 0x0200000000000000 ),
+    LIT64( 0x0400000000000000 ),
+    LIT64( 0x0800000000000000 ),
+    LIT64( 0x1000000000000000 ),
+    LIT64( 0x2000000000000000 ),
+    LIT64( 0x4000000000000000 ),
+    LIT64( 0x6000000000000000 ),
+    LIT64( 0x7000000000000000 ),
+    LIT64( 0x7800000000000000 ),
+    LIT64( 0x7C00000000000000 ),
+    LIT64( 0x7E00000000000000 ),
+    LIT64( 0x7F00000000000000 ),
+    LIT64( 0x7F80000000000000 ),
+    LIT64( 0x7FC0000000000000 ),
+    LIT64( 0x7FE0000000000000 ),
+    LIT64( 0x7FF0000000000000 ),
+    LIT64( 0x7FF8000000000000 ),
+    LIT64( 0x7FFC000000000000 ),
+    LIT64( 0x7FFE000000000000 ),
+    LIT64( 0x7FFF000000000000 ),
+    LIT64( 0x7FFF800000000000 ),
+    LIT64( 0x7FFFC00000000000 ),
+    LIT64( 0x7FFFE00000000000 ),
+    LIT64( 0x7FFFF00000000000 ),
+    LIT64( 0x7FFFF80000000000 ),
+    LIT64( 0x7FFFFC0000000000 ),
+    LIT64( 0x7FFFFE0000000000 ),
+    LIT64( 0x7FFFFF0000000000 ),
+    LIT64( 0x7FFFFF8000000000 ),
+    LIT64( 0x7FFFFFC000000000 ),
+    LIT64( 0x7FFFFFE000000000 ),
+    LIT64( 0x7FFFFFF000000000 ),
+    LIT64( 0x7FFFFFF800000000 ),
+    LIT64( 0x7FFFFFFC00000000 ),
+    LIT64( 0x7FFFFFFE00000000 ),
+    LIT64( 0x7FFFFFFF00000000 ),
+    LIT64( 0x7FFFFFFF80000000 ),
+    LIT64( 0x7FFFFFFFC0000000 ),
+    LIT64( 0x7FFFFFFFE0000000 ),
+    LIT64( 0x7FFFFFFFF0000000 ),
+    LIT64( 0x7FFFFFFFF8000000 ),
+    LIT64( 0x7FFFFFFFFC000000 ),
+    LIT64( 0x7FFFFFFFFE000000 ),
+    LIT64( 0x7FFFFFFFFF000000 ),
+    LIT64( 0x7FFFFFFFFF800000 ),
+    LIT64( 0x7FFFFFFFFFC00000 ),
+    LIT64( 0x7FFFFFFFFFE00000 ),
+    LIT64( 0x7FFFFFFFFFF00000 ),
+    LIT64( 0x7FFFFFFFFFF80000 ),
+    LIT64( 0x7FFFFFFFFFFC0000 ),
+    LIT64( 0x7FFFFFFFFFFE0000 ),
+    LIT64( 0x7FFFFFFFFFFF0000 ),
+    LIT64( 0x7FFFFFFFFFFF8000 ),
+    LIT64( 0x7FFFFFFFFFFFC000 ),
+    LIT64( 0x7FFFFFFFFFFFE000 ),
+    LIT64( 0x7FFFFFFFFFFFF000 ),
+    LIT64( 0x7FFFFFFFFFFFF800 ),
+    LIT64( 0x7FFFFFFFFFFFFC00 ),
+    LIT64( 0x7FFFFFFFFFFFFE00 ),
+    LIT64( 0x7FFFFFFFFFFFFF00 ),
+    LIT64( 0x7FFFFFFFFFFFFF80 ),
+    LIT64( 0x7FFFFFFFFFFFFFC0 ),
+    LIT64( 0x7FFFFFFFFFFFFFE0 ),
+    LIT64( 0x7FFFFFFFFFFFFFF0 ),
+    LIT64( 0x7FFFFFFFFFFFFFF8 ),
+    LIT64( 0x7FFFFFFFFFFFFFFC ),
+    LIT64( 0x7FFFFFFFFFFFFFFE ),
+    LIT64( 0x7FFFFFFFFFFFFFFF ),
+    LIT64( 0x7FFFFFFFFFFFFFFD ),
+    LIT64( 0x7FFFFFFFFFFFFFFB ),
+    LIT64( 0x7FFFFFFFFFFFFFF7 ),
+    LIT64( 0x7FFFFFFFFFFFFFEF ),
+    LIT64( 0x7FFFFFFFFFFFFFDF ),
+    LIT64( 0x7FFFFFFFFFFFFFBF ),
+    LIT64( 0x7FFFFFFFFFFFFF7F ),
+    LIT64( 0x7FFFFFFFFFFFFEFF ),
+    LIT64( 0x7FFFFFFFFFFFFDFF ),
+    LIT64( 0x7FFFFFFFFFFFFBFF ),
+    LIT64( 0x7FFFFFFFFFFFF7FF ),
+    LIT64( 0x7FFFFFFFFFFFEFFF ),
+    LIT64( 0x7FFFFFFFFFFFDFFF ),
+    LIT64( 0x7FFFFFFFFFFFBFFF ),
+    LIT64( 0x7FFFFFFFFFFF7FFF ),
+    LIT64( 0x7FFFFFFFFFFEFFFF ),
+    LIT64( 0x7FFFFFFFFFFDFFFF ),
+    LIT64( 0x7FFFFFFFFFFBFFFF ),
+    LIT64( 0x7FFFFFFFFFF7FFFF ),
+    LIT64( 0x7FFFFFFFFFEFFFFF ),
+    LIT64( 0x7FFFFFFFFFDFFFFF ),
+    LIT64( 0x7FFFFFFFFFBFFFFF ),
+    LIT64( 0x7FFFFFFFFF7FFFFF ),
+    LIT64( 0x7FFFFFFFFEFFFFFF ),
+    LIT64( 0x7FFFFFFFFDFFFFFF ),
+    LIT64( 0x7FFFFFFFFBFFFFFF ),
+    LIT64( 0x7FFFFFFFF7FFFFFF ),
+    LIT64( 0x7FFFFFFFEFFFFFFF ),
+    LIT64( 0x7FFFFFFFDFFFFFFF ),
+    LIT64( 0x7FFFFFFFBFFFFFFF ),
+    LIT64( 0x7FFFFFFF7FFFFFFF ),
+    LIT64( 0x7FFFFFFEFFFFFFFF ),
+    LIT64( 0x7FFFFFFDFFFFFFFF ),
+    LIT64( 0x7FFFFFFBFFFFFFFF ),
+    LIT64( 0x7FFFFFF7FFFFFFFF ),
+    LIT64( 0x7FFFFFEFFFFFFFFF ),
+    LIT64( 0x7FFFFFDFFFFFFFFF ),
+    LIT64( 0x7FFFFFBFFFFFFFFF ),
+    LIT64( 0x7FFFFF7FFFFFFFFF ),
+    LIT64( 0x7FFFFEFFFFFFFFFF ),
+    LIT64( 0x7FFFFDFFFFFFFFFF ),
+    LIT64( 0x7FFFFBFFFFFFFFFF ),
+    LIT64( 0x7FFFF7FFFFFFFFFF ),
+    LIT64( 0x7FFFEFFFFFFFFFFF ),
+    LIT64( 0x7FFFDFFFFFFFFFFF ),
+    LIT64( 0x7FFFBFFFFFFFFFFF ),
+    LIT64( 0x7FFF7FFFFFFFFFFF ),
+    LIT64( 0x7FFEFFFFFFFFFFFF ),
+    LIT64( 0x7FFDFFFFFFFFFFFF ),
+    LIT64( 0x7FFBFFFFFFFFFFFF ),
+    LIT64( 0x7FF7FFFFFFFFFFFF ),
+    LIT64( 0x7FEFFFFFFFFFFFFF ),
+    LIT64( 0x7FDFFFFFFFFFFFFF ),
+    LIT64( 0x7FBFFFFFFFFFFFFF ),
+    LIT64( 0x7F7FFFFFFFFFFFFF ),
+    LIT64( 0x7EFFFFFFFFFFFFFF ),
+    LIT64( 0x7DFFFFFFFFFFFFFF ),
+    LIT64( 0x7BFFFFFFFFFFFFFF ),
+    LIT64( 0x77FFFFFFFFFFFFFF ),
+    LIT64( 0x6FFFFFFFFFFFFFFF ),
+    LIT64( 0x5FFFFFFFFFFFFFFF ),
+    LIT64( 0x3FFFFFFFFFFFFFFF ),
+    LIT64( 0x1FFFFFFFFFFFFFFF ),
+    LIT64( 0x0FFFFFFFFFFFFFFF ),
+    LIT64( 0x07FFFFFFFFFFFFFF ),
+    LIT64( 0x03FFFFFFFFFFFFFF ),
+    LIT64( 0x01FFFFFFFFFFFFFF ),
+    LIT64( 0x00FFFFFFFFFFFFFF ),
+    LIT64( 0x007FFFFFFFFFFFFF ),
+    LIT64( 0x003FFFFFFFFFFFFF ),
+    LIT64( 0x001FFFFFFFFFFFFF ),
+    LIT64( 0x000FFFFFFFFFFFFF ),
+    LIT64( 0x0007FFFFFFFFFFFF ),
+    LIT64( 0x0003FFFFFFFFFFFF ),
+    LIT64( 0x0001FFFFFFFFFFFF ),
+    LIT64( 0x0000FFFFFFFFFFFF ),
+    LIT64( 0x00007FFFFFFFFFFF ),
+    LIT64( 0x00003FFFFFFFFFFF ),
+    LIT64( 0x00001FFFFFFFFFFF ),
+    LIT64( 0x00000FFFFFFFFFFF ),
+    LIT64( 0x000007FFFFFFFFFF ),
+    LIT64( 0x000003FFFFFFFFFF ),
+    LIT64( 0x000001FFFFFFFFFF ),
+    LIT64( 0x000000FFFFFFFFFF ),
+    LIT64( 0x0000007FFFFFFFFF ),
+    LIT64( 0x0000003FFFFFFFFF ),
+    LIT64( 0x0000001FFFFFFFFF ),
+    LIT64( 0x0000000FFFFFFFFF ),
+    LIT64( 0x00000007FFFFFFFF ),
+    LIT64( 0x00000003FFFFFFFF ),
+    LIT64( 0x00000001FFFFFFFF ),
+    LIT64( 0x00000000FFFFFFFF ),
+    LIT64( 0x000000007FFFFFFF ),
+    LIT64( 0x000000003FFFFFFF ),
+    LIT64( 0x000000001FFFFFFF ),
+    LIT64( 0x000000000FFFFFFF ),
+    LIT64( 0x0000000007FFFFFF ),
+    LIT64( 0x0000000003FFFFFF ),
+    LIT64( 0x0000000001FFFFFF ),
+    LIT64( 0x0000000000FFFFFF ),
+    LIT64( 0x00000000007FFFFF ),
+    LIT64( 0x00000000003FFFFF ),
+    LIT64( 0x00000000001FFFFF ),
+    LIT64( 0x00000000000FFFFF ),
+    LIT64( 0x000000000007FFFF ),
+    LIT64( 0x000000000003FFFF ),
+    LIT64( 0x000000000001FFFF ),
+    LIT64( 0x000000000000FFFF ),
+    LIT64( 0x0000000000007FFF ),
+    LIT64( 0x0000000000003FFF ),
+    LIT64( 0x0000000000001FFF ),
+    LIT64( 0x0000000000000FFF ),
+    LIT64( 0x00000000000007FF ),
+    LIT64( 0x00000000000003FF ),
+    LIT64( 0x00000000000001FF ),
+    LIT64( 0x00000000000000FF ),
+    LIT64( 0x000000000000007F ),
+    LIT64( 0x000000000000003F ),
+    LIT64( 0x000000000000001F ),
+    LIT64( 0x000000000000000F ),
+    LIT64( 0x0000000000000007 ),
+    LIT64( 0x0000000000000003 )
+};
+
+static const uint32 floatx80NumQInP1 = floatx80NumQIn * floatx80NumP1;
+static const uint32 floatx80NumQOutP1 = floatx80NumQOut * floatx80NumP1;
+
+static floatx80 floatx80NextQInP1( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    floatx80 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = floatx80P1[ sigNum ];
+    z.high = floatx80QIn[ expNum ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    ++sigNum;
+    if ( floatx80NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( floatx80NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static floatx80 floatx80NextQOutP1( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    floatx80 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = floatx80P1[ sigNum ];
+    z.high = floatx80QOut[ expNum ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    ++sigNum;
+    if ( floatx80NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( floatx80NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static const uint32 floatx80NumQInP2 = floatx80NumQIn * floatx80NumP2;
+static const uint32 floatx80NumQOutP2 = floatx80NumQOut * floatx80NumP2;
+
+static floatx80 floatx80NextQInP2( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    floatx80 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = floatx80P2[ sigNum ];
+    z.high = floatx80QIn[ expNum ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    ++sigNum;
+    if ( floatx80NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( floatx80NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static floatx80 floatx80NextQOutP2( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    floatx80 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = floatx80P2[ sigNum ];
+    z.high = floatx80QOut[ expNum ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    ++sigNum;
+    if ( floatx80NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( floatx80NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static floatx80 floatx80RandomQOutP3( void )
+{
+    floatx80 z;
+
+    z.low =
+          (   floatx80P2[ randomUint8() % floatx80NumP2 ]
+            + floatx80P2[ randomUint8() % floatx80NumP2 ] )
+        & LIT64( 0x7FFFFFFFFFFFFFFF );
+    z.high = floatx80QOut[ randomUint8() % floatx80NumQOut ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    return z;
+
+}
+
+static floatx80 floatx80RandomQOutPInf( void )
+{
+    floatx80 z;
+
+    z.low = randomUint64() & LIT64( 0x7FFFFFFFFFFFFFFF );
+    z.high = floatx80QOut[ randomUint8() % floatx80NumQOut ];
+    if ( z.high & 0x7FFF ) z.low |= LIT64( 0x8000000000000000 );
+    return z;
+
+}
+
+enum {
+    floatx80NumQInfWeightMasks = 14
+};
+
+static const uint16 floatx80QInfWeightMasks[ floatx80NumQInfWeightMasks ] = {
+    0x7FFF,
+    0x7FFF,
+    0x3FFF,
+    0x1FFF,
+    0x07FF,
+    0x07FF,
+    0x03FF,
+    0x01FF,
+    0x00FF,
+    0x007F,
+    0x003F,
+    0x001F,
+    0x000F,
+    0x0007
+};
+
+static const uint16 floatx80QInfWeightOffsets[ floatx80NumQInfWeightMasks ] = {
+    0x0000,
+    0x0000,
+    0x2000,
+    0x3000,
+    0x3800,
+    0x3C00,
+    0x3E00,
+    0x3F00,
+    0x3F80,
+    0x3FC0,
+    0x3FE0,
+    0x3FF0,
+    0x3FF8,
+    0x3FFC
+};
+
+static floatx80 floatx80RandomQInfP3( void )
+{
+    int8 weightMaskNum;
+    floatx80 z;
+
+    z.low =
+          (   floatx80P2[ randomUint8() % floatx80NumP2 ]
+            + floatx80P2[ randomUint8() % floatx80NumP2 ] )
+        & LIT64( 0x7FFFFFFFFFFFFFFF );
+    weightMaskNum = randomUint8() % floatx80NumQInfWeightMasks;
+    z.high =
+          randomUint16() & floatx80QInfWeightMasks[ weightMaskNum ]
+        + floatx80QInfWeightOffsets[ weightMaskNum ];
+    if ( z.high ) z.low |= LIT64( 0x8000000000000000 );
+    z.high |= ( (uint16) ( randomUint8() & 1 ) )<<15;
+    return z;
+
+}
+
+static floatx80 floatx80RandomQInfPInf( void )
+{
+    int8 weightMaskNum;
+    floatx80 z;
+
+    z.low = randomUint64() & LIT64( 0x7FFFFFFFFFFFFFFF );
+    weightMaskNum = randomUint8() % floatx80NumQInfWeightMasks;
+    z.high =
+          randomUint16() & floatx80QInfWeightMasks[ weightMaskNum ]
+        + floatx80QInfWeightOffsets[ weightMaskNum ];
+    if ( z.high ) z.low |= LIT64( 0x8000000000000000 );
+    z.high |= ( (uint16) ( randomUint8() & 1 ) )<<15;
+    return z;
+
+}
+
+static floatx80 floatx80Random( void )
+{
+
+    switch ( randomUint8() & 7 ) {
+     case 0:
+     case 1:
+     case 2:
+        return floatx80RandomQOutP3();
+     case 3:
+        return floatx80RandomQOutPInf();
+     case 4:
+     case 5:
+     case 6:
+        return floatx80RandomQInfP3();
+     case 7:
+        return floatx80RandomQInfPInf();
+    }
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+enum {
+    float128NumQIn  =  22,
+    float128NumQOut =  78,
+    float128NumP1   =   4,
+    float128NumP2   = 443
+};
+
+static const uint64 float128QIn[ float128NumQIn ] = {
+    LIT64( 0x0000000000000000 ),	/* positive, subnormal		*/
+    LIT64( 0x0001000000000000 ),	/* positive, -16382		*/
+    LIT64( 0x3F8E000000000000 ),	/* positive,   -113		*/
+    LIT64( 0x3FFD000000000000 ),	/* positive,     -2		*/
+    LIT64( 0x3FFE000000000000 ),	/* positive,     -1		*/
+    LIT64( 0x3FFF000000000000 ),	/* positive,      0		*/
+    LIT64( 0x4000000000000000 ),	/* positive,      1		*/
+    LIT64( 0x4001000000000000 ),	/* positive,      2		*/
+    LIT64( 0x4070000000000000 ),	/* positive,    113		*/
+    LIT64( 0x7FFE000000000000 ),	/* positive,  16383		*/
+    LIT64( 0x7FFF000000000000 ),	/* positive, infinity or NaN	*/
+    LIT64( 0x8000000000000000 ),	/* negative, subnormal		*/
+    LIT64( 0x8001000000000000 ),	/* negative, -16382		*/
+    LIT64( 0xBF8E000000000000 ),	/* negative,   -113		*/
+    LIT64( 0xBFFD000000000000 ),	/* negative,     -2		*/
+    LIT64( 0xBFFE000000000000 ),	/* negative,     -1		*/
+    LIT64( 0xBFFF000000000000 ),	/* negative,      0		*/
+    LIT64( 0xC000000000000000 ),	/* negative,      1		*/
+    LIT64( 0xC001000000000000 ),	/* negative,      2		*/
+    LIT64( 0xC070000000000000 ),	/* negative,    113		*/
+    LIT64( 0xFFFE000000000000 ),	/* negative,  16383		*/
+    LIT64( 0xFFFF000000000000 )		/* negative, infinity or NaN	*/
+};
+
+static const uint64 float128QOut[ float128NumQOut ] = {
+    LIT64( 0x0000000000000000 ),	/* positive, subnormal		*/
+    LIT64( 0x0001000000000000 ),	/* positive, -16382		*/
+    LIT64( 0x0002000000000000 ),	/* positive, -16381		*/
+    LIT64( 0x3BFE000000000000 ),	/* positive,  -1025		*/
+    LIT64( 0x3BFF000000000000 ),	/* positive,  -1024		*/
+    LIT64( 0x3C00000000000000 ),	/* positive,  -1023		*/
+    LIT64( 0x3C01000000000000 ),	/* positive,  -1022		*/
+    LIT64( 0x3F7E000000000000 ),	/* positive,   -129		*/
+    LIT64( 0x3F7F000000000000 ),	/* positive,   -128		*/
+    LIT64( 0x3F80000000000000 ),	/* positive,   -127		*/
+    LIT64( 0x3F81000000000000 ),	/* positive,   -126		*/
+    LIT64( 0x3F8E000000000000 ),	/* positive,   -113		*/
+    LIT64( 0x3FFB000000000000 ),	/* positive,     -4		*/
+    LIT64( 0x3FFC000000000000 ),	/* positive,     -3		*/
+    LIT64( 0x3FFD000000000000 ),	/* positive,     -2		*/
+    LIT64( 0x3FFE000000000000 ),	/* positive,     -1		*/
+    LIT64( 0x3FFF000000000000 ),	/* positive,      0		*/
+    LIT64( 0x4000000000000000 ),	/* positive,      1		*/
+    LIT64( 0x4001000000000000 ),	/* positive,      2		*/
+    LIT64( 0x4002000000000000 ),	/* positive,      3		*/
+    LIT64( 0x4003000000000000 ),	/* positive,      4		*/
+    LIT64( 0x401C000000000000 ),	/* positive,     29		*/
+    LIT64( 0x401D000000000000 ),	/* positive,     30		*/
+    LIT64( 0x401E000000000000 ),	/* positive,     31		*/
+    LIT64( 0x401F000000000000 ),	/* positive,     32		*/
+    LIT64( 0x403C000000000000 ),	/* positive,     61		*/
+    LIT64( 0x403D000000000000 ),	/* positive,     62		*/
+    LIT64( 0x403E000000000000 ),	/* positive,     63		*/
+    LIT64( 0x403F000000000000 ),	/* positive,     64		*/
+    LIT64( 0x4070000000000000 ),	/* positive,    113		*/
+    LIT64( 0x407E000000000000 ),	/* positive,    127		*/
+    LIT64( 0x407F000000000000 ),	/* positive,    128		*/
+    LIT64( 0x4080000000000000 ),	/* positive,    129		*/
+    LIT64( 0x43FE000000000000 ),	/* positive,   1023		*/
+    LIT64( 0x43FF000000000000 ),	/* positive,   1024		*/
+    LIT64( 0x4400000000000000 ),	/* positive,   1025		*/
+    LIT64( 0x7FFD000000000000 ),	/* positive,  16382		*/
+    LIT64( 0x7FFE000000000000 ),	/* positive,  16383		*/
+    LIT64( 0x7FFF000000000000 ),	/* positive, infinity or NaN	*/
+    LIT64( 0x8000000000000000 ),	/* negative, subnormal		*/
+    LIT64( 0x8001000000000000 ),	/* negative, -16382		*/
+    LIT64( 0x8002000000000000 ),	/* negative, -16381		*/
+    LIT64( 0xBBFE000000000000 ),	/* negative,  -1025		*/
+    LIT64( 0xBBFF000000000000 ),	/* negative,  -1024		*/
+    LIT64( 0xBC00000000000000 ),	/* negative,  -1023		*/
+    LIT64( 0xBC01000000000000 ),	/* negative,  -1022		*/
+    LIT64( 0xBF7E000000000000 ),	/* negative,   -129		*/
+    LIT64( 0xBF7F000000000000 ),	/* negative,   -128		*/
+    LIT64( 0xBF80000000000000 ),	/* negative,   -127		*/
+    LIT64( 0xBF81000000000000 ),	/* negative,   -126		*/
+    LIT64( 0xBF8E000000000000 ),	/* negative,   -113		*/
+    LIT64( 0xBFFB000000000000 ),	/* negative,     -4		*/
+    LIT64( 0xBFFC000000000000 ),	/* negative,     -3		*/
+    LIT64( 0xBFFD000000000000 ),	/* negative,     -2		*/
+    LIT64( 0xBFFE000000000000 ),	/* negative,     -1		*/
+    LIT64( 0xBFFF000000000000 ),	/* negative,      0		*/
+    LIT64( 0xC000000000000000 ),	/* negative,      1		*/
+    LIT64( 0xC001000000000000 ),	/* negative,      2		*/
+    LIT64( 0xC002000000000000 ),	/* negative,      3		*/
+    LIT64( 0xC003000000000000 ),	/* negative,      4		*/
+    LIT64( 0xC01C000000000000 ),	/* negative,     29		*/
+    LIT64( 0xC01D000000000000 ),	/* negative,     30		*/
+    LIT64( 0xC01E000000000000 ),	/* negative,     31		*/
+    LIT64( 0xC01F000000000000 ),	/* negative,     32		*/
+    LIT64( 0xC03C000000000000 ),	/* negative,     61		*/
+    LIT64( 0xC03D000000000000 ),	/* negative,     62		*/
+    LIT64( 0xC03E000000000000 ),	/* negative,     63		*/
+    LIT64( 0xC03F000000000000 ),	/* negative,     64		*/
+    LIT64( 0xC070000000000000 ),	/* negative,    113		*/
+    LIT64( 0xC07E000000000000 ),	/* negative,    127		*/
+    LIT64( 0xC07F000000000000 ),	/* negative,    128		*/
+    LIT64( 0xC080000000000000 ),	/* negative,    129		*/
+    LIT64( 0xC3FE000000000000 ),	/* negative,   1023		*/
+    LIT64( 0xC3FF000000000000 ),	/* negative,   1024		*/
+    LIT64( 0xC400000000000000 ),	/* negative,   1025		*/
+    LIT64( 0xFFFD000000000000 ),	/* negative,  16382		*/
+    LIT64( 0xFFFE000000000000 ),	/* negative,  16383		*/
+    LIT64( 0xFFFF000000000000 )		/* negative, infinity or NaN	*/
+};
+
+static const struct { bits64 high, low; } float128P1[ float128NumP1 ] = {
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000001 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFE ) }
+};
+
+static const struct { bits64 high, low; } float128P2[ float128NumP2 ] = {
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000001 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000002 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000004 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000008 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000010 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000020 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000040 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000080 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000100 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000200 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000400 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000800 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000001000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000002000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000004000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000008000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000010000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000020000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000040000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000080000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000100000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000200000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000400000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000800000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000001000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000002000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000004000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000008000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000010000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000020000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000040000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000080000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000100000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000200000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000400000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000800000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000001000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000002000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000004000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000008000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000010000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000020000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000040000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000080000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000100000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000200000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000400000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000800000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0001000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0002000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0004000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0008000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0010000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0020000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0040000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0080000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0100000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0200000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0400000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0800000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x1000000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x2000000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x4000000000000000 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x8000000000000000 ) },
+    { LIT64( 0x0000000000000001 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000002 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000004 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000008 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000010 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000020 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000040 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000080 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000100 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000200 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000400 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000000800 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000001000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000002000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000004000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000008000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000010000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000020000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000040000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000080000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000100000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000200000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000400000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000000800000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000001000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000002000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000004000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000008000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000010000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000020000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000040000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000080000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000100000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000200000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000400000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000000800000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000001000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000002000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000004000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000008000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000010000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000020000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000040000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000080000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000100000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000200000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000400000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000800000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000C00000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000E00000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000F00000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000F80000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FC0000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FE0000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FF0000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FF8000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFC000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFE000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFF000000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFF800000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFC00000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFE00000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFF00000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFF80000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFC0000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFE0000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFF0000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFF8000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFC000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFE000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFF000000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFF800000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFC00000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFE00000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFF00000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFF80000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFC0000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFE0000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFF0000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFF8000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFC000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFE000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFF000 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFF800 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFC00 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFE00 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFF00 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFF80 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFC0 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFE0 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFF0 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFF8 ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFC ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFE ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0x0000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0x8000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xC000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xE000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xF000000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xF800000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFC00000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFE00000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFF00000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFF80000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFC0000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFE0000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFF0000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFF8000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFC000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFE000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFF000000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFF800000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFC00000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFE00000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFF00000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFF80000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFC0000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFE0000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFF0000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFF8000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFC000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFE000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFF000000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFF800000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFC00000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFE00000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFF00000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFF80000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFC0000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFE0000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFF0000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFF8000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFC000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFE000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFF000000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFF800000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFC00000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFE00000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFF00000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFF80000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFC0000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFE0000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFF0000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFF8000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFC000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFE000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFF000 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFF800 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFC00 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFE00 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFF00 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFF80 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFC0 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFE0 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFF0 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFF8 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFC ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFE ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFD ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFB ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFF7 ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFEF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFDF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFBF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFF7F ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFEFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFDFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFBFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFF7FF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFEFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFDFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFBFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFF7FFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFEFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFDFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFBFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFF7FFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFEFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFDFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFBFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFF7FFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFEFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFDFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFBFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFF7FFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFEFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFDFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFFBFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFF7FFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFEFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFDFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFFBFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFF7FFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFEFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFDFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFFBFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFF7FFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFEFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFDFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFFBFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFF7FFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFEFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFDFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFFBFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFF7FFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFEFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFDFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFFBFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFF7FFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFEFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFDFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFFBFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFF7FFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFEFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFDFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xFBFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xF7FFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xEFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xDFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0xBFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFF ), LIT64( 0x7FFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFD ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFFB ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFF7 ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFEF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFDF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFFBF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFF7F ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFEFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFDFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFFBFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFF7FF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFEFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFDFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFFBFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFF7FFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFEFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFDFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFFBFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFF7FFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFEFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFDFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFFBFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFF7FFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFEFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFDFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFFBFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFF7FFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFEFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFDFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFFBFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFF7FFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFEFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFDFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFFBFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFF7FFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFEFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFDFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FFBFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FF7FFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FEFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FDFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000FBFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000F7FFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000EFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000DFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000BFFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00007FFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00003FFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00001FFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000FFFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000007FFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000003FFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000001FFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000FFFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000007FFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000003FFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000001FFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000FFFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000007FFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000003FFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000001FFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000FFFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000007FFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000003FFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000001FFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000FFFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000007FFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000003FFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000001FFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000FFFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000007FFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000003FFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000001FFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000000FFFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000007FFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000003FFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000001FFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000000FFFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000007FFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000003FFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000001FFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000FFF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000000007FF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000000003FF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000000001FF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x00000000000000FF ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000000007F ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000000003F ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000000001F ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x000000000000000F ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000007 ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000003 ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000001 ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0xFFFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x7FFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x3FFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x1FFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0FFFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x07FFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x03FFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x01FFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00FFFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x007FFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x003FFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x001FFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000FFFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0007FFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0003FFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0001FFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000FFFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00007FFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00003FFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00001FFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000FFFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000007FFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000003FFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000001FFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000FFFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000007FFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000003FFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000001FFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000FFFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000007FFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000003FFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000001FFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000FFFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000007FFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000003FFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000001FFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000FFFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000007FFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000003FFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000001FFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000FFFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000007FFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000003FFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000001FFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000000FFFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000007FFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000003FFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000001FFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000000FFFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000007FFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000003FFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000001FFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000FFF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000000007FF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000000003FF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000000001FF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x00000000000000FF ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000000007F ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000000003F ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000000001F ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x000000000000000F ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000007 ) },
+    { LIT64( 0x0000000000000000 ), LIT64( 0x0000000000000003 ) }
+};
+
+static const uint32 float128NumQInP1 = float128NumQIn * float128NumP1;
+static const uint32 float128NumQOutP1 = float128NumQOut * float128NumP1;
+
+static float128 float128NextQInP1( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    float128 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = float128P1[ sigNum ].low;
+    z.high = float128QIn[ expNum ] | float128P1[ sigNum ].high;
+    ++sigNum;
+    if ( float128NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float128NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float128 float128NextQOutP1( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    float128 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = float128P1[ sigNum ].low;
+    z.high = float128QOut[ expNum ] | float128P1[ sigNum ].high;
+    ++sigNum;
+    if ( float128NumP1 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float128NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static const uint32 float128NumQInP2 = float128NumQIn * float128NumP2;
+static const uint32 float128NumQOutP2 = float128NumQOut * float128NumP2;
+
+static float128 float128NextQInP2( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    float128 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = float128P2[ sigNum ].low;
+    z.high = float128QIn[ expNum ] | float128P2[ sigNum ].high;
+    ++sigNum;
+    if ( float128NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float128NumQIn <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float128 float128NextQOutP2( sequenceT *sequencePtr )
+{
+    int16 expNum, sigNum;
+    float128 z;
+
+    sigNum = sequencePtr->term1Num;
+    expNum = sequencePtr->expNum;
+    z.low = float128P2[ sigNum ].low;
+    z.high = float128QOut[ expNum ] | float128P2[ sigNum ].high;
+    ++sigNum;
+    if ( float128NumP2 <= sigNum ) {
+        sigNum = 0;
+        ++expNum;
+        if ( float128NumQOut <= expNum ) {
+            expNum = 0;
+            sequencePtr->done = TRUE;
+        }
+        sequencePtr->expNum = expNum;
+    }
+    sequencePtr->term1Num = sigNum;
+    return z;
+
+}
+
+static float128 float128RandomQOutP3( void )
+{
+    int16 sigNum1, sigNum2;
+    uint64 sig1Low, sig2Low;
+    float128 z;
+
+    sigNum1 = randomUint8() % float128NumP2;
+    sigNum2 = randomUint8() % float128NumP2;
+    sig1Low = float128P2[ sigNum1 ].low;
+    sig2Low = float128P2[ sigNum2 ].low;
+    z.low = sig1Low + sig2Low;
+    z.high =
+          float128QOut[ randomUint8() % float128NumQOut ]
+        | (   (   float128P2[ sigNum1 ].high
+                + float128P2[ sigNum2 ].high
+                + ( z.low < sig1Low )
+              )
+            & LIT64( 0x0000FFFFFFFFFFFF )
+          );
+    return z;
+
+}
+
+static float128 float128RandomQOutPInf( void )
+{
+    float128 z;
+
+    z.low = randomUint64();
+    z.high =
+          float128QOut[ randomUint8() % float128NumQOut ]
+        | ( randomUint64() & LIT64( 0x0000FFFFFFFFFFFF ) );
+    return z;
+
+}
+
+enum {
+    float128NumQInfWeightMasks = 14
+};
+
+static const uint64 float128QInfWeightMasks[ float128NumQInfWeightMasks ] = {
+    LIT64( 0x7FFF000000000000 ),
+    LIT64( 0x7FFF000000000000 ),
+    LIT64( 0x3FFF000000000000 ),
+    LIT64( 0x1FFF000000000000 ),
+    LIT64( 0x07FF000000000000 ),
+    LIT64( 0x07FF000000000000 ),
+    LIT64( 0x03FF000000000000 ),
+    LIT64( 0x01FF000000000000 ),
+    LIT64( 0x00FF000000000000 ),
+    LIT64( 0x007F000000000000 ),
+    LIT64( 0x003F000000000000 ),
+    LIT64( 0x001F000000000000 ),
+    LIT64( 0x000F000000000000 ),
+    LIT64( 0x0007000000000000 )
+};
+
+static const uint64 float128QInfWeightOffsets[ float128NumQInfWeightMasks ] = {
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0x0000000000000000 ),
+    LIT64( 0x2000000000000000 ),
+    LIT64( 0x3000000000000000 ),
+    LIT64( 0x3800000000000000 ),
+    LIT64( 0x3C00000000000000 ),
+    LIT64( 0x3E00000000000000 ),
+    LIT64( 0x3F00000000000000 ),
+    LIT64( 0x3F80000000000000 ),
+    LIT64( 0x3FC0000000000000 ),
+    LIT64( 0x3FE0000000000000 ),
+    LIT64( 0x3FF0000000000000 ),
+    LIT64( 0x3FF8000000000000 ),
+    LIT64( 0x3FFC000000000000 )
+};
+
+static float128 float128RandomQInfP3( void )
+{
+    int16 sigNum1, sigNum2;
+    uint64 sig1Low, sig2Low;
+    int8 weightMaskNum;
+    float128 z;
+
+    sigNum1 = randomUint8() % float128NumP2;
+    sigNum2 = randomUint8() % float128NumP2;
+    sig1Low = float128P2[ sigNum1 ].low;
+    sig2Low = float128P2[ sigNum2 ].low;
+    z.low = sig1Low + sig2Low;
+    weightMaskNum = randomUint8() % float128NumQInfWeightMasks;
+    z.high =
+          ( ( (uint64) ( randomUint8() & 1 ) )<<63 )
+        | (   (   ( ( (uint64) randomUint16() )<<48 )
+                & float128QInfWeightMasks[ weightMaskNum ] )
+            + float128QInfWeightOffsets[ weightMaskNum ]
+          )
+        | (   (   float128P2[ sigNum1 ].high
+                + float128P2[ sigNum2 ].high
+                + ( z.low < sig1Low )
+              )
+            & LIT64( 0x0000FFFFFFFFFFFF )
+          );
+    return z;
+
+}
+
+static float128 float128RandomQInfPInf( void )
+{
+    int8 weightMaskNum;
+    float128 z;
+
+    weightMaskNum = randomUint8() % float128NumQInfWeightMasks;
+    z.low = randomUint64();
+    z.high =
+          ( ( (uint64) ( randomUint8() & 1 ) )<<63 )
+        | (   (   ( ( (uint64) randomUint16() )<<48 )
+                & float128QInfWeightMasks[ weightMaskNum ] )
+            + float128QInfWeightOffsets[ weightMaskNum ]
+          )
+        | ( randomUint64() & LIT64( 0x0000FFFFFFFFFFFF ) );
+    return z;
+
+}
+
+static float128 float128Random( void )
+{
+
+    switch ( randomUint8() & 7 ) {
+     case 0:
+     case 1:
+     case 2:
+        return float128RandomQOutP3();
+     case 3:
+        return float128RandomQOutPInf();
+     case 4:
+     case 5:
+     case 6:
+        return float128RandomQInfP3();
+     case 7:
+        return float128RandomQInfPInf();
+    }
+
+}
+
+#endif
+
+static int8 level = 0;
+
+void testCases_setLevel( int8 levelIn )
+{
+
+    if ( ( levelIn < 1 ) || ( 2 < levelIn ) ) {
+        fail( "Invalid testing level: %d", levelIn );
+    }
+    level = levelIn;
+
+}
+
+static int8 sequenceType;
+static sequenceT sequenceA, sequenceB;
+static int8 subcase;
+
+uint32 testCases_total;
+flag testCases_done;
+
+static float32 current_a_float32;
+static float32 current_b_float32;
+static float64 current_a_float64;
+static float64 current_b_float64;
+#ifdef FLOATX80
+static floatx80 current_a_floatx80;
+static floatx80 current_b_floatx80;
+#endif
+#ifdef FLOAT128
+static float128 current_a_float128;
+static float128 current_b_float128;
+#endif
+
+void testCases_initSequence( int8 sequenceTypeIn )
+{
+
+    sequenceType = sequenceTypeIn;
+    sequenceA.term2Num = 0;
+    sequenceA.term1Num = 0;
+    sequenceA.expNum = 0;
+    sequenceA.done = FALSE;
+    sequenceB.term2Num = 0;
+    sequenceB.term1Num = 0;
+    sequenceB.expNum = 0;
+    sequenceB.done = FALSE;
+    subcase = 0;
+    switch ( level ) {
+     case 1:
+        switch ( sequenceTypeIn ) {
+         case testCases_sequence_a_int32:
+            testCases_total = 3 * int32NumP1;
+            break;
+#ifdef BITS64
+         case testCases_sequence_a_int64:
+            testCases_total = 3 * int64NumP1;
+            break;
+#endif
+         case testCases_sequence_a_float32:
+            testCases_total = 3 * float32NumQOutP1;
+            break;
+         case testCases_sequence_ab_float32:
+            testCases_total = 6 * float32NumQInP1 * float32NumQInP1;
+            current_a_float32 = float32NextQInP1( &sequenceA );
+            break;
+         case testCases_sequence_a_float64:
+            testCases_total = 3 * float64NumQOutP1;
+            break;
+         case testCases_sequence_ab_float64:
+            testCases_total = 6 * float64NumQInP1 * float64NumQInP1;
+            current_a_float64 = float64NextQInP1( &sequenceA );
+            break;
+#ifdef FLOATX80
+         case testCases_sequence_a_floatx80:
+            testCases_total = 3 * floatx80NumQOutP1;
+            break;
+         case testCases_sequence_ab_floatx80:
+            testCases_total = 6 * floatx80NumQInP1 * floatx80NumQInP1;
+            current_a_floatx80 = floatx80NextQInP1( &sequenceA );
+            break;
+#endif
+#ifdef FLOAT128
+         case testCases_sequence_a_float128:
+            testCases_total = 3 * float128NumQOutP1;
+            break;
+         case testCases_sequence_ab_float128:
+            testCases_total = 6 * float128NumQInP1 * float128NumQInP1;
+            current_a_float128 = float128NextQInP1( &sequenceA );
+            break;
+#endif
+        }
+        break;
+     case 2:
+        switch ( sequenceTypeIn ) {
+         case testCases_sequence_a_int32:
+            testCases_total = 2 * int32NumP2;
+            break;
+#ifdef BITS64
+         case testCases_sequence_a_int64:
+            testCases_total = 2 * int64NumP2;
+            break;
+#endif
+         case testCases_sequence_a_float32:
+            testCases_total = 2 * float32NumQOutP2;
+            break;
+         case testCases_sequence_ab_float32:
+            testCases_total = 2 * float32NumQInP2 * float32NumQInP2;
+            current_a_float32 = float32NextQInP2( &sequenceA );
+            break;
+         case testCases_sequence_a_float64:
+            testCases_total = 2 * float64NumQOutP2;
+            break;
+         case testCases_sequence_ab_float64:
+            testCases_total = 2 * float64NumQInP2 * float64NumQInP2;
+            current_a_float64 = float64NextQInP2( &sequenceA );
+            break;
+#ifdef FLOATX80
+         case testCases_sequence_a_floatx80:
+            testCases_total = 2 * floatx80NumQOutP2;
+            break;
+         case testCases_sequence_ab_floatx80:
+            testCases_total = 2 * floatx80NumQInP2 * floatx80NumQInP2;
+            current_a_floatx80 = floatx80NextQInP2( &sequenceA );
+            break;
+#endif
+#ifdef FLOAT128
+         case testCases_sequence_a_float128:
+            testCases_total = 2 * float128NumQOutP2;
+            break;
+         case testCases_sequence_ab_float128:
+            testCases_total = 2 * float128NumQInP2 * float128NumQInP2;
+            current_a_float128 = float128NextQInP2( &sequenceA );
+            break;
+#endif
+        }
+        break;
+    }
+    testCases_done = FALSE;
+
+}
+
+int32 testCases_a_int32;
+#ifdef BITS64
+int64 testCases_a_int64;
+#endif
+float32 testCases_a_float32;
+float32 testCases_b_float32;
+float64 testCases_a_float64;
+float64 testCases_b_float64;
+#ifdef FLOATX80
+floatx80 testCases_a_floatx80;
+floatx80 testCases_b_floatx80;
+#endif
+#ifdef FLOAT128
+float128 testCases_a_float128;
+float128 testCases_b_float128;
+#endif
+
+void testCases_next( void )
+{
+
+    switch ( level ) {
+     case 1:
+        switch ( sequenceType ) {
+         case testCases_sequence_a_int32:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_int32 = int32RandomP3();
+                break;
+             case 1:
+                testCases_a_int32 = int32RandomPInf();
+                break;
+             case 2:
+                testCases_a_int32 = int32NextP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#ifdef BITS64
+         case testCases_sequence_a_int64:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_int64 = int64RandomP3();
+                break;
+             case 1:
+                testCases_a_int64 = int64RandomPInf();
+                break;
+             case 2:
+                testCases_a_int64 = int64NextP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+         case testCases_sequence_a_float32:
+            switch ( subcase ) {
+             case 0:
+             case 1:
+                testCases_a_float32 = float32Random();
+                break;
+             case 2:
+                testCases_a_float32 = float32NextQOutP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float32:
+            switch ( subcase ) {
+             case 0:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float32 = float32NextQInP1( &sequenceA );
+                }
+                current_b_float32 = float32NextQInP1( &sequenceB );
+             case 2:
+             case 4:
+                testCases_a_float32 = float32Random();
+                testCases_b_float32 = float32Random();
+                break;
+             case 1:
+                testCases_a_float32 = current_a_float32;
+                testCases_b_float32 = float32Random();
+                break;
+             case 3:
+                testCases_a_float32 = float32Random();
+                testCases_b_float32 = current_b_float32;
+                break;
+             case 5:
+                testCases_a_float32 = current_a_float32;
+                testCases_b_float32 = current_b_float32;
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_a_float64:
+            switch ( subcase ) {
+             case 0:
+             case 1:
+                testCases_a_float64 = float64Random();
+                break;
+             case 2:
+                testCases_a_float64 = float64NextQOutP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float64:
+            switch ( subcase ) {
+             case 0:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float64 = float64NextQInP1( &sequenceA );
+                }
+                current_b_float64 = float64NextQInP1( &sequenceB );
+             case 2:
+             case 4:
+                testCases_a_float64 = float64Random();
+                testCases_b_float64 = float64Random();
+                break;
+             case 1:
+                testCases_a_float64 = current_a_float64;
+                testCases_b_float64 = float64Random();
+                break;
+             case 3:
+                testCases_a_float64 = float64Random();
+                testCases_b_float64 = current_b_float64;
+                break;
+             case 5:
+                testCases_a_float64 = current_a_float64;
+                testCases_b_float64 = current_b_float64;
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#ifdef FLOATX80
+         case testCases_sequence_a_floatx80:
+            switch ( subcase ) {
+             case 0:
+             case 1:
+                testCases_a_floatx80 = floatx80Random();
+                break;
+             case 2:
+                testCases_a_floatx80 = floatx80NextQOutP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_floatx80:
+            switch ( subcase ) {
+             case 0:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_floatx80 = floatx80NextQInP1( &sequenceA );
+                }
+                current_b_floatx80 = floatx80NextQInP1( &sequenceB );
+             case 2:
+             case 4:
+                testCases_a_floatx80 = floatx80Random();
+                testCases_b_floatx80 = floatx80Random();
+                break;
+             case 1:
+                testCases_a_floatx80 = current_a_floatx80;
+                testCases_b_floatx80 = floatx80Random();
+                break;
+             case 3:
+                testCases_a_floatx80 = floatx80Random();
+                testCases_b_floatx80 = current_b_floatx80;
+                break;
+             case 5:
+                testCases_a_floatx80 = current_a_floatx80;
+                testCases_b_floatx80 = current_b_floatx80;
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+#ifdef FLOAT128
+         case testCases_sequence_a_float128:
+            switch ( subcase ) {
+             case 0:
+             case 1:
+                testCases_a_float128 = float128Random();
+                break;
+             case 2:
+                testCases_a_float128 = float128NextQOutP1( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float128:
+            switch ( subcase ) {
+             case 0:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float128 = float128NextQInP1( &sequenceA );
+                }
+                current_b_float128 = float128NextQInP1( &sequenceB );
+             case 2:
+             case 4:
+                testCases_a_float128 = float128Random();
+                testCases_b_float128 = float128Random();
+                break;
+             case 1:
+                testCases_a_float128 = current_a_float128;
+                testCases_b_float128 = float128Random();
+                break;
+             case 3:
+                testCases_a_float128 = float128Random();
+                testCases_b_float128 = current_b_float128;
+                break;
+             case 5:
+                testCases_a_float128 = current_a_float128;
+                testCases_b_float128 = current_b_float128;
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+        }
+        break;
+     case 2:
+        switch ( sequenceType ) {
+         case testCases_sequence_a_int32:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_int32 = int32RandomP3();
+                break;
+             case 2:
+                testCases_a_int32 = int32RandomPInf();
+                break;
+             case 3:
+                subcase = -1;
+             case 1:
+                testCases_a_int32 = int32NextP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                break;
+            }
+            ++subcase;
+            break;
+#ifdef BITS64
+         case testCases_sequence_a_int64:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_int64 = int64RandomP3();
+                break;
+             case 2:
+                testCases_a_int64 = int64RandomPInf();
+                break;
+             case 3:
+                subcase = -1;
+             case 1:
+                testCases_a_int64 = int64NextP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+         case testCases_sequence_a_float32:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float32 = float32Random();
+                break;
+             case 1:
+                testCases_a_float32 = float32NextQOutP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float32:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float32 = float32Random();
+                testCases_b_float32 = float32Random();
+                break;
+             case 1:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float32 = float32NextQInP2( &sequenceA );
+                }
+                testCases_a_float32 = current_a_float32;
+                testCases_b_float32 = float32NextQInP2( &sequenceB );
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_a_float64:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float64 = float64Random();
+                break;
+             case 1:
+                testCases_a_float64 = float64NextQOutP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float64:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float64 = float64Random();
+                testCases_b_float64 = float64Random();
+                break;
+             case 1:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float64 = float64NextQInP2( &sequenceA );
+                }
+                testCases_a_float64 = current_a_float64;
+                testCases_b_float64 = float64NextQInP2( &sequenceB );
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#ifdef FLOATX80
+         case testCases_sequence_a_floatx80:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_floatx80 = floatx80Random();
+                break;
+             case 1:
+                testCases_a_floatx80 = floatx80NextQOutP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_floatx80:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_floatx80 = floatx80Random();
+                testCases_b_floatx80 = floatx80Random();
+                break;
+             case 1:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_floatx80 = floatx80NextQInP2( &sequenceA );
+                }
+                testCases_a_floatx80 = current_a_floatx80;
+                testCases_b_floatx80 = floatx80NextQInP2( &sequenceB );
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+#ifdef FLOAT128
+         case testCases_sequence_a_float128:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float128 = float128Random();
+                break;
+             case 1:
+                testCases_a_float128 = float128NextQOutP2( &sequenceA );
+                testCases_done = sequenceA.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+         case testCases_sequence_ab_float128:
+            switch ( subcase ) {
+             case 0:
+                testCases_a_float128 = float128Random();
+                testCases_b_float128 = float128Random();
+                break;
+             case 1:
+                if ( sequenceB.done ) {
+                    sequenceB.done = FALSE;
+                    current_a_float128 = float128NextQInP2( &sequenceA );
+                }
+                testCases_a_float128 = current_a_float128;
+                testCases_b_float128 = float128NextQInP2( &sequenceB );
+                testCases_done = sequenceA.done & sequenceB.done;
+                subcase = -1;
+                break;
+            }
+            ++subcase;
+            break;
+#endif
+        }
+        break;
+    }
+
+}
+
diff --git a/testfloat/testCases.h b/testfloat/testCases.h
new file mode 100644
index 00000000000..ed6ea2be46b
--- /dev/null
+++ b/testfloat/testCases.h
@@ -0,0 +1,69 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+void testCases_setLevel( int8 );
+
+void testCases_initSequence( int8 );
+enum {
+    testCases_sequence_a_int32,
+#ifdef BITS64
+    testCases_sequence_a_int64,
+#endif
+    testCases_sequence_a_float32,
+    testCases_sequence_ab_float32,
+    testCases_sequence_a_float64,
+    testCases_sequence_ab_float64,
+#ifdef FLOATX80
+    testCases_sequence_a_floatx80,
+    testCases_sequence_ab_floatx80,
+#endif
+#ifdef FLOAT128
+    testCases_sequence_a_float128,
+    testCases_sequence_ab_float128,
+#endif
+};
+
+extern uint32 testCases_total;
+extern flag testCases_done;
+
+void testCases_next( void );
+
+extern int32 testCases_a_int32;
+#ifdef BITS64
+extern int64 testCases_a_int64;
+#endif
+extern float32 testCases_a_float32;
+extern float32 testCases_b_float32;
+extern float64 testCases_a_float64;
+extern float64 testCases_b_float64;
+#ifdef FLOATX80
+extern floatx80 testCases_a_floatx80;
+extern floatx80 testCases_b_floatx80;
+#endif
+#ifdef FLOAT128
+extern float128 testCases_a_float128;
+extern float128 testCases_b_float128;
+#endif
+
diff --git a/testfloat/testFunction.c b/testfloat/testFunction.c
new file mode 100644
index 00000000000..687563b0d75
--- /dev/null
+++ b/testfloat/testFunction.c
@@ -0,0 +1,1149 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include "milieu.h"
+#include "softfloat.h"
+#include "testCases.h"
+#include "testLoops.h"
+#include "systmodes.h"
+#include "systflags.h"
+#include "systfloat.h"
+#include "testFunction.h"
+
+const functionT functions[ NUM_FUNCTIONS ] = {
+    { 0, 0, 0, 0 },
+    { "int32_to_float32",                1, FALSE, TRUE  },
+    { "int32_to_float64",                1, FALSE, FALSE },
+    { "int32_to_floatx80",               1, FALSE, FALSE },
+    { "int32_to_float128",               1, FALSE, FALSE },
+    { "int64_to_float32",                1, FALSE, TRUE  },
+    { "int64_to_float64",                1, FALSE, TRUE  },
+    { "int64_to_floatx80",               1, FALSE, FALSE },
+    { "int64_to_float128",               1, FALSE, FALSE },
+    { "float32_to_int32",                1, FALSE, TRUE  },
+    { "float32_to_int32_round_to_zero",  1, FALSE, FALSE },
+    { "float32_to_int64",                1, FALSE, TRUE  },
+    { "float32_to_int64_round_to_zero",  1, FALSE, FALSE },
+    { "float32_to_float64",              1, FALSE, FALSE },
+    { "float32_to_floatx80",             1, FALSE, FALSE },
+    { "float32_to_float128",             1, FALSE, FALSE },
+    { "float32_round_to_int",            1, FALSE, TRUE  },
+    { "float32_add",                     2, FALSE, TRUE  },
+    { "float32_sub",                     2, FALSE, TRUE  },
+    { "float32_mul",                     2, FALSE, TRUE  },
+    { "float32_div",                     2, FALSE, TRUE  },
+    { "float32_rem",                     2, FALSE, FALSE },
+    { "float32_sqrt",                    1, FALSE, TRUE  },
+    { "float32_eq",                      2, FALSE, FALSE },
+    { "float32_le",                      2, FALSE, FALSE },
+    { "float32_lt",                      2, FALSE, FALSE },
+    { "float32_eq_signaling",            2, FALSE, FALSE },
+    { "float32_le_quiet",                2, FALSE, FALSE },
+    { "float32_lt_quiet",                2, FALSE, FALSE },
+    { "float64_to_int32",                1, FALSE, TRUE  },
+    { "float64_to_int32_round_to_zero",  1, FALSE, FALSE },
+    { "float64_to_int64",                1, FALSE, TRUE  },
+    { "float64_to_int64_round_to_zero",  1, FALSE, FALSE },
+    { "float64_to_float32",              1, FALSE, TRUE  },
+    { "float64_to_floatx80",             1, FALSE, FALSE },
+    { "float64_to_float128",             1, FALSE, FALSE },
+    { "float64_round_to_int",            1, FALSE, TRUE  },
+    { "float64_add",                     2, FALSE, TRUE  },
+    { "float64_sub",                     2, FALSE, TRUE  },
+    { "float64_mul",                     2, FALSE, TRUE  },
+    { "float64_div",                     2, FALSE, TRUE  },
+    { "float64_rem",                     2, FALSE, FALSE },
+    { "float64_sqrt",                    1, FALSE, TRUE  },
+    { "float64_eq",                      2, FALSE, FALSE },
+    { "float64_le",                      2, FALSE, FALSE },
+    { "float64_lt",                      2, FALSE, FALSE },
+    { "float64_eq_signaling",            2, FALSE, FALSE },
+    { "float64_le_quiet",                2, FALSE, FALSE },
+    { "float64_lt_quiet",                2, FALSE, FALSE },
+    { "floatx80_to_int32",               1, FALSE, TRUE  },
+    { "floatx80_to_int32_round_to_zero", 1, FALSE, FALSE },
+    { "floatx80_to_int64",               1, FALSE, TRUE  },
+    { "floatx80_to_int64_round_to_zero", 1, FALSE, FALSE },
+    { "floatx80_to_float32",             1, FALSE, TRUE  },
+    { "floatx80_to_float64",             1, FALSE, TRUE  },
+    { "floatx80_to_float128",            1, FALSE, FALSE },
+    { "floatx80_round_to_int",           1, FALSE, TRUE  },
+    { "floatx80_add",                    2, TRUE,  TRUE  },
+    { "floatx80_sub",                    2, TRUE,  TRUE  },
+    { "floatx80_mul",                    2, TRUE,  TRUE  },
+    { "floatx80_div",                    2, TRUE,  TRUE  },
+    { "floatx80_rem",                    2, FALSE, FALSE },
+    { "floatx80_sqrt",                   1, TRUE,  TRUE  },
+    { "floatx80_eq",                     2, FALSE, FALSE },
+    { "floatx80_le",                     2, FALSE, FALSE },
+    { "floatx80_lt",                     2, FALSE, FALSE },
+    { "floatx80_eq_signaling",           2, FALSE, FALSE },
+    { "floatx80_le_quiet",               2, FALSE, FALSE },
+    { "floatx80_lt_quiet",               2, FALSE, FALSE },
+    { "float128_to_int32",               1, FALSE, TRUE  },
+    { "float128_to_int32_round_to_zero", 1, FALSE, FALSE },
+    { "float128_to_int64",               1, FALSE, TRUE  },
+    { "float128_to_int64_round_to_zero", 1, FALSE, FALSE },
+    { "float128_to_float32",             1, FALSE, TRUE  },
+    { "float128_to_float64",             1, FALSE, TRUE  },
+    { "float128_to_floatx80",            1, FALSE, TRUE  },
+    { "float128_round_to_int",           1, FALSE, TRUE  },
+    { "float128_add",                    2, FALSE, TRUE  },
+    { "float128_sub",                    2, FALSE, TRUE  },
+    { "float128_mul",                    2, FALSE, TRUE  },
+    { "float128_div",                    2, FALSE, TRUE  },
+    { "float128_rem",                    2, FALSE, FALSE },
+    { "float128_sqrt",                   1, FALSE, TRUE  },
+    { "float128_eq",                     2, FALSE, FALSE },
+    { "float128_le",                     2, FALSE, FALSE },
+    { "float128_lt",                     2, FALSE, FALSE },
+    { "float128_eq_signaling",           2, FALSE, FALSE },
+    { "float128_le_quiet",               2, FALSE, FALSE },
+    { "float128_lt_quiet",               2, FALSE, FALSE },
+};
+
+const flag functionExists[ NUM_FUNCTIONS ] = {
+    0,
+#ifdef SYST_INT32_TO_FLOAT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT32_TO_FLOAT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT32_TO_FLOATX80
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT32_TO_FLOAT128
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT64_TO_FLOAT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT64_TO_FLOAT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT64_TO_FLOATX80
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_INT64_TO_FLOAT128
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_INT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_INT32_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_INT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_INT64_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_FLOAT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_FLOATX80
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_TO_FLOAT128
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_ROUND_TO_INT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_ADD
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_SUB
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_MUL
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_DIV
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_REM
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_SQRT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_EQ
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_LE
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_LT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_EQ_SIGNALING
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_LE_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT32_LT_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_INT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_INT32_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_INT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_INT64_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_FLOAT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_FLOATX80
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_TO_FLOAT128
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_ROUND_TO_INT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_ADD
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_SUB
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_MUL
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_DIV
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_REM
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_SQRT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_EQ
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_LE
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_LT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_EQ_SIGNALING
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_LE_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT64_LT_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_INT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_INT32_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_INT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_INT64_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT128
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_ROUND_TO_INT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_ADD
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_SUB
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_MUL
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_DIV
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_REM
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_SQRT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_EQ
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_LE
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_LT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_EQ_SIGNALING
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_LE_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOATX80_LT_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_INT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_INT32_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_INT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_INT64_ROUND_TO_ZERO
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_FLOAT32
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_FLOAT64
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_TO_FLOATX80
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_ROUND_TO_INT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_ADD
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_SUB
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_MUL
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_DIV
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_REM
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_SQRT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_EQ
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_LE
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_LT
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_EQ_SIGNALING
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_LE_QUIET
+    1,
+#else
+    0,
+#endif
+#ifdef SYST_FLOAT128_LT_QUIET
+    1,
+#else
+    0,
+#endif
+};
+
+static void
+ testFunctionVariety(
+     uint8 functionCode, int8 roundingPrecision, int8 roundingMode )
+{
+    uint8 roundingCode;
+
+    functionName = functions[ functionCode ].name;
+#ifdef FLOATX80
+    if ( roundingPrecision == 32 ) {
+        roundingPrecisionName = "32";
+    }
+    else if ( roundingPrecision == 64 ) {
+        roundingPrecisionName = "64";
+    }
+    else if ( roundingPrecision == 80 ) {
+        roundingPrecisionName = "80";
+    }
+    else {
+        roundingPrecision = 80;
+        roundingPrecisionName = 0;
+    }
+    floatx80_rounding_precision = roundingPrecision;
+    syst_float_set_rounding_precision( roundingPrecision );
+#endif
+    switch ( roundingMode ) {
+     case 0:
+        roundingModeName = 0;
+        roundingCode = float_round_nearest_even;
+        break;
+     case ROUND_NEAREST_EVEN:
+        roundingModeName = "nearest_even";
+        roundingCode = float_round_nearest_even;
+        break;
+     case ROUND_TO_ZERO:
+        roundingModeName = "to_zero";
+        roundingCode = float_round_to_zero;
+        break;
+     case ROUND_DOWN:
+        roundingModeName = "down";
+        roundingCode = float_round_down;
+        break;
+     case ROUND_UP:
+        roundingModeName = "up";
+        roundingCode = float_round_up;
+        break;
+    }
+    float_rounding_mode = roundingCode;
+    syst_float_set_rounding_mode( roundingCode );
+    fputs( "Testing ", stderr );
+    writeFunctionName( stderr );
+    fputs( ".\n", stderr );
+    switch ( functionCode ) {
+#ifdef SYST_INT32_TO_FLOAT32
+     case INT32_TO_FLOAT32:
+        test_a_int32_z_float32( int32_to_float32, syst_int32_to_float32 );
+        break;
+#endif
+#ifdef SYST_INT32_TO_FLOAT64
+     case INT32_TO_FLOAT64:
+        test_a_int32_z_float64( int32_to_float64, syst_int32_to_float64 );
+        break;
+#endif
+#ifdef SYST_INT32_TO_FLOATX80
+     case INT32_TO_FLOATX80:
+        test_a_int32_z_floatx80( int32_to_floatx80, syst_int32_to_floatx80 );
+        break;
+#endif
+#ifdef SYST_INT32_TO_FLOAT128
+     case INT32_TO_FLOAT128:
+        test_a_int32_z_float128( int32_to_float128, syst_int32_to_float128 );
+        break;
+#endif
+#ifdef SYST_INT64_TO_FLOAT32
+     case INT64_TO_FLOAT32:
+        test_a_int64_z_float32( int64_to_float32, syst_int64_to_float32 );
+        break;
+#endif
+#ifdef SYST_INT64_TO_FLOAT64
+     case INT64_TO_FLOAT64:
+        test_a_int64_z_float64( int64_to_float64, syst_int64_to_float64 );
+        break;
+#endif
+#ifdef SYST_INT64_TO_FLOATX80
+     case INT64_TO_FLOATX80:
+        test_a_int64_z_floatx80( int64_to_floatx80, syst_int64_to_floatx80 );
+        break;
+#endif
+#ifdef SYST_INT64_TO_FLOAT128
+     case INT64_TO_FLOAT128:
+        test_a_int64_z_float128( int64_to_float128, syst_int64_to_float128 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_INT32
+     case FLOAT32_TO_INT32:
+        test_a_float32_z_int32( float32_to_int32, syst_float32_to_int32 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_INT32_ROUND_TO_ZERO
+     case FLOAT32_TO_INT32_ROUND_TO_ZERO:
+        test_a_float32_z_int32(
+            float32_to_int32_round_to_zero,
+            syst_float32_to_int32_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_INT64
+     case FLOAT32_TO_INT64:
+        test_a_float32_z_int64( float32_to_int64, syst_float32_to_int64 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_INT64_ROUND_TO_ZERO
+     case FLOAT32_TO_INT64_ROUND_TO_ZERO:
+        test_a_float32_z_int64(
+            float32_to_int64_round_to_zero,
+            syst_float32_to_int64_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_FLOAT64
+     case FLOAT32_TO_FLOAT64:
+        test_a_float32_z_float64(
+            float32_to_float64, syst_float32_to_float64 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_FLOATX80
+     case FLOAT32_TO_FLOATX80:
+        test_a_float32_z_floatx80(
+            float32_to_floatx80, syst_float32_to_floatx80 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_TO_FLOAT128
+     case FLOAT32_TO_FLOAT128:
+        test_a_float32_z_float128(
+            float32_to_float128, syst_float32_to_float128 );
+        break;
+#endif
+#ifdef SYST_FLOAT32_ROUND_TO_INT
+     case FLOAT32_ROUND_TO_INT:
+        test_az_float32( float32_round_to_int, syst_float32_round_to_int );
+        break;
+#endif
+#ifdef SYST_FLOAT32_ADD
+     case FLOAT32_ADD:
+        test_abz_float32( float32_add, syst_float32_add );
+        break;
+#endif
+#ifdef SYST_FLOAT32_SUB
+     case FLOAT32_SUB:
+        test_abz_float32( float32_sub, syst_float32_sub );
+        break;
+#endif
+#ifdef SYST_FLOAT32_MUL
+     case FLOAT32_MUL:
+        test_abz_float32( float32_mul, syst_float32_mul );
+        break;
+#endif
+#ifdef SYST_FLOAT32_DIV
+     case FLOAT32_DIV:
+        test_abz_float32( float32_div, syst_float32_div );
+        break;
+#endif
+#ifdef SYST_FLOAT32_REM
+     case FLOAT32_REM:
+        test_abz_float32( float32_rem, syst_float32_rem );
+        break;
+#endif
+#ifdef SYST_FLOAT32_SQRT
+     case FLOAT32_SQRT:
+        test_az_float32( float32_sqrt, syst_float32_sqrt );
+        break;
+#endif
+#ifdef SYST_FLOAT32_EQ
+     case FLOAT32_EQ:
+        test_ab_float32_z_flag( float32_eq, syst_float32_eq );
+        break;
+#endif
+#ifdef SYST_FLOAT32_LE
+     case FLOAT32_LE:
+        test_ab_float32_z_flag( float32_le, syst_float32_le );
+        break;
+#endif
+#ifdef SYST_FLOAT32_LT
+     case FLOAT32_LT:
+        test_ab_float32_z_flag( float32_lt, syst_float32_lt );
+        break;
+#endif
+#ifdef SYST_FLOAT32_EQ_SIGNALING
+     case FLOAT32_EQ_SIGNALING:
+        test_ab_float32_z_flag(
+            float32_eq_signaling, syst_float32_eq_signaling );
+        break;
+#endif
+#ifdef SYST_FLOAT32_LE_QUIET
+     case FLOAT32_LE_QUIET:
+        test_ab_float32_z_flag( float32_le_quiet, syst_float32_le_quiet );
+        break;
+#endif
+#ifdef SYST_FLOAT32_LT_QUIET
+     case FLOAT32_LT_QUIET:
+        test_ab_float32_z_flag( float32_lt_quiet, syst_float32_lt_quiet );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_INT32
+     case FLOAT64_TO_INT32:
+        test_a_float64_z_int32( float64_to_int32, syst_float64_to_int32 );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_INT32_ROUND_TO_ZERO
+     case FLOAT64_TO_INT32_ROUND_TO_ZERO:
+        test_a_float64_z_int32(
+            float64_to_int32_round_to_zero,
+            syst_float64_to_int32_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_INT64
+     case FLOAT64_TO_INT64:
+        test_a_float64_z_int64( float64_to_int64, syst_float64_to_int64 );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_INT64_ROUND_TO_ZERO
+     case FLOAT64_TO_INT64_ROUND_TO_ZERO:
+        test_a_float64_z_int64(
+            float64_to_int64_round_to_zero,
+            syst_float64_to_int64_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_FLOAT32
+     case FLOAT64_TO_FLOAT32:
+        test_a_float64_z_float32(
+            float64_to_float32, syst_float64_to_float32 );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_FLOATX80
+     case FLOAT64_TO_FLOATX80:
+        test_a_float64_z_floatx80(
+            float64_to_floatx80, syst_float64_to_floatx80 );
+        break;
+#endif
+#ifdef SYST_FLOAT64_TO_FLOAT128
+     case FLOAT64_TO_FLOAT128:
+        test_a_float64_z_float128(
+            float64_to_float128, syst_float64_to_float128 );
+        break;
+#endif
+#ifdef SYST_FLOAT64_ROUND_TO_INT
+     case FLOAT64_ROUND_TO_INT:
+        test_az_float64( float64_round_to_int, syst_float64_round_to_int );
+        break;
+#endif
+#ifdef SYST_FLOAT64_ADD
+     case FLOAT64_ADD:
+        test_abz_float64( float64_add, syst_float64_add );
+        break;
+#endif
+#ifdef SYST_FLOAT64_SUB
+     case FLOAT64_SUB:
+        test_abz_float64( float64_sub, syst_float64_sub );
+        break;
+#endif
+#ifdef SYST_FLOAT64_MUL
+     case FLOAT64_MUL:
+        test_abz_float64( float64_mul, syst_float64_mul );
+        break;
+#endif
+#ifdef SYST_FLOAT64_DIV
+     case FLOAT64_DIV:
+        test_abz_float64( float64_div, syst_float64_div );
+        break;
+#endif
+#ifdef SYST_FLOAT64_REM
+     case FLOAT64_REM:
+        test_abz_float64( float64_rem, syst_float64_rem );
+        break;
+#endif
+#ifdef SYST_FLOAT64_SQRT
+     case FLOAT64_SQRT:
+        test_az_float64( float64_sqrt, syst_float64_sqrt );
+        break;
+#endif
+#ifdef SYST_FLOAT64_EQ
+     case FLOAT64_EQ:
+        test_ab_float64_z_flag( float64_eq, syst_float64_eq );
+        break;
+#endif
+#ifdef SYST_FLOAT64_LE
+     case FLOAT64_LE:
+        test_ab_float64_z_flag( float64_le, syst_float64_le );
+        break;
+#endif
+#ifdef SYST_FLOAT64_LT
+     case FLOAT64_LT:
+        test_ab_float64_z_flag( float64_lt, syst_float64_lt );
+        break;
+#endif
+#ifdef SYST_FLOAT64_EQ_SIGNALING
+     case FLOAT64_EQ_SIGNALING:
+        test_ab_float64_z_flag(
+            float64_eq_signaling, syst_float64_eq_signaling );
+        break;
+#endif
+#ifdef SYST_FLOAT64_LE_QUIET
+     case FLOAT64_LE_QUIET:
+        test_ab_float64_z_flag( float64_le_quiet, syst_float64_le_quiet );
+        break;
+#endif
+#ifdef SYST_FLOAT64_LT_QUIET
+     case FLOAT64_LT_QUIET:
+        test_ab_float64_z_flag( float64_lt_quiet, syst_float64_lt_quiet );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_INT32
+     case FLOATX80_TO_INT32:
+        test_a_floatx80_z_int32( floatx80_to_int32, syst_floatx80_to_int32 );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_INT32_ROUND_TO_ZERO
+     case FLOATX80_TO_INT32_ROUND_TO_ZERO:
+        test_a_floatx80_z_int32(
+            floatx80_to_int32_round_to_zero,
+            syst_floatx80_to_int32_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_INT64
+     case FLOATX80_TO_INT64:
+        test_a_floatx80_z_int64( floatx80_to_int64, syst_floatx80_to_int64 );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_INT64_ROUND_TO_ZERO
+     case FLOATX80_TO_INT64_ROUND_TO_ZERO:
+        test_a_floatx80_z_int64(
+            floatx80_to_int64_round_to_zero,
+            syst_floatx80_to_int64_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT32
+     case FLOATX80_TO_FLOAT32:
+        test_a_floatx80_z_float32(
+            floatx80_to_float32, syst_floatx80_to_float32 );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT64
+     case FLOATX80_TO_FLOAT64:
+        test_a_floatx80_z_float64(
+            floatx80_to_float64, syst_floatx80_to_float64 );
+        break;
+#endif
+#ifdef SYST_FLOATX80_TO_FLOAT128
+     case FLOATX80_TO_FLOAT128:
+        test_a_floatx80_z_float128(
+            floatx80_to_float128, syst_floatx80_to_float128 );
+        break;
+#endif
+#ifdef SYST_FLOATX80_ROUND_TO_INT
+     case FLOATX80_ROUND_TO_INT:
+        test_az_floatx80( floatx80_round_to_int, syst_floatx80_round_to_int );
+        break;
+#endif
+#ifdef SYST_FLOATX80_ADD
+     case FLOATX80_ADD:
+        test_abz_floatx80( floatx80_add, syst_floatx80_add );
+        break;
+#endif
+#ifdef SYST_FLOATX80_SUB
+     case FLOATX80_SUB:
+        test_abz_floatx80( floatx80_sub, syst_floatx80_sub );
+        break;
+#endif
+#ifdef SYST_FLOATX80_MUL
+     case FLOATX80_MUL:
+        test_abz_floatx80( floatx80_mul, syst_floatx80_mul );
+        break;
+#endif
+#ifdef SYST_FLOATX80_DIV
+     case FLOATX80_DIV:
+        test_abz_floatx80( floatx80_div, syst_floatx80_div );
+        break;
+#endif
+#ifdef SYST_FLOATX80_REM
+     case FLOATX80_REM:
+        test_abz_floatx80( floatx80_rem, syst_floatx80_rem );
+        break;
+#endif
+#ifdef SYST_FLOATX80_SQRT
+     case FLOATX80_SQRT:
+        test_az_floatx80( floatx80_sqrt, syst_floatx80_sqrt );
+        break;
+#endif
+#ifdef SYST_FLOATX80_EQ
+     case FLOATX80_EQ:
+        test_ab_floatx80_z_flag( floatx80_eq, syst_floatx80_eq );
+        break;
+#endif
+#ifdef SYST_FLOATX80_LE
+     case FLOATX80_LE:
+        test_ab_floatx80_z_flag( floatx80_le, syst_floatx80_le );
+        break;
+#endif
+#ifdef SYST_FLOATX80_LT
+     case FLOATX80_LT:
+        test_ab_floatx80_z_flag( floatx80_lt, syst_floatx80_lt );
+        break;
+#endif
+#ifdef SYST_FLOATX80_EQ_SIGNALING
+     case FLOATX80_EQ_SIGNALING:
+        test_ab_floatx80_z_flag(
+            floatx80_eq_signaling, syst_floatx80_eq_signaling );
+        break;
+#endif
+#ifdef SYST_FLOATX80_LE_QUIET
+     case FLOATX80_LE_QUIET:
+        test_ab_floatx80_z_flag( floatx80_le_quiet, syst_floatx80_le_quiet );
+        break;
+#endif
+#ifdef SYST_FLOATX80_LT_QUIET
+     case FLOATX80_LT_QUIET:
+        test_ab_floatx80_z_flag( floatx80_lt_quiet, syst_floatx80_lt_quiet );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_INT32
+     case FLOAT128_TO_INT32:
+        test_a_float128_z_int32( float128_to_int32, syst_float128_to_int32 );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_INT32_ROUND_TO_ZERO
+     case FLOAT128_TO_INT32_ROUND_TO_ZERO:
+        test_a_float128_z_int32(
+            float128_to_int32_round_to_zero,
+            syst_float128_to_int32_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_INT64
+     case FLOAT128_TO_INT64:
+        test_a_float128_z_int64( float128_to_int64, syst_float128_to_int64 );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_INT64_ROUND_TO_ZERO
+     case FLOAT128_TO_INT64_ROUND_TO_ZERO:
+        test_a_float128_z_int64(
+            float128_to_int64_round_to_zero,
+            syst_float128_to_int64_round_to_zero
+        );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_FLOAT32
+     case FLOAT128_TO_FLOAT32:
+        test_a_float128_z_float32(
+            float128_to_float32, syst_float128_to_float32 );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_FLOAT64
+     case FLOAT128_TO_FLOAT64:
+        test_a_float128_z_float64(
+            float128_to_float64, syst_float128_to_float64 );
+        break;
+#endif
+#ifdef SYST_FLOAT128_TO_FLOATX80
+     case FLOAT128_TO_FLOATX80:
+        test_a_float128_z_floatx80(
+            float128_to_floatx80, syst_float128_to_floatx80 );
+        break;
+#endif
+#ifdef SYST_FLOAT128_ROUND_TO_INT
+     case FLOAT128_ROUND_TO_INT:
+        test_az_float128( float128_round_to_int, syst_float128_round_to_int );
+        break;
+#endif
+#ifdef SYST_FLOAT128_ADD
+     case FLOAT128_ADD:
+        test_abz_float128( float128_add, syst_float128_add );
+        break;
+#endif
+#ifdef SYST_FLOAT128_SUB
+     case FLOAT128_SUB:
+        test_abz_float128( float128_sub, syst_float128_sub );
+        break;
+#endif
+#ifdef SYST_FLOAT128_MUL
+     case FLOAT128_MUL:
+        test_abz_float128( float128_mul, syst_float128_mul );
+        break;
+#endif
+#ifdef SYST_FLOAT128_DIV
+     case FLOAT128_DIV:
+        test_abz_float128( float128_div, syst_float128_div );
+        break;
+#endif
+#ifdef SYST_FLOAT128_REM
+     case FLOAT128_REM:
+        test_abz_float128( float128_rem, syst_float128_rem );
+        break;
+#endif
+#ifdef SYST_FLOAT128_SQRT
+     case FLOAT128_SQRT:
+        test_az_float128( float128_sqrt, syst_float128_sqrt );
+        break;
+#endif
+#ifdef SYST_FLOAT128_EQ
+     case FLOAT128_EQ:
+        test_ab_float128_z_flag( float128_eq, syst_float128_eq );
+        break;
+#endif
+#ifdef SYST_FLOAT128_LE
+     case FLOAT128_LE:
+        test_ab_float128_z_flag( float128_le, syst_float128_le );
+        break;
+#endif
+#ifdef SYST_FLOAT128_LT
+     case FLOAT128_LT:
+        test_ab_float128_z_flag( float128_lt, syst_float128_lt );
+        break;
+#endif
+#ifdef SYST_FLOAT128_EQ_SIGNALING
+     case FLOAT128_EQ_SIGNALING:
+        test_ab_float128_z_flag(
+            float128_eq_signaling, syst_float128_eq_signaling );
+        break;
+#endif
+#ifdef SYST_FLOAT128_LE_QUIET
+     case FLOAT128_LE_QUIET:
+        test_ab_float128_z_flag( float128_le_quiet, syst_float128_le_quiet );
+        break;
+#endif
+#ifdef SYST_FLOAT128_LT_QUIET
+     case FLOAT128_LT_QUIET:
+        test_ab_float128_z_flag( float128_lt_quiet, syst_float128_lt_quiet );
+        break;
+#endif
+    }
+    if ( ( errorStop && anyErrors ) || stop ) exitWithStatus();
+
+}
+
+void
+ testFunction(
+     uint8 functionCode, int8 roundingPrecisionIn, int8 roundingModeIn )
+{
+    int8 roundingPrecision, roundingMode;
+
+    roundingPrecision = 32;
+    for (;;) {
+        if ( ! functions[ functionCode ].roundingPrecision ) {
+            roundingPrecision = 0;
+        }
+        else if ( roundingPrecisionIn ) {
+            roundingPrecision = roundingPrecisionIn;
+        }
+        for ( roundingMode = 1;
+              roundingMode < NUM_ROUNDINGMODES;
+              ++roundingMode
+            ) {
+            if ( ! functions[ functionCode ].roundingMode ) {
+                roundingMode = 0;
+            }
+            else if ( roundingModeIn ) {
+                roundingMode = roundingModeIn;
+            }
+            testFunctionVariety(
+                functionCode, roundingPrecision, roundingMode );
+            if ( roundingModeIn || ! roundingMode ) break;
+        }
+        if ( roundingPrecisionIn || ! roundingPrecision ) break;
+        if ( roundingPrecision == 80 ) {
+            break;
+        }
+        else if ( roundingPrecision == 64 ) {
+            roundingPrecision = 80;
+        }
+        else if ( roundingPrecision == 32 ) {
+            roundingPrecision = 64;
+        }
+    }
+
+}
+
diff --git a/testfloat/testFunction.h b/testfloat/testFunction.h
new file mode 100644
index 00000000000..04bf856046d
--- /dev/null
+++ b/testfloat/testFunction.h
@@ -0,0 +1,135 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+enum {
+    INT32_TO_FLOAT32 = 1,
+    INT32_TO_FLOAT64,
+    INT32_TO_FLOATX80,
+    INT32_TO_FLOAT128,
+    INT64_TO_FLOAT32,
+    INT64_TO_FLOAT64,
+    INT64_TO_FLOATX80,
+    INT64_TO_FLOAT128,
+    FLOAT32_TO_INT32,
+    FLOAT32_TO_INT32_ROUND_TO_ZERO,
+    FLOAT32_TO_INT64,
+    FLOAT32_TO_INT64_ROUND_TO_ZERO,
+    FLOAT32_TO_FLOAT64,
+    FLOAT32_TO_FLOATX80,
+    FLOAT32_TO_FLOAT128,
+    FLOAT32_ROUND_TO_INT,
+    FLOAT32_ADD,
+    FLOAT32_SUB,
+    FLOAT32_MUL,
+    FLOAT32_DIV,
+    FLOAT32_REM,
+    FLOAT32_SQRT,
+    FLOAT32_EQ,
+    FLOAT32_LE,
+    FLOAT32_LT,
+    FLOAT32_EQ_SIGNALING,
+    FLOAT32_LE_QUIET,
+    FLOAT32_LT_QUIET,
+    FLOAT64_TO_INT32,
+    FLOAT64_TO_INT32_ROUND_TO_ZERO,
+    FLOAT64_TO_INT64,
+    FLOAT64_TO_INT64_ROUND_TO_ZERO,
+    FLOAT64_TO_FLOAT32,
+    FLOAT64_TO_FLOATX80,
+    FLOAT64_TO_FLOAT128,
+    FLOAT64_ROUND_TO_INT,
+    FLOAT64_ADD,
+    FLOAT64_SUB,
+    FLOAT64_MUL,
+    FLOAT64_DIV,
+    FLOAT64_REM,
+    FLOAT64_SQRT,
+    FLOAT64_EQ,
+    FLOAT64_LE,
+    FLOAT64_LT,
+    FLOAT64_EQ_SIGNALING,
+    FLOAT64_LE_QUIET,
+    FLOAT64_LT_QUIET,
+    FLOATX80_TO_INT32,
+    FLOATX80_TO_INT32_ROUND_TO_ZERO,
+    FLOATX80_TO_INT64,
+    FLOATX80_TO_INT64_ROUND_TO_ZERO,
+    FLOATX80_TO_FLOAT32,
+    FLOATX80_TO_FLOAT64,
+    FLOATX80_TO_FLOAT128,
+    FLOATX80_ROUND_TO_INT,
+    FLOATX80_ADD,
+    FLOATX80_SUB,
+    FLOATX80_MUL,
+    FLOATX80_DIV,
+    FLOATX80_REM,
+    FLOATX80_SQRT,
+    FLOATX80_EQ,
+    FLOATX80_LE,
+    FLOATX80_LT,
+    FLOATX80_EQ_SIGNALING,
+    FLOATX80_LE_QUIET,
+    FLOATX80_LT_QUIET,
+    FLOAT128_TO_INT32,
+    FLOAT128_TO_INT32_ROUND_TO_ZERO,
+    FLOAT128_TO_INT64,
+    FLOAT128_TO_INT64_ROUND_TO_ZERO,
+    FLOAT128_TO_FLOAT32,
+    FLOAT128_TO_FLOAT64,
+    FLOAT128_TO_FLOATX80,
+    FLOAT128_ROUND_TO_INT,
+    FLOAT128_ADD,
+    FLOAT128_SUB,
+    FLOAT128_MUL,
+    FLOAT128_DIV,
+    FLOAT128_REM,
+    FLOAT128_SQRT,
+    FLOAT128_EQ,
+    FLOAT128_LE,
+    FLOAT128_LT,
+    FLOAT128_EQ_SIGNALING,
+    FLOAT128_LE_QUIET,
+    FLOAT128_LT_QUIET,
+    NUM_FUNCTIONS
+};
+
+typedef struct {
+    char *name;
+    int8 numInputs;
+    flag roundingPrecision, roundingMode;
+} functionT;
+extern const functionT functions[ NUM_FUNCTIONS ];
+extern const flag functionExists[ NUM_FUNCTIONS ];
+
+enum {
+    ROUND_NEAREST_EVEN = 1,
+    ROUND_TO_ZERO,
+    ROUND_DOWN,
+    ROUND_UP,
+    NUM_ROUNDINGMODES
+};
+
+void testFunction( uint8, int8, int8 );
+
diff --git a/testfloat/testLoops.c b/testfloat/testLoops.c
new file mode 100644
index 00000000000..8ba92f313a4
--- /dev/null
+++ b/testfloat/testLoops.c
@@ -0,0 +1,2713 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdlib.h>
+#include <stdio.h>
+#include "milieu.h"
+#include "softfloat.h"
+#include "testCases.h"
+#include "writeHex.h"
+#include "testLoops.h"
+
+volatile flag stop = FALSE;
+
+char *trueName, *testName;
+flag forever, errorStop;
+uint32 maxErrorCount = 0;
+flag checkNaNs = FALSE;
+int8 *trueFlagsPtr;
+int8 ( *testFlagsFunctionPtr )( void );
+char *functionName;
+char *roundingPrecisionName, *roundingModeName, *tininessModeName;
+flag anyErrors = FALSE;
+
+void writeFunctionName( FILE *stream )
+{
+
+    fputs( functionName, stream );
+    if ( roundingModeName ) {
+        if ( roundingPrecisionName ) {
+            fputs( ", precision ", stream );
+            fputs( roundingPrecisionName, stream );
+        }
+        fputs( ", rounding ", stream );
+        fputs( roundingModeName, stream );
+        if ( tininessModeName ) {
+            fputs( ", tininess ", stream );
+            fputs( tininessModeName, stream );
+            fputs( " rounding", stream );
+        }
+    }
+
+}
+
+void exitWithStatus( void )
+{
+
+    exit( anyErrors ? EXIT_FAILURE : EXIT_SUCCESS );
+
+}
+
+static uint32 tenthousandsCount, errorCount = 0;
+
+static void writeTestsTotal( void )
+{
+
+    if ( forever ) {
+        fputs( "Unbounded tests.\n", stderr );
+    }
+    else {
+        fprintf( stderr, "\r%d tests total.\n", testCases_total );
+    }
+
+}
+
+static void writeTestsPerformed( int16 count )
+{
+
+    if ( tenthousandsCount ) {
+        fprintf(
+            stderr, "\r%d%04d tests performed", tenthousandsCount, count );
+    }
+    else {
+        fprintf( stderr, "\r%d tests performed", count );
+    }
+    if ( errorCount ) {
+        fprintf(
+            stderr,
+            "; %d error%s found.\n",
+            errorCount,
+            ( errorCount == 1 ) ? "" : "s"
+        );
+    }
+    else {
+        fputs( ".\n", stderr );
+        fputs( "No errors found in ", stdout );
+        writeFunctionName( stdout );
+        fputs( ".\n", stdout );
+        fflush( stdout );
+    }
+
+}
+
+static void checkEarlyExit( void )
+{
+
+    ++tenthousandsCount;
+    if ( stop ) {
+        writeTestsPerformed( 0 );
+        exitWithStatus();
+    }
+    fprintf( stderr, "\r%3d0000", tenthousandsCount );
+
+}
+
+static void writeErrorFound( int16 count )
+{
+
+    fputc( '\r', stderr );
+    if ( errorCount == 1 ) {
+        fputs( "Errors found in ", stdout );
+        writeFunctionName( stdout );
+        fputs( ":\n", stdout );
+    }
+    if ( stop ) {
+        writeTestsPerformed( count );
+        exitWithStatus();
+    }
+    anyErrors = TRUE;
+
+}
+
+INLINE void writeInput_a_int32( void )
+{
+
+    writeHex_bits32( testCases_a_int32, stdout );
+
+}
+
+#ifdef BITS64
+
+INLINE void writeInput_a_int64( void )
+{
+
+    writeHex_bits64( testCases_a_int64, stdout );
+
+}
+
+#endif
+
+INLINE void writeInput_a_float32( void )
+{
+
+    writeHex_float32( testCases_a_float32, stdout );
+
+}
+
+static void writeInputs_ab_float32( void )
+{
+
+    writeHex_float32( testCases_a_float32, stdout );
+    fputs( "  ", stdout );
+    writeHex_float32( testCases_b_float32, stdout );
+
+}
+
+INLINE void writeInput_a_float64( void )
+{
+
+    writeHex_float64( testCases_a_float64, stdout );
+
+}
+
+static void writeInputs_ab_float64( void )
+{
+
+    writeHex_float64( testCases_a_float64, stdout );
+    fputs( "  ", stdout );
+    writeHex_float64( testCases_b_float64, stdout );
+
+}
+
+#ifdef FLOATX80
+
+INLINE void writeInput_a_floatx80( void )
+{
+
+    writeHex_floatx80( testCases_a_floatx80, stdout );
+
+}
+
+static void writeInputs_ab_floatx80( void )
+{
+
+    writeHex_floatx80( testCases_a_floatx80, stdout );
+    fputs( "  ", stdout );
+    writeHex_floatx80( testCases_b_floatx80, stdout );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+INLINE void writeInput_a_float128( void )
+{
+
+    writeHex_float128( testCases_a_float128, stdout );
+
+}
+
+static void writeInputs_ab_float128( void )
+{
+
+    writeHex_float128( testCases_a_float128, stdout );
+    fputs( "  ", stdout );
+    writeHex_float128( testCases_b_float128, stdout );
+
+}
+
+#endif
+
+static void
+ writeOutputs_z_flag(
+     flag trueZ, uint8 trueFlags, flag testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_flag( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_flag( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+static void
+ writeOutputs_z_int32(
+     int32 trueZ, uint8 trueFlags, int32 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_bits32( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_bits32( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+#ifdef BITS64
+
+static void
+ writeOutputs_z_int64(
+     int64 trueZ, uint8 trueFlags, int64 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_bits64( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_bits64( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+#endif
+
+static void
+ writeOutputs_z_float32(
+     float32 trueZ, uint8 trueFlags, float32 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float32( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float32( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+static void
+ writeOutputs_z_float64(
+     float64 trueZ, uint8 trueFlags, float64 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float64( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float64( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+#ifdef FLOATX80
+
+static void
+ writeOutputs_z_floatx80(
+     floatx80 trueZ, uint8 trueFlags, floatx80 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_floatx80( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "  ", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_floatx80( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+static void
+ writeOutputs_z_float128(
+     float128 trueZ, uint8 trueFlags, float128 testZ, uint8 testFlags )
+{
+
+    fputs( trueName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float128( trueZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( trueFlags, stdout );
+    fputs( "\n\t", stdout );
+    fputs( testName, stdout );
+    fputs( ": ", stdout );
+    writeHex_float128( testZ, stdout );
+    fputc( ' ', stdout );
+    writeHex_float_flags( testFlags, stdout );
+    fputc( '\n', stdout );
+
+}
+
+#endif
+
+INLINE flag float32_isNaN( float32 a )
+{
+
+    return 0x7F800000 < ( a & 0x7FFFFFFF );
+
+}
+
+#ifdef BITS64
+
+INLINE flag float64_same( float64 a, float64 b )
+{
+
+    return a == b;
+
+}
+
+INLINE flag float64_isNaN( float64 a )
+{
+
+    return LIT64( 0x7FF0000000000000 ) < ( a & LIT64( 0x7FFFFFFFFFFFFFFF ) );
+
+}
+
+#else
+
+INLINE flag float64_same( float64 a, float64 b )
+{
+
+    return ( a.high == b.high ) && ( a.low == b.low );
+
+}
+
+INLINE flag float64_isNaN( float64 a )
+{
+    bits32 absAHigh;
+
+    absAHigh = a.high & 0x7FFFFFFF;
+    return
+        ( 0x7FF00000 < absAHigh ) || ( ( absAHigh == 0x7FF00000 ) && a.low );
+
+}
+
+#endif
+
+#ifdef FLOATX80
+
+INLINE flag floatx80_same( floatx80 a, floatx80 b )
+{
+
+    return ( a.high == b.high ) && ( a.low == b.low );
+
+}
+
+INLINE flag floatx80_isNaN( floatx80 a )
+{
+
+    return ( ( a.high & 0x7FFF ) == 0x7FFF ) && a.low;
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+INLINE flag float128_same( float128 a, float128 b )
+{
+
+    return ( a.high == b.high ) && ( a.low == b.low );
+
+}
+
+INLINE flag float128_isNaN( float128 a )
+{
+    bits64 absAHigh;
+
+    absAHigh = a.high & LIT64( 0x7FFFFFFFFFFFFFFF );
+    return
+           ( LIT64( 0x7FFF000000000000 ) < absAHigh )
+        || ( ( absAHigh == LIT64( 0x7FFF000000000000 ) ) && a.low );
+
+}
+
+#endif
+
+void
+ test_a_int32_z_float32(
+     float32 trueFunction( int32 ), float32 testFunction( int32 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int32();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_a_int32_z_float64(
+     float64 trueFunction( int32 ), float64 testFunction( int32 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int32();
+                fputs( "  ", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_int32_z_floatx80(
+     floatx80 trueFunction( int32 ), floatx80 testFunction( int32 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int32();
+                fputs( "  ", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void
+ test_a_int32_z_float128(
+     float128 trueFunction( int32 ), float128 testFunction( int32 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int32();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#ifdef BITS64
+
+void
+ test_a_int64_z_float32(
+     float32 trueFunction( int64 ), float32 testFunction( int64 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int64();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_a_int64_z_float64(
+     float64 trueFunction( int64 ), float64 testFunction( int64 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int64();
+                fputs( "  ", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_int64_z_floatx80(
+     floatx80 trueFunction( int64 ), floatx80 testFunction( int64 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int64();
+                fputs( "  ", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void
+ test_a_int64_z_float128(
+     float128 trueFunction( int64 ), float128 testFunction( int64 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_int64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_int64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_int64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_int64();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#endif
+
+void
+ test_a_float32_z_int32(
+     int32 trueFunction( float32 ), int32 testFunction( float32 ) )
+{
+    int16 count;
+    int32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == 0x7FFFFFFF )
+                 && (    ( testZ == 0x7FFFFFFF )
+                      || ( testZ == (sbits32) 0x80000000 ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_int32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef BITS64
+
+void
+ test_a_float32_z_int64(
+     int64 trueFunction( float32 ), int64 testFunction( float32 ) )
+{
+    int16 count;
+    int64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                 && (    ( testZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                      || ( testZ == (sbits64) LIT64( 0x8000000000000000 ) ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_int64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_a_float32_z_float64(
+     float64 trueFunction( float32 ), float64 testFunction( float32 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_float32_z_floatx80(
+     floatx80 trueFunction( float32 ), floatx80 testFunction( float32 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void
+ test_a_float32_z_float128(
+     float128 trueFunction( float32 ), float128 testFunction( float32 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_az_float32(
+     float32 trueFunction( float32 ), float32 testFunction( float32 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float32_is_signaling_nan( testCases_a_float32 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_ab_float32_z_flag(
+     flag trueFunction( float32, float32 ),
+     flag testFunction( float32, float32 )
+ )
+{
+    int16 count;
+    flag trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32, testCases_b_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32, testCases_b_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float32_is_signaling_nan( testCases_a_float32 )
+                      || float32_is_signaling_nan( testCases_b_float32 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_flag( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+void
+ test_abz_float32(
+     float32 trueFunction( float32, float32 ),
+     float32 testFunction( float32, float32 )
+ )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float32 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float32, testCases_b_float32 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float32, testCases_b_float32 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float32_is_signaling_nan( testCases_a_float32 )
+                      || float32_is_signaling_nan( testCases_b_float32 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float32();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+void
+ test_a_float64_z_int32(
+     int32 trueFunction( float64 ), int32 testFunction( float64 ) )
+{
+    int16 count;
+    int32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == 0x7FFFFFFF )
+                 && (    ( testZ == 0x7FFFFFFF )
+                      || ( testZ == (sbits32) 0x80000000 ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "  ", stdout );
+                writeOutputs_z_int32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef BITS64
+
+void
+ test_a_float64_z_int64(
+     int64 trueFunction( float64 ), int64 testFunction( float64 ) )
+{
+    int16 count;
+    int64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                 && (    ( testZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                      || ( testZ == (sbits64) LIT64( 0x8000000000000000 ) ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "  ", stdout );
+                writeOutputs_z_int64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_a_float64_z_float32(
+     float32 trueFunction( float64 ), float32 testFunction( float64 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_float64_z_floatx80(
+     floatx80 trueFunction( float64 ), floatx80 testFunction( float64 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void
+ test_a_float64_z_float128(
+     float128 trueFunction( float64 ), float128 testFunction( float64 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_az_float64(
+     float64 trueFunction( float64 ), float64 testFunction( float64 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float64_is_signaling_nan( testCases_a_float64 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float64();
+                fputs( "  ", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_ab_float64_z_flag(
+     flag trueFunction( float64, float64 ),
+     flag testFunction( float64, float64 )
+ )
+{
+    int16 count;
+    flag trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64, testCases_b_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64, testCases_b_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float64_is_signaling_nan( testCases_a_float64 )
+                      || float64_is_signaling_nan( testCases_b_float64 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float64();
+                fputs( "  ", stdout );
+                writeOutputs_z_flag( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+void
+ test_abz_float64(
+     float64 trueFunction( float64, float64 ),
+     float64 testFunction( float64, float64 )
+ )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float64 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float64, testCases_b_float64 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float64, testCases_b_float64 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float64_is_signaling_nan( testCases_a_float64 )
+                      || float64_is_signaling_nan( testCases_b_float64 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float64();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_floatx80_z_int32(
+     int32 trueFunction( floatx80 ), int32 testFunction( floatx80 ) )
+{
+    int16 count;
+    int32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == 0x7FFFFFFF )
+                 && (    ( testZ == 0x7FFFFFFF )
+                      || ( testZ == (sbits32) 0x80000000 ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "  ", stdout );
+                writeOutputs_z_int32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef BITS64
+
+void
+ test_a_floatx80_z_int64(
+     int64 trueFunction( floatx80 ), int64 testFunction( floatx80 ) )
+{
+    int16 count;
+    int64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                 && (    ( testZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                      || ( testZ == (sbits64) LIT64( 0x8000000000000000 ) ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "  ", stdout );
+                writeOutputs_z_int64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_a_floatx80_z_float32(
+     float32 trueFunction( floatx80 ), float32 testFunction( floatx80 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_a_floatx80_z_float64(
+     float64 trueFunction( floatx80 ), float64 testFunction( floatx80 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOAT128
+
+void
+ test_a_floatx80_z_float128(
+     float128 trueFunction( floatx80 ), float128 testFunction( floatx80 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_az_floatx80(
+     floatx80 trueFunction( floatx80 ), floatx80 testFunction( floatx80 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && floatx80_is_signaling_nan( testCases_a_floatx80 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_floatx80();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_ab_floatx80_z_flag(
+     flag trueFunction( floatx80, floatx80 ),
+     flag testFunction( floatx80, floatx80 )
+ )
+{
+    int16 count;
+    flag trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80, testCases_b_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80, testCases_b_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    floatx80_is_signaling_nan( testCases_a_floatx80 )
+                      || floatx80_is_signaling_nan( testCases_b_floatx80 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_floatx80();
+                fputs( "  ", stdout );
+                writeOutputs_z_flag( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+void
+ test_abz_floatx80(
+     floatx80 trueFunction( floatx80, floatx80 ),
+     floatx80 testFunction( floatx80, floatx80 )
+ )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_floatx80 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_floatx80, testCases_b_floatx80 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_floatx80, testCases_b_floatx80 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    floatx80_is_signaling_nan( testCases_a_floatx80 )
+                      || floatx80_is_signaling_nan( testCases_b_floatx80 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_floatx80();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void
+ test_a_float128_z_int32(
+     int32 trueFunction( float128 ), int32 testFunction( float128 ) )
+{
+    int16 count;
+    int32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == 0x7FFFFFFF )
+                 && (    ( testZ == 0x7FFFFFFF )
+                      || ( testZ == (sbits32) 0x80000000 ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "  ", stdout );
+                writeOutputs_z_int32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef BITS64
+
+void
+ test_a_float128_z_int64(
+     int64 trueFunction( float128 ), int64 testFunction( float128 ) )
+{
+    int16 count;
+    int64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ( trueZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                 && (    ( testZ == LIT64( 0x7FFFFFFFFFFFFFFF ) )
+                      || ( testZ == (sbits64) LIT64( 0x8000000000000000 ) ) )
+                 && ( trueFlags == float_flag_invalid )
+                 && ( testFlags == float_flag_invalid )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_int64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_a_float128_z_float32(
+     float32 trueFunction( float128 ), float32 testFunction( float128 ) )
+{
+    int16 count;
+    float32 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float32_isNaN( trueZ )
+                 && float32_isNaN( testZ )
+                 && ! float32_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "  ", stdout );
+                writeOutputs_z_float32( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_a_float128_z_float64(
+     float64 trueFunction( float128 ), float64 testFunction( float128 ) )
+{
+    int16 count;
+    float64 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float64_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float64_isNaN( trueZ )
+                 && float64_isNaN( testZ )
+                 && ! float64_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float64( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#ifdef FLOATX80
+
+void
+ test_a_float128_z_floatx80(
+     floatx80 trueFunction( float128 ), floatx80 testFunction( float128 ) )
+{
+    int16 count;
+    floatx80 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! floatx80_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && floatx80_isNaN( trueZ )
+                 && floatx80_isNaN( testZ )
+                 && ! floatx80_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_floatx80( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+#endif
+
+void
+ test_az_float128(
+     float128 trueFunction( float128 ), float128 testFunction( float128 ) )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_a_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && float128_is_signaling_nan( testCases_a_float128 ) ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInput_a_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+
+}
+
+void
+ test_ab_float128_z_flag(
+     flag trueFunction( float128, float128 ),
+     flag testFunction( float128, float128 )
+ )
+{
+    int16 count;
+    flag trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128, testCases_b_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128, testCases_b_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float128_is_signaling_nan( testCases_a_float128 )
+                      || float128_is_signaling_nan( testCases_b_float128 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if ( ( trueZ != testZ ) || ( trueFlags != testFlags ) ) {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_flag( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+void
+ test_abz_float128(
+     float128 trueFunction( float128, float128 ),
+     float128 testFunction( float128, float128 )
+ )
+{
+    int16 count;
+    float128 trueZ, testZ;
+    uint8 trueFlags, testFlags;
+
+    errorCount = 0;
+    tenthousandsCount = 0;
+    count = 10000;
+    testCases_initSequence( testCases_sequence_ab_float128 );
+    writeTestsTotal();
+    while ( ! testCases_done || forever ) {
+        testCases_next();
+        *trueFlagsPtr = 0;
+        trueZ = trueFunction( testCases_a_float128, testCases_b_float128 );
+        trueFlags = *trueFlagsPtr;
+        (void) testFlagsFunctionPtr();
+        testZ = testFunction( testCases_a_float128, testCases_b_float128 );
+        testFlags = testFlagsFunctionPtr();
+        --count;
+        if ( count == 0 ) {
+            checkEarlyExit();
+            count = 10000;
+        }
+        if ( ! float128_same( trueZ, testZ ) || ( trueFlags != testFlags ) ) {
+            if (    ! checkNaNs
+                 && (    float128_is_signaling_nan( testCases_a_float128 )
+                      || float128_is_signaling_nan( testCases_b_float128 ) )
+               ) {
+                trueFlags |= float_flag_invalid;
+            }
+            if (    ! checkNaNs
+                 && float128_isNaN( trueZ )
+                 && float128_isNaN( testZ )
+                 && ! float128_is_signaling_nan( testZ )
+                 && ( trueFlags == testFlags )
+               ) {
+                /* no problem */
+            }
+            else {
+                ++errorCount;
+                writeErrorFound( 10000 - count );
+                writeInputs_ab_float128();
+                fputs( "\n\t", stdout );
+                writeOutputs_z_float128( trueZ, trueFlags, testZ, testFlags );
+                fflush( stdout );
+                if ( errorCount == maxErrorCount ) goto exit;
+            }
+        }
+    }
+ exit:
+    writeTestsPerformed( 10000 - count );
+    return;
+
+}
+
+#endif
+
diff --git a/testfloat/testLoops.h b/testfloat/testLoops.h
new file mode 100644
index 00000000000..c3b08477f0e
--- /dev/null
+++ b/testfloat/testLoops.h
@@ -0,0 +1,143 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdio.h>
+
+extern volatile flag stop;
+
+extern char *trueName, *testName;
+extern flag forever, errorStop;
+extern uint32 maxErrorCount;
+extern flag checkNaNs;
+extern int8 *trueFlagsPtr;
+extern int8 ( *testFlagsFunctionPtr )( void );
+extern char *functionName;
+extern char *roundingPrecisionName, *roundingModeName, *tininessModeName;
+extern flag anyErrors;
+
+void writeFunctionName( FILE * );
+void exitWithStatus( void );
+
+void test_a_int32_z_float32( float32 ( int32 ), float32 ( int32 ) );
+void test_a_int32_z_float64( float64 ( int32 ), float64 ( int32 ) );
+#ifdef FLOATX80
+void test_a_int32_z_floatx80( floatx80 ( int32 ), floatx80 ( int32 ) );
+#endif
+#ifdef FLOAT128
+void test_a_int32_z_float128( float128 ( int32 ), float128 ( int32 ) );
+#endif
+#ifdef BITS64
+void test_a_int64_z_float32( float32 ( int64 ), float32 ( int64 ) );
+void test_a_int64_z_float64( float64 ( int64 ), float64 ( int64 ) );
+#ifdef FLOATX80
+void test_a_int64_z_floatx80( floatx80 ( int64 ), floatx80 ( int64 ) );
+#endif
+#ifdef FLOAT128
+void test_a_int64_z_float128( float128 ( int64 ), float128 ( int64 ) );
+#endif
+#endif
+
+void test_a_float32_z_int32( int32 ( float32 ), int32 ( float32 ) );
+#ifdef BITS64
+void test_a_float32_z_int64( int64 ( float32 ), int64 ( float32 ) );
+#endif
+void test_a_float32_z_float64( float64 ( float32 ), float64 ( float32 ) );
+#ifdef FLOATX80
+void test_a_float32_z_floatx80( floatx80 ( float32 ), floatx80 ( float32 ) );
+#endif
+#ifdef FLOAT128
+void test_a_float32_z_float128( float128 ( float32 ), float128 ( float32 ) );
+#endif
+void test_az_float32( float32 ( float32 ), float32 ( float32 ) );
+void
+ test_ab_float32_z_flag(
+     flag ( float32, float32 ), flag ( float32, float32 ) );
+void
+ test_abz_float32(
+     float32 ( float32, float32 ), float32 ( float32, float32 ) );
+
+void test_a_float64_z_int32( int32 ( float64 ), int32 ( float64 ) );
+#ifdef BITS64
+void test_a_float64_z_int64( int64 ( float64 ), int64 ( float64 ) );
+#endif
+void test_a_float64_z_float32( float32 ( float64 ), float32 ( float64 ) );
+#ifdef FLOATX80
+void test_a_float64_z_floatx80( floatx80 ( float64 ), floatx80 ( float64 ) );
+#endif
+#ifdef FLOAT128
+void test_a_float64_z_float128( float128 ( float64 ), float128 ( float64 ) );
+#endif
+void test_az_float64( float64 ( float64 ), float64 ( float64 ) );
+void
+ test_ab_float64_z_flag(
+     flag ( float64, float64 ), flag ( float64, float64 ) );
+void
+ test_abz_float64(
+     float64 ( float64, float64 ), float64 ( float64, float64 ) );
+
+#ifdef FLOATX80
+
+void test_a_floatx80_z_int32( int32 ( floatx80 ), int32 ( floatx80 ) );
+#ifdef BITS64
+void test_a_floatx80_z_int64( int64 ( floatx80 ), int64 ( floatx80 ) );
+#endif
+void test_a_floatx80_z_float32( float32 ( floatx80 ), float32 ( floatx80 ) );
+void test_a_floatx80_z_float64( float64 ( floatx80 ), float64 ( floatx80 ) );
+#ifdef FLOAT128
+void
+ test_a_floatx80_z_float128( float128 ( floatx80 ), float128 ( floatx80 ) );
+#endif
+void test_az_floatx80( floatx80 ( floatx80 ), floatx80 ( floatx80 ) );
+void
+ test_ab_floatx80_z_flag(
+     flag ( floatx80, floatx80 ), flag ( floatx80, floatx80 ) );
+void
+ test_abz_floatx80(
+     floatx80 ( floatx80, floatx80 ), floatx80 ( floatx80, floatx80 ) );
+
+#endif
+
+#ifdef FLOAT128
+
+void test_a_float128_z_int32( int32 ( float128 ), int32 ( float128 ) );
+#ifdef BITS64
+void test_a_float128_z_int64( int64 ( float128 ), int64 ( float128 ) );
+#endif
+void test_a_float128_z_float32( float32 ( float128 ), float32 ( float128 ) );
+void test_a_float128_z_float64( float64 ( float128 ), float64 ( float128 ) );
+#ifdef FLOATX80
+void
+ test_a_float128_z_floatx80( floatx80 ( float128 ), floatx80 ( float128 ) );
+#endif
+void test_az_float128( float128 ( float128 ), float128 ( float128 ) );
+void
+ test_ab_float128_z_flag(
+     flag ( float128, float128 ), flag ( float128, float128 ) );
+void
+ test_abz_float128(
+     float128 ( float128, float128 ), float128 ( float128, float128 ) );
+
+#endif
+
diff --git a/testfloat/testfloat-history.txt b/testfloat/testfloat-history.txt
new file mode 100644
index 00000000000..61520b3b678
--- /dev/null
+++ b/testfloat/testfloat-history.txt
@@ -0,0 +1,57 @@
+
+History of Major Changes to TestFloat, up to Release 2a
+
+John R. Hauser
+1998 December 17
+
+
+The TestFloat releases parallel those of SoftFloat, on which TestFloat is
+based.  Each TestFloat release also incorporates all bug fixes from the
+corresponding release of SoftFloat.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Release 2a (1998 December)
+
+-- Added support for testing conversions between floating-point and 64-bit
+   integers.
+
+-- Improved the makefiles.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Release 2 (1997 June)
+
+-- Integrated the generation of test cases and the checking of system
+   results into a single program.  (Before they were separate programs,
+   normally joined by explicit command-line pipes.)
+
+-- Improved the sequence of test cases.
+
+-- Added support for testing extended double precision and quadruple
+   precision.
+
+-- Made program output more readable, and added new command arguments.
+
+-- Reduced dependence on the quality of the standard `random' function for
+   generating test cases.  (Previously naively expected `random' to be able
+   to generate good random bits for the entire machine word width.)
+
+-- Created `testsoftfloat', with its own simpler complete software floating-
+   point (``slowfloat'') for comparison purposes.
+
+-- Made some changes to the source file structure, including renaming
+   `environment.h' to `milieu.h' (to avoid confusion with environment
+   variables).
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Release 1a (1996 July)
+
+-- Added the `-tininessbefore' and `-tininessafter' options to control
+   whether tininess should be detected before or after rounding.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Release 1 (1996 July)
+
+-- Original release.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
diff --git a/testfloat/testfloat-source.txt b/testfloat/testfloat-source.txt
new file mode 100644
index 00000000000..b8f7e9bb3ed
--- /dev/null
+++ b/testfloat/testfloat-source.txt
@@ -0,0 +1,444 @@
+
+TestFloat Release 2a Source Documentation
+
+John R. Hauser
+1998 December 16
+
+
+-------------------------------------------------------------------------------
+Introduction
+
+TestFloat is a program for testing that a floating-point implementation
+conforms to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
+All standard operations supported by the system can be tested, except for
+conversions to and from decimal.  Any of the following machine formats can
+be tested:  single precision, double precision, extended double precision,
+and/or quadruple precision.  Testing extended double-precision or quadruple-
+precision formats requires a C compiler that supports 64-bit integer
+arithmetic.
+
+This document gives information needed for compiling and/or porting
+TestFloat.
+
+The source code for TestFloat is intended to be relatively machine-
+independent.  TestFloat is written in C, and should be compilable using
+any ISO/ANSI C compiler.  At the time of this writing, the program has
+been successfully compiled using the GNU C Compiler (`gcc') for several
+platforms.  Because ISO/ANSI C does not provide access to some features
+of IEC/IEEE floating-point such as the exception flags, porting TestFloat
+unfortunately involves some machine-dependent coding.
+
+TestFloat depends on SoftFloat, which is a software implementation of
+floating-point that conforms to the IEC/IEEE Standard.  SoftFloat is not
+included with the TestFloat sources.  It can be obtained from the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/SoftFloat.html'.
+
+In addition to a program for testing a machine's floating-point, the
+TestFloat package includes a variant for testing SoftFloat called
+`testsoftfloat'.  The sources for both programs are intermixed, and both are
+described here.
+
+The first release of TestFloat (Release 1) was called _FloatTest_.  The old
+name has been obsolete for some time.
+
+
+-------------------------------------------------------------------------------
+Limitations
+
+TestFloat as written requires an ISO/ANSI-style C compiler.  No attempt has
+been made to accomodate compilers that are not ISO-conformant.  Older ``K&R-
+style'' compilers are not adequate for compiling TestFloat.  All testing I
+have done so far has been with the GNU C Compiler.  Compilation with other
+compilers should be possible but has not been tested.
+
+The TestFloat sources assume that source code file names can be longer than
+8 characters.  In order to compile under an MS-DOS-style system, many of the
+source files will need to be renamed, and the source and makefiles edited
+appropriately.  Once compiled, the TestFloat program does not depend on the
+existence of long file names.
+
+The underlying machine is assumed to be binary with a word size that is a
+power of 2.  Bytes are 8 bits.  Testing of extended double-precision and
+quadruple-precision formats depends on the C compiler implementing a 64-bit
+integer type.  If the largest integer type supported by the C compiler is
+32 bits, only single- and double-precision operations can be tested.
+
+
+-------------------------------------------------------------------------------
+Contents
+
+    Introduction
+    Limitations
+    Contents
+    Legal Notice
+    TestFloat Source Directory Structure
+    Target-Independent Modules
+    Target-Specific Modules
+    Target-Specific Header Files
+        processors/*.h
+        testfloat/*/milieu.h
+    Target-Specific Floating-Point Subroutines
+    Steps to Creating the TestFloat Executables
+    Improving the Random Number Generator
+    Contact Information
+
+
+
+-------------------------------------------------------------------------------
+Legal Notice
+
+TestFloat was written by John R. Hauser.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+
+-------------------------------------------------------------------------------
+TestFloat Source Directory Structure
+
+Because TestFloat is targeted to multiple platforms, its source code
+is slightly scattered between target-specific and target-independent
+directories and files.  The directory structure is as follows:
+
+    processors
+    testfloat
+        templates
+        386-Win32-gcc
+        SPARC-Solaris-gcc
+
+The two topmost directories and their contents are:
+
+    testfloat    - Most of the source code needed for TestFloat.
+    processors   - Target-specific header files that are not specific to
+                       TestFloat.
+
+Within the `testfloat' directory are subdirectories for each of the
+targeted platforms.  The TestFloat source code is distributed with targets
+`386-Win32-gcc' and `SPARC-Solaris-gcc' (and perhaps others) already
+prepared.  These can be used as examples for porting to new targets.  Source
+files that are not within these target-specific subdirectories are intended
+to be target-independent.
+
+The naming convention used for the target-specific directories is
+`<processor>-<executable-type>-<compiler>'.  The names of the supplied
+target directories should be interpreted as follows:
+
+  <processor>:
+    386          - Intel 386-compatible processor.
+    SPARC        - SPARC processor (as used by Sun machines).
+  <executable-type>:
+    Win32        - Microsoft Win32 executable.
+    Solaris      - Sun Solaris executable.
+  <compiler>:
+    gcc          - GNU C Compiler.
+
+You do not need to maintain this convention if you do not want to.
+
+Alongside the supplied target-specific directories there is a `templates'
+directory containing a set of ``generic'' target-specific source files.
+A new target directory can be created by copying the `templates' directory
+and editing the files inside.  (Complete instructions for porting TestFloat
+to a new target are in the section _Steps_to_Creating_the_TestFloat_
+_Executables_.)  Note that the `templates' directory will not work as a
+target directory without some editing.  To avoid confusion, it would be wise
+to refrain from editing the files inside `templates' directly.
+
+In addition to the distributed sources, TestFloat depends on the existence
+of an appropriately-compiled SoftFloat binary and the corresponding header
+file `softfloat.h'.  SoftFloat is not included with the TestFloat sources.
+It can be obtained from the Web page `http://HTTP.CS.Berkeley.EDU/~jhauser/
+arithmetic/SoftFloat.html'.
+
+As distributed, the makefiles for TestFloat assume the existence of three
+sibling directories:
+
+    processors
+    softfloat
+    testfloat
+
+Only the `processors' and `testfloat' directories are included in the
+TestFloat package.  The `softfloat' directory is assumed to contain a
+target-specific subdirectory within which the SoftFloat header file and
+compiled binary can be found.  (See the source documentation accompanying
+SoftFloat.)  The `processors' directory distributed with TestFloat is
+intended to be identical to that included with the SoftFloat source.
+
+These are the defaults, but other organizations of the sources are possible.
+The TestFloat makefiles and `milieu.h' files (see below) are easily edited
+to accomodate other arrangements.
+
+
+-------------------------------------------------------------------------------
+Target-Independent Modules
+
+The TestFloat program is composed of a number of modules, some target-
+specific and some target-independent.  The target-independent modules are as
+follows:
+
+-- The `fail' module provides a common routine for writing an error message
+   and aborting.
+
+-- The `random' module generates random integer values.
+
+-- The `writeHex' module defines routines for writing the various types in
+   the hexadecimal form used by TestFloat.
+
+-- The `testCases' module generates test cases for the various types.
+
+-- The `testLoops' module contains various routines for exercising two
+   implementations of a function and reporting any differences observed.
+
+-- The `slowfloat' module provides the simple floating-point implementation
+   used by `testsoftfloat' for comparing against SoftFloat.  The heart
+   of `slowfloat' is found in either `slowfloat-32' or `slowfloat-64',
+   depending on whether the `BITS64' macro is defined.
+
+-- The `systfloat' module gives a SoftFloat-like interface to the machine's
+   floating-point.
+
+-- The `testFunction' module implements `testfloat's main loop for testing a
+   function for all of the relevant rounding modes and rounding precisions.
+   (The `testsoftfloat' program contains its own version of this code.)
+
+-- The `testfloat' and `testsoftfloat' modules are the main modules for the
+   `testfloat' and `testsoftfloat' programs.
+
+Except possibly for `systfloat', these modules should not need to be
+modified.
+
+The `systfloat' module uses the floating-point operations of the C language
+to access a machine's floating-point.  Unfortunately, some IEC/IEEE
+floating-point operations are not accessible within ISO/ANSI C.  The
+following machine functions cannot be tested unless an alternate `systfloat'
+module is provided:
+
+    <float>_to_int32 (rounded according to rounding mode)
+    <float>_to_int64 (rounded according to rounding mode)
+    <float>_round_to_int
+    <float>_rem
+    <float>_sqrt, except float64_sqrt
+    <float>_eq_signaling
+    <float>_le_quiet
+    <float>_lt_quiet
+
+The `-list' option to `testfloat' will show the operations the program is
+prepared to test.  The section _Target-Specific_Floating-Point_Subroutines_
+later in this document explains how to create a target-specific `systfloat'
+module to change the set of testable functions.
+
+
+-------------------------------------------------------------------------------
+Target-Specific Modules
+
+No target-specific modules are needed for `testsoftfloat'.
+
+The `testfloat' program uses two target-specific modules:
+
+-- The `systmodes' module defines functions for setting the modes
+   controlling the system's floating-point, including the rounding mode and
+   the rounding precision for extended double precision.
+
+-- The `systflags' module provides a function for clearing and examining the
+   system's floating-point exception flags.
+
+These modules must be supplied for each target.  They can be implemented in
+any way desired, so long as all is reflected in the target's makefile.  For
+the targets that come with the distributed source, each of these modules is
+implemented as a single assembly language or C language source file.
+
+
+-------------------------------------------------------------------------------
+Target-Specific Header Files
+
+The purpose of the two target-specific header files is detailed below.
+In the following, the `*' symbol is used in place of the name of a specific
+target, such as `386-Win32-gcc' or `SPARC-Solaris-gcc', or in place of some
+other text as explained below.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+processors/*.h
+
+The target-specific `processors' header file defines integer types
+of various sizes, and also defines certain C preprocessor macros that
+characterize the target.  The two examples supplied are `386-gcc.h' and
+`SPARC-gcc.h'.  The naming convention used for processor header files is
+`<processor>-<compiler>.h'.  The `processors' header file used to compile
+TestFloat should be the same as that used to compile SoftFloat.
+
+If 64-bit integers are supported by the compiler, the macro name `BITS64'
+should be defined here along with the corresponding 64-bit integer
+types.  In addition, the function-like macro `LIT64' must be defined for
+constructing 64-bit integer literals (constants).  The `LIT64' macro is used
+consistently in the TestFloat code to annotate 64-bit literals.
+
+If an inlining attribute (such as an `inline' keyword) is provided by the
+compiler, the macro `INLINE' should be defined to the appropriate keyword.
+If not, `INLINE' can be set to the keyword `static'.  The `INLINE' macro
+appears in the TestFloat source code before every function that should be
+inlined by the compiler.
+
+For maximum flexibility, the TestFloat source files do not include the
+`processors' header file directly; rather, this file is included by the
+target-specific `milieu.h' header, and `milieu.h' is included by the source
+files.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+testfloat/*/milieu.h
+
+The `milieu.h' header file provides declarations that are needed to
+compile TestFloat.  In particular, it is through this header file that
+the appropriate `processors' header is included to characterize the target
+processor.  In addition, deviations from ISO/ANSI C by the compiler (such as
+names not properly declared in system header files) are corrected in this
+header if possible.
+
+If the preprocessor macro `BITS64' is defined in the `processors' header
+file but only the 32-bit version of SoftFloat is actually used, the `BITS64'
+macro should be undefined here after the `processors' header has defined it.
+
+If the C compiler implements the `long double' floating-point type of C
+as extended double precision, then `LONG_DOUBLE_IS_FLOATX80' should be
+defined here.  Alternatively, if the C `long double' type is implemented as
+quadruple precision, `LONG_DOUBLE_IS_FLOAT128' should be defined.  At most
+one of these macros should be defined.  A C compiler is allowed to implement
+`long double' the same as `double', in which case neither of these macros
+should be defined.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Target-Specific Floating-Point Subroutines
+
+This section applies only to `testfloat' and not to `testsoftfloat'.
+
+By default, TestFloat tests a machine's floating-point by testing the
+floating-point operations of the C language.  Unfortunately, some IEC/IEEE
+floating-point operations are not defined within ISO/ANSI C.  If a machine
+implements such ``non-C'' operations, target-specific subroutines for
+the operations can be supplied to allow TestFloat to test these machine
+features.  Typically, such subroutines will need to be written in assembly
+language, although equivalent functions can sometimes be found among the
+system's software libraries.
+
+The following machine functions cannot be tested by TestFloat unless target-
+specific subroutines are supplied for them:
+
+    <float>_to_int32 (rounded according to rounding mode)
+    <float>_to_int64 (rounded according to rounding mode)
+    <float>_round_to_int
+    <float>_rem
+    <float>_sqrt, except float64_sqrt
+    <float>_eq_signaling
+    <float>_le_quiet
+    <float>_lt_quiet
+
+In addition to these, none of the `floatx80' functions can be tested by
+default if the C `long double' type is something other than extended double
+precision; and likewise, none of the `float128' functions can be tested by
+default if `long double' is not quadruple precision.  Since `long double'
+cannot be both extended double precision and quadruple precision at the
+same time, at least one of these types cannot be tested by TestFloat without
+appropriate subroutines being supplied for that type.  (On the other hand,
+few systems implement _both_ extended double-precision and quadruple-
+precision floating-point; and unless a system does implement both, it does
+not need both tested.)
+
+Note that the `-list' option to `testfloat' will show the operations
+TestFloat is prepared to test.
+
+TestFloat's `systfloat' module supplies the system version of the functions
+to be tested.  The names of the `systfloat' subroutines are the same as the
+function names used as arguments to the `testfloat' command but with `syst_'
+prefixed--thus, for example, `syst_float32_add' and `syst_int32_to_float32'.
+The default `systfloat' module maps these system functions to the standard
+C operations; so `syst_float32_add', for example, is implemented using the
+C `+' operation for the single-precision `float' type.  For each system
+function supplied by `systfloat', a corresponding `SYST_<function>'
+preprocessor macro is defined in `systfloat.h' to indicate that the function
+exists to be tested (e.g., `SYST_FLOAT32_ADD').  The `systfloat.h' header
+file also declares function prototypes for the `systfloat' functions.
+
+(The `systfloat.h' file that comes with the TestFloat package declares
+prototypes for all of the possible `systfloat' functions, whether defined in
+`systfloat' or not.  There is no penalty for declaring a function prototype
+that is never used.)
+
+A target-specific version of the `systfloat' module can easily be created to
+replace the generic one.  This in fact has been done for the example targets
+`386-Win32-gcc' and `SPARC-Solaris-gcc'.  For each target, an assembly
+language `systfloat.S' has been created in the target directory along with
+a corresponding `systfloat.h' header file defining the `SYST_<function>'
+macros for the functions implemented.  The makefiles of the targets have
+been edited to use these target-specific versions of `systfloat' rather than
+the generic one.
+
+The `systfloat' modules of the example targets have been written entirely
+in assembly language in order to bypass any peculiarities of the C compiler.
+Although this is probably a good idea, it is certainly not required.
+
+
+-------------------------------------------------------------------------------
+Steps to Creating the TestFloat Executables
+
+Porting and/or compiling TestFloat involves the following steps:
+
+1. Port SoftFloat and create a SoftFloat binary.  (Refer to the
+   documentation accompanying SoftFloat.)
+
+2. If one does not already exist, create an appropriate target-specific
+   subdirectory under `testfloat' by copying the given `templates'
+   directory.  The remaining steps occur within the target-specific
+   subdirectory.
+
+3. Edit the files `milieu.h' and `Makefile' to reflect the current
+   environment.
+
+4. Make `testsoftfloat' by executing `make testsoftfloat' (or `make
+   testsoftfloat.exe', or whatever the `testsoftfloat' executable is
+   called).  Verify that SoftFloat is working correctly by testing it with
+   `testsoftfloat'.
+
+If you only wanted `testsoftfloat', you are done.  The steps for `testfloat'
+continue:
+
+5. In the target-specific subdirectory, implement the `systmodes' and
+   `systflags' modules.  (The `syst_float_set_rounding_precision' function
+   need not do anything if the system does not support extended double
+   precision.)
+
+6. If the target machine supports standard floating-point functions that are
+   not accessible within ISO/ANSI C, or if the C compiler cannot be trusted
+   to use the machine's floating-point directly, create a target-specific
+   `systfloat' module.
+
+7. In the target-specific subdirectory, execute `make'.
+
+
+-------------------------------------------------------------------------------
+Improving the Random Number Generator
+
+If you are serious about using TestFloat for testing floating-point, you
+should consider replacing the supplied `random.c' with a better target-
+specific one.  The standard C `rand' function is rather poor on some
+systems, and consequently `random.c' has been written to assume very little
+about the quality of `rand'.  As a result, the `rand' function is called
+more frequently than it might need to be, shortening the time before
+the random number generator repeats, and possibly wasting time as well.
+If `rand' is better on your system, or if another better random number
+generator is available (such as `rand48' on most Unix systems), TestFloat
+can be improved by overriding the given `random.c' with a target-specific
+one.
+
+
+-------------------------------------------------------------------------------
+Contact Information
+
+At the time of this writing, the most up-to-date information about
+TestFloat and the latest release can be found at the Web page `http://
+HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+
diff --git a/testfloat/testfloat.c b/testfloat/testfloat.c
new file mode 100644
index 00000000000..ea9e8f4c09d
--- /dev/null
+++ b/testfloat/testfloat.c
@@ -0,0 +1,295 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdlib.h>
+#include <signal.h>
+#include <string.h>
+#include "milieu.h"
+#include "fail.h"
+#include "softfloat.h"
+#include "testCases.h"
+#include "testLoops.h"
+#include "systflags.h"
+#include "testFunction.h"
+
+static void catchSIGINT( int signalCode )
+{
+
+    if ( stop ) exit( EXIT_FAILURE );
+    stop = TRUE;
+
+}
+
+main( int argc, char **argv )
+{
+    char *argPtr;
+    flag functionArgument;
+    uint8 functionCode;
+    int8 operands, roundingPrecision, roundingMode;
+
+    fail_programName = "testfloat";
+    if ( argc <= 1 ) goto writeHelpMessage;
+    testCases_setLevel( 1 );
+    trueName = "soft";
+    testName = "syst";
+    errorStop = FALSE;
+    forever = FALSE;
+    maxErrorCount = 20;
+    trueFlagsPtr = &float_exception_flags;
+    testFlagsFunctionPtr = syst_float_flags_clear;
+    tininessModeName = 0;
+    functionArgument = FALSE;
+    functionCode = 0;
+    operands = 0;
+    roundingPrecision = 0;
+    roundingMode = 0;
+    --argc;
+    ++argv;
+    while ( argc && ( argPtr = argv[ 0 ] ) ) {
+        if ( argPtr[ 0 ] == '-' ) ++argPtr;
+        if ( strcmp( argPtr, "help" ) == 0 ) {
+ writeHelpMessage:
+            fputs(
+"testfloat [<option>...] <function>\n"
+"  <option>:  (* is default)\n"
+"    -help            --Write this message and exit.\n"
+"    -list            --List all testable functions and exit.\n"
+"    -level <num>     --Testing level <num> (1 or 2).\n"
+" *  -level 1\n"
+"    -errors <num>    --Stop each function test after <num> errors.\n"
+" *  -errors 20\n"
+"    -errorstop       --Exit after first function with any error.\n"
+"    -forever         --Test one function repeatedly (implies `-level 2').\n"
+"    -checkNaNs       --Check for bitwise correctness of NaN results.\n"
+#ifdef FLOATX80
+"    -precision32     --Only test rounding precision equivalent to float32.\n"
+"    -precision64     --Only test rounding precision equivalent to float64.\n"
+"    -precision80     --Only test maximum rounding precision.\n"
+#endif
+"    -nearesteven     --Only test rounding to nearest/even.\n"
+"    -tozero          --Only test rounding to zero.\n"
+"    -down            --Only test rounding down.\n"
+"    -up              --Only test rounding up.\n"
+"    -tininessbefore  --Underflow tininess detected before rounding.\n"
+"    -tininessafter   --Underflow tininess detected after rounding.\n"
+"  <function>:\n"
+"    int32_to_<float>                 <float>_add   <float>_eq\n"
+"    <float>_to_int32                 <float>_sub   <float>_le\n"
+"    <float>_to_int32_round_to_zero   <float>_mul   <float>_lt\n"
+#ifdef BITS64
+"    int64_to_<float>                 <float>_div   <float>_eq_signaling\n"
+"    <float>_to_int64                 <float>_rem   <float>_le_quiet\n"
+"    <float>_to_int64_round_to_zero                 <float>_lt_quiet\n"
+"    <float>_to_<float>\n"
+"    <float>_round_to_int\n"
+"    <float>_sqrt\n"
+#else
+"    <float>_to_<float>               <float>_div   <float>_eq_signaling\n"
+"    <float>_round_to_int             <float>_rem   <float>_le_quiet\n"
+"    <float>_sqrt                                   <float>_lt_quiet\n"
+#endif
+"    -all1            --All 1-operand functions.\n"
+"    -all2            --All 2-operand functions.\n"
+"    -all             --All functions.\n"
+"  <float>:\n"
+"    float32          --Single precision.\n"
+"    float64          --Double precision.\n"
+#ifdef FLOATX80
+"    floatx80         --Extended double precision.\n"
+#endif
+#ifdef FLOAT128
+"    float128         --Quadruple precision.\n"
+#endif
+                ,
+                stdout
+            );
+            return EXIT_SUCCESS;
+        }
+        else if ( strcmp( argPtr, "list" ) == 0 ) {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( functionExists[ functionCode ] ) {
+                    puts( functions[ functionCode ].name );
+                }
+            }
+            return EXIT_SUCCESS;
+        }
+        else if ( strcmp( argPtr, "level" ) == 0 ) {
+            if ( argc < 2 ) goto optionError;
+            testCases_setLevel( atoi( argv[ 1 ] ) );
+            --argc;
+            ++argv;
+        }
+        else if ( strcmp( argPtr, "level1" ) == 0 ) {
+            testCases_setLevel( 1 );
+        }
+        else if ( strcmp( argPtr, "level2" ) == 0 ) {
+            testCases_setLevel( 2 );
+        }
+        else if ( strcmp( argPtr, "errors" ) == 0 ) {
+            if ( argc < 2 ) {
+     optionError:
+                fail( "`%s' option requires numeric argument", argv[ 0 ] );
+            }
+            maxErrorCount = atoi( argv[ 1 ] );
+            --argc;
+            ++argv;
+        }
+        else if ( strcmp( argPtr, "errorstop" ) == 0 ) {
+            errorStop = TRUE;
+        }
+        else if ( strcmp( argPtr, "forever" ) == 0 ) {
+            testCases_setLevel( 2 );
+            forever = TRUE;
+        }
+        else if (    ( strcmp( argPtr, "checkNaNs" ) == 0 )
+                  || ( strcmp( argPtr, "checknans" ) == 0 ) ) {
+            checkNaNs = TRUE;
+        }
+#ifdef FLOATX80
+        else if ( strcmp( argPtr, "precision32" ) == 0 ) {
+            roundingPrecision = 32;
+        }
+        else if ( strcmp( argPtr, "precision64" ) == 0 ) {
+            roundingPrecision = 64;
+        }
+        else if ( strcmp( argPtr, "precision80" ) == 0 ) {
+            roundingPrecision = 80;
+        }
+#endif
+        else if (    ( strcmp( argPtr, "nearesteven" ) == 0 )
+                  || ( strcmp( argPtr, "nearest_even" ) == 0 ) ) {
+            roundingMode = ROUND_NEAREST_EVEN;
+        }
+        else if (    ( strcmp( argPtr, "tozero" ) == 0 )
+                  || ( strcmp( argPtr, "to_zero" ) == 0 ) ) {
+            roundingMode = ROUND_TO_ZERO;
+        }
+        else if ( strcmp( argPtr, "down" ) == 0 ) {
+            roundingMode = ROUND_DOWN;
+        }
+        else if ( strcmp( argPtr, "up" ) == 0 ) {
+            roundingMode = ROUND_UP;
+        }
+        else if ( strcmp( argPtr, "tininessbefore" ) == 0 ) {
+            float_detect_tininess = float_tininess_before_rounding;
+        }
+        else if ( strcmp( argPtr, "tininessafter" ) == 0 ) {
+            float_detect_tininess = float_tininess_after_rounding;
+        }
+        else if ( strcmp( argPtr, "all1" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 1;
+        }
+        else if ( strcmp( argPtr, "all2" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 2;
+        }
+        else if ( strcmp( argPtr, "all" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 0;
+        }
+        else {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( strcmp( argPtr, functions[ functionCode ].name ) == 0 ) {
+                    break;
+                }
+            }
+            if ( functionCode == NUM_FUNCTIONS ) {
+                fail( "Invalid option or function `%s'", argv[ 0 ] );
+            }
+            if ( ! functionExists[ functionCode ] ) {
+                fail(
+                    "Function `%s' is not supported or cannot be tested",
+                    argPtr
+                );
+            }
+            functionArgument = TRUE;
+        }
+        --argc;
+        ++argv;
+    }
+    if ( ! functionArgument ) fail( "Function argument required" );
+    (void) signal( SIGINT, catchSIGINT );
+    (void) signal( SIGTERM, catchSIGINT );
+    if ( functionCode ) {
+        if ( forever ) {
+            if ( ! roundingPrecision ) roundingPrecision = 80;
+            if ( ! roundingMode ) roundingMode = ROUND_NEAREST_EVEN;
+        }
+        testFunction( functionCode, roundingPrecision, roundingMode );
+    }
+    else {
+        if ( forever ) {
+            fail( "Can only test one function with `-forever' option" );
+        }
+        if ( operands == 1 ) {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if (    functionExists[ functionCode ]
+                     && ( functions[ functionCode ].numInputs == 1 ) ) {
+                    testFunction(
+                        functionCode, roundingPrecision, roundingMode );
+                }
+            }
+        }
+        else if ( operands == 2 ) {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if (    functionExists[ functionCode ]
+                     && ( functions[ functionCode ].numInputs == 2 ) ) {
+                    testFunction(
+                        functionCode, roundingPrecision, roundingMode );
+                }
+            }
+        }
+        else {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( functionExists[ functionCode ] ) {
+                    testFunction(
+                        functionCode, roundingPrecision, roundingMode );
+                }
+            }
+        }
+    }
+    exitWithStatus();
+
+}
+
diff --git a/testfloat/testfloat.txt b/testfloat/testfloat.txt
new file mode 100644
index 00000000000..6e72c1d89fd
--- /dev/null
+++ b/testfloat/testfloat.txt
@@ -0,0 +1,771 @@
+
+TestFloat Release 2a General Documentation
+
+John R. Hauser
+1998 December 16
+
+
+-------------------------------------------------------------------------------
+Introduction
+
+TestFloat is a program for testing that a floating-point implementation
+conforms to the IEC/IEEE Standard for Binary Floating-Point Arithmetic.
+All standard operations supported by the system can be tested, except for
+conversions to and from decimal.  Any of the following machine formats can
+be tested:  single precision, double precision, extended double precision,
+and/or quadruple precision.
+
+TestFloat actually comes in two variants:  one is a program for testing
+a machine's floating-point, and the other is a program for testing
+the SoftFloat software implementation of floating-point.  (Information
+about SoftFloat can be found at the SoftFloat Web page, `http://
+HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/SoftFloat.html'.)  The version that
+tests SoftFloat is expected to be of interest only to people compiling the
+SoftFloat sources.  However, because the two versions share much in common,
+they are discussed together in all the TestFloat documentation.
+
+This document explains how to use the TestFloat programs.  It does not
+attempt to define or explain the IEC/IEEE Standard for floating-point.
+Details about the standard are available elsewhere.
+
+The first release of TestFloat (Release 1) was called _FloatTest_.  The old
+name has been obsolete for some time.
+
+
+-------------------------------------------------------------------------------
+Limitations
+
+TestFloat's output is not always easily interpreted.  Detailed knowledge
+of the IEC/IEEE Standard and its vagaries is needed to use TestFloat
+responsibly.
+
+TestFloat performs relatively simple tests designed to check the fundamental
+soundness of the floating-point under test.  TestFloat may also at times
+manage to find rarer and more subtle bugs, but it will probably only find
+such bugs by accident.  Software that purposefully seeks out various kinds
+of subtle floating-point bugs can be found through links posted on the
+TestFloat Web page (`http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/
+TestFloat.html').
+
+
+-------------------------------------------------------------------------------
+Contents
+
+    Introduction
+    Limitations
+    Contents
+    Legal Notice
+    What TestFloat Does
+    Executing TestFloat
+    Functions Tested by TestFloat
+        Conversion Functions
+        Standard Arithmetic Functions
+        Remainder and Round-to-Integer Functions
+        Comparison Functions
+    Interpreting TestFloat Output
+    Variations Allowed by the IEC/IEEE Standard
+        Underflow
+        NaNs
+        Conversions to Integer
+    TestFloat Options
+        -help
+        -list
+        -level <num>
+        -errors <num>
+        -errorstop
+        -forever
+        -checkNaNs
+        -precision32, -precision64, -precision80
+        -nearesteven, -tozero, -down, -up
+        -tininessbefore, -tininessafter
+    Function Sets
+    Contact Information
+
+
+
+-------------------------------------------------------------------------------
+Legal Notice
+
+TestFloat was written by John R. Hauser.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+
+-------------------------------------------------------------------------------
+What TestFloat Does
+
+TestFloat tests a system's floating-point by comparing its behavior with
+that of TestFloat's own internal floating-point implemented in software.
+For each operation tested, TestFloat generates a large number of test cases,
+made up of simple pattern tests intermixed with weighted random inputs.
+The cases generated should be adequate for testing carry chain propagations,
+plus the rounding of adds, subtracts, multiplies, and simple operations like
+conversions.  TestFloat makes a point of checking all boundary cases of the
+arithmetic, including underflows, overflows, invalid operations, subnormal
+inputs, zeros (positive and negative), infinities, and NaNs.  For the
+interesting operations like adds and multiplies, literally millions of test
+cases can be checked.
+
+TestFloat is not remarkably good at testing difficult rounding cases for
+divisions and square roots.  It also makes no attempt to find bugs specific
+to SRT divisions and the like (such as the infamous Pentium divide bug).
+Software that tests for such failures can be found through links on the
+TestFloat Web page, `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/
+TestFloat.html'.
+
+NOTE!
+It is the responsibility of the user to verify that the discrepancies
+TestFloat finds actually represent faults in the system being tested.
+Advice to help with this task is provided later in this document.
+Furthermore, even if TestFloat finds no fault with a floating-point
+implementation, that in no way guarantees that the implementation is bug-
+free.
+
+For each operation, TestFloat can test all four rounding modes required
+by the IEC/IEEE Standard.  TestFloat verifies not only that the numeric
+results of an operation are correct, but also that the proper floating-point
+exception flags are raised.  All five exception flags are tested, including
+the inexact flag.  TestFloat does not attempt to verify that the floating-
+point exception flags are actually implemented as sticky flags.
+
+For machines that implement extended double precision with rounding
+precision control (such as Intel's 80x86), TestFloat can test the add,
+subtract, multiply, divide, and square root functions at all the standard
+rounding precisions.  The rounding precision can be set equivalent to single
+precision, to double precision, or to the full extended double precision.
+Rounding precision control can only be applied to the extended double-
+precision format and only for the five standard arithmetic operations:  add,
+subtract, multiply, divide, and square root.  Other functions can be tested
+only at full precision.
+
+As a rule, TestFloat is not particular about the bit patterns of NaNs that
+appear as function results.  Any NaN is considered as good a result as
+another.  This laxness can be overridden so that TestFloat checks for
+particular bit patterns within NaN results.  See the sections _Variations_
+_Allowed_by_the_IEC/IEEE_Standard_ and _TestFloat_Options_ for details.
+
+Not all IEC/IEEE Standard functions are supported by all machines.
+TestFloat can only test functions that exist on the machine.  But even if
+a function is supported by the machine, TestFloat may still not be able
+to test the function if it is not accessible through standard ISO C (the
+programming language in which TestFloat is written) and if the person who
+compiled TestFloat did not provide an alternate means for TestFloat to
+invoke the machine function.
+
+TestFloat compares a machine's floating-point against the SoftFloat software
+implementation of floating-point, also written by me.  SoftFloat is built
+into the TestFloat executable and does not need to be supplied by the user.
+If SoftFloat is wanted for some other reason (to compile a new version
+of TestFloat, for instance), it can be found separately at the Web page
+`http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/SoftFloat.html'.
+
+For testing SoftFloat itself, the TestFloat package includes a program that
+compares SoftFloat's floating-point against _another_ software floating-
+point implementation.  The second software floating-point is simpler and
+slower than SoftFloat, and is completely independent of SoftFloat.  Although
+the second software floating-point cannot be guaranteed to be bug-free, the
+chance that it would mimic any of SoftFloat's bugs is remote.  Consequently,
+an error in one or the other floating-point version should appear as an
+unexpected discrepancy between the two implementations.  Note that testing
+SoftFloat should only be necessary when compiling a new TestFloat executable
+or when compiling SoftFloat for some other reason.
+
+
+-------------------------------------------------------------------------------
+Executing TestFloat
+
+TestFloat is intended to be executed from a command line interpreter.  The
+`testfloat' program is invoked as follows:
+
+    testfloat [<option>...] <function>
+
+Here square brackets ([]) indicate optional items, while angled brackets
+(<>) denote parameters to be filled in.
+
+The `<function>' argument is a name like `float32_add' or `float64_to_int32'.
+The complete list of function names is given in the next section,
+_Functions_Tested_by_TestFloat_.  It is also possible to test all machine
+functions in a single invocation.  The various options to TestFloat are
+detailed in the section _TestFloat_Options_ later in this document.  If
+`testfloat' is executed without any arguments, a summary of TestFloat usage
+is written.
+
+TestFloat will ordinarily test a function for all four rounding modes, one
+after the other.  If the rounding mode is not supposed to have any affect
+on the results--for instance, some operations do not require rounding--only
+the nearest/even rounding mode is checked.  For extended double-precision
+operations affected by rounding precision control, TestFloat also tests all
+three rounding precision modes, one after the other.  Testing can be limited
+to a single rounding mode and/or rounding precision with appropriate options
+(see _TestFloat_Options_).
+
+As it executes, TestFloat writes status information to the standard error
+output, which should be the screen by default.  In order for this status to
+be displayed properly, the standard error stream should not be redirected
+to a file.  The discrepancies TestFloat finds are written to the standard
+output stream, which is easily redirected to a file if desired.  Ordinarily,
+the errors TestFloat reports and the ongoing status information appear
+intermixed on the same screen.
+
+The version of TestFloat for testing SoftFloat is called `testsoftfloat'.
+It is invoked the same as `testfloat',
+
+    testsoftfloat [<option>...] <function>
+
+and operates similarly.
+
+
+-------------------------------------------------------------------------------
+Functions Tested by TestFloat
+
+TestFloat tests all operations required by the IEC/IEEE Standard except for
+conversions to and from decimal.  The operations are
+
+-- Conversions among the supported floating-point formats, and also between
+   integers (32-bit and 64-bit) and any of the floating-point formats.
+
+-- The usual add, subtract, multiply, divide, and square root operations
+   for all supported floating-point formats.
+
+-- For each format, the floating-point remainder operation defined by the
+   IEC/IEEE Standard.
+
+-- For each floating-point format, a ``round to integer'' operation that
+   rounds to the nearest integer value in the same format.  (The floating-
+   point formats can hold integer values, of course.)
+
+-- Comparisons between two values in the same floating-point format.
+
+Detailed information about these functions is given below.  In the function
+names used by TestFloat, single precision is called `float32', double
+precision is `float64', extended double precision is `floatx80', and
+quadruple precision is `float128'.  TestFloat uses the same names for
+functions as SoftFloat.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Conversion Functions
+
+All conversions among the floating-point formats and all conversion between
+a floating-point format and 32-bit and 64-bit signed integers can be tested.
+The conversion functions are:
+
+   int32_to_float32      int64_to_float32
+   int32_to_float64      int64_to_float32
+   int32_to_floatx80     int64_to_floatx80
+   int32_to_float128     int64_to_float128
+
+   float32_to_int32      float32_to_int64
+   float32_to_int32      float64_to_int64
+   floatx80_to_int32     floatx80_to_int64
+   float128_to_int32     float128_to_int64
+
+   float32_to_float64    float32_to_floatx80   float32_to_float128
+   float64_to_float32    float64_to_floatx80   float64_to_float128
+   floatx80_to_float32   floatx80_to_float64   floatx80_to_float128
+   float128_to_float32   float128_to_float64   float128_to_floatx80
+
+These conversions all round according to the current rounding mode as
+necessary.  Conversions from a smaller to a larger floating-point format are
+always exact and so require no rounding.  Conversions from 32-bit integers
+to double precision or to any larger floating-point format are also exact,
+and likewise for conversions from 64-bit integers to extended double and
+quadruple precisions.
+
+ISO/ANSI C requires that conversions to integers be rounded toward zero.
+Such conversions can be tested with the following functions that ignore any
+rounding mode:
+
+   float32_to_int32_round_to_zero    float32_to_int64_round_to_zero
+   float64_to_int32_round_to_zero    float64_to_int64_round_to_zero
+   floatx80_to_int32_round_to_zero   floatx80_to_int64_round_to_zero
+   float128_to_int32_round_to_zero   float128_to_int64_round_to_zero
+
+TestFloat assumes that conversions from floating-point to integer should
+raise the invalid exception if the source value cannot be rounded to a
+representable integer of the desired size (32 or 64 bits).  If such a
+conversion overflows, TestFloat expects the largest integer with the same
+sign as the operand to be returned.  If the floating-point operand is a NaN,
+TestFloat allows either the largest postive or largest negative integer to
+be returned.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Standard Arithmetic Functions
+
+The following standard arithmetic functions can be tested:
+
+   float32_add    float32_sub    float32_mul    float32_div    float32_sqrt
+   float64_add    float64_sub    float64_mul    float64_div    float64_sqrt
+   floatx80_add   floatx80_sub   floatx80_mul   floatx80_div   floatx80_sqrt
+   float128_add   float128_sub   float128_mul   float128_div   float128_sqrt
+
+The extended double-precision (`floatx80') functions can be rounded to
+reduced precision under rounding precision control.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Remainder and Round-to-Integer Functions
+
+For each format, TestFloat can test the IEC/IEEE Standard remainder and
+round-to-integer functions.  The remainder functions are:
+
+   float32_rem
+   float64_rem
+   floatx80_rem
+   float128_rem
+
+The round-to-integer functions are:
+
+   float32_round_to_int
+   float64_round_to_int
+   floatx80_round_to_int
+   float128_round_to_int
+
+The remainder functions are always exact and so do not require rounding.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Comparison Functions
+
+The following floating-point comparison functions can be tested:
+
+   float32_eq    float32_le    float32_lt
+   float64_eq    float64_le    float64_lt
+   floatx80_eq   floatx80_le   floatx80_lt
+   float128_eq   float128_le   float128_lt
+
+The abbreviation `eq' stands for ``equal'' (=); `le' stands for ``less than
+or equal'' (<=); and `lt' stands for ``less than'' (<).
+
+The IEC/IEEE Standard specifies that the less-than-or-equal and less-than
+functions raise the invalid exception if either input is any kind of NaN.
+The equal functions, for their part, are defined not to raise the invalid
+exception on quiet NaNs.  For completeness, the following additional
+functions can be tested if supported:
+
+   float32_eq_signaling    float32_le_quiet    float32_lt_quiet
+   float64_eq_signaling    float64_le_quiet    float64_lt_quiet
+   floatx80_eq_signaling   floatx80_le_quiet   floatx80_lt_quiet
+   float128_eq_signaling   float128_le_quiet   float128_lt_quiet
+
+The `signaling' equal functions are identical to the standard functions
+except that the invalid exception should be raised for any NaN input.
+Likewise, the `quiet' comparison functions should be identical to their
+counterparts except that the invalid exception is not raised for quiet NaNs.
+
+Obviously, no comparison functions ever require rounding.  Any rounding mode
+is ignored.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Interpreting TestFloat Output
+
+The ``errors'' reported by TestFloat may or may not really represent errors
+in the system being tested.  For each test case tried, TestFloat performs
+the same floating-point operation for the two implementations being compared
+and reports any unexpected difference in the results.  The two results could
+differ for several reasons:
+
+-- The IEC/IEEE Standard allows for some variation in how conforming
+   floating-point behaves.  Two implementations can occasionally give
+   different results without either being incorrect.
+
+-- The trusted floating-point emulation could be faulty.  This could be
+   because there is a bug in the way the enulation is coded, or because a
+   mistake was made when the code was compiled for the current system.
+
+-- TestFloat may not work properly, reporting discrepancies that do not
+   exist.
+
+-- Lastly, the floating-point being tested could actually be faulty.
+
+It is the responsibility of the user to determine the causes for the
+discrepancies TestFloat reports.  Making this determination can require
+detailed knowledge about the IEC/IEEE Standard.  Assuming TestFloat is
+working properly, any differences found will be due to either the first or
+last of these reasons.  Variations in the IEC/IEEE Standard that could lead
+to false error reports are discussed in the section _Variations_Allowed_by_
+_the_IEC/IEEE_Standard_.
+
+For each error (or apparent error) TestFloat reports, a line of text
+is written to the default output.  If a line would be longer than 79
+characters, it is divided.  The first part of each error line begins in the
+leftmost column, and any subsequent ``continuation'' lines are indented with
+a tab.
+
+Each error reported by `testfloat' is of the form:
+
+    <inputs>  soft: <output-from-emulation>  syst: <output-from-system>
+
+The `<inputs>' are the inputs to the operation.  Each output is shown as a
+pair:  the result value first, followed by the exception flags.  The `soft'
+label stands for ``software'' (or ``SoftFloat''), while `syst' stands for
+``system,'' the machine's floating-point.
+
+For example, two typical error lines could be
+
+    800.7FFF00  87F.000100  soft: 001.000000 ....x  syst: 001.000000 ...ux
+    081.000004  000.1FFFFF  soft: 001.000000 ....x  syst: 001.000000 ...ux
+
+In the first line, the inputs are `800.7FFF00' and `87F.000100'.  The
+internal emulation result is `001.000000' with flags `....x', and the
+system result is the same but with flags `...ux'.  All the items composed of
+hexadecimal digits and a single period represent floating-point values (here
+single precision).  These cases were reported as errors because the flag
+results differ.
+
+In addition to the exception flags, there are seven data types that may
+be represented.  Four are floating-point types:  single precision, double
+precision, extended double precision, and quadruple precision.  The
+remaining three types are 32-bit and 64-bit two's-complement integers and
+Boolean values (the results of comparison operations).  Boolean values are
+represented as a single character, either a `0' or a `1'.  32-bit integers
+are written as 8 hexadecimal digits in two's-complement form.  Thus,
+`FFFFFFFF' is -1, and `7FFFFFFF' is the largest positive 32-bit integer.
+64-bit integers are the same except with 16 hexadecimal digits.
+
+Floating-point values are written in a correspondingly primitive form.
+Double-precision values are represented by 16 hexadecimal digits that give
+the raw bits of the floating-point encoding.  A period separates the 3rd and
+4th hexadecimal digits to mark the division between the exponent bits and
+fraction bits.  Some notable double-precision values include:
+
+    000.0000000000000    +0
+    3FF.0000000000000     1
+    400.0000000000000     2
+    7FF.0000000000000    +infinity
+
+    800.0000000000000    -0
+    BFF.0000000000000    -1
+    C00.0000000000000    -2
+    FFF.0000000000000    -infinity
+
+    3FE.FFFFFFFFFFFFF    largest representable number preceding +1
+
+The following categories are easily distinguished (assuming the `x's are not
+all 0):
+
+    000.xxxxxxxxxxxxx    positive subnormal (denormalized) numbers
+    7FF.xxxxxxxxxxxxx    positive NaNs
+    800.xxxxxxxxxxxxx    negative subnormal numbers
+    FFF.xxxxxxxxxxxxx    negative NaNs
+
+Quadruple-precision values are written the same except with 4 hexadecimal
+digits for the sign and exponent and 28 for the fraction.  Notable values
+include:
+
+    0000.0000000000000000000000000000    +0
+    3FFF.0000000000000000000000000000     1
+    4000.0000000000000000000000000000     2
+    7FFF.0000000000000000000000000000    +infinity
+
+    8000.0000000000000000000000000000    -0
+    BFFF.0000000000000000000000000000    -1
+    C000.0000000000000000000000000000    -2
+    FFFF.0000000000000000000000000000    -infinity
+
+    3FFE.FFFFFFFFFFFFFFFFFFFFFFFFFFFF    largest representable number
+                                             preceding +1
+
+Extended double-precision values are a little unusual in that the leading
+significand bit is not hidden as with other formats.  When correctly
+encoded, the leading significand bit of an extended double-precision value
+will be 0 if the value is zero or subnormal, and will be 1 otherwise.
+Hence, the same values listed above appear in extended double-precision as
+follows (note the leading `8' digit in the significands):
+
+    0000.0000000000000000    +0
+    3FFF.8000000000000000     1
+    4000.8000000000000000     2
+    7FFF.8000000000000000    +infinity
+
+    8000.0000000000000000    -0
+    BFFF.8000000000000000    -1
+    C000.8000000000000000    -2
+    FFFF.8000000000000000    -infinity
+
+    3FFE.FFFFFFFFFFFFFFFF    largest representable number preceding +1
+
+The representation of single-precision values is unusual for a different
+reason.  Because the subfields of standard single-precision do not fall
+on neat 4-bit boundaries, single-precision outputs are slightly perturbed.
+These are written as 9 hexadecimal digits, with a period separating the 3rd
+and 4th hexadecimal digits.  Broken out into bits, the 9 hexademical digits
+cover the single-precision subfields as follows:
+
+    x000 .... ....  .  .... .... .... .... .... ....    sign       (1 bit)
+    .... xxxx xxxx  .  .... .... .... .... .... ....    exponent   (8 bits)
+    .... .... ....  .  0xxx xxxx xxxx xxxx xxxx xxxx    fraction  (23 bits)
+
+As shown in this schematic, the first hexadecimal digit contains only
+the sign, and will be either `0' or `8'.  The next two digits give the
+biased exponent as an 8-bit integer.  This is followed by a period and
+6 hexadecimal digits of fraction.  The most significant hexadecimal digit
+of the fraction can be at most a `7'.
+
+Notable single-precision values include:
+
+    000.000000    +0
+    07F.000000     1
+    080.000000     2
+    0FF.000000    +infinity
+
+    800.000000    -0
+    87F.000000    -1
+    880.000000    -2
+    8FF.000000    -infinity
+
+    07E.7FFFFF    largest representable number preceding +1
+
+Again, certain categories are easily distinguished (assuming the `x's are
+not all 0):
+
+    000.xxxxxx    positive subnormal (denormalized) numbers
+    0FF.xxxxxx    positive NaNs
+    800.xxxxxx    negative subnormal numbers
+    8FF.xxxxxx    negative NaNs
+
+Lastly, exception flag values are represented by five characters, one
+character per flag.  Each flag is written as either a letter or a period
+(`.') according to whether the flag was set or not by the operation.  A
+period indicates the flag was not set.  The letter used to indicate a set
+flag depends on the flag:
+
+    v    invalid flag
+    z    division-by-zero flag
+    o    overflow flag
+    u    underflow flag
+    x    inexact flag
+
+For example, the notation `...ux' indicates that the underflow and inexact
+exception flags were set and that the other three flags (invalid, division-
+by-zero, and overflow) were not set.  The exception flags are always shown
+following the value returned as the result of the operation.
+
+The output from `testsoftfloat' is of the same form, except that the results
+are labeled `true' and `soft':
+
+    <inputs>  true: <simple-software-result>  soft: <SoftFloat-result>
+
+The ``true'' result is from the simpler, slower software floating-point,
+which, although not necessarily correct, is more likely to be right than
+the SoftFloat (`soft') result.
+
+
+-------------------------------------------------------------------------------
+Variations Allowed by the IEC/IEEE Standard
+
+The IEC/IEEE Standard admits some variation among conforming
+implementations.  Because TestFloat expects the two implementations being
+compared to deliver bit-for-bit identical results under most circumstances,
+this leeway in the standard can result in false errors being reported if
+the two implementations do not make the same choices everywhere the standard
+provides an option.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Underflow
+
+The standard specifies that the underflow exception flag is to be raised
+when two conditions are met simultaneously:  (1) _tininess_ and (2) _loss_
+_of_accuracy_.  A result is tiny when its magnitude is nonzero yet smaller
+than any normalized floating-point number.  The standard allows tininess to
+be determined either before or after a result is rounded to the destination
+precision.  If tininess is detected before rounding, some borderline cases
+will be flagged as underflows even though the result after rounding actually
+lies within the normal floating-point range.  By detecting tininess after
+rounding, a system can avoid some unnecessary signaling of underflow.
+
+Loss of accuracy occurs when the subnormal format is not sufficient
+to represent an underflowed result accurately.  The standard allows
+loss of accuracy to be detected either as an _inexact_result_ or as a
+_denormalization_loss_.  If loss of accuracy is detected as an inexact
+result, the underflow flag is raised whenever an underflowed quantity
+cannot be exactly represented in the subnormal format (that is, whenever the
+inexact flag is also raised).  A denormalization loss, on the other hand,
+occurs only when the subnormal format is not able to represent the result
+that would have been returned if the destination format had infinite range.
+Some underflowed results are inexact but do not suffer a denormalization
+loss.  By detecting loss of accuracy as a denormalization loss, a system can
+once again avoid some unnecessary signaling of underflow.
+
+The `-tininessbefore' and `-tininessafter' options can be used to control
+whether TestFloat expects tininess on underflow to be detected before or
+after rounding.  (See _TestFloat_Options_ below.)  One or the other is
+selected as the default when TestFloat is compiled, but these command
+options allow the default to be overridden.
+
+Most (possibly all) systems detect loss of accuracy as an inexact result.
+The current version of TestFloat can only test for this case.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+NaNs
+
+The IEC/IEEE Standard gives the floating-point formats a large number of
+NaN encodings and specifies that NaNs are to be returned as results under
+certain conditions.  However, the standard allows an implementation almost
+complete freedom over _which_ NaN to return in each situation.
+
+By default, TestFloat does not check the bit patterns of NaN results.  When
+the result of an operation should be a NaN, any NaN is considered as good
+as another.  This laxness can be overridden with the `-checkNaNs' option.
+(See _TestFloat_Options_ below.)  In order for this option to be sensible,
+TestFloat must have been compiled so that its internal floating-point
+implementation (SoftFloat) generates the proper NaN results for the system
+being tested.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+Conversions to Integer
+
+Conversion of a floating-point value to an integer format will fail if the
+source value is a NaN or if it is too large.  The IEC/IEEE Standard does not
+specify what value should be returned as the integer result in these cases.
+Moreover, according to the standard, the invalid exception can be raised or
+an unspecified alternative mechanism may be used to signal such cases.
+
+TestFloat assumes that conversions to integer will raise the invalid
+exception if the source value cannot be rounded to a representable integer.
+When the conversion overflows, TestFloat expects the largest integer with
+the same sign as the operand to be returned.  If the floating-point operand
+is a NaN, TestFloat allows either the largest postive or largest negative
+integer to be returned.  The current version of TestFloat provides no means
+to alter these conventions.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+TestFloat Options
+
+The `testfloat' (and `testsoftfloat') program accepts several command
+options.  If mutually contradictory options are given, the last one has
+priority.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-help
+
+The `-help' option causes a summary of program usage to be written, after
+which the program exits.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-list
+
+The `-list' option causes a list of testable functions to be written,
+after which the program exits.  Some machines do not implement all of the
+functions TestFloat can test, plus it may not be possible to test functions
+that are inaccessible from the C language.
+
+The `testsoftfloat' program does not have this option.  All SoftFloat
+functions can be tested by `testsoftfloat'.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-level <num>
+
+The `-level' option sets the level of testing.  The argument to `-level' can
+be either 1 or 2.  The default is level 1.  Level 2 performs many more tests
+than level 1.  Testing at level 2 can take as much as a day (even longer for
+`testsoftfloat'), but can reveal bugs not found by level 1.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-errors <num>
+
+The `-errors' option instructs TestFloat to report no more than the
+specified number of errors for any combination of function, rounding mode,
+etc.  The argument to `-errors' must be a nonnegative decimal number.  Once
+the specified number of error reports has been generated, TestFloat ends the
+current test and begins the next one, if any.  The default is `-errors 20'.
+
+Against intuition, `-errors 0' causes TestFloat to report every error it
+finds.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-errorstop
+
+The `-errorstop' option causes the program to exit after the first function
+for which any errors are reported.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-forever
+
+The `-forever' option causes a single operation to be repeatedly tested.
+Only one rounding mode and/or rounding precision can be tested in a single
+invocation.  If not specified, the rounding mode defaults to nearest/even.
+For extended double-precision operations, the rounding precision defaults
+to full extended double precision.  The testing level is set to 2 by this
+option.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-checkNaNs
+
+The `-checkNaNs' option causes TestFloat to verify the bitwise correctness
+of NaN results.  In order for this option to be sensible, TestFloat must
+have been compiled so that its internal floating-point implementation
+(SoftFloat) generates the proper NaN results for the system being tested.
+
+This option is not available to `testsoftfloat'.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-precision32, -precision64, -precision80
+
+For extended double-precision functions affected by rounding precision
+control, the `-precision32' option restricts testing to only the cases
+in which rounding precision is equivalent to single precision.  The other
+rounding precision options are not tested.  Likewise, the `-precision64'
+and `-precision80' options fix the rounding precision equivalent to double
+precision or extended double precision, respectively.  These options are
+ignored for functions not affected by rounding precision control.
+
+These options are not available if extended double precision is not
+supported by the machine or if extended double precision functions cannot be
+tested.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-nearesteven, -tozero, -down, -up
+
+The `-nearesteven' option restricts testing to only the cases in which the
+rounding mode is nearest/even.  The other rounding mode options are not
+tested.  Likewise, `-tozero' forces rounding to zero; `-down' forces
+rounding down; and `-up' forces rounding up.  These options are ignored for
+functions that are exact and thus do not round.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+-tininessbefore, -tininessafter
+
+The `-tininessbefore' option indicates that the system detects tininess
+on underflow before rounding.  The `-tininessafter' option indicates that
+tininess is detected after rounding.  TestFloat alters its expectations
+accordingly.  These options override the default selected when TestFloat was
+compiled.  Choosing the wrong one of these two options should cause error
+reports for some (not all) functions.
+
+For `testsoftfloat', these options operate more like the rounding precision
+and rounding mode options, in that they restrict the tests performed by
+`testsoftfloat'.  By default, `testsoftfloat' tests both cases for any
+function for which there is a difference.
+
+- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
+
+
+-------------------------------------------------------------------------------
+Function Sets
+
+Just as TestFloat can test an operation for all four rounding modes in
+sequence, multiple operations can be tested with a single invocation of
+TestFloat.  Three sets are recognized:  `-all1', `-all2', and `-all'.  The
+set `-all1' comprises all one-operand functions; `-all2' is all two-operand
+functions; and `-all' is all functions.  A function set can be used in place
+of a function name in the TestFloat command line, such as
+
+    testfloat [<option>...] -all
+
+
+-------------------------------------------------------------------------------
+Contact Information
+
+At the time of this writing, the most up-to-date information about
+TestFloat and the latest release can be found at the Web page `http://
+HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+
diff --git a/testfloat/testsoftfloat.c b/testfloat/testsoftfloat.c
new file mode 100644
index 00000000000..07a46a58008
--- /dev/null
+++ b/testfloat/testsoftfloat.c
@@ -0,0 +1,1040 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdlib.h>
+#include <signal.h>
+#include <string.h>
+#include "milieu.h"
+#include "fail.h"
+#include "softfloat.h"
+#include "slowfloat.h"
+#include "testCases.h"
+#include "testLoops.h"
+
+static void catchSIGINT( int signalCode )
+{
+
+    if ( stop ) exit( EXIT_FAILURE );
+    stop = TRUE;
+
+}
+
+int8 clearFlags( void )
+{
+    int8 flags;
+
+    flags = float_exception_flags;
+    float_exception_flags = 0;
+    return flags;
+
+}
+
+enum {
+    INT32_TO_FLOAT32 = 1,
+    INT32_TO_FLOAT64,
+#ifdef FLOATX80
+    INT32_TO_FLOATX80,
+#endif
+#ifdef FLOAT128
+    INT32_TO_FLOAT128,
+#endif
+#ifdef BITS64
+    INT64_TO_FLOAT32,
+    INT64_TO_FLOAT64,
+#ifdef FLOATX80
+    INT64_TO_FLOATX80,
+#endif
+#ifdef FLOAT128
+    INT64_TO_FLOAT128,
+#endif
+#endif
+    FLOAT32_TO_INT32,
+    FLOAT32_TO_INT32_ROUND_TO_ZERO,
+#ifdef BITS64
+    FLOAT32_TO_INT64,
+    FLOAT32_TO_INT64_ROUND_TO_ZERO,
+#endif
+    FLOAT32_TO_FLOAT64,
+#ifdef FLOATX80
+    FLOAT32_TO_FLOATX80,
+#endif
+#ifdef FLOAT128
+    FLOAT32_TO_FLOAT128,
+#endif
+    FLOAT32_ROUND_TO_INT,
+    FLOAT32_ADD,
+    FLOAT32_SUB,
+    FLOAT32_MUL,
+    FLOAT32_DIV,
+    FLOAT32_REM,
+    FLOAT32_SQRT,
+    FLOAT32_EQ,
+    FLOAT32_LE,
+    FLOAT32_LT,
+    FLOAT32_EQ_SIGNALING,
+    FLOAT32_LE_QUIET,
+    FLOAT32_LT_QUIET,
+    FLOAT64_TO_INT32,
+    FLOAT64_TO_INT32_ROUND_TO_ZERO,
+#ifdef BITS64
+    FLOAT64_TO_INT64,
+    FLOAT64_TO_INT64_ROUND_TO_ZERO,
+#endif
+    FLOAT64_TO_FLOAT32,
+#ifdef FLOATX80
+    FLOAT64_TO_FLOATX80,
+#endif
+#ifdef FLOAT128
+    FLOAT64_TO_FLOAT128,
+#endif
+    FLOAT64_ROUND_TO_INT,
+    FLOAT64_ADD,
+    FLOAT64_SUB,
+    FLOAT64_MUL,
+    FLOAT64_DIV,
+    FLOAT64_REM,
+    FLOAT64_SQRT,
+    FLOAT64_EQ,
+    FLOAT64_LE,
+    FLOAT64_LT,
+    FLOAT64_EQ_SIGNALING,
+    FLOAT64_LE_QUIET,
+    FLOAT64_LT_QUIET,
+#ifdef FLOATX80
+    FLOATX80_TO_INT32,
+    FLOATX80_TO_INT32_ROUND_TO_ZERO,
+#ifdef BITS64
+    FLOATX80_TO_INT64,
+    FLOATX80_TO_INT64_ROUND_TO_ZERO,
+#endif
+    FLOATX80_TO_FLOAT32,
+    FLOATX80_TO_FLOAT64,
+#ifdef FLOAT128
+    FLOATX80_TO_FLOAT128,
+#endif
+    FLOATX80_ROUND_TO_INT,
+    FLOATX80_ADD,
+    FLOATX80_SUB,
+    FLOATX80_MUL,
+    FLOATX80_DIV,
+    FLOATX80_REM,
+    FLOATX80_SQRT,
+    FLOATX80_EQ,
+    FLOATX80_LE,
+    FLOATX80_LT,
+    FLOATX80_EQ_SIGNALING,
+    FLOATX80_LE_QUIET,
+    FLOATX80_LT_QUIET,
+#endif
+#ifdef FLOAT128
+    FLOAT128_TO_INT32,
+    FLOAT128_TO_INT32_ROUND_TO_ZERO,
+#ifdef BITS64
+    FLOAT128_TO_INT64,
+    FLOAT128_TO_INT64_ROUND_TO_ZERO,
+#endif
+    FLOAT128_TO_FLOAT32,
+    FLOAT128_TO_FLOAT64,
+#ifdef FLOATX80
+    FLOAT128_TO_FLOATX80,
+#endif
+    FLOAT128_ROUND_TO_INT,
+    FLOAT128_ADD,
+    FLOAT128_SUB,
+    FLOAT128_MUL,
+    FLOAT128_DIV,
+    FLOAT128_REM,
+    FLOAT128_SQRT,
+    FLOAT128_EQ,
+    FLOAT128_LE,
+    FLOAT128_LT,
+    FLOAT128_EQ_SIGNALING,
+    FLOAT128_LE_QUIET,
+    FLOAT128_LT_QUIET,
+#endif
+    NUM_FUNCTIONS
+};
+static struct {
+    char *name;
+    int8 numInputs;
+    flag roundingPrecision, roundingMode;
+    flag tininessMode, tininessModeAtReducedPrecision;
+} functions[ NUM_FUNCTIONS ] = {
+    { 0, 0, 0, 0, 0, 0 },
+    { "int32_to_float32",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "int32_to_float64",                1, FALSE, FALSE, FALSE, FALSE },
+#ifdef FLOATX80
+    { "int32_to_floatx80",               1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef FLOAT128
+    { "int32_to_float128",               1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef BITS64
+    { "int64_to_float32",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "int64_to_float64",                1, FALSE, TRUE,  FALSE, FALSE },
+#ifdef FLOATX80
+    { "int64_to_floatx80",               1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef FLOAT128
+    { "int64_to_float128",               1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#endif
+    { "float32_to_int32",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_to_int32_round_to_zero",  1, FALSE, FALSE, FALSE, FALSE },
+#ifdef BITS64
+    { "float32_to_int64",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_to_int64_round_to_zero",  1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "float32_to_float64",              1, FALSE, FALSE, FALSE, FALSE },
+#ifdef FLOATX80
+    { "float32_to_floatx80",             1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef FLOAT128
+    { "float32_to_float128",             1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "float32_round_to_int",            1, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_add",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_sub",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_mul",                     2, FALSE, TRUE,  TRUE,  FALSE },
+    { "float32_div",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_rem",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_sqrt",                    1, FALSE, TRUE,  FALSE, FALSE },
+    { "float32_eq",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_le",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_lt",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_eq_signaling",            2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_le_quiet",                2, FALSE, FALSE, FALSE, FALSE },
+    { "float32_lt_quiet",                2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_to_int32",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_to_int32_round_to_zero",  1, FALSE, FALSE, FALSE, FALSE },
+#ifdef BITS64
+    { "float64_to_int64",                1, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_to_int64_round_to_zero",  1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "float64_to_float32",              1, FALSE, TRUE,  TRUE,  FALSE },
+#ifdef FLOATX80
+    { "float64_to_floatx80",             1, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef FLOAT128
+    { "float64_to_float128",             1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "float64_round_to_int",            1, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_add",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_sub",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_mul",                     2, FALSE, TRUE,  TRUE,  FALSE },
+    { "float64_div",                     2, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_rem",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_sqrt",                    1, FALSE, TRUE,  FALSE, FALSE },
+    { "float64_eq",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_le",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_lt",                      2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_eq_signaling",            2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_le_quiet",                2, FALSE, FALSE, FALSE, FALSE },
+    { "float64_lt_quiet",                2, FALSE, FALSE, FALSE, FALSE },
+#ifdef FLOATX80
+    { "floatx80_to_int32",               1, FALSE, TRUE,  FALSE, FALSE },
+    { "floatx80_to_int32_round_to_zero", 1, FALSE, FALSE, FALSE, FALSE },
+#ifdef BITS64
+    { "floatx80_to_int64",               1, FALSE, TRUE,  FALSE, FALSE },
+    { "floatx80_to_int64_round_to_zero", 1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "floatx80_to_float32",             1, FALSE, TRUE,  TRUE,  FALSE },
+    { "floatx80_to_float64",             1, FALSE, TRUE,  TRUE,  FALSE },
+#ifdef FLOAT128
+    { "floatx80_to_float128",            1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "floatx80_round_to_int",           1, FALSE, TRUE,  FALSE, FALSE },
+    { "floatx80_add",                    2, TRUE,  TRUE,  FALSE, TRUE  },
+    { "floatx80_sub",                    2, TRUE,  TRUE,  FALSE, TRUE  },
+    { "floatx80_mul",                    2, TRUE,  TRUE,  TRUE,  TRUE  },
+    { "floatx80_div",                    2, TRUE,  TRUE,  FALSE, TRUE  },
+    { "floatx80_rem",                    2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_sqrt",                   1, TRUE,  TRUE,  FALSE, FALSE },
+    { "floatx80_eq",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_le",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_lt",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_eq_signaling",           2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_le_quiet",               2, FALSE, FALSE, FALSE, FALSE },
+    { "floatx80_lt_quiet",               2, FALSE, FALSE, FALSE, FALSE },
+#endif
+#ifdef FLOAT128
+    { "float128_to_int32",               1, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_to_int32_round_to_zero", 1, FALSE, FALSE, FALSE, FALSE },
+#ifdef BITS64
+    { "float128_to_int64",               1, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_to_int64_round_to_zero", 1, FALSE, FALSE, FALSE, FALSE },
+#endif
+    { "float128_to_float32",             1, FALSE, TRUE,  TRUE,  FALSE },
+    { "float128_to_float64",             1, FALSE, TRUE,  TRUE,  FALSE },
+#ifdef FLOATX80
+    { "float128_to_floatx80",            1, FALSE, TRUE,  TRUE,  FALSE },
+#endif
+    { "float128_round_to_int",           1, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_add",                    2, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_sub",                    2, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_mul",                    2, FALSE, TRUE,  TRUE,  FALSE },
+    { "float128_div",                    2, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_rem",                    2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_sqrt",                   1, FALSE, TRUE,  FALSE, FALSE },
+    { "float128_eq",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_le",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_lt",                     2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_eq_signaling",           2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_le_quiet",               2, FALSE, FALSE, FALSE, FALSE },
+    { "float128_lt_quiet",               2, FALSE, FALSE, FALSE, FALSE },
+#endif
+};
+
+enum {
+    ROUND_NEAREST_EVEN = 1,
+    ROUND_TO_ZERO,
+    ROUND_DOWN,
+    ROUND_UP,
+    NUM_ROUNDINGMODES
+};
+enum {
+    TININESS_BEFORE_ROUNDING = 1,
+    TININESS_AFTER_ROUNDING,
+    NUM_TININESSMODES
+};
+
+static void
+ testFunctionVariety(
+     uint8 functionCode,
+     int8 roundingPrecision,
+     int8 roundingMode,
+     int8 tininessMode
+ )
+{
+    uint8 roundingCode;
+    int8 tininessCode;
+
+    functionName = functions[ functionCode ].name;
+    if ( roundingPrecision == 32 ) {
+        roundingPrecisionName = "32";
+    }
+    else if ( roundingPrecision == 64 ) {
+        roundingPrecisionName = "64";
+    }
+    else if ( roundingPrecision == 80 ) {
+        roundingPrecisionName = "80";
+    }
+    else {
+        roundingPrecisionName = 0;
+    }
+#ifdef FLOATX80
+    floatx80_rounding_precision = roundingPrecision;
+    slow_floatx80_rounding_precision = roundingPrecision;
+#endif
+    switch ( roundingMode ) {
+     case 0:
+        roundingModeName = 0;
+        roundingCode = float_round_nearest_even;
+        break;
+     case ROUND_NEAREST_EVEN:
+        roundingModeName = "nearest_even";
+        roundingCode = float_round_nearest_even;
+        break;
+     case ROUND_TO_ZERO:
+        roundingModeName = "to_zero";
+        roundingCode = float_round_to_zero;
+        break;
+     case ROUND_DOWN:
+        roundingModeName = "down";
+        roundingCode = float_round_down;
+        break;
+     case ROUND_UP:
+        roundingModeName = "up";
+        roundingCode = float_round_up;
+        break;
+    }
+    float_rounding_mode = roundingCode;
+    slow_float_rounding_mode = roundingCode;
+    switch ( tininessMode ) {
+     case 0:
+        tininessModeName = 0;
+        tininessCode = float_tininess_after_rounding;
+        break;
+     case TININESS_BEFORE_ROUNDING:
+        tininessModeName = "before";
+        tininessCode = float_tininess_before_rounding;
+        break;
+     case TININESS_AFTER_ROUNDING:
+        tininessModeName = "after";
+        tininessCode = float_tininess_after_rounding;
+        break;
+    }
+    float_detect_tininess = tininessCode;
+    slow_float_detect_tininess = tininessCode;
+    fputs( "Testing ", stderr );
+    writeFunctionName( stderr );
+    fputs( ".\n", stderr );
+    switch ( functionCode ) {
+     case INT32_TO_FLOAT32:
+        test_a_int32_z_float32( slow_int32_to_float32, int32_to_float32 );
+        break;
+     case INT32_TO_FLOAT64:
+        test_a_int32_z_float64( slow_int32_to_float64, int32_to_float64 );
+        break;
+#ifdef FLOATX80
+     case INT32_TO_FLOATX80:
+        test_a_int32_z_floatx80( slow_int32_to_floatx80, int32_to_floatx80 );
+        break;
+#endif
+#ifdef FLOAT128
+     case INT32_TO_FLOAT128:
+        test_a_int32_z_float128( slow_int32_to_float128, int32_to_float128 );
+        break;
+#endif
+#ifdef BITS64
+     case INT64_TO_FLOAT32:
+        test_a_int64_z_float32( slow_int64_to_float32, int64_to_float32 );
+        break;
+     case INT64_TO_FLOAT64:
+        test_a_int64_z_float64( slow_int64_to_float64, int64_to_float64 );
+        break;
+#ifdef FLOATX80
+     case INT64_TO_FLOATX80:
+        test_a_int64_z_floatx80( slow_int64_to_floatx80, int64_to_floatx80 );
+        break;
+#endif
+#ifdef FLOAT128
+     case INT64_TO_FLOAT128:
+        test_a_int64_z_float128( slow_int64_to_float128, int64_to_float128 );
+        break;
+#endif
+#endif
+     case FLOAT32_TO_INT32:
+        test_a_float32_z_int32( slow_float32_to_int32, float32_to_int32 );
+        break;
+     case FLOAT32_TO_INT32_ROUND_TO_ZERO:
+        test_a_float32_z_int32(
+            slow_float32_to_int32_round_to_zero,
+            float32_to_int32_round_to_zero
+        );
+        break;
+#ifdef BITS64
+     case FLOAT32_TO_INT64:
+        test_a_float32_z_int64( slow_float32_to_int64, float32_to_int64 );
+        break;
+     case FLOAT32_TO_INT64_ROUND_TO_ZERO:
+        test_a_float32_z_int64(
+            slow_float32_to_int64_round_to_zero,
+            float32_to_int64_round_to_zero
+        );
+        break;
+#endif
+     case FLOAT32_TO_FLOAT64:
+        test_a_float32_z_float64(
+            slow_float32_to_float64, float32_to_float64 );
+        break;
+#ifdef FLOATX80
+     case FLOAT32_TO_FLOATX80:
+        test_a_float32_z_floatx80(
+            slow_float32_to_floatx80, float32_to_floatx80 );
+        break;
+#endif
+#ifdef FLOAT128
+     case FLOAT32_TO_FLOAT128:
+        test_a_float32_z_float128(
+            slow_float32_to_float128, float32_to_float128 );
+        break;
+#endif
+     case FLOAT32_ROUND_TO_INT:
+        test_az_float32( slow_float32_round_to_int, float32_round_to_int );
+        break;
+     case FLOAT32_ADD:
+        test_abz_float32( slow_float32_add, float32_add );
+        break;
+     case FLOAT32_SUB:
+        test_abz_float32( slow_float32_sub, float32_sub );
+        break;
+     case FLOAT32_MUL:
+        test_abz_float32( slow_float32_mul, float32_mul );
+        break;
+     case FLOAT32_DIV:
+        test_abz_float32( slow_float32_div, float32_div );
+        break;
+     case FLOAT32_REM:
+        test_abz_float32( slow_float32_rem, float32_rem );
+        break;
+     case FLOAT32_SQRT:
+        test_az_float32( slow_float32_sqrt, float32_sqrt );
+        break;
+     case FLOAT32_EQ:
+        test_ab_float32_z_flag( slow_float32_eq, float32_eq );
+        break;
+     case FLOAT32_LE:
+        test_ab_float32_z_flag( slow_float32_le, float32_le );
+        break;
+     case FLOAT32_LT:
+        test_ab_float32_z_flag( slow_float32_lt, float32_lt );
+        break;
+     case FLOAT32_EQ_SIGNALING:
+        test_ab_float32_z_flag(
+            slow_float32_eq_signaling, float32_eq_signaling );
+        break;
+     case FLOAT32_LE_QUIET:
+        test_ab_float32_z_flag( slow_float32_le_quiet, float32_le_quiet );
+        break;
+     case FLOAT32_LT_QUIET:
+        test_ab_float32_z_flag( slow_float32_lt_quiet, float32_lt_quiet );
+        break;
+     case FLOAT64_TO_INT32:
+        test_a_float64_z_int32( slow_float64_to_int32, float64_to_int32 );
+        break;
+     case FLOAT64_TO_INT32_ROUND_TO_ZERO:
+        test_a_float64_z_int32(
+            slow_float64_to_int32_round_to_zero,
+            float64_to_int32_round_to_zero
+        );
+        break;
+#ifdef BITS64
+     case FLOAT64_TO_INT64:
+        test_a_float64_z_int64( slow_float64_to_int64, float64_to_int64 );
+        break;
+     case FLOAT64_TO_INT64_ROUND_TO_ZERO:
+        test_a_float64_z_int64(
+            slow_float64_to_int64_round_to_zero,
+            float64_to_int64_round_to_zero
+        );
+        break;
+#endif
+     case FLOAT64_TO_FLOAT32:
+        test_a_float64_z_float32(
+            slow_float64_to_float32, float64_to_float32 );
+        break;
+#ifdef FLOATX80
+     case FLOAT64_TO_FLOATX80:
+        test_a_float64_z_floatx80(
+            slow_float64_to_floatx80, float64_to_floatx80 );
+        break;
+#endif
+#ifdef FLOAT128
+     case FLOAT64_TO_FLOAT128:
+        test_a_float64_z_float128(
+            slow_float64_to_float128, float64_to_float128 );
+        break;
+#endif
+     case FLOAT64_ROUND_TO_INT:
+        test_az_float64( slow_float64_round_to_int, float64_round_to_int );
+        break;
+     case FLOAT64_ADD:
+        test_abz_float64( slow_float64_add, float64_add );
+        break;
+     case FLOAT64_SUB:
+        test_abz_float64( slow_float64_sub, float64_sub );
+        break;
+     case FLOAT64_MUL:
+        test_abz_float64( slow_float64_mul, float64_mul );
+        break;
+     case FLOAT64_DIV:
+        test_abz_float64( slow_float64_div, float64_div );
+        break;
+     case FLOAT64_REM:
+        test_abz_float64( slow_float64_rem, float64_rem );
+        break;
+     case FLOAT64_SQRT:
+        test_az_float64( slow_float64_sqrt, float64_sqrt );
+        break;
+     case FLOAT64_EQ:
+        test_ab_float64_z_flag( slow_float64_eq, float64_eq );
+        break;
+     case FLOAT64_LE:
+        test_ab_float64_z_flag( slow_float64_le, float64_le );
+        break;
+     case FLOAT64_LT:
+        test_ab_float64_z_flag( slow_float64_lt, float64_lt );
+        break;
+     case FLOAT64_EQ_SIGNALING:
+        test_ab_float64_z_flag(
+            slow_float64_eq_signaling, float64_eq_signaling );
+        break;
+     case FLOAT64_LE_QUIET:
+        test_ab_float64_z_flag( slow_float64_le_quiet, float64_le_quiet );
+        break;
+     case FLOAT64_LT_QUIET:
+        test_ab_float64_z_flag( slow_float64_lt_quiet, float64_lt_quiet );
+        break;
+#ifdef FLOATX80
+     case FLOATX80_TO_INT32:
+        test_a_floatx80_z_int32( slow_floatx80_to_int32, floatx80_to_int32 );
+        break;
+     case FLOATX80_TO_INT32_ROUND_TO_ZERO:
+        test_a_floatx80_z_int32(
+            slow_floatx80_to_int32_round_to_zero,
+            floatx80_to_int32_round_to_zero
+        );
+        break;
+#ifdef BITS64
+     case FLOATX80_TO_INT64:
+        test_a_floatx80_z_int64( slow_floatx80_to_int64, floatx80_to_int64 );
+        break;
+     case FLOATX80_TO_INT64_ROUND_TO_ZERO:
+        test_a_floatx80_z_int64(
+            slow_floatx80_to_int64_round_to_zero,
+            floatx80_to_int64_round_to_zero
+        );
+        break;
+#endif
+     case FLOATX80_TO_FLOAT32:
+        test_a_floatx80_z_float32(
+            slow_floatx80_to_float32, floatx80_to_float32 );
+        break;
+     case FLOATX80_TO_FLOAT64:
+        test_a_floatx80_z_float64(
+            slow_floatx80_to_float64, floatx80_to_float64 );
+        break;
+#ifdef FLOAT128
+     case FLOATX80_TO_FLOAT128:
+        test_a_floatx80_z_float128(
+            slow_floatx80_to_float128, floatx80_to_float128 );
+        break;
+#endif
+     case FLOATX80_ROUND_TO_INT:
+        test_az_floatx80( slow_floatx80_round_to_int, floatx80_round_to_int );
+        break;
+     case FLOATX80_ADD:
+        test_abz_floatx80( slow_floatx80_add, floatx80_add );
+        break;
+     case FLOATX80_SUB:
+        test_abz_floatx80( slow_floatx80_sub, floatx80_sub );
+        break;
+     case FLOATX80_MUL:
+        test_abz_floatx80( slow_floatx80_mul, floatx80_mul );
+        break;
+     case FLOATX80_DIV:
+        test_abz_floatx80( slow_floatx80_div, floatx80_div );
+        break;
+     case FLOATX80_REM:
+        test_abz_floatx80( slow_floatx80_rem, floatx80_rem );
+        break;
+     case FLOATX80_SQRT:
+        test_az_floatx80( slow_floatx80_sqrt, floatx80_sqrt );
+        break;
+     case FLOATX80_EQ:
+        test_ab_floatx80_z_flag( slow_floatx80_eq, floatx80_eq );
+        break;
+     case FLOATX80_LE:
+        test_ab_floatx80_z_flag( slow_floatx80_le, floatx80_le );
+        break;
+     case FLOATX80_LT:
+        test_ab_floatx80_z_flag( slow_floatx80_lt, floatx80_lt );
+        break;
+     case FLOATX80_EQ_SIGNALING:
+        test_ab_floatx80_z_flag(
+            slow_floatx80_eq_signaling, floatx80_eq_signaling );
+        break;
+     case FLOATX80_LE_QUIET:
+        test_ab_floatx80_z_flag( slow_floatx80_le_quiet, floatx80_le_quiet );
+        break;
+     case FLOATX80_LT_QUIET:
+        test_ab_floatx80_z_flag( slow_floatx80_lt_quiet, floatx80_lt_quiet );
+        break;
+#endif
+#ifdef FLOAT128
+     case FLOAT128_TO_INT32:
+        test_a_float128_z_int32( slow_float128_to_int32, float128_to_int32 );
+        break;
+     case FLOAT128_TO_INT32_ROUND_TO_ZERO:
+        test_a_float128_z_int32(
+            slow_float128_to_int32_round_to_zero,
+            float128_to_int32_round_to_zero
+        );
+        break;
+#ifdef BITS64
+     case FLOAT128_TO_INT64:
+        test_a_float128_z_int64( slow_float128_to_int64, float128_to_int64 );
+        break;
+     case FLOAT128_TO_INT64_ROUND_TO_ZERO:
+        test_a_float128_z_int64(
+            slow_float128_to_int64_round_to_zero,
+            float128_to_int64_round_to_zero
+        );
+        break;
+#endif
+     case FLOAT128_TO_FLOAT32:
+        test_a_float128_z_float32(
+            slow_float128_to_float32, float128_to_float32 );
+        break;
+     case FLOAT128_TO_FLOAT64:
+        test_a_float128_z_float64(
+            slow_float128_to_float64, float128_to_float64 );
+        break;
+#ifdef FLOATX80
+     case FLOAT128_TO_FLOATX80:
+        test_a_float128_z_floatx80(
+            slow_float128_to_floatx80, float128_to_floatx80 );
+        break;
+#endif
+     case FLOAT128_ROUND_TO_INT:
+        test_az_float128( slow_float128_round_to_int, float128_round_to_int );
+        break;
+     case FLOAT128_ADD:
+        test_abz_float128( slow_float128_add, float128_add );
+        break;
+     case FLOAT128_SUB:
+        test_abz_float128( slow_float128_sub, float128_sub );
+        break;
+     case FLOAT128_MUL:
+        test_abz_float128( slow_float128_mul, float128_mul );
+        break;
+     case FLOAT128_DIV:
+        test_abz_float128( slow_float128_div, float128_div );
+        break;
+     case FLOAT128_REM:
+        test_abz_float128( slow_float128_rem, float128_rem );
+        break;
+     case FLOAT128_SQRT:
+        test_az_float128( slow_float128_sqrt, float128_sqrt );
+        break;
+     case FLOAT128_EQ:
+        test_ab_float128_z_flag( slow_float128_eq, float128_eq );
+        break;
+     case FLOAT128_LE:
+        test_ab_float128_z_flag( slow_float128_le, float128_le );
+        break;
+     case FLOAT128_LT:
+        test_ab_float128_z_flag( slow_float128_lt, float128_lt );
+        break;
+     case FLOAT128_EQ_SIGNALING:
+        test_ab_float128_z_flag(
+            slow_float128_eq_signaling, float128_eq_signaling );
+        break;
+     case FLOAT128_LE_QUIET:
+        test_ab_float128_z_flag( slow_float128_le_quiet, float128_le_quiet );
+        break;
+     case FLOAT128_LT_QUIET:
+        test_ab_float128_z_flag( slow_float128_lt_quiet, float128_lt_quiet );
+        break;
+#endif
+    }
+    if ( ( errorStop && anyErrors ) || stop ) exitWithStatus();
+
+}
+
+static void
+ testFunction(
+     uint8 functionCode,
+     int8 roundingPrecisionIn,
+     int8 roundingModeIn,
+     int8 tininessModeIn
+ )
+{
+    int8 roundingPrecision, roundingMode, tininessMode;
+
+    roundingPrecision = 32;
+    for (;;) {
+        if ( ! functions[ functionCode ].roundingPrecision ) {
+            roundingPrecision = 0;
+        }
+        else if ( roundingPrecisionIn ) {
+            roundingPrecision = roundingPrecisionIn;
+        }
+        for ( roundingMode = 1;
+              roundingMode < NUM_ROUNDINGMODES;
+              ++roundingMode
+            ) {
+            if ( ! functions[ functionCode ].roundingMode ) {
+                roundingMode = 0;
+            }
+            else if ( roundingModeIn ) {
+                roundingMode = roundingModeIn;
+            }
+            for ( tininessMode = 1;
+                  tininessMode < NUM_TININESSMODES;
+                  ++tininessMode
+                ) {
+                if (    ( roundingPrecision == 32 )
+                     || ( roundingPrecision == 64 ) ) {
+                    if ( ! functions[ functionCode ]
+                               .tininessModeAtReducedPrecision
+                       ) {
+                        tininessMode = 0;
+                    }
+                    else if ( tininessModeIn ) {
+                        tininessMode = tininessModeIn;
+                    }
+                }
+                else {
+                    if ( ! functions[ functionCode ].tininessMode ) {
+                        tininessMode = 0;
+                    }
+                    else if ( tininessModeIn ) {
+                        tininessMode = tininessModeIn;
+                    }
+                }
+                testFunctionVariety(
+                    functionCode, roundingPrecision, roundingMode, tininessMode
+                );
+                if ( tininessModeIn || ! tininessMode ) break;
+            }
+            if ( roundingModeIn || ! roundingMode ) break;
+        }
+        if ( roundingPrecisionIn || ! roundingPrecision ) break;
+        if ( roundingPrecision == 80 ) {
+            break;
+        }
+        else if ( roundingPrecision == 64 ) {
+            roundingPrecision = 80;
+        }
+        else if ( roundingPrecision == 32 ) {
+            roundingPrecision = 64;
+        }
+    }
+
+}
+
+main( int argc, char **argv )
+{
+    char *argPtr;
+    flag functionArgument;
+    uint8 functionCode;
+    int8 operands, roundingPrecision, roundingMode, tininessMode;
+
+    fail_programName = "testsoftfloat";
+    if ( argc <= 1 ) goto writeHelpMessage;
+    testCases_setLevel( 1 );
+    trueName = "true";
+    testName = "soft";
+    errorStop = FALSE;
+    forever = FALSE;
+    maxErrorCount = 20;
+    trueFlagsPtr = &slow_float_exception_flags;
+    testFlagsFunctionPtr = clearFlags;
+    functionArgument = FALSE;
+    functionCode = 0;
+    operands = 0;
+    roundingPrecision = 0;
+    roundingMode = 0;
+    tininessMode = 0;
+    --argc;
+    ++argv;
+    while ( argc && ( argPtr = argv[ 0 ] ) ) {
+        if ( argPtr[ 0 ] == '-' ) ++argPtr;
+        if ( strcmp( argPtr, "help" ) == 0 ) {
+ writeHelpMessage:
+            fputs(
+"testsoftfloat [<option>...] <function>\n"
+"  <option>:  (* is default)\n"
+"    -help            --Write this message and exit.\n"
+"    -level <num>     --Testing level <num> (1 or 2).\n"
+" *  -level 1\n"
+"    -errors <num>    --Stop each function test after <num> errors.\n"
+" *  -errors 20\n"
+"    -errorstop       --Exit after first function with any error.\n"
+"    -forever         --Test one function repeatedly (implies `-level 2').\n"
+#ifdef FLOATX80
+"    -precision32     --Only test rounding precision equivalent to float32.\n"
+"    -precision64     --Only test rounding precision equivalent to float64.\n"
+"    -precision80     --Only test maximum rounding precision.\n"
+#endif
+"    -nearesteven     --Only test rounding to nearest/even.\n"
+"    -tozero          --Only test rounding to zero.\n"
+"    -down            --Only test rounding down.\n"
+"    -up              --Only test rounding up.\n"
+"    -tininessbefore  --Only test underflow tininess before rounding.\n"
+"    -tininessafter   --Only test underflow tininess after rounding.\n"
+"  <function>:\n"
+"    int32_to_<float>                 <float>_add   <float>_eq\n"
+"    <float>_to_int32                 <float>_sub   <float>_le\n"
+"    <float>_to_int32_round_to_zero   <float>_mul   <float>_lt\n"
+#ifdef BITS64
+"    int64_to_<float>                 <float>_div   <float>_eq_signaling\n"
+"    <float>_to_int64                 <float>_rem   <float>_le_quiet\n"
+"    <float>_to_int64_round_to_zero                 <float>_lt_quiet\n"
+"    <float>_to_<float>\n"
+"    <float>_round_to_int\n"
+"    <float>_sqrt\n"
+#else
+"    <float>_to_<float>               <float>_div   <float>_eq_signaling\n"
+"    <float>_round_to_int             <float>_rem   <float>_le_quiet\n"
+"    <float>_sqrt                                   <float>_lt_quiet\n"
+#endif
+"    -all1            --All 1-operand functions.\n"
+"    -all2            --All 2-operand functions.\n"
+"    -all             --All functions.\n"
+"  <float>:\n"
+"    float32          --Single precision.\n"
+"    float64          --Double precision.\n"
+#ifdef FLOATX80
+"    floatx80         --Extended double precision.\n"
+#endif
+#ifdef FLOAT128
+"    float128         --Quadruple precision.\n"
+#endif
+                ,
+                stdout
+            );
+            return EXIT_SUCCESS;
+        }
+        else if ( strcmp( argPtr, "level" ) == 0 ) {
+            if ( argc < 2 ) goto optionError;
+            testCases_setLevel( atoi( argv[ 1 ] ) );
+            --argc;
+            ++argv;
+        }
+        else if ( strcmp( argPtr, "level1" ) == 0 ) {
+            testCases_setLevel( 1 );
+        }
+        else if ( strcmp( argPtr, "level2" ) == 0 ) {
+            testCases_setLevel( 2 );
+        }
+        else if ( strcmp( argPtr, "errors" ) == 0 ) {
+            if ( argc < 2 ) {
+     optionError:
+                fail( "`%s' option requires numeric argument", argv[ 0 ] );
+            }
+            maxErrorCount = atoi( argv[ 1 ] );
+            --argc;
+            ++argv;
+        }
+        else if ( strcmp( argPtr, "errorstop" ) == 0 ) {
+            errorStop = TRUE;
+        }
+        else if ( strcmp( argPtr, "forever" ) == 0 ) {
+            testCases_setLevel( 2 );
+            forever = TRUE;
+        }
+#ifdef FLOATX80
+        else if ( strcmp( argPtr, "precision32" ) == 0 ) {
+            roundingPrecision = 32;
+        }
+        else if ( strcmp( argPtr, "precision64" ) == 0 ) {
+            roundingPrecision = 64;
+        }
+        else if ( strcmp( argPtr, "precision80" ) == 0 ) {
+            roundingPrecision = 80;
+        }
+#endif
+        else if (    ( strcmp( argPtr, "nearesteven" ) == 0 )
+                  || ( strcmp( argPtr, "nearest_even" ) == 0 ) ) {
+            roundingMode = ROUND_NEAREST_EVEN;
+        }
+        else if (    ( strcmp( argPtr, "tozero" ) == 0 )
+                  || ( strcmp( argPtr, "to_zero" ) == 0 ) ) {
+            roundingMode = ROUND_TO_ZERO;
+        }
+        else if ( strcmp( argPtr, "down" ) == 0 ) {
+            roundingMode = ROUND_DOWN;
+        }
+        else if ( strcmp( argPtr, "up" ) == 0 ) {
+            roundingMode = ROUND_UP;
+        }
+        else if ( strcmp( argPtr, "tininessbefore" ) == 0 ) {
+            tininessMode = TININESS_BEFORE_ROUNDING;
+        }
+        else if ( strcmp( argPtr, "tininessafter" ) == 0 ) {
+            tininessMode = TININESS_AFTER_ROUNDING;
+        }
+        else if ( strcmp( argPtr, "all1" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 1;
+        }
+        else if ( strcmp( argPtr, "all2" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 2;
+        }
+        else if ( strcmp( argPtr, "all" ) == 0 ) {
+            functionArgument = TRUE;
+            functionCode = 0;
+            operands = 0;
+        }
+        else {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( strcmp( argPtr, functions[ functionCode ].name ) == 0 ) {
+                    break;
+                }
+            }
+            if ( functionCode == NUM_FUNCTIONS ) {
+                fail( "Invalid option or function `%s'", argv[ 0 ] );
+            }
+            functionArgument = TRUE;
+        }
+        --argc;
+        ++argv;
+    }
+    if ( ! functionArgument ) fail( "Function argument required" );
+    (void) signal( SIGINT, catchSIGINT );
+    (void) signal( SIGTERM, catchSIGINT );
+    if ( functionCode ) {
+        if ( forever ) {
+            if ( ! roundingPrecision ) roundingPrecision = 80;
+            if ( ! roundingMode ) roundingMode = ROUND_NEAREST_EVEN;
+        }
+        testFunction(
+            functionCode, roundingPrecision, roundingMode, tininessMode );
+    }
+    else {
+        if ( operands == 1 ) {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( functions[ functionCode ].numInputs == 1 ) {
+                    testFunction(
+                        functionCode,
+                        roundingPrecision,
+                        roundingMode,
+                        tininessMode
+                    );
+                }
+            }
+        }
+        else if ( operands == 2 ) {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                if ( functions[ functionCode ].numInputs == 2 ) {
+                    testFunction(
+                        functionCode,
+                        roundingPrecision,
+                        roundingMode,
+                        tininessMode
+                    );
+                }
+            }
+        }
+        else {
+            for ( functionCode = 1;
+                  functionCode < NUM_FUNCTIONS;
+                  ++functionCode
+                ) {
+                testFunction(
+                    functionCode, roundingPrecision, roundingMode, tininessMode
+                );
+            }
+        }
+    }
+    exitWithStatus();
+
+}
+
diff --git a/testfloat/writeHex.c b/testfloat/writeHex.c
new file mode 100644
index 00000000000..f304aed631d
--- /dev/null
+++ b/testfloat/writeHex.c
@@ -0,0 +1,183 @@
+
+/*
+===============================================================================
+
+This C source file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdio.h>
+#include "milieu.h"
+#include "softfloat.h"
+#include "writeHex.h"
+
+void writeHex_flag( flag a, FILE *stream )
+{
+
+    fputc( a ? '1' : '0', stream );
+
+}
+
+static void writeHex_bits8( bits8 a, FILE *stream )
+{
+    int digit;
+
+    digit = ( a>>4 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = a & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+
+}
+
+static void writeHex_bits12( int16 a, FILE *stream )
+{
+    int digit;
+
+    digit = ( a>>8 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = ( a>>4 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = a & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+
+}
+
+static void writeHex_bits16( bits16 a, FILE *stream )
+{
+    int digit;
+
+    digit = ( a>>12 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = ( a>>8 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = ( a>>4 ) & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+    digit = a & 0xF;
+    if ( 9 < digit ) digit += 'A' - ( '0' + 10 );
+    fputc( '0' + digit, stream );
+
+}
+
+void writeHex_bits32( bits32 a, FILE *stream )
+{
+
+    writeHex_bits16( a>>16, stream );
+    writeHex_bits16( a, stream );
+
+}
+
+#ifdef BITS64
+
+void writeHex_bits64( bits64 a, FILE *stream )
+{
+
+    writeHex_bits32( a>>32, stream );
+    writeHex_bits32( a, stream );
+
+}
+
+#endif
+
+void writeHex_float32( float32 a, FILE *stream )
+{
+
+    fputc( ( ( (sbits32) a ) < 0 ) ? '8' : '0', stream );
+    writeHex_bits8( a>>23, stream );
+    fputc( '.', stream );
+    writeHex_bits8( ( a>>16 ) & 0x7F, stream );
+    writeHex_bits16( a, stream );
+
+}
+
+#ifdef BITS64
+
+void writeHex_float64( float64 a, FILE *stream )
+{
+
+    writeHex_bits12( a>>52, stream );
+    fputc( '.', stream );
+    writeHex_bits12( a>>40, stream );
+    writeHex_bits8( a>>32, stream );
+    writeHex_bits32( a, stream );
+
+}
+
+#else
+
+void writeHex_float64( float64 a, FILE *stream )
+{
+
+    writeHex_bits12( a.high>>20, stream );
+    fputc( '.', stream );
+    writeHex_bits12( a.high>>8, stream );
+    writeHex_bits8( a.high, stream );
+    writeHex_bits32( a.low, stream );
+
+}
+
+#endif
+
+#ifdef FLOATX80
+
+void writeHex_floatx80( floatx80 a, FILE *stream )
+{
+
+    writeHex_bits16( a.high, stream );
+    fputc( '.', stream );
+    writeHex_bits64( a.low, stream );
+
+}
+
+#endif
+
+#ifdef FLOAT128
+
+void writeHex_float128( float128 a, FILE *stream )
+{
+
+    writeHex_bits16( a.high>>48, stream );
+    fputc( '.', stream );
+    writeHex_bits16( a.high>>32, stream );
+    writeHex_bits32( a.high, stream );
+    writeHex_bits64( a.low, stream );
+
+}
+
+#endif
+
+void writeHex_float_flags( uint8 flags, FILE *stream )
+{
+
+    fputc( flags & float_flag_invalid   ? 'v' : '.', stream );
+    fputc( flags & float_flag_divbyzero ? 'z' : '.', stream );
+    fputc( flags & float_flag_overflow  ? 'o' : '.', stream );
+    fputc( flags & float_flag_underflow ? 'u' : '.', stream );
+    fputc( flags & float_flag_inexact   ? 'x' : '.', stream );
+
+}
+
diff --git a/testfloat/writeHex.h b/testfloat/writeHex.h
new file mode 100644
index 00000000000..0dde44c43d3
--- /dev/null
+++ b/testfloat/writeHex.h
@@ -0,0 +1,42 @@
+
+/*
+===============================================================================
+
+This C header file is part of TestFloat, Release 2a, a package of programs
+for testing the correctness of floating-point arithmetic complying to the
+IEC/IEEE Standard for Floating-Point.
+
+Written by John R. Hauser.  More information is available through the Web
+page `http://HTTP.CS.Berkeley.EDU/~jhauser/arithmetic/TestFloat.html'.
+
+THIS SOFTWARE IS DISTRIBUTED AS IS, FOR FREE.  Although reasonable effort
+has been made to avoid it, THIS SOFTWARE MAY CONTAIN FAULTS THAT WILL AT
+TIMES RESULT IN INCORRECT BEHAVIOR.  USE OF THIS SOFTWARE IS RESTRICTED TO
+PERSONS AND ORGANIZATIONS WHO CAN AND WILL TAKE FULL RESPONSIBILITY FOR ANY
+AND ALL LOSSES, COSTS, OR OTHER PROBLEMS ARISING FROM ITS USE.
+
+Derivative works are acceptable, even for commercial purposes, so long as
+(1) they include prominent notice that the work is derivative, and (2) they
+include prominent notice akin to these four paragraphs for those parts of
+this code that are retained.
+
+===============================================================================
+*/
+
+#include <stdio.h>
+
+void writeHex_flag( flag, FILE * );
+void writeHex_bits32( bits32, FILE * );
+#ifdef BITS64
+void writeHex_bits64( bits64, FILE * );
+#endif
+void writeHex_float32( float32, FILE * );
+void writeHex_float64( float64, FILE * );
+#ifdef FLOATX80
+void writeHex_floatx80( floatx80, FILE * );
+#endif
+#ifdef FLOAT128
+void writeHex_float128( float128, FILE * );
+#endif
+void writeHex_float_flags( uint8, FILE * );
+

From f79727118c771153044ad18212527a9fe8146ae2 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Tue, 20 Apr 2010 21:04:57 +0000
Subject: [PATCH 171/532] Fix another instance of lapic_cyclic_clock_func.

---
 sys/x86/isa/clock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/x86/isa/clock.c b/sys/x86/isa/clock.c
index 83026d6c258..36dc36e68c0 100644
--- a/sys/x86/isa/clock.c
+++ b/sys/x86/isa/clock.c
@@ -186,8 +186,8 @@ clkintr(struct trapframe *frame)
 	 * timers.
 	 */
 	int cpu = PCPU_GET(cpuid);
-	if (lapic_cyclic_clock_func[cpu] != NULL)
-		(*lapic_cyclic_clock_func[cpu])(frame);
+	if (cyclic_clock_func[cpu] != NULL)
+		(*cyclic_clock_func[cpu])(frame);
 #endif
 
 	if (using_atrtc_timer) {

From 86c02aae879814b98145d5a7611cf3831e2ba3a5 Mon Sep 17 00:00:00 2001
From: Ryan Stone <rstone@FreeBSD.org>
Date: Tue, 20 Apr 2010 22:43:53 +0000
Subject: [PATCH 172/532] add new committer rstone to committers-src.dot

---
 share/misc/committers-src.dot | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/share/misc/committers-src.dot b/share/misc/committers-src.dot
index f7e4a326e66..75f88ad17c7 100644
--- a/share/misc/committers-src.dot
+++ b/share/misc/committers-src.dot
@@ -170,6 +170,7 @@ roberto [label="Ollivier Robert\nroberto@FreeBSD.org\n1995/02/22"]
 rpaulo [label="Rui Paulo\nrpaulo@FreeBSD.org\n2007/09/25"]
 rrs [label="Randall R Stewart\nrrs@FreeBSD.org\n2007/02/08"]
 rse [label="Ralf S. Engelschall\nrse@FreeBSD.org\n1997/07/31"]
+rstone [label="Ryan Stone\nrstone@FreeBSD.org\n2010/04/19"]
 ru [label="Ruslan Ermilov\nru@FreeBSD.org\n1999/05/27"]
 rwatson [label="Robert N. M. Watson\nrwatson@FreeBSD.org\n1999/12/16"]
 sam [label="Sam Leffler\nsam@FreeBSD.org\n2002/07/02"]
@@ -271,6 +272,8 @@ ed -> uqs
 eivind -> des
 eivind -> rwatson
 
+emaste -> rstone
+
 emax -> markus
 
 gallatin -> ticso
@@ -342,6 +345,7 @@ jkh -> yar
 
 jkoshy -> kaiw
 jkoshy -> fabient
+jkoshy -> rstone
 
 jlemon -> bmilekic
 jlemon -> brooks

From 0e4ae43e17351d26d597e0ef1e25310d2ee0b195 Mon Sep 17 00:00:00 2001
From: Robert Watson <rwatson@FreeBSD.org>
Date: Wed, 21 Apr 2010 00:52:55 +0000
Subject: [PATCH 173/532] Merge @176820, @176822, @177156 to tcpp from P4 to
 HEAD:

  Improve accuracy of connection data transfer math.

  Disable Nagle's algorithm to avoid delaying transfers of data --
  will want to refine this to combine payload with header transfer,
  however.

  Now that we're running w/o Nagle, try to send the initial data
  burst with the header in a single TCP segment.

  Prefer %zu to %ju for size_t.

MFC after:	1 week
Sponsored by:	Juniper, Inc.
---
 tools/tools/netrate/tcpp/tcpp_client.c | 47 +++++++++++++++++++-------
 tools/tools/netrate/tcpp/tcpp_server.c |  5 +++
 2 files changed, 39 insertions(+), 13 deletions(-)

diff --git a/tools/tools/netrate/tcpp/tcpp_client.c b/tools/tools/netrate/tcpp/tcpp_client.c
index 15df06eda77..6c3e800366b 100644
--- a/tools/tools/netrate/tcpp/tcpp_client.c
+++ b/tools/tools/netrate/tcpp/tcpp_client.c
@@ -33,9 +33,11 @@
 #include <sys/socket.h>
 #include <sys/sysctl.h>
 #include <sys/time.h>
+#include <sys/uio.h>
 #include <sys/wait.h>
 
 #include <netinet/in.h>
+#include <netinet/tcp.h>
 
 #include <err.h>
 #include <errno.h>
@@ -82,6 +84,7 @@ static int			 kq;
 static int			 started;	/* Number started so far. */
 static int			 finished;	/* Number finished so far. */
 static int			 counter;	/* IP number offset. */
+static uint64_t			 payload_len;
 
 static struct connection *
 tcpp_client_newconn(void)
@@ -109,6 +112,9 @@ tcpp_client_newconn(void)
 	i = 1;
 	if (setsockopt(fd, SOL_SOCKET, SO_NOSIGPIPE, &i, sizeof(i)) < 0)
 		err(-1, "setsockopt");
+	i = 1;
+	if (setsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &i, sizeof(i)) < 0)
+		err(-1, "setsockopt");
 #if 0
 	i = 1;
 	if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &i, sizeof(i)) < 0)
@@ -131,7 +137,7 @@ tcpp_client_newconn(void)
 	conn->conn_magic = CONNECTION_MAGIC;
 	conn->conn_fd = fd;
 	conn->conn_header.th_magic = TCPP_MAGIC;
-	conn->conn_header.th_len = bflag;
+	conn->conn_header.th_len = payload_len;
 	tcpp_header_encode(&conn->conn_header);
 
 	EV_SET(&kev, fd, EVFILT_WRITE, EV_ADD, 0, 0, conn);
@@ -156,16 +162,22 @@ static void
 tcpp_client_handleconn(struct kevent *kev)
 {
 	struct connection *conn;
-	ssize_t len;
+	struct iovec iov[2];
+	ssize_t len, header_left;
 
 	conn = kev->udata;
 	if (conn->conn_magic != CONNECTION_MAGIC)
 		errx(-1, "tcpp_client_handleconn: magic");
 
 	if (conn->conn_header_sent < sizeof(conn->conn_header)) {
-		len = write(conn->conn_fd, ((u_char *)&conn->conn_header) +
-		    conn->conn_header_sent, sizeof(conn->conn_header) -
-		    conn->conn_header_sent);
+		header_left = sizeof(conn->conn_header) -
+		    conn->conn_header_sent;
+		iov[0].iov_base = ((u_char *)&conn->conn_header) +
+		    conn->conn_header_sent;
+		iov[0].iov_len = header_left;
+		iov[1].iov_base = buffer;
+		iov[1].iov_len = min(sizeof(buffer), payload_len);
+		len = writev(conn->conn_fd, iov, 2);
 		if (len < 0) {
 			tcpp_client_closeconn(conn);
 			err(-1, "tcpp_client_handleconn: header write");
@@ -175,10 +187,14 @@ tcpp_client_handleconn(struct kevent *kev)
 			errx(-1, "tcpp_client_handleconn: header write "
 			    "premature EOF");
 		}
-		conn->conn_header_sent += len;
+		if (len > header_left) {
+			conn->conn_data_sent += (len - header_left);
+			conn->conn_header_sent += header_left;
+		} else
+			conn->conn_header_sent += len;
 	} else {
 		len = write(conn->conn_fd, buffer, min(sizeof(buffer),
-		    bflag - conn->conn_data_sent));
+		    payload_len - conn->conn_data_sent));
 		if (len < 0) {
 			tcpp_client_closeconn(conn);
 			err(-1, "tcpp_client_handleconn: data write");
@@ -189,12 +205,12 @@ tcpp_client_handleconn(struct kevent *kev)
 			    "premature EOF");
 		}
 		conn->conn_data_sent += len;
-		if (conn->conn_data_sent >= bflag) {
-			/*
-			 * All is well.
-			 */
-			tcpp_client_closeconn(conn);
-		}
+	}
+	if (conn->conn_data_sent >= payload_len) {
+		/*
+		 * All is well.
+		 */
+		tcpp_client_closeconn(conn);
 	}
 }
 
@@ -261,6 +277,11 @@ tcpp_client(void)
 	pid_t pid;
 	int i, failed, status;
 
+	if (bflag < sizeof(struct tcpp_header))
+		errx(-1, "Can't use -b less than %zu\n",
+		   sizeof(struct tcpp_header));
+	payload_len = bflag - sizeof(struct tcpp_header);
+
 	pid_list = malloc(sizeof(*pid_list) * pflag);
 	if (pid_list == NULL)
 		err(-1, "malloc pid_list");
diff --git a/tools/tools/netrate/tcpp/tcpp_server.c b/tools/tools/netrate/tcpp/tcpp_server.c
index 8101005b2bd..76f5e8d5152 100644
--- a/tools/tools/netrate/tcpp/tcpp_server.c
+++ b/tools/tools/netrate/tcpp/tcpp_server.c
@@ -37,6 +37,7 @@
 #include <sys/wait.h>
 
 #include <netinet/in.h>
+#include <netinet/tcp.h>
 
 #include <err.h>
 #include <fcntl.h>
@@ -239,6 +240,10 @@ tcpp_server_worker(int workernum)
 	if (setsockopt(listen_sock, SOL_SOCKET, SO_REUSEPORT, &i, sizeof(i))
 	    < 0)
 		err(-1, "setsockopt");
+	i = 1;
+	if (setsockopt(listen_sock, IPPROTO_TCP, TCP_NODELAY, &i, sizeof(i))
+	    < 0)
+		err(-1, "setsockopt");
 	if (bind(listen_sock, (struct sockaddr *)&localipbase,
 	    sizeof(localipbase)) < 0)
 		err(-1, "bind");

From 611049ae75c16fe84438c398381ac6f184ce5b37 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Wed, 21 Apr 2010 01:13:08 +0000
Subject: [PATCH 174/532] When CPUTYPE is defined to any value, on amd64
 platform "mmx" is available through MACHINE_CPU, indicating the CPU supports
 that feature, as done by revision 138685.

This changeset adds "mmx" into the default amd64 MACHINE_CPU list
when no CPUTYPE is specified to provide consistent behavior.

PR:		amd64/145593
Submitted by:	mm
MFC after:	2 weeks
---
 share/mk/bsd.cpu.mk | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/share/mk/bsd.cpu.mk b/share/mk/bsd.cpu.mk
index 35661b30807..86f9aed0d5d 100644
--- a/share/mk/bsd.cpu.mk
+++ b/share/mk/bsd.cpu.mk
@@ -9,7 +9,7 @@ _CPUCFLAGS =
 . if ${MACHINE_ARCH} == "i386"
 MACHINE_CPU = i486
 . elif ${MACHINE_ARCH} == "amd64"
-MACHINE_CPU = amd64 sse2 sse
+MACHINE_CPU = amd64 sse2 sse mmx
 . elif ${MACHINE_ARCH} == "ia64"
 MACHINE_CPU = itanium
 . elif ${MACHINE_ARCH} == "powerpc"

From 3d65a7ed79b2668712ceeee6ba12e43edcab8143 Mon Sep 17 00:00:00 2001
From: Randi Harper <randi@FreeBSD.org>
Date: Wed, 21 Apr 2010 08:18:06 +0000
Subject: [PATCH 175/532] Add myself to committers-src.dot.

Approved by:	cperciva (mentor)
---
 share/misc/committers-src.dot | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/share/misc/committers-src.dot b/share/misc/committers-src.dot
index 75f88ad17c7..144d08ab68e 100644
--- a/share/misc/committers-src.dot
+++ b/share/misc/committers-src.dot
@@ -162,6 +162,7 @@ pjd [label="Pawel Jakub Dawidek\npjd@FreeBSD.org\n2004/02/02"]
 ps [label="Paul Saab\nps@FreeBSD.org\n2000/02/23"]
 qingli [label="Qing Li\nqingli@FreeBSD.org\n2005/04/13"]
 rafan [label="Rong-En Fan\nrafan@FreeBSD.org\n2007/01/31"]
+randi [label="Randi Harper\nrandi@FreeBSD.org\n2010/04/20"]
 remko [label="Remko Lodder\nremko@FreeBSD.org\n2007/02/23"]
 rik [label="Roman Kurakin\nrik@FreeBSD.org\n2003/12/18"]
 rink [label="Rink Springer\nrink@FreeBSD.org\n2006/01/16"]
@@ -245,6 +246,7 @@ bz -> jamie
 bz -> syrinx
 
 cperciva -> flz
+cperciva -> randi
 cperciva -> simon
 
 csjp -> bushman

From 1c044382c345293e593c6d9f8daeaef78019ce40 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Wed, 21 Apr 2010 10:21:34 +0000
Subject: [PATCH 176/532] Avoid memory access after free.  Use the (shortend)
 copy for the ipsec mtu lookup as well.

PR:		kern/145736
Submitted by:	Peter Molnar (peter molnar.cc)
MFC after:	3 days
---
 sys/netinet/ip_input.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/netinet/ip_input.c b/sys/netinet/ip_input.c
index 8be51fb992d..fdef6457b3e 100644
--- a/sys/netinet/ip_input.c
+++ b/sys/netinet/ip_input.c
@@ -1590,7 +1590,7 @@ ip_forward(struct mbuf *m, int srcrt)
 		 * If IPsec is configured for this path,
 		 * override any possibly mtu value set by ip_output.
 		 */ 
-		mtu = ip_ipsec_mtu(m, mtu);
+		mtu = ip_ipsec_mtu(mcopy, mtu);
 #endif /* IPSEC */
 		/*
 		 * If the MTU was set before make sure we are below the

From 94c6c6ba672c567c414259879481d4f20a35727a Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 11:17:16 +0000
Subject: [PATCH 177/532] As was done in r155238 for i386 and in r155239 for
 amd64, clear the carry flag for ia32 binary executed on amd64 host in
 get_mcontext().

PR:	kern/92110 (one more time)
Reported by:	stas
MFC after:	1 week
---
 sys/amd64/ia32/ia32_signal.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/amd64/ia32/ia32_signal.c b/sys/amd64/ia32/ia32_signal.c
index 8033443ffad..859b9923f68 100644
--- a/sys/amd64/ia32/ia32_signal.c
+++ b/sys/amd64/ia32/ia32_signal.c
@@ -141,9 +141,11 @@ ia32_get_mcontext(struct thread *td, struct ia32_mcontext *mcp, int flags)
 	mcp->mc_esi = tp->tf_rsi;
 	mcp->mc_ebp = tp->tf_rbp;
 	mcp->mc_isp = tp->tf_rsp;
+	mcp->mc_eflags = tp->tf_rflags;
 	if (flags & GET_MC_CLEAR_RET) {
 		mcp->mc_eax = 0;
 		mcp->mc_edx = 0;
+		mcp->mc_eflags &= ~PSL_C;
 	} else {
 		mcp->mc_eax = tp->tf_rax;
 		mcp->mc_edx = tp->tf_rdx;
@@ -152,7 +154,6 @@ ia32_get_mcontext(struct thread *td, struct ia32_mcontext *mcp, int flags)
 	mcp->mc_ecx = tp->tf_rcx;
 	mcp->mc_eip = tp->tf_rip;
 	mcp->mc_cs = tp->tf_cs;
-	mcp->mc_eflags = tp->tf_rflags;
 	mcp->mc_esp = tp->tf_rsp;
 	mcp->mc_ss = tp->tf_ss;
 	mcp->mc_len = sizeof(*mcp);

From 45ea4bc76f7d57924f15b2c2104d057054baf659 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 21 Apr 2010 11:28:13 +0000
Subject: [PATCH 178/532] Comply with the new cyclic dtrace module variable
 name (cyclic_clock_func).

---
 sys/pc98/cbus/clock.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/pc98/cbus/clock.c b/sys/pc98/cbus/clock.c
index 3028c074d15..8c5a7d2af75 100644
--- a/sys/pc98/cbus/clock.c
+++ b/sys/pc98/cbus/clock.c
@@ -175,8 +175,8 @@ clkintr(struct trapframe *frame)
 	 * timers.
 	 */
 	int cpu = PCPU_GET(cpuid);
-	if (lapic_cyclic_clock_func[cpu] != NULL)
-		(*lapic_cyclic_clock_func[cpu])(frame);
+	if (cyclic_clock_func[cpu] != NULL)
+		(*cyclic_clock_func[cpu])(frame);
 #endif
 
 #ifdef SMP

From cb5a415375015ad7af565f4e4c0f9f583f342cae Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Wed, 21 Apr 2010 11:50:13 +0000
Subject: [PATCH 179/532] Apply threshold filter to root node in calltree view.

MFC after: 3days
---
 usr.sbin/pmcstat/pmcpl_calltree.c | 15 ++++++++++-----
 1 file changed, 10 insertions(+), 5 deletions(-)

diff --git a/usr.sbin/pmcstat/pmcpl_calltree.c b/usr.sbin/pmcstat/pmcpl_calltree.c
index 4f62c62158b..f8ceece158b 100644
--- a/usr.sbin/pmcstat/pmcpl_calltree.c
+++ b/usr.sbin/pmcstat/pmcpl_calltree.c
@@ -499,9 +499,10 @@ void
 pmcpl_ct_topdisplay(void)
 {
 	int i, x, y, pmcin;
-	struct pmcpl_ct_sample rsamples;
+	struct pmcpl_ct_sample r, *rsamples;
 
-	pmcpl_ct_samples_root(&rsamples);
+	rsamples = &r;
+	pmcpl_ct_samples_root(rsamples);
 
 	PMCSTAT_PRINTW("%-10.10s %s\n", "IMAGE", "CALLTREE");
 
@@ -524,16 +525,20 @@ pmcpl_ct_topdisplay(void)
 			if (PMCPL_CT_SAMPLE(pmcin,
 			    &pmcpl_ct_root->pct_arc[i].pcta_samples) == 0)
 				continue;
+			if (PMCPL_CT_SAMPLEP(pmcin,
+			    &pmcpl_ct_root->pct_arc[i].pcta_samples) <=
+			    pmcstat_threshold)
+				continue;
 			if (pmcpl_ct_node_dumptop(pmcin,
 			        pmcpl_ct_root->pct_arc[i].pcta_child,
-			        &rsamples, x, &y, pmcstat_displayheight - 2)) {
+			        rsamples, x, &y, pmcstat_displayheight - 2)) {
 				break;
 			}
 		}
 
-		pmcpl_ct_node_printtop(&rsamples, pmcin, y);
+		pmcpl_ct_node_printtop(rsamples, pmcin, y);
 	}
-	pmcpl_ct_samples_free(&rsamples);
+	pmcpl_ct_samples_free(rsamples);
 }
 
 /*

From 97293c583326ef1d7818eee9aacf15b757d47404 Mon Sep 17 00:00:00 2001
From: Randi Harper <randi@FreeBSD.org>
Date: Wed, 21 Apr 2010 12:02:45 +0000
Subject: [PATCH 180/532] Fix bug introduced in r198317: Don't leak a file
 descriptor after scanning for network devices.

Approved by:	cperciva (mentor)
MFC after:	1 week (original commit + bugfix)
---
 usr.sbin/sysinstall/tcpip.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/usr.sbin/sysinstall/tcpip.c b/usr.sbin/sysinstall/tcpip.c
index 75a40e87512..1c1d6766cd2 100644
--- a/usr.sbin/sysinstall/tcpip.c
+++ b/usr.sbin/sysinstall/tcpip.c
@@ -683,6 +683,8 @@ tcpDeviceScan(void)
 		}
 	}
 
+	close(s);
+
 	freeifaddrs(ifap);
 
 	return (NULL);

From ee9c0bb32bd8df6307042c141a3de84edbaa1c92 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Wed, 21 Apr 2010 12:17:01 +0000
Subject: [PATCH 181/532] indent.pro example: actually install the sample file

---
 etc/mtree/BSD.usr.dist  | 2 ++
 share/examples/Makefile | 2 ++
 2 files changed, 4 insertions(+)

diff --git a/etc/mtree/BSD.usr.dist b/etc/mtree/BSD.usr.dist
index 87f7ca39767..da3ad541b72 100644
--- a/etc/mtree/BSD.usr.dist
+++ b/etc/mtree/BSD.usr.dist
@@ -219,6 +219,8 @@
             ..
             ibcs2
             ..
+            indent
+            ..
             ipfilter
             ..
             ipfw
diff --git a/share/examples/Makefile b/share/examples/Makefile
index 99d92c0d99a..7d011e3a9cd 100644
--- a/share/examples/Makefile
+++ b/share/examples/Makefile
@@ -15,6 +15,7 @@ LDIRS=	BSD_daemon \
 	find_interface \
 	hast \
 	ibcs2 \
+	indent \
 	ipfw \
 	kld \
 	libvgl \
@@ -77,6 +78,7 @@ XFILES=	BSD_daemon/FreeBSD.pfa \
 	hast/vip-up.sh \
 	ibcs2/README \
 	ibcs2/hello.uu \
+	indent/indent.pro \
 	ipfw/change_rules.sh \
 	kld/Makefile \
 	kld/cdev/Makefile \

From 585b8e0f8a233e6b3e74182d03163657c5630a5c Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 16:38:37 +0000
Subject: [PATCH 182/532] Move realpath(3) prototype to a POSIX section.

Noted by:	bde
MFC after:	2 weeks
---
 include/stdlib.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/include/stdlib.h b/include/stdlib.h
index b833203c65b..3c48f1746f8 100644
--- a/include/stdlib.h
+++ b/include/stdlib.h
@@ -155,6 +155,7 @@ void	 _Exit(int) __dead2;
 #if __POSIX_VISIBLE /* >= ??? */
 int	 posix_memalign(void **, size_t, size_t); /* (ADV) */
 int	 rand_r(unsigned *);			/* (TSF) */
+char	*realpath(const char * __restrict, char * __restrict);
 int	 setenv(const char *, const char *, int);
 int	 unsetenv(const char *);
 #endif
@@ -201,7 +202,6 @@ int	 posix_openpt(int);
 char	*ptsname(int);
 int	 putenv(char *);
 long	 random(void);
-char	*realpath(const char * __restrict, char * __restrict);
 unsigned short
 	*seed48(unsigned short[3]);
 #ifndef _SETKEY_DECLARED

From 7e36a19d5175b1fe085903ee15900175b9418559 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 16:41:02 +0000
Subject: [PATCH 183/532] Add standards section, improve wording, taking into
 account the handling of NULL and changed type in declaration.

Suggested by:	bde
MFC after:	2 weeks
---
 lib/libc/stdlib/realpath.3 | 14 ++++++++------
 1 file changed, 8 insertions(+), 6 deletions(-)

diff --git a/lib/libc/stdlib/realpath.3 b/lib/libc/stdlib/realpath.3
index 371ee3b2804..aee65f7600f 100644
--- a/lib/libc/stdlib/realpath.3
+++ b/lib/libc/stdlib/realpath.3
@@ -94,14 +94,11 @@ If an error occurs,
 .Fn realpath
 returns
 .Dv NULL ,
-and
+and if
 .Fa resolved_path
-contains the pathname which caused the problem.
-If
-.Fa resolved_path
-was
+is not
 .Dv NULL ,
-then information of the failed pathname component is lost.
+the array that it points to contains the pathname which caused the problem.
 .Sh ERRORS
 The function
 .Fn realpath
@@ -126,6 +123,11 @@ when given a relative
 .Fa pathname .
 .Sh "SEE ALSO"
 .Xr getcwd 3
+.Sh STANDARDS
+The
+.Fn realpath
+function conforms to
+.St -p1003.1-2001 .
 .Sh HISTORY
 The
 .Fn realpath

From 698de07abbb0ecdbccb24a68483d380300fd78cd Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Wed, 21 Apr 2010 17:38:16 +0000
Subject: [PATCH 184/532] Use correct bus_dma_tag_t for TX frames.

Reported by:	Andreas Nilsson <andrnils at gmail.com>
Approved by:	rpaulo (mentor)
MFC after:	3 days
---
 sys/dev/iwn/if_iwn.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sys/dev/iwn/if_iwn.c b/sys/dev/iwn/if_iwn.c
index 2e60da3bf1b..09c1efdbe71 100644
--- a/sys/dev/iwn/if_iwn.c
+++ b/sys/dev/iwn/if_iwn.c
@@ -2282,6 +2282,7 @@ iwn4965_tx_done(struct iwn_softc *sc, struct iwn_rx_desc *desc,
     struct iwn_rx_data *data)
 {
 	struct iwn4965_tx_stat *stat = (struct iwn4965_tx_stat *)(desc + 1);
+	struct iwn_tx_ring *ring = &sc->txq[desc->qid & 0xf];
 
 	DPRINTF(sc, IWN_DEBUG_XMIT, "%s: "
 	    "qid %d idx %d retries %d nkill %d rate %x duration %d status %x\n",
@@ -2289,7 +2290,7 @@ iwn4965_tx_done(struct iwn_softc *sc, struct iwn_rx_desc *desc,
 	    stat->btkillcnt, stat->rate, le16toh(stat->duration),
 	    le32toh(stat->status));
 
-	bus_dmamap_sync(sc->rxq.data_dmat, data->map, BUS_DMASYNC_POSTREAD);
+	bus_dmamap_sync(ring->data_dmat, data->map, BUS_DMASYNC_POSTREAD);
 	iwn_tx_done(sc, desc, stat->ackfailcnt, le32toh(stat->status) & 0xff);
 }
 
@@ -2298,6 +2299,7 @@ iwn5000_tx_done(struct iwn_softc *sc, struct iwn_rx_desc *desc,
     struct iwn_rx_data *data)
 {
 	struct iwn5000_tx_stat *stat = (struct iwn5000_tx_stat *)(desc + 1);
+	struct iwn_tx_ring *ring = &sc->txq[desc->qid & 0xf];
 
 	DPRINTF(sc, IWN_DEBUG_XMIT, "%s: "
 	    "qid %d idx %d retries %d nkill %d rate %x duration %d status %x\n",
@@ -2310,7 +2312,7 @@ iwn5000_tx_done(struct iwn_softc *sc, struct iwn_rx_desc *desc,
 	iwn5000_reset_sched(sc, desc->qid & 0xf, desc->idx);
 #endif
 
-	bus_dmamap_sync(sc->rxq.data_dmat, data->map, BUS_DMASYNC_POSTREAD);
+	bus_dmamap_sync(ring->data_dmat, data->map, BUS_DMASYNC_POSTREAD);
 	iwn_tx_done(sc, desc, stat->ackfailcnt, le16toh(stat->status) & 0xff);
 }
 

From aff644837180af24e083c9b877d96d8e2605581e Mon Sep 17 00:00:00 2001
From: Rebecca Cran <brucec@FreeBSD.org>
Date: Wed, 21 Apr 2010 18:55:44 +0000
Subject: [PATCH 185/532] Make the "Q = Finish" text visible when running
 sysinstall as a normal application.  Reword the boot manager screen to try
 and avoid confusion.

Reviewed by:	randi
Approved by:	rrs (mentor)
MFC after:	1 week
---
 usr.sbin/sysinstall/disks.c |  7 +++++--
 usr.sbin/sysinstall/menus.c | 16 ++++++----------
 2 files changed, 11 insertions(+), 12 deletions(-)

diff --git a/usr.sbin/sysinstall/disks.c b/usr.sbin/sysinstall/disks.c
index a1c09b659ef..c0fd12a518b 100644
--- a/usr.sbin/sysinstall/disks.c
+++ b/usr.sbin/sysinstall/disks.c
@@ -207,9 +207,12 @@ print_command_summary(void)
     mvprintw(14, 0, "The following commands are supported (in upper or lower case):");
     mvprintw(16, 0, "A = Use Entire Disk   G = set Drive Geometry   C = Create Slice");
     mvprintw(17, 0, "D = Delete Slice      Z = Toggle Size Units    S = Set Bootable   | = Expert m.");
-    mvprintw(18, 0, "T = Change Type       U = Undo All Changes     Q = Finish");
+    mvprintw(18, 0, "T = Change Type       U = Undo All Changes");
+
     if (!RunningAsInit)
-	mvprintw(18, 47, "W = Write Changes");
+	mvprintw(18, 47, "W = Write Changes  Q = Finish");
+    else
+	mvprintw(18, 47, "Q = Finish");
     mvprintw(21, 0, "Use F1 or ? to get more help, arrow keys to select.");
     move(0, 0);
 }
diff --git a/usr.sbin/sysinstall/menus.c b/usr.sbin/sysinstall/menus.c
index 11510368dce..65cab4cdaf4 100644
--- a/usr.sbin/sysinstall/menus.c
+++ b/usr.sbin/sysinstall/menus.c
@@ -1175,20 +1175,16 @@ DMenu MenuMBRType = {
     "at boot time.  If you have more than one drive and want to boot\n"
     "from the second one, the boot manager will also make it possible\n"
     "to do so (limitations in the PC BIOS usually prevent this otherwise).\n"
-    "If you will only have FreeBSD on the machine the boot manager is\n"
-    "not needed and it slows down the boot while offering you the choice\n"
-    "of which operating system to boot.  If you do not want a boot\n"
-    "manager, or wish to replace an existing one, select \"standard\".\n"
-    "If you would prefer your Master Boot Record remain untouched then\n"
-    "select \"None\".\n\n"
-    "  NOTE:  PC-DOS users will almost certainly require \"None\"!",
-    "Press F1 to read about drive setup",
+    "If you have other operating systems installed and would like a choice when\n"
+    "booting, choose \"BootMgr\". If you would prefer to keep your existing\n"
+    "boot manager, select \"None\".\n\n",
+    "",    
     "drives",
-    { { "Standard",	"Install a standard MBR (no boot manager)",
+    { { "Standard",	"Install a standard MBR (non-interactive boot manager)",
 	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 1 },
       { "BootMgr",	"Install the FreeBSD Boot Manager",
 	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 0 },
-      { "None",		"Leave the Master Boot Record untouched",
+      { "None",		"Do not install a boot manager",
 	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 2 },
       { NULL } },
 };

From 4e64643386fa2011f313d4f0c7f85ffb4d3d3b29 Mon Sep 17 00:00:00 2001
From: Rebecca Cran <brucec@FreeBSD.org>
Date: Wed, 21 Apr 2010 19:03:34 +0000
Subject: [PATCH 186/532] Remove the reference to DD mode, and replace with
 Expert mode. Also, make sure the "Q = Finish" text is visible. Reword the
 boot manager screen to try and avoid confusion, and make the order of the
 menu items match that in sysinstall.

PR:		bin/142916
Submitted by:	Jeremy Chadwick <freebsd at jdc.parodius.com>
Reviewed by:	randi
Approved by:	rrs (mentor)
MFC after:	1 week
---
 usr.sbin/sade/disks.c |  7 +++----
 usr.sbin/sade/menus.c | 23 +++++++++++------------
 2 files changed, 14 insertions(+), 16 deletions(-)

diff --git a/usr.sbin/sade/disks.c b/usr.sbin/sade/disks.c
index b802e124dd6..3158ff34c5c 100644
--- a/usr.sbin/sade/disks.c
+++ b/usr.sbin/sade/disks.c
@@ -202,10 +202,9 @@ static void
 print_command_summary(void)
 {
     mvprintw(14, 0, "The following commands are supported (in upper or lower case):");
-    mvprintw(16, 0, "A = Use Entire Disk   G = set Drive Geometry   C = Create Slice   F = `DD' mode");
-    mvprintw(17, 0, "D = Delete Slice      Z = Toggle Size Units    S = Set Bootable   | = Wizard m.");
-    mvprintw(18, 0, "T = Change Type       U = Undo All Changes     Q = Finish");
-    mvprintw(18, 47, "W = Write Changes");
+    mvprintw(16, 0, "A = Use Entire Disk   G = set Drive Geometry   C = Create Slice");
+    mvprintw(17, 0, "D = Delete Slice      Z = Toggle Size Units    S = Set Bootable   | = Expert m.");
+    mvprintw(18, 0, "T = Change Type       U = Undo All Changes     W = Write Changes  Q = Finish");
     mvprintw(21, 0, "Use F1 or ? to get more help, arrow keys to select.");
     move(0, 0);
 }
diff --git a/usr.sbin/sade/menus.c b/usr.sbin/sade/menus.c
index f48c772d88e..df518046561 100644
--- a/usr.sbin/sade/menus.c
+++ b/usr.sbin/sade/menus.c
@@ -92,24 +92,23 @@ DMenu MenuIPLType = {
 DMenu MenuMBRType = {
     DMENU_NORMAL_TYPE | DMENU_SELECTION_RETURNS,
     "overwrite me",		/* will be disk specific label */
-    "FreeBSD comes with a boot selector that allows you to easily\n"
+    "FreeBSD comes with a boot manager that allows you to easily\n"
     "select between FreeBSD and any other operating systems on your machine\n"
     "at boot time.  If you have more than one drive and want to boot\n"
-    "from the second one, the boot selector will also make it possible\n"
+    "from the second one, the boot manager will also make it possible\n"
     "to do so (limitations in the PC BIOS usually prevent this otherwise).\n"
-    "If you do not want a boot selector, or wish to replace an existing\n"
-    "one, select \"standard\".  If you would prefer your Master Boot\n"
-    "Record to remain untouched then select \"None\".\n\n"
-    "  NOTE:  PC-DOS users will almost certainly require \"None\"!",
-    "Press F1 to read about drive setup",
+    "If you have other operating systems installed and would like a choice when\n"
+    "booting, choose \"BootMgr\". If you would prefer to keep your existing\n"
+    "boot manager, select \"None\".\n\n",
+    "",
     "drives",
-    { { "BootMgr",	"Install the FreeBSD Boot Manager",
-	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, 0, 0, 0, 0 },
-      { "Standard",	"Install a standard MBR (no boot manager)",
+    { { "Standard",	"Install a standard MBR (non-interactive boot manager)",
 	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 1 },
-      { "None",		"Leave the Master Boot Record untouched",
+      { "BootMgr",	"Install the FreeBSD boot manager",
+	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 0 },
+      { "None",		"Do not install a boot manager",
 	dmenuRadioCheck, dmenuSetValue, NULL, &BootMgr, '(', '*', ')', 2 },
-      { NULL, NULL, NULL, NULL, NULL, NULL, 0, 0, 0, 0 } },
+      { NULL, NULL, NULL, NULL, NULL, NULL, 0, 0, 0, 0 } }
 };
 #endif /* PC98 */
 #endif /* __i386__ */

From 9847e91b0bc53e8f6fa1736f1a82b0de0e335141 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 19:28:01 +0000
Subject: [PATCH 187/532] Extract the code to copy-out struct rusage32 from
 struct rusage into the new function.

Reviewed by:	jhb
MFC after:	1 week
---
 sys/compat/freebsd32/freebsd32_misc.c | 56 ++++++++++++---------------
 sys/compat/freebsd32/freebsd32_util.h |  5 ++-
 2 files changed, 28 insertions(+), 33 deletions(-)

diff --git a/sys/compat/freebsd32/freebsd32_misc.c b/sys/compat/freebsd32/freebsd32_misc.c
index 7cc27be5ff3..f0fde2b3c14 100644
--- a/sys/compat/freebsd32/freebsd32_misc.c
+++ b/sys/compat/freebsd32/freebsd32_misc.c
@@ -129,6 +129,28 @@ static int freebsd32_kevent_copyin(void *arg, struct kevent *kevp, int count);
 #define RETVAL_LO 0	
 #endif
 
+void
+freebsd32_rusage_out(const struct rusage *s, struct rusage32 *s32)
+{
+
+	TV_CP(*s, *s32, ru_utime);
+	TV_CP(*s, *s32, ru_stime);
+	CP(*s, *s32, ru_maxrss);
+	CP(*s, *s32, ru_ixrss);
+	CP(*s, *s32, ru_idrss);
+	CP(*s, *s32, ru_isrss);
+	CP(*s, *s32, ru_minflt);
+	CP(*s, *s32, ru_majflt);
+	CP(*s, *s32, ru_nswap);
+	CP(*s, *s32, ru_inblock);
+	CP(*s, *s32, ru_oublock);
+	CP(*s, *s32, ru_msgsnd);
+	CP(*s, *s32, ru_msgrcv);
+	CP(*s, *s32, ru_nsignals);
+	CP(*s, *s32, ru_nvcsw);
+	CP(*s, *s32, ru_nivcsw);
+}
+
 int
 freebsd32_wait4(struct thread *td, struct freebsd32_wait4_args *uap)
 {
@@ -146,22 +168,7 @@ freebsd32_wait4(struct thread *td, struct freebsd32_wait4_args *uap)
 	if (uap->status != NULL)
 		error = copyout(&status, uap->status, sizeof(status));
 	if (uap->rusage != NULL && error == 0) {
-		TV_CP(ru, ru32, ru_utime);
-		TV_CP(ru, ru32, ru_stime);
-		CP(ru, ru32, ru_maxrss);
-		CP(ru, ru32, ru_ixrss);
-		CP(ru, ru32, ru_idrss);
-		CP(ru, ru32, ru_isrss);
-		CP(ru, ru32, ru_minflt);
-		CP(ru, ru32, ru_majflt);
-		CP(ru, ru32, ru_nswap);
-		CP(ru, ru32, ru_inblock);
-		CP(ru, ru32, ru_oublock);
-		CP(ru, ru32, ru_msgsnd);
-		CP(ru, ru32, ru_msgrcv);
-		CP(ru, ru32, ru_nsignals);
-		CP(ru, ru32, ru_nvcsw);
-		CP(ru, ru32, ru_nivcsw);
+		freebsd32_rusage_out(&ru, &ru32);
 		error = copyout(&ru32, uap->rusage, sizeof(ru32));
 	}
 	return (error);
@@ -755,22 +762,7 @@ freebsd32_getrusage(struct thread *td, struct freebsd32_getrusage_args *uap)
 	if (error)
 		return (error);
 	if (uap->rusage != NULL) {
-		TV_CP(s, s32, ru_utime);
-		TV_CP(s, s32, ru_stime);
-		CP(s, s32, ru_maxrss);
-		CP(s, s32, ru_ixrss);
-		CP(s, s32, ru_idrss);
-		CP(s, s32, ru_isrss);
-		CP(s, s32, ru_minflt);
-		CP(s, s32, ru_majflt);
-		CP(s, s32, ru_nswap);
-		CP(s, s32, ru_inblock);
-		CP(s, s32, ru_oublock);
-		CP(s, s32, ru_msgsnd);
-		CP(s, s32, ru_msgrcv);
-		CP(s, s32, ru_nsignals);
-		CP(s, s32, ru_nvcsw);
-		CP(s, s32, ru_nivcsw);
+		freebsd32_rusage_out(&s, &s32);
 		error = copyout(&s32, uap->rusage, sizeof(s32));
 	}
 	return (error);
diff --git a/sys/compat/freebsd32/freebsd32_util.h b/sys/compat/freebsd32/freebsd32_util.h
index f5ccbf16e04..3ac676d9902 100644
--- a/sys/compat/freebsd32/freebsd32_util.h
+++ b/sys/compat/freebsd32/freebsd32_util.h
@@ -94,9 +94,12 @@ int    syscall32_module_handler(struct module *mod, int what, void *arg);
 int    syscall32_helper_register(struct syscall_helper_data *sd);
 int    syscall32_helper_unregister(struct syscall_helper_data *sd);
 
-register_t *freebsd32_copyout_strings(struct image_params *imgp);
 struct iovec32;
+struct rusage32;
+register_t *freebsd32_copyout_strings(struct image_params *imgp);
 int	freebsd32_copyiniov(struct iovec32 *iovp, u_int iovcnt,
 	    struct iovec **iov, int error);
+void	freebsd32_rusage_out(const struct rusage *s, struct rusage32 *s32);
+
 
 #endif /* !_COMPAT_FREEBSD32_FREEBSD32_UTIL_H_ */

From 05e06d1157d93d2722d680e7017f34a1c568d985 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 19:32:00 +0000
Subject: [PATCH 188/532] Provide compat32 shims for kinfo_proc sysctl. This
 allows 32bit ps(1) to mostly work on 64bit host.

The work is based on an original patch submitted by emaste, obtained
from Sandvine's source tree.

Reviewed by:	jhb
MFC after:	1 week
---
 sys/compat/freebsd32/freebsd32.h |  82 +++++++++++++++++++
 sys/kern/kern_proc.c             | 134 ++++++++++++++++++++++++++++++-
 2 files changed, 212 insertions(+), 4 deletions(-)

diff --git a/sys/compat/freebsd32/freebsd32.h b/sys/compat/freebsd32/freebsd32.h
index e74da64ecf1..aad550e16b5 100644
--- a/sys/compat/freebsd32/freebsd32.h
+++ b/sys/compat/freebsd32/freebsd32.h
@@ -31,6 +31,7 @@
 
 #include <sys/procfs.h>
 #include <sys/socket.h>
+#include <sys/user.h>
 
 #define PTRIN(v)	(void *)(uintptr_t) (v)
 #define PTROUT(v)	(u_int32_t)(uintptr_t) (v)
@@ -229,4 +230,85 @@ struct mq_attr32 {
 	int	__reserved[4];
 };
 
+struct kinfo_proc32 {
+	int	ki_structsize;
+	int	ki_layout;
+	uint32_t ki_args;
+	uint32_t ki_paddr;
+	uint32_t ki_addr;
+	uint32_t ki_tracep;
+	uint32_t ki_textvp;
+	uint32_t ki_fd;
+	uint32_t ki_vmspace;
+	uint32_t ki_wchan;
+	pid_t	ki_pid;
+	pid_t	ki_ppid;
+	pid_t	ki_pgid;
+	pid_t	ki_tpgid;
+	pid_t	ki_sid;
+	pid_t	ki_tsid;
+	short	ki_jobc;
+	short	ki_spare_short1;
+	dev_t	ki_tdev;
+	sigset_t ki_siglist;
+	sigset_t ki_sigmask;
+	sigset_t ki_sigignore;
+	sigset_t ki_sigcatch;
+	uid_t	ki_uid;
+	uid_t	ki_ruid;
+	uid_t	ki_svuid;
+	gid_t	ki_rgid;
+	gid_t	ki_svgid;
+	short	ki_ngroups;
+	short	ki_spare_short2;
+	gid_t 	ki_groups[KI_NGROUPS];
+	uint32_t ki_size;
+	int32_t ki_rssize;
+	int32_t ki_swrss;
+	int32_t ki_tsize;
+	int32_t ki_dsize;
+	int32_t ki_ssize;
+	u_short	ki_xstat;
+	u_short	ki_acflag;
+	fixpt_t	ki_pctcpu;
+	u_int	ki_estcpu;
+	u_int	ki_slptime;
+	u_int	ki_swtime;
+	int	ki_spareint1;
+	u_int64_t ki_runtime;
+	struct	timeval32 ki_start;
+	struct	timeval32 ki_childtime;
+	int	ki_flag;
+	int	ki_kiflag;
+	int	ki_traceflag;
+	char	ki_stat;
+	signed char ki_nice;
+	char	ki_lock;
+	char	ki_rqindex;
+	u_char	ki_oncpu;
+	u_char	ki_lastcpu;
+	char	ki_ocomm[OCOMMLEN+1];
+	char	ki_wmesg[WMESGLEN+1];
+	char	ki_login[LOGNAMELEN+1];
+	char	ki_lockname[LOCKNAMELEN+1];
+	char	ki_comm[COMMLEN+1];
+	char	ki_emul[KI_EMULNAMELEN+1];
+	char	ki_sparestrings[68];
+	int	ki_spareints[KI_NSPARE_INT];
+	u_int	ki_cr_flags;
+	int	ki_jid;
+	int	ki_numthreads;
+	lwpid_t	ki_tid;
+	struct	priority ki_pri;
+	struct	rusage32 ki_rusage;
+	struct	rusage32 ki_rusage_ch;
+	uint32_t ki_pcb;
+	uint32_t ki_kstack;
+	uint32_t ki_udata;
+	uint32_t ki_spareptrs[KI_NSPARE_PTR];	/* spare room for growth */
+	int	ki_sparelongs[KI_NSPARE_LONG];
+	int	ki_sflag;
+	int	ki_tdflags;
+};
+
 #endif /* !_COMPAT_FREEBSD32_FREEBSD32_H_ */
diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index c41909d9dc5..c5a42a1bf99 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -42,6 +42,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/kernel.h>
+#include <sys/limits.h>
 #include <sys/lock.h>
 #include <sys/malloc.h>
 #include <sys/mount.h>
@@ -79,6 +80,11 @@ __FBSDID("$FreeBSD$");
 #include <vm/vm_object.h>
 #include <vm/uma.h>
 
+#ifdef COMPAT_FREEBSD32
+#include <compat/freebsd32/freebsd32.h>
+#include <compat/freebsd32/freebsd32_util.h>
+#endif
+
 SDT_PROVIDER_DEFINE(proc);
 SDT_PROBE_DEFINE(proc, kernel, ctor, entry);
 SDT_PROBE_ARGTYPE(proc, kernel, ctor, entry, 0, "struct proc *");
@@ -969,6 +975,128 @@ zpfind(pid_t pid)
 #define KERN_PROC_ZOMBMASK	0x3
 #define KERN_PROC_NOTHREADS	0x4
 
+#ifdef COMPAT_FREEBSD32
+
+/*
+ * This function is typically used to copy out the kernel address, so
+ * it can be replaced by assignment of zero.
+ */
+static inline uint32_t
+ptr32_trim(void *ptr)
+{
+	uintptr_t uptr;
+
+	uptr = (uintptr_t)ptr;
+	return ((uptr > UINT_MAX) ? 0 : uptr);
+}
+
+#define PTRTRIM_CP(src,dst,fld) \
+	do { (dst).fld = ptr32_trim((src).fld); } while (0)
+
+static void
+freebsd32_kinfo_proc_out(const struct kinfo_proc *ki, struct kinfo_proc32 *ki32)
+{
+	int i;
+
+	bzero(ki32, sizeof(struct kinfo_proc32));
+	ki32->ki_structsize = sizeof(struct kinfo_proc32);
+	CP(*ki, *ki32, ki_layout);
+	PTRTRIM_CP(*ki, *ki32, ki_args);
+	PTRTRIM_CP(*ki, *ki32, ki_paddr);
+	PTRTRIM_CP(*ki, *ki32, ki_addr);
+	PTRTRIM_CP(*ki, *ki32, ki_tracep);
+	PTRTRIM_CP(*ki, *ki32, ki_textvp);
+	PTRTRIM_CP(*ki, *ki32, ki_fd);
+	PTRTRIM_CP(*ki, *ki32, ki_vmspace);
+	PTRTRIM_CP(*ki, *ki32, ki_wchan);
+	CP(*ki, *ki32, ki_pid);
+	CP(*ki, *ki32, ki_ppid);
+	CP(*ki, *ki32, ki_pgid);
+	CP(*ki, *ki32, ki_tpgid);
+	CP(*ki, *ki32, ki_sid);
+	CP(*ki, *ki32, ki_tsid);
+	CP(*ki, *ki32, ki_jobc);
+	CP(*ki, *ki32, ki_tdev);
+	CP(*ki, *ki32, ki_siglist);
+	CP(*ki, *ki32, ki_sigmask);
+	CP(*ki, *ki32, ki_sigignore);
+	CP(*ki, *ki32, ki_sigcatch);
+	CP(*ki, *ki32, ki_uid);
+	CP(*ki, *ki32, ki_ruid);
+	CP(*ki, *ki32, ki_svuid);
+	CP(*ki, *ki32, ki_rgid);
+	CP(*ki, *ki32, ki_svgid);
+	CP(*ki, *ki32, ki_ngroups);
+	for (i = 0; i < KI_NGROUPS; i++)
+		CP(*ki, *ki32, ki_groups[i]);
+	CP(*ki, *ki32, ki_size);
+	CP(*ki, *ki32, ki_rssize);
+	CP(*ki, *ki32, ki_swrss);
+	CP(*ki, *ki32, ki_tsize);
+	CP(*ki, *ki32, ki_dsize);
+	CP(*ki, *ki32, ki_ssize);
+	CP(*ki, *ki32, ki_xstat);
+	CP(*ki, *ki32, ki_acflag);
+	CP(*ki, *ki32, ki_pctcpu);
+	CP(*ki, *ki32, ki_estcpu);
+	CP(*ki, *ki32, ki_slptime);
+	CP(*ki, *ki32, ki_swtime);
+	CP(*ki, *ki32, ki_runtime);
+	TV_CP(*ki, *ki32, ki_start);
+	TV_CP(*ki, *ki32, ki_childtime);
+	CP(*ki, *ki32, ki_flag);
+	CP(*ki, *ki32, ki_kiflag);
+	CP(*ki, *ki32, ki_traceflag);
+	CP(*ki, *ki32, ki_stat);
+	CP(*ki, *ki32, ki_nice);
+	CP(*ki, *ki32, ki_lock);
+	CP(*ki, *ki32, ki_rqindex);
+	CP(*ki, *ki32, ki_oncpu);
+	CP(*ki, *ki32, ki_lastcpu);
+	bcopy(ki->ki_ocomm, ki32->ki_ocomm, OCOMMLEN + 1);
+	bcopy(ki->ki_wmesg, ki32->ki_wmesg, WMESGLEN + 1);
+	bcopy(ki->ki_login, ki32->ki_login, LOGNAMELEN + 1);
+	bcopy(ki->ki_lockname, ki32->ki_lockname, LOCKNAMELEN + 1);
+	bcopy(ki->ki_comm, ki32->ki_comm, COMMLEN + 1);
+	bcopy(ki->ki_emul, ki32->ki_emul, KI_EMULNAMELEN + 1);
+	CP(*ki, *ki32, ki_cr_flags);
+	CP(*ki, *ki32, ki_jid);
+	CP(*ki, *ki32, ki_numthreads);
+	CP(*ki, *ki32, ki_tid);
+	CP(*ki, *ki32, ki_pri);
+	freebsd32_rusage_out(&ki->ki_rusage, &ki32->ki_rusage);
+	freebsd32_rusage_out(&ki->ki_rusage_ch, &ki32->ki_rusage_ch);
+	PTRTRIM_CP(*ki, *ki32, ki_pcb);
+	PTRTRIM_CP(*ki, *ki32, ki_kstack);
+	PTRTRIM_CP(*ki, *ki32, ki_udata);
+	CP(*ki, *ki32, ki_sflag);
+	CP(*ki, *ki32, ki_tdflags);
+}
+
+static int
+sysctl_out_proc_copyout(struct kinfo_proc *ki, struct sysctl_req *req)
+{
+	struct kinfo_proc32 ki32;
+	int error;
+
+	if (req->flags & SCTL_MASK32) {
+		freebsd32_kinfo_proc_out(ki, &ki32);
+		error = SYSCTL_OUT(req, (caddr_t)&ki32,
+		    sizeof(struct kinfo_proc32));
+	} else
+		error = SYSCTL_OUT(req, (caddr_t)ki,
+		    sizeof(struct kinfo_proc));
+	return (error);
+}
+#else
+static int
+sysctl_out_proc_copyout(struct kinfo_proc *ki, struct sysctl_req *req)
+{
+
+	return (SYSCTL_OUT(req, (caddr_t)&ki, sizeof(struct kinfo_proc)));
+}
+#endif
+
 /*
  * Must be called with the process locked and will return with it unlocked.
  */
@@ -986,13 +1114,11 @@ sysctl_out_proc(struct proc *p, struct sysctl_req *req, int flags)
 
 	fill_kinfo_proc(p, &kinfo_proc);
 	if (flags & KERN_PROC_NOTHREADS)
-		error = SYSCTL_OUT(req, (caddr_t)&kinfo_proc,
-		    sizeof(kinfo_proc));
+		error = sysctl_out_proc_copyout(&kinfo_proc, req);
 	else {
 		FOREACH_THREAD_IN_PROC(p, td) {
 			fill_kinfo_thread(td, &kinfo_proc, 1);
-			error = SYSCTL_OUT(req, (caddr_t)&kinfo_proc,
-			    sizeof(kinfo_proc));
+			error = sysctl_out_proc_copyout(&kinfo_proc, req);
 			if (error)
 				break;
 		}

From 431c9bf2d5ca4e10a4b744e51dfe3145f5fe7b6e Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 19:35:56 +0000
Subject: [PATCH 189/532] C language does not has references, it provides
 pointers.

Suggested by:	bde
MFC after:	2 weeks
---
 lib/libc/stdlib/realpath.3 | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/libc/stdlib/realpath.3 b/lib/libc/stdlib/realpath.3
index aee65f7600f..166bb129c24 100644
--- a/lib/libc/stdlib/realpath.3
+++ b/lib/libc/stdlib/realpath.3
@@ -56,13 +56,13 @@ and
 in
 .Fa pathname ,
 and copies the resulting absolute pathname into
-the memory referenced by
+the memory pointed to by
 .Fa resolved_path .
 The
 .Fa resolved_path
 argument
 .Em must
-refer to a buffer capable of storing at least
+point to a buffer capable of storing at least
 .Dv PATH_MAX
 characters, or be
 .Dv NULL .

From b4bf2ac1ac93250a084d8de15b48cf351c1bd4c1 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 21 Apr 2010 20:04:42 +0000
Subject: [PATCH 190/532] Fix typo.

Submitted by:	emaste
Pointy hat to:	kib (who needs much bigger wardrobe)
MFC after:	1 week
---
 sys/kern/kern_proc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index c5a42a1bf99..d4943c09d8f 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -1093,7 +1093,7 @@ static int
 sysctl_out_proc_copyout(struct kinfo_proc *ki, struct sysctl_req *req)
 {
 
-	return (SYSCTL_OUT(req, (caddr_t)&ki, sizeof(struct kinfo_proc)));
+	return (SYSCTL_OUT(req, (caddr_t)ki, sizeof(struct kinfo_proc)));
 }
 #endif
 

From 64451dd71ddcce10f7e28dc3d4a93297ee6c01ee Mon Sep 17 00:00:00 2001
From: Ana Kukec <anchie@FreeBSD.org>
Date: Wed, 21 Apr 2010 20:05:49 +0000
Subject: [PATCH 191/532] Fix of r206904, \nanchie is needed instead of
 \anchie.

Approved by:	bz (mentor)
Reported by:	danfe, zec
---
 share/misc/committers-src.dot | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/share/misc/committers-src.dot b/share/misc/committers-src.dot
index 144d08ab68e..ebecb351bd8 100644
--- a/share/misc/committers-src.dot
+++ b/share/misc/committers-src.dot
@@ -55,7 +55,7 @@ node [color=lightblue2, style=filled, bgcolor=black];
 ache [label="Andrey Chernov\nache@FreeBSD.org\n1993/10/31"]
 akiyama [label="Shunsuke Akiyama\nakiyama@FreeBSD.org\n2000/06/19"]
 ambrisko [label="Doug Ambrisko\nambrisko@FreeBSD.org\n2001/12/19"]
-anchie [label="Ana Kukec\anchie@FreeBSD.org\n2010/04/14"]
+anchie [label="Ana Kukec\nanchie@FreeBSD.org\n2010/04/14"]
 andre [label="Andre Oppermann\nandre@FreeBSD.org\n2003/11/12"]
 anholt [label="Eric Anholt\nanholt@FreeBSD.org\n2002/04/22"]
 antoine [label="Antoine Brodin\nantoine@FreeBSD.org\n2008/02/03"]

From 3acd904d851f93d821cab1e367ac11ac566d9995 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Wed, 21 Apr 2010 21:51:14 +0000
Subject: [PATCH 192/532] Change usb devd events from fake attach to a notify.
 The ugen device is not a proper device_t so it faked the devctl event to
 appear like one, this is now a notify which allows more information to be
 passed.

We notify for both the device attach/detach and for each usb interface. A devd
rule can now match on the interface properties, including composite devices
which may have a uvideo interface and also usound and possibly uhid too.

An example to match a umass device with a scsi subclass and BBB protocol would be

notify 100 {
	match "system"          "USB";
	match "subsystem"       "INTERFACE";
	match "type"            "ATTACH";
	match "intclass"        "0x08";
	match "intsubclass"     "0x06";
	match "intprotocol"     "0x50";
	action ...
};

The old attach devctl event has been retained for the moment to make merging to
8.1 easier. This was never compatible with 7.x or earlier due to the ugen regex
change needed.

Reviewed by:	warner
MFC after:	1 week
---
 etc/devd.conf            |  14 ++--
 etc/devd/uath.conf       | 156 +++++++++++++++++++++++----------------
 sys/dev/usb/usb_device.c |  98 +++++++++++++++++++++++-
 3 files changed, 193 insertions(+), 75 deletions(-)

diff --git a/etc/devd.conf b/etc/devd.conf
index 696ac7c5f1a..8827e3b4e03 100644
--- a/etc/devd.conf
+++ b/etc/devd.conf
@@ -138,12 +138,14 @@ attach 100 {
 # This entry starts the ColdSync tool in daemon mode. Make sure you have an up
 # to date /usr/local/etc/palms. We override the 'listen' settings for port and
 # type in /usr/local/etc/coldsync.conf.
-attach 100 {
-	device-name "ugen[0-9]+";
-	match "vendor" "0x082d";
-	match "product" "0x0100";
-	match "release" "0x0100";
-	action "/usr/local/bin/coldsync -md -p /dev/$device-name -t usb";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x082d";
+	match "product"		"0x0100";
+	match "release"		"0x0100";
+	action "/usr/local/bin/coldsync -md -p /dev/$cdev -t usb";
 };
 
 #
diff --git a/etc/devd/uath.conf b/etc/devd/uath.conf
index 33bee69bb17..dc4019cefbc 100644
--- a/etc/devd/uath.conf
+++ b/etc/devd/uath.conf
@@ -4,117 +4,143 @@
 
 # Accton
 #   SMCWUSB-G and SMCWUSBT-G2
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x083a";
-	match "product" "(0x4505|0x4507)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x083a";
+	match "product"		"(0x4505|0x4507)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Atheros Communications
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x168c";
-	match "product" "0x0002";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x168c";
+	match "product"		"0x0002";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Atheros Communications
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x0cf3";
-	match "product" "(0x0002|0x0004|0x0006)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x0cf3";
+	match "product"		"(0x0002|0x0004|0x0006)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Conceptronic
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x0d8e";
-	match "product" "(0x7802|0x7812)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x0d8e";
+	match "product"		"(0x7802|0x7812)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # D-Link
 #   DWL-AG132, DWL-G132 and DWL-AG122
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x2001";
-	match "product" "(0x3a01|0x3a03|0x3a05)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x2001";
+	match "product"		"(0x3a01|0x3a03|0x3a05)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # D-Link
 #  DWA-120
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x07d1";
-	match "product" "0x3a0c";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x07d1";
+	match "product"		"0x3a0c";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Gigaset
 #   SMCWUSBT-G
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x1690";
-	match "product" "(0x0711|0x0713)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x1690";
+	match "product"		"(0x0711|0x0713)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Global Sun Technology
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x16ab";
-	match "product" "(0x7802|0x7812)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x16ab";
+	match "product"		"(0x7802|0x7812)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # BayNETGEAR
 #   WG111U
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x0846";
-	match "product" "0x4301";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x0846";
+	match "product"		"0x4301";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Netgear
 #   WG111T and WPN111
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x1385";
-	match "product" "(0x4251|0x5f01)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x1385";
+	match "product"		"(0x4251|0x5f01)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # U-MEDIA Communications
 #   TEW-444UB and AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x157e";
-	match "product" "(0x3007|0x3206)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x157e";
+	match "product"		"(0x3007|0x3206)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Wistron NeWeb
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x1435";
-	match "product" "(0x0827|0x0829)";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x1435";
+	match "product"		"(0x0827|0x0829)";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
 
 # Z-Com
 #   AR5523
-attach 100 {
-	device-name "ugen[0-9.]+";
-	match "vendor" "0x0cde";
-	match "product" "0x0013";
-	action "/usr/sbin/uathload -d /dev/$device-name";
+notify 100 {
+	match "system"		"USB";
+	match "subsystem"	"DEVICE";
+	match "type"		"ATTACH";
+	match "vendor"		"0x0cde";
+	match "product"		"0x0013";
+	action "/usr/sbin/uathload -d /dev/$cdev";
 };
diff --git a/sys/dev/usb/usb_device.c b/sys/dev/usb/usb_device.c
index d7f0c31c8d3..f0106ec9eec 100644
--- a/sys/dev/usb/usb_device.c
+++ b/sys/dev/usb/usb_device.c
@@ -45,6 +45,7 @@
 #include <sys/priv.h>
 #include <sys/conf.h>
 #include <sys/fcntl.h>
+#include <sys/sbuf.h>
 
 #include <dev/usb/usb.h>
 #include <dev/usb/usbdi.h>
@@ -1834,7 +1835,7 @@ config_done:
 	printf("%s: <%s> at %s\n", udev->ugen_name, udev->manufacturer,
 	    device_get_nameunit(udev->bus->bdev));
 
-	usb_notify_addq("+", udev);
+	usb_notify_addq("ATTACH", udev);
 #endif
 done:
 	if (err) {
@@ -1980,7 +1981,7 @@ usb_free_device(struct usb_device *udev, uint8_t flag)
 	usb_set_device_state(udev, USB_STATE_DETACHED);
 
 #if USB_HAVE_UGEN
-	usb_notify_addq("-", udev);
+	usb_notify_addq("DETACH", udev);
 
 	printf("%s: <%s> at %s (disconnected)\n", udev->ugen_name,
 	    udev->manufacturer, device_get_nameunit(bus->bdev));
@@ -2347,13 +2348,23 @@ usbd_get_device_index(struct usb_device *udev)
  *
  * This function will generate events for dev.
  *------------------------------------------------------------------------*/
+#ifndef BURN_BRIDGES
 static void
-usb_notify_addq(const char *type, struct usb_device *udev)
+usb_notify_addq_compat(const char *type, struct usb_device *udev)
 {
 	char *data = NULL;
+	const char *ntype;
 	struct malloc_type *mt;
 	const size_t buf_size = 512;
 
+	/* Convert notify type */
+	if (strcmp(type, "ATTACH") == 0)
+		ntype = "+";
+	else if (strcmp(type, "DETACH") == 0)
+		ntype = "-";
+	else
+		return;
+
 	mtx_lock(&malloc_mtx);
 	mt = malloc_desc2type("bus");	/* XXX M_BUS */
 	mtx_unlock(&malloc_mtx);
@@ -2378,7 +2389,7 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 	    "port=%u "
 	    "on "
 	    "%s\n",
-	    type,
+	    ntype,
 	    udev->ugen_name,
 	    UGETW(udev->ddesc.idVendor),
 	    UGETW(udev->ddesc.idProduct),
@@ -2393,6 +2404,85 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 
 	devctl_queue_data(data);
 }
+#endif
+
+static void
+usb_notify_addq(const char *type, struct usb_device *udev)
+{
+	struct usb_interface *iface;
+	struct sbuf *sb;
+	int i;
+
+#ifndef BURN_BRIDGES
+	usb_notify_addq_compat(type, udev);
+#endif
+
+	/* announce the device */
+	sb = sbuf_new_auto();
+	sbuf_printf(sb,
+	    "cdev=%s "
+	    "vendor=0x%04x "
+	    "product=0x%04x "
+	    "devclass=0x%02x "
+	    "devsubclass=0x%02x "
+	    "sernum=\"%s\" "
+	    "release=0x%04x "
+	    "port=%u "
+	    "parent=%s\n",
+	    udev->ugen_name,
+	    UGETW(udev->ddesc.idVendor),
+	    UGETW(udev->ddesc.idProduct),
+	    udev->ddesc.bDeviceClass,
+	    udev->ddesc.bDeviceSubClass,
+	    udev->serial,
+	    UGETW(udev->ddesc.bcdDevice),
+	    udev->port_no,
+	    udev->parent_hub != NULL ?
+	    udev->parent_hub->ugen_name :
+	    device_get_nameunit(device_get_parent(udev->bus->bdev)));
+	sbuf_finish(sb);
+	devctl_notify("USB", "DEVICE", type, sbuf_data(sb));
+	sbuf_delete(sb);
+
+	/* announce each interface */
+	for (i = 0; i < USB_IFACE_MAX; i++) {
+		iface = usbd_get_iface(udev, i);
+		if (iface == NULL)
+			break;		/* end of interfaces */
+		if (iface->idesc == NULL)
+			continue;	/* no interface descriptor */
+
+		sb = sbuf_new_auto();
+		sbuf_printf(sb,
+		    "cdev=%s "
+		    "vendor=0x%04x "
+		    "product=0x%04x "
+		    "devclass=0x%02x "
+		    "devsubclass=0x%02x "
+		    "sernum=\"%s\" "
+		    "release=0x%04x "
+		    "interface=%d "
+		    "endpoints=%d "
+		    "intclass=0x%02x "
+		    "intsubclass=0x%02x "
+		    "intprotocol=0x%02x\n",
+		    udev->ugen_name,
+		    UGETW(udev->ddesc.idVendor),
+		    UGETW(udev->ddesc.idProduct),
+		    udev->ddesc.bDeviceClass,
+		    udev->ddesc.bDeviceSubClass,
+		    udev->serial,
+		    UGETW(udev->ddesc.bcdDevice),
+		    iface->idesc->bInterfaceNumber,
+		    iface->idesc->bNumEndpoints,
+		    iface->idesc->bInterfaceClass,
+		    iface->idesc->bInterfaceSubClass,
+		    iface->idesc->bInterfaceProtocol);
+		sbuf_finish(sb);
+		devctl_notify("USB", "INTERFACE", type, sbuf_data(sb));
+		sbuf_delete(sb);
+	}
+}
 
 /*------------------------------------------------------------------------*
  *	usb_fifo_free_wrap

From dc00aa5dd6bd72b9152d634540a6194b267be977 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Wed, 21 Apr 2010 21:57:03 +0000
Subject: [PATCH 193/532] ln: Allow a trailing slash when creating a link to a
 directory.

In the 'ln source... directory' synopsis, the basename of each source
determines the name of the created link. Determine this using basename(3)
instead of strrchr(..., '/') which is incorrect if the pathname ends in a
slash.

The patch is somewhat changed to allow for basename(3) implementations that
change the passed pathname, and to fix the -w option's checking also.
The code to compare directory entries only applies to hard links, which
cannot be created to directories using ln.

Example:
  ln -s /etc/defaults/ /tmp
This should create a symlink named defaults.

PR:		121568
Submitted by:	Ighighi
MFC after:	1 week
---
 bin/ln/ln.c | 27 +++++++++++++--------------
 1 file changed, 13 insertions(+), 14 deletions(-)

diff --git a/bin/ln/ln.c b/bin/ln/ln.c
index d77939291f7..fc3afc13308 100644
--- a/bin/ln/ln.c
+++ b/bin/ln/ln.c
@@ -47,6 +47,7 @@ __FBSDID("$FreeBSD$");
 #include <err.h>
 #include <errno.h>
 #include <fcntl.h>
+#include <libgen.h>
 #include <limits.h>
 #include <stdio.h>
 #include <stdlib.h>
@@ -226,6 +227,7 @@ linkit(const char *source, const char *target, int isdir)
 	int ch, exists, first;
 	char path[PATH_MAX];
 	char wbuf[PATH_MAX];
+	char bbuf[PATH_MAX];
 
 	if (!sflag) {
 		/* If source doesn't exist, quit now. */
@@ -248,11 +250,9 @@ linkit(const char *source, const char *target, int isdir)
 	if (isdir ||
 	    (lstat(target, &sb) == 0 && S_ISDIR(sb.st_mode)) ||
 	    (!hflag && stat(target, &sb) == 0 && S_ISDIR(sb.st_mode))) {
-		if ((p = strrchr(source, '/')) == NULL)
-			p = source;
-		else
-			++p;
-		if (snprintf(path, sizeof(path), "%s/%s", target, p) >=
+		if (strlcpy(bbuf, source, sizeof(bbuf)) >= sizeof(bbuf) ||
+		    (p = basename(bbuf)) == NULL ||
+		    snprintf(path, sizeof(path), "%s/%s", target, p) >=
 		    (ssize_t)sizeof(path)) {
 			errno = ENAMETOOLONG;
 			warn("%s", source);
@@ -276,15 +276,14 @@ linkit(const char *source, const char *target, int isdir)
 			 * absolute path of the source, by appending `source'
 			 * to the parent directory of the target.
 			 */
-			p = strrchr(target, '/');
-			if (p != NULL)
-				p++;
-			else
-				p = target;
-			(void)snprintf(wbuf, sizeof(wbuf), "%.*s%s",
-			    (int)(p - target), target, source);
-			if (stat(wbuf, &sb) != 0)
-				warn("warning: %s", source);
+			strlcpy(bbuf, target, sizeof(bbuf));
+			p = dirname(bbuf);
+			if (p != NULL) {
+				(void)snprintf(wbuf, sizeof(wbuf), "%s/%s",
+						p, source);
+				if (stat(wbuf, &sb) != 0)
+					warn("warning: %s", source);
+			}
 		}
 	}
 

From 6639021699a01a2f2e5aa1c9e91917175551676e Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Wed, 21 Apr 2010 23:03:26 +0000
Subject: [PATCH 194/532] Also add the usb mode to the devd string as the usb
 controller can work in both host or device (gadget) modes.

Suggested by:	HPS
---
 sys/dev/usb/usb_device.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sys/dev/usb/usb_device.c b/sys/dev/usb/usb_device.c
index f0106ec9eec..2f3a402d4ac 100644
--- a/sys/dev/usb/usb_device.c
+++ b/sys/dev/usb/usb_device.c
@@ -2427,6 +2427,7 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 	    "devsubclass=0x%02x "
 	    "sernum=\"%s\" "
 	    "release=0x%04x "
+	    "mode=%s "
 	    "port=%u "
 	    "parent=%s\n",
 	    udev->ugen_name,
@@ -2436,6 +2437,7 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 	    udev->ddesc.bDeviceSubClass,
 	    udev->serial,
 	    UGETW(udev->ddesc.bcdDevice),
+	    (udev->flags.usb_mode == USB_MODE_HOST) ? "host" : "device",
 	    udev->port_no,
 	    udev->parent_hub != NULL ?
 	    udev->parent_hub->ugen_name :
@@ -2461,6 +2463,7 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 		    "devsubclass=0x%02x "
 		    "sernum=\"%s\" "
 		    "release=0x%04x "
+		    "mode=%s "
 		    "interface=%d "
 		    "endpoints=%d "
 		    "intclass=0x%02x "
@@ -2473,6 +2476,7 @@ usb_notify_addq(const char *type, struct usb_device *udev)
 		    udev->ddesc.bDeviceSubClass,
 		    udev->serial,
 		    UGETW(udev->ddesc.bcdDevice),
+		    (udev->flags.usb_mode == USB_MODE_HOST) ? "host" : "device",
 		    iface->idesc->bInterfaceNumber,
 		    iface->idesc->bNumEndpoints,
 		    iface->idesc->bInterfaceClass,

From 7ab4cc53fd3e51f8336a23daf073191f7b1d024e Mon Sep 17 00:00:00 2001
From: Alexander Leidinger <netchild@FreeBSD.org>
Date: Thu, 22 Apr 2010 09:20:17 +0000
Subject: [PATCH 195/532] Revert r206179 (by imp) and do something similar
 which is more consistent with all other corresponding CTF places by changing
 the corresponding code which is generated by config(8). Or in short, move the
 '@' from the variable definition to the use of the variable. [1]

While I'm here break up a long line. [2]

Discussed with:	imp [1,2], bde [2]
---
 sys/conf/kern.pre.mk         | 3 ++-
 usr.sbin/config/mkmakefile.c | 2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)

diff --git a/sys/conf/kern.pre.mk b/sys/conf/kern.pre.mk
index d55b3797a34..b70f5f164f1 100644
--- a/sys/conf/kern.pre.mk
+++ b/sys/conf/kern.pre.mk
@@ -128,7 +128,8 @@ NORMAL_C_NOWERROR= ${CC} -c ${CFLAGS} ${PROF} ${.IMPSRC}
 NORMAL_M= ${AWK} -f $S/tools/makeobjops.awk ${.IMPSRC} -c ; \
 	  ${CC} -c ${CFLAGS} ${WERROR} ${PROF} ${.PREFIX}.c
 
-NORMAL_CTFCONVERT= @[ -z "${CTFCONVERT}" -o -n "${NO_CTF}" ] || ${CTFCONVERT} ${CTFFLAGS} ${.TARGET}
+NORMAL_CTFCONVERT= [ -z "${CTFCONVERT}" -o -n "${NO_CTF}" ] || \
+		   ${CTFCONVERT} ${CTFFLAGS} ${.TARGET}
 
 NORMAL_LINT=	${LINT} ${LINTFLAGS} ${CFLAGS:M-[DIU]*} ${.IMPSRC}
 
diff --git a/usr.sbin/config/mkmakefile.c b/usr.sbin/config/mkmakefile.c
index f0d37faadc8..7845089cf55 100644
--- a/usr.sbin/config/mkmakefile.c
+++ b/usr.sbin/config/mkmakefile.c
@@ -741,7 +741,7 @@ do_rules(FILE *f)
 				printf("config: don't know rules for %s\n", np);
 				break;
 			}
-			snprintf(cmd, sizeof(cmd), "${%s_%c%s}\n.if defined(NORMAL_CTFCONVERT) && !empty(NORMAL_CTFCONVERT)\n\t${NORMAL_CTFCONVERT}\n.endif", ftype,
+			snprintf(cmd, sizeof(cmd), "${%s_%c%s}\n\t@${NORMAL_CTFCONVERT}", ftype,
 			    toupper(och),
 			    ftp->f_flags & NOWERROR ? "_NOWERROR" : "");
 			compilewith = cmd;

From 69745d38867c90f8d887090fa805c84f6a0e2248 Mon Sep 17 00:00:00 2001
From: Alexander Leidinger <netchild@FreeBSD.org>
Date: Thu, 22 Apr 2010 09:24:01 +0000
Subject: [PATCH 196/532] Do the ctfmerge (= all symbols, not only symbols from
 one object file) for KLDs too.

Noticed by:	np
---
 sys/conf/kmod.mk | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sys/conf/kmod.mk b/sys/conf/kmod.mk
index 526eec74bcc..18bfbb12b3b 100644
--- a/sys/conf/kmod.mk
+++ b/sys/conf/kmod.mk
@@ -137,6 +137,10 @@ CFLAGS+=	-mlongcall -fno-omit-frame-pointer
 CFLAGS+=	-G0 -fno-pic -mno-abicalls -mlong-calls
 .endif
 
+.if defined(DEBUG) || defined(DEBUG_FLAGS)
+CTFFLAGS+=	-g
+.endif
+
 .if defined(FIRMWS)
 .if !exists(@)
 ${KMOD:S/$/.c/}: @
@@ -202,6 +206,7 @@ ${KMOD}.kld: ${OBJS}
 ${FULLPROG}: ${OBJS}
 .endif
 	${LD} ${LDFLAGS} -r -d -o ${.TARGET} ${OBJS}
+	@[ -z "${CTFMERGE}" -o -n "${NO_CTF}" ] || ${CTFMERGE} ${CTFFLAGS} -o ${.TARGET} ${OBJS}
 .if defined(EXPORT_SYMS)
 .if ${EXPORT_SYMS} != YES
 .if ${EXPORT_SYMS} == NO

From 096924c9b7f56d9a16b6fde1ac77b6c739550649 Mon Sep 17 00:00:00 2001
From: Alexander Leidinger <netchild@FreeBSD.org>
Date: Thu, 22 Apr 2010 09:30:02 +0000
Subject: [PATCH 197/532] Remove explicit setting of NO_CTF in WMAKEENV and in
 the make call for the buildkernel. This way makeoptions WITH_CTF=yes not only
 works when compiling the tradtitional way, but also when using buildkernel.
 This does not enable the CTF part of the world, it still defaults to without
 CTF info.

The cross/build-tools/bootstrap targets are not affected by this, they
still have and should keep the explicit NO_CTF.

Notified by:	np
---
 Makefile.inc1 | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/Makefile.inc1 b/Makefile.inc1
index 1c5d892468b..e54ae5ca925 100644
--- a/Makefile.inc1
+++ b/Makefile.inc1
@@ -259,7 +259,7 @@ WMAKEENV=	${CROSSENV} \
 		VERSION="${VERSION}" \
 		INSTALL="sh ${.CURDIR}/tools/install.sh" \
 		PATH=${TMPPATH}
-.if ${MK_CDDL} == "no" || defined(NO_CTF)
+.if ${MK_CDDL} == "no"
 WMAKEENV+=	NO_CTF=1
 .endif
 WMAKE=		${WMAKEENV} ${MAKE} -f Makefile.inc1 DESTDIR=${WORLDTMP}
@@ -780,7 +780,7 @@ buildkernel:
 	@echo "--------------------------------------------------------------"
 	cd ${KRNLOBJDIR}/${_kernel}; \
 	    MAKESRCPATH=${KERNSRCDIR}/dev/aic7xxx/aicasm \
-	    ${MAKE} SSP_CFLAGS= -DNO_CPU_CFLAGS -DNO_CTF \
+	    ${MAKE} SSP_CFLAGS= -DNO_CPU_CFLAGS \
 	    -f ${KERNSRCDIR}/dev/aic7xxx/aicasm/Makefile
 # XXX - Gratuitously builds aicasm in the ``makeoptions NO_MODULES'' case.
 .if !defined(MODULES_WITH_WORLD) && !defined(NO_MODULES) && exists(${KERNSRCDIR}/modules)

From 9c03c0d88c108ccddde58f4703674df3aa406690 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Thu, 22 Apr 2010 18:21:25 +0000
Subject: [PATCH 198/532] Rework how drm maps are handled.

 * On 32 bit platforms we steal the upper 4 bits of the map handle
   to store a unique map id.
 * On 64 bit platforms we steal the upper 24 bits.

Resolves issues where the offsets that are handed to mmap may overlap the VRAM on some cards.

Tested on: radeon, intel, mga, and via.

This will break nouveau.  I will spin new patches shortly.
---
 sys/dev/drm/drmP.h         | 42 ++++++++++++++++++++-----------------
 sys/dev/drm/drm_bufs.c     | 37 ++++++++++++++++++--------------
 sys/dev/drm/drm_context.c  |  2 +-
 sys/dev/drm/drm_drv.c      |  7 +++++++
 sys/dev/drm/drm_memory.c   |  2 +-
 sys/dev/drm/drm_scatter.c  |  7 +++++--
 sys/dev/drm/drm_sysctl.c   | 10 +++++----
 sys/dev/drm/drm_vm.c       | 43 +++++++++++++++++++++-----------------
 sys/dev/drm/i915_dma.c     | 14 ++++++-------
 sys/dev/drm/mach64_dma.c   |  8 +++----
 sys/dev/drm/mga_dma.c      | 22 +++++++++----------
 sys/dev/drm/mga_warp.c     |  4 ++--
 sys/dev/drm/r128_cce.c     | 19 +++++++++--------
 sys/dev/drm/r128_state.c   |  4 ++--
 sys/dev/drm/r600_blit.c    | 17 +++++++--------
 sys/dev/drm/r600_cp.c      | 40 ++++++++++++++++++-----------------
 sys/dev/drm/radeon_cp.c    | 36 +++++++++++++++----------------
 sys/dev/drm/radeon_cs.c    |  2 +-
 sys/dev/drm/radeon_state.c |  8 +++----
 sys/dev/drm/savage_bci.c   | 28 ++++++++++++-------------
 sys/dev/drm/via_dma.c      | 13 +++++++++---
 sys/dev/drm/via_map.c      |  2 +-
 sys/dev/drm/via_mm.c       |  2 --
 23 files changed, 201 insertions(+), 168 deletions(-)

diff --git a/sys/dev/drm/drmP.h b/sys/dev/drm/drmP.h
index f89e75032ec..1204974dc56 100644
--- a/sys/dev/drm/drmP.h
+++ b/sys/dev/drm/drmP.h
@@ -239,22 +239,22 @@ typedef u_int8_t u8;
 #define DRM_MEMORYBARRIER()		mb()
 
 #define DRM_READ8(map, offset)						\
-	*(volatile u_int8_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int8_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset))
 #define DRM_READ16(map, offset)						\
-	*(volatile u_int16_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int16_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset))
 #define DRM_READ32(map, offset)						\
-	*(volatile u_int32_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int32_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset))
 #define DRM_WRITE8(map, offset, val)					\
-	*(volatile u_int8_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int8_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset)) = val
 #define DRM_WRITE16(map, offset, val)					\
-	*(volatile u_int16_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int16_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset)) = val
 #define DRM_WRITE32(map, offset, val)					\
-	*(volatile u_int32_t *)(((vm_offset_t)(map)->handle) +		\
+	*(volatile u_int32_t *)(((vm_offset_t)(map)->virtual) +		\
 	    (vm_offset_t)(offset)) = val
 
 #define DRM_VERIFYAREA_READ( uaddr, size )		\
@@ -481,18 +481,21 @@ typedef struct drm_sg_mem {
 	struct drm_dma_handle	 *dmah;		/* Handle to PCI memory  */
 } drm_sg_mem_t;
 
+#define DRM_MAP_HANDLE_BITS	(sizeof(void *) == 4 ? 4 : 24)
+#define DRM_MAP_HANDLE_SHIFT	(sizeof(void *) * 8 - DRM_MAP_HANDLE_BITS)
 typedef TAILQ_HEAD(drm_map_list, drm_local_map) drm_map_list_t;
 
 typedef struct drm_local_map {
-	unsigned long	offset;	 /* Physical address (0 for SAREA)*/
-	unsigned long	size;	 /* Physical size (bytes)	    */
-	enum drm_map_type	type;	 /* Type of memory mapped		    */
-	enum drm_map_flags	flags;	 /* Flags				    */
-	void		*handle; /* User-space: "Handle" to pass to mmap    */
-				 /* Kernel-space: kernel-virtual address    */
-	int		mtrr;	 /* Boolean: MTRR used */
-				 /* Private data			    */
-	int		rid;	 /* PCI resource ID for bus_space */
+	unsigned long offset;	  /* Physical address (0 for SAREA)       */
+	unsigned long size;	  /* Physical size (bytes)                */
+	enum drm_map_type type;	  /* Type of memory mapped                */
+	enum drm_map_flags flags; /* Flags                                */
+	void *handle;		  /* User-space: "Handle" to pass to mmap */
+				  /* Kernel-space: kernel-virtual address */
+	int mtrr;		  /* Boolean: MTRR used                   */
+				  /* Private data                         */
+	int rid;		  /* PCI resource ID for bus_space        */
+	void *virtual;		  /* Kernel-space: kernel-virtual address */
 	struct resource *bsr;
 	bus_space_tag_t bst;
 	bus_space_handle_t bsh;
@@ -643,6 +646,7 @@ struct drm_device {
 
 	/* Linked list of mappable regions. Protected by dev_lock */
 	drm_map_list_t	  maplist;
+	struct unrhdr	  *map_unrhdr;
 
 	drm_local_map_t	  **context_sareas;
 	int		  max_context;
@@ -973,17 +977,17 @@ drm_free(void *pt, size_t size, struct malloc_type *area)
 static __inline__ void
 drm_core_ioremap_wc(struct drm_local_map *map, struct drm_device *dev)
 {
-	map->handle = drm_ioremap_wc(dev, map);
+	map->virtual = drm_ioremap_wc(dev, map);
 }
 static __inline__ void
 drm_core_ioremap(struct drm_local_map *map, struct drm_device *dev)
 {
-	map->handle = drm_ioremap(dev, map);
+	map->virtual = drm_ioremap(dev, map);
 }
 static __inline__ void
 drm_core_ioremapfree(struct drm_local_map *map, struct drm_device *dev)
 {
-	if ( map->handle && map->size )
+	if ( map->virtual && map->size )
 		drm_ioremapfree(map);
 }
 
@@ -994,7 +998,7 @@ drm_core_findmap(struct drm_device *dev, unsigned long offset)
 
 	DRM_SPINLOCK_ASSERT(&dev->dev_lock);
 	TAILQ_FOREACH(map, &dev->maplist, link) {
-		if (map->offset == offset)
+		if (offset == (unsigned long)map->handle)
 			return map;
 	}
 	return NULL;
diff --git a/sys/dev/drm/drm_bufs.c b/sys/dev/drm/drm_bufs.c
index bd31b0ae31a..9c1ab5c62b9 100644
--- a/sys/dev/drm/drm_bufs.c
+++ b/sys/dev/drm/drm_bufs.c
@@ -156,10 +156,12 @@ int drm_addmap(struct drm_device * dev, unsigned long offset,
 	map->size = size;
 	map->type = type;
 	map->flags = flags;
+	map->handle = (void *)((unsigned long)alloc_unr(dev->map_unrhdr) <<
+	    DRM_MAP_HANDLE_SHIFT);
 
 	switch (map->type) {
 	case _DRM_REGISTERS:
-		map->handle = drm_ioremap(dev, map);
+		map->virtual = drm_ioremap(dev, map);
 		if (!(map->flags & _DRM_WRITE_COMBINING))
 			break;
 		/* FALLTHROUGH */
@@ -168,25 +170,25 @@ int drm_addmap(struct drm_device * dev, unsigned long offset,
 			map->mtrr = 1;
 		break;
 	case _DRM_SHM:
-		map->handle = malloc(map->size, DRM_MEM_MAPS, M_NOWAIT);
+		map->virtual = malloc(map->size, DRM_MEM_MAPS, M_NOWAIT);
 		DRM_DEBUG("%lu %d %p\n",
-		    map->size, drm_order(map->size), map->handle);
-		if (!map->handle) {
+		    map->size, drm_order(map->size), map->virtual);
+		if (!map->virtual) {
 			free(map, DRM_MEM_MAPS);
 			DRM_LOCK();
 			return ENOMEM;
 		}
-		map->offset = (unsigned long)map->handle;
+		map->offset = (unsigned long)map->virtual;
 		if (map->flags & _DRM_CONTAINS_LOCK) {
 			/* Prevent a 2nd X Server from creating a 2nd lock */
 			DRM_LOCK();
 			if (dev->lock.hw_lock != NULL) {
 				DRM_UNLOCK();
-				free(map->handle, DRM_MEM_MAPS);
+				free(map->virtual, DRM_MEM_MAPS);
 				free(map, DRM_MEM_MAPS);
 				return EBUSY;
 			}
-			dev->lock.hw_lock = map->handle; /* Pointer to lock */
+			dev->lock.hw_lock = map->virtual; /* Pointer to lock */
 			DRM_UNLOCK();
 		}
 		break;
@@ -224,7 +226,8 @@ int drm_addmap(struct drm_device * dev, unsigned long offset,
 			DRM_LOCK();
 			return EINVAL;
 		}
-		map->offset += dev->sg->handle;
+		map->virtual = (void *)(dev->sg->handle + offset);
+		map->offset = dev->sg->handle + offset;
 		break;
 	case _DRM_CONSISTENT:
 		/* Unfortunately, we don't get any alignment specification from
@@ -242,7 +245,7 @@ int drm_addmap(struct drm_device * dev, unsigned long offset,
 			DRM_LOCK();
 			return ENOMEM;
 		}
-		map->handle = map->dmah->vaddr;
+		map->virtual = map->dmah->vaddr;
 		map->offset = map->dmah->busaddr;
 		break;
 	default:
@@ -291,11 +294,7 @@ int drm_addmap_ioctl(struct drm_device *dev, void *data,
 	request->type = map->type;
 	request->flags = map->flags;
 	request->mtrr   = map->mtrr;
-	request->handle = map->handle;
-
-	if (request->type != _DRM_SHM) {
-		request->handle = (void *)request->offset;
-	}
+	request->handle = (void *)map->handle;
 
 	return 0;
 }
@@ -324,7 +323,7 @@ void drm_rmmap(struct drm_device *dev, drm_local_map_t *map)
 		}
 		break;
 	case _DRM_SHM:
-		free(map->handle, DRM_MEM_MAPS);
+		free(map->virtual, DRM_MEM_MAPS);
 		break;
 	case _DRM_AGP:
 	case _DRM_SCATTER_GATHER:
@@ -342,6 +341,12 @@ void drm_rmmap(struct drm_device *dev, drm_local_map_t *map)
 		    map->bsr);
 	}
 
+	DRM_UNLOCK();
+	if (map->handle)
+		free_unr(dev->map_unrhdr, (unsigned long)map->handle >>
+		    DRM_MAP_HANDLE_SHIFT);
+	DRM_LOCK();
+
 	free(map, DRM_MEM_MAPS);
 }
 
@@ -1054,7 +1059,7 @@ int drm_mapbufs(struct drm_device *dev, void *data, struct drm_file *file_priv)
 			goto done;
 		}
 		size = round_page(map->size);
-		foff = map->offset;
+		foff = (unsigned long)map->handle;
 	} else {
 		size = round_page(dma->byte_count),
 		foff = 0;
diff --git a/sys/dev/drm/drm_context.c b/sys/dev/drm/drm_context.c
index 398008bc187..3c13b43f9a7 100644
--- a/sys/dev/drm/drm_context.c
+++ b/sys/dev/drm/drm_context.c
@@ -147,7 +147,7 @@ int drm_getsareactx(struct drm_device *dev, void *data,
 	map = dev->context_sareas[request->ctx_id];
 	DRM_UNLOCK();
 
-	request->handle = map->handle;
+	request->handle = (void *)map->handle;
 
 	return 0;
 }
diff --git a/sys/dev/drm/drm_drv.c b/sys/dev/drm/drm_drv.c
index c690c34dfa3..8d9bc69e0c4 100644
--- a/sys/dev/drm/drm_drv.c
+++ b/sys/dev/drm/drm_drv.c
@@ -434,6 +434,12 @@ static int drm_load(struct drm_device *dev)
 	DRM_DEBUG("\n");
 
 	TAILQ_INIT(&dev->maplist);
+	dev->map_unrhdr = new_unrhdr(1, ((1 << DRM_MAP_HANDLE_BITS) - 1), NULL);
+	if (dev->map_unrhdr == NULL) {
+		DRM_ERROR("Couldn't allocate map number allocator\n");
+		return EINVAL;
+	}
+
 
 	drm_mem_init();
 	drm_sysctl_init(dev);
@@ -565,6 +571,7 @@ static void drm_unload(struct drm_device *dev)
 	}
 
 	delete_unrhdr(dev->drw_unrhdr);
+	delete_unrhdr(dev->map_unrhdr);
 
 	drm_mem_uninit();
 
diff --git a/sys/dev/drm/drm_memory.c b/sys/dev/drm/drm_memory.c
index 8eaf4d1e9e2..415b7741f71 100644
--- a/sys/dev/drm/drm_memory.c
+++ b/sys/dev/drm/drm_memory.c
@@ -83,7 +83,7 @@ void *drm_ioremap(struct drm_device *dev, drm_local_map_t *map)
 
 void drm_ioremapfree(drm_local_map_t *map)
 {
-	pmap_unmapdev((vm_offset_t) map->handle, map->size);
+	pmap_unmapdev((vm_offset_t) map->virtual, map->size);
 }
 
 int
diff --git a/sys/dev/drm/drm_scatter.c b/sys/dev/drm/drm_scatter.c
index b3ab63b630c..181531a2233 100644
--- a/sys/dev/drm/drm_scatter.c
+++ b/sys/dev/drm/drm_scatter.c
@@ -100,11 +100,14 @@ drm_sg_alloc(struct drm_device *dev, struct drm_scatter_gather *request)
 	dev->sg = entry;
 	DRM_UNLOCK();
 
-	DRM_DEBUG("handle=%08lx, kva=%p, contents=%08lx\n", entry->handle,
-	    entry->virtual, *(unsigned long *)entry->virtual);
+	pmap_change_attr((vm_offset_t)dmah->vaddr, request->size,
+	    PAT_WRITE_COMBINING);
 
 	request->handle = entry->handle;
 
+	DRM_DEBUG("handle=%08lx, kva=%p, contents=%08lx\n", entry->handle,
+	    entry->virtual, *(unsigned long *)entry->virtual);
+
 	return 0;
 }
 
diff --git a/sys/dev/drm/drm_sysctl.c b/sys/dev/drm/drm_sysctl.c
index cc332833fd6..4d9b0e8f0e0 100644
--- a/sys/dev/drm/drm_sysctl.c
+++ b/sys/dev/drm/drm_sysctl.c
@@ -188,7 +188,7 @@ static int drm_vm_info DRM_SYSCTL_HANDLER_ARGS
 	DRM_UNLOCK();
 
 	DRM_SYSCTL_PRINT("\nslot offset	        size       "
-	    "type flags address            mtrr\n");
+	    "type flags address            handle mtrr\n");
 
 	for (i = 0; i < mapcount; i++) {
 		map = &tempmaps[i];
@@ -204,9 +204,11 @@ static int drm_vm_info DRM_SYSCTL_HANDLER_ARGS
 			yesno = "yes";
 
 		DRM_SYSCTL_PRINT(
-		    "%4d 0x%016lx 0x%08lx %4.4s  0x%02x 0x%016lx %s\n", i,
-		    map->offset, map->size, type, map->flags,
-		    (unsigned long)map->handle, yesno);
+		    "%4d 0x%016lx 0x%08lx %4.4s  0x%02x 0x%016lx %6d %s\n",
+		    i, map->offset, map->size, type, map->flags,
+		    (unsigned long)map->virtual,
+		    (unsigned int)((unsigned long)map->handle >>
+		    DRM_MAP_HANDLE_SHIFT), yesno);
 	}
 	SYSCTL_OUT(req, "", 1);
 
diff --git a/sys/dev/drm/drm_vm.c b/sys/dev/drm/drm_vm.c
index 8f260fd9b10..798685607af 100644
--- a/sys/dev/drm/drm_vm.c
+++ b/sys/dev/drm/drm_vm.c
@@ -54,6 +54,7 @@ int drm_mmap(struct cdev *kdev, vm_ooffset_t offset, vm_paddr_t *paddr,
 	if (file_priv && !file_priv->authenticated)
 		return EACCES;
 
+	DRM_DEBUG("called with offset %016jx\n", offset);
 	if (dev->dma && offset < ptoa(dev->dma->page_count)) {
 		drm_device_dma_t *dma = dev->dma;
 
@@ -72,31 +73,31 @@ int drm_mmap(struct cdev *kdev, vm_ooffset_t offset, vm_paddr_t *paddr,
 		}
 	}
 
-				/* A sequential search of a linked list is
-				   fine here because: 1) there will only be
-				   about 5-10 entries in the list and, 2) a
-				   DRI client only has to do this mapping
-				   once, so it doesn't have to be optimized
-				   for performance, even if the list was a
-				   bit longer. */
+	/* A sequential search of a linked list is
+	   fine here because: 1) there will only be
+	   about 5-10 entries in the list and, 2) a
+	   DRI client only has to do this mapping
+	   once, so it doesn't have to be optimized
+	   for performance, even if the list was a
+	   bit longer.
+	*/
 	DRM_LOCK();
 	TAILQ_FOREACH(map, &dev->maplist, link) {
-		if (offset >= map->offset && offset < map->offset + map->size)
+		if (offset >> DRM_MAP_HANDLE_SHIFT ==
+		    (unsigned long)map->handle >> DRM_MAP_HANDLE_SHIFT)
 			break;
 	}
 
 	if (map == NULL) {
-		DRM_DEBUG("Can't find map, requested offset = %016lx\n",
-		    (unsigned long)offset);
+		DRM_DEBUG("Can't find map, request offset = %016jx\n", offset);
 		TAILQ_FOREACH(map, &dev->maplist, link) {
 			DRM_DEBUG("map offset = %016lx, handle = %016lx\n",
-			    (unsigned long)map->offset,
-			    (unsigned long)map->handle);
+			    map->offset, (unsigned long)map->handle);
 		}
 		DRM_UNLOCK();
 		return -1;
 	}
-	if (((map->flags&_DRM_RESTRICTED) && !DRM_SUSER(DRM_CURPROC))) {
+	if (((map->flags & _DRM_RESTRICTED) && !DRM_SUSER(DRM_CURPROC))) {
 		DRM_UNLOCK();
 		DRM_DEBUG("restricted map\n");
 		return -1;
@@ -104,18 +105,22 @@ int drm_mmap(struct cdev *kdev, vm_ooffset_t offset, vm_paddr_t *paddr,
 	type = map->type;
 	DRM_UNLOCK();
 
+	offset = offset & ((1ULL << DRM_MAP_HANDLE_SHIFT) - 1);
+
 	switch (type) {
 	case _DRM_FRAME_BUFFER:
-	case _DRM_REGISTERS:
 	case _DRM_AGP:
-		phys = offset;
-		break;
-	case _DRM_CONSISTENT:
-		phys = vtophys((char *)map->handle + (offset - map->offset));
+		*memattr = VM_MEMATTR_WRITE_COMBINING;
+		/* FALLTHROUGH */
+	case _DRM_REGISTERS:
+		phys = map->offset + offset;
 		break;
 	case _DRM_SCATTER_GATHER:
+		*memattr = VM_MEMATTR_WRITE_COMBINING;
+		/* FALLTHROUGH */
+	case _DRM_CONSISTENT:
 	case _DRM_SHM:
-		phys = vtophys(offset);
+		phys = vtophys((char *)map->virtual + offset);
 		break;
 	default:
 		DRM_ERROR("bad map type %d\n", type);
diff --git a/sys/dev/drm/i915_dma.c b/sys/dev/drm/i915_dma.c
index 386c058b351..7f8ddc194ed 100644
--- a/sys/dev/drm/i915_dma.c
+++ b/sys/dev/drm/i915_dma.c
@@ -151,7 +151,7 @@ static int i915_dma_cleanup(struct drm_device * dev)
 	if (dev_priv->ring.virtual_start) {
 		drm_core_ioremapfree(&dev_priv->ring.map, dev);
 		dev_priv->ring.virtual_start = NULL;
-		dev_priv->ring.map.handle = NULL;
+		dev_priv->ring.map.virtual = NULL;
 		dev_priv->ring.map.size = 0;
 	}
 
@@ -174,7 +174,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init)
 	}
 
 	dev_priv->sarea_priv = (drm_i915_sarea_t *)
-	    ((u8 *) dev_priv->sarea->handle + init->sarea_priv_offset);
+	    ((u8 *) dev_priv->sarea->virtual + init->sarea_priv_offset);
 
 	if (init->ring_size != 0) {
 		if (dev_priv->ring.ring_obj != NULL) {
@@ -195,7 +195,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init)
 
 		drm_core_ioremap_wc(&dev_priv->ring.map, dev);
 
-		if (dev_priv->ring.map.handle == NULL) {
+		if (dev_priv->ring.map.virtual == NULL) {
 			i915_dma_cleanup(dev);
 			DRM_ERROR("can not ioremap virtual address for"
 				  " ring buffer\n");
@@ -203,7 +203,7 @@ static int i915_initialize(struct drm_device * dev, drm_i915_init_t * init)
 		}
 	}
 
-	dev_priv->ring.virtual_start = dev_priv->ring.map.handle;
+	dev_priv->ring.virtual_start = dev_priv->ring.map.virtual;
 
 	dev_priv->cpp = init->cpp;
 	dev_priv->back_offset = init->back_offset;
@@ -229,7 +229,7 @@ static int i915_dma_resume(struct drm_device * dev)
 		return -EINVAL;
 	}
 
-	if (dev_priv->ring.map.handle == NULL) {
+	if (dev_priv->ring.map.virtual == NULL) {
 		DRM_ERROR("can not ioremap virtual address for"
 			  " ring buffer\n");
 		return -ENOMEM;
@@ -823,14 +823,14 @@ static int i915_set_status_page(struct drm_device *dev, void *data,
 	dev_priv->hws_map.mtrr = 0;
 
 	drm_core_ioremap_wc(&dev_priv->hws_map, dev);
-	if (dev_priv->hws_map.handle == NULL) {
+	if (dev_priv->hws_map.virtual == NULL) {
 		i915_dma_cleanup(dev);
 		dev_priv->status_gfx_addr = 0;
 		DRM_ERROR("can not ioremap virtual address for"
 				" G33 hw status page\n");
 		return -ENOMEM;
 	}
-	dev_priv->hw_status_page = dev_priv->hws_map.handle;
+	dev_priv->hw_status_page = dev_priv->hws_map.virtual;
 
 	memset(dev_priv->hw_status_page, 0, PAGE_SIZE);
 	I915_WRITE(HWS_PGA, dev_priv->status_gfx_addr);
diff --git a/sys/dev/drm/mach64_dma.c b/sys/dev/drm/mach64_dma.c
index 6068c85b3fe..9aa0faa1981 100644
--- a/sys/dev/drm/mach64_dma.c
+++ b/sys/dev/drm/mach64_dma.c
@@ -1078,11 +1078,11 @@ static int mach64_do_dma_init(struct drm_device * dev, drm_mach64_init_t * init)
 	}
 
 	dev_priv->sarea_priv = (drm_mach64_sarea_t *)
-	    ((u8 *) dev_priv->sarea->handle + init->sarea_priv_offset);
+	    ((u8 *) dev_priv->sarea->virtual + init->sarea_priv_offset);
 
 	if (!dev_priv->is_pci) {
 		drm_core_ioremap(dev_priv->ring_map, dev);
-		if (!dev_priv->ring_map->handle) {
+		if (!dev_priv->ring_map->virtual) {
 			DRM_ERROR("can not ioremap virtual address for"
 				  " descriptor ring\n");
 			dev->dev_private = (void *)dev_priv;
@@ -1103,7 +1103,7 @@ static int mach64_do_dma_init(struct drm_device * dev, drm_mach64_init_t * init)
 		dev_priv->dev_buffers = dev->agp_buffer_map;
 
 		drm_core_ioremap(dev->agp_buffer_map, dev);
-		if (!dev->agp_buffer_map->handle) {
+		if (!dev->agp_buffer_map->virtual) {
 			DRM_ERROR("can not ioremap virtual address for"
 				  " dma buffer\n");
 			dev->dev_private = (void *)dev_priv;
@@ -1147,7 +1147,7 @@ static int mach64_do_dma_init(struct drm_device * dev, drm_mach64_init_t * init)
 	}
 
 	dev_priv->ring.size = 0x4000;	/* 16KB */
-	dev_priv->ring.start = dev_priv->ring_map->handle;
+	dev_priv->ring.start = dev_priv->ring_map->virtual;
 	dev_priv->ring.start_addr = (u32) dev_priv->ring_map->offset;
 
 	memset(dev_priv->ring.start, 0, dev_priv->ring.size);
diff --git a/sys/dev/drm/mga_dma.c b/sys/dev/drm/mga_dma.c
index e081e2c72c3..71775b65b19 100644
--- a/sys/dev/drm/mga_dma.c
+++ b/sys/dev/drm/mga_dma.c
@@ -585,11 +585,11 @@ static int mga_do_agp_dma_bootstrap(struct drm_device *dev,
 	drm_core_ioremap(dev_priv->primary, dev);
 	drm_core_ioremap(dev->agp_buffer_map, dev);
 
-	if (!dev_priv->warp->handle ||
-	    !dev_priv->primary->handle || !dev->agp_buffer_map->handle) {
+	if (!dev_priv->warp->virtual ||
+	    !dev_priv->primary->virtual || !dev->agp_buffer_map->virtual) {
 		DRM_ERROR("failed to ioremap agp regions! (%p, %p, %p)\n",
-			  dev_priv->warp->handle, dev_priv->primary->handle,
-			  dev->agp_buffer_map->handle);
+			  dev_priv->warp->virtual, dev_priv->primary->virtual,
+			  dev->agp_buffer_map->virtual);
 		return -ENOMEM;
 	}
 
@@ -878,14 +878,14 @@ static int mga_do_init_dma(struct drm_device * dev, drm_mga_init_t * init)
 	}
 
 	dev_priv->sarea_priv =
-	    (drm_mga_sarea_t *) ((u8 *) dev_priv->sarea->handle +
+	    (drm_mga_sarea_t *) ((u8 *) dev_priv->sarea->virtual +
 				 init->sarea_priv_offset);
 
-	if (!dev_priv->warp->handle ||
-	    !dev_priv->primary->handle ||
+	if (!dev_priv->warp->virtual ||
+	    !dev_priv->primary->virtual ||
 	    ((dev_priv->dma_access != 0) &&
 	     ((dev->agp_buffer_map == NULL) ||
-	      (dev->agp_buffer_map->handle == NULL)))) {
+	      (dev->agp_buffer_map->virtual == NULL)))) {
 		DRM_ERROR("failed to ioremap agp regions!\n");
 		return -ENOMEM;
 	}
@@ -902,7 +902,7 @@ static int mga_do_init_dma(struct drm_device * dev, drm_mga_init_t * init)
 		return ret;
 	}
 
-	dev_priv->prim.status = (u32 *) dev_priv->status->handle;
+	dev_priv->prim.status = (u32 *) dev_priv->status->virtual;
 
 	mga_do_wait_for_idle(dev_priv);
 
@@ -910,8 +910,8 @@ static int mga_do_init_dma(struct drm_device * dev, drm_mga_init_t * init)
 	 */
 	MGA_WRITE(MGA_PRIMADDRESS, dev_priv->primary->offset | MGA_DMA_GENERAL);
 
-	dev_priv->prim.start = (u8 *) dev_priv->primary->handle;
-	dev_priv->prim.end = ((u8 *) dev_priv->primary->handle
+	dev_priv->prim.start = (u8 *) dev_priv->primary->virtual;
+	dev_priv->prim.end = ((u8 *) dev_priv->primary->virtual
 			      + dev_priv->primary->size);
 	dev_priv->prim.size = dev_priv->primary->size;
 
diff --git a/sys/dev/drm/mga_warp.c b/sys/dev/drm/mga_warp.c
index dd3e7346931..98c1615f509 100644
--- a/sys/dev/drm/mga_warp.c
+++ b/sys/dev/drm/mga_warp.c
@@ -96,7 +96,7 @@ unsigned int mga_warp_microcode_size(const drm_mga_private_t * dev_priv)
 
 static int mga_warp_install_g400_microcode(drm_mga_private_t * dev_priv)
 {
-	unsigned char *vcbase = dev_priv->warp->handle;
+	unsigned char *vcbase = dev_priv->warp->virtual;
 	unsigned long pcbase = dev_priv->warp->offset;
 
 	memset(dev_priv->warp_pipe_phys, 0, sizeof(dev_priv->warp_pipe_phys));
@@ -124,7 +124,7 @@ static int mga_warp_install_g400_microcode(drm_mga_private_t * dev_priv)
 
 static int mga_warp_install_g200_microcode(drm_mga_private_t * dev_priv)
 {
-	unsigned char *vcbase = dev_priv->warp->handle;
+	unsigned char *vcbase = dev_priv->warp->virtual;
 	unsigned long pcbase = dev_priv->warp->offset;
 
 	memset(dev_priv->warp_pipe_phys, 0, sizeof(dev_priv->warp_pipe_phys));
diff --git a/sys/dev/drm/r128_cce.c b/sys/dev/drm/r128_cce.c
index c799bd1d804..7d6a5683456 100644
--- a/sys/dev/drm/r128_cce.c
+++ b/sys/dev/drm/r128_cce.c
@@ -509,7 +509,7 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
 	}
 
 	dev_priv->sarea_priv =
-	    (drm_r128_sarea_t *) ((u8 *) dev_priv->sarea->handle +
+	    (drm_r128_sarea_t *) ((u8 *) dev_priv->sarea->virtual +
 				  init->sarea_priv_offset);
 
 #if __OS_HAS_AGP
@@ -517,9 +517,9 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
 		drm_core_ioremap(dev_priv->cce_ring, dev);
 		drm_core_ioremap(dev_priv->ring_rptr, dev);
 		drm_core_ioremap(dev->agp_buffer_map, dev);
-		if (!dev_priv->cce_ring->handle ||
-		    !dev_priv->ring_rptr->handle ||
-		    !dev->agp_buffer_map->handle) {
+		if (!dev_priv->cce_ring->virtual ||
+		    !dev_priv->ring_rptr->virtual ||
+		    !dev->agp_buffer_map->virtual) {
 			DRM_ERROR("Could not ioremap agp regions!\n");
 			dev->dev_private = (void *)dev_priv;
 			r128_do_cleanup_cce(dev);
@@ -528,10 +528,11 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
 	} else
 #endif
 	{
-		dev_priv->cce_ring->handle = (void *)dev_priv->cce_ring->offset;
-		dev_priv->ring_rptr->handle =
+		dev_priv->cce_ring->virtual =
+		    (void *)dev_priv->cce_ring->offset;
+		dev_priv->ring_rptr->virtual =
 		    (void *)dev_priv->ring_rptr->offset;
-		dev->agp_buffer_map->handle =
+		dev->agp_buffer_map->virtual =
 		    (void *)dev->agp_buffer_map->offset;
 	}
 
@@ -542,8 +543,8 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
 #endif
 		dev_priv->cce_buffers_offset = (unsigned long)dev->sg->virtual;
 
-	dev_priv->ring.start = (u32 *) dev_priv->cce_ring->handle;
-	dev_priv->ring.end = ((u32 *) dev_priv->cce_ring->handle
+	dev_priv->ring.start = (u32 *) dev_priv->cce_ring->virtual;
+	dev_priv->ring.end = ((u32 *) dev_priv->cce_ring->virtual
 			      + init->ring_size / sizeof(u32));
 	dev_priv->ring.size = init->ring_size;
 	dev_priv->ring.size_l2qw = drm_order(init->ring_size / 8);
diff --git a/sys/dev/drm/r128_state.c b/sys/dev/drm/r128_state.c
index 6aef11cdb79..9632dec2a68 100644
--- a/sys/dev/drm/r128_state.c
+++ b/sys/dev/drm/r128_state.c
@@ -657,7 +657,7 @@ static void r128_cce_dispatch_indirect(struct drm_device * dev,
 		 */
 		if (dwords & 1) {
 			u32 *data = (u32 *)
-			    ((char *)dev->agp_buffer_map->handle
+			    ((char *)dev->agp_buffer_map->virtual
 			     + buf->offset + start);
 			data[dwords++] = cpu_to_le32(R128_CCE_PACKET2);
 		}
@@ -722,7 +722,7 @@ static void r128_cce_dispatch_indices(struct drm_device * dev,
 
 		dwords = (end - start + 3) / sizeof(u32);
 
-		data = (u32 *) ((char *)dev->agp_buffer_map->handle
+		data = (u32 *) ((char *)dev->agp_buffer_map->virtual
 				+ buf->offset + start);
 
 		data[0] = cpu_to_le32(CCE_PACKET3(R128_3D_RNDR_GEN_INDX_PRIM,
diff --git a/sys/dev/drm/r600_blit.c b/sys/dev/drm/r600_blit.c
index a26f2c445b9..d3c41ae9fc5 100644
--- a/sys/dev/drm/r600_blit.c
+++ b/sys/dev/drm/r600_blit.c
@@ -1290,8 +1290,8 @@ set_shaders(struct drm_device *dev)
 	DRM_DEBUG("\n");
 
 	/* load shaders */
-	vs = (u32 *) ((char *)dev->agp_buffer_map->handle + dev_priv->blit_vb->offset);
-	ps = (u32 *) ((char *)dev->agp_buffer_map->handle + dev_priv->blit_vb->offset + 256);
+	vs = (u32 *) ((char *)dev->agp_buffer_map->virtual + dev_priv->blit_vb->offset);
+	ps = (u32 *) ((char *)dev->agp_buffer_map->virtual + dev_priv->blit_vb->offset + 256);
 
 	shader_size = sizeof(r6xx_vs) / 4;
 	for (i= 0; i < shader_size; i++)
@@ -1718,11 +1718,10 @@ r600_blit_copy(struct drm_device *dev,
 	u64 vb_addr;
 	u32 *vb;
 
-	vb = (u32 *) ((char *)dev->agp_buffer_map->handle +
+	vb = (u32 *) ((char *)dev->agp_buffer_map->virtual +
 	    dev_priv->blit_vb->offset + dev_priv->blit_vb->used);
-	DRM_DEBUG("src=0x%016llx, dst=0x%016llx, size=%d\n",
-	    (unsigned long long)src_gpu_addr,
-	    (unsigned long long)dst_gpu_addr, size_bytes);
+	DRM_DEBUG("src=0x%016jx, dst=0x%016jx, size=%d\n",
+	    src_gpu_addr, dst_gpu_addr, size_bytes);
 
 	if ((size_bytes & 3) || (src_gpu_addr & 3) || (dst_gpu_addr & 3)) {
 		max_bytes = 8192;
@@ -1759,7 +1758,7 @@ r600_blit_copy(struct drm_device *dev,
 				if (!dev_priv->blit_vb)
 					return;
 				set_shaders(dev);
-				vb = (u32 *) ((char *)dev->agp_buffer_map->handle +
+				vb = (u32 *) ((char *)dev->agp_buffer_map->virtual +
 				    dev_priv->blit_vb->offset + dev_priv->blit_vb->used);
 			}
 
@@ -1849,7 +1848,7 @@ r600_blit_copy(struct drm_device *dev,
 				if (!dev_priv->blit_vb)
 					return;
 				set_shaders(dev);
-				vb = (u32 *) ((char *)dev->agp_buffer_map->handle +
+				vb = (u32 *) ((char *)dev->agp_buffer_map->virtual +
 				    dev_priv->blit_vb->offset + dev_priv->blit_vb->used);
 			}
 
@@ -1928,7 +1927,7 @@ r600_blit_swap(struct drm_device *dev,
 			return;
 		set_shaders(dev);
 	}
-	vb = (u32 *) ((char *)dev->agp_buffer_map->handle +
+	vb = (u32 *) ((char *)dev->agp_buffer_map->virtual +
 	    dev_priv->blit_vb->offset + dev_priv->blit_vb->used);
 
 	sx2 = sx + w;
diff --git a/sys/dev/drm/r600_cp.c b/sys/dev/drm/r600_cp.c
index 2a4a6bd873c..d9345d86a19 100644
--- a/sys/dev/drm/r600_cp.c
+++ b/sys/dev/drm/r600_cp.c
@@ -1914,7 +1914,7 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	}
 
 	dev_priv->sarea_priv =
-	    (drm_radeon_sarea_t *) ((u8 *) dev_priv->sarea->handle +
+	    (drm_radeon_sarea_t *) ((u8 *) dev_priv->sarea->virtual +
 				    init->sarea_priv_offset);
 
 #if __OS_HAS_AGP
@@ -1923,9 +1923,9 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
 		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
 		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
-		if (!dev_priv->cp_ring->handle ||
-		    !dev_priv->ring_rptr->handle ||
-		    !dev->agp_buffer_map->handle) {
+		if (!dev_priv->cp_ring->virtual ||
+		    !dev_priv->ring_rptr->virtual ||
+		    !dev->agp_buffer_map->virtual) {
 			DRM_ERROR("could not find ioremap agp regions!\n");
 			r600_do_cleanup_cp(dev);
 			return -EINVAL;
@@ -1933,18 +1933,19 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	} else
 #endif
 	{
-		dev_priv->cp_ring->handle = (void *)dev_priv->cp_ring->offset;
-		dev_priv->ring_rptr->handle =
+		dev_priv->cp_ring->virtual =
+		    (void *)dev_priv->cp_ring->offset;
+		dev_priv->ring_rptr->virtual =
 		    (void *)dev_priv->ring_rptr->offset;
-		dev->agp_buffer_map->handle =
+		dev->agp_buffer_map->virtual =
 		    (void *)dev->agp_buffer_map->offset;
 
-		DRM_DEBUG("dev_priv->cp_ring->handle %p\n",
-			  dev_priv->cp_ring->handle);
-		DRM_DEBUG("dev_priv->ring_rptr->handle %p\n",
-			  dev_priv->ring_rptr->handle);
-		DRM_DEBUG("dev->agp_buffer_map->handle %p\n",
-			  dev->agp_buffer_map->handle);
+		DRM_DEBUG("dev_priv->cp_ring->virtual %p\n",
+			  dev_priv->cp_ring->virtual);
+		DRM_DEBUG("dev_priv->ring_rptr->virtual %p\n",
+			  dev_priv->ring_rptr->virtual);
+		DRM_DEBUG("dev->agp_buffer_map->virtual %p\n",
+			  dev->agp_buffer_map->virtual);
 	}
 
 	dev_priv->fb_location = (radeon_read_fb_location(dev_priv) & 0xffff) << 24;
@@ -2024,8 +2025,8 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	DRM_DEBUG("dev_priv->gart_buffers_offset 0x%08lx\n",
 		  dev_priv->gart_buffers_offset);
 
-	dev_priv->ring.start = (u32 *) dev_priv->cp_ring->handle;
-	dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->handle
+	dev_priv->ring.start = (u32 *) dev_priv->cp_ring->virtual;
+	dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->virtual
 			      + init->ring_size / sizeof(u32));
 	dev_priv->ring.size = init->ring_size;
 	dev_priv->ring.size_l2qw = drm_order(init->ring_size / 8);
@@ -2064,14 +2065,14 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 			dev_priv->gart_info.table_size;
 
 		drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
-		if (!dev_priv->gart_info.mapping.handle) {
+		if (!dev_priv->gart_info.mapping.virtual) {
 			DRM_ERROR("ioremap failed.\n");
 			r600_do_cleanup_cp(dev);
 			return -EINVAL;
 		}
 
 		dev_priv->gart_info.addr =
-			dev_priv->gart_info.mapping.handle;
+			dev_priv->gart_info.mapping.virtual;
 
 		DRM_DEBUG("Setting phys_pci_gart to %p %08lX\n",
 			  dev_priv->gart_info.addr,
@@ -2219,7 +2220,7 @@ int r600_cp_dispatch_indirect(struct drm_device *dev,
 		 */
 		while (dwords & 0xf) {
 			u32 *data = (u32 *)
-			    ((char *)dev->agp_buffer_map->handle
+			    ((char *)dev->agp_buffer_map->virtual
 			     + buf->offset + start);
 			data[dwords++] = RADEON_CP_PACKET2;
 		}
@@ -2343,7 +2344,8 @@ int r600_cp_dispatch_texture(struct drm_device * dev,
 		/* Dispatch the indirect buffer.
 		 */
 		buffer =
-		    (u32 *) ((char *)dev->agp_buffer_map->handle + buf->offset);
+		    (u32 *) ((char *)dev->agp_buffer_map->virtual +
+		    buf->offset);
 
 		if (DRM_COPY_FROM_USER(buffer, data, pass_size)) {
 			DRM_ERROR("EFAULT on pad, %d bytes\n", pass_size);
diff --git a/sys/dev/drm/radeon_cp.c b/sys/dev/drm/radeon_cp.c
index 2e85f115a6d..a6e609db1ae 100644
--- a/sys/dev/drm/radeon_cp.c
+++ b/sys/dev/drm/radeon_cp.c
@@ -53,7 +53,7 @@ u32 radeon_read_ring_rptr(drm_radeon_private_t *dev_priv, u32 off)
 		val = DRM_READ32(dev_priv->ring_rptr, off);
 	} else {
 		val = *(((volatile u32 *)
-			 dev_priv->ring_rptr->handle) +
+			 dev_priv->ring_rptr->virtual) +
 			(off / sizeof(u32)));
 		val = le32_to_cpu(val);
 	}
@@ -77,7 +77,7 @@ void radeon_write_ring_rptr(drm_radeon_private_t *dev_priv, u32 off, u32 val)
 	if (dev_priv->flags & RADEON_IS_AGP)
 		DRM_WRITE32(dev_priv->ring_rptr, off, val);
 	else
-		*(((volatile u32 *) dev_priv->ring_rptr->handle) +
+		*(((volatile u32 *) dev_priv->ring_rptr->virtual) +
 		  (off / sizeof(u32))) = cpu_to_le32(val);
 }
 
@@ -1278,7 +1278,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	}
 
 	dev_priv->sarea_priv =
-	    (drm_radeon_sarea_t *) ((u8 *) dev_priv->sarea->handle +
+	    (drm_radeon_sarea_t *) ((u8 *) dev_priv->sarea->virtual +
 				    init->sarea_priv_offset);
 
 #if __OS_HAS_AGP
@@ -1286,9 +1286,9 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 		drm_core_ioremap_wc(dev_priv->cp_ring, dev);
 		drm_core_ioremap_wc(dev_priv->ring_rptr, dev);
 		drm_core_ioremap_wc(dev->agp_buffer_map, dev);
-		if (!dev_priv->cp_ring->handle ||
-		    !dev_priv->ring_rptr->handle ||
-		    !dev->agp_buffer_map->handle) {
+		if (!dev_priv->cp_ring->virtual ||
+		    !dev_priv->ring_rptr->virtual ||
+		    !dev->agp_buffer_map->virtual) {
 			DRM_ERROR("could not find ioremap agp regions!\n");
 			radeon_do_cleanup_cp(dev);
 			return -EINVAL;
@@ -1296,19 +1296,19 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	} else
 #endif
 	{
-		dev_priv->cp_ring->handle =
+		dev_priv->cp_ring->virtual =
 			(void *)(unsigned long)dev_priv->cp_ring->offset;
-		dev_priv->ring_rptr->handle =
+		dev_priv->ring_rptr->virtual =
 			(void *)(unsigned long)dev_priv->ring_rptr->offset;
-		dev->agp_buffer_map->handle =
+		dev->agp_buffer_map->virtual =
 			(void *)(unsigned long)dev->agp_buffer_map->offset;
 
-		DRM_DEBUG("dev_priv->cp_ring->handle %p\n",
-			  dev_priv->cp_ring->handle);
-		DRM_DEBUG("dev_priv->ring_rptr->handle %p\n",
-			  dev_priv->ring_rptr->handle);
-		DRM_DEBUG("dev->agp_buffer_map->handle %p\n",
-			  dev->agp_buffer_map->handle);
+		DRM_DEBUG("dev_priv->cp_ring->virtual %p\n",
+			  dev_priv->cp_ring->virtual);
+		DRM_DEBUG("dev_priv->ring_rptr->virtual %p\n",
+			  dev_priv->ring_rptr->virtual);
+		DRM_DEBUG("dev->agp_buffer_map->virtual %p\n",
+			  dev->agp_buffer_map->virtual);
 	}
 
 	dev_priv->fb_location = (radeon_read_fb_location(dev_priv) & 0xffff) << 16;
@@ -1386,8 +1386,8 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 	DRM_DEBUG("dev_priv->gart_buffers_offset 0x%lx\n",
 		  dev_priv->gart_buffers_offset);
 
-	dev_priv->ring.start = (u32 *) dev_priv->cp_ring->handle;
-	dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->handle
+	dev_priv->ring.start = (u32 *) dev_priv->cp_ring->virtual;
+	dev_priv->ring.end = ((u32 *) dev_priv->cp_ring->virtual
 			      + init->ring_size / sizeof(u32));
 	dev_priv->ring.size = init->ring_size;
 	dev_priv->ring.size_l2qw = drm_order(init->ring_size / 8);
@@ -1423,7 +1423,7 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 
 			drm_core_ioremap_wc(&dev_priv->gart_info.mapping, dev);
 			dev_priv->gart_info.addr =
-			    dev_priv->gart_info.mapping.handle;
+			    dev_priv->gart_info.mapping.virtual;
 
 			if (dev_priv->flags & RADEON_IS_PCIE)
 				dev_priv->gart_info.gart_reg_if = DRM_ATI_GART_PCIE;
diff --git a/sys/dev/drm/radeon_cs.c b/sys/dev/drm/radeon_cs.c
index b523126b947..14fe2fc2afb 100644
--- a/sys/dev/drm/radeon_cs.c
+++ b/sys/dev/drm/radeon_cs.c
@@ -821,7 +821,7 @@ static int r600_ib_get(struct drm_radeon_cs_parser *parser)
 	}
 	buf->file_priv = parser->file_priv;
 	dev_priv->cs_buf = buf;
-	parser->ib = (void *)((vm_offset_t)dev->agp_buffer_map->handle +
+	parser->ib = (void *)((vm_offset_t)dev->agp_buffer_map->virtual +
 	    buf->offset);
 
 	return 0;
diff --git a/sys/dev/drm/radeon_state.c b/sys/dev/drm/radeon_state.c
index fd8388ffa59..806150128dd 100644
--- a/sys/dev/drm/radeon_state.c
+++ b/sys/dev/drm/radeon_state.c
@@ -1420,7 +1420,7 @@ static void radeon_cp_dispatch_swap(struct drm_device *dev)
 static void radeon_cp_dispatch_flip(struct drm_device *dev)
 {
 	drm_radeon_private_t *dev_priv = dev->dev_private;
-	struct drm_sarea *sarea = (struct drm_sarea *)dev_priv->sarea->handle;
+	struct drm_sarea *sarea = (struct drm_sarea *)dev_priv->sarea->virtual;
 	int offset = (dev_priv->sarea_priv->pfCurrentPage == 1)
 	    ? dev_priv->front_offset : dev_priv->back_offset;
 	RING_LOCALS;
@@ -1582,7 +1582,7 @@ static void radeon_cp_dispatch_indirect(struct drm_device * dev,
 		 */
 		if (dwords & 1) {
 			u32 *data = (u32 *)
-			    ((char *)dev->agp_buffer_map->handle
+			    ((char *)dev->agp_buffer_map->virtual
 			     + buf->offset + start);
 			data[dwords++] = RADEON_CP_PACKET2;
 		}
@@ -1629,7 +1629,7 @@ static void radeon_cp_dispatch_indices(struct drm_device *dev,
 
 	dwords = (prim->finish - prim->start + 3) / sizeof(u32);
 
-	data = (u32 *) ((char *)dev->agp_buffer_map->handle +
+	data = (u32 *) ((char *)dev->agp_buffer_map->virtual +
 			elt_buf->offset + prim->start);
 
 	data[0] = CP_PACKET3(RADEON_3D_RNDR_GEN_INDX_PRIM, dwords - 2);
@@ -1781,7 +1781,7 @@ static int radeon_cp_dispatch_texture(struct drm_device * dev,
 		/* Dispatch the indirect buffer.
 		 */
 		buffer =
-		    (u32 *) ((char *)dev->agp_buffer_map->handle + buf->offset);
+		    (u32 *) ((char *)dev->agp_buffer_map->virtual + buf->offset);
 		dwords = size / 4;
 
 #define RADEON_COPY_MT(_buf, _data, _width) \
diff --git a/sys/dev/drm/savage_bci.c b/sys/dev/drm/savage_bci.c
index 4168ddf819a..0f8d66e1289 100644
--- a/sys/dev/drm/savage_bci.c
+++ b/sys/dev/drm/savage_bci.c
@@ -376,7 +376,7 @@ uint32_t *savage_dma_alloc(drm_savage_private_t *dev_priv, unsigned int n)
 		  cur, dev_priv->dma_pages[cur].used, n, rest, nr_pages);
 
 	if (cur + nr_pages < dev_priv->nr_dma_pages) {
-		dma_ptr = (uint32_t *)dev_priv->cmd_dma->handle +
+		dma_ptr = (uint32_t *)dev_priv->cmd_dma->virtual +
 		    cur * SAVAGE_DMA_PAGE_SIZE + dev_priv->dma_pages[cur].used;
 		if (n < rest)
 			rest = n;
@@ -392,7 +392,7 @@ uint32_t *savage_dma_alloc(drm_savage_private_t *dev_priv, unsigned int n)
 			dev_priv->dma_pages[i].used = 0;
 			dev_priv->dma_pages[i].flushed = 0;
 		}
-		dma_ptr = (uint32_t *)dev_priv->cmd_dma->handle;
+		dma_ptr = (uint32_t *)dev_priv->cmd_dma->virtual;
 		dev_priv->first_dma_page = cur = 0;
 	}
 	for (i = cur; nr_pages > 0; ++i, --nr_pages) {
@@ -443,7 +443,7 @@ static void savage_dma_flush(drm_savage_private_t *dev_priv)
 
 	/* pad with noops */
 	if (pad) {
-		uint32_t *dma_ptr = (uint32_t *)dev_priv->cmd_dma->handle +
+		uint32_t *dma_ptr = (uint32_t *)dev_priv->cmd_dma->virtual +
 		    cur * SAVAGE_DMA_PAGE_SIZE + dev_priv->dma_pages[cur].used;
 		dev_priv->dma_pages[cur].used += pad;
 		while (pad != 0) {
@@ -517,7 +517,7 @@ static void savage_fake_dma_flush(drm_savage_private_t *dev_priv)
 	for (i = dev_priv->first_dma_page;
 	     i <= dev_priv->current_dma_page && dev_priv->dma_pages[i].used;
 	     ++i) {
-		uint32_t *dma_ptr = (uint32_t *)dev_priv->cmd_dma->handle +
+		uint32_t *dma_ptr = (uint32_t *)dev_priv->cmd_dma->virtual +
 			i * SAVAGE_DMA_PAGE_SIZE;
 #if SAVAGE_DMA_DEBUG
 		/* Sanity check: all pages except the last one must be full. */
@@ -784,7 +784,7 @@ static int savage_do_init_bci(struct drm_device *dev, drm_savage_init_t *init)
 				return -EINVAL;
 			}
 			drm_core_ioremap(dev_priv->cmd_dma, dev);
-			if (!dev_priv->cmd_dma->handle) {
+			if (!dev_priv->cmd_dma->virtual) {
 				DRM_ERROR("failed to ioremap command "
 					  "DMA region!\n");
 				savage_do_cleanup_bci(dev);
@@ -806,9 +806,9 @@ static int savage_do_init_bci(struct drm_device *dev, drm_savage_init_t *init)
 		dev_priv->fake_dma.offset = 0;
 		dev_priv->fake_dma.size = SAVAGE_FAKE_DMA_SIZE;
 		dev_priv->fake_dma.type = _DRM_SHM;
-		dev_priv->fake_dma.handle = drm_alloc(SAVAGE_FAKE_DMA_SIZE,
+		dev_priv->fake_dma.virtual = drm_alloc(SAVAGE_FAKE_DMA_SIZE,
 						      DRM_MEM_DRIVER);
-		if (!dev_priv->fake_dma.handle) {
+		if (!dev_priv->fake_dma.virtual) {
 			DRM_ERROR("could not allocate faked DMA buffer!\n");
 			savage_do_cleanup_bci(dev);
 			return -ENOMEM;
@@ -818,7 +818,7 @@ static int savage_do_init_bci(struct drm_device *dev, drm_savage_init_t *init)
 	}
 
 	dev_priv->sarea_priv =
-		(drm_savage_sarea_t *)((uint8_t *)dev_priv->sarea->handle +
+		(drm_savage_sarea_t *)((uint8_t *)dev_priv->sarea->virtual +
 				       init->sarea_priv_offset);
 
 	/* setup bitmap descriptors */
@@ -857,7 +857,7 @@ static int savage_do_init_bci(struct drm_device *dev, drm_savage_init_t *init)
 	dev_priv->event_counter = 0;
 	dev_priv->event_wrap = 0;
 	dev_priv->bci_ptr = (volatile uint32_t *)
-	    ((uint8_t *)dev_priv->mmio->handle + SAVAGE_BCI_OFFSET);
+	    ((uint8_t *)dev_priv->mmio->virtual + SAVAGE_BCI_OFFSET);
 	if (S3_SAVAGE3D_SERIES(dev_priv->chipset)) {
 		dev_priv->status_used_mask = SAVAGE_FIFO_USED_MASK_S3D;
 	} else {
@@ -865,7 +865,7 @@ static int savage_do_init_bci(struct drm_device *dev, drm_savage_init_t *init)
 	}
 	if (dev_priv->status != NULL) {
 		dev_priv->status_ptr =
-			(volatile uint32_t *)dev_priv->status->handle;
+			(volatile uint32_t *)dev_priv->status->virtual;
 		dev_priv->wait_fifo = savage_bci_wait_fifo_shadow;
 		dev_priv->wait_evnt = savage_bci_wait_event_shadow;
 		dev_priv->status_ptr[1023] = dev_priv->event_counter;
@@ -905,16 +905,16 @@ static int savage_do_cleanup_bci(struct drm_device *dev)
 	drm_savage_private_t *dev_priv = dev->dev_private;
 
 	if (dev_priv->cmd_dma == &dev_priv->fake_dma) {
-		if (dev_priv->fake_dma.handle)
-			drm_free(dev_priv->fake_dma.handle,
+		if (dev_priv->fake_dma.virtual)
+			drm_free(dev_priv->fake_dma.virtual,
 				 SAVAGE_FAKE_DMA_SIZE, DRM_MEM_DRIVER);
-	} else if (dev_priv->cmd_dma && dev_priv->cmd_dma->handle &&
+	} else if (dev_priv->cmd_dma && dev_priv->cmd_dma->virtual &&
 		   dev_priv->cmd_dma->type == _DRM_AGP &&
 		   dev_priv->dma_type == SAVAGE_DMA_AGP)
 		drm_core_ioremapfree(dev_priv->cmd_dma, dev);
 
 	if (dev_priv->dma_type == SAVAGE_DMA_AGP &&
-	    dev->agp_buffer_map && dev->agp_buffer_map->handle) {
+	    dev->agp_buffer_map && dev->agp_buffer_map->virtual) {
 		drm_core_ioremapfree(dev->agp_buffer_map, dev);
 		/* make sure the next instance (which may be running
 		 * in PCI mode) doesn't try to use an old
diff --git a/sys/dev/drm/via_dma.c b/sys/dev/drm/via_dma.c
index e1af2938072..6f435fe5986 100644
--- a/sys/dev/drm/via_dma.c
+++ b/sys/dev/drm/via_dma.c
@@ -158,6 +158,9 @@ static inline uint32_t *via_check_dma(drm_via_private_t * dev_priv,
 
 int via_dma_cleanup(struct drm_device * dev)
 {
+	drm_via_blitq_t *blitq;
+	int i;
+
 	if (dev->dev_private) {
 		drm_via_private_t *dev_priv =
 		    (drm_via_private_t *) dev->dev_private;
@@ -169,6 +172,10 @@ int via_dma_cleanup(struct drm_device * dev)
 			dev_priv->ring.virtual_start = NULL;
 		}
 
+		for (i=0; i< VIA_NUM_BLIT_ENGINES; ++i) {
+			blitq = dev_priv->blit_queues + i;
+			mtx_destroy(&blitq->blit_lock);
+		}
 	}
 
 	return 0;
@@ -206,14 +213,14 @@ static int via_initialize(struct drm_device * dev,
 
 	drm_core_ioremap_wc(&dev_priv->ring.map, dev);
 
-	if (dev_priv->ring.map.handle == NULL) {
+	if (dev_priv->ring.map.virtual == NULL) {
 		via_dma_cleanup(dev);
 		DRM_ERROR("can not ioremap virtual address for"
 			  " ring buffer\n");
 		return -ENOMEM;
 	}
 
-	dev_priv->ring.virtual_start = dev_priv->ring.map.handle;
+	dev_priv->ring.virtual_start = dev_priv->ring.map.virtual;
 
 	dev_priv->dma_ptr = dev_priv->ring.virtual_start;
 	dev_priv->dma_low = 0;
@@ -222,7 +229,7 @@ static int via_initialize(struct drm_device * dev,
 	dev_priv->dma_offset = init->offset;
 	dev_priv->last_pause_ptr = NULL;
 	dev_priv->hw_addr_ptr =
-		(volatile uint32_t *)((char *)dev_priv->mmio->handle +
+		(volatile uint32_t *)((char *)dev_priv->mmio->virtual +
 		init->reg_pause_addr);
 
 	via_cmdbuf_start(dev_priv);
diff --git a/sys/dev/drm/via_map.c b/sys/dev/drm/via_map.c
index 5504a6b198d..687b8be8013 100644
--- a/sys/dev/drm/via_map.c
+++ b/sys/dev/drm/via_map.c
@@ -59,7 +59,7 @@ static int via_do_init_map(struct drm_device * dev, drm_via_init_t * init)
 	}
 
 	dev_priv->sarea_priv =
-	    (drm_via_sarea_t *) ((u8 *) dev_priv->sarea->handle +
+	    (drm_via_sarea_t *) ((u8 *) dev_priv->sarea->virtual +
 				 init->sarea_priv_offset);
 
 	dev_priv->agpAddr = init->agpAddr;
diff --git a/sys/dev/drm/via_mm.c b/sys/dev/drm/via_mm.c
index 9aaee9f4445..6dc185ba125 100644
--- a/sys/dev/drm/via_mm.c
+++ b/sys/dev/drm/via_mm.c
@@ -45,7 +45,6 @@ int via_agp_init(struct drm_device *dev, void *data, struct drm_file *file_priv)
 
 	ret = drm_sman_set_range(&dev_priv->sman, VIA_MEM_AGP, 0,
 				 agp->size >> VIA_MM_ALIGN_SHIFT);
-
 	if (ret) {
 		DRM_ERROR("AGP memory manager initialisation error\n");
 		return ret;
@@ -66,7 +65,6 @@ int via_fb_init(struct drm_device *dev, void *data, struct drm_file *file_priv)
 
 	ret = drm_sman_set_range(&dev_priv->sman, VIA_MEM_VIDEO, 0,
 				 fb->size >> VIA_MM_ALIGN_SHIFT);
-
 	if (ret) {
 		DRM_ERROR("VRAM memory manager initialisation error\n");
 		return ret;

From 2ee6488c02506b8a645055a08dec655e5ddfbae8 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Thu, 22 Apr 2010 18:44:23 +0000
Subject: [PATCH 199/532] re-write scatter gather memory allocation yet
 again...

This time, abandon the use of busdma and start interacting with the VM
system directly.  Make use of the new kmem_alloc_attr() which allows us
to easily allocate non-contiguous pages to back the GART table.  This
should help a lot when starting or restarting X after the system has
been running for a while and memory has become fragmented.

MFC after:	2 weeks
---
 sys/dev/drm/drmP.h        |  13 +++--
 sys/dev/drm/drm_bufs.c    |   6 +-
 sys/dev/drm/drm_scatter.c | 119 ++++++++++++--------------------------
 sys/dev/drm/r128_cce.c    |   5 +-
 sys/dev/drm/r600_cp.c     |  15 ++---
 sys/dev/drm/radeon_cp.c   |  15 ++---
 6 files changed, 61 insertions(+), 112 deletions(-)

diff --git a/sys/dev/drm/drmP.h b/sys/dev/drm/drmP.h
index 1204974dc56..26bda43659d 100644
--- a/sys/dev/drm/drmP.h
+++ b/sys/dev/drm/drmP.h
@@ -49,6 +49,7 @@ struct drm_file;
 #include <sys/module.h>
 #include <sys/systm.h>
 #include <sys/conf.h>
+#include <sys/sglist.h>
 #include <sys/stat.h>
 #if __FreeBSD_version >= 700000
 #include <sys/priv.h>
@@ -68,10 +69,13 @@ struct drm_file;
 #include <vm/vm.h>
 #include <vm/pmap.h>
 #include <vm/vm_extern.h>
+#include <vm/vm_kern.h>
 #include <vm/vm_map.h>
 #include <vm/vm_object.h>
 #include <vm/vm_page.h>
+#include <vm/vm_pager.h>
 #include <vm/vm_param.h>
+#include <vm/vm_phys.h>
 #include <machine/param.h>
 #include <machine/pmap.h>
 #include <machine/bus.h>
@@ -474,11 +478,10 @@ typedef struct drm_agp_head {
 } drm_agp_head_t;
 
 typedef struct drm_sg_mem {
-	unsigned long		  handle;
-	void			 *virtual;
-	int			  pages;
-	dma_addr_t		 *busaddr;
-	struct drm_dma_handle	 *dmah;		/* Handle to PCI memory  */
+	vm_offset_t vaddr;
+	vm_paddr_t *busaddr;
+	vm_pindex_t pages;
+	vm_object_t obj;
 } drm_sg_mem_t;
 
 #define DRM_MAP_HANDLE_BITS	(sizeof(void *) == 4 ? 4 : 24)
diff --git a/sys/dev/drm/drm_bufs.c b/sys/dev/drm/drm_bufs.c
index 9c1ab5c62b9..2d27cd4b88b 100644
--- a/sys/dev/drm/drm_bufs.c
+++ b/sys/dev/drm/drm_bufs.c
@@ -226,8 +226,8 @@ int drm_addmap(struct drm_device * dev, unsigned long offset,
 			DRM_LOCK();
 			return EINVAL;
 		}
-		map->virtual = (void *)(dev->sg->handle + offset);
-		map->offset = dev->sg->handle + offset;
+		map->virtual = (void *)(dev->sg->vaddr + offset);
+		map->offset = dev->sg->vaddr + offset;
 		break;
 	case _DRM_CONSISTENT:
 		/* Unfortunately, we don't get any alignment specification from
@@ -744,7 +744,7 @@ static int drm_do_addbufs_sg(struct drm_device *dev, struct drm_buf_desc *reques
 
 		buf->offset  = (dma->byte_count + offset);
 		buf->bus_address = agp_offset + offset;
-		buf->address = (void *)(agp_offset + offset + dev->sg->handle);
+		buf->address = (void *)(agp_offset + offset + dev->sg->vaddr);
 		buf->next    = NULL;
 		buf->pending = 0;
 		buf->file_priv = NULL;
diff --git a/sys/dev/drm/drm_scatter.c b/sys/dev/drm/drm_scatter.c
index 181531a2233..9a1a4b1e1d3 100644
--- a/sys/dev/drm/drm_scatter.c
+++ b/sys/dev/drm/drm_scatter.c
@@ -1,5 +1,5 @@
 /*-
- * Copyright 2000 VA Linux Systems, Inc., Sunnyvale, California.
+ * Copyright (c) 2009 Robert C. Noland III <rnoland@FreeBSD.org>
  * All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -20,11 +20,6 @@
  * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
  * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
  * DEALINGS IN THE SOFTWARE.
- *
- * Authors:
- *   Gareth Hughes <gareth@valinux.com>
- *   Eric Anholt <anholt@FreeBSD.org>
- *
  */
 
 #include <sys/cdefs.h>
@@ -32,99 +27,58 @@ __FBSDID("$FreeBSD$");
 
 /** @file drm_scatter.c
  * Allocation of memory for scatter-gather mappings by the graphics chip.
- *
  * The memory allocated here is then made into an aperture in the card
- * by drm_ati_pcigart_init().
+ * by mapping the pages into the GART.
  */
 
 #include "dev/drm/drmP.h"
 
-static void drm_sg_alloc_cb(void *arg, bus_dma_segment_t *segs,
-			    int nsegs, int error);
-
 int
 drm_sg_alloc(struct drm_device *dev, struct drm_scatter_gather *request)
 {
 	struct drm_sg_mem *entry;
-	struct drm_dma_handle *dmah;
-	int ret;
+	vm_size_t size;
+	vm_pindex_t pindex;
 
 	if (dev->sg)
 		return EINVAL;
 
-	entry = malloc(sizeof(*entry), DRM_MEM_SGLISTS, M_WAITOK | M_ZERO);
-	entry->pages = round_page(request->size) / PAGE_SIZE;
-	DRM_DEBUG("sg size=%ld pages=%d\n", request->size, entry->pages);
+	DRM_DEBUG("request size=%ld\n", request->size);
 
+	entry = malloc(sizeof(*entry), DRM_MEM_DRIVER, M_WAITOK | M_ZERO);
+
+	size = round_page(request->size);
+	entry->pages = OFF_TO_IDX(size);
 	entry->busaddr = malloc(entry->pages * sizeof(*entry->busaddr),
-	    DRM_MEM_PAGES, M_WAITOK | M_ZERO);
-	dmah = malloc(sizeof(struct drm_dma_handle), DRM_MEM_DMA,
-	    M_WAITOK | M_ZERO);
-	entry->dmah = dmah;
+	    DRM_MEM_SGLISTS, M_WAITOK | M_ZERO);
 
-	ret = bus_dma_tag_create(NULL, PAGE_SIZE, 0, /* tag, align, boundary */
-	    BUS_SPACE_MAXADDR_32BIT, BUS_SPACE_MAXADDR, /* lowaddr, highaddr */
-	    NULL, NULL, /* filtfunc, filtfuncargs */
-	    request->size, entry->pages, /* maxsize, nsegs */
-	    PAGE_SIZE, 0, /* maxsegsize, flags */
-	    NULL, NULL, /* lockfunc, lockfuncargs */
-	    &dmah->tag);
-	if (ret != 0) {
+	entry->vaddr = kmem_alloc_attr(kernel_map, size, M_WAITOK | M_ZERO,
+	    0, BUS_SPACE_MAXADDR_32BIT, VM_MEMATTR_WRITE_COMBINING);
+	if (entry->vaddr == 0) {
 		drm_sg_cleanup(entry);
-		return ENOMEM;
+		return (ENOMEM);
 	}
 
-	ret = bus_dmamem_alloc(dmah->tag, &dmah->vaddr,
-	    BUS_DMA_WAITOK | BUS_DMA_ZERO | BUS_DMA_NOCACHE, &dmah->map);
-	if (ret != 0) {
-		drm_sg_cleanup(entry);
-		return ENOMEM;
-	}
-
-	entry->handle = (unsigned long)dmah->vaddr;
-	entry->virtual = dmah->vaddr;
-
-	ret = bus_dmamap_load(dmah->tag, dmah->map, dmah->vaddr,
-	    request->size, drm_sg_alloc_cb, entry, BUS_DMA_NOWAIT);
-	if (ret != 0) {
-		drm_sg_cleanup(entry);
-		return ENOMEM;
+	for(pindex = 0; pindex < entry->pages; pindex++) {
+		entry->busaddr[pindex] =
+		    vtophys(entry->vaddr + IDX_TO_OFF(pindex));
 	}
 
 	DRM_LOCK();
 	if (dev->sg) {
 		DRM_UNLOCK();
 		drm_sg_cleanup(entry);
-		return EINVAL;
+		return (EINVAL);
 	}
 	dev->sg = entry;
 	DRM_UNLOCK();
 
-	pmap_change_attr((vm_offset_t)dmah->vaddr, request->size,
-	    PAT_WRITE_COMBINING);
+	request->handle = entry->vaddr;
 
-	request->handle = entry->handle;
+	DRM_DEBUG("allocated %ju pages @ 0x%08zx, contents=%08lx\n",
+	    entry->pages, entry->vaddr, *(unsigned long *)entry->vaddr);
 
-	DRM_DEBUG("handle=%08lx, kva=%p, contents=%08lx\n", entry->handle,
-	    entry->virtual, *(unsigned long *)entry->virtual);
-
-	return 0;
-}
-
-static void
-drm_sg_alloc_cb(void *arg, bus_dma_segment_t *segs, int nsegs, int error)
-{
-	struct drm_sg_mem *entry = arg;
-	int i;
-
-	if (error != 0)
-	    return;
-
-	for(i = 0 ; i < nsegs ; i++) {
-		entry->busaddr[i] = segs[i].ds_addr;
-		DRM_DEBUG("segment %d @ 0x%016lx\n", i,
-		    (unsigned long)segs[i].ds_addr);
-	}
+	return (0);
 }
 
 int
@@ -135,23 +89,22 @@ drm_sg_alloc_ioctl(struct drm_device *dev, void *data,
 
 	DRM_DEBUG("\n");
 
-	return drm_sg_alloc(dev, request);
+	return (drm_sg_alloc(dev, request));
 }
 
 void
 drm_sg_cleanup(struct drm_sg_mem *entry)
 {
-	struct drm_dma_handle *dmah = entry->dmah;
+	if (entry == NULL)
+		return;
 
-	if (dmah->map != NULL)
-		bus_dmamap_unload(dmah->tag, dmah->map);
-	if (dmah->vaddr != NULL)
-		bus_dmamem_free(dmah->tag, dmah->vaddr, dmah->map);
-	if (dmah->tag != NULL)
-		bus_dma_tag_destroy(dmah->tag);
-	free(dmah, DRM_MEM_DMA);
-	free(entry->busaddr, DRM_MEM_PAGES);
-	free(entry, DRM_MEM_SGLISTS);
+	if (entry->vaddr != 0)
+		kmem_free(kernel_map, entry->vaddr, IDX_TO_OFF(entry->pages));
+
+	free(entry->busaddr, DRM_MEM_SGLISTS);
+	free(entry, DRM_MEM_DRIVER);
+
+	return;
 }
 
 int
@@ -165,12 +118,12 @@ drm_sg_free(struct drm_device *dev, void *data, struct drm_file *file_priv)
 	dev->sg = NULL;
 	DRM_UNLOCK();
 
-	if (!entry || entry->handle != request->handle)
-		return EINVAL;
+	if (!entry || entry->vaddr != request->handle)
+		return (EINVAL);
 
-	DRM_DEBUG("sg free virtual = 0x%lx\n", entry->handle);
+	DRM_DEBUG("free 0x%zx\n", entry->vaddr);
 
 	drm_sg_cleanup(entry);
 
-	return 0;
+	return (0);
 }
diff --git a/sys/dev/drm/r128_cce.c b/sys/dev/drm/r128_cce.c
index 7d6a5683456..2bda4a58fd3 100644
--- a/sys/dev/drm/r128_cce.c
+++ b/sys/dev/drm/r128_cce.c
@@ -327,8 +327,7 @@ static void r128_cce_init_ring_buffer(struct drm_device * dev,
 		ring_start = dev_priv->cce_ring->offset - dev->agp->base;
 	else
 #endif
-		ring_start = dev_priv->cce_ring->offset -
-				(unsigned long)dev->sg->virtual;
+		ring_start = dev_priv->cce_ring->offset - dev->sg->vaddr;
 
 	R128_WRITE(R128_PM4_BUFFER_OFFSET, ring_start | R128_AGP_OFFSET);
 
@@ -541,7 +540,7 @@ static int r128_do_init_cce(struct drm_device * dev, drm_r128_init_t * init)
 		dev_priv->cce_buffers_offset = dev->agp->base;
 	else
 #endif
-		dev_priv->cce_buffers_offset = (unsigned long)dev->sg->virtual;
+		dev_priv->cce_buffers_offset = dev->sg->vaddr;
 
 	dev_priv->ring.start = (u32 *) dev_priv->cce_ring->virtual;
 	dev_priv->ring.end = ((u32 *) dev_priv->cce_ring->virtual
diff --git a/sys/dev/drm/r600_cp.c b/sys/dev/drm/r600_cp.c
index d9345d86a19..f5c3b97d793 100644
--- a/sys/dev/drm/r600_cp.c
+++ b/sys/dev/drm/r600_cp.c
@@ -1670,9 +1670,8 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 	} else
 #endif
 	{
-		rptr_addr = dev_priv->ring_rptr->offset
-			- ((unsigned long) dev->sg->virtual)
-			+ dev_priv->gart_vm_start;
+		rptr_addr = dev_priv->ring_rptr->offset - dev->sg->vaddr +
+		    dev_priv->gart_vm_start;
 	}
 	RADEON_WRITE(R600_CP_RB_RPTR_ADDR,
 		     rptr_addr & 0xffffffff);
@@ -1706,9 +1705,8 @@ static void r600_cp_init_ring_buffer(struct drm_device *dev,
 			      + dev_priv->gart_vm_start);
 	} else
 #endif
-		ring_start = (dev_priv->cp_ring->offset
-			      - (unsigned long)dev->sg->virtual
-			      + dev_priv->gart_vm_start);
+		ring_start = dev_priv->cp_ring->offset - dev->sg->vaddr +
+		    dev_priv->gart_vm_start;
 
 	RADEON_WRITE(R600_CP_RB_BASE, ring_start >> 8);
 
@@ -2012,9 +2010,8 @@ int r600_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 						 + dev_priv->gart_vm_start);
 	else
 #endif
-		dev_priv->gart_buffers_offset = (dev->agp_buffer_map->offset
-						 - (unsigned long)dev->sg->virtual
-						 + dev_priv->gart_vm_start);
+		dev_priv->gart_buffers_offset = dev->agp_buffer_map->offset -
+		    dev->sg->vaddr + dev_priv->gart_vm_start;
 
 	DRM_DEBUG("fb 0x%08x size %d\n",
 		  (unsigned int) dev_priv->fb_location,
diff --git a/sys/dev/drm/radeon_cp.c b/sys/dev/drm/radeon_cp.c
index a6e609db1ae..0a486afa4bd 100644
--- a/sys/dev/drm/radeon_cp.c
+++ b/sys/dev/drm/radeon_cp.c
@@ -720,9 +720,8 @@ static void radeon_cp_init_ring_buffer(struct drm_device * dev,
 			      + dev_priv->gart_vm_start);
 	} else
 #endif
-		ring_start = (dev_priv->cp_ring->offset
-			      - (unsigned long)dev->sg->virtual
-			      + dev_priv->gart_vm_start);
+		ring_start = (dev_priv->cp_ring->offset - dev->sg->vaddr +
+		    dev_priv->gart_vm_start);
 
 	RADEON_WRITE(RADEON_CP_RB_BASE, ring_start);
 
@@ -744,9 +743,8 @@ static void radeon_cp_init_ring_buffer(struct drm_device * dev,
 #endif
 	{
 		RADEON_WRITE(RADEON_CP_RB_RPTR_ADDR,
-			     dev_priv->ring_rptr->offset
-			     - ((unsigned long) dev->sg->virtual)
-			     + dev_priv->gart_vm_start);
+		    dev_priv->ring_rptr->offset - dev->sg->vaddr +
+		    dev_priv->gart_vm_start);
 	}
 
 	/* Set ring buffer size */
@@ -1377,9 +1375,8 @@ static int radeon_do_init_cp(struct drm_device *dev, drm_radeon_init_t *init,
 						 + dev_priv->gart_vm_start);
 	else
 #endif
-		dev_priv->gart_buffers_offset = (dev->agp_buffer_map->offset
-					- (unsigned long)dev->sg->virtual
-					+ dev_priv->gart_vm_start);
+		dev_priv->gart_buffers_offset = dev->agp_buffer_map->offset -
+		    dev->sg->vaddr + dev_priv->gart_vm_start;
 
 	DRM_DEBUG("dev_priv->gart_size %d\n", dev_priv->gart_size);
 	DRM_DEBUG("dev_priv->gart_vm_start 0x%x\n", dev_priv->gart_vm_start);

From 7af9c09a61db245f8eebb8a91ba211e4675e1e6a Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 22 Apr 2010 18:47:23 +0000
Subject: [PATCH 200/532] Allow to modify directory's content even if the
 ZFS_NOUNLINK (SF_NOUNLINK, sunlnk) flag is set. We only deny dirctory's
 removal or rename.

PR:		kern/143343
Reported by:	marck
MFC after:	3 days
---
 .../contrib/opensolaris/uts/common/fs/zfs/zfs_acl.c | 13 +++++++++++++
 1 file changed, 13 insertions(+)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_acl.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_acl.c
index eb93721e167..9825d83fb67 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_acl.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_acl.c
@@ -2235,11 +2235,24 @@ zfs_zaccess_common(znode_t *zp, uint32_t v4_mode, uint32_t *working_mode,
 		return (EPERM);
 	}
 
+#ifdef sun
 	if ((v4_mode & (ACE_DELETE | ACE_DELETE_CHILD)) &&
 	    (zp->z_phys->zp_flags & ZFS_NOUNLINK)) {
 		*check_privs = B_FALSE;
 		return (EPERM);
 	}
+#else
+	/*
+	 * In FreeBSD we allow to modify directory's content is ZFS_NOUNLINK
+	 * (sunlnk) is set. We just don't allow directory removal, which is
+	 * handled in zfs_zaccess_delete().
+	 */
+	if ((v4_mode & ACE_DELETE) &&
+	    (zp->z_phys->zp_flags & ZFS_NOUNLINK)) {
+		*check_privs = B_FALSE;
+		return (EPERM);
+	}
+#endif
 
 	if (((v4_mode & (ACE_READ_DATA|ACE_EXECUTE)) &&
 	    (zp->z_phys->zp_flags & ZFS_AV_QUARANTINED))) {

From a003741e38224ac0e8f0b03a5a3cb250581b7da5 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Thu, 22 Apr 2010 18:47:30 +0000
Subject: [PATCH 201/532] Now that we properly set write-combining on the pages
 that back the GART, disable snooping on radeons.

MFC after:	2 weeks
---
 sys/dev/drm/ati_pcigart.c | 11 +++++++----
 sys/dev/drm/r600_cp.c     |  2 +-
 2 files changed, 8 insertions(+), 5 deletions(-)

diff --git a/sys/dev/drm/ati_pcigart.c b/sys/dev/drm/ati_pcigart.c
index e3fecd1071b..063288ed64d 100644
--- a/sys/dev/drm/ati_pcigart.c
+++ b/sys/dev/drm/ati_pcigart.c
@@ -39,8 +39,9 @@ __FBSDID("$FreeBSD$");
 #define ATI_PCIGART_PAGE_SIZE		4096	/* PCI GART page size */
 #define ATI_PCIGART_PAGE_MASK		(~(ATI_PCIGART_PAGE_SIZE-1))
 
-#define ATI_PCIE_WRITE 0x4
-#define ATI_PCIE_READ 0x8
+#define ATI_GART_NOSNOOP	0x1
+#define ATI_GART_WRITE		0x4
+#define ATI_GART_READ		0x8
 
 static void
 drm_ati_alloc_pcigart_table_cb(void *arg, bus_dma_segment_t *segs,
@@ -196,13 +197,15 @@ drm_ati_pcigart_init(struct drm_device *dev,
 			case DRM_ATI_GART_IGP:
 				page_base |=
 				    (upper_32_bits(entry_addr) & 0xff) << 4;
-				page_base |= 0xc;
+				page_base |= ATI_GART_READ | ATI_GART_WRITE;
+				page_base |= ATI_GART_NOSNOOP;
 				break;
 			case DRM_ATI_GART_PCIE:
 				page_base >>= 8;
 				page_base |=
 				    (upper_32_bits(entry_addr) & 0xff) << 24;
-				page_base |= ATI_PCIE_READ | ATI_PCIE_WRITE;
+				page_base |= ATI_GART_READ | ATI_GART_WRITE;
+				page_base |= ATI_GART_NOSNOOP;
 				break;
 			default:
 			case DRM_ATI_GART_PCI:
diff --git a/sys/dev/drm/r600_cp.c b/sys/dev/drm/r600_cp.c
index f5c3b97d793..c2b8770acdd 100644
--- a/sys/dev/drm/r600_cp.c
+++ b/sys/dev/drm/r600_cp.c
@@ -180,7 +180,7 @@ int r600_page_table_init(struct drm_device *dev)
 		entry_addr = entry->busaddr[i];
 		for (j = 0; j < (PAGE_SIZE / ATI_PCIGART_PAGE_SIZE); j++) {
 			page_base = (u64) entry_addr & ATI_PCIGART_PAGE_MASK;
-			page_base |= R600_PTE_VALID | R600_PTE_SYSTEM | R600_PTE_SNOOPED;
+			page_base |= R600_PTE_VALID | R600_PTE_SYSTEM;
 			page_base |= R600_PTE_READABLE | R600_PTE_WRITEABLE;
 
 			*pci_gart = page_base;

From 36df4f8d056ab10837fc2de35fc7de7ce6add900 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 22 Apr 2010 19:18:10 +0000
Subject: [PATCH 202/532] Fix compilation with WITHOUT_CRYPT or WITHOUT_OPENSSL
 options.

Reported by:	Andrei V. Lavreniyuk <andy.lavr@reactor-xg.kiev.ua>
MFC after:	3 days
---
 sbin/hastctl/Makefile   |  9 +++++++--
 sbin/hastd/Makefile     | 10 +++++++---
 sbin/hastd/hast_proto.c |  8 ++++++++
 3 files changed, 22 insertions(+), 5 deletions(-)

diff --git a/sbin/hastctl/Makefile b/sbin/hastctl/Makefile
index e0d6b5b7223..301493cec79 100644
--- a/sbin/hastctl/Makefile
+++ b/sbin/hastctl/Makefile
@@ -25,8 +25,13 @@ CFLAGS+=-DINET6
 # This is needed to have WARNS > 1.
 CFLAGS+=-DYY_NO_UNPUT
 
-DPADD=	${LIBCRYPTO} ${LIBL}
-LDADD=	-lcrypto -ll
+DPADD=	${LIBL}
+LDADD=	-ll
+.if ${MK_OPENSSL} != "no"
+DPADD+=	${LIBCRYPTO}
+LDADD+=	-lcrypto
+CFLAGS+=-DHAVE_CRYPTO
+.endif
 
 YFLAGS+=-v
 
diff --git a/sbin/hastd/Makefile b/sbin/hastd/Makefile
index 8153c3a0f4b..bfb62f4f11f 100644
--- a/sbin/hastd/Makefile
+++ b/sbin/hastd/Makefile
@@ -26,9 +26,13 @@ CFLAGS+=-DINET6
 # This is needed to have WARNS > 1.
 CFLAGS+=-DYY_NO_UNPUT
 
-DPADD=	${LIBCRYPTO} ${LIBGEOM} ${LIBBSDXML} ${LIBSBUF} ${LIBL} \
-	${LIBPTHREAD} ${LIBUTIL}
-LDADD=	-lcrypto -lgeom -lbsdxml -lsbuf -ll -lpthread -lutil
+DPADD=	${LIBGEOM} ${LIBBSDXML} ${LIBSBUF} ${LIBL} ${LIBPTHREAD} ${LIBUTIL}
+LDADD=	-lgeom -lbsdxml -lsbuf -ll -lpthread -lutil
+.if ${MK_OPENSSL} != "no"
+DPADD+=	${LIBCRYPTO}
+LDADD+=	-lcrypto
+CFLAGS+=-DHAVE_CRYPTO
+.endif
 
 YFLAGS+=-v
 
diff --git a/sbin/hastd/hast_proto.c b/sbin/hastd/hast_proto.c
index 6e660069ef9..348dfc887bb 100644
--- a/sbin/hastd/hast_proto.c
+++ b/sbin/hastd/hast_proto.c
@@ -37,7 +37,9 @@ __FBSDID("$FreeBSD$");
 #include <string.h>
 #include <strings.h>
 
+#ifdef HAVE_CRYPTO
 #include <openssl/sha.h>
+#endif
 
 #include <hast.h>
 #include <ebuf.h>
@@ -67,14 +69,18 @@ static int compression_send(struct hast_resource *res, struct nv *nv,
     void **datap, size_t *sizep, bool *freedatap);
 static int compression_recv(struct hast_resource *res, struct nv *nv,
     void **datap, size_t *sizep, bool *freedatap);
+#ifdef HAVE_CRYPTO
 static int checksum_send(struct hast_resource *res, struct nv *nv,
     void **datap, size_t *sizep, bool *freedatap);
 static int checksum_recv(struct hast_resource *res, struct nv *nv,
     void **datap, size_t *sizep, bool *freedatap);
+#endif
 
 static struct hast_pipe_stage pipeline[] = {
 	{ "compression", compression_send, compression_recv },
+#ifdef HAVE_CRYPTO
 	{ "checksum", checksum_send, checksum_recv }
+#endif
 };
 
 static int
@@ -161,6 +167,7 @@ compression_recv(struct hast_resource *res, struct nv *nv, void **datap,
 	return (0);
 }
 
+#ifdef HAVE_CRYPTO
 static int
 checksum_send(struct hast_resource *res, struct nv *nv, void **datap,
     size_t *sizep, bool *freedatap __unused)
@@ -221,6 +228,7 @@ checksum_recv(struct hast_resource *res, struct nv *nv, void **datap,
 
 	return (0);
 }
+#endif	/* HAVE_CRYPTO */
 
 /*
  * Send the given nv structure via conn.

From a1a667ecf1788ec552a5c5c8f4d83786d764eaa1 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 22 Apr 2010 20:25:07 +0000
Subject: [PATCH 203/532] Intialize interrupt moderation control register. The
 magic value was chosen by lots of trial and errors. The chosen value shows
 good interrupt moderation without additional latency. Without this change,
 controller can generate more than 140k interrupts per second under high
 network load.

Submitted by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index c4fb5c9bc8b..6c737bd3f42 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1551,10 +1551,13 @@ sge_init_locked(struct sge_softc *sc)
 	/*
 	 * XXX Try to mitigate interrupts.
 	 */
+	CSR_WRITE_4(sc, IntrControl, 0x08880000);
+#ifdef notyet
 	if (sc->sge_intrcontrol != 0)
 		CSR_WRITE_4(sc, IntrControl, sc->sge_intrcontrol);
 	if (sc->sge_intrtimer != 0)
 		CSR_WRITE_4(sc, IntrTimer, sc->sge_intrtimer);
+#endif
 
 	/*
 	 * Clear and enable interrupts.

From 139e9ad4d642e178015ae0af21af73edb32fc009 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 22 Apr 2010 20:46:39 +0000
Subject: [PATCH 204/532] Document the new USB notification types.

---
 sbin/devd/devd.conf.5 | 56 ++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 55 insertions(+), 1 deletion(-)

diff --git a/sbin/devd/devd.conf.5 b/sbin/devd/devd.conf.5
index f7efefb20eb..17c5e099b56 100644
--- a/sbin/devd/devd.conf.5
+++ b/sbin/devd/devd.conf.5
@@ -250,18 +250,40 @@ CIS-vendor.
 Device class.
 .It Li device
 Device ID.
+.It Li devclass
+Device Class (USB)
+.It Li devsubclass
+Device Sub-class (USB)
 .It Li device-name
 Name of attached/detached device.
+.It Li endpoints
+Endpoint count (USB)
 .It Li function
 Card functions.
+.It Li interface
+Interface ID (USB)
+.It Li intclass
+Interface Class (USB)
+.It Li intprotocol
+Interface Protocol  (USB)
+.It Li intsubclass
+Interface Sub-class (USB)
 .It Li manufacturer
 Manufacturer ID (pccard).
+.It Li mode
+Peripheral mode (USB)
 .It Li notify
 Match the value of the
 .Dq Li notify
 variable.
+.It Li parent
+Parent device
+.It Li port
+Hub port number (USB)
 .It Li product
-Product ID (pccard).
+Product ID (pccard/USB).
+.It Li release
+Hardware revision (USB)
 .It Li serial
 Serial Number (USB).
 .It Li slot
@@ -342,6 +364,27 @@ The
 node is destroyed.
 .El
 .El
+.It Li USB
+Events related to the USB subsystem.
+.Bl -tag -width ".Sy Subsystem" -compact
+.It Sy Subsystem
+.It Li DEVICE
+.Bl -tag -width ".Li DETACH" -compact
+.It Sy Type
+.It Li ATTACH
+USB device is attached to the system.
+.It Li DETACH
+USB device is detached from the system.
+.El
+.It Li INTERFACE
+.Bl -tag -width ".Li DETACH" -compact
+.It Sy Type
+.It Li ATTACH
+USB interface is attached from a device.
+.It Li DETACH
+USB interface is detached from a device.
+.El
+.El
 .It Li coretemp
 Events related to the
 .Xr coretemp 4
@@ -460,6 +503,17 @@ notify 0 {
 	action "logger Lid opened, the sleeper must awaken!";
 };
 
+#
+# Match a USB device type
+#
+notify 0 {
+	match "system"			"USB";
+	match "subsystem"		"INTERFACE";
+	match "type"			"ATTACH";
+	match "intclass"		"0x0e";
+	action "logger USB video device attached";
+};
+
 #
 # Try to configure ath and wi devices with pccard_ether
 # as they are attached.

From b850ecc180aa37c044f17d6e3c964691cfc30514 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 22 Apr 2010 21:31:34 +0000
Subject: [PATCH 205/532] Change USB_DEBUG to #ifdef and allow it to be turned
 off. Previously this had the illusion of a tunable setting but was always
 turned on regardless.

MFC after:	1 week
---
 sys/amd64/conf/GENERIC             |  1 +
 sys/arm/conf/DB-78XXX              |  1 +
 sys/arm/conf/DB-88F5XXX            |  1 +
 sys/arm/conf/DB-88F6XXX            |  1 +
 sys/arm/conf/HL200                 |  1 +
 sys/arm/conf/KB920X                |  1 +
 sys/arm/conf/LN2410SBC             |  1 +
 sys/arm/conf/SHEEVAPLUG            |  1 +
 sys/dev/sound/usb/uaudio.c         | 22 ++++++++---------
 sys/dev/usb/controller/ehci.c      | 38 +++++++++++++++---------------
 sys/dev/usb/controller/uhci.c      | 34 +++++++++++++-------------
 sys/dev/usb/controller/uss820dci.c |  2 +-
 sys/dev/usb/input/atp.c            | 12 +++++-----
 sys/dev/usb/input/uhid.c           |  2 +-
 sys/dev/usb/input/ukbd.c           |  6 ++---
 sys/dev/usb/input/ums.c            |  6 +++--
 sys/dev/usb/misc/udbp.c            |  2 +-
 sys/dev/usb/net/if_aue.c           |  2 +-
 sys/dev/usb/net/if_axe.c           |  2 +-
 sys/dev/usb/net/if_cdce.c          |  2 +-
 sys/dev/usb/net/if_cue.c           |  2 +-
 sys/dev/usb/net/if_kue.c           |  2 +-
 sys/dev/usb/net/if_rue.c           |  2 +-
 sys/dev/usb/net/if_udav.c          |  2 +-
 sys/dev/usb/serial/u3g.c           |  2 +-
 sys/dev/usb/serial/ubsa.c          |  5 ++--
 sys/dev/usb/serial/ubser.c         |  2 +-
 sys/dev/usb/serial/uchcom.c        |  2 +-
 sys/dev/usb/serial/uftdi.c         |  2 +-
 sys/dev/usb/serial/ulpt.c          |  2 +-
 sys/dev/usb/serial/umodem.c        |  2 +-
 sys/dev/usb/serial/umoscom.c       |  2 +-
 sys/dev/usb/serial/uplcom.c        |  2 +-
 sys/dev/usb/serial/usb_serial.c    |  2 +-
 sys/dev/usb/serial/uslcom.c        |  2 +-
 sys/dev/usb/serial/uvisor.c        |  4 ++--
 sys/dev/usb/serial/uvscom.c        |  2 +-
 sys/dev/usb/storage/umass.c        |  8 +++----
 sys/dev/usb/storage/urio.c         |  2 +-
 sys/dev/usb/storage/ustorage_fs.c  |  2 +-
 sys/dev/usb/usb_debug.h            |  2 +-
 sys/dev/usb/usb_freebsd.h          |  4 ----
 sys/dev/usb/usb_request.c          | 12 +++++-----
 sys/dev/usb/usb_transfer.c         |  2 +-
 sys/dev/usb/wlan/if_rum.c          |  2 +-
 sys/dev/usb/wlan/if_run.c          |  2 +-
 sys/dev/usb/wlan/if_ural.c         |  2 +-
 sys/dev/usb/wlan/if_zyd.c          |  2 +-
 sys/i386/conf/GENERIC              |  1 +
 sys/i386/conf/XBOX                 |  1 +
 sys/ia64/conf/GENERIC              |  1 +
 sys/mips/conf/SENTRY5              |  1 +
 sys/mips/conf/XLR                  |  1 +
 sys/pc98/conf/GENERIC              |  1 +
 sys/powerpc/conf/GENERIC           |  1 +
 sys/powerpc/conf/MPC85XX           |  1 +
 sys/sparc64/conf/GENERIC           |  1 +
 sys/sun4v/conf/GENERIC             |  1 +
 58 files changed, 121 insertions(+), 106 deletions(-)

diff --git a/sys/amd64/conf/GENERIC b/sys/amd64/conf/GENERIC
index f6d0a600bf2..ddfb1ab387b 100644
--- a/sys/amd64/conf/GENERIC
+++ b/sys/amd64/conf/GENERIC
@@ -281,6 +281,7 @@ device		firmware	# firmware assist module
 device		bpf		# Berkeley packet filter
 
 # USB support
+#options 	USB_DEBUG	# enable debug msgs
 device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		ehci		# EHCI PCI->USB interface (USB 2.0)
diff --git a/sys/arm/conf/DB-78XXX b/sys/arm/conf/DB-78XXX
index 939ae2b96c5..b82f1185087 100644
--- a/sys/arm/conf/DB-78XXX
+++ b/sys/arm/conf/DB-78XXX
@@ -67,6 +67,7 @@ device		e1000phy
 device		bpf
 
 # USB
+options 	USB_DEBUG	# enable debug msgs
 device		usb
 device		ehci
 device		umass
diff --git a/sys/arm/conf/DB-88F5XXX b/sys/arm/conf/DB-88F5XXX
index 41f32b1586a..edb5fe8c252 100644
--- a/sys/arm/conf/DB-88F5XXX
+++ b/sys/arm/conf/DB-88F5XXX
@@ -74,6 +74,7 @@ device		iicbus
 device		ds133x
 
 # USB
+options 	USB_DEBUG	# enable debug msgs
 device		usb
 device		ehci
 device		umass
diff --git a/sys/arm/conf/DB-88F6XXX b/sys/arm/conf/DB-88F6XXX
index c0857fda624..1b4b7fd5ca5 100644
--- a/sys/arm/conf/DB-88F6XXX
+++ b/sys/arm/conf/DB-88F6XXX
@@ -67,6 +67,7 @@ device		e1000phy
 device		bpf
 
 # USB
+options 	USB_DEBUG	# enable debug msgs
 device		usb
 device		ehci
 device		umass
diff --git a/sys/arm/conf/HL200 b/sys/arm/conf/HL200
index f50f64cd622..dd46a61c2fd 100644
--- a/sys/arm/conf/HL200
+++ b/sys/arm/conf/HL200
@@ -95,6 +95,7 @@ device		icee
 
 device		bpf
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 device		ohci		# OHCI localbus->USB interface
 device		usb		# USB Bus (required)
 #device		udbp		# USB Double Bulk Pipe devices
diff --git a/sys/arm/conf/KB920X b/sys/arm/conf/KB920X
index 336228df419..f47e9acf09b 100644
--- a/sys/arm/conf/KB920X
+++ b/sys/arm/conf/KB920X
@@ -96,6 +96,7 @@ device		icee
 
 device		bpf
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 device		ohci		# OHCI localbus->USB interface
 device		usb		# USB Bus (required)
 #device		udbp		# USB Double Bulk Pipe devices
diff --git a/sys/arm/conf/LN2410SBC b/sys/arm/conf/LN2410SBC
index 5a14f591c8c..e4f3a5445c6 100644
--- a/sys/arm/conf/LN2410SBC
+++ b/sys/arm/conf/LN2410SBC
@@ -79,6 +79,7 @@ options 	WITNESS_SKIPSPIN	#Don't run witness on spinlocks for speed
 
 device		md
 
+options 	USB_DEBUG	# enable debug msgs
 device		usb
 device		ohci
 device		umass
diff --git a/sys/arm/conf/SHEEVAPLUG b/sys/arm/conf/SHEEVAPLUG
index d11b34d57b3..dcd82e5549d 100644
--- a/sys/arm/conf/SHEEVAPLUG
+++ b/sys/arm/conf/SHEEVAPLUG
@@ -61,6 +61,7 @@ options		DEVICE_POLLING
 device		vlan
 
 # USB
+options 	USB_DEBUG	# enable debug msgs
 device		usb
 device		ehci
 device		umass
diff --git a/sys/dev/sound/usb/uaudio.c b/sys/dev/sound/usb/uaudio.c
index 069a8c04887..16132667d5a 100644
--- a/sys/dev/sound/usb/uaudio.c
+++ b/sys/dev/sound/usb/uaudio.c
@@ -91,7 +91,7 @@ static int uaudio_default_rate = 0;		/* use rate list */
 static int uaudio_default_bits = 32;
 static int uaudio_default_channels = 0;		/* use default */
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uaudio_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uaudio, CTLFLAG_RW, 0, "USB uaudio");
@@ -321,7 +321,7 @@ static const struct uaudio_format uaudio_formats[] = {
 #define	UAC_RECORD	3
 #define	UAC_NCLASSES	4
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static const char *uac_names[] = {
 	"outputs", "inputs", "equalization", "record"
 };
@@ -406,7 +406,7 @@ static void	umidi_init(device_t dev);
 static int32_t	umidi_probe(device_t dev);
 static int32_t	umidi_detach(device_t dev);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void	uaudio_chan_dump_ep_desc(
 		    const usb_endpoint_descriptor_audio_t *);
 static void	uaudio_mixer_dump_cluster(uint8_t,
@@ -780,7 +780,7 @@ uaudio_detach(device_t dev)
  * AS - Audio Stream - routines
  *========================================================================*/
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void
 uaudio_chan_dump_ep_desc(const usb_endpoint_descriptor_audio_t *ed)
 {
@@ -1019,7 +1019,7 @@ uaudio_chan_fill_info_sub(struct uaudio_softc *sc, struct usb_device *udev,
 				if ((chan->valid == 0) && usbd_get_iface(udev, curidx)) {
 
 					chan->valid = 1;
-#if USB_DEBUG
+#ifdef USB_DEBUG
 					uaudio_chan_dump_ep_desc(ed1);
 					uaudio_chan_dump_ep_desc(ed2);
 
@@ -1689,7 +1689,7 @@ uaudio_mixer_add_ctl(struct uaudio_softc *sc, struct uaudio_mixer_node *mc)
 
 	uaudio_mixer_add_ctl_sub(sc, mc);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uaudio_debug > 2) {
 		uint8_t i;
 
@@ -1708,7 +1708,7 @@ static void
 uaudio_mixer_add_input(struct uaudio_softc *sc,
     const struct uaudio_terminal_node *iot, int id)
 {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	const struct usb_audio_input_terminal *d = iot[id].u.it;
 
 	DPRINTFN(3, "bTerminalId=%d wTerminalType=0x%04x "
@@ -1724,7 +1724,7 @@ static void
 uaudio_mixer_add_output(struct uaudio_softc *sc,
     const struct uaudio_terminal_node *iot, int id)
 {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	const struct usb_audio_output_terminal *d = iot[id].u.ot;
 
 	DPRINTFN(3, "bTerminalId=%d wTerminalType=0x%04x "
@@ -2257,7 +2257,7 @@ error:
 	return (NULL);
 }
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void
 uaudio_mixer_dump_cluster(uint8_t id, const struct uaudio_terminal_node *iot)
 {
@@ -2350,7 +2350,7 @@ done:
 	return (r);
 }
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 
 struct uaudio_tt_to_string {
 	uint16_t terminal_type;
@@ -2856,7 +2856,7 @@ uaudio_mixer_fill_info(struct uaudio_softc *sc, struct usb_device *udev,
 		(iot + i)->root = iot;
 	} while (i--);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	i = ID_max;
 	do {
 		uint8_t j;
diff --git a/sys/dev/usb/controller/ehci.c b/sys/dev/usb/controller/ehci.c
index 28ad987b31d..12cfe53680c 100644
--- a/sys/dev/usb/controller/ehci.c
+++ b/sys/dev/usb/controller/ehci.c
@@ -89,7 +89,7 @@ __FBSDID("$FreeBSD$");
    ((ehci_softc_t *)(((uint8_t *)(bus)) - \
     ((uint8_t *)&(((ehci_softc_t *)0)->sc_bus))))
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ehcidebug = 0;
 static int ehcinohighspeed = 0;
 static int ehciiaadbug = 0;
@@ -258,7 +258,7 @@ ehci_init(ehci_softc_t *sc)
 	usb_callout_init_mtx(&sc->sc_tmo_pcd, &sc->sc_bus.bus_mtx, 0);
 	usb_callout_init_mtx(&sc->sc_tmo_poll, &sc->sc_bus.bus_mtx, 0);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ehciiaadbug)
 		sc->sc_flags |= EHCI_SCFLG_IAADBUG;
 	if (ehcilostintrbug)
@@ -486,7 +486,7 @@ ehci_init(ehci_softc_t *sc)
 
 	usb_bus_mem_flush_all(&sc->sc_bus, &ehci_iterate_hw_softc);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ehcidebug) {
 		ehci_dump_sqh(sc, sc->sc_async_p_last);
 	}
@@ -685,7 +685,7 @@ ehci_shutdown(ehci_softc_t *sc)
 	}
 }
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void
 ehci_dump_regs(ehci_softc_t *sc)
 {
@@ -1229,7 +1229,7 @@ ehci_non_isoc_done_sub(struct usb_xfer *xfer)
 
 	xfer->td_transfer_cache = td;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (status & EHCI_QTD_STATERRS) {
 		DPRINTFN(11, "error, addr=%d, endpt=0x%02x, frame=0x%02x"
 		    "status=%s%s%s%s%s%s%s%s\n",
@@ -1260,7 +1260,7 @@ ehci_non_isoc_done(struct usb_xfer *xfer)
 	DPRINTFN(13, "xfer=%p endpoint=%p transfer done\n",
 	    xfer, xfer->endpoint);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ehcidebug > 10) {
 		ehci_softc_t *sc = EHCI_BUS2SC(xfer->xroot->bus);
 
@@ -1527,7 +1527,7 @@ ehci_interrupt(ehci_softc_t *sc)
 
 	DPRINTFN(16, "real interrupt\n");
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ehcidebug > 15) {
 		ehci_dump_regs(sc);
 	}
@@ -1548,7 +1548,7 @@ ehci_interrupt(ehci_softc_t *sc)
 	if (status & EHCI_STS_HSE) {
 		printf("%s: unrecoverable error, "
 		    "controller halted\n", __FUNCTION__);
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		ehci_dump_regs(sc);
 		ehci_dump_isoc(sc);
 #endif
@@ -1978,7 +1978,7 @@ ehci_setup_standard_chain(struct usb_xfer *xfer, ehci_qh_t **qh_last)
 
 	xfer->td_transfer_last = td;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ehcidebug > 8) {
 		DPRINTF("nexttog=%d; data before transfer:\n",
 		    xfer->endpoint->toggle_next);
@@ -2106,7 +2106,7 @@ ehci_isoc_fs_done(ehci_softc_t *sc, struct usb_xfer *xfer)
 		if (pp_last >= &sc->sc_isoc_fs_p_last[EHCI_VIRTUAL_FRAMELIST_COUNT]) {
 			pp_last = &sc->sc_isoc_fs_p_last[0];
 		}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (ehcidebug > 15) {
 			DPRINTF("isoc FS-TD\n");
 			ehci_dump_sitd(sc, td);
@@ -2160,7 +2160,7 @@ ehci_isoc_hs_done(ehci_softc_t *sc, struct usb_xfer *xfer)
 		if (pp_last >= &sc->sc_isoc_hs_p_last[EHCI_VIRTUAL_FRAMELIST_COUNT]) {
 			pp_last = &sc->sc_isoc_hs_p_last[0];
 		}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (ehcidebug > 15) {
 			DPRINTF("isoc HS-TD\n");
 			ehci_dump_itd(sc, td);
@@ -2224,7 +2224,7 @@ ehci_device_done(struct usb_xfer *xfer, usb_error_t error)
 
 	if ((methods == &ehci_device_bulk_methods) ||
 	    (methods == &ehci_device_ctrl_methods)) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (ehcidebug > 8) {
 			DPRINTF("nexttog=%d; data after transfer:\n",
 			    xfer->endpoint->toggle_next);
@@ -2509,7 +2509,7 @@ ehci_device_isoc_fs_enter(struct usb_xfer *xfer)
 	uint8_t sb;
 	uint8_t error;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	uint8_t once = 1;
 
 #endif
@@ -2593,7 +2593,7 @@ ehci_device_isoc_fs_enter(struct usb_xfer *xfer)
 		/* reuse sitd_portaddr and sitd_back from last transfer */
 
 		if (*plen > xfer->max_frame_size) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (once) {
 				once = 0;
 				printf("%s: frame length(%d) exceeds %d "
@@ -2683,7 +2683,7 @@ ehci_device_isoc_fs_enter(struct usb_xfer *xfer)
 		}
 		usb_pc_cpu_flush(td->page_cache);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (ehcidebug > 15) {
 			DPRINTF("FS-TD %d\n", nframes);
 			ehci_dump_sitd(sc, td);
@@ -2800,7 +2800,7 @@ ehci_device_isoc_hs_enter(struct usb_xfer *xfer)
 	uint8_t td_no;
 	uint8_t page_no;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	uint8_t once = 1;
 
 #endif
@@ -2878,7 +2878,7 @@ ehci_device_isoc_hs_enter(struct usb_xfer *xfer)
 		}
 		/* range check */
 		if (*plen > xfer->max_frame_size) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (once) {
 				once = 0;
 				printf("%s: frame length(%d) exceeds %d bytes "
@@ -2962,7 +2962,7 @@ ehci_device_isoc_hs_enter(struct usb_xfer *xfer)
 				td->itd_status[td_no - 1] |= htohc32(sc, EHCI_ITD_IOC);
 			}
 			usb_pc_cpu_flush(td->page_cache);
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (ehcidebug > 15) {
 				DPRINTF("HS-TD %d\n", nframes);
 				ehci_dump_itd(sc, td);
@@ -3398,7 +3398,7 @@ ehci_roothub_exec(struct usb_device *udev,
 			break;
 		case UHF_PORT_RESET:
 			DPRINTFN(6, "reset port %d\n", index);
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (ehcinohighspeed) {
 				/*
 				 * Connect USB device to companion
diff --git a/sys/dev/usb/controller/uhci.c b/sys/dev/usb/controller/uhci.c
index 837a26f16a1..f87907c4925 100644
--- a/sys/dev/usb/controller/uhci.c
+++ b/sys/dev/usb/controller/uhci.c
@@ -82,7 +82,7 @@ __FBSDID("$FreeBSD$");
    ((uhci_softc_t *)(((uint8_t *)(bus)) - \
     ((uint8_t *)&(((uhci_softc_t *)0)->sc_bus))))
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uhcidebug = 0;
 static int uhcinoloop = 0;
 
@@ -459,7 +459,7 @@ uhci_init(uhci_softc_t *sc)
 
 	usb_callout_init_mtx(&sc->sc_root_intr, &sc->sc_bus.bus_mtx, 0);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 2) {
 		uhci_dumpregs(sc);
 	}
@@ -668,7 +668,7 @@ uhci_suspend(uhci_softc_t *sc)
 {
 	USB_BUS_LOCK(&sc->sc_bus);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 2) {
 		uhci_dumpregs(sc);
 	}
@@ -712,7 +712,7 @@ uhci_resume(uhci_softc_t *sc)
 
 	uhci_start(sc);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 2) {
 		uhci_dumpregs(sc);
 	}
@@ -724,7 +724,7 @@ uhci_resume(uhci_softc_t *sc)
 	uhci_do_poll(&sc->sc_bus);
 }
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void
 uhci_dumpregs(uhci_softc_t *sc)
 {
@@ -855,7 +855,7 @@ uhci_add_loop(uhci_softc_t *sc)
 	struct uhci_qh *qh_lst;
 	struct uhci_qh *qh_rec;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcinoloop) {
 		return;
 	}
@@ -878,7 +878,7 @@ uhci_rem_loop(uhci_softc_t *sc)
 {
 	struct uhci_qh *qh_lst;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcinoloop) {
 		return;
 	}
@@ -1046,7 +1046,7 @@ uhci_isoc_done(uhci_softc_t *sc, struct usb_xfer *xfer)
 		if (pp_last >= &sc->sc_isoc_p_last[UHCI_VFRAMELIST_COUNT]) {
 			pp_last = &sc->sc_isoc_p_last[0];
 		}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (uhcidebug > 5) {
 			DPRINTF("isoc TD\n");
 			uhci_dump_td(td);
@@ -1177,7 +1177,7 @@ uhci_non_isoc_done_sub(struct usb_xfer *xfer)
 
 	xfer->endpoint->toggle_next = (token & UHCI_TD_SET_DT(1)) ? 0 : 1;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (status & UHCI_TD_ERROR) {
 		DPRINTFN(11, "error, addr=%d, endpt=0x%02x, frame=0x%02x "
 		    "status=%s%s%s%s%s%s%s%s%s%s%s\n",
@@ -1207,7 +1207,7 @@ uhci_non_isoc_done(struct usb_xfer *xfer)
 	DPRINTFN(13, "xfer=%p endpoint=%p transfer done\n",
 	    xfer, xfer->endpoint);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 10) {
 		uhci_dump_tds(xfer->td_transfer_first);
 	}
@@ -1446,7 +1446,7 @@ uhci_interrupt(uhci_softc_t *sc)
 
 	DPRINTFN(16, "real interrupt\n");
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 15) {
 		uhci_dumpregs(sc);
 	}
@@ -1460,7 +1460,7 @@ uhci_interrupt(uhci_softc_t *sc)
 	    UHCI_STS_HCPE | UHCI_STS_HCH)) {
 
 		if (status & UHCI_STS_RD) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			printf("%s: resume detect\n",
 			    __FUNCTION__);
 #endif
@@ -1477,7 +1477,7 @@ uhci_interrupt(uhci_softc_t *sc)
 			/* no acknowledge needed */
 			DPRINTF("%s: host controller halted\n",
 			    __FUNCTION__);
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (uhcidebug > 0) {
 				uhci_dump_all(sc);
 			}
@@ -1839,7 +1839,7 @@ uhci_setup_standard_chain(struct usb_xfer *xfer)
 
 	xfer->td_transfer_last = td;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (uhcidebug > 8) {
 		DPRINTF("nexttog=%d; data before transfer:\n",
 		    xfer->endpoint->toggle_next);
@@ -2155,7 +2155,7 @@ uhci_device_isoc_enter(struct usb_xfer *xfer)
 	uint32_t temp;
 	uint32_t *plen;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	uint8_t once = 1;
 
 #endif
@@ -2227,7 +2227,7 @@ uhci_device_isoc_enter(struct usb_xfer *xfer)
 			pp_last = &sc->sc_isoc_p_last[0];
 		}
 		if (*plen > xfer->max_frame_size) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			if (once) {
 				once = 0;
 				printf("%s: frame length(%d) exceeds %d "
@@ -2279,7 +2279,7 @@ uhci_device_isoc_enter(struct usb_xfer *xfer)
 
 		usb_pc_cpu_flush(td->page_cache);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (uhcidebug > 5) {
 			DPRINTF("TD %d\n", nframes);
 			uhci_dump_td(td);
diff --git a/sys/dev/usb/controller/uss820dci.c b/sys/dev/usb/controller/uss820dci.c
index c910c1f21dd..74221695f84 100644
--- a/sys/dev/usb/controller/uss820dci.c
+++ b/sys/dev/usb/controller/uss820dci.c
@@ -77,7 +77,7 @@
 #define	USS820_DCI_PC2SC(pc) \
    USS820_DCI_BUS2SC(USB_DMATAG_TO_XROOT((pc)->tag_parent)->bus)
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uss820dcidebug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uss820dci, CTLFLAG_RW, 0, "USB uss820dci");
diff --git a/sys/dev/usb/input/atp.c b/sys/dev/usb/input/atp.c
index 6c0ce2c7036..c0fe6d4b386 100644
--- a/sys/dev/usb/input/atp.c
+++ b/sys/dev/usb/input/atp.c
@@ -116,7 +116,7 @@ __FBSDID("$FreeBSD$");
 /* Tunables */
 SYSCTL_NODE(_hw_usb, OID_AUTO, atp, CTLFLAG_RW, 0, "USB atp");
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 enum atp_log_level {
 	ATP_LLEVEL_DISABLED = 0,
 	ATP_LLEVEL_ERROR,
@@ -126,7 +126,7 @@ enum atp_log_level {
 static int atp_debug = ATP_LLEVEL_ERROR; /* the default is to only log errors */
 SYSCTL_INT(_hw_usb_atp, OID_AUTO, debug, CTLFLAG_RW,
     &atp_debug, ATP_LLEVEL_ERROR, "ATP debug level");
-#endif /* #if USB_DEBUG */
+#endif /* USB_DEBUG */
 
 static u_int atp_touch_timeout = ATP_TOUCH_TIMEOUT;
 SYSCTL_INT(_hw_usb_atp, OID_AUTO, touch_timeout, CTLFLAG_RW, &atp_touch_timeout,
@@ -1055,7 +1055,7 @@ atp_update_strokes(struct atp_softc *sc, atp_pspan *pspans_x,
 		if (pspans_y[j].matched == FALSE) break;
 	}
 	if ((i < n_xpspans) && (j < n_ypspans)) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		if (atp_debug >= ATP_LLEVEL_INFO) {
 			printf("unmatched pspans:");
 			for (; i < n_xpspans; i++) {
@@ -1072,7 +1072,7 @@ atp_update_strokes(struct atp_softc *sc, atp_pspan *pspans_x,
 			}
 			printf("\n");
 		}
-#endif /* #if USB_DEBUG */
+#endif /* USB_DEBUG */
 		if ((n_xpspans == 1) && (n_ypspans == 1))
 			/* The common case of a single pair of new pspans. */
 			atp_add_stroke(sc, &pspans_x[0], &pspans_y[0]);
@@ -1082,7 +1082,7 @@ atp_update_strokes(struct atp_softc *sc, atp_pspan *pspans_x,
 			    pspans_y, n_ypspans);
 	}
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (atp_debug >= ATP_LLEVEL_INFO) {
 		for (i = 0; i < sc->sc_n_strokes; i++) {
 			atp_stroke *stroke = &sc->sc_strokes[i];
@@ -1110,7 +1110,7 @@ atp_update_strokes(struct atp_softc *sc, atp_pspan *pspans_x,
 		if (sc->sc_n_strokes)
 			printf("\n");
 	}
-#endif /* #if USB_DEBUG */
+#endif /* USB_DEBUG */
 
 	return (movement);
 }
diff --git a/sys/dev/usb/input/uhid.c b/sys/dev/usb/input/uhid.c
index a471991f62e..d36ecf36240 100644
--- a/sys/dev/usb/input/uhid.c
+++ b/sys/dev/usb/input/uhid.c
@@ -76,7 +76,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/input/usb_rdesc.h>
 #include <dev/usb/quirk/usb_quirk.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uhid_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uhid, CTLFLAG_RW, 0, "USB uhid");
diff --git a/sys/dev/usb/input/ukbd.c b/sys/dev/usb/input/ukbd.c
index f446db7a28c..034a781b2fe 100644
--- a/sys/dev/usb/input/ukbd.c
+++ b/sys/dev/usb/input/ukbd.c
@@ -88,7 +88,7 @@ __FBSDID("$FreeBSD$");
 /* the following file must be included after "ukbdmap.h" */
 #include <dev/kbd/kbdtables.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ukbd_debug = 0;
 static int ukbd_no_leds = 0;
 
@@ -612,7 +612,7 @@ ukbd_intr_callback(struct usb_xfer *xfer, usb_error_t error)
 				apple_fn = 1;
 			else
 				apple_fn = 0;
-#if USB_DEBUG
+#ifdef USB_DEBUG
 			DPRINTF("apple_eject=%u apple_fn=%u\n",
 			    apple_eject, apple_fn);
 
@@ -678,7 +678,7 @@ ukbd_set_leds_callback(struct usb_xfer *xfer, usb_error_t error)
 	uint8_t buf[2];
 	struct ukbd_softc *sc = usbd_xfer_softc(xfer);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (ukbd_no_leds)
 		return;
 #endif
diff --git a/sys/dev/usb/input/ums.c b/sys/dev/usb/input/ums.c
index a4385b95d11..e1e37becb81 100644
--- a/sys/dev/usb/input/ums.c
+++ b/sys/dev/usb/input/ums.c
@@ -74,7 +74,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/tty.h>
 #include <sys/mouse.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ums_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ums, CTLFLAG_RW, 0, "USB ums");
@@ -494,7 +494,9 @@ ums_attach(device_t dev)
 	int err;
 	uint16_t d_len;
 	uint8_t i;
+#ifdef USB_DEBUG
 	uint8_t j;
+#endif
 
 	DPRINTFN(11, "sc=%p\n", sc);
 
@@ -588,7 +590,7 @@ ums_attach(device_t dev)
 	free(d_ptr, M_TEMP);
 	d_ptr = NULL;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	for (j = 0; j < UMS_INFO_MAX; j++) {
 		info = &sc->sc_info[j];
 
diff --git a/sys/dev/usb/misc/udbp.c b/sys/dev/usb/misc/udbp.c
index ada6b3c0a29..5eef3103c31 100644
--- a/sys/dev/usb/misc/udbp.c
+++ b/sys/dev/usb/misc/udbp.c
@@ -94,7 +94,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/misc/udbp.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int udbp_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, udbp, CTLFLAG_RW, 0, "USB udbp");
diff --git a/sys/dev/usb/net/if_aue.c b/sys/dev/usb/net/if_aue.c
index b508474e468..a8c0a548b51 100644
--- a/sys/dev/usb/net/if_aue.c
+++ b/sys/dev/usb/net/if_aue.c
@@ -100,7 +100,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/net/usb_ethernet.h>
 #include <dev/usb/net/if_auereg.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int aue_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, aue, CTLFLAG_RW, 0, "USB aue");
diff --git a/sys/dev/usb/net/if_axe.c b/sys/dev/usb/net/if_axe.c
index 9772f401b68..e255c855c94 100644
--- a/sys/dev/usb/net/if_axe.c
+++ b/sys/dev/usb/net/if_axe.c
@@ -123,7 +123,7 @@ __FBSDID("$FreeBSD$");
  */
 #define AXE_178_MAX_FRAME_BURST	1
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int axe_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, axe, CTLFLAG_RW, 0, "USB axe");
diff --git a/sys/dev/usb/net/if_cdce.c b/sys/dev/usb/net/if_cdce.c
index 2fcb0ff0b5b..b5e7fd464bd 100644
--- a/sys/dev/usb/net/if_cdce.c
+++ b/sys/dev/usb/net/if_cdce.c
@@ -108,7 +108,7 @@ static uether_fn_t cdce_setpromisc;
 
 static uint32_t	cdce_m_crc32(struct mbuf *, uint32_t, uint32_t);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int cdce_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, cdce, CTLFLAG_RW, 0, "USB CDC-Ethernet");
diff --git a/sys/dev/usb/net/if_cue.c b/sys/dev/usb/net/if_cue.c
index 05ff1a54a88..e26b29f6fd2 100644
--- a/sys/dev/usb/net/if_cue.c
+++ b/sys/dev/usb/net/if_cue.c
@@ -122,7 +122,7 @@ static int	cue_getmac(struct cue_softc *, void *);
 static uint32_t	cue_mchash(const uint8_t *);
 static void	cue_reset(struct cue_softc *);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int cue_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, cue, CTLFLAG_RW, 0, "USB cue");
diff --git a/sys/dev/usb/net/if_kue.c b/sys/dev/usb/net/if_kue.c
index 5d35da40bd5..4eee0945d9f 100644
--- a/sys/dev/usb/net/if_kue.c
+++ b/sys/dev/usb/net/if_kue.c
@@ -163,7 +163,7 @@ static int	kue_ctl(struct kue_softc *, uint8_t, uint8_t, uint16_t,
 static int	kue_load_fw(struct kue_softc *);
 static void	kue_reset(struct kue_softc *);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int kue_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, kue, CTLFLAG_RW, 0, "USB kue");
diff --git a/sys/dev/usb/net/if_rue.c b/sys/dev/usb/net/if_rue.c
index 3e77305aa97..f0d16086b20 100644
--- a/sys/dev/usb/net/if_rue.c
+++ b/sys/dev/usb/net/if_rue.c
@@ -97,7 +97,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/net/usb_ethernet.h>
 #include <dev/usb/net/if_ruereg.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int rue_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, rue, CTLFLAG_RW, 0, "USB rue");
diff --git a/sys/dev/usb/net/if_udav.c b/sys/dev/usb/net/if_udav.c
index ae30b142da6..f56e9b03829 100644
--- a/sys/dev/usb/net/if_udav.c
+++ b/sys/dev/usb/net/if_udav.c
@@ -185,7 +185,7 @@ static const struct usb_ether_methods udav_ue_methods = {
 	.ue_mii_sts = udav_ifmedia_status,
 };
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int udav_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, udav, CTLFLAG_RW, 0, "USB udav");
diff --git a/sys/dev/usb/serial/u3g.c b/sys/dev/usb/serial/u3g.c
index 7ac1bd3310d..ffe88126dfa 100644
--- a/sys/dev/usb/serial/u3g.c
+++ b/sys/dev/usb/serial/u3g.c
@@ -63,7 +63,7 @@
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int u3g_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, u3g, CTLFLAG_RW, 0, "USB 3g");
diff --git a/sys/dev/usb/serial/ubsa.c b/sys/dev/usb/serial/ubsa.c
index 32639fcc403..58175094922 100644
--- a/sys/dev/usb/serial/ubsa.c
+++ b/sys/dev/usb/serial/ubsa.c
@@ -93,7 +93,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ubsa_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ubsa, CTLFLAG_RW, 0, "USB ubsa");
@@ -405,9 +405,8 @@ ubsa_cfg_set_break(struct ucom_softc *ucom, uint8_t onoff)
 static int
 ubsa_pre_param(struct ucom_softc *ucom, struct termios *t)
 {
-	struct ubsa_softc *sc = ucom->sc_parent;
 
-	DPRINTF("sc = %p\n", sc);
+	DPRINTF("sc = %p\n", ucom->sc_parent);
 
 	switch (t->c_ospeed) {
 	case B0:
diff --git a/sys/dev/usb/serial/ubser.c b/sys/dev/usb/serial/ubser.c
index 06c96c026e5..3f2dc2d30c7 100644
--- a/sys/dev/usb/serial/ubser.c
+++ b/sys/dev/usb/serial/ubser.c
@@ -115,7 +115,7 @@ __FBSDID("$FreeBSD$");
 #define	VENDOR_SET_BREAK		0x02
 #define	VENDOR_CLEAR_BREAK		0x03
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ubser_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ubser, CTLFLAG_RW, 0, "USB ubser");
diff --git a/sys/dev/usb/serial/uchcom.c b/sys/dev/usb/serial/uchcom.c
index 9fea8492fca..92f3a922345 100644
--- a/sys/dev/usb/serial/uchcom.c
+++ b/sys/dev/usb/serial/uchcom.c
@@ -101,7 +101,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uchcom_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uchcom, CTLFLAG_RW, 0, "USB uchcom");
diff --git a/sys/dev/usb/serial/uftdi.c b/sys/dev/usb/serial/uftdi.c
index 76a08de30a0..b9d1d347f29 100644
--- a/sys/dev/usb/serial/uftdi.c
+++ b/sys/dev/usb/serial/uftdi.c
@@ -73,7 +73,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/serial/usb_serial.h>
 #include <dev/usb/serial/uftdi_reg.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uftdi_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uftdi, CTLFLAG_RW, 0, "USB uftdi");
diff --git a/sys/dev/usb/serial/ulpt.c b/sys/dev/usb/serial/ulpt.c
index f053f6d3751..58dc367237b 100644
--- a/sys/dev/usb/serial/ulpt.c
+++ b/sys/dev/usb/serial/ulpt.c
@@ -72,7 +72,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/usb_debug.h>
 #include <dev/usb/usb_process.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ulpt_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ulpt, CTLFLAG_RW, 0, "USB ulpt");
diff --git a/sys/dev/usb/serial/umodem.c b/sys/dev/usb/serial/umodem.c
index 39afdad893f..a7d00c9d23a 100644
--- a/sys/dev/usb/serial/umodem.c
+++ b/sys/dev/usb/serial/umodem.c
@@ -116,7 +116,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int umodem_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, umodem, CTLFLAG_RW, 0, "USB umodem");
diff --git a/sys/dev/usb/serial/umoscom.c b/sys/dev/usb/serial/umoscom.c
index 0481b192c8b..3a36a44f71c 100644
--- a/sys/dev/usb/serial/umoscom.c
+++ b/sys/dev/usb/serial/umoscom.c
@@ -48,7 +48,7 @@
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int umoscom_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, umoscom, CTLFLAG_RW, 0, "USB umoscom");
diff --git a/sys/dev/usb/serial/uplcom.c b/sys/dev/usb/serial/uplcom.c
index c5d58e46dcb..ae88805bbbb 100644
--- a/sys/dev/usb/serial/uplcom.c
+++ b/sys/dev/usb/serial/uplcom.c
@@ -116,7 +116,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uplcom_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uplcom, CTLFLAG_RW, 0, "USB uplcom");
diff --git a/sys/dev/usb/serial/usb_serial.c b/sys/dev/usb/serial/usb_serial.c
index 871ae54190b..6573d8e0c2e 100644
--- a/sys/dev/usb/serial/usb_serial.c
+++ b/sys/dev/usb/serial/usb_serial.c
@@ -104,7 +104,7 @@ __FBSDID("$FreeBSD$");
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ucom, CTLFLAG_RW, 0, "USB ucom");
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ucom_debug = 0;
 
 SYSCTL_INT(_hw_usb_ucom, OID_AUTO, debug, CTLFLAG_RW,
diff --git a/sys/dev/usb/serial/uslcom.c b/sys/dev/usb/serial/uslcom.c
index d97cc2c542c..f20c12ac147 100644
--- a/sys/dev/usb/serial/uslcom.c
+++ b/sys/dev/usb/serial/uslcom.c
@@ -50,7 +50,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uslcom_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uslcom, CTLFLAG_RW, 0, "USB uslcom");
diff --git a/sys/dev/usb/serial/uvisor.c b/sys/dev/usb/serial/uvisor.c
index f0e4257a372..77ff31f5bd2 100644
--- a/sys/dev/usb/serial/uvisor.c
+++ b/sys/dev/usb/serial/uvisor.c
@@ -78,7 +78,7 @@
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uvisor_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uvisor, CTLFLAG_RW, 0, "USB uvisor");
@@ -392,7 +392,7 @@ uvisor_init(struct uvisor_softc *sc, struct usb_device *udev, struct usb_config
 			goto done;
 		}
 	}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (sc->sc_flag & UVISOR_FLAG_VISOR) {
 		uint16_t i, np;
 		const char *desc;
diff --git a/sys/dev/usb/serial/uvscom.c b/sys/dev/usb/serial/uvscom.c
index 4e3ff576422..f220587da12 100644
--- a/sys/dev/usb/serial/uvscom.c
+++ b/sys/dev/usb/serial/uvscom.c
@@ -68,7 +68,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/serial/usb_serial.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int uvscom_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, uvscom, CTLFLAG_RW, 0, "USB uvscom");
diff --git a/sys/dev/usb/storage/umass.c b/sys/dev/usb/storage/umass.c
index b14bb47e8cb..a66357b5db5 100644
--- a/sys/dev/usb/storage/umass.c
+++ b/sys/dev/usb/storage/umass.c
@@ -146,7 +146,7 @@ __FBSDID("$FreeBSD$");
 #define	UMASS_USB_FLAGS
 #endif
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 #define	DIF(m, x)				\
   do {						\
     if (umass_debug & (m)) { x ; }		\
@@ -488,7 +488,7 @@ static uint8_t	umass_no_transform(struct umass_softc *, uint8_t *, uint8_t);
 static uint8_t	umass_std_transform(struct umass_softc *, union ccb *, uint8_t
 		    *, uint8_t);
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void	umass_bbb_dump_cbw(struct umass_softc *, umass_bbb_cbw_t *);
 static void	umass_bbb_dump_csw(struct umass_softc *, umass_bbb_csw_t *);
 static void	umass_cbi_dump_cmd(struct umass_softc *, void *, uint8_t);
@@ -917,7 +917,7 @@ umass_attach(device_t dev)
 	}
 	sc->sc_iface_no = id->bInterfaceNumber;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	device_printf(dev, " ");
 
 	switch (sc->sc_proto & UMASS_PROTO_COMMAND) {
@@ -3012,7 +3012,7 @@ umass_std_transform(struct umass_softc *sc, union ccb *ccb,
 	return (1);
 }
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static void
 umass_bbb_dump_cbw(struct umass_softc *sc, umass_bbb_cbw_t *cbw)
 {
diff --git a/sys/dev/usb/storage/urio.c b/sys/dev/usb/storage/urio.c
index 403c4c2b216..1aef8469f43 100644
--- a/sys/dev/usb/storage/urio.c
+++ b/sys/dev/usb/storage/urio.c
@@ -78,7 +78,7 @@ __FBSDID("$FreeBSD$");
 
 #include <dev/usb/storage/rio500_usb.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int urio_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, urio, CTLFLAG_RW, 0, "USB urio");
diff --git a/sys/dev/usb/storage/ustorage_fs.c b/sys/dev/usb/storage/ustorage_fs.c
index 10047e11c22..52cfd6eb286 100644
--- a/sys/dev/usb/storage/ustorage_fs.c
+++ b/sys/dev/usb/storage/ustorage_fs.c
@@ -64,7 +64,7 @@
 #define	USB_DEBUG_VAR ustorage_fs_debug
 #include <dev/usb/usb_debug.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ustorage_fs_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ustorage_fs, CTLFLAG_RW, 0, "USB ustorage_fs");
diff --git a/sys/dev/usb/usb_debug.h b/sys/dev/usb/usb_debug.h
index b6bfbcfca1a..8718c89bef9 100644
--- a/sys/dev/usb/usb_debug.h
+++ b/sys/dev/usb/usb_debug.h
@@ -34,7 +34,7 @@ extern int usb_debug;
 
 /* Check if USB debugging is enabled. */
 #ifdef USB_DEBUG_VAR
-#if (USB_DEBUG != 0)
+#ifdef USB_DEBUG
 #define	DPRINTFN(n,fmt,...) do {		\
   if ((USB_DEBUG_VAR) >= (n)) {			\
     printf("%s: " fmt,				\
diff --git a/sys/dev/usb/usb_freebsd.h b/sys/dev/usb/usb_freebsd.h
index 1f34317d667..8a008cd47dd 100644
--- a/sys/dev/usb/usb_freebsd.h
+++ b/sys/dev/usb/usb_freebsd.h
@@ -57,10 +57,6 @@
 #define	USB_HUB_MAX_DEPTH	5
 #define	USB_EP0_BUFSIZE		1024	/* bytes */
 
-#ifndef USB_DEBUG
-#define USB_DEBUG 1
-#endif
-
 typedef uint32_t usb_timeout_t;		/* milliseconds */
 typedef uint32_t usb_frlength_t;	/* bytes */
 typedef uint32_t usb_frcount_t;		/* units */
diff --git a/sys/dev/usb/usb_request.c b/sys/dev/usb/usb_request.c
index 03745faa7a5..32c57788773 100644
--- a/sys/dev/usb/usb_request.c
+++ b/sys/dev/usb/usb_request.c
@@ -68,7 +68,7 @@
 #include <dev/usb/usb_bus.h>
 #include <sys/ctype.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int usb_pr_poll_delay = USB_PORT_RESET_DELAY;
 static int usb_pr_recovery_delay = USB_PORT_RESET_RECOVERY;
 static int usb_ss_delay = 0;
@@ -433,7 +433,7 @@ usbd_do_request_flags(struct usb_device *udev, struct mtx *mtx,
 		} else {
 			if (xfer->frlengths[0] == 0) {
 				if (xfer->flags.manual_status) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 					int temp;
 
 					temp = usb_ss_delay;
@@ -603,7 +603,7 @@ usbd_req_reset_port(struct usb_device *udev, struct mtx *mtx, uint8_t port)
 	usb_error_t err;
 	uint16_t n;
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	uint16_t pr_poll_delay;
 	uint16_t pr_recovery_delay;
 
@@ -612,7 +612,7 @@ usbd_req_reset_port(struct usb_device *udev, struct mtx *mtx, uint8_t port)
 	if (err) {
 		goto done;
 	}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	/* range check input parameters */
 	pr_poll_delay = usb_pr_poll_delay;
 	if (pr_poll_delay < 1) {
@@ -627,7 +627,7 @@ usbd_req_reset_port(struct usb_device *udev, struct mtx *mtx, uint8_t port)
 #endif
 	n = 0;
 	while (1) {
-#if USB_DEBUG
+#ifdef USB_DEBUG
 		/* wait for the device to recover from reset */
 		usb_pause_mtx(mtx, USB_MS_TO_TICKS(pr_poll_delay));
 		n += pr_poll_delay;
@@ -666,7 +666,7 @@ usbd_req_reset_port(struct usb_device *udev, struct mtx *mtx, uint8_t port)
 		err = USB_ERR_TIMEOUT;
 		goto done;
 	}
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	/* wait for the device to recover from reset */
 	usb_pause_mtx(mtx, USB_MS_TO_TICKS(pr_recovery_delay));
 #else
diff --git a/sys/dev/usb/usb_transfer.c b/sys/dev/usb/usb_transfer.c
index 9be1c0f4bde..dc39fc64b8e 100644
--- a/sys/dev/usb/usb_transfer.c
+++ b/sys/dev/usb/usb_transfer.c
@@ -1418,7 +1418,7 @@ usbd_transfer_submit(struct usb_xfer *xfer)
 	    xfer, xfer->endpoint, xfer->nframes, USB_GET_DATA_ISREAD(xfer) ?
 	    "read" : "write");
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 	if (USB_DEBUG_VAR > 0) {
 		USB_BUS_LOCK(bus);
 
diff --git a/sys/dev/usb/wlan/if_rum.c b/sys/dev/usb/wlan/if_rum.c
index 5b3441b0118..4a5fbae2779 100644
--- a/sys/dev/usb/wlan/if_rum.c
+++ b/sys/dev/usb/wlan/if_rum.c
@@ -77,7 +77,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/wlan/if_rumvar.h>
 #include <dev/usb/wlan/if_rumfw.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int rum_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, rum, CTLFLAG_RW, 0, "USB rum");
diff --git a/sys/dev/usb/wlan/if_run.c b/sys/dev/usb/wlan/if_run.c
index 06d3764ab54..3ab607fa342 100644
--- a/sys/dev/usb/wlan/if_run.c
+++ b/sys/dev/usb/wlan/if_run.c
@@ -79,7 +79,7 @@ __FBSDID("$FreeBSD$");
 
 #define nitems(_a)      (sizeof((_a)) / sizeof((_a)[0]))
 
-#if	USB_DEBUG
+#ifdef	USB_DEBUG
 #define RUN_DEBUG
 #endif
 
diff --git a/sys/dev/usb/wlan/if_ural.c b/sys/dev/usb/wlan/if_ural.c
index fc907bcb346..4a4378b90ef 100644
--- a/sys/dev/usb/wlan/if_ural.c
+++ b/sys/dev/usb/wlan/if_ural.c
@@ -78,7 +78,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/wlan/if_uralreg.h>
 #include <dev/usb/wlan/if_uralvar.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int ural_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, ural, CTLFLAG_RW, 0, "USB ural");
diff --git a/sys/dev/usb/wlan/if_zyd.c b/sys/dev/usb/wlan/if_zyd.c
index ee143d632ad..a0b8179a4a3 100644
--- a/sys/dev/usb/wlan/if_zyd.c
+++ b/sys/dev/usb/wlan/if_zyd.c
@@ -75,7 +75,7 @@ __FBSDID("$FreeBSD$");
 #include <dev/usb/wlan/if_zydreg.h>
 #include <dev/usb/wlan/if_zydfw.h>
 
-#if USB_DEBUG
+#ifdef USB_DEBUG
 static int zyd_debug = 0;
 
 SYSCTL_NODE(_hw_usb, OID_AUTO, zyd, CTLFLAG_RW, 0, "USB zyd");
diff --git a/sys/i386/conf/GENERIC b/sys/i386/conf/GENERIC
index 2bc296342c5..24f5aabeb9f 100644
--- a/sys/i386/conf/GENERIC
+++ b/sys/i386/conf/GENERIC
@@ -294,6 +294,7 @@ device		firmware	# firmware assist module
 device		bpf		# Berkeley packet filter
 
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		ehci		# EHCI PCI->USB interface (USB 2.0)
diff --git a/sys/i386/conf/XBOX b/sys/i386/conf/XBOX
index 96719659ed1..2ba5738c008 100644
--- a/sys/i386/conf/XBOX
+++ b/sys/i386/conf/XBOX
@@ -82,6 +82,7 @@ device		pty		# BSD-style compatibility pseudo ttys
 device		bpf		# Berkeley packet filter
 
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 #device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		usb		# USB Bus (required)
diff --git a/sys/ia64/conf/GENERIC b/sys/ia64/conf/GENERIC
index 9b53bf3f1f7..cf06a295860 100644
--- a/sys/ia64/conf/GENERIC
+++ b/sys/ia64/conf/GENERIC
@@ -116,6 +116,7 @@ device		ida		# Compaq Smart RAID
 device		mlx		# Mylex DAC960 family
 
 # USB host controllers and peripherals
+options 	USB_DEBUG	# enable debug msgs
 device		ehci		# EHCI host controller
 device		ohci		# OHCI PCI->USB interface
 device		uhci		# UHCI PCI->USB interface
diff --git a/sys/mips/conf/SENTRY5 b/sys/mips/conf/SENTRY5
index a4e32c07d83..c38e45f20a3 100644
--- a/sys/mips/conf/SENTRY5
+++ b/sys/mips/conf/SENTRY5
@@ -75,6 +75,7 @@ device		pci			# siba_pcib
 #device		ath_hal			# pci chip support
 #options		AH_SUPPORT_AR5416	# enable AR5416 tx/rx descriptors
 
+options 	USB_DEBUG		# enable debug msgs
 device		usb			# USB Bus (required)
 device		uhci			# UHCI PCI->USB interface
 device		ehci			# EHCI PCI->USB interface (USB 2.0)
diff --git a/sys/mips/conf/XLR b/sys/mips/conf/XLR
index d743c1b33f0..a0737b3b618 100644
--- a/sys/mips/conf/XLR
+++ b/sys/mips/conf/XLR
@@ -132,6 +132,7 @@ device          scbus
 #device          ohci            # OHCI PCI->USB interface
 device          ehci            # EHCI PCI->USB interface (USB 2.0)
 device          usb             # USB Bus (required)
+options 	USB_DEBUG	# enable debug msgs
 #device         udbp            # USB Double Bulk Pipe devices
 #device          ugen            # Generic
 #device          uhid            # "Human Interface Devices"
diff --git a/sys/pc98/conf/GENERIC b/sys/pc98/conf/GENERIC
index 85239cbdcc3..699091ecd75 100644
--- a/sys/pc98/conf/GENERIC
+++ b/sys/pc98/conf/GENERIC
@@ -247,6 +247,7 @@ device		firmware	# firmware assist module
 device		bpf		# Berkeley packet filter
 
 # USB support
+#options 	USB_DEBUG	# enable debug msgs
 #device		uhci		# UHCI PCI->USB interface
 #device		ohci		# OHCI PCI->USB interface
 #device		ehci		# EHCI PCI->USB interface (USB 2.0)
diff --git a/sys/powerpc/conf/GENERIC b/sys/powerpc/conf/GENERIC
index 5c76272f1e0..86f56cc1dc7 100644
--- a/sys/powerpc/conf/GENERIC
+++ b/sys/powerpc/conf/GENERIC
@@ -144,6 +144,7 @@ device		firmware	# firmware assist module
 device		bpf		#Berkeley packet filter
 
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		ehci		# EHCI PCI->USB interface
diff --git a/sys/powerpc/conf/MPC85XX b/sys/powerpc/conf/MPC85XX
index 51d4e8d5342..a2a8de4a9a3 100644
--- a/sys/powerpc/conf/MPC85XX
+++ b/sys/powerpc/conf/MPC85XX
@@ -79,6 +79,7 @@ device		sec
 device		tsec
 device		tun
 device		uart
+options 	USB_DEBUG	# enable debug msgs
 #device		uhci
 device		umass
 device		usb
diff --git a/sys/sparc64/conf/GENERIC b/sys/sparc64/conf/GENERIC
index be3cfd3f9ab..9ce2d6c4cb8 100644
--- a/sys/sparc64/conf/GENERIC
+++ b/sys/sparc64/conf/GENERIC
@@ -226,6 +226,7 @@ device		firmware	# firmware assist module
 device		bpf		# Berkeley packet filter
 
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		ehci		# EHCI PCI->USB interface (USB 2.0)
diff --git a/sys/sun4v/conf/GENERIC b/sys/sun4v/conf/GENERIC
index 4c7cd5bf0c0..26ec3afa4ce 100644
--- a/sys/sun4v/conf/GENERIC
+++ b/sys/sun4v/conf/GENERIC
@@ -183,6 +183,7 @@ device		faith		# IPv6-to-IPv4 relaying (translation)
 device		bpf		# Berkeley packet filter
 
 # USB support
+options 	USB_DEBUG	# enable debug msgs
 #device		uhci		# UHCI PCI->USB interface
 #device		ohci		# OHCI PCI->USB interface
 device		usb		# USB Bus (required)

From 0d92ac2a9240a76521908970dd2ea0406c973ba6 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 22 Apr 2010 21:41:50 +0000
Subject: [PATCH 206/532] Use SX_DUPOK rather than making the string unique.

---
 sys/dev/usb/usb_device.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/usb/usb_device.c b/sys/dev/usb/usb_device.c
index 2f3a402d4ac..bc386474907 100644
--- a/sys/dev/usb/usb_device.c
+++ b/sys/dev/usb/usb_device.c
@@ -1491,10 +1491,10 @@ usb_alloc_device(device_t parent_dev, struct usb_bus *bus,
 		return (NULL);
 	}
 	/* initialise our SX-lock */
-	sx_init(udev->default_sx, "0123456789ABCDEF - USB device SX lock" + depth);
+	sx_init_flags(udev->default_sx, "USB device SX lock", SX_DUPOK);
 
 	/* initialise our SX-lock */
-	sx_init(udev->default_sx + 1, "0123456789ABCDEF - USB config SX lock" + depth);
+	sx_init_flags(udev->default_sx + 1, "USB config SX lock", SX_DUPOK);
 
 	cv_init(udev->default_cv, "WCTRL");
 	cv_init(udev->default_cv + 1, "UGONE");

From 91cd92400fc6fa107e8b2b60140d2675e9f78020 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 22 Apr 2010 22:00:16 +0000
Subject: [PATCH 207/532] Properly name the sxlocks, mutexes and condvars.

---
 sys/dev/usb/usb_dev.c      |  8 ++++----
 sys/dev/usb/usb_device.c   | 38 +++++++++++++++++++-------------------
 sys/dev/usb/usb_device.h   |  8 +++++---
 sys/dev/usb/usb_request.c  |  8 ++++----
 sys/dev/usb/usb_transfer.c |  2 +-
 5 files changed, 33 insertions(+), 31 deletions(-)

diff --git a/sys/dev/usb/usb_dev.c b/sys/dev/usb/usb_dev.c
index dffabad29b9..a696a2652be 100644
--- a/sys/dev/usb/usb_dev.c
+++ b/sys/dev/usb/usb_dev.c
@@ -284,7 +284,7 @@ error:
 		usbd_enum_unlock(cpd->udev);
 
 		if (--(cpd->udev->refcount) == 0) {
-			cv_signal(cpd->udev->default_cv + 1);
+			cv_signal(&cpd->udev->ref_cv);
 		}
 	}
 	mtx_unlock(&usb_ref_lock);
@@ -352,7 +352,7 @@ usb_unref_device(struct usb_cdev_privdata *cpd,
 	}
 	if (crd->is_uref) {
 		if (--(cpd->udev->refcount) == 0) {
-			cv_signal(cpd->udev->default_cv + 1);
+			cv_signal(&cpd->udev->ref_cv);
 		}
 		crd->is_uref = 0;
 	}
@@ -500,7 +500,7 @@ usb_fifo_create(struct usb_cdev_privdata *cpd,
 		/* update some fields */
 		f->fifo_index = n + USB_FIFO_TX;
 		f->dev_ep_index = e;
-		f->priv_mtx = udev->default_mtx;
+		f->priv_mtx = &udev->device_mtx;
 		f->priv_sc0 = ep;
 		f->methods = &usb_ugen_methods;
 		f->iface_index = ep->iface_index;
@@ -527,7 +527,7 @@ usb_fifo_create(struct usb_cdev_privdata *cpd,
 		/* update some fields */
 		f->fifo_index = n + USB_FIFO_RX;
 		f->dev_ep_index = e;
-		f->priv_mtx = udev->default_mtx;
+		f->priv_mtx = &udev->device_mtx;
 		f->priv_sc0 = ep;
 		f->methods = &usb_ugen_methods;
 		f->iface_index = ep->iface_index;
diff --git a/sys/dev/usb/usb_device.c b/sys/dev/usb/usb_device.c
index bc386474907..98fbb908bdd 100644
--- a/sys/dev/usb/usb_device.c
+++ b/sys/dev/usb/usb_device.c
@@ -655,7 +655,7 @@ usb_config_parse(struct usb_device *udev, uint8_t iface_index, uint8_t cmd)
 		goto cleanup;
 
 	if (cmd == USB_CFG_INIT) {
-		sx_assert(udev->default_sx + 1, SA_LOCKED);
+		sx_assert(&udev->enum_sx, SA_LOCKED);
 
 		/* check for in-use endpoints */
 
@@ -1062,7 +1062,7 @@ usb_detach_device(struct usb_device *udev, uint8_t iface_index,
 	}
 	DPRINTFN(4, "udev=%p\n", udev);
 
-	sx_assert(udev->default_sx + 1, SA_LOCKED);
+	sx_assert(&udev->enum_sx, SA_LOCKED);
 
 	/*
 	 * First detach the child to give the child's detach routine a
@@ -1380,7 +1380,7 @@ usb_suspend_resume(struct usb_device *udev, uint8_t do_suspend)
 	}
 	DPRINTFN(4, "udev=%p do_suspend=%d\n", udev, do_suspend);
 
-	sx_assert(udev->default_sx + 1, SA_LOCKED);
+	sx_assert(&udev->enum_sx, SA_LOCKED);
 
 	USB_BUS_LOCK(udev->bus);
 	/* filter the suspend events */
@@ -1419,13 +1419,13 @@ usbd_clear_stall_proc(struct usb_proc_msg *_pm)
 
 	/* Change lock */
 	USB_BUS_UNLOCK(udev->bus);
-	mtx_lock(udev->default_mtx);
+	mtx_lock(&udev->device_mtx);
 
 	/* Start clear stall callback */
 	usbd_transfer_start(udev->default_xfer[1]);
 
 	/* Change lock */
-	mtx_unlock(udev->default_mtx);
+	mtx_unlock(&udev->device_mtx);
 	USB_BUS_LOCK(udev->bus);
 }
 
@@ -1491,16 +1491,16 @@ usb_alloc_device(device_t parent_dev, struct usb_bus *bus,
 		return (NULL);
 	}
 	/* initialise our SX-lock */
-	sx_init_flags(udev->default_sx, "USB device SX lock", SX_DUPOK);
+	sx_init_flags(&udev->ctrl_sx, "USB device SX lock", SX_DUPOK);
 
 	/* initialise our SX-lock */
-	sx_init_flags(udev->default_sx + 1, "USB config SX lock", SX_DUPOK);
+	sx_init_flags(&udev->enum_sx, "USB config SX lock", SX_DUPOK);
 
-	cv_init(udev->default_cv, "WCTRL");
-	cv_init(udev->default_cv + 1, "UGONE");
+	cv_init(&udev->ctrlreq_cv, "WCTRL");
+	cv_init(&udev->ref_cv, "UGONE");
 
 	/* initialise our mutex */
-	mtx_init(udev->default_mtx, "USB device mutex", NULL, MTX_DEF);
+	mtx_init(&udev->device_mtx, "USB device mutex", NULL, MTX_DEF);
 
 	/* initialise generic clear stall */
 	udev->cs_msg[0].hdr.pm_callback = &usbd_clear_stall_proc;
@@ -2005,7 +2005,7 @@ usb_free_device(struct usb_device *udev, uint8_t flag)
 	mtx_lock(&usb_ref_lock);
 	udev->refcount--;
 	while (udev->refcount != 0) {
-		cv_wait(udev->default_cv + 1, &usb_ref_lock);
+		cv_wait(&udev->ref_cv, &usb_ref_lock);
 	}
 	mtx_unlock(&usb_ref_lock);
 
@@ -2036,13 +2036,13 @@ usb_free_device(struct usb_device *udev, uint8_t flag)
 	    &udev->cs_msg[0], &udev->cs_msg[1]);
 	USB_BUS_UNLOCK(udev->bus);
 
-	sx_destroy(udev->default_sx);
-	sx_destroy(udev->default_sx + 1);
+	sx_destroy(&udev->ctrl_sx);
+	sx_destroy(&udev->enum_sx);
 
-	cv_destroy(udev->default_cv);
-	cv_destroy(udev->default_cv + 1);
+	cv_destroy(&udev->ctrlreq_cv);
+	cv_destroy(&udev->ref_cv);
 
-	mtx_destroy(udev->default_mtx);
+	mtx_destroy(&udev->device_mtx);
 #if USB_HAVE_UGEN
 	KASSERT(LIST_FIRST(&udev->pd_list) == NULL, ("leaked cdev entries"));
 #endif
@@ -2588,7 +2588,7 @@ usbd_device_attached(struct usb_device *udev)
 void
 usbd_enum_lock(struct usb_device *udev)
 {
-	sx_xlock(udev->default_sx + 1);
+	sx_xlock(&udev->enum_sx);
 	/* 
 	 * NEWBUS LOCK NOTE: We should check if any parent SX locks
 	 * are locked before locking Giant. Else the lock can be
@@ -2603,7 +2603,7 @@ void
 usbd_enum_unlock(struct usb_device *udev)
 {
 	mtx_unlock(&Giant);
-	sx_xunlock(udev->default_sx + 1);
+	sx_xunlock(&udev->enum_sx);
 }
 
 /*
@@ -2614,5 +2614,5 @@ usbd_enum_unlock(struct usb_device *udev)
 uint8_t
 usbd_enum_is_locked(struct usb_device *udev)
 {
-	return (sx_xlocked(udev->default_sx + 1));
+	return (sx_xlocked(&udev->enum_sx));
 }
diff --git a/sys/dev/usb/usb_device.h b/sys/dev/usb/usb_device.h
index 3afdecf3ae9..f497949f09c 100644
--- a/sys/dev/usb/usb_device.h
+++ b/sys/dev/usb/usb_device.h
@@ -113,9 +113,11 @@ struct usb_power_save {
 struct usb_device {
 	struct usb_clear_stall_msg cs_msg[2];	/* generic clear stall
 						 * messages */
-	struct sx default_sx[2];
-	struct mtx default_mtx[1];
-	struct cv default_cv[2];
+	struct sx ctrl_sx;
+	struct sx enum_sx;
+	struct mtx device_mtx;
+	struct cv ctrlreq_cv;
+	struct cv ref_cv;
 	struct usb_interface *ifaces;
 	struct usb_endpoint default_ep;	/* Control Endpoint 0 */
 	struct usb_endpoint *endpoints;
diff --git a/sys/dev/usb/usb_request.c b/sys/dev/usb/usb_request.c
index 32c57788773..a5ae56babee 100644
--- a/sys/dev/usb/usb_request.c
+++ b/sys/dev/usb/usb_request.c
@@ -99,7 +99,7 @@ usbd_do_request_callback(struct usb_xfer *xfer, usb_error_t error)
 		usbd_transfer_submit(xfer);
 		break;
 	default:
-		cv_signal(xfer->xroot->udev->default_cv);
+		cv_signal(&xfer->xroot->udev->ctrlreq_cv);
 		break;
 	}
 }
@@ -319,7 +319,7 @@ usbd_do_request_flags(struct usb_device *udev, struct mtx *mtx,
 	 * is achieved when multiple threads are involved:
 	 */
 
-	sx_xlock(udev->default_sx);
+	sx_xlock(&udev->ctrl_sx);
 
 	hr_func = usbd_get_hr_func(udev);
 
@@ -457,7 +457,7 @@ usbd_do_request_flags(struct usb_device *udev, struct mtx *mtx,
 		usbd_transfer_start(xfer);
 
 		while (usbd_transfer_pending(xfer)) {
-			cv_wait(udev->default_cv,
+			cv_wait(&udev->ctrlreq_cv,
 			    xfer->xroot->xfer_mtx);
 		}
 
@@ -534,7 +534,7 @@ usbd_do_request_flags(struct usb_device *udev, struct mtx *mtx,
 	USB_XFER_UNLOCK(xfer);
 
 done:
-	sx_xunlock(udev->default_sx);
+	sx_xunlock(&udev->ctrl_sx);
 
 	if (mtx) {
 		mtx_lock(mtx);
diff --git a/sys/dev/usb/usb_transfer.c b/sys/dev/usb/usb_transfer.c
index dc39fc64b8e..ce32f72e6d3 100644
--- a/sys/dev/usb/usb_transfer.c
+++ b/sys/dev/usb/usb_transfer.c
@@ -2821,7 +2821,7 @@ repeat:
 	iface_index = 0;
 	if (usbd_transfer_setup(udev, &iface_index,
 	    udev->default_xfer, usb_control_ep_cfg, USB_DEFAULT_XFER_MAX, NULL,
-	    udev->default_mtx)) {
+	    &udev->device_mtx)) {
 		DPRINTFN(0, "could not setup default "
 		    "USB transfer\n");
 	} else {

From 5b3bb704f7310e9437574b71cb5232f2f5392b48 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 22 Apr 2010 22:15:08 +0000
Subject: [PATCH 208/532] Use a more obvious prefix for the USB control
 (endpoint 0) transfers rather than default_*.

---
 sys/dev/usb/usb_compat_linux.c |  2 +-
 sys/dev/usb/usb_dev.c          |  2 +-
 sys/dev/usb/usb_device.c       | 46 +++++++++++++++++-----------------
 sys/dev/usb/usb_device.h       | 10 ++++----
 sys/dev/usb/usb_generic.c      | 18 ++++++-------
 sys/dev/usb/usb_hub.c          |  2 +-
 sys/dev/usb/usb_request.c      |  4 +--
 sys/dev/usb/usb_transfer.c     | 24 +++++++++---------
 sys/dev/usb/usb_transfer.h     |  2 +-
 9 files changed, 55 insertions(+), 55 deletions(-)

diff --git a/sys/dev/usb/usb_compat_linux.c b/sys/dev/usb/usb_compat_linux.c
index fc28a1c5060..89aaa8f7aac 100644
--- a/sys/dev/usb/usb_compat_linux.c
+++ b/sys/dev/usb/usb_compat_linux.c
@@ -971,7 +971,7 @@ usb_linux_create_usb_device(struct usb_device *udev, device_t dev)
 			udev->devnum = device_get_unit(dev);
 			bcopy(&udev->ddesc, &udev->descriptor,
 			    sizeof(udev->descriptor));
-			bcopy(udev->default_ep.edesc, &udev->ep0.desc,
+			bcopy(udev->ctrl_ep.edesc, &udev->ep0.desc,
 			    sizeof(udev->ep0.desc));
 		}
 	}
diff --git a/sys/dev/usb/usb_dev.c b/sys/dev/usb/usb_dev.c
index a696a2652be..2ac5b2e4edf 100644
--- a/sys/dev/usb/usb_dev.c
+++ b/sys/dev/usb/usb_dev.c
@@ -615,7 +615,7 @@ usb_dev_get_ep(struct usb_device *udev, uint8_t ep_index, uint8_t dir)
 	uint8_t ep_dir;
 
 	if (ep_index == 0) {
-		ep = &udev->default_ep;
+		ep = &udev->ctrl_ep;
 	} else {
 		if (dir == USB_FIFO_RX) {
 			if (udev->flags.usb_mode == USB_MODE_HOST) {
diff --git a/sys/dev/usb/usb_device.c b/sys/dev/usb/usb_device.c
index 98fbb908bdd..95020de0ac3 100644
--- a/sys/dev/usb/usb_device.c
+++ b/sys/dev/usb/usb_device.c
@@ -179,9 +179,9 @@ usbd_get_ep_by_addr(struct usb_device *udev, uint8_t ea_val)
 	/*
 	 * The default endpoint is always present and is checked separately:
 	 */
-	if ((udev->default_ep.edesc) &&
-	    ((udev->default_ep.edesc->bEndpointAddress & EA_MASK) == ea_val)) {
-		ep = &udev->default_ep;
+	if ((udev->ctrl_ep.edesc) &&
+	    ((udev->ctrl_ep.edesc->bEndpointAddress & EA_MASK) == ea_val)) {
+		ep = &udev->ctrl_ep;
 		goto found;
 	}
 	return (NULL);
@@ -297,11 +297,11 @@ usbd_get_endpoint(struct usb_device *udev, uint8_t iface_index,
 	 * address" and "any direction" returns the first endpoint of the
 	 * interface. "iface_index" and "direction" is ignored:
 	 */
-	if ((udev->default_ep.edesc) &&
-	    ((udev->default_ep.edesc->bEndpointAddress & ea_mask) == ea_val) &&
-	    ((udev->default_ep.edesc->bmAttributes & type_mask) == type_val) &&
+	if ((udev->ctrl_ep.edesc) &&
+	    ((udev->ctrl_ep.edesc->bEndpointAddress & ea_mask) == ea_val) &&
+	    ((udev->ctrl_ep.edesc->bmAttributes & type_mask) == type_val) &&
 	    (!index)) {
-		ep = &udev->default_ep;
+		ep = &udev->ctrl_ep;
 		goto found;
 	}
 	return (NULL);
@@ -1422,7 +1422,7 @@ usbd_clear_stall_proc(struct usb_proc_msg *_pm)
 	mtx_lock(&udev->device_mtx);
 
 	/* Start clear stall callback */
-	usbd_transfer_start(udev->default_xfer[1]);
+	usbd_transfer_start(udev->ctrl_xfer[1]);
 
 	/* Change lock */
 	mtx_unlock(&udev->device_mtx);
@@ -1529,13 +1529,13 @@ usb_alloc_device(device_t parent_dev, struct usb_bus *bus,
 	udev->refcount = 1;
 
 	/* set up default endpoint descriptor */
-	udev->default_ep_desc.bLength = sizeof(udev->default_ep_desc);
-	udev->default_ep_desc.bDescriptorType = UDESC_ENDPOINT;
-	udev->default_ep_desc.bEndpointAddress = USB_CONTROL_ENDPOINT;
-	udev->default_ep_desc.bmAttributes = UE_CONTROL;
-	udev->default_ep_desc.wMaxPacketSize[0] = USB_MAX_IPACKET;
-	udev->default_ep_desc.wMaxPacketSize[1] = 0;
-	udev->default_ep_desc.bInterval = 0;
+	udev->ctrl_ep_desc.bLength = sizeof(udev->ctrl_ep_desc);
+	udev->ctrl_ep_desc.bDescriptorType = UDESC_ENDPOINT;
+	udev->ctrl_ep_desc.bEndpointAddress = USB_CONTROL_ENDPOINT;
+	udev->ctrl_ep_desc.bmAttributes = UE_CONTROL;
+	udev->ctrl_ep_desc.wMaxPacketSize[0] = USB_MAX_IPACKET;
+	udev->ctrl_ep_desc.wMaxPacketSize[1] = 0;
+	udev->ctrl_ep_desc.bInterval = 0;
 	udev->ddesc.bMaxPacketSize = USB_MAX_IPACKET;
 
 	udev->speed = speed;
@@ -1559,8 +1559,8 @@ usb_alloc_device(device_t parent_dev, struct usb_bus *bus,
 
 	/* init the default endpoint */
 	usb_init_endpoint(udev, 0,
-	    &udev->default_ep_desc,
-	    &udev->default_ep);
+	    &udev->ctrl_ep_desc,
+	    &udev->ctrl_ep);
 
 	/* set device index */
 	udev->device_index = device_index;
@@ -1573,10 +1573,10 @@ usb_alloc_device(device_t parent_dev, struct usb_bus *bus,
 	LIST_INIT(&udev->pd_list);
 
 	/* Create the control endpoint device */
-	udev->default_dev = usb_make_dev(udev, 0, FREAD|FWRITE);
+	udev->ctrl_dev = usb_make_dev(udev, 0, FREAD|FWRITE);
 
 	/* Create a link from /dev/ugenX.X to the default endpoint */
-	make_dev_alias(udev->default_dev, udev->ugen_name);
+	make_dev_alias(udev->ctrl_dev, udev->ugen_name);
 #endif
 	if (udev->flags.usb_mode == USB_MODE_HOST) {
 
@@ -2009,20 +2009,20 @@ usb_free_device(struct usb_device *udev, uint8_t flag)
 	}
 	mtx_unlock(&usb_ref_lock);
 
-	destroy_dev_sched_cb(udev->default_dev, usb_cdev_cleanup,
-	    udev->default_dev->si_drv1);
+	destroy_dev_sched_cb(udev->ctrl_dev, usb_cdev_cleanup,
+	    udev->ctrl_dev->si_drv1);
 #endif
 
 	if (udev->flags.usb_mode == USB_MODE_DEVICE) {
 		/* stop receiving any control transfers (Device Side Mode) */
-		usbd_transfer_unsetup(udev->default_xfer, USB_DEFAULT_XFER_MAX);
+		usbd_transfer_unsetup(udev->ctrl_xfer, USB_CTRL_XFER_MAX);
 	}
 
 	/* the following will get the device unconfigured in software */
 	usb_unconfigure(udev, USB_UNCFG_FLAG_FREE_EP0);
 
 	/* unsetup any leftover default USB transfers */
-	usbd_transfer_unsetup(udev->default_xfer, USB_DEFAULT_XFER_MAX);
+	usbd_transfer_unsetup(udev->ctrl_xfer, USB_CTRL_XFER_MAX);
 
 	/* template unsetup, if any */
 	(usb_temp_unsetup_p) (udev);
diff --git a/sys/dev/usb/usb_device.h b/sys/dev/usb/usb_device.h
index f497949f09c..08b9fd7ef4f 100644
--- a/sys/dev/usb/usb_device.h
+++ b/sys/dev/usb/usb_device.h
@@ -30,7 +30,7 @@
 struct usb_symlink;		/* UGEN */
 struct usb_device;		/* linux compat */
 
-#define	USB_DEFAULT_XFER_MAX 2
+#define	USB_CTRL_XFER_MAX 2
 
 /* "usb_parse_config()" commands */
 
@@ -119,7 +119,7 @@ struct usb_device {
 	struct cv ctrlreq_cv;
 	struct cv ref_cv;
 	struct usb_interface *ifaces;
-	struct usb_endpoint default_ep;	/* Control Endpoint 0 */
+	struct usb_endpoint ctrl_ep;	/* Control Endpoint 0 */
 	struct usb_endpoint *endpoints;
 	struct usb_power_save pwr_save;/* power save data */
 	struct usb_bus *bus;		/* our USB BUS */
@@ -128,13 +128,13 @@ struct usb_device {
 	struct usb_device *parent_hs_hub;	/* high-speed parent HUB */
 	struct usb_config_descriptor *cdesc;	/* full config descr */
 	struct usb_hub *hub;		/* only if this is a hub */
-	struct usb_xfer *default_xfer[USB_DEFAULT_XFER_MAX];
+	struct usb_xfer *ctrl_xfer[USB_CTRL_XFER_MAX];
 	struct usb_temp_data *usb_template_ptr;
 	struct usb_endpoint *ep_curr;	/* current clear stall endpoint */
 #if USB_HAVE_UGEN
 	struct usb_fifo *fifo[USB_FIFO_MAX];
 	struct usb_symlink *ugen_symlink;	/* our generic symlink */
-	struct cdev *default_dev;	/* Control Endpoint 0 device node */
+	struct cdev *ctrl_dev;	/* Control Endpoint 0 device node */
 	LIST_HEAD(,usb_fs_privdata) pd_list;
 	char	ugen_name[20];		/* name of ugenX.X device */
 #endif
@@ -166,7 +166,7 @@ struct usb_device {
 
 	struct usb_device_flags flags;
 
-	struct usb_endpoint_descriptor default_ep_desc;	/* for endpoint 0 */
+	struct usb_endpoint_descriptor ctrl_ep_desc;	/* for endpoint 0 */
 	struct usb_device_descriptor ddesc;	/* device descriptor */
 
 	char	*serial;		/* serial number */
diff --git a/sys/dev/usb/usb_generic.c b/sys/dev/usb/usb_generic.c
index fb7d5df0fcb..db274683440 100644
--- a/sys/dev/usb/usb_generic.c
+++ b/sys/dev/usb/usb_generic.c
@@ -81,11 +81,11 @@
 
 static usb_callback_t ugen_read_clear_stall_callback;
 static usb_callback_t ugen_write_clear_stall_callback;
-static usb_callback_t ugen_default_read_callback;
-static usb_callback_t ugen_default_write_callback;
+static usb_callback_t ugen_ctrl_read_callback;
+static usb_callback_t ugen_ctrl_write_callback;
 static usb_callback_t ugen_isoc_read_callback;
 static usb_callback_t ugen_isoc_write_callback;
-static usb_callback_t ugen_default_fs_callback;
+static usb_callback_t ugen_ctrl_fs_callback;
 
 static usb_fifo_open_t ugen_open;
 static usb_fifo_close_t ugen_close;
@@ -265,7 +265,7 @@ ugen_open_pipe_write(struct usb_fifo *f)
 		if (f->flag_short) {
 			usb_config[0].flags.force_short_xfer = 1;
 		}
-		usb_config[0].callback = &ugen_default_write_callback;
+		usb_config[0].callback = &ugen_ctrl_write_callback;
 		usb_config[0].timeout = f->timeout;
 		usb_config[0].frames = 1;
 		usb_config[0].bufsize = f->bufsize;
@@ -335,7 +335,7 @@ ugen_open_pipe_read(struct usb_fifo *f)
 		}
 		usb_config[0].timeout = f->timeout;
 		usb_config[0].frames = 1;
-		usb_config[0].callback = &ugen_default_read_callback;
+		usb_config[0].callback = &ugen_ctrl_read_callback;
 		usb_config[0].bufsize = f->bufsize;
 
 		if (ugen_transfer_setup(f, usb_config, 2)) {
@@ -401,7 +401,7 @@ ugen_stop_io(struct usb_fifo *f)
 }
 
 static void
-ugen_default_read_callback(struct usb_xfer *xfer, usb_error_t error)
+ugen_ctrl_read_callback(struct usb_xfer *xfer, usb_error_t error)
 {
 	struct usb_fifo *f = usbd_xfer_softc(xfer);
 	struct usb_mbuf *m;
@@ -453,7 +453,7 @@ ugen_default_read_callback(struct usb_xfer *xfer, usb_error_t error)
 }
 
 static void
-ugen_default_write_callback(struct usb_xfer *xfer, usb_error_t error)
+ugen_ctrl_write_callback(struct usb_xfer *xfer, usb_error_t error)
 {
 	struct usb_fifo *f = usbd_xfer_softc(xfer);
 	usb_frlength_t actlen;
@@ -1480,7 +1480,7 @@ ugen_ioctl(struct usb_fifo *f, u_long cmd, void *addr, int fflags)
 		usb_config[0].direction = ed->bEndpointAddress & (UE_DIR_OUT | UE_DIR_IN);
 		usb_config[0].interval = USB_DEFAULT_INTERVAL;
 		usb_config[0].flags.proxy_buffer = 1;
-		usb_config[0].callback = &ugen_default_fs_callback;
+		usb_config[0].callback = &ugen_ctrl_fs_callback;
 		usb_config[0].timeout = 0;	/* no timeout */
 		usb_config[0].frames = u.popen->max_frames;
 		usb_config[0].bufsize = u.popen->max_bufsize;
@@ -2201,7 +2201,7 @@ ugen_ioctl_post(struct usb_fifo *f, u_long cmd, void *addr, int fflags)
 }
 
 static void
-ugen_default_fs_callback(struct usb_xfer *xfer, usb_error_t error)
+ugen_ctrl_fs_callback(struct usb_xfer *xfer, usb_error_t error)
 {
 	;				/* workaround for a bug in "indent" */
 
diff --git a/sys/dev/usb/usb_hub.c b/sys/dev/usb/usb_hub.c
index 97797963269..c31f8fa449c 100644
--- a/sys/dev/usb/usb_hub.c
+++ b/sys/dev/usb/usb_hub.c
@@ -246,7 +246,7 @@ uhub_explore_sub(struct uhub_softc *sc, struct usb_port *up)
 	/* start control transfer, if device mode */
 
 	if (child->flags.usb_mode == USB_MODE_DEVICE) {
-		usbd_default_transfer_setup(child);
+		usbd_ctrl_transfer_setup(child);
 	}
 	/* if a HUB becomes present, do a recursive HUB explore */
 
diff --git a/sys/dev/usb/usb_request.c b/sys/dev/usb/usb_request.c
index a5ae56babee..6150cb1fa14 100644
--- a/sys/dev/usb/usb_request.c
+++ b/sys/dev/usb/usb_request.c
@@ -374,9 +374,9 @@ usbd_do_request_flags(struct usb_device *udev, struct mtx *mtx,
 	/*
 	 * Setup a new USB transfer or use the existing one, if any:
 	 */
-	usbd_default_transfer_setup(udev);
+	usbd_ctrl_transfer_setup(udev);
 
-	xfer = udev->default_xfer[0];
+	xfer = udev->ctrl_xfer[0];
 	if (xfer == NULL) {
 		/* most likely out of memory */
 		err = USB_ERR_NOMEM;
diff --git a/sys/dev/usb/usb_transfer.c b/sys/dev/usb/usb_transfer.c
index ce32f72e6d3..535d12bfdef 100644
--- a/sys/dev/usb/usb_transfer.c
+++ b/sys/dev/usb/usb_transfer.c
@@ -72,7 +72,7 @@ struct usb_std_packet_size {
 
 static usb_callback_t usb_request_callback;
 
-static const struct usb_config usb_control_ep_cfg[USB_DEFAULT_XFER_MAX] = {
+static const struct usb_config usb_control_ep_cfg[USB_CTRL_XFER_MAX] = {
 
 	/* This transfer is used for generic control endpoint transfers */
 
@@ -2433,8 +2433,8 @@ usbd_pipe_start(struct usb_xfer_queue *pq)
 			if (udev->flags.usb_mode == USB_MODE_DEVICE) {
 				(udev->bus->methods->set_stall) (
 				    udev, NULL, ep, &did_stall);
-			} else if (udev->default_xfer[1]) {
-				info = udev->default_xfer[1]->xroot;
+			} else if (udev->ctrl_xfer[1]) {
+				info = udev->ctrl_xfer[1]->xroot;
 				usb_proc_msignal(
 				    &info->bus->non_giant_callback_proc,
 				    &udev->cs_msg[0], &udev->cs_msg[1]);
@@ -2757,13 +2757,13 @@ usb_command_wrapper(struct usb_xfer_queue *pq, struct usb_xfer *xfer)
 }
 
 /*------------------------------------------------------------------------*
- *	usbd_default_transfer_setup
+ *	usbd_ctrl_transfer_setup
  *
  * This function is used to setup the default USB control endpoint
  * transfer.
  *------------------------------------------------------------------------*/
 void
-usbd_default_transfer_setup(struct usb_device *udev)
+usbd_ctrl_transfer_setup(struct usb_device *udev)
 {
 	struct usb_xfer *xfer;
 	uint8_t no_resetup;
@@ -2774,12 +2774,12 @@ usbd_default_transfer_setup(struct usb_device *udev)
 		return;
 repeat:
 
-	xfer = udev->default_xfer[0];
+	xfer = udev->ctrl_xfer[0];
 	if (xfer) {
 		USB_XFER_LOCK(xfer);
 		no_resetup =
 		    ((xfer->address == udev->address) &&
-		    (udev->default_ep_desc.wMaxPacketSize[0] ==
+		    (udev->ctrl_ep_desc.wMaxPacketSize[0] ==
 		    udev->ddesc.bMaxPacketSize));
 		if (udev->flags.usb_mode == USB_MODE_DEVICE) {
 			if (no_resetup) {
@@ -2806,13 +2806,13 @@ repeat:
 	/*
 	 * Update wMaxPacketSize for the default control endpoint:
 	 */
-	udev->default_ep_desc.wMaxPacketSize[0] =
+	udev->ctrl_ep_desc.wMaxPacketSize[0] =
 	    udev->ddesc.bMaxPacketSize;
 
 	/*
 	 * Unsetup any existing USB transfer:
 	 */
-	usbd_transfer_unsetup(udev->default_xfer, USB_DEFAULT_XFER_MAX);
+	usbd_transfer_unsetup(udev->ctrl_xfer, USB_CTRL_XFER_MAX);
 
 	/*
 	 * Try to setup a new USB transfer for the
@@ -2820,7 +2820,7 @@ repeat:
 	 */
 	iface_index = 0;
 	if (usbd_transfer_setup(udev, &iface_index,
-	    udev->default_xfer, usb_control_ep_cfg, USB_DEFAULT_XFER_MAX, NULL,
+	    udev->ctrl_xfer, usb_control_ep_cfg, USB_CTRL_XFER_MAX, NULL,
 	    &udev->device_mtx)) {
 		DPRINTFN(0, "could not setup default "
 		    "USB transfer\n");
@@ -3001,13 +3001,13 @@ usbd_transfer_poll(struct usb_xfer **ppxfer, uint16_t max)
 		USB_BUS_LOCK(xroot->bus);
 
 		/* check for clear stall */
-		if (udev->default_xfer[1] != NULL) {
+		if (udev->ctrl_xfer[1] != NULL) {
 
 			/* poll clear stall start */
 			pm = &udev->cs_msg[0].hdr;
 			(pm->pm_callback) (pm);
 			/* poll clear stall done thread */
-			pm = &udev->default_xfer[1]->
+			pm = &udev->ctrl_xfer[1]->
 			    xroot->done_m[0].hdr;
 			(pm->pm_callback) (pm);
 		}
diff --git a/sys/dev/usb/usb_transfer.h b/sys/dev/usb/usb_transfer.h
index 27f3afffa63..6e08df09ef8 100644
--- a/sys/dev/usb/usb_transfer.h
+++ b/sys/dev/usb/usb_transfer.h
@@ -123,7 +123,7 @@ void	usbd_transfer_done(struct usb_xfer *xfer, usb_error_t error);
 void	usbd_transfer_enqueue(struct usb_xfer_queue *pq,
 	    struct usb_xfer *xfer);
 void	usbd_transfer_setup_sub(struct usb_setup_params *parm);
-void	usbd_default_transfer_setup(struct usb_device *udev);
+void	usbd_ctrl_transfer_setup(struct usb_device *udev);
 void	usbd_clear_data_toggle(struct usb_device *udev,
 	    struct usb_endpoint *ep);
 usb_callback_t usbd_do_request_callback;

From b834123032063d5b86ac7d4ee5506a75a27eacc1 Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Thu, 22 Apr 2010 23:47:19 +0000
Subject: [PATCH 209/532] If a conditional jump instruction has the same jt and
 jf, do not perform the test and jump unconditionally.

---
 sys/amd64/amd64/bpf_jit_machdep.c | 35 ++++++++++++++++++++++---------
 sys/amd64/amd64/bpf_jit_machdep.h |  6 ++++++
 sys/i386/i386/bpf_jit_machdep.c   | 35 ++++++++++++++++++++++---------
 sys/i386/i386/bpf_jit_machdep.h   |  6 ++++++
 4 files changed, 62 insertions(+), 20 deletions(-)

diff --git a/sys/amd64/amd64/bpf_jit_machdep.c b/sys/amd64/amd64/bpf_jit_machdep.c
index 6a5793eb81a..fe861d29227 100644
--- a/sys/amd64/amd64/bpf_jit_machdep.c
+++ b/sys/amd64/amd64/bpf_jit_machdep.c
@@ -419,62 +419,77 @@ bpf_jit_compile(struct bpf_insn *prog, u_int nins, size_t *size)
 				break;
 
 			case BPF_JMP|BPF_JA:
-				JMP(stream.refs[stream.bpf_pc + ins->k] -
-				    stream.refs[stream.bpf_pc]);
+				JUMP(ins->k);
 				break;
 
 			case BPF_JMP|BPF_JGT|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JA, JBE);
 				break;
 
 			case BPF_JMP|BPF_JGE|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JAE, JB);
 				break;
 
 			case BPF_JMP|BPF_JEQ|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JE, JNE);
 				break;
 
 			case BPF_JMP|BPF_JSET|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				TESTid(ins->k, EAX);
 				JCC(JNE, JE);
 				break;
 
 			case BPF_JMP|BPF_JGT|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JA, JBE);
 				break;
 
 			case BPF_JMP|BPF_JGE|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JAE, JB);
 				break;
 
 			case BPF_JMP|BPF_JEQ|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JE, JNE);
 				break;
 
 			case BPF_JMP|BPF_JSET|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				TESTrd(EDX, EAX);
 				JCC(JNE, JE);
 				break;
diff --git a/sys/amd64/amd64/bpf_jit_machdep.h b/sys/amd64/amd64/bpf_jit_machdep.h
index aa7f342fb63..01c251f6db9 100644
--- a/sys/amd64/amd64/bpf_jit_machdep.h
+++ b/sys/amd64/amd64/bpf_jit_machdep.h
@@ -473,4 +473,10 @@ typedef void (*emit_func)(bpf_bin_stream *stream, u_int value, u_int n);
 	}								\
 } while (0)
 
+#define	JUMP(off) do {							\
+	if ((off) != 0)							\
+		JMP(stream.refs[stream.bpf_pc + (off)] -		\
+		    stream.refs[stream.bpf_pc]);			\
+} while (0)
+
 #endif	/* _BPF_JIT_MACHDEP_H_ */
diff --git a/sys/i386/i386/bpf_jit_machdep.c b/sys/i386/i386/bpf_jit_machdep.c
index e9d9ecabe40..4c2946f9f6b 100644
--- a/sys/i386/i386/bpf_jit_machdep.c
+++ b/sys/i386/i386/bpf_jit_machdep.c
@@ -440,62 +440,77 @@ bpf_jit_compile(struct bpf_insn *prog, u_int nins, size_t *size)
 				break;
 
 			case BPF_JMP|BPF_JA:
-				JMP(stream.refs[stream.bpf_pc + ins->k] -
-				    stream.refs[stream.bpf_pc]);
+				JUMP(ins->k);
 				break;
 
 			case BPF_JMP|BPF_JGT|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JA, JBE);
 				break;
 
 			case BPF_JMP|BPF_JGE|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JAE, JB);
 				break;
 
 			case BPF_JMP|BPF_JEQ|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPid(ins->k, EAX);
 				JCC(JE, JNE);
 				break;
 
 			case BPF_JMP|BPF_JSET|BPF_K:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				TESTid(ins->k, EAX);
 				JCC(JNE, JE);
 				break;
 
 			case BPF_JMP|BPF_JGT|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JA, JBE);
 				break;
 
 			case BPF_JMP|BPF_JGE|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JAE, JB);
 				break;
 
 			case BPF_JMP|BPF_JEQ|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				CMPrd(EDX, EAX);
 				JCC(JE, JNE);
 				break;
 
 			case BPF_JMP|BPF_JSET|BPF_X:
-				if (ins->jt == 0 && ins->jf == 0)
+				if (ins->jt == ins->jf) {
+					JUMP(ins->jt);
 					break;
+				}
 				TESTrd(EDX, EAX);
 				JCC(JNE, JE);
 				break;
diff --git a/sys/i386/i386/bpf_jit_machdep.h b/sys/i386/i386/bpf_jit_machdep.h
index e82f68a2ec7..4ae5494a72a 100644
--- a/sys/i386/i386/bpf_jit_machdep.h
+++ b/sys/i386/i386/bpf_jit_machdep.h
@@ -418,4 +418,10 @@ typedef void (*emit_func)(bpf_bin_stream *stream, u_int value, u_int n);
 	}								\
 } while (0)
 
+#define	JUMP(off) do {							\
+	if ((off) != 0)							\
+		JMP(stream.refs[stream.bpf_pc + (off)] -		\
+		    stream.refs[stream.bpf_pc]);			\
+} while (0)
+
 #endif	/* _BPF_JIT_MACHDEP_H_ */

From 67c5c2d2d81f0ed3aee6edda8bcb3f8bd4e7ffce Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Thu, 22 Apr 2010 23:51:01 +0000
Subject: [PATCH 210/532] When the experimental NFS client is handling an NFSv4
 server reboot with delegations enabled, the recovery could fail if the renew
 thread is trying to return a delegation, since it will not do the recovery.
 This patch fixes the above by having nfscl_recalldeleg() fail with the I/O
 operations returning EIO, so that they will be attempted later. Most of the
 patch consists of adding an argument to various functions to indicate the
 delegation recall case where this needs to be done.

MFC after:	1 week
---
 sys/fs/nfs/nfs_var.h            |  4 ++--
 sys/fs/nfsclient/nfs.h          |  6 ++++--
 sys/fs/nfsclient/nfs_clbio.c    | 31 +++++++++++++++++----------
 sys/fs/nfsclient/nfs_clnfsiod.c |  6 ++++--
 sys/fs/nfsclient/nfs_clnode.c   |  2 +-
 sys/fs/nfsclient/nfs_clrpcops.c | 16 ++++++++++----
 sys/fs/nfsclient/nfs_clstate.c  | 25 ++++++++++++++++------
 sys/fs/nfsclient/nfs_clvnops.c  | 37 ++++++++++++++++++++++++++-------
 8 files changed, 91 insertions(+), 36 deletions(-)

diff --git a/sys/fs/nfs/nfs_var.h b/sys/fs/nfs/nfs_var.h
index 498511c5939..5388e13f549 100644
--- a/sys/fs/nfs/nfs_var.h
+++ b/sys/fs/nfs/nfs_var.h
@@ -369,7 +369,7 @@ int nfsrpc_readlink(vnode_t, struct uio *, struct ucred *,
 int nfsrpc_read(vnode_t, struct uio *, struct ucred *, NFSPROC_T *,
     struct nfsvattr *, int *, void *);
 int nfsrpc_write(vnode_t, struct uio *, int *, u_char *,
-    struct ucred *, NFSPROC_T *, struct nfsvattr *, int *, void *);
+    struct ucred *, NFSPROC_T *, struct nfsvattr *, int *, void *, int);
 int nfsrpc_mknod(vnode_t, char *, int, struct vattr *, u_int32_t,
     enum vtype, struct ucred *, NFSPROC_T *, struct nfsvattr *,
     struct nfsvattr *, struct nfsfh **, int *, int *, void *);
@@ -502,7 +502,7 @@ int nfscl_maperr(NFSPROC_T *, int, uid_t, gid_t);
 void nfscl_init(void);
 
 /* nfs_clbio.c */
-int ncl_flush(vnode_t, int, struct ucred *, NFSPROC_T *, int);
+int ncl_flush(vnode_t, int, struct ucred *, NFSPROC_T *, int, int);
 
 /* nfs_clnode.c */
 void ncl_invalcaches(vnode_t);
diff --git a/sys/fs/nfsclient/nfs.h b/sys/fs/nfsclient/nfs.h
index 4b542869d62..c6071aface4 100644
--- a/sys/fs/nfsclient/nfs.h
+++ b/sys/fs/nfsclient/nfs.h
@@ -79,14 +79,16 @@ int ncl_biowrite(struct vnode *, struct uio *, int, struct ucred *);
 int ncl_vinvalbuf(struct vnode *, int, struct thread *, int);
 int ncl_asyncio(struct nfsmount *, struct buf *, struct ucred *,
     struct thread *);
-int ncl_doio(struct vnode *, struct buf *, struct ucred *, struct thread *);
+int ncl_doio(struct vnode *, struct buf *, struct ucred *, struct thread *,
+    int);
 void ncl_nhinit(void);
 void ncl_nhuninit(void);
 void ncl_nodelock(struct nfsnode *);
 void ncl_nodeunlock(struct nfsnode *);
 int ncl_getattrcache(struct vnode *, struct vattr *);
 int ncl_readrpc(struct vnode *, struct uio *, struct ucred *);
-int ncl_writerpc(struct vnode *, struct uio *, struct ucred *, int *, int *);
+int ncl_writerpc(struct vnode *, struct uio *, struct ucred *, int *, int *,
+    int);
 int ncl_readlinkrpc(struct vnode *, struct uio *, struct ucred *);
 int ncl_readdirrpc(struct vnode *, struct uio *, struct ucred *,
     struct thread *);
diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index d0dd2cc2e9f..2401c887be7 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -336,7 +336,7 @@ ncl_putpages(struct vop_putpages_args *ap)
 	else
 	    iomode = NFSWRITE_FILESYNC;
 
-	error = ncl_writerpc(vp, &uio, cred, &iomode, &must_commit);
+	error = ncl_writerpc(vp, &uio, cred, &iomode, &must_commit, 0);
 
 	pmap_qremove(kva, npages);
 	relpbuf(bp, &ncl_pbuf_freecnt);
@@ -554,7 +554,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred)
 		if ((bp->b_flags & B_CACHE) == 0) {
 		    bp->b_iocmd = BIO_READ;
 		    vfs_busy_pages(bp, 0);
-		    error = ncl_doio(vp, bp, cred, td);
+		    error = ncl_doio(vp, bp, cred, td, 0);
 		    if (error) {
 			brelse(bp);
 			return (error);
@@ -583,7 +583,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred)
 		if ((bp->b_flags & B_CACHE) == 0) {
 		    bp->b_iocmd = BIO_READ;
 		    vfs_busy_pages(bp, 0);
-		    error = ncl_doio(vp, bp, cred, td);
+		    error = ncl_doio(vp, bp, cred, td, 0);
 		    if (error) {
 			bp->b_ioflags |= BIO_ERROR;
 			brelse(bp);
@@ -609,7 +609,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred)
 		if ((bp->b_flags & B_CACHE) == 0) {
 		    bp->b_iocmd = BIO_READ;
 		    vfs_busy_pages(bp, 0);
-		    error = ncl_doio(vp, bp, cred, td);
+		    error = ncl_doio(vp, bp, cred, td, 0);
 		    if (error) {
 			    brelse(bp);
 		    }
@@ -638,7 +638,7 @@ ncl_bioread(struct vnode *vp, struct uio *uio, int ioflag, struct ucred *cred)
 			    if ((bp->b_flags & B_CACHE) == 0) {
 				    bp->b_iocmd = BIO_READ;
 				    vfs_busy_pages(bp, 0);
-				    error = ncl_doio(vp, bp, cred, td);
+				    error = ncl_doio(vp, bp, cred, td, 0);
 				    /*
 				     * no error + B_INVAL == directory EOF,
 				     * use the block.
@@ -771,7 +771,7 @@ do_sync:
 			uio.uio_td = td;
 			iomode = NFSWRITE_FILESYNC;
 			error = ncl_writerpc(vp, &uio, cred, &iomode,
-			    &must_commit);
+			    &must_commit, 0);
 			KASSERT((must_commit == 0), 
 				("ncl_directio_write: Did not commit write"));
 			if (error)
@@ -1122,7 +1122,7 @@ again:
 		if ((bp->b_flags & B_CACHE) == 0) {
 			bp->b_iocmd = BIO_READ;
 			vfs_busy_pages(bp, 0);
-			error = ncl_doio(vp, bp, cred, td);
+			error = ncl_doio(vp, bp, cred, td, 0);
 			if (error) {
 				brelse(bp);
 				break;
@@ -1523,7 +1523,7 @@ ncl_doio_directwrite(struct buf *bp)
 	
 	iomode = NFSWRITE_FILESYNC;
 	uiop->uio_td = NULL; /* NULL since we're in nfsiod */
-	ncl_writerpc(bp->b_vp, uiop, bp->b_wcred, &iomode, &must_commit);
+	ncl_writerpc(bp->b_vp, uiop, bp->b_wcred, &iomode, &must_commit, 0);
 	KASSERT((must_commit == 0), ("ncl_doio_directwrite: Did not commit write"));
 	free(iov_base, M_NFSDIRECTIO);
 	free(uiop->uio_iov, M_NFSDIRECTIO);
@@ -1550,7 +1550,8 @@ ncl_doio_directwrite(struct buf *bp)
  * synchronously or from an nfsiod.
  */
 int
-ncl_doio(struct vnode *vp, struct buf *bp, struct ucred *cr, struct thread *td)
+ncl_doio(struct vnode *vp, struct buf *bp, struct ucred *cr, struct thread *td,
+    int called_from_strategy)
 {
 	struct uio *uiop;
 	struct nfsnode *np;
@@ -1695,7 +1696,8 @@ ncl_doio(struct vnode *vp, struct buf *bp, struct ucred *cr, struct thread *td)
 		else
 		    iomode = NFSWRITE_FILESYNC;
 
-		error = ncl_writerpc(vp, uiop, cr, &iomode, &must_commit);
+		error = ncl_writerpc(vp, uiop, cr, &iomode, &must_commit,
+		    called_from_strategy);
 
 		/*
 		 * When setting B_NEEDCOMMIT also set B_CLUSTEROK to try
@@ -1732,6 +1734,12 @@ ncl_doio(struct vnode *vp, struct buf *bp, struct ucred *cr, struct thread *td)
 		 * the block is reused. This is indicated by setting
 		 * the B_DELWRI and B_NEEDCOMMIT flags.
 		 *
+		 * EIO is returned by ncl_writerpc() to indicate a recoverable
+		 * write error and is handled as above, except that
+		 * B_EINTR isn't set. One cause of this is a stale stateid
+		 * error for the RPC that indicates recovery is required,
+		 * when called with called_from_strategy != 0.
+		 *
 		 * If the buffer is marked B_PAGING, it does not reside on
 		 * the vp's paging queues so we cannot call bdirty().  The
 		 * bp in this case is not an NFS cache block so we should
@@ -1760,7 +1768,8 @@ ncl_doio(struct vnode *vp, struct buf *bp, struct ucred *cr, struct thread *td)
 			    bdirty(bp);
 			    bp->b_flags &= ~B_DONE;
 			}
-			if (error && (bp->b_flags & B_ASYNC) == 0)
+			if ((error == EINTR || error == ETIMEDOUT) &&
+			    (bp->b_flags & B_ASYNC) == 0)
 			    bp->b_flags |= B_EINTR;
 			splx(s);
 	    	} else {
diff --git a/sys/fs/nfsclient/nfs_clnfsiod.c b/sys/fs/nfsclient/nfs_clnfsiod.c
index 6649fc032a3..62ea4f8fd4e 100644
--- a/sys/fs/nfsclient/nfs_clnfsiod.c
+++ b/sys/fs/nfsclient/nfs_clnfsiod.c
@@ -278,9 +278,11 @@ nfssvc_iod(void *instance)
 			(void)ncl_doio_directwrite(bp);
 		} else {
 			if (bp->b_iocmd == BIO_READ)
-				(void) ncl_doio(bp->b_vp, bp, bp->b_rcred, NULL);
+				(void) ncl_doio(bp->b_vp, bp, bp->b_rcred,
+				    NULL, 0);
 			else
-				(void) ncl_doio(bp->b_vp, bp, bp->b_wcred, NULL);
+				(void) ncl_doio(bp->b_vp, bp, bp->b_wcred,
+				    NULL, 0);
 		}
 		mtx_lock(&ncl_iod_mutex);
 		/*
diff --git a/sys/fs/nfsclient/nfs_clnode.c b/sys/fs/nfsclient/nfs_clnode.c
index 6b2aa7a5f2a..c1337427fd2 100644
--- a/sys/fs/nfsclient/nfs_clnode.c
+++ b/sys/fs/nfsclient/nfs_clnode.c
@@ -199,7 +199,7 @@ ncl_inactive(struct vop_inactive_args *ap)
 		 * available for the writes.
 		 */
 		if (nfscl_mustflush(vp))
-			(void) ncl_flush(vp, MNT_WAIT, NULL, ap->a_td, 1);
+			(void) ncl_flush(vp, MNT_WAIT, NULL, ap->a_td, 1, 0);
 		(void) nfsrpc_close(vp, 1, ap->a_td);
 	}
 
diff --git a/sys/fs/nfsclient/nfs_clrpcops.c b/sys/fs/nfsclient/nfs_clrpcops.c
index a7015f5226b..63c15c10fd9 100644
--- a/sys/fs/nfsclient/nfs_clrpcops.c
+++ b/sys/fs/nfsclient/nfs_clrpcops.c
@@ -1346,11 +1346,16 @@ nfsmout:
 
 /*
  * nfs write operation
+ * When called_from_strategy != 0, it should return EIO for an error that
+ * indicates recovery is in progress, so that the buffer will be left
+ * dirty and be written back to the server later. If it loops around,
+ * the recovery thread could get stuck waiting for the buffer and recovery
+ * will then deadlock.
  */
 APPLESTATIC int
 nfsrpc_write(vnode_t vp, struct uio *uiop, int *iomode, u_char *verfp,
     struct ucred *cred, NFSPROC_T *p, struct nfsvattr *nap, int *attrflagp,
-    void *stuff)
+    void *stuff, int called_from_strategy)
 {
 	int error, expireret = 0, retrycnt, nostateid;
 	u_int32_t clidrev = 0;
@@ -1410,12 +1415,15 @@ nfscl_dumpstate(nmp, 1, 1, 0, 0);
 			expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
 		}
 		retrycnt++;
-	} while (error == NFSERR_GRACE || error == NFSERR_STALESTATEID ||
-	    error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY ||
+	} while (error == NFSERR_GRACE || error == NFSERR_DELAY ||
+	    ((error == NFSERR_STALESTATEID ||
+	      error == NFSERR_STALEDONTRECOVER) && called_from_strategy == 0) ||
 	    (error == NFSERR_OLDSTATEID && retrycnt < 20) ||
 	    ((error == NFSERR_EXPIRED || error == NFSERR_BADSTATEID) &&
 	     expireret == 0 && clidrev != 0 && retrycnt < 4));
-	if (error && retrycnt >= 4)
+	if (error != 0 && (retrycnt >= 4 ||
+	    ((error == NFSERR_STALESTATEID ||
+	      error == NFSERR_STALEDONTRECOVER) && called_from_strategy != 0)))
 		error = EIO;
 	if (NFSHASNFSV4(nmp) && p == NULL)
 		NFSFREECRED(newcred);
diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index 78b596640cf..cf8c922ae43 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -139,7 +139,7 @@ static void nfscl_freedeleg(struct nfscldeleghead *, struct nfscldeleg *);
 static int nfscl_errmap(struct nfsrv_descript *);
 static void nfscl_cleanup_common(struct nfsclclient *, u_int8_t *);
 static int nfscl_recalldeleg(struct nfsclclient *, struct nfsmount *,
-    struct nfscldeleg *, vnode_t, struct ucred *, NFSPROC_T *);
+    struct nfscldeleg *, vnode_t, struct ucred *, NFSPROC_T *, int);
 static void nfscl_freeopenowner(struct nfsclowner *, int);
 static void nfscl_cleandeleg(struct nfscldeleg *);
 static int nfscl_trydelegreturn(struct nfscldeleg *, struct ucred *,
@@ -2469,7 +2469,7 @@ tryagain:
 				NFSUNLOCKCLSTATE();
 				newnfs_copycred(&dp->nfsdl_cred, cred);
 				ret = nfscl_recalldeleg(clp, clp->nfsc_nmp, dp,
-				    NULL, cred, p);
+				    NULL, cred, p, 1);
 				if (!ret) {
 				    nfscl_cleandeleg(dp);
 				    TAILQ_REMOVE(&clp->nfsc_deleg, dp,
@@ -3309,7 +3309,8 @@ nfscl_lockt(vnode_t vp, struct nfsclclient *clp, u_int64_t off,
  */
 static int
 nfscl_recalldeleg(struct nfsclclient *clp, struct nfsmount *nmp,
-    struct nfscldeleg *dp, vnode_t vp, struct ucred *cred, NFSPROC_T *p)
+    struct nfscldeleg *dp, vnode_t vp, struct ucred *cred, NFSPROC_T *p,
+    int called_from_renewthread)
 {
 	struct nfsclowner *owp, *lowp, *nowp;
 	struct nfsclopen *op, *lop;
@@ -3343,6 +3344,7 @@ nfscl_recalldeleg(struct nfsclclient *clp, struct nfsmount *nmp,
 	 * Ok, if it's a write delegation, flush data to the server, so
 	 * that close/open consistency is retained.
 	 */
+	ret = 0;
 	NFSLOCKNODE(np);
 	if ((dp->nfsdl_flags & NFSCLDL_WRITE) && (np->n_flag & NMODIFIED)) {
 #ifdef APPLE
@@ -3351,7 +3353,8 @@ nfscl_recalldeleg(struct nfsclclient *clp, struct nfsmount *nmp,
 		np->n_flag |= NDELEGRECALL;
 #endif
 		NFSUNLOCKNODE(np);
-		(void) ncl_flush(vp, MNT_WAIT, cred, p, 1);
+		ret = ncl_flush(vp, MNT_WAIT, cred, p, 1,
+		    called_from_renewthread);
 		NFSLOCKNODE(np);
 #ifdef APPLE
 		OSBitAndAtomic((int32_t)~(NMODIFIED | NDELEGRECALL), (UInt32 *)&np->n_flag);
@@ -3360,6 +3363,16 @@ nfscl_recalldeleg(struct nfsclclient *clp, struct nfsmount *nmp,
 #endif
 	}
 	NFSUNLOCKNODE(np);
+	if (ret == EIO && called_from_renewthread != 0) {
+		/*
+		 * If the flush failed with EIO for the renew thread,
+		 * return now, so that the dirty buffer will be flushed
+		 * later.
+		 */
+		if (gotvp != 0)
+			vrele(vp);
+		return (ret);
+	}
 
 	/*
 	 * Now, for each openowner with opens issued locally, move them
@@ -3857,7 +3870,7 @@ nfscl_removedeleg(vnode_t vp, NFSPROC_T *p, nfsv4stateid_t *stp)
 			NFSUNLOCKCLSTATE();
 			cred = newnfs_getcred();
 			newnfs_copycred(&dp->nfsdl_cred, cred);
-			(void) nfscl_recalldeleg(clp, nmp, dp, vp, cred, p);
+			(void) nfscl_recalldeleg(clp, nmp, dp, vp, cred, p, 0);
 			NFSFREECRED(cred);
 			triedrecall = 1;
 			NFSLOCKCLSTATE();
@@ -3955,7 +3968,7 @@ nfscl_renamedeleg(vnode_t fvp, nfsv4stateid_t *fstp, int *gotfdp, vnode_t tvp,
 			NFSUNLOCKCLSTATE();
 			cred = newnfs_getcred();
 			newnfs_copycred(&dp->nfsdl_cred, cred);
-			(void) nfscl_recalldeleg(clp, nmp, dp, fvp, cred, p);
+			(void) nfscl_recalldeleg(clp, nmp, dp, fvp, cred, p, 0);
 			NFSFREECRED(cred);
 			triedrecall = 1;
 			NFSLOCKCLSTATE();
diff --git a/sys/fs/nfsclient/nfs_clvnops.c b/sys/fs/nfsclient/nfs_clvnops.c
index 3be823f1969..4e844f1024b 100644
--- a/sys/fs/nfsclient/nfs_clvnops.c
+++ b/sys/fs/nfsclient/nfs_clvnops.c
@@ -670,13 +670,13 @@ nfs_close(struct vop_close_args *ap)
 		     * traditional vnode locking implemented for Vnode Ops.
 		     */
 		    int cm = newnfs_commit_on_close ? 1 : 0;
-		    error = ncl_flush(vp, MNT_WAIT, cred, ap->a_td, cm);
+		    error = ncl_flush(vp, MNT_WAIT, cred, ap->a_td, cm, 0);
 		    /* np->n_flag &= ~NMODIFIED; */
 		} else if (NFS_ISV4(vp)) { 
 			if (nfscl_mustflush(vp)) {
 				int cm = newnfs_commit_on_close ? 1 : 0;
 				error = ncl_flush(vp, MNT_WAIT, cred, ap->a_td,
-				    cm);
+				    cm, 0);
 				/*
 				 * as above w.r.t races when clearing
 				 * NMODIFIED.
@@ -1306,7 +1306,7 @@ ncl_readrpc(struct vnode *vp, struct uio *uiop, struct ucred *cred)
  */
 int
 ncl_writerpc(struct vnode *vp, struct uio *uiop, struct ucred *cred,
-	     int *iomode, int *must_commit)
+    int *iomode, int *must_commit, int called_from_strategy)
 {
 	struct nfsvattr nfsva;
 	int error = 0, attrflag, ret;
@@ -1315,7 +1315,7 @@ ncl_writerpc(struct vnode *vp, struct uio *uiop, struct ucred *cred,
 
 	*must_commit = 0;
 	error = nfsrpc_write(vp, uiop, iomode, verf, cred,
-	    uiop->uio_td, &nfsva, &attrflag, NULL);
+	    uiop->uio_td, &nfsva, &attrflag, NULL, called_from_strategy);
 	NFSLOCKMNT(nmp);
 	if (!error && NFSHASWRITEVERF(nmp) &&
 	    NFSBCMP(verf, nmp->nm_verf, NFSX_VERF)) {
@@ -2473,7 +2473,7 @@ nfs_strategy(struct vop_strategy_args *ap)
 	 */
 	if ((bp->b_flags & B_ASYNC) == 0 ||
 	    ncl_asyncio(VFSTONFS(ap->a_vp->v_mount), bp, NOCRED, curthread))
-		(void)ncl_doio(ap->a_vp, bp, cr, curthread);
+		(void) ncl_doio(ap->a_vp, bp, cr, curthread, 1);
 	return (0);
 }
 
@@ -2484,17 +2484,20 @@ nfs_strategy(struct vop_strategy_args *ap)
 static int
 nfs_fsync(struct vop_fsync_args *ap)
 {
-	return (ncl_flush(ap->a_vp, ap->a_waitfor, NULL, ap->a_td, 1));
+	return (ncl_flush(ap->a_vp, ap->a_waitfor, NULL, ap->a_td, 1, 0));
 }
 
 /*
  * Flush all the blocks associated with a vnode.
  * 	Walk through the buffer pool and push any dirty pages
  *	associated with the vnode.
+ * If the called_from_renewthread argument is TRUE, it has been called
+ * from the NFSv4 renew thread and, as such, cannot block indefinitely
+ * waiting for a buffer write to complete.
  */
 int
 ncl_flush(struct vnode *vp, int waitfor, struct ucred *cred, struct thread *td,
-    int commit)
+    int commit, int called_from_renewthread)
 {
 	struct nfsnode *np = VTONFS(vp);
 	struct buf *bp;
@@ -2513,6 +2516,8 @@ ncl_flush(struct vnode *vp, int waitfor, struct ucred *cred, struct thread *td,
 	struct buf *bvec_on_stack[NFS_COMMITBVECSIZ];
 	int bvecsize = 0, bveccount;
 
+	if (called_from_renewthread != 0)
+		slptimeo = hz;
 	if (nmp->nm_flag & NFSMNT_INT)
 		slpflag = NFS_PCATCH;
 	if (!commit)
@@ -2708,6 +2713,14 @@ loop:
 				error = 0;
 				goto loop;
 			}
+			if (called_from_renewthread != 0) {
+				/*
+				 * Return EIO so the flush will be retried
+				 * later.
+				 */
+				error = EIO;
+				goto done;
+			}
 			if (newnfs_sigintr(nmp, td)) {
 				error = EINTR;
 				goto done;
@@ -2747,6 +2760,14 @@ loop:
 			error = bufobj_wwait(bo, slpflag, slptimeo);
 			if (error) {
 			    BO_UNLOCK(bo);
+			    if (called_from_renewthread != 0) {
+				/*
+				 * Return EIO so that the flush will be
+				 * retried later.
+				 */
+				error = EIO;
+				goto done;
+			    }
 			    error = newnfs_sigintr(nmp, td);
 			    if (error)
 				goto done;
@@ -2838,7 +2859,7 @@ nfs_advlock(struct vop_advlock_args *ap)
 		 */
 		if (ap->a_op == F_UNLCK &&
 		    nfscl_checkwritelocked(vp, ap->a_fl, cred, td))
-			(void) ncl_flush(vp, MNT_WAIT, cred, td, 1);
+			(void) ncl_flush(vp, MNT_WAIT, cred, td, 1, 0);
 
 		/*
 		 * Loop around doing the lock op, while a blocking lock

From efa50b120d62f93d0d764b1f5553afe36bc7172c Mon Sep 17 00:00:00 2001
From: Neel Natu <neel@FreeBSD.org>
Date: Fri, 23 Apr 2010 01:34:01 +0000
Subject: [PATCH 211/532] Fix compilation error.

tick.c:298:5: error: "KDTRACE_HOOKS" is not defined
---
 sys/mips/mips/tick.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/mips/mips/tick.c b/sys/mips/mips/tick.c
index 9a9108c0b83..5931a5b103d 100644
--- a/sys/mips/mips/tick.c
+++ b/sys/mips/mips/tick.c
@@ -295,7 +295,7 @@ clock_intr(void *arg)
 	 */
 	if (delta > cycles_per_hz)
 		delta = cycles_per_hz;
-#if KDTRACE_HOOKS
+#ifdef KDTRACE_HOOKS
 	/*
 	 * If the DTrace hooks are configured and a callback function
 	 * has been registered, then call it to process the high speed

From 3f71c319f4771ce63279b8464aee835cdeca6846 Mon Sep 17 00:00:00 2001
From: Marcel Moolenaar <marcel@FreeBSD.org>
Date: Fri, 23 Apr 2010 03:11:39 +0000
Subject: [PATCH 212/532] Implement the resize verb and add support for
 resizing partitions for all schemes but EBR. Quality work by Andrey!

Submitted by:	"Andrey V. Elsukov" <bu7cher@yandex.ru>
---
 sys/geom/part/g_part.c       | 85 ++++++++++++++++++++++++++++++++++--
 sys/geom/part/g_part_apm.c   | 16 +++++++
 sys/geom/part/g_part_bsd.c   | 16 +++++++
 sys/geom/part/g_part_gpt.c   | 16 +++++++
 sys/geom/part/g_part_if.m    | 14 ++++++
 sys/geom/part/g_part_mbr.c   | 28 ++++++++++++
 sys/geom/part/g_part_pc98.c  | 28 ++++++++++++
 sys/geom/part/g_part_vtoc8.c | 23 ++++++++++
 8 files changed, 222 insertions(+), 4 deletions(-)

diff --git a/sys/geom/part/g_part.c b/sys/geom/part/g_part.c
index 0434ace0cfe..9cddd0dc9fd 100644
--- a/sys/geom/part/g_part.c
+++ b/sys/geom/part/g_part.c
@@ -971,9 +971,85 @@ g_part_ctl_recover(struct gctl_req *req, struct g_part_parms *gpp)
 static int
 g_part_ctl_resize(struct gctl_req *req, struct g_part_parms *gpp)
 {
-	gctl_error(req, "%d verb 'resize'", ENOSYS);
-	return (ENOSYS);
-} 
+	struct g_geom *gp;
+	struct g_provider *pp;
+	struct g_part_entry *pe, *entry;
+	struct g_part_table *table;
+	struct sbuf *sb;
+	quad_t end;
+	int error;
+
+	gp = gpp->gpp_geom;
+	G_PART_TRACE((G_T_TOPOLOGY, "%s(%s)", __func__, gp->name));
+	g_topology_assert();
+	table = gp->softc;
+
+	/* check gpp_index */
+	LIST_FOREACH(entry, &table->gpt_entry, gpe_entry) {
+		if (entry->gpe_deleted || entry->gpe_internal)
+			continue;
+		if (entry->gpe_index == gpp->gpp_index)
+			break;
+	}
+	if (entry == NULL) {
+		gctl_error(req, "%d index '%d'", ENOENT, gpp->gpp_index);
+		return (ENOENT);
+	}
+
+	/* check gpp_size */
+	end = entry->gpe_start + gpp->gpp_size - 1;
+	if (gpp->gpp_size < 1 || end > table->gpt_last) {
+		gctl_error(req, "%d size '%jd'", EINVAL,
+		    (intmax_t)gpp->gpp_size);
+		return (EINVAL);
+	}
+
+	LIST_FOREACH(pe, &table->gpt_entry, gpe_entry) {
+		if (pe->gpe_deleted || pe->gpe_internal || pe == entry)
+			continue;
+		if (end >= pe->gpe_start && end <= pe->gpe_end) {
+			gctl_error(req, "%d end '%jd'", ENOSPC,
+			    (intmax_t)end);
+			return (ENOSPC);
+		}
+		if (entry->gpe_start < pe->gpe_start && end > pe->gpe_end) {
+			gctl_error(req, "%d size '%jd'", ENOSPC,
+			    (intmax_t)gpp->gpp_size);
+			return (ENOSPC);
+		}
+	}
+
+	pp = entry->gpe_pp;
+	if ((g_debugflags & 16) == 0 &&
+	    (pp->acr > 0 || pp->acw > 0 || pp->ace > 0)) {
+		gctl_error(req, "%d", EBUSY);
+		return (EBUSY);
+	}
+
+	error = G_PART_RESIZE(table, entry, gpp);
+	if (error) {
+		gctl_error(req, "%d", error);
+		return (error);
+	}
+
+	if (!entry->gpe_created)
+		entry->gpe_modified = 1;
+
+	/* update mediasize of changed provider */
+	pp->mediasize = (entry->gpe_end - entry->gpe_start + 1) *
+		pp->sectorsize;
+
+	/* Provide feedback if so requested. */
+	if (gpp->gpp_parms & G_PART_PARM_OUTPUT) {
+		sb = sbuf_new_auto();
+		G_PART_FULLNAME(table, entry, sb, gp->name);
+		sbuf_cat(sb, " resized\n");
+		sbuf_finish(sb);
+		gctl_set_param(req, "output", sbuf_data(sb), sbuf_len(sb) + 1);
+		sbuf_delete(sb);
+	}
+	return (0);
+}
 
 static int
 g_part_ctl_setunset(struct gctl_req *req, struct g_part_parms *gpp,
@@ -1194,7 +1270,8 @@ g_part_ctlreq(struct gctl_req *req, struct g_class *mp, const char *verb)
 			mparms |= G_PART_PARM_GEOM;
 		} else if (!strcmp(verb, "resize")) {
 			ctlreq = G_PART_CTL_RESIZE;
-			mparms |= G_PART_PARM_GEOM | G_PART_PARM_INDEX;
+			mparms |= G_PART_PARM_GEOM | G_PART_PARM_INDEX |
+			    G_PART_PARM_SIZE;
 		}
 		break;
 	case 's':
diff --git a/sys/geom/part/g_part_apm.c b/sys/geom/part/g_part_apm.c
index 030a0e69c86..c80925a4f53 100644
--- a/sys/geom/part/g_part_apm.c
+++ b/sys/geom/part/g_part_apm.c
@@ -74,6 +74,8 @@ static int g_part_apm_read(struct g_part_table *, struct g_consumer *);
 static const char *g_part_apm_type(struct g_part_table *, struct g_part_entry *,
     char *, size_t);
 static int g_part_apm_write(struct g_part_table *, struct g_consumer *);
+static int g_part_apm_resize(struct g_part_table *, struct g_part_entry *,
+    struct g_part_parms *);
 
 static kobj_method_t g_part_apm_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_apm_add),
@@ -82,6 +84,7 @@ static kobj_method_t g_part_apm_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_apm_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_apm_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_apm_modify),
+	KOBJMETHOD(g_part_resize,	g_part_apm_resize),
 	KOBJMETHOD(g_part_name,		g_part_apm_name),
 	KOBJMETHOD(g_part_probe,	g_part_apm_probe),
 	KOBJMETHOD(g_part_read,		g_part_apm_read),
@@ -318,6 +321,19 @@ g_part_apm_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_apm_resize(struct g_part_table *basetable,
+    struct g_part_entry *baseentry, struct g_part_parms *gpp)
+{
+	struct g_part_apm_entry *entry;
+
+	entry = (struct g_part_apm_entry *)baseentry;
+	baseentry->gpe_end = baseentry->gpe_start + gpp->gpp_size - 1;
+	entry->ent.ent_size = gpp->gpp_size;
+
+	return (0);
+}
+
 static const char *
 g_part_apm_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)
diff --git a/sys/geom/part/g_part_bsd.c b/sys/geom/part/g_part_bsd.c
index 18a74aa0c79..e99f0bbf821 100644
--- a/sys/geom/part/g_part_bsd.c
+++ b/sys/geom/part/g_part_bsd.c
@@ -73,6 +73,8 @@ static int g_part_bsd_read(struct g_part_table *, struct g_consumer *);
 static const char *g_part_bsd_type(struct g_part_table *, struct g_part_entry *,
     char *, size_t);
 static int g_part_bsd_write(struct g_part_table *, struct g_consumer *);
+static int g_part_bsd_resize(struct g_part_table *, struct g_part_entry *,
+    struct g_part_parms *);
 
 static kobj_method_t g_part_bsd_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_bsd_add),
@@ -82,6 +84,7 @@ static kobj_method_t g_part_bsd_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_bsd_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_bsd_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_bsd_modify),
+	KOBJMETHOD(g_part_resize,	g_part_bsd_resize),
 	KOBJMETHOD(g_part_name,		g_part_bsd_name),
 	KOBJMETHOD(g_part_probe,	g_part_bsd_probe),
 	KOBJMETHOD(g_part_read,		g_part_bsd_read),
@@ -288,6 +291,19 @@ g_part_bsd_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_bsd_resize(struct g_part_table *basetable,
+    struct g_part_entry *baseentry, struct g_part_parms *gpp)
+{
+	struct g_part_bsd_entry *entry;
+
+	entry = (struct g_part_bsd_entry *)baseentry;
+	baseentry->gpe_end = baseentry->gpe_start + gpp->gpp_size - 1;
+	entry->part.p_size = gpp->gpp_size;
+
+	return (0);
+}
+
 static const char *
 g_part_bsd_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)
diff --git a/sys/geom/part/g_part_gpt.c b/sys/geom/part/g_part_gpt.c
index b0c959852b6..a6f74907102 100644
--- a/sys/geom/part/g_part_gpt.c
+++ b/sys/geom/part/g_part_gpt.c
@@ -103,6 +103,8 @@ static int g_part_gpt_read(struct g_part_table *, struct g_consumer *);
 static const char *g_part_gpt_type(struct g_part_table *, struct g_part_entry *,
     char *, size_t);
 static int g_part_gpt_write(struct g_part_table *, struct g_consumer *);
+static int g_part_gpt_resize(struct g_part_table *, struct g_part_entry *,
+    struct g_part_parms *);
 
 static kobj_method_t g_part_gpt_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_gpt_add),
@@ -112,6 +114,7 @@ static kobj_method_t g_part_gpt_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_gpt_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_gpt_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_gpt_modify),
+	KOBJMETHOD(g_part_resize,	g_part_gpt_resize),
 	KOBJMETHOD(g_part_name,		g_part_gpt_name),
 	KOBJMETHOD(g_part_probe,	g_part_gpt_probe),
 	KOBJMETHOD(g_part_read,		g_part_gpt_read),
@@ -550,6 +553,19 @@ g_part_gpt_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_gpt_resize(struct g_part_table *basetable,
+    struct g_part_entry *baseentry, struct g_part_parms *gpp)
+{
+	struct g_part_gpt_entry *entry;
+	entry = (struct g_part_gpt_entry *)baseentry;
+
+	baseentry->gpe_end = baseentry->gpe_start + gpp->gpp_size - 1;
+	entry->ent.ent_lba_end = baseentry->gpe_end;
+
+	return (0);
+}
+
 static const char *
 g_part_gpt_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)
diff --git a/sys/geom/part/g_part_if.m b/sys/geom/part/g_part_if.m
index 8252f0cadc3..04440fef122 100644
--- a/sys/geom/part/g_part_if.m
+++ b/sys/geom/part/g_part_if.m
@@ -58,6 +58,13 @@ CODE {
 	{
 		return (0);
 	}
+
+	static int
+	default_resize(struct g_part_table *t __unused,
+	    struct g_part_entry *e __unused, struct g_part_parms *p __unused)
+	{
+		return (ENOSYS);
+	}
 };
 
 # add() - scheme specific processing for the add verb.
@@ -114,6 +121,13 @@ METHOD int modify {
 	struct g_part_parms *gpp;
 };
 
+# resize() - scheme specific processing for the resize verb.
+METHOD int resize {
+	struct g_part_table *table;
+	struct g_part_entry *entry;
+	struct g_part_parms *gpp;
+} DEFAULT default_resize;
+
 # name() - return the name of the given partition entry.
 # Typical names are "p1", "s0" or "c".
 METHOD const char * name {
diff --git a/sys/geom/part/g_part_mbr.c b/sys/geom/part/g_part_mbr.c
index 72d0ecbfe10..8b5ba2779a6 100644
--- a/sys/geom/part/g_part_mbr.c
+++ b/sys/geom/part/g_part_mbr.c
@@ -76,6 +76,8 @@ static int g_part_mbr_setunset(struct g_part_table *, struct g_part_entry *,
 static const char *g_part_mbr_type(struct g_part_table *, struct g_part_entry *,
     char *, size_t);
 static int g_part_mbr_write(struct g_part_table *, struct g_consumer *);
+static int g_part_mbr_resize(struct g_part_table *, struct g_part_entry *,
+    struct g_part_parms *);
 
 static kobj_method_t g_part_mbr_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_mbr_add),
@@ -85,6 +87,7 @@ static kobj_method_t g_part_mbr_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_mbr_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_mbr_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_mbr_modify),
+	KOBJMETHOD(g_part_resize,	g_part_mbr_resize),
 	KOBJMETHOD(g_part_name,		g_part_mbr_name),
 	KOBJMETHOD(g_part_probe,	g_part_mbr_probe),
 	KOBJMETHOD(g_part_read,		g_part_mbr_read),
@@ -302,6 +305,31 @@ g_part_mbr_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_mbr_resize(struct g_part_table *basetable,
+    struct g_part_entry *baseentry, struct g_part_parms *gpp)
+{
+	struct g_part_mbr_entry *entry;
+	uint32_t size, sectors;
+
+	sectors = basetable->gpt_sectors;
+	size = gpp->gpp_size;
+
+	if (size < sectors)
+		return (EINVAL);
+	if (size % sectors)
+		size = size - (size % sectors);
+	if (size < sectors)
+		return (EINVAL);
+
+	entry = (struct g_part_mbr_entry *)baseentry;
+	baseentry->gpe_end = baseentry->gpe_start + size - 1;
+	entry->ent.dp_size = size;
+	mbr_set_chs(basetable, baseentry->gpe_end, &entry->ent.dp_ecyl,
+	    &entry->ent.dp_ehd, &entry->ent.dp_esect);
+	return (0);
+}
+
 static const char *
 g_part_mbr_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)
diff --git a/sys/geom/part/g_part_pc98.c b/sys/geom/part/g_part_pc98.c
index 8857f9b98f0..ab83662a4f9 100644
--- a/sys/geom/part/g_part_pc98.c
+++ b/sys/geom/part/g_part_pc98.c
@@ -77,6 +77,8 @@ static int g_part_pc98_setunset(struct g_part_table *, struct g_part_entry *,
 static const char *g_part_pc98_type(struct g_part_table *,
     struct g_part_entry *, char *, size_t);
 static int g_part_pc98_write(struct g_part_table *, struct g_consumer *);
+static int g_part_pc98_resize(struct g_part_table *, struct g_part_entry *,  
+    struct g_part_parms *);
 
 static kobj_method_t g_part_pc98_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_pc98_add),
@@ -86,6 +88,7 @@ static kobj_method_t g_part_pc98_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_pc98_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_pc98_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_pc98_modify),
+	KOBJMETHOD(g_part_resize,	g_part_pc98_resize),
 	KOBJMETHOD(g_part_name,		g_part_pc98_name),
 	KOBJMETHOD(g_part_probe,	g_part_pc98_probe),
 	KOBJMETHOD(g_part_read,		g_part_pc98_read),
@@ -308,6 +311,31 @@ g_part_pc98_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_pc98_resize(struct g_part_table *basetable,
+    struct g_part_entry *baseentry, struct g_part_parms *gpp)
+{
+	struct g_part_pc98_entry *entry;
+	uint32_t size, cyl;
+
+	cyl = basetable->gpt_heads * basetable->gpt_sectors;
+	size = gpp->gpp_size;
+
+	if (size < cyl)
+		return (EINVAL);
+	if (size % cyl)
+		size = size - (size % cyl);
+	if (size < cyl)
+		return (EINVAL);
+
+	entry = (struct g_part_pc98_entry *)baseentry;
+	baseentry->gpe_end = baseentry->gpe_start + size - 1;
+	pc98_set_chs(basetable, baseentry->gpe_end, &entry->ent.dp_ecyl,
+	    &entry->ent.dp_ehd, &entry->ent.dp_esect);
+
+	return (0);
+}
+
 static const char *
 g_part_pc98_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)
diff --git a/sys/geom/part/g_part_vtoc8.c b/sys/geom/part/g_part_vtoc8.c
index 0108e97a093..2cabcfdfc2c 100644
--- a/sys/geom/part/g_part_vtoc8.c
+++ b/sys/geom/part/g_part_vtoc8.c
@@ -67,6 +67,8 @@ static int g_part_vtoc8_read(struct g_part_table *, struct g_consumer *);
 static const char *g_part_vtoc8_type(struct g_part_table *, struct g_part_entry *,
     char *, size_t);
 static int g_part_vtoc8_write(struct g_part_table *, struct g_consumer *);
+static int g_part_vtoc8_resize(struct g_part_table *, struct g_part_entry *,  
+    struct g_part_parms *);
 
 static kobj_method_t g_part_vtoc8_methods[] = {
 	KOBJMETHOD(g_part_add,		g_part_vtoc8_add),
@@ -75,6 +77,7 @@ static kobj_method_t g_part_vtoc8_methods[] = {
 	KOBJMETHOD(g_part_dumpconf,	g_part_vtoc8_dumpconf),
 	KOBJMETHOD(g_part_dumpto,	g_part_vtoc8_dumpto),
 	KOBJMETHOD(g_part_modify,	g_part_vtoc8_modify),
+	KOBJMETHOD(g_part_resize,	g_part_vtoc8_resize),
 	KOBJMETHOD(g_part_name,		g_part_vtoc8_name),
 	KOBJMETHOD(g_part_probe,	g_part_vtoc8_probe),
 	KOBJMETHOD(g_part_read,		g_part_vtoc8_read),
@@ -294,6 +297,26 @@ g_part_vtoc8_modify(struct g_part_table *basetable,
 	return (0);
 }
 
+static int
+g_part_vtoc8_resize(struct g_part_table *basetable,
+    struct g_part_entry *entry, struct g_part_parms *gpp)
+{
+	struct g_part_vtoc8_table *table;
+	uint64_t size;
+
+	table = (struct g_part_vtoc8_table *)basetable;
+	size = gpp->gpp_size;
+	if (size % table->secpercyl)
+		size = size - (size % table->secpercyl);
+	if (size < table->secpercyl)
+		return (EINVAL);
+
+	entry->gpe_end = entry->gpe_start + size - 1;
+	be32enc(&table->vtoc.map[entry->gpe_index - 1].nblks, size);
+
+	return (0);
+}
+
 static const char *
 g_part_vtoc8_name(struct g_part_table *table, struct g_part_entry *baseentry,
     char *buf, size_t bufsz)

From 783310e6d9ad66127ae3fc49e27f6994c993366e Mon Sep 17 00:00:00 2001
From: Marcel Moolenaar <marcel@FreeBSD.org>
Date: Fri, 23 Apr 2010 03:14:04 +0000
Subject: [PATCH 213/532] Implement the resize command for resizing partitions.
 Without new size, the partition in question is resized to fill all available
 space. Quality work by Andrey!

Submitted by:	"Andrey V. Elsukov" <bu7cher@yandex.ru>
---
 sbin/geom/class/part/geom_part.c | 102 +++++++++++++++++++++++++++++++
 sbin/geom/class/part/gpart.8     |  33 +++++++++-
 2 files changed, 134 insertions(+), 1 deletion(-)

diff --git a/sbin/geom/class/part/geom_part.c b/sbin/geom/class/part/geom_part.c
index e2a045e3cfa..3ddbd7aa5cb 100644
--- a/sbin/geom/class/part/geom_part.c
+++ b/sbin/geom/class/part/geom_part.c
@@ -133,6 +133,13 @@ struct g_command PUBSYM(class_commands)[] = {
 		G_OPT_SENTINEL },
 	  "geom", NULL
         },
+	{ "resize", 0, gpart_issue, {
+		{ 's', "size", autofill, G_TYPE_ASCLBA },
+		{ 'i', index_param, NULL, G_TYPE_ASCNUM },
+		{ 'f', "flags", flags, G_TYPE_STRING },
+		G_OPT_SENTINEL },
+	  "geom", NULL
+	},
 	G_CMD_SENTINEL
 };
 
@@ -242,6 +249,99 @@ fmtattrib(struct gprovider *pp)
 	return (buf);
 }
 
+static int
+gpart_autofill_resize(struct gctl_req *req)
+{
+	struct gmesh mesh;
+	struct gclass *cp;
+	struct ggeom *gp;
+	struct gprovider *pp;
+	unsigned long long last, size, start, new_size;
+	unsigned long long lba, new_lba;
+	const char *s;
+	char *val;
+	int error, idx;
+
+	s = gctl_get_ascii(req, "size");
+	if (*s == '*')
+		new_size = (unsigned long long)atoll(s);
+	else
+		return (0);
+
+	s = gctl_get_ascii(req, index_param);
+	idx = strtol(s, &val, 10);
+	if (idx < 1 || *s == '\0' || *val != '\0')
+		errx(EXIT_FAILURE, "invalid partition index");
+
+	error = geom_gettree(&mesh);
+	if (error)
+		return (error);
+	s = gctl_get_ascii(req, "class");
+	if (s == NULL)
+		abort();
+	cp = find_class(&mesh, s);
+	if (cp == NULL)
+		errx(EXIT_FAILURE, "Class %s not found.", s);
+	s = gctl_get_ascii(req, "geom");
+	if (s == NULL)
+		abort();
+	gp = find_geom(cp, s);
+	if (gp == NULL)
+		errx(EXIT_FAILURE, "No such geom: %s.", s);
+	last = atoll(find_geomcfg(gp, "last"));
+
+	LIST_FOREACH(pp, &gp->lg_provider, lg_provider) {
+		s = find_provcfg(pp, "index");
+		if (s == NULL)
+			continue;
+		if (atoi(s) == idx)
+			break;
+	}
+	if (pp == NULL)
+		errx(EXIT_FAILURE, "invalid partition index");
+
+	s = find_provcfg(pp, "start");
+	if (s == NULL) {
+		s = find_provcfg(pp, "offset");
+		start = atoll(s) / pp->lg_sectorsize;
+	} else
+		start = atoll(s);
+	s = find_provcfg(pp, "end");
+	if (s == NULL) {
+		s = find_provcfg(pp, "length");
+		lba = start + atoll(s) / pp->lg_sectorsize;
+	} else
+		lba = atoll(s) + 1;
+
+	if (lba > last)
+		return (ENOSPC);
+	size = lba - start;
+	pp = find_provider(gp, lba);
+	if (pp == NULL)
+		new_size = last - start + 1;
+	else {
+		s = find_provcfg(pp, "start");
+		if (s == NULL) {
+			s = find_provcfg(pp, "offset");
+			new_lba = atoll(s) / pp->lg_sectorsize;
+		} else
+			new_lba = atoll(s);
+		/* Is there any free space between current and
+		 * next providers?
+		 */
+		if (new_lba > lba)
+			new_size = new_lba - start;
+		else
+			return (ENOSPC);
+	}
+	asprintf(&val, "%llu", new_size);
+	if (val == NULL)
+		return (ENOMEM);
+	gctl_change_param(req, "size", -1, val);
+
+	return (0);
+}
+
 static int
 gpart_autofill(struct gctl_req *req)
 {
@@ -257,6 +357,8 @@ gpart_autofill(struct gctl_req *req)
 	int error, has_size, has_start;
 
 	s = gctl_get_ascii(req, "verb");
+	if (strcmp(s, "resize") == 0)
+		return gpart_autofill_resize(req);
 	if (strcmp(s, "add") != 0)
 		return (0);
 
diff --git a/sbin/geom/class/part/gpart.8 b/sbin/geom/class/part/gpart.8
index 3de658c775b..0e6a1b97f90 100644
--- a/sbin/geom/class/part/gpart.8
+++ b/sbin/geom/class/part/gpart.8
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd November 18, 2008
+.Dd April 22, 2010
 .Dt GPART 8
 .Os
 .Sh NAME
@@ -120,6 +120,13 @@ utility:
 .Op Fl t Ar type
 .Op Fl f Ar flags
 .Ar geom
+.\" ==== RESIZE ====
+.Nm
+.Cm resize
+.Fl i Ar index
+.Op Fl s Ar size
+.Op Fl f Ar flags
+.Ar geom
 .\" ==== SET ====
 .Nm
 .Cm set
@@ -325,6 +332,30 @@ See the section entitled
 below for a discussion
 about its use.
 .El
+.\" ==== RESIZE ====
+.It Cm resize
+Resize a partition from geom
+.Ar geom
+and further identified by the
+.Fl i Ar index
+option. New partition size is expressed in logical block
+numbers and can be given by the 
+.Fl s Ar size
+option. If
+.Fl s
+option is ommited then new size is automatically calculated
+to maximum available from given geom
+.Ar geom .
+.Pp
+Additional options include:
+.Bl -tag -width 10n
+.It Fl f Ar flags
+Additional operational flags.
+See the section entitled
+.Sx "OPERATIONAL FLAGS"
+below for a discussion
+about its use.
+.El
 .\" ==== SET ====
 .It Cm set
 Set the named attribute on the partition entry.

From 665a080d7181b963d1842dcbc212ee7456ddce4a Mon Sep 17 00:00:00 2001
From: Marcel Moolenaar <marcel@FreeBSD.org>
Date: Fri, 23 Apr 2010 03:37:39 +0000
Subject: [PATCH 214/532] Add and describe GEOM_PART_EBR.

---
 sbin/geom/class/part/gpart.8 | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sbin/geom/class/part/gpart.8 b/sbin/geom/class/part/gpart.8
index 0e6a1b97f90..66557f305c9 100644
--- a/sbin/geom/class/part/gpart.8
+++ b/sbin/geom/class/part/gpart.8
@@ -37,6 +37,7 @@ lines in your kernel configuration file:
 .Bd -ragged -offset indent
 .Cd "options GEOM_PART_APM"
 .Cd "options GEOM_PART_BSD"
+.Cd "options GEOM_PART_EBR"
 .Cd "options GEOM_PART_GPT"
 .Cd "options GEOM_PART_MBR"
 .Cd "options GEOM_PART_PC98"
@@ -53,6 +54,10 @@ option adds support for the traditional
 .Bx
 disklabel.
 The
+.Dv GEOM_PART_EBR
+option adds support for the Extended Boot Record (EBR),
+which is used to define a logical partition.
+The
 .Dv GEOM_PART_GPT
 option adds support for the GUID Partition Table (GPT)
 found on Intel Itanium computers and Intel-based Macintosh computers.

From f31e6c7f26cf0980192132d1b7bfdf743490f1ad Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Fri, 23 Apr 2010 08:19:47 +0000
Subject: [PATCH 215/532] * Fix compilation when using SCTP_AUDITING_ENABLED. *
 Fix delaying of SACK by taking out old optimization code   which does not
 optimize anymore. * Fix fast retransmission of chunks abandoned by the  
 "number of retransmissions" policy.

MFC after: 3 days.
---
 sys/netinet/sctp_asconf.c |  2 +-
 sys/netinet/sctp_indata.c | 58 ++++++++++++++-------------------------
 sys/netinet/sctputil.c    |  6 ++--
 3 files changed, 24 insertions(+), 42 deletions(-)

diff --git a/sys/netinet/sctp_asconf.c b/sys/netinet/sctp_asconf.c
index f7de414ed1a..2d16ef01bae 100644
--- a/sys/netinet/sctp_asconf.c
+++ b/sys/netinet/sctp_asconf.c
@@ -1113,7 +1113,7 @@ sctp_assoc_immediate_retrans(struct sctp_tcb *stcb, struct sctp_nets *dstnet)
 		}
 		SCTP_TCB_LOCK_ASSERT(stcb);
 #ifdef SCTP_AUDITING_ENABLED
-		sctp_auditing(4, stcb->sctp_ep, stcb->asoc.deleted_primary);
+		sctp_auditing(4, stcb->sctp_ep, stcb, stcb->asoc.deleted_primary);
 #endif
 		sctp_chunk_output(stcb->sctp_ep, stcb, SCTP_OUTPUT_FROM_T3, SCTP_SO_NOT_LOCKED);
 		if ((stcb->asoc.num_send_timers_up == 0) &&
diff --git a/sys/netinet/sctp_indata.c b/sys/netinet/sctp_indata.c
index c1430a53f40..4b00fc3fc2c 100644
--- a/sys/netinet/sctp_indata.c
+++ b/sys/netinet/sctp_indata.c
@@ -2823,25 +2823,7 @@ sctp_process_data(struct mbuf **mm, int iphlen, int *offset, int length,
 		stcb->asoc.send_sack = 1;
 	}
 	/* Start a sack timer or QUEUE a SACK for sending */
-	if ((stcb->asoc.cumulative_tsn == stcb->asoc.highest_tsn_inside_nr_map) &&
-	    (stcb->asoc.nr_mapping_array[0] != 0xff)) {
-		if ((stcb->asoc.data_pkts_seen >= stcb->asoc.sack_freq) ||
-		    (stcb->asoc.delayed_ack == 0) ||
-		    (stcb->asoc.numduptsns) ||
-		    (stcb->asoc.send_sack == 1)) {
-			if (SCTP_OS_TIMER_PENDING(&stcb->asoc.dack_timer.timer)) {
-				(void)SCTP_OS_TIMER_STOP(&stcb->asoc.dack_timer.timer);
-			}
-			sctp_send_sack(stcb);
-		} else {
-			if (!SCTP_OS_TIMER_PENDING(&stcb->asoc.dack_timer.timer)) {
-				sctp_timer_start(SCTP_TIMER_TYPE_RECV,
-				    stcb->sctp_ep, stcb, NULL);
-			}
-		}
-	} else {
-		sctp_sack_check(stcb, was_a_gap, &abort_flag);
-	}
+	sctp_sack_check(stcb, was_a_gap, &abort_flag);
 	if (abort_flag)
 		return (2);
 
@@ -3532,6 +3514,25 @@ sctp_strike_gap_ack_chunks(struct sctp_tcb *stcb, struct sctp_association *asoc,
 		if (tp1->sent == SCTP_DATAGRAM_RESEND) {
 			struct sctp_nets *alt;
 
+			if ((stcb->asoc.peer_supports_prsctp) &&
+			    (PR_SCTP_RTX_ENABLED(tp1->flags))) {
+				/*
+				 * Has it been retransmitted tv_sec times? -
+				 * we store the retran count there.
+				 */
+				if (tp1->snd_count > tp1->rec.data.timetodrop.tv_sec) {
+					/* Yes, so drop it */
+					if (tp1->data != NULL) {
+						(void)sctp_release_pr_sctp_chunk(stcb, tp1,
+						    (SCTP_RESPONSE_TO_USER_REQ | SCTP_NOTIFY_DATAGRAM_SENT),
+						    SCTP_SO_NOT_LOCKED);
+					}
+					/* Make sure to flag we had a FR */
+					tp1->whoTo->net_ack++;
+					tp1 = TAILQ_NEXT(tp1, sctp_next);
+					continue;
+				}
+			}
 			/* fix counts and things */
 			if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_FLIGHT_LOGGING_ENABLE) {
 				sctp_misc_ints(SCTP_FLIGHT_LOG_DOWN_RSND,
@@ -3554,25 +3555,6 @@ sctp_strike_gap_ack_chunks(struct sctp_tcb *stcb, struct sctp_association *asoc,
 			/* remove from the total flight */
 			sctp_total_flight_decrease(stcb, tp1);
 
-			if ((stcb->asoc.peer_supports_prsctp) &&
-			    (PR_SCTP_RTX_ENABLED(tp1->flags))) {
-				/*
-				 * Has it been retransmitted tv_sec times? -
-				 * we store the retran count there.
-				 */
-				if (tp1->snd_count > tp1->rec.data.timetodrop.tv_sec) {
-					/* Yes, so drop it */
-					if (tp1->data != NULL) {
-						(void)sctp_release_pr_sctp_chunk(stcb, tp1,
-						    (SCTP_RESPONSE_TO_USER_REQ | SCTP_NOTIFY_DATAGRAM_SENT),
-						    SCTP_SO_NOT_LOCKED);
-					}
-					/* Make sure to flag we had a FR */
-					tp1->whoTo->net_ack++;
-					tp1 = TAILQ_NEXT(tp1, sctp_next);
-					continue;
-				}
-			}
 			/* printf("OK, we are now ready to FR this guy\n"); */
 			if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_FR_LOGGING_ENABLE) {
 				sctp_log_fr(tp1->rec.data.TSN_seq, tp1->snd_count,
diff --git a/sys/netinet/sctputil.c b/sys/netinet/sctputil.c
index 26122ce4abf..04f76732cf1 100644
--- a/sys/netinet/sctputil.c
+++ b/sys/netinet/sctputil.c
@@ -674,7 +674,7 @@ sctp_auditing(int from, struct sctp_inpcb *inp, struct sctp_tcb *stcb,
 			sctp_audit_indx = 0;
 		}
 		rep = 1;
-		SCTP_PRINTF("tot_flt_book:%d\n", tot_book);
+		SCTP_PRINTF("tot_flt_book:%d\n", tot_book_cnt);
 
 		stcb->asoc.total_flight_count = tot_book_cnt;
 	}
@@ -703,8 +703,8 @@ sctp_auditing(int from, struct sctp_inpcb *inp, struct sctp_tcb *stcb,
 				}
 			}
 			if (lnet->flight_size != tot_out) {
-				SCTP_PRINTF("net:%x flight was %d corrected to %d\n",
-				    (uint32_t) lnet, lnet->flight_size,
+				SCTP_PRINTF("net:%p flight was %d corrected to %d\n",
+				    lnet, lnet->flight_size,
 				    tot_out);
 				lnet->flight_size = tot_out;
 			}

From 762c7db2e8f714203a9766492a631d0ea310bd78 Mon Sep 17 00:00:00 2001
From: Florent Thoumie <flz@FreeBSD.org>
Date: Fri, 23 Apr 2010 11:07:43 +0000
Subject: [PATCH 216/532] - Take libinstall.a out of pkg_install and make it a
 proper shared library. - Rework the wrapper support to check libpkg version
 as well as pkg_install version. - Add libfetch to _prebuild_libs. - There are
 no new features introduced.

Notes: the API is not stable, so basically, do not use libpkg in your
projects for now. Also there's no manpage for libpkg yet, because the API
will change drastically. I repeat, do not use libpkg for now.
---
 Makefile.inc1                                 |  5 +-
 lib/Makefile                                  |  5 ++
 lib/libpkg/Makefile                           | 47 ++++++++++
 .../pkg_install/lib => lib/libpkg}/deps.c     |  2 +-
 .../pkg_install/lib => lib/libpkg}/exec.c     |  2 +-
 .../pkg_install/lib => lib/libpkg}/file.c     |  2 +-
 .../pkg_install/lib => lib/libpkg}/global.c   |  2 +-
 .../pkg_install/lib => lib/libpkg}/match.c    |  4 +-
 .../pkg_install/lib => lib/libpkg}/msg.c      |  9 +-
 .../pkg_install/lib => lib/libpkg}/pen.c      | 17 ++--
 .../pkg_install/lib/lib.h => lib/libpkg/pkg.h | 12 ++-
 lib/libpkg/pkgwrap.c                          | 90 +++++++++++++++++++
 .../pkg_install/lib => lib/libpkg}/plist.c    |  2 +-
 .../pkg_install/lib => lib/libpkg}/str.c      |  2 +-
 .../pkg_install/lib => lib/libpkg}/url.c      |  4 +-
 .../pkg_install/lib => lib/libpkg}/version.c  | 12 ++-
 share/mk/bsd.libnames.mk                      |  1 +
 usr.sbin/pkg_install/Makefile                 |  4 +-
 usr.sbin/pkg_install/Makefile.inc             | 13 +--
 usr.sbin/pkg_install/add/Makefile             |  7 +-
 usr.sbin/pkg_install/add/extract.c            |  2 +-
 usr.sbin/pkg_install/add/futil.c              |  2 +-
 usr.sbin/pkg_install/add/main.c               |  4 +-
 usr.sbin/pkg_install/add/perform.c            |  2 +-
 usr.sbin/pkg_install/create/Makefile          |  6 +-
 usr.sbin/pkg_install/create/main.c            |  4 +-
 usr.sbin/pkg_install/create/perform.c         |  2 +-
 usr.sbin/pkg_install/create/pl.c              |  2 +-
 usr.sbin/pkg_install/delete/Makefile          |  5 --
 usr.sbin/pkg_install/delete/main.c            |  4 +-
 usr.sbin/pkg_install/delete/perform.c         |  2 +-
 usr.sbin/pkg_install/info/Makefile            |  6 +-
 usr.sbin/pkg_install/info/info.h              | 39 ++++----
 usr.sbin/pkg_install/info/main.c              | 14 ++-
 usr.sbin/pkg_install/info/perform.c           |  2 +-
 usr.sbin/pkg_install/info/show.c              |  2 +-
 usr.sbin/pkg_install/lib/Makefile             | 11 ---
 usr.sbin/pkg_install/lib/pkgwrap.c            | 89 ------------------
 usr.sbin/pkg_install/updating/Makefile        |  5 --
 usr.sbin/pkg_install/updating/main.c          |  4 +-
 usr.sbin/pkg_install/version/Makefile         |  5 --
 usr.sbin/pkg_install/version/main.c           |  4 +-
 usr.sbin/pkg_install/version/perform.c        |  2 +-
 43 files changed, 250 insertions(+), 210 deletions(-)
 create mode 100644 lib/libpkg/Makefile
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/deps.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/exec.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/file.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/global.c (98%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/match.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/msg.c (91%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/pen.c (90%)
 rename usr.sbin/pkg_install/lib/lib.h => lib/libpkg/pkg.h (96%)
 create mode 100644 lib/libpkg/pkgwrap.c
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/plist.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/str.c (99%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/url.c (98%)
 rename {usr.sbin/pkg_install/lib => lib/libpkg}/version.c (98%)
 delete mode 100644 usr.sbin/pkg_install/lib/Makefile
 delete mode 100644 usr.sbin/pkg_install/lib/pkgwrap.c

diff --git a/Makefile.inc1 b/Makefile.inc1
index e54ae5ca925..df4604aeb71 100644
--- a/Makefile.inc1
+++ b/Makefile.inc1
@@ -1111,7 +1111,7 @@ _prebuild_libs=	${_kerberos5_lib_libasn1} ${_kerberos5_lib_libheimntlm} \
 		${_kerberos5_lib_libhx509} ${_kerberos5_lib_libkrb5} \
 		${_kerberos5_lib_libroken} \
 		lib/libbz2 lib/libcom_err lib/libcrypt \
-		lib/libexpat \
+		lib/libexpat lib/libfetch \
 		${_lib_libgssapi} ${_lib_libipx} \
 		lib/libkiconv lib/libkvm lib/libmd \
 		lib/ncurses/ncurses lib/ncurses/ncursesw \
@@ -1138,6 +1138,7 @@ _cddl_lib= cddl/lib
 _secure_lib_libcrypto= secure/lib/libcrypto
 _secure_lib_libssl= secure/lib/libssl
 lib/libradius__L secure/lib/libssl__L: secure/lib/libcrypto__L
+lib/libfetch__L: secure/lib/libcrypto__L secure/lib/libssl__L lib/libmd__L
 .if ${MK_OPENSSH} != "no"
 _secure_lib_libssh= secure/lib/libssh
 secure/lib/libssh__L: lib/libz__L secure/lib/libcrypto__L lib/libcrypt__L
@@ -1173,7 +1174,7 @@ _lib_libypclnt=	lib/libypclnt
 .endif
 
 .if ${MK_OPENSSL} == "no"
-lib/libradius__L: lib/libmd__L
+lib/libfetch__L lib/libradius__L: lib/libmd__L
 .endif
 
 .for _lib in ${_prereq_libs}
diff --git a/lib/Makefile b/lib/Makefile
index f3025b1f236..68f163e9e19 100644
--- a/lib/Makefile
+++ b/lib/Makefile
@@ -85,6 +85,7 @@ SUBDIR=	${SUBDIR_ORDERED} \
 	libopie \
 	libpam \
 	libpcap \
+	${_libpkg} \
 	${_libpmc} \
 	libproc \
 	librt \
@@ -197,6 +198,10 @@ _libmp=		libmp
 _libpmc=	libpmc
 .endif
 
+.if ${MK_PKGTOOLS} != "no"
+_libpkg=	libpkg
+.endif
+
 .if ${MK_SENDMAIL} != "no"
 _libmilter=	libmilter
 _libsm=		libsm
diff --git a/lib/libpkg/Makefile b/lib/libpkg/Makefile
new file mode 100644
index 00000000000..8f2b325f098
--- /dev/null
+++ b/lib/libpkg/Makefile
@@ -0,0 +1,47 @@
+# $FreeBSD$
+
+.include <bsd.own.mk>
+
+LIB=		pkg
+
+SHLIBDIR?=	/usr/lib
+SHLIB_MAJOR=	0
+
+SRCS=		deps.c		\
+		exec.c		\
+		file.c		\
+		global.c	\
+		match.c		\
+		msg.c		\
+		pen.c		\
+		pkgwrap.c	\
+		plist.c		\
+		str.c		\
+		url.c		\
+		version.c
+INCS=		pkg.h
+
+CFLAGS+=	-DYES_I_KNOW_THE_API_IS_RUBBISH_AND_IS_DOOMED_TO_CHANGE
+
+DPADD=		${LIBFETCH} ${LIBMD} ${LIBUTIL}
+LDADD=		-lfetch -lmd -lutil
+
+.if ${MK_OPENSSL} != "no"
+DPADD+=         ${LIBSSL} ${LIBCRYPTO}
+LDADD+=         -lssl -lcrypto
+.endif
+
+WARNS?=		3
+
+DATE!= grep LIBPKG_VERSION ${.CURDIR}/pkg.h | sed 's|.*[      ]||'
+
+distfile: clean
+	@(cd ${.CURDIR}/..; \
+		cp -r libpkg libpkg-${DATE}; \
+		tar -czf libpkg/libpkg-${DATE}.tar.gz \
+			--exclude .#* --exclude *~ --exclude CVS \
+			--exclude .svn --exclude libpkg-*.tar.gz \
+			libpkg-${DATE}; \
+		rm -rf libpkg-${DATE})
+
+.include <bsd.lib.mk>
diff --git a/usr.sbin/pkg_install/lib/deps.c b/lib/libpkg/deps.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/deps.c
rename to lib/libpkg/deps.c
index 66f44a9c6fe..101d0460219 100644
--- a/usr.sbin/pkg_install/lib/deps.c
+++ b/lib/libpkg/deps.c
@@ -22,7 +22,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <stdio.h>
 
diff --git a/usr.sbin/pkg_install/lib/exec.c b/lib/libpkg/exec.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/exec.c
rename to lib/libpkg/exec.c
index fc8220c5f39..04891d3cc09 100644
--- a/usr.sbin/pkg_install/lib/exec.c
+++ b/lib/libpkg/exec.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 
 /*
diff --git a/usr.sbin/pkg_install/lib/file.c b/lib/libpkg/file.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/file.c
rename to lib/libpkg/file.c
index c7ab9d6b297..7c95f995a2f 100644
--- a/usr.sbin/pkg_install/lib/file.c
+++ b/lib/libpkg/file.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <pwd.h>
 #include <time.h>
diff --git a/usr.sbin/pkg_install/lib/global.c b/lib/libpkg/global.c
similarity index 98%
rename from usr.sbin/pkg_install/lib/global.c
rename to lib/libpkg/global.c
index e136ec86a6f..8103126e693 100644
--- a/usr.sbin/pkg_install/lib/global.c
+++ b/lib/libpkg/global.c
@@ -22,7 +22,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 
 /* These are global for all utils */
 Boolean	Quiet		= FALSE;
diff --git a/usr.sbin/pkg_install/lib/match.c b/lib/libpkg/match.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/match.c
rename to lib/libpkg/match.c
index 6c1b2bfcf00..ba6544224dc 100644
--- a/usr.sbin/pkg_install/lib/match.c
+++ b/lib/libpkg/match.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <fnmatch.h>
 #include <fts.h>
@@ -292,7 +292,7 @@ matchallbyorigin(const char **origins, int *retval)
 		break;
 	    }
 	}
-	if (cmd != PLIST_ORIGIN && ( Verbose || 0 != strncmp("bsdpan-", installed[i], 7 ) ) )
+	if (cmd != PLIST_ORIGIN && 0 != strncmp("bsdpan-", installed[i], 7))
 	    warnx("package %s has no origin recorded", installed[i]);
 	fclose(fp);
     }
diff --git a/usr.sbin/pkg_install/lib/msg.c b/lib/libpkg/msg.c
similarity index 91%
rename from usr.sbin/pkg_install/lib/msg.c
rename to lib/libpkg/msg.c
index 5b17624016d..0d25ad19900 100644
--- a/usr.sbin/pkg_install/lib/msg.c
+++ b/lib/libpkg/msg.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <paths.h>
 
@@ -61,12 +61,7 @@ y_or_n(Boolean def, const char *msg, ...)
 	else
 	    fprintf(stderr, " [no]? ");
 	fflush(stderr);
-	if (AutoAnswer) {
-	    ch = (AutoAnswer == YES) ? 'Y' : 'N';
-	    fprintf(stderr, "%c\n", ch);
-	}
-	else
-	    ch = toupper(fgetc(tty));
+	ch = toupper(fgetc(tty));
 	if (ch == '\n')
 	    ch = (def) ? 'Y' : 'N';
     }
diff --git a/usr.sbin/pkg_install/lib/pen.c b/lib/libpkg/pen.c
similarity index 90%
rename from usr.sbin/pkg_install/lib/pen.c
rename to lib/libpkg/pen.c
index 2b405a3a43a..6e30445b759 100644
--- a/usr.sbin/pkg_install/lib/pen.c
+++ b/lib/libpkg/pen.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <libutil.h>
 #include <libgen.h>
@@ -103,7 +103,7 @@ popPen(char *pen)
 const char *
 make_playpen(char *pen, off_t sz)
 {
-    char humbuf1[6], humbuf2[6];
+    char humbuf[6];
     char cwd[FILENAME_MAX];
 
     if (!find_play_pen(pen, sz))
@@ -114,22 +114,15 @@ make_playpen(char *pen, off_t sz)
 	errx(2, "%s: can't mktemp '%s'", __func__, pen);
     }
 
-    if (Verbose) {
-	if (sz) {
-	    humanize_number(humbuf1, sizeof humbuf1, sz, "", HN_AUTOSCALE,
-	        HN_NOSPACE);
-	    humanize_number(humbuf2, sizeof humbuf2, min_free(pen),
-	        "", HN_AUTOSCALE, HN_NOSPACE);
-	    fprintf(stderr, "Requested space: %s bytes, free space: %s bytes in %s\n", humbuf1, humbuf2, pen);
-	}
-    }
+    humanize_number(humbuf, sizeof humbuf, sz, "", HN_AUTOSCALE, HN_NOSPACE);
 
     if (min_free(pen) < sz) {
 	rmdir(pen);
 	cleanup(0);
 	errx(2, "%s: not enough free space to create '%s'.\n"
 	     "Please set your PKG_TMPDIR environment variable to a location\n"
-	     "with more space and\ntry the command again", __func__, pen);
+	     "with at least %s and try the command again",
+	     __func__, humbuf, pen);
     }
 
     if (!getcwd(cwd, FILENAME_MAX)) {
diff --git a/usr.sbin/pkg_install/lib/lib.h b/lib/libpkg/pkg.h
similarity index 96%
rename from usr.sbin/pkg_install/lib/lib.h
rename to lib/libpkg/pkg.h
index 089bfa3630b..12e35623bc8 100644
--- a/usr.sbin/pkg_install/lib/lib.h
+++ b/lib/libpkg/pkg.h
@@ -23,6 +23,10 @@
 #ifndef _INST_LIB_LIB_H_
 #define _INST_LIB_LIB_H_
 
+#ifndef YES_I_KNOW_THE_API_IS_RUBBISH_AND_IS_DOOMED_TO_CHANGE
+#error "You obviously have no idea what you're doing."
+#endif
+
 /* Includes */
 #include <sys/param.h>
 #include <sys/file.h>
@@ -96,13 +100,12 @@
 #define PKG_PREFIX_VNAME	"PKG_PREFIX"
 
 /*
- * Version of the package tools - increase whenever you make a change
+ * Version of the package library - increase whenever you make a change
  * in the code that is not cosmetic only.
  */
-#define PKG_INSTALL_VERSION	20100403
+#define LIBPKG_VERSION 20100423
 
 #define PKG_WRAPCONF_FNAME	"/var/db/pkg_install.conf"
-#define main(argc, argv)	real_main(argc, argv)
 
 /* Version numbers to assist with changes in package file format */
 #define PLIST_FMT_VER_MAJOR	1
@@ -214,7 +217,7 @@ Boolean 	make_preserve_name(char *, int, const char *, const char *);
 
 /* For all */
 int		pkg_perform(char **);
-int		real_main(int, char **);
+void		pkg_wrap(long, char **);
 
 /* Query installed packages */
 char		**matchinstalled(match_t, char **, int *);
@@ -229,6 +232,7 @@ int		chkifdepends(const char *, const char *);
 int		requiredby(const char *, struct reqr_by_head **, Boolean, Boolean);
 
 /* Version */
+int		libpkg_version(void);
 int		verscmp(Package *, int, int);
 int		version_cmp(const char *, const char *);
 
diff --git a/lib/libpkg/pkgwrap.c b/lib/libpkg/pkgwrap.c
new file mode 100644
index 00000000000..9c8c0c3d0a2
--- /dev/null
+++ b/lib/libpkg/pkgwrap.c
@@ -0,0 +1,90 @@
+/*
+ * FreeBSD install - a package for the installation and maintenance
+ * of non-core utilities.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * Maxim Sobolev
+ * 8 September 2002
+ *
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include "pkg.h"
+#include <ctype.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <unistd.h>
+
+extern char **environ;
+
+void
+pkg_wrap(long curver, char **argv)
+{
+    FILE* f;
+    char ver[9];			/* Format is: 'YYYYMMDD\0' */
+    char buffer[FILENAME_MAX+10];	/* Format is: 'YYYYMMDD <path>' */
+    char cmd[FILENAME_MAX+5];		/* Format is: '<path> -PPq' */
+    char *path, *cp;
+    long ptver, lpver;
+
+    if (getenv("PKG_NOWRAP") != NULL)
+	goto nowrap;
+
+    setenv("PKG_NOWRAP", "1", 1);
+
+    /* Get alternative location for package tools. */
+    if ((f = fopen(PKG_WRAPCONF_FNAME, "r")) == NULL) {
+	goto nowrap;
+    } else {
+	if (get_string(buffer, FILENAME_MAX+9, f) == NULL) {
+	    goto nowrap;
+	} else {
+	    if ((path = strrchr(buffer, ' ')) == NULL) {
+		goto nowrap;
+	    } else {
+		*path++ = '\0';
+	    }
+	}
+    }
+
+    if ((cp = strrchr(argv[0], '/')) == NULL) {
+	cp = argv[0];
+    } else {
+	cp++;
+    }
+
+    /* Get version of the other pkg_install and libpkg */
+    snprintf(cmd, FILENAME_MAX+10, "%s/%s -PPq", path, cp);
+    if ((f = popen(cmd, "r")) == NULL) {
+	perror("popen()");
+	goto nowrap;
+    } else {
+	if (get_string(ver, 9, f) == NULL)
+	    goto nowrap; 
+	else
+	    ptver = strtol(ver, NULL, 10);
+	if (get_string(ver, 9, f) == NULL)
+	    goto nowrap;
+	else
+	    lpver = strtol(ver, NULL, 10);
+	pclose(f);
+    }
+
+    if ((lpver >= LIBPKG_VERSION) && (ptver > curver)) {
+	snprintf(cmd, FILENAME_MAX, "%s/%s", path, cp);
+	execve(cmd, argv, environ);
+    }
+
+nowrap:
+    unsetenv("PKG_NOWRAP");
+}
diff --git a/usr.sbin/pkg_install/lib/plist.c b/lib/libpkg/plist.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/plist.c
rename to lib/libpkg/plist.c
index 545b541c3f1..b14ac20aafc 100644
--- a/usr.sbin/pkg_install/lib/plist.c
+++ b/lib/libpkg/plist.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <md5.h>
 
diff --git a/usr.sbin/pkg_install/lib/str.c b/lib/libpkg/str.c
similarity index 99%
rename from usr.sbin/pkg_install/lib/str.c
rename to lib/libpkg/str.c
index 0d9e288b757..c26e6cce660 100644
--- a/usr.sbin/pkg_install/lib/str.c
+++ b/lib/libpkg/str.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 
 char *
 strconcat(const char *s1, const char *s2)
diff --git a/usr.sbin/pkg_install/lib/url.c b/lib/libpkg/url.c
similarity index 98%
rename from usr.sbin/pkg_install/lib/url.c
rename to lib/libpkg/url.c
index 8c553471199..4c108499428 100644
--- a/usr.sbin/pkg_install/lib/url.c
+++ b/lib/libpkg/url.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 #include <fetch.h>
 #include <libgen.h>
@@ -109,7 +109,7 @@ fileGetURL(const char *base, const char *spec, int keep_package)
 	printf("Error: Unable to get %s: %s\n",
 	       fname, fetchLastErrString);
 	/* If the fetch fails, yank the package. */
-	if (keep_package && unlink(pkg) < 0 && Verbose) {
+	if (keep_package && unlink(pkg) < 0) {
 	    warnx("failed to remove partially fetched package: %s", pkg);
 	}
 	return NULL;
diff --git a/usr.sbin/pkg_install/lib/version.c b/lib/libpkg/version.c
similarity index 98%
rename from usr.sbin/pkg_install/lib/version.c
rename to lib/libpkg/version.c
index d9c4fe7e070..1b7bb5bdbd9 100644
--- a/usr.sbin/pkg_install/lib/version.c
+++ b/lib/libpkg/version.c
@@ -19,9 +19,19 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include "pkg.h"
 #include <err.h>
 
+/*
+ * This routine could easily go somewhere else.
+ *
+ */
+int
+libpkg_version(void)
+{
+	return LIBPKG_VERSION;
+}
+
 /*
  * Routines to assist with PLIST_FMT_VER numbers in the packing
  * lists.
diff --git a/share/mk/bsd.libnames.mk b/share/mk/bsd.libnames.mk
index c80fe234279..70f9c063921 100644
--- a/share/mk/bsd.libnames.mk
+++ b/share/mk/bsd.libnames.mk
@@ -128,6 +128,7 @@ MINUSLPAM+=	-lypclnt
 
 LIBPANEL?=	${DESTDIR}${LIBDIR}/libpanel.a
 LIBPCAP?=	${DESTDIR}${LIBDIR}/libpcap.a
+LIBPKG?=	${DESTDIR}${LIBDIR}/libpkg.a
 LIBPMC?=	${DESTDIR}${LIBDIR}/libpmc.a
 LIBPROC?=	${DESTDIR}${LIBDIR}/libproc.a
 LIBPTHREAD?=	${DESTDIR}${LIBDIR}/libpthread.a
diff --git a/usr.sbin/pkg_install/Makefile b/usr.sbin/pkg_install/Makefile
index bf1a2135cc7..0aa19415441 100644
--- a/usr.sbin/pkg_install/Makefile
+++ b/usr.sbin/pkg_install/Makefile
@@ -2,11 +2,11 @@
 
 .include <bsd.own.mk>
 
-SUBDIR=	lib add create delete info updating version
+SUBDIR=	add create delete info updating version
 
 .include <bsd.subdir.mk>
 
-DATE!=	grep PKG_INSTALL_VERSION ${.CURDIR}/lib/lib.h | sed 's|.*[ 	]||'
+DATE!=	grep PKG_INSTALL_VERSION ${.CURDIR}/Makefile.inc | sed 's|.*=||'
 
 distfile: clean
 	@(cd ${.CURDIR}/..; \
diff --git a/usr.sbin/pkg_install/Makefile.inc b/usr.sbin/pkg_install/Makefile.inc
index 2fa20aa1c39..5c9e2661a58 100644
--- a/usr.sbin/pkg_install/Makefile.inc
+++ b/usr.sbin/pkg_install/Makefile.inc
@@ -2,16 +2,11 @@
 
 .include <bsd.own.mk>
 
-LIBINSTALL=	${.OBJDIR}/../lib/libinstall.a
+CFLAGS+=	-DPKG_INSTALL_VERSION=20100423
+CFLAGS+=	-DYES_I_KNOW_THE_API_IS_RUBBISH_AND_IS_DOOMED_TO_CHANGE
 
-DPADD+=		${LIBUTIL}
-LDADD+=		-lutil
-
-.if ${MK_OPENSSL} != "no" && \
-    defined(LDADD) && ${LDADD:M-lfetch} != ""
-DPADD+=		${LIBSSL} ${LIBCRYPTO}
-LDADD+=		-lssl -lcrypto
-.endif
+DPADD+=		${LIBPKG}
+LDADD+=		-lpkg
 
 # Inherit BINDIR from one level up.
 .include "../Makefile.inc"
diff --git a/usr.sbin/pkg_install/add/Makefile b/usr.sbin/pkg_install/add/Makefile
index 89988e78b22..03a43f068b7 100644
--- a/usr.sbin/pkg_install/add/Makefile
+++ b/usr.sbin/pkg_install/add/Makefile
@@ -1,14 +1,11 @@
 # $FreeBSD$
 
+.include <bsd.own.mk>
+
 PROG=	pkg_add
 SRCS=	main.c perform.c futil.c extract.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WARNS?=	3
 WFORMAT?=	1
 
-DPADD=	${LIBINSTALL} ${LIBFETCH} ${LIBMD}
-LDADD=	${LIBINSTALL} -lfetch -lmd
-
 .include <bsd.prog.mk>
diff --git a/usr.sbin/pkg_install/add/extract.c b/usr.sbin/pkg_install/add/extract.c
index 732a13fea44..6c2c4fc7b92 100644
--- a/usr.sbin/pkg_install/add/extract.c
+++ b/usr.sbin/pkg_install/add/extract.c
@@ -23,7 +23,7 @@ __FBSDID("$FreeBSD$");
 
 #include <ctype.h>
 #include <err.h>
-#include "lib.h"
+#include <pkg.h>
 #include "add.h"
 
 
diff --git a/usr.sbin/pkg_install/add/futil.c b/usr.sbin/pkg_install/add/futil.c
index ca095ea7ae2..c5253202ca5 100644
--- a/usr.sbin/pkg_install/add/futil.c
+++ b/usr.sbin/pkg_install/add/futil.c
@@ -22,7 +22,7 @@
 __FBSDID("$FreeBSD$");
 
 #include <err.h>
-#include "lib.h"
+#include <pkg.h>
 #include "add.h"
 
 /*
diff --git a/usr.sbin/pkg_install/add/main.c b/usr.sbin/pkg_install/add/main.c
index d9b381bdd9b..762b61a442e 100644
--- a/usr.sbin/pkg_install/add/main.c
+++ b/usr.sbin/pkg_install/add/main.c
@@ -26,7 +26,7 @@ __FBSDID("$FreeBSD$");
 #include <err.h>
 #include <getopt.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "add.h"
 
 char	*Prefix		= NULL;
@@ -128,6 +128,8 @@ main(int argc, char **argv)
     static char temppackageroot[MAXPATHLEN];
     static char pkgaddpath[MAXPATHLEN];
 
+    pkg_wrap(PKG_INSTALL_VERSION, argv);
+
     if (*argv[0] != '/' && strchr(argv[0], '/') != NULL)
 	PkgAddCmd = realpath(argv[0], pkgaddpath);
     else
diff --git a/usr.sbin/pkg_install/add/perform.c b/usr.sbin/pkg_install/add/perform.c
index b23cd5232a3..653a1d6296f 100644
--- a/usr.sbin/pkg_install/add/perform.c
+++ b/usr.sbin/pkg_install/add/perform.c
@@ -23,7 +23,7 @@ __FBSDID("$FreeBSD$");
 
 #include <err.h>
 #include <paths.h>
-#include "lib.h"
+#include <pkg.h>
 #include "add.h"
 
 #include <libgen.h>
diff --git a/usr.sbin/pkg_install/create/Makefile b/usr.sbin/pkg_install/create/Makefile
index 42718c6f927..333403778fe 100644
--- a/usr.sbin/pkg_install/create/Makefile
+++ b/usr.sbin/pkg_install/create/Makefile
@@ -3,12 +3,10 @@
 PROG=	pkg_create
 SRCS=	main.c perform.c pl.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WARNS?=	3
 WFORMAT?=	1
 
-DPADD=	${LIBINSTALL} ${LIBMD}
-LDADD=	${LIBINSTALL} -lmd
+DPADD=	${LIBMD}
+LDADD=	-lmd
 
 .include <bsd.prog.mk>
diff --git a/usr.sbin/pkg_install/create/main.c b/usr.sbin/pkg_install/create/main.c
index d85392dc9c7..e463d1f295b 100644
--- a/usr.sbin/pkg_install/create/main.c
+++ b/usr.sbin/pkg_install/create/main.c
@@ -15,7 +15,7 @@ __FBSDID("$FreeBSD$");
 #include <getopt.h>
 #include <err.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "create.h"
 
 match_t	MatchType	= MATCH_GLOB;
@@ -72,6 +72,8 @@ main(int argc, char **argv)
     int ch;
     char **pkgs, **start, *tmp;
 
+    pkg_wrap(PKG_INSTALL_VERSION, argv);
+
     pkgs = start = argv;
     while ((ch = getopt_long(argc, argv, opts, longopts, NULL)) != -1)
 	switch(ch) {
diff --git a/usr.sbin/pkg_install/create/perform.c b/usr.sbin/pkg_install/create/perform.c
index e4f0d2eae58..0e14095c815 100644
--- a/usr.sbin/pkg_install/create/perform.c
+++ b/usr.sbin/pkg_install/create/perform.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include <pkg.h>
 #include "create.h"
 
 #include <err.h>
diff --git a/usr.sbin/pkg_install/create/pl.c b/usr.sbin/pkg_install/create/pl.c
index 18bbaf24734..fe62d421014 100644
--- a/usr.sbin/pkg_install/create/pl.c
+++ b/usr.sbin/pkg_install/create/pl.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include <pkg.h>
 #include "create.h"
 #include <errno.h>
 #include <err.h>
diff --git a/usr.sbin/pkg_install/delete/Makefile b/usr.sbin/pkg_install/delete/Makefile
index c9a0fdebea2..4f3b3906e63 100644
--- a/usr.sbin/pkg_install/delete/Makefile
+++ b/usr.sbin/pkg_install/delete/Makefile
@@ -3,11 +3,6 @@
 PROG=	pkg_delete
 SRCS=	main.c perform.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WFORMAT?=	1
 
-DPADD=	${LIBINSTALL} ${LIBMD}
-LDADD=	${LIBINSTALL} -lmd
-
 .include <bsd.prog.mk>
diff --git a/usr.sbin/pkg_install/delete/main.c b/usr.sbin/pkg_install/delete/main.c
index f09a432f2bb..7677fe9d44b 100644
--- a/usr.sbin/pkg_install/delete/main.c
+++ b/usr.sbin/pkg_install/delete/main.c
@@ -27,7 +27,7 @@ __FBSDID("$FreeBSD$");
 #include <getopt.h>
 #include <err.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "delete.h"
 
 char	*Prefix		= NULL;
@@ -67,6 +67,8 @@ main(int argc, char **argv)
     const char *tmp;
     struct stat stat_s;
 
+    pkg_wrap(PKG_INSTALL_VERSION, argv);
+
     pkgs = start = argv;
     while ((ch = getopt_long(argc, argv, opts, longopts, NULL)) != -1)
 	switch(ch) {
diff --git a/usr.sbin/pkg_install/delete/perform.c b/usr.sbin/pkg_install/delete/perform.c
index 2557dbcd105..02f9717a091 100644
--- a/usr.sbin/pkg_install/delete/perform.c
+++ b/usr.sbin/pkg_install/delete/perform.c
@@ -22,7 +22,7 @@
 __FBSDID("$FreeBSD$");
 
 #include <err.h>
-#include "lib.h"
+#include <pkg.h>
 #include "delete.h"
 
 static int pkg_do(char *);
diff --git a/usr.sbin/pkg_install/info/Makefile b/usr.sbin/pkg_install/info/Makefile
index 485cb226321..ba3909d17da 100644
--- a/usr.sbin/pkg_install/info/Makefile
+++ b/usr.sbin/pkg_install/info/Makefile
@@ -3,11 +3,9 @@
 PROG=	pkg_info
 SRCS=	main.c perform.c show.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WFORMAT?=	1
 
-DPADD=	${LIBINSTALL} ${LIBFETCH} ${LIBMD}
-LDADD=	${LIBINSTALL} -lfetch -lmd
+DPADD=	${LIBMD}
+LDADD=	-lmd
 
 .include <bsd.prog.mk>
diff --git a/usr.sbin/pkg_install/info/info.h b/usr.sbin/pkg_install/info/info.h
index 46e29b1202d..68ec7107a3c 100644
--- a/usr.sbin/pkg_install/info/info.h
+++ b/usr.sbin/pkg_install/info/info.h
@@ -33,25 +33,26 @@
 #define MAXNAMESIZE  20
 #endif
 
-#define SHOW_COMMENT	0x00001
-#define SHOW_DESC	0x00002
-#define SHOW_PLIST	0x00004
-#define SHOW_INSTALL	0x00008
-#define SHOW_DEINSTALL	0x00010
-#define SHOW_REQUIRE	0x00020
-#define SHOW_PREFIX	0x00040
-#define SHOW_INDEX	0x00080
-#define SHOW_FILES	0x00100
-#define SHOW_DISPLAY	0x00200
-#define SHOW_REQBY	0x00400
-#define SHOW_MTREE	0x00800
-#define SHOW_SIZE	0x01000
-#define SHOW_ORIGIN	0x02000
-#define SHOW_CKSUM	0x04000
-#define SHOW_FMTREV	0x08000
-#define SHOW_PTREV	0x10000
-#define SHOW_DEPEND	0x20000
-#define SHOW_PKGNAME	0x40000
+#define SHOW_COMMENT	0x000001
+#define SHOW_DESC	0x000002
+#define SHOW_PLIST	0x000004
+#define SHOW_INSTALL	0x000008
+#define SHOW_DEINSTALL	0x000010
+#define SHOW_REQUIRE	0x000020
+#define SHOW_PREFIX	0x000040
+#define SHOW_INDEX	0x000080
+#define SHOW_FILES	0x000100
+#define SHOW_DISPLAY	0x000200
+#define SHOW_REQBY	0x000400
+#define SHOW_MTREE	0x000800
+#define SHOW_SIZE	0x001000
+#define SHOW_ORIGIN	0x002000
+#define SHOW_CKSUM	0x004000
+#define SHOW_FMTREV	0x008000
+#define SHOW_PTREV	0x010000
+#define SHOW_DEPEND	0x020000
+#define SHOW_PKGNAME	0x040000
+#define SHOW_LPREV	0x100000
 
 struct which_entry {
     TAILQ_ENTRY(which_entry) next;
diff --git a/usr.sbin/pkg_install/info/main.c b/usr.sbin/pkg_install/info/main.c
index 2de638e4cad..ca4e477f999 100644
--- a/usr.sbin/pkg_install/info/main.c
+++ b/usr.sbin/pkg_install/info/main.c
@@ -25,7 +25,7 @@ __FBSDID("$FreeBSD$");
 #include <getopt.h>
 #include <err.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "info.h"
 
 int	Flags		= 0;
@@ -68,6 +68,8 @@ main(int argc, char **argv)
     char **pkgs, **start;
     char *pkgs_split;
 
+    pkg_wrap(PKG_INSTALL_VERSION, argv);
+
     whead = malloc(sizeof(struct which_head));
     if (whead == NULL)
 	err(2, NULL);
@@ -225,7 +227,10 @@ main(int argc, char **argv)
 	    }
 
 	case 'P':
-	    Flags = SHOW_PTREV;
+	    if (Flags & SHOW_PTREV)
+		Flags |= SHOW_LPREV;
+	    else
+		Flags = SHOW_PTREV;
 	    break;
 
 	case 'h':
@@ -242,6 +247,11 @@ main(int argc, char **argv)
 	if (!Quiet)
 	    printf("Package tools revision: ");
 	printf("%d\n", PKG_INSTALL_VERSION);
+	if (Flags & SHOW_LPREV) {
+	    if (!Quiet)
+		printf("Libpkg revision: ");
+	    printf("%d\n", libpkg_version());
+	}
 	exit(0);
     }
 
diff --git a/usr.sbin/pkg_install/info/perform.c b/usr.sbin/pkg_install/info/perform.c
index 09cad78ad3f..d2956120822 100644
--- a/usr.sbin/pkg_install/info/perform.c
+++ b/usr.sbin/pkg_install/info/perform.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include <pkg.h>
 #include "info.h"
 #include <err.h>
 #include <signal.h>
diff --git a/usr.sbin/pkg_install/info/show.c b/usr.sbin/pkg_install/info/show.c
index c65c312a506..0d3b4ad73d2 100644
--- a/usr.sbin/pkg_install/info/show.c
+++ b/usr.sbin/pkg_install/info/show.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include <pkg.h>
 #include "info.h"
 #include <err.h>
 #include <stdlib.h>
diff --git a/usr.sbin/pkg_install/lib/Makefile b/usr.sbin/pkg_install/lib/Makefile
deleted file mode 100644
index 84a41b8d7f1..00000000000
--- a/usr.sbin/pkg_install/lib/Makefile
+++ /dev/null
@@ -1,11 +0,0 @@
-# $FreeBSD$
-
-LIB=	install
-INTERNALLIB=
-SRCS=	file.c msg.c plist.c str.c exec.c global.c pen.c match.c \
-	deps.c version.c pkgwrap.c url.c
-
-WARNS?=	3
-WFORMAT?=	1
-
-.include <bsd.lib.mk>
diff --git a/usr.sbin/pkg_install/lib/pkgwrap.c b/usr.sbin/pkg_install/lib/pkgwrap.c
deleted file mode 100644
index cbd15cd7b44..00000000000
--- a/usr.sbin/pkg_install/lib/pkgwrap.c
+++ /dev/null
@@ -1,89 +0,0 @@
-/*
- * FreeBSD install - a package for the installation and maintenance
- * of non-core utilities.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- *
- * Maxim Sobolev
- * 8 September 2002
- *
- */
-
-#include <sys/cdefs.h>
-__FBSDID("$FreeBSD$");
-
-#include "lib.h"
-#include <ctype.h>
-#include <stdio.h>
-#include <stdlib.h>
-#include <unistd.h>
-
-#undef main
-
-#define SEPARATORS " \t"
-
-extern char **environ;
-
-int
-main(int argc, char **argv)
-{
-    FILE *f;
-    char buffer[FILENAME_MAX], *cp, *verstr;
-    int len;
-
-    if (getenv("PKG_NOWRAP") != NULL)
-	goto nowrap;
-    f = fopen(PKG_WRAPCONF_FNAME, "r");
-    if (f == NULL)
-	goto nowrap;
-    cp = fgets(buffer, 256, f);
-    fclose(f);
-    if (cp == NULL)
-	goto nowrap;
-    len = strlen(cp);
-    if (cp[len - 1] == '\n')
-	cp[len - 1] = '\0';
-    while (strchr(SEPARATORS, *cp) != NULL)
-	cp++;
-    verstr = cp;
-    cp = strpbrk(cp, SEPARATORS);
-    if (cp == NULL)
-	goto nowrap;
-    *cp = '\0';
-    for (cp = verstr; *cp != '\0'; cp++)
-	if (isdigit(*cp) == 0)
-	    goto nowrap;
-    if (atoi(verstr) < PKG_INSTALL_VERSION)
-	goto nowrap;
-    cp++;
-    while (*cp != '\0' && strchr(SEPARATORS, *cp) != NULL)
-	cp++;
-    if (*cp == '\0')
-	goto nowrap;
-    bcopy(cp, buffer, strlen(cp) + 1);
-    cp = strpbrk(buffer, SEPARATORS);
-    if (cp != NULL)
-	*cp = '\0';
-    if (!isdir(buffer))
-	goto nowrap;
-    cp = strrchr(argv[0], '/');
-    if (cp == NULL)
-	cp = argv[0];
-    else
-	cp++;
-    strlcat(buffer, "/", sizeof(buffer));
-    strlcat(buffer, cp, sizeof(buffer));
-    setenv("PKG_NOWRAP", "1", 1);
-    execve(buffer, argv, environ);
-
-nowrap:
-    unsetenv("PKG_NOWRAP");
-    return(real_main(argc, argv));
-}
diff --git a/usr.sbin/pkg_install/updating/Makefile b/usr.sbin/pkg_install/updating/Makefile
index b0d3689779e..f5b7525ddb5 100644
--- a/usr.sbin/pkg_install/updating/Makefile
+++ b/usr.sbin/pkg_install/updating/Makefile
@@ -3,11 +3,6 @@
 PROG=	pkg_updating
 SRCS=	main.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WFORMAT?= 1
 
-DPADD=	${LIBINSTALL} ${LIBFETCH} ${LIBMD}
-LDADD=	${LIBINSTALL} -lfetch -lmd
-
 .include <bsd.prog.mk>
diff --git a/usr.sbin/pkg_install/updating/main.c b/usr.sbin/pkg_install/updating/main.c
index 0ab2ec0fe8e..993ccd52310 100644
--- a/usr.sbin/pkg_install/updating/main.c
+++ b/usr.sbin/pkg_install/updating/main.c
@@ -19,7 +19,7 @@ __FBSDID("$FreeBSD$");
 #include <sysexits.h>
 #include <getopt.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "pathnames.h"
 
 typedef struct installedport {
@@ -87,6 +87,8 @@ main(int argc, char *argv[])
 	DIR *dir;
 	FILE *fd;
 
+	pkg_wrap(PKG_INSTALL_VERSION, argv);
+
 	while ((ch = getopt_long(argc, argv, opts, longopts, NULL)) != -1) {
 		switch (ch) {
 			case 'd':
diff --git a/usr.sbin/pkg_install/version/Makefile b/usr.sbin/pkg_install/version/Makefile
index 3e1d7a522f2..fb079e34c7a 100644
--- a/usr.sbin/pkg_install/version/Makefile
+++ b/usr.sbin/pkg_install/version/Makefile
@@ -3,13 +3,8 @@
 PROG=	pkg_version
 SRCS=	main.c perform.c
 
-CFLAGS+= -I${.CURDIR}/../lib
-
 WFORMAT?=	1
 
-DPADD=	${LIBINSTALL} ${LIBFETCH} ${LIBMD}
-LDADD=	${LIBINSTALL} -lfetch -lmd
-
 test:
 	sh ${.CURDIR}/test-pkg_version.sh
 
diff --git a/usr.sbin/pkg_install/version/main.c b/usr.sbin/pkg_install/version/main.c
index cad8583c23b..42384970603 100644
--- a/usr.sbin/pkg_install/version/main.c
+++ b/usr.sbin/pkg_install/version/main.c
@@ -25,7 +25,7 @@ __FBSDID("$FreeBSD$");
 #include <getopt.h>
 #include <err.h>
 
-#include "lib.h"
+#include <pkg.h>
 #include "version.h"
 
 char	*LimitChars = NULL;
@@ -58,6 +58,8 @@ main(int argc, char **argv)
 {
     int ch, cmp = 0;
 
+    pkg_wrap(PKG_INSTALL_VERSION, argv);
+
     if (argc == 4 && !strcmp(argv[1], "-t")) {
 	cmp = version_cmp(argv[2], argv[3]);
 	printf(cmp > 0 ? ">\n" : (cmp < 0 ? "<\n" : "=\n"));
diff --git a/usr.sbin/pkg_install/version/perform.c b/usr.sbin/pkg_install/version/perform.c
index dc85696cfc1..79575a32ba2 100644
--- a/usr.sbin/pkg_install/version/perform.c
+++ b/usr.sbin/pkg_install/version/perform.c
@@ -21,7 +21,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
-#include "lib.h"
+#include <pkg.h>
 #include "version.h"
 #include <err.h>
 #include <fetch.h>

From 2f826fdf53c298a18569d06a639063283858dc84 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Fri, 23 Apr 2010 14:32:58 +0000
Subject: [PATCH 217/532] Remove one zero from the double-0. This code doesn't
 have a license to kill.

MFC after:	3 days
---
 sys/kern/kern_descrip.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/kern/kern_descrip.c b/sys/kern/kern_descrip.c
index 676de650524..302ca5eeaf4 100644
--- a/sys/kern/kern_descrip.c
+++ b/sys/kern/kern_descrip.c
@@ -2896,7 +2896,7 @@ sysctl_kern_proc_ofiledesc(SYSCTL_HANDLER_ARGS)
 				free(sa, M_SONAME);
 			}
 			if (so->so_proto->pr_usrreqs->pru_peeraddr(so, &sa)
-			    == 00 && sa->sa_len <= sizeof(kif->kf_sa_peer)) {
+			    == 0 && sa->sa_len <= sizeof(kif->kf_sa_peer)) {
 				bcopy(sa, &kif->kf_sa_peer, sa->sa_len);
 				free(sa, M_SONAME);
 			}
@@ -3149,7 +3149,7 @@ sysctl_kern_proc_filedesc(SYSCTL_HANDLER_ARGS)
 				free(sa, M_SONAME);
 			}
 			if (so->so_proto->pr_usrreqs->pru_peeraddr(so, &sa)
-			    == 00 && sa->sa_len <= sizeof(kif->kf_sa_peer)) {
+			    == 0 && sa->sa_len <= sizeof(kif->kf_sa_peer)) {
 				bcopy(sa, &kif->kf_sa_peer, sa->sa_len);
 				free(sa, M_SONAME);
 			}

From 05d841b38b42704c92dd656dbffc988600ea7656 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Fri, 23 Apr 2010 14:48:30 +0000
Subject: [PATCH 218/532] Address some WITNESS panics that occur when using the
 via driver.

Some of these cases should be safe in a non-atomic fashion, however
since all of the driver ioctls are locked, a lot of work is required to
fix it correctly.  Just don't sleep now.

MFC after:	2 weeks
---
 sys/dev/drm/drmP.h        |  2 +-
 sys/dev/drm/drm_hashtab.c |  3 ++-
 sys/dev/drm/drm_mm.c      |  3 ++-
 sys/dev/drm/drm_sman.c    | 10 +++++-----
 4 files changed, 10 insertions(+), 8 deletions(-)

diff --git a/sys/dev/drm/drmP.h b/sys/dev/drm/drmP.h
index 26bda43659d..3699460a706 100644
--- a/sys/dev/drm/drmP.h
+++ b/sys/dev/drm/drmP.h
@@ -228,7 +228,7 @@ enum {
 #define DRM_MTRR_WC		MDF_WRITECOMBINE
 #define jiffies			ticks
 
-typedef unsigned long dma_addr_t;
+typedef vm_paddr_t dma_addr_t;
 typedef u_int64_t u64;
 typedef u_int32_t u32;
 typedef u_int16_t u16;
diff --git a/sys/dev/drm/drm_hashtab.c b/sys/dev/drm/drm_hashtab.c
index e98f1026cf5..360c02bfada 100644
--- a/sys/dev/drm/drm_hashtab.c
+++ b/sys/dev/drm/drm_hashtab.c
@@ -46,7 +46,8 @@ int drm_ht_create(struct drm_open_hash *ht, unsigned int order)
 	ht->size = 1 << order;
 	ht->order = order;
 	ht->table = NULL;
-	ht->table = hashinit(ht->size, DRM_MEM_HASHTAB, &ht->mask);
+	ht->table = hashinit_flags(ht->size, DRM_MEM_HASHTAB, &ht->mask,
+	    HASH_NOWAIT);
 	if (!ht->table) {
 		DRM_ERROR("Out of memory for hash table\n");
 		return -ENOMEM;
diff --git a/sys/dev/drm/drm_mm.c b/sys/dev/drm/drm_mm.c
index 344436fd22d..bab36c1117a 100644
--- a/sys/dev/drm/drm_mm.c
+++ b/sys/dev/drm/drm_mm.c
@@ -333,7 +333,8 @@ int drm_mm_init(struct drm_mm * mm, unsigned long start, unsigned long size)
 	mm->num_unused = 0;
 	mtx_init(&mm->unused_lock, "drm_unused", NULL, MTX_DEF);
 
-	return drm_mm_create_tail_node(mm, start, size, 0);
+	/* XXX This could be non-atomic but gets called from a locked path */
+	return drm_mm_create_tail_node(mm, start, size, 1);
 }
 
 void drm_mm_takedown(struct drm_mm * mm)
diff --git a/sys/dev/drm/drm_sman.c b/sys/dev/drm/drm_sman.c
index 9f1132ce2b2..32f9eb9e465 100644
--- a/sys/dev/drm/drm_sman.c
+++ b/sys/dev/drm/drm_sman.c
@@ -96,7 +96,8 @@ static void *drm_sman_mm_allocate(void *private, unsigned long size,
 	if (!tmp) {
 		return NULL;
 	}
-	tmp = drm_mm_get_block(tmp, size, alignment);
+	/* This could be non-atomic, but we are called from a locked path */
+	tmp = drm_mm_get_block_atomic(tmp, size, alignment);
 	return tmp;
 }
 
@@ -131,7 +132,7 @@ drm_sman_set_range(struct drm_sman * sman, unsigned int manager,
 	KASSERT(manager < sman->num_managers, ("Invalid manager"));
 
 	sman_mm = &sman->mm[manager];
-	mm = malloc(sizeof(*mm), DRM_MEM_MM, M_WAITOK | M_ZERO);
+	mm = malloc(sizeof(*mm), DRM_MEM_MM, M_NOWAIT | M_ZERO);
 	if (!mm) {
 		return -ENOMEM;
 	}
@@ -174,7 +175,7 @@ static struct drm_owner_item *drm_sman_get_owner_item(struct drm_sman * sman,
 				      owner_hash);
 	}
 
-	owner_item = malloc(sizeof(*owner_item), DRM_MEM_MM, M_WAITOK | M_ZERO);
+	owner_item = malloc(sizeof(*owner_item), DRM_MEM_MM, M_NOWAIT | M_ZERO);
 	if (!owner_item)
 		goto out;
 
@@ -206,12 +207,11 @@ struct drm_memblock_item *drm_sman_alloc(struct drm_sman *sman, unsigned int man
 
 	sman_mm = &sman->mm[manager];
 	tmp = sman_mm->allocate(sman_mm->private, size, alignment);
-
 	if (!tmp) {
 		return NULL;
 	}
 
-	memblock = malloc(sizeof(*memblock), DRM_MEM_MM, M_WAITOK | M_ZERO);
+	memblock = malloc(sizeof(*memblock), DRM_MEM_MM, M_NOWAIT | M_ZERO);
 	DRM_DEBUG("allocated mem_block %p\n", memblock);
 	if (!memblock)
 		goto out;

From e9615a1a0ee558bd2b76ed087a3f88ea8c72b9ac Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 23 Apr 2010 17:26:49 +0000
Subject: [PATCH 219/532] sh: Add some more tests for ${v#...} and ${v%...}.

These pass on stable/8 as well.
---
 tools/regression/bin/sh/expansion/trim3.0 | 46 +++++++++++++++++++++++
 1 file changed, 46 insertions(+)
 create mode 100644 tools/regression/bin/sh/expansion/trim3.0

diff --git a/tools/regression/bin/sh/expansion/trim3.0 b/tools/regression/bin/sh/expansion/trim3.0
new file mode 100644
index 00000000000..b89a04140c3
--- /dev/null
+++ b/tools/regression/bin/sh/expansion/trim3.0
@@ -0,0 +1,46 @@
+# $FreeBSD$
+
+e= q='?' a='*' t=texttext s='ast*que?non' p='/et[c]/' w='a b c' b='{{(#)}}'
+h='##' c='\\\\'
+failures=''
+ok=''
+
+testcase() {
+	code="$1"
+	expected="$2"
+	oIFS="$IFS"
+	eval "$code"
+	IFS='|'
+	result="$#|$*"
+	IFS="$oIFS"
+	if [ "x$result" = "x$expected" ]; then
+		ok=x$ok
+	else
+		failures=x$failures
+		echo "For $code, expected $expected actual $result"
+	fi
+}
+
+# This doesn't make much sense, but it fails in dash so I'm adding it here:
+testcase 'set -- "${w%${w#???}}"'		'1|a b'
+
+testcase 'set -- ${p#/et[}'			'1|c]/'
+testcase 'set -- "${p#/et[}"'			'1|c]/'
+testcase 'set -- "${p%${p#????}}"'		'1|/et['
+
+testcase 'set -- ${b%'\'}\''}'			'1|{{(#)}'
+
+testcase 'set -- ${c#\\}'			'1|\\\'
+testcase 'set -- ${c#\\\\}'			'1|\\'
+testcase 'set -- ${c#\\\\\\}'			'1|\'
+testcase 'set -- ${c#\\\\\\\\}'			'0|'
+testcase 'set -- "${c#\\}"'			'1|\\\'
+testcase 'set -- "${c#\\\\}"'			'1|\\'
+testcase 'set -- "${c#\\\\\\}"'			'1|\'
+testcase 'set -- "${c#\\\\\\\\}"'		'1|'
+testcase 'set -- "${c#"$c"}"'			'1|'
+testcase 'set -- ${c#"$c"}'			'0|'
+testcase 'set -- "${c%"$c"}"'			'1|'
+testcase 'set -- ${c%"$c"}'			'0|'
+
+test "x$failures" = x

From 562d4fbb7c9b1c5a8c88f5975908b4e27ebd8569 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 23 Apr 2010 18:53:17 +0000
Subject: [PATCH 220/532] Large memory mappings are always CPU local and always
 done with interrupts disabled.  Be doubly-sure that we don't try to do a TLB
 shootdown on SMP systems for those mappings.

Submitted by:	C. Jayachandran
---
 sys/mips/mips/pmap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index e3b7daf2ba7..32c0f929797 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -251,11 +251,11 @@ caddr_t virtual_sys_start = (caddr_t)0;
 #define	PMAP_LMEM_UNMAP()						\
 	pte = pmap_pte(kernel_pmap, sysm->base);			\
 	*pte = PTE_G;							\
-	pmap_invalidate_page(kernel_pmap, sysm->base);			\
+	pmap_TLB_invalidate_kernel(sysm->base);				\
 	sysm->valid1 = 0;						\
 	pte = pmap_pte(kernel_pmap, sysm->base + PAGE_SIZE);		\
 	*pte = PTE_G;							\
-	pmap_invalidate_page(kernel_pmap, sysm->base + PAGE_SIZE);	\
+	pmap_TLB_invalidate_kernel(sysm->base + PAGE_SIZE);		\
 	sysm->valid2 = 0;						\
 	sched_unpin();							\
 	intr_restore(intr);						\

From 38f836dbb3c7404a4f3effae33076b149fd157d1 Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 23 Apr 2010 19:20:56 +0000
Subject: [PATCH 221/532] o) Remove default MAXMEM on SWARM; pmap can readily
 use lmem for >512M    physical addresses. o) Set a local maxmem in
 sb_machdep.c to avoid trying to use pages over 2^64    under 32-bit ABIs. 
 Our pmap needs corrected to use vm_paddr_t consistently,    then we can make
 vm_paddr_t 64-bit under 32-bit ABIs and add code in pmap    to limit
 phys_avail by the maximum PFN that a 32-bit PTE can hold.

---
 sys/mips/conf/SWARM          |  6 ------
 sys/mips/sibyte/sb_machdep.c | 11 +++++++++++
 2 files changed, 11 insertions(+), 6 deletions(-)

diff --git a/sys/mips/conf/SWARM b/sys/mips/conf/SWARM
index 3189bcbcb9c..041c94da983 100644
--- a/sys/mips/conf/SWARM
+++ b/sys/mips/conf/SWARM
@@ -8,12 +8,6 @@ options		CPU_SB1
 files		"../sibyte/files.sibyte"
 hints		"SWARM.hints"
 
-#
-# 32-bit kernel cannot deal with physical memory beyond 4GB
-# XXX pmap assumes that all the memory can be mapped using KSEG0
-#
-options		MAXMEM=512*1024
-
 options		CFE
 options		CFE_CONSOLE
 options		CFE_ENV
diff --git a/sys/mips/sibyte/sb_machdep.c b/sys/mips/sibyte/sb_machdep.c
index dca2869a5f2..559bf741502 100644
--- a/sys/mips/sibyte/sb_machdep.c
+++ b/sys/mips/sibyte/sb_machdep.c
@@ -158,6 +158,17 @@ mips_init(void)
 	TUNABLE_INT_FETCH("hw.physmem", &tmp);
 	maxmem = (uint64_t)tmp * 1024;
 
+	/*
+	 * XXX
+	 * If we used vm_paddr_t consistently in pmap, etc., we could
+	 * use 64-bit page numbers on !n64 systems, too, like i386
+	 * does with PAE.
+	 */
+#if !defined(__mips_n64)
+	if (maxmem == 0 || maxmem > 0xffffffff)
+		maxmem = 0xffffffff;
+#endif
+
 #ifdef CFE
 	/*
 	 * Query DRAM memory map from CFE.

From c805ff09d5367d8781df85c3641f529033ac30fe Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Fri, 23 Apr 2010 19:48:31 +0000
Subject: [PATCH 222/532] Build some nops into CLEAR_STATUS here to make sure
 that the following instructions can't be interrupted.

---
 sys/mips/mips/exception.S | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sys/mips/mips/exception.S b/sys/mips/mips/exception.S
index aa2a0a6b5e9..d6fbc9cbd44 100644
--- a/sys/mips/mips/exception.S
+++ b/sys/mips/mips/exception.S
@@ -230,13 +230,15 @@ SlowFault:
 	or      a0, a0, a2	        ; \
 	li      a2, ~(MIPS_SR_INT_IE | MIPS_SR_EXL | SR_KSU_USER)   ; \
 	and     a0, a0, a2              ; \
-        mtc0    a0, COP_0_STATUS_REG    
+        mtc0    a0, COP_0_STATUS_REG    ; \
+	ITLBNOPFIX
 #else
 #define CLEAR_STATUS \
 	mfc0    a0, COP_0_STATUS_REG    ;\
 	li      a2, ~(MIPS_SR_INT_IE | MIPS_SR_EXL | SR_KSU_USER)   ; \
 	and     a0, a0, a2              ; \
-	mtc0	a0, COP_0_STATUS_REG
+	mtc0	a0, COP_0_STATUS_REG	; \
+	ITLBNOPFIX
 #endif
 
 /*

From ed9dd82fa253046e161b98fca9db198a1f267313 Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Fri, 23 Apr 2010 22:42:49 +0000
Subject: [PATCH 223/532] Adjust a test case and make it more jump optimization
 neutral for JIT case.

---
 tools/regression/bpf/bpf_filter/tests/test0083.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/tools/regression/bpf/bpf_filter/tests/test0083.h b/tools/regression/bpf/bpf_filter/tests/test0083.h
index debe719e670..a55780f9c5e 100644
--- a/tools/regression/bpf/bpf_filter/tests/test0083.h
+++ b/tools/regression/bpf/bpf_filter/tests/test0083.h
@@ -6,7 +6,7 @@
 
 /* BPF program */
 struct bpf_insn pc[] = {
-	BPF_JUMP(BPF_JMP+BPF_JA, 0, 0, 0),
+	BPF_STMT(BPF_LD|BPF_IMM, 0),
 };
 
 /* Packet */

From 07b9cc2f468471a723504ac789201a84f6dcd38a Mon Sep 17 00:00:00 2001
From: Juli Mallett <jmallett@FreeBSD.org>
Date: Sat, 24 Apr 2010 03:11:35 +0000
Subject: [PATCH 224/532] Most MIPS systems have a comparatively-sparse
 physical memory layout.  Switch to using the sparse physseg layout in the VM
 system.

---
 sys/mips/include/vmparam.h | 4 ++--
 sys/mips/mips/pmap.c       | 7 ++++---
 2 files changed, 6 insertions(+), 5 deletions(-)

diff --git a/sys/mips/include/vmparam.h b/sys/mips/include/vmparam.h
index bed5e36398e..3b9b12e43d0 100644
--- a/sys/mips/include/vmparam.h
+++ b/sys/mips/include/vmparam.h
@@ -152,9 +152,9 @@
 #define	VM_PHYSSEG_MAX		32
 
 /*
- * The physical address space is densely populated.
+ * The physical address space is sparsely populated.
  */
-#define	VM_PHYSSEG_DENSE
+#define	VM_PHYSSEG_SPARSE
 
 /*
  * Create three free page pools: VM_FREEPOOL_DEFAULT is the default pool
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 32c0f929797..9c28f23804f 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -2381,8 +2381,7 @@ pmap_remove_pages(pmap_t pmap)
 		*pte = is_kernel_pmap(pmap) ? PTE_G : 0;
 
 		m = PHYS_TO_VM_PAGE(mips_tlbpfn_to_paddr(tpte));
-
-		KASSERT(m < &vm_page_array[vm_page_array_size],
+		KASSERT(m != NULL,
 		    ("pmap_remove_pages: bad tpte %x", tpte));
 
 		pv->pv_pmap->pm_stats.resident_count--;
@@ -2984,10 +2983,12 @@ page_is_managed(vm_offset_t pa)
 {
 	vm_offset_t pgnum = mips_btop(pa);
 
-	if (pgnum >= first_page && (pgnum < (first_page + vm_page_array_size))) {
+	if (pgnum >= first_page) {
 		vm_page_t m;
 
 		m = PHYS_TO_VM_PAGE(pa);
+		if (m == NULL)
+			return 0;
 		if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) == 0)
 			return 1;
 	}

From 113db2dddb719510679d71ff41679afe30c8b3d1 Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Sat, 24 Apr 2010 07:05:35 +0000
Subject: [PATCH 225/532]  - Merge soft-updates journaling from
 projects/suj/head into head.  This    brings in support for an optional
 intent log which eliminates the need    for background fsck on unclean
 shutdown.

Sponsored by:   iXsystems, Yahoo!, and Juniper.
With help from: McKusick and Peter Holm
---
 lib/libufs/Makefile             |    7 +-
 lib/libufs/cgroup.c             |  140 +-
 lib/libufs/inode.c              |   16 +
 lib/libufs/libufs.h             |   16 +
 lib/libufs/sblock.c             |   50 +
 lib/libufs/type.c               |    5 +
 sbin/dumpfs/dumpfs.c            |    4 +-
 sbin/fsck_ffs/Makefile          |    3 +-
 sbin/fsck_ffs/fsck.h            |    5 +-
 sbin/fsck_ffs/gjournal.c        |  293 +-
 sbin/fsck_ffs/main.c            |   27 +-
 sbin/fsck_ffs/pass5.c           |    1 +
 sbin/fsck_ffs/suj.c             | 4699 ++++++++++++++++++++
 sbin/fsdb/fsdb.c                |   11 +-
 sbin/fsdb/fsdbutil.c            |   26 +-
 sbin/tunefs/tunefs.8            |    9 +-
 sbin/tunefs/tunefs.c            |  567 ++-
 sys/kern/vfs_bio.c              |   28 +-
 sys/kern/vfs_subr.c             |    1 +
 sys/sys/buf.h                   |    3 +-
 sys/sys/mount.h                 |    4 +-
 sys/ufs/ffs/ffs_alloc.c         |  252 +-
 sys/ufs/ffs/ffs_balloc.c        |   13 +-
 sys/ufs/ffs/ffs_extern.h        |   24 +-
 sys/ufs/ffs/ffs_inode.c         |  132 +-
 sys/ufs/ffs/ffs_snapshot.c      |   66 +-
 sys/ufs/ffs/ffs_softdep.c       | 7306 ++++++++++++++++++++++++++-----
 sys/ufs/ffs/ffs_subr.c          |  130 +-
 sys/ufs/ffs/ffs_vfsops.c        |   68 +-
 sys/ufs/ffs/ffs_vnops.c         |    1 +
 sys/ufs/ffs/fs.h                |  135 +-
 sys/ufs/ffs/softdep.h           |  446 +-
 sys/ufs/ufs/dinode.h            |    9 +-
 sys/ufs/ufs/inode.h             |    3 +-
 sys/ufs/ufs/ufs_dirhash.c       |    2 -
 sys/ufs/ufs/ufs_extern.h        |   24 +-
 sys/ufs/ufs/ufs_lookup.c        |  184 +-
 sys/ufs/ufs/ufs_vnops.c         |  583 +--
 sys/ufs/ufs/ufsmount.h          |    9 +
 usr.sbin/makefs/ffs/ffs_bswap.c |    2 -
 40 files changed, 13165 insertions(+), 2139 deletions(-)
 create mode 100644 sbin/fsck_ffs/suj.c

diff --git a/lib/libufs/Makefile b/lib/libufs/Makefile
index c9232ef54ad..1dfc2429a1f 100644
--- a/lib/libufs/Makefile
+++ b/lib/libufs/Makefile
@@ -3,7 +3,7 @@
 LIB=	ufs
 SHLIBDIR?= /lib
 
-SRCS=	block.c cgroup.c inode.c sblock.c type.c
+SRCS=	block.c cgroup.c inode.c sblock.c type.c ffs_subr.c ffs_tables.c
 INCS=	libufs.h
 
 MAN=	bread.3 cgread.3 libufs.3 sbread.3 ufs_disk_close.3
@@ -16,8 +16,11 @@ MLINKS+= ufs_disk_close.3 ufs_disk_fillout.3
 MLINKS+= ufs_disk_close.3 ufs_disk_fillout_blank.3
 MLINKS+= ufs_disk_close.3 ufs_disk_write.3
 
-WARNS?=	3
+.PATH:  ${.CURDIR}/../../sys/ufs/ffs
 
+WARNS?=	2
+
+DEBUG_FLAGS = -g
 CFLAGS+= -D_LIBUFS
 .if defined(LIBUFS_DEBUG)
 CFLAGS+= -D_LIBUFS_DEBUGGING
diff --git a/lib/libufs/cgroup.c b/lib/libufs/cgroup.c
index 28e5ea87a43..b0cb4a7d63d 100644
--- a/lib/libufs/cgroup.c
+++ b/lib/libufs/cgroup.c
@@ -40,11 +40,143 @@ __FBSDID("$FreeBSD$");
 #include <errno.h>
 #include <fcntl.h>
 #include <stdio.h>
+#include <stdlib.h>
 #include <string.h>
 #include <unistd.h>
 
 #include <libufs.h>
 
+ufs2_daddr_t
+cgballoc(struct uufsd *disk)
+{
+	u_int8_t *blksfree;
+	struct cg *cgp;
+	struct fs *fs;
+	long bno;
+
+	fs = &disk->d_fs;
+	cgp = &disk->d_cg;
+	blksfree = cg_blksfree(cgp);
+	for (bno = 0; bno < fs->fs_fpg / fs->fs_frag; bno++)
+		if (ffs_isblock(fs, blksfree, bno))
+			goto gotit;
+	return (0);
+gotit:
+	fs->fs_cs(fs, cgp->cg_cgx).cs_nbfree--;
+	ffs_clrblock(fs, blksfree, (long)bno);
+	ffs_clusteracct(fs, cgp, bno, -1);
+	cgp->cg_cs.cs_nbfree--;
+	fs->fs_cstotal.cs_nbfree--;
+	fs->fs_fmod = 1;
+	return (cgbase(fs, cgp->cg_cgx) + blkstofrags(fs, bno));
+}
+
+int
+cgbfree(struct uufsd *disk, ufs2_daddr_t bno, long size)
+{
+	u_int8_t *blksfree;
+	struct fs *fs;
+	struct cg *cgp;
+	ufs1_daddr_t fragno, cgbno;
+	int i, cg, blk, frags, bbase;
+
+	fs = &disk->d_fs;
+	cg = dtog(fs, bno);
+	if (cgread1(disk, cg) != 1)
+		return (-1);
+	cgp = &disk->d_cg;
+	cgbno = dtogd(fs, bno);
+	blksfree = cg_blksfree(cgp);
+	if (size == fs->fs_bsize) {
+		fragno = fragstoblks(fs, cgbno);
+		ffs_setblock(fs, blksfree, fragno);
+		ffs_clusteracct(fs, cgp, fragno, 1);
+		cgp->cg_cs.cs_nbfree++;
+		fs->fs_cstotal.cs_nbfree++;
+		fs->fs_cs(fs, cg).cs_nbfree++;
+	} else {
+		bbase = cgbno - fragnum(fs, cgbno);
+		/*
+		 * decrement the counts associated with the old frags
+		 */
+		blk = blkmap(fs, blksfree, bbase);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, -1);
+		/*
+		 * deallocate the fragment
+		 */
+		frags = numfrags(fs, size);
+		for (i = 0; i < frags; i++)
+			setbit(blksfree, cgbno + i);
+		cgp->cg_cs.cs_nffree += i;
+		fs->fs_cstotal.cs_nffree += i;
+		fs->fs_cs(fs, cg).cs_nffree += i;
+		/*
+		 * add back in counts associated with the new frags
+		 */
+		blk = blkmap(fs, blksfree, bbase);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, 1);
+		/*
+		 * if a complete block has been reassembled, account for it
+		 */
+		fragno = fragstoblks(fs, bbase);
+		if (ffs_isblock(fs, blksfree, fragno)) {
+			cgp->cg_cs.cs_nffree -= fs->fs_frag;
+			fs->fs_cstotal.cs_nffree -= fs->fs_frag;
+			fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag;
+			ffs_clusteracct(fs, cgp, fragno, 1);
+			cgp->cg_cs.cs_nbfree++;
+			fs->fs_cstotal.cs_nbfree++;
+			fs->fs_cs(fs, cg).cs_nbfree++;
+		}
+	}
+	return cgwrite(disk);
+}
+
+ino_t
+cgialloc(struct uufsd *disk)
+{
+	struct ufs2_dinode *dp2;
+	u_int8_t *inosused;
+	struct cg *cgp;
+	struct fs *fs;
+	ino_t ino;
+	int i;
+
+	fs = &disk->d_fs;
+	cgp = &disk->d_cg;
+	inosused = cg_inosused(cgp);
+	for (ino = 0; ino < fs->fs_ipg / NBBY; ino++)
+		if (isclr(inosused, ino))
+			goto gotit;
+	return (0);
+gotit:
+	if (fs->fs_magic == FS_UFS2_MAGIC &&
+	    ino + INOPB(fs) > cgp->cg_initediblk &&
+	    cgp->cg_initediblk < cgp->cg_niblk) {
+		char block[MAXBSIZE];
+		bzero(block, (int)fs->fs_bsize);
+		dp2 = (struct ufs2_dinode *)&block;
+		for (i = 0; i < INOPB(fs); i++) {
+			dp2->di_gen = arc4random() / 2 + 1;
+			dp2++;
+		}
+		if (bwrite(disk, ino_to_fsba(fs,
+		    cgp->cg_cgx * fs->fs_ipg + cgp->cg_initediblk),
+		    block, fs->fs_bsize))
+			return (0);
+		cgp->cg_initediblk += INOPB(fs);
+	}
+
+	setbit(inosused, ino);
+	cgp->cg_irotor = ino;
+	cgp->cg_cs.cs_nifree--;
+	fs->fs_cstotal.cs_nifree--;
+	fs->fs_cs(fs, cgp->cg_cgx).cs_nifree--;
+	fs->fs_fmod = 1;
+
+	return (ino + (cgp->cg_cgx * fs->fs_ipg));
+}
+
 int
 cgread(struct uufsd *disk)
 {
@@ -55,14 +187,12 @@ int
 cgread1(struct uufsd *disk, int c)
 {
 	struct fs *fs;
-	off_t ccg;
 
 	fs = &disk->d_fs;
 
 	if ((unsigned)c >= fs->fs_ncg) {
 		return (0);
 	}
-	ccg = fsbtodb(fs, cgtod(fs, c)) * disk->d_bsize;
 	if (bread(disk, fsbtodb(fs, cgtod(fs, c)), disk->d_cgunion.d_buf,
 	    fs->fs_bsize) == -1) {
 		ERROR(disk, "unable to read cylinder group");
@@ -72,6 +202,12 @@ cgread1(struct uufsd *disk, int c)
 	return (1);
 }
 
+int
+cgwrite(struct uufsd *disk)
+{
+	return (cgwrite1(disk, disk->d_lcg));
+}
+
 int
 cgwrite1(struct uufsd *disk, int c)
 {
diff --git a/lib/libufs/inode.c b/lib/libufs/inode.c
index d8bef61f891..6d94582b3e5 100644
--- a/lib/libufs/inode.c
+++ b/lib/libufs/inode.c
@@ -93,3 +93,19 @@ gotit:	switch (disk->d_ufs) {
 	ERROR(disk, "unknown UFS filesystem type");
 	return (-1);
 }
+
+int
+putino(struct uufsd *disk)
+{
+	struct fs *fs;
+
+	fs = &disk->d_fs;
+	if (disk->d_inoblock == NULL) {
+		ERROR(disk, "No inode block allocated");
+		return (-1);
+	}
+	if (bwrite(disk, fsbtodb(fs, ino_to_fsba(&disk->d_fs, disk->d_inomin)),
+	    disk->d_inoblock, disk->d_fs.fs_bsize) <= 0)
+		return (-1);
+	return (0);
+}
diff --git a/lib/libufs/libufs.h b/lib/libufs/libufs.h
index 42a64f76e2b..c3541a0c8e6 100644
--- a/lib/libufs/libufs.h
+++ b/lib/libufs/libufs.h
@@ -71,6 +71,7 @@ struct uufsd {
 	int d_fd;		/* raw device file descriptor */
 	long d_bsize;		/* device bsize */
 	ufs2_daddr_t d_sblock;	/* superblock location */
+	struct csum *d_sbcsum;	/* Superblock summary info */
 	caddr_t d_inoblock;	/* inode block */
 	ino_t d_inomin;		/* low inode */
 	ino_t d_inomax;		/* high inode */
@@ -109,14 +110,19 @@ int berase(struct uufsd *, ufs2_daddr_t, ufs2_daddr_t);
 /*
  * cgroup.c
  */
+ufs2_daddr_t cgballoc(struct uufsd *);
+int cgbfree(struct uufsd *, ufs2_daddr_t, long);
+ino_t cgialloc(struct uufsd *);
 int cgread(struct uufsd *);
 int cgread1(struct uufsd *, int);
+int cgwrite(struct uufsd *);
 int cgwrite1(struct uufsd *, int);
 
 /*
  * inode.c
  */
 int getino(struct uufsd *, void **, ino_t, int *);
+int putino(struct uufsd *);
 
 /*
  * sblock.c
@@ -132,6 +138,16 @@ int ufs_disk_fillout(struct uufsd *, const char *);
 int ufs_disk_fillout_blank(struct uufsd *, const char *);
 int ufs_disk_write(struct uufsd *);
 
+/*
+ * ffs_subr.c
+ */
+void	ffs_clrblock(struct fs *, u_char *, ufs1_daddr_t);
+void	ffs_clusteracct(struct fs *, struct cg *, ufs1_daddr_t, int);
+void	ffs_fragacct(struct fs *, int, int32_t [], int);
+int	ffs_isblock(struct fs *, u_char *, ufs1_daddr_t);
+int	ffs_isfreeblock(struct fs *, u_char *, ufs1_daddr_t);
+void	ffs_setblock(struct fs *, u_char *, ufs1_daddr_t);
+
 __END_DECLS
 
 #endif	/* __LIBUFS_H__ */
diff --git a/lib/libufs/sblock.c b/lib/libufs/sblock.c
index 8986290039f..d6bec3ed182 100644
--- a/lib/libufs/sblock.c
+++ b/lib/libufs/sblock.c
@@ -40,6 +40,7 @@ __FBSDID("$FreeBSD$");
 #include <errno.h>
 #include <stdio.h>
 #include <string.h>
+#include <stdlib.h>
 #include <unistd.h>
 
 #include <libufs.h>
@@ -49,8 +50,11 @@ static int superblocks[] = SBLOCKSEARCH;
 int
 sbread(struct uufsd *disk)
 {
+	uint8_t block[MAXBSIZE];
 	struct fs *fs;
 	int sb, superblock;
+	int i, size, blks;
+	uint8_t *space;
 
 	ERROR(disk, NULL);
 
@@ -86,6 +90,34 @@ sbread(struct uufsd *disk)
 	}
 	disk->d_bsize = fs->fs_fsize / fsbtodb(fs, 1);
 	disk->d_sblock = superblock / disk->d_bsize;
+	/*
+	 * Read in the superblock summary information.
+	 */
+	size = fs->fs_cssize;
+	blks = howmany(size, fs->fs_fsize);
+	size += fs->fs_ncg * sizeof(int32_t);
+	space = malloc(size);
+	if (space == NULL) {
+		ERROR(disk, "failed to allocate space for summary information");
+		return (-1);
+	}
+	fs->fs_csp = (struct csum *)space;
+	for (i = 0; i < blks; i += fs->fs_frag) {
+		size = fs->fs_bsize;
+		if (i + fs->fs_frag > blks)
+			size = (blks - i) * fs->fs_fsize;
+		if (bread(disk, fsbtodb(fs, fs->fs_csaddr + i), block, size)
+		    == -1) {
+			ERROR(disk, "Failed to read sb summary information");
+			free(fs->fs_csp);
+			return (-1);
+		}
+		bcopy(block, space, size);
+		space += size;
+	}
+	fs->fs_maxcluster = (uint32_t *)space;
+	disk->d_sbcsum = fs->fs_csp;
+
 	return (0);
 }
 
@@ -93,6 +125,8 @@ int
 sbwrite(struct uufsd *disk, int all)
 {
 	struct fs *fs;
+	int blks, size;
+	uint8_t *space;
 	unsigned i;
 
 	ERROR(disk, NULL);
@@ -107,6 +141,22 @@ sbwrite(struct uufsd *disk, int all)
 		ERROR(disk, "failed to write superblock");
 		return (-1);
 	}
+	/*
+	 * Write superblock summary information.
+	 */
+	blks = howmany(fs->fs_cssize, fs->fs_fsize);
+	space = (uint8_t *)disk->d_sbcsum;
+	for (i = 0; i < blks; i += fs->fs_frag) {
+		size = fs->fs_bsize;
+		if (i + fs->fs_frag > blks)
+			size = (blks - i) * fs->fs_fsize;
+		if (bwrite(disk, fsbtodb(fs, fs->fs_csaddr + i), space, size)
+		    == -1) {
+			ERROR(disk, "Failed to write sb summary information");
+			return (-1);
+		}
+		space += size;
+	}
 	if (all) {
 		for (i = 0; i < fs->fs_ncg; i++)
 			if (bwrite(disk, fsbtodb(fs, cgsblock(fs, i)),
diff --git a/lib/libufs/type.c b/lib/libufs/type.c
index 8a553bedbc9..05904b9e40b 100644
--- a/lib/libufs/type.c
+++ b/lib/libufs/type.c
@@ -66,6 +66,10 @@ ufs_disk_close(struct uufsd *disk)
 		free((char *)(uintptr_t)disk->d_name);
 		disk->d_name = NULL;
 	}
+	if (disk->d_sbcsum != NULL) {
+		free(disk->d_sbcsum);
+		disk->d_sbcsum = NULL;
+	}
 	return (0);
 }
 
@@ -156,6 +160,7 @@ again:	if ((ret = stat(name, &st)) < 0) {
 	disk->d_mine = 0;
 	disk->d_ufs = 0;
 	disk->d_error = NULL;
+	disk->d_sbcsum = NULL;
 
 	if (oname != name) {
 		name = strdup(name);
diff --git a/sbin/dumpfs/dumpfs.c b/sbin/dumpfs/dumpfs.c
index e4b59950401..38c05f605ad 100644
--- a/sbin/dumpfs/dumpfs.c
+++ b/sbin/dumpfs/dumpfs.c
@@ -238,7 +238,7 @@ dumpfs(const char *name)
 	if (fsflags & FS_UNCLEAN)
 		printf("unclean ");
 	if (fsflags & FS_DOSOFTDEP)
-		printf("soft-updates ");
+		printf("soft-updates%s ", (fsflags & FS_SUJ) ? "+journal" : "");
 	if (fsflags & FS_NEEDSFSCK)
 		printf("needs fsck run ");
 	if (fsflags & FS_INDEXDIRS)
@@ -255,7 +255,7 @@ dumpfs(const char *name)
 		printf("nfsv4acls ");
 	fsflags &= ~(FS_UNCLEAN | FS_DOSOFTDEP | FS_NEEDSFSCK | FS_INDEXDIRS |
 		     FS_ACLS | FS_MULTILABEL | FS_GJOURNAL | FS_FLAGS_UPDATED |
-		     FS_NFS4ACLS);
+		     FS_NFS4ACLS | FS_SUJ);
 	if (fsflags != 0)
 		printf("unknown flags (%#x)", fsflags);
 	putchar('\n');
diff --git a/sbin/fsck_ffs/Makefile b/sbin/fsck_ffs/Makefile
index aaae685df14..db2930bea62 100644
--- a/sbin/fsck_ffs/Makefile
+++ b/sbin/fsck_ffs/Makefile
@@ -7,8 +7,7 @@ LINKS+=	${BINDIR}/fsck_ffs ${BINDIR}/fsck_4.2bsd
 MAN=	fsck_ffs.8
 MLINKS=	fsck_ffs.8 fsck_ufs.8 fsck_ffs.8 fsck_4.2bsd.8
 SRCS=	dir.c ea.c fsutil.c inode.c main.c pass1.c pass1b.c pass2.c pass3.c \
-	pass4.c pass5.c setup.c utilities.c ffs_subr.c ffs_tables.c gjournal.c \
-	getmntopts.c
+	pass4.c pass5.c setup.c suj.c utilities.c gjournal.c getmntopts.c
 DPADD=	${LIBUFS}
 LDADD=	-lufs
 WARNS?=	2
diff --git a/sbin/fsck_ffs/fsck.h b/sbin/fsck_ffs/fsck.h
index ad7fe138ed6..6dad6b7318b 100644
--- a/sbin/fsck_ffs/fsck.h
+++ b/sbin/fsck_ffs/fsck.h
@@ -347,10 +347,6 @@ void		direrror(ino_t ino, const char *errmesg);
 int		dirscan(struct inodesc *);
 int		dofix(struct inodesc *, const char *msg);
 int		eascan(struct inodesc *, struct ufs2_dinode *dp);
-void		ffs_clrblock(struct fs *, u_char *, ufs1_daddr_t);
-void		ffs_fragacct(struct fs *, int, int32_t [], int);
-int		ffs_isblock(struct fs *, u_char *, ufs1_daddr_t);
-void		ffs_setblock(struct fs *, u_char *, ufs1_daddr_t);
 void		fileerror(ino_t cwd, ino_t ino, const char *errmesg);
 int		findino(struct inodesc *);
 int		findname(struct inodesc *);
@@ -392,3 +388,4 @@ void		sblock_init(void);
 void		setinodebuf(ino_t);
 int		setup(char *dev);
 void		gjournal_check(const char *filesys);
+int		suj_check(const char *filesys);
diff --git a/sbin/fsck_ffs/gjournal.c b/sbin/fsck_ffs/gjournal.c
index bd887cab850..10c32c00c07 100644
--- a/sbin/fsck_ffs/gjournal.c
+++ b/sbin/fsck_ffs/gjournal.c
@@ -95,27 +95,6 @@ struct ufs2_dinode ufs2_zino;
 
 static void putcgs(void);
 
-/*
- * Write current block of inodes.
- */
-static int
-putino(struct uufsd *disk, ino_t inode)
-{
-	caddr_t inoblock;
-	struct fs *fs;
-	ssize_t ret;
-
-	fs = &disk->d_fs;
-	inoblock = disk->d_inoblock;
-
-	assert(inoblock != NULL);
-	assert(inode >= disk->d_inomin && inode <= disk->d_inomax);
-	ret = bwrite(disk, fsbtodb(fs, ino_to_fsba(fs, inode)), inoblock,
-	    fs->fs_bsize);
-
-	return (ret == -1 ? -1 : 0);
-}
-
 /*
  * Return cylinder group from the cache or load it if it is not in the
  * cache yet.
@@ -242,13 +221,11 @@ cancelcgs(void)
 #endif
 
 /*
- * Open the given provider, load statistics.
+ * Open the given provider, load superblock.
  */
 static void
-getdisk(void)
+opendisk(void)
 {
-	int i;
-
 	if (disk != NULL)
 		return;
 	disk = malloc(sizeof(*disk));
@@ -259,24 +236,6 @@ getdisk(void)
 		    disk->d_error);
 	}
 	fs = &disk->d_fs;
-	fs->fs_csp = malloc((size_t)fs->fs_cssize);
-	if (fs->fs_csp == NULL)
-		err(1, "malloc(%zu)", (size_t)fs->fs_cssize);
-	bzero(fs->fs_csp, (size_t)fs->fs_cssize);
-	for (i = 0; i < fs->fs_cssize; i += fs->fs_bsize) {
-		if (bread(disk, fsbtodb(fs, fs->fs_csaddr + numfrags(fs, i)),
-		    (void *)(((char *)fs->fs_csp) + i),
-		    (size_t)(fs->fs_cssize - i < fs->fs_bsize ? fs->fs_cssize - i : fs->fs_bsize)) == -1) {
-			err(1, "bread: %s", disk->d_error);
-		}
-	}
-	if (fs->fs_contigsumsize > 0) {
-		fs->fs_maxcluster = malloc(fs->fs_ncg * sizeof(int32_t));
-		if (fs->fs_maxcluster == NULL)
-			err(1, "malloc(%zu)", fs->fs_ncg * sizeof(int32_t));
-		for (i = 0; i < fs->fs_ncg; i++)
-			fs->fs_maxcluster[i] = fs->fs_contigsumsize;
-	}
 }
 
 /*
@@ -286,11 +245,6 @@ static void
 closedisk(void)
 {
 
-	free(fs->fs_csp);
-	if (fs->fs_contigsumsize > 0) {
-		free(fs->fs_maxcluster);
-		fs->fs_maxcluster = NULL;
-	}
 	fs->fs_clean = 1;
 	if (sbwrite(disk, 0) == -1)
 		err(1, "sbwrite(%s)", devnam);
@@ -301,227 +255,6 @@ closedisk(void)
 	fs = NULL;
 }
 
-/*
- * Write the statistics back, call closedisk().
- */
-static void
-putdisk(void)
-{
-	int i;
-
-	assert(disk != NULL && fs != NULL);
-	for (i = 0; i < fs->fs_cssize; i += fs->fs_bsize) {
-		if (bwrite(disk, fsbtodb(fs, fs->fs_csaddr + numfrags(fs, i)),
-		    (void *)(((char *)fs->fs_csp) + i),
-		    (size_t)(fs->fs_cssize - i < fs->fs_bsize ? fs->fs_cssize - i : fs->fs_bsize)) == -1) {
-			err(1, "bwrite: %s", disk->d_error);
-		}
-	}
-	closedisk();
-}
-
-#if 0
-/*
- * Free memory, close the disk, but don't write anything back.
- */
-static void
-canceldisk(void)
-{
-	int i;
-
-	assert(disk != NULL && fs != NULL);
-	free(fs->fs_csp);
-	if (fs->fs_contigsumsize > 0)
-		free(fs->fs_maxcluster);
-	if (ufs_disk_close(disk) == -1)
-		err(1, "ufs_disk_close(%s)", devnam);
-	free(disk);
-	disk = NULL;
-	fs = NULL;
-}
-#endif
-
-static int
-isblock(unsigned char *cp, ufs1_daddr_t h)
-{
-	unsigned char mask;
-
-	switch ((int)fs->fs_frag) {
-	case 8:
-		return (cp[h] == 0xff);
-	case 4:
-		mask = 0x0f << ((h & 0x1) << 2);
-		return ((cp[h >> 1] & mask) == mask);
-	case 2:
-		mask = 0x03 << ((h & 0x3) << 1);
-		return ((cp[h >> 2] & mask) == mask);
-	case 1:
-		mask = 0x01 << (h & 0x7);
-		return ((cp[h >> 3] & mask) == mask);
-	default:
-		assert(!"isblock: invalid number of fragments");
-	}
-	return (0);
-}
-
-/*
- * put a block into the map
- */
-static void
-setblock(unsigned char *cp, ufs1_daddr_t h)
-{
-
-	switch ((int)fs->fs_frag) {
-	case 8:
-		cp[h] = 0xff;
-		return;
-	case 4:
-		cp[h >> 1] |= (0x0f << ((h & 0x1) << 2));
-		return;
-	case 2:
-		cp[h >> 2] |= (0x03 << ((h & 0x3) << 1));
-		return;
-	case 1:
-		cp[h >> 3] |= (0x01 << (h & 0x7));
-		return;
-	default:
-		assert(!"setblock: invalid number of fragments");
-	}
-}
-
-/*
- * check if a block is free
- */
-static int
-isfreeblock(u_char *cp, ufs1_daddr_t h)
-{
-
-	switch ((int)fs->fs_frag) {
-	case 8:
-		return (cp[h] == 0);
-	case 4:
-		return ((cp[h >> 1] & (0x0f << ((h & 0x1) << 2))) == 0);
-	case 2:
-		return ((cp[h >> 2] & (0x03 << ((h & 0x3) << 1))) == 0);
-	case 1:
-		return ((cp[h >> 3] & (0x01 << (h & 0x7))) == 0);
-	default:
-		assert(!"isfreeblock: invalid number of fragments");
-	}
-	return (0);
-}
-
-/*
- * Update the frsum fields to reflect addition or deletion
- * of some frags.
- */
-void
-fragacct(int fragmap, int32_t fraglist[], int cnt)
-{
-	int inblk;
-	int field, subfield;
-	int siz, pos;
-
-	inblk = (int)(fragtbl[fs->fs_frag][fragmap]) << 1;
-	fragmap <<= 1;
-	for (siz = 1; siz < fs->fs_frag; siz++) {
-		if ((inblk & (1 << (siz + (fs->fs_frag % NBBY)))) == 0)
-			continue;
-		field = around[siz];
-		subfield = inside[siz];
-		for (pos = siz; pos <= fs->fs_frag; pos++) {
-			if ((fragmap & field) == subfield) {
-				fraglist[siz] += cnt;
-				pos += siz;
-				field <<= siz;
-				subfield <<= siz;
-			}
-			field <<= 1;
-			subfield <<= 1;
-		}
-	}
-}
-
-static void
-clusteracct(struct cg *cgp, ufs1_daddr_t blkno)
-{
-	int32_t *sump;
-	int32_t *lp;
-	u_char *freemapp, *mapp;
-	int i, start, end, forw, back, map, bit;
-
-	if (fs->fs_contigsumsize <= 0)
-		return;
-	freemapp = cg_clustersfree(cgp);
-	sump = cg_clustersum(cgp);
-	/*
-	 * Clear the actual block.
-	 */
-	setbit(freemapp, blkno);
-	/*
-	 * Find the size of the cluster going forward.
-	 */
-	start = blkno + 1;
-	end = start + fs->fs_contigsumsize;
-	if (end >= cgp->cg_nclusterblks)
-		end = cgp->cg_nclusterblks;
-	mapp = &freemapp[start / NBBY];
-	map = *mapp++;
-	bit = 1 << (start % NBBY);
-	for (i = start; i < end; i++) {
-		if ((map & bit) == 0)
-			break;
-		if ((i & (NBBY - 1)) != (NBBY - 1)) {
-			bit <<= 1;
-		} else {
-			map = *mapp++;
-			bit = 1;
-		}
-	}
-	forw = i - start;
-	/*
-	 * Find the size of the cluster going backward.
-	 */
-	start = blkno - 1;
-	end = start - fs->fs_contigsumsize;
-	if (end < 0)
-		end = -1;
-	mapp = &freemapp[start / NBBY];
-	map = *mapp--;
-	bit = 1 << (start % NBBY);
-	for (i = start; i > end; i--) {
-		if ((map & bit) == 0)
-			break;
-		if ((i & (NBBY - 1)) != 0) {
-			bit >>= 1;
-		} else {
-			map = *mapp--;
-			bit = 1 << (NBBY - 1);
-		}
-	}
-	back = start - i;
-	/*
-	 * Account for old cluster and the possibly new forward and
-	 * back clusters.
-	 */
-	i = back + forw + 1;
-	if (i > fs->fs_contigsumsize)
-		i = fs->fs_contigsumsize;
-	sump[i]++;
-	if (back > 0)
-		sump[back]--;
-	if (forw > 0)
-		sump[forw]--;
-	/*
-	 * Update cluster summary information.
-	 */
-	lp = &sump[fs->fs_contigsumsize];
-	for (i = fs->fs_contigsumsize; i > 0; i--)
-		if (*lp-- > 0)
-			break;
-	fs->fs_maxcluster[cgp->cg_cgx] = i;
-}
-
 static void
 blkfree(ufs2_daddr_t bno, long size)
 {
@@ -539,10 +272,10 @@ blkfree(ufs2_daddr_t bno, long size)
 	blksfree = cg_blksfree(cgp);
 	if (size == fs->fs_bsize) {
 		fragno = fragstoblks(fs, cgbno);
-		if (!isfreeblock(blksfree, fragno))
+		if (!ffs_isfreeblock(fs, blksfree, fragno))
 			assert(!"blkfree: freeing free block");
-		setblock(blksfree, fragno);
-		clusteracct(cgp, fragno);
+		ffs_setblock(fs, blksfree, fragno);
+		ffs_clusteracct(fs, cgp, fragno, 1);
 		cgp->cg_cs.cs_nbfree++;
 		fs->fs_cstotal.cs_nbfree++;
 		fs->fs_cs(fs, cg).cs_nbfree++;
@@ -552,7 +285,7 @@ blkfree(ufs2_daddr_t bno, long size)
 		 * decrement the counts associated with the old frags
 		 */
 		blk = blkmap(fs, blksfree, bbase);
-		fragacct(blk, cgp->cg_frsum, -1);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, -1);
 		/*
 		 * deallocate the fragment
 		 */
@@ -569,16 +302,16 @@ blkfree(ufs2_daddr_t bno, long size)
 		 * add back in counts associated with the new frags
 		 */
 		blk = blkmap(fs, blksfree, bbase);
-		fragacct(blk, cgp->cg_frsum, 1);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, 1);
 		/*
 		 * if a complete block has been reassembled, account for it
 		 */
 		fragno = fragstoblks(fs, bbase);
-		if (isblock(blksfree, fragno)) {
+		if (ffs_isblock(fs, blksfree, fragno)) {
 			cgp->cg_cs.cs_nffree -= fs->fs_frag;
 			fs->fs_cstotal.cs_nffree -= fs->fs_frag;
 			fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag;
-			clusteracct(cgp, fragno);
+			ffs_clusteracct(fs, cgp, fragno, 1);
 			cgp->cg_cs.cs_nbfree++;
 			fs->fs_cstotal.cs_nbfree++;
 			fs->fs_cs(fs, cg).cs_nbfree++;
@@ -599,7 +332,7 @@ freeindir(ufs2_daddr_t blk, int level)
 	if (bread(disk, fsbtodb(fs, blk), (void *)&sblks, (size_t)fs->fs_bsize) == -1)
 		err(1, "bread: %s", disk->d_error);
 	blks = (ufs2_daddr_t *)&sblks;
-	for (i = 0; i < howmany(fs->fs_bsize, sizeof(ufs2_daddr_t)); i++) {
+	for (i = 0; i < NINDIR(fs); i++) {
 		if (blks[i] == 0)
 			break;
 		if (level == 0)
@@ -671,7 +404,7 @@ gjournal_check(const char *filesys)
 	int cg, mode;
 
 	devnam = filesys;
-	getdisk();
+	opendisk();
 	/* Are there any unreferenced inodes in this file system? */
 	if (fs->fs_unrefs == 0) {
 		//printf("No unreferenced inodes.\n");
@@ -747,7 +480,7 @@ gjournal_check(const char *filesys)
 			/* Zero-fill the inode. */
 			*dino = ufs2_zino;
 			/* Write the inode back. */
-			if (putino(disk, ino) == -1)
+			if (putino(disk) == -1)
 				err(1, "putino(cg=%d ino=%d)", cg, ino);
 			if (cgp->cg_unrefs == 0) {
 				//printf("No more unreferenced inodes in cg=%d.\n", cg);
@@ -772,5 +505,5 @@ gjournal_check(const char *filesys)
 	/* Write back modified cylinder groups. */
 	putcgs();
 	/* Write back updated statistics and super-block. */
-	putdisk();
+	closedisk();
 }
diff --git a/sbin/fsck_ffs/main.c b/sbin/fsck_ffs/main.c
index 66edd63e21f..e9a97043160 100644
--- a/sbin/fsck_ffs/main.c
+++ b/sbin/fsck_ffs/main.c
@@ -242,8 +242,9 @@ checkfilesys(char *filesys)
 		if ((fsreadfd = open(filesys, O_RDONLY)) < 0 || readsb(0) == 0)
 			exit(3);	/* Cannot read superblock */
 		close(fsreadfd);
-		if (sblock.fs_flags & FS_NEEDSFSCK)
-			exit(4);	/* Earlier background failed */
+		/* Earlier background failed or journaled */
+		if (sblock.fs_flags & (FS_NEEDSFSCK | FS_SUJ))
+			exit(4);
 		if ((sblock.fs_flags & FS_DOSOFTDEP) == 0)
 			exit(5);	/* Not running soft updates */
 		size = MIBSIZE;
@@ -299,7 +300,7 @@ checkfilesys(char *filesys)
 			pfatal("MOUNTED READ-ONLY, CANNOT RUN IN BACKGROUND\n");
 		} else if ((fsreadfd = open(filesys, O_RDONLY)) >= 0) {
 			if (readsb(0) != 0) {
-				if (sblock.fs_flags & FS_NEEDSFSCK) {
+				if (sblock.fs_flags & (FS_NEEDSFSCK | FS_SUJ)) {
 					bkgrdflag = 0;
 					pfatal("UNEXPECTED INCONSISTENCY, %s\n",
 					    "CANNOT RUN IN BACKGROUND\n");
@@ -384,6 +385,26 @@ checkfilesys(char *filesys)
 		    sblock.fs_cstotal.cs_nffree * 100.0 / sblock.fs_dsize);
 		return (0);
 	}
+	/*
+	 * Determine if we can and should do journal recovery.
+	 */
+	if ((sblock.fs_flags & (FS_SUJ | FS_NEEDSFSCK)) == FS_SUJ) {
+		if (preen || reply("USE JOURNAL?")) {
+			if (suj_check(filesys) == 0) {
+				if (chkdoreload(mntp) == 0)
+					exit(0);
+				exit(4);
+			}
+			/* suj_check failed, fall through. */
+		}
+		printf("** Skipping journal, falling through to full fsck\n");
+		/*
+		 * Write the superblock so we don't try to recover the
+		 * journal on another pass.
+		 */
+		sblock.fs_mtime = time(NULL);
+		sbdirty();
+	}
 	
 	/*
 	 * Cleared if any questions answered no. Used to decide if
diff --git a/sbin/fsck_ffs/pass5.c b/sbin/fsck_ffs/pass5.c
index 173156efaf5..639ce0f8abd 100644
--- a/sbin/fsck_ffs/pass5.c
+++ b/sbin/fsck_ffs/pass5.c
@@ -45,6 +45,7 @@ __FBSDID("$FreeBSD$");
 #include <inttypes.h>
 #include <limits.h>
 #include <string.h>
+#include <libufs.h>
 
 #include "fsck.h"
 
diff --git a/sbin/fsck_ffs/suj.c b/sbin/fsck_ffs/suj.c
new file mode 100644
index 00000000000..84608f164e4
--- /dev/null
+++ b/sbin/fsck_ffs/suj.c
@@ -0,0 +1,4699 @@
+/*-
+ * Copyright 2009, 2010 Jeffrey W. Roberson <jeff@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/disklabel.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+
+#include <ufs/ufs/ufsmount.h>
+#include <ufs/ufs/dinode.h>
+#include <ufs/ufs/dir.h>
+#include <ufs/ffs/fs.h>
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <libufs.h>
+#include <string.h>
+#include <strings.h>
+#include <err.h>
+#include <assert.h>
+
+#include "fsck.h"
+
+#define	DOTDOT_OFFSET	DIRECTSIZ(1)
+#define	SUJ_HASHSIZE	2048
+#define	SUJ_HASHMASK	(SUJ_HASHSIZE - 1)
+#define	SUJ_HASH(x)	((x * 2654435761) & SUJ_HASHMASK)
+
+struct suj_seg {
+	TAILQ_ENTRY(suj_seg) ss_next;
+	struct jsegrec	ss_rec;
+	uint8_t		*ss_blk;
+};
+
+struct suj_rec {
+	TAILQ_ENTRY(suj_rec) sr_next;
+	union jrec	*sr_rec;
+};
+TAILQ_HEAD(srechd, suj_rec);
+
+struct suj_ino {
+	LIST_ENTRY(suj_ino)	si_next;
+	struct srechd		si_recs;
+	struct srechd		si_newrecs;
+	struct srechd		si_movs;
+	struct jtrncrec		*si_trunc;
+	ino_t			si_ino;
+	char			si_skipparent;
+	char			si_hasrecs;
+	char			si_blkadj;
+	char			si_linkadj;
+	int			si_mode;
+	nlink_t			si_nlinkadj;
+	nlink_t			si_nlink;
+	nlink_t			si_dotlinks;
+};
+LIST_HEAD(inohd, suj_ino);
+
+struct suj_blk {
+	LIST_ENTRY(suj_blk)	sb_next;
+	struct srechd		sb_recs;
+	ufs2_daddr_t		sb_blk;
+};
+LIST_HEAD(blkhd, suj_blk);
+
+struct data_blk {
+	LIST_ENTRY(data_blk)	db_next;
+	uint8_t			*db_buf;
+	ufs2_daddr_t		db_blk;
+	int			db_size;
+	int			db_dirty;
+};
+
+struct ino_blk {
+	LIST_ENTRY(ino_blk)	ib_next;
+	uint8_t			*ib_buf;
+	int			ib_dirty;
+	ufs2_daddr_t		ib_blk;
+};
+LIST_HEAD(iblkhd, ino_blk);
+
+struct suj_cg {
+	LIST_ENTRY(suj_cg)	sc_next;
+	struct blkhd		sc_blkhash[SUJ_HASHSIZE];
+	struct inohd		sc_inohash[SUJ_HASHSIZE];
+	struct iblkhd		sc_iblkhash[SUJ_HASHSIZE];
+	struct ino_blk		*sc_lastiblk;
+	struct suj_ino		*sc_lastino;
+	struct suj_blk		*sc_lastblk;
+	uint8_t			*sc_cgbuf;
+	struct cg		*sc_cgp;
+	int			sc_dirty;
+	int			sc_cgx;
+};
+
+LIST_HEAD(cghd, suj_cg) cghash[SUJ_HASHSIZE];
+LIST_HEAD(dblkhd, data_blk) dbhash[SUJ_HASHSIZE];
+struct suj_cg *lastcg;
+struct data_blk *lastblk;
+
+TAILQ_HEAD(seghd, suj_seg) allsegs;
+uint64_t oldseq;
+static struct uufsd *disk = NULL;
+static struct fs *fs = NULL;
+ino_t sujino;
+
+/*
+ * Summary statistics.
+ */
+uint64_t freefrags;
+uint64_t freeblocks;
+uint64_t freeinos;
+uint64_t freedir;
+uint64_t jbytes;
+uint64_t jrecs;
+
+typedef void (*ino_visitor)(ino_t, ufs_lbn_t, ufs2_daddr_t, int);
+static void ino_trunc(ino_t, off_t);
+static void ino_decr(ino_t);
+static void ino_adjust(struct suj_ino *);
+static void ino_build(struct suj_ino *);
+static int blk_isfree(ufs2_daddr_t);
+
+static void *
+errmalloc(size_t n)
+{
+	void *a;
+
+	a = malloc(n);
+	if (a == NULL)
+		errx(1, "malloc(%zu)", n);
+	return (a);
+}
+
+/*
+ * Open the given provider, load superblock.
+ */
+static void
+opendisk(const char *devnam)
+{
+	if (disk != NULL)
+		return;
+	disk = malloc(sizeof(*disk));
+	if (disk == NULL)
+		errx(1, "malloc(%zu)", sizeof(*disk));
+	if (ufs_disk_fillout(disk, devnam) == -1) {
+		err(1, "ufs_disk_fillout(%s) failed: %s", devnam,
+		    disk->d_error);
+	}
+	fs = &disk->d_fs;
+}
+
+/*
+ * Mark file system as clean, write the super-block back, close the disk.
+ */
+static void
+closedisk(const char *devnam)
+{
+	struct csum *cgsum;
+	int i;
+
+	/*
+	 * Recompute the fs summary info from correct cs summaries.
+	 */
+	bzero(&fs->fs_cstotal, sizeof(struct csum_total));
+	for (i = 0; i < fs->fs_ncg; i++) {
+		cgsum = &fs->fs_cs(fs, i);
+		fs->fs_cstotal.cs_nffree += cgsum->cs_nffree;
+		fs->fs_cstotal.cs_nbfree += cgsum->cs_nbfree;
+		fs->fs_cstotal.cs_nifree += cgsum->cs_nifree;
+		fs->fs_cstotal.cs_ndir += cgsum->cs_ndir;
+	}
+	fs->fs_pendinginodes = 0;
+	fs->fs_pendingblocks = 0;
+	fs->fs_clean = 1;
+	fs->fs_time = time(NULL);
+	fs->fs_mtime = time(NULL);
+	if (sbwrite(disk, 0) == -1)
+		err(1, "sbwrite(%s)", devnam);
+	if (ufs_disk_close(disk) == -1)
+		err(1, "ufs_disk_close(%s)", devnam);
+	free(disk);
+	disk = NULL;
+	fs = NULL;
+}
+
+/*
+ * Lookup a cg by number in the hash so we can keep track of which cgs
+ * need stats rebuilt.
+ */
+static struct suj_cg *
+cg_lookup(int cgx)
+{
+	struct cghd *hd;
+	struct suj_cg *sc;
+
+	if (cgx < 0 || cgx >= fs->fs_ncg) {
+		abort();
+		errx(1, "Bad cg number %d", cgx);
+	}
+	if (lastcg && lastcg->sc_cgx == cgx)
+		return (lastcg);
+	hd = &cghash[SUJ_HASH(cgx)];
+	LIST_FOREACH(sc, hd, sc_next)
+		if (sc->sc_cgx == cgx) {
+			lastcg = sc;
+			return (sc);
+		}
+	sc = errmalloc(sizeof(*sc));
+	bzero(sc, sizeof(*sc));
+	sc->sc_cgbuf = errmalloc(fs->fs_bsize);
+	sc->sc_cgp = (struct cg *)sc->sc_cgbuf;
+	sc->sc_cgx = cgx;
+	LIST_INSERT_HEAD(hd, sc, sc_next);
+	if (bread(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
+	    fs->fs_bsize) == -1)
+		err(1, "Unable to read cylinder group %d", sc->sc_cgx);
+
+	return (sc);
+}
+
+/*
+ * Lookup an inode number in the hash and allocate a suj_ino if it does
+ * not exist.
+ */
+static struct suj_ino *
+ino_lookup(ino_t ino, int creat)
+{
+	struct suj_ino *sino;
+	struct inohd *hd;
+	struct suj_cg *sc;
+
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	if (sc->sc_lastino && sc->sc_lastino->si_ino == ino)
+		return (sc->sc_lastino);
+	hd = &sc->sc_inohash[SUJ_HASH(ino)];
+	LIST_FOREACH(sino, hd, si_next)
+		if (sino->si_ino == ino)
+			return (sino);
+	if (creat == 0)
+		return (NULL);
+	sino = errmalloc(sizeof(*sino));
+	bzero(sino, sizeof(*sino));
+	sino->si_ino = ino;
+	TAILQ_INIT(&sino->si_recs);
+	TAILQ_INIT(&sino->si_newrecs);
+	TAILQ_INIT(&sino->si_movs);
+	LIST_INSERT_HEAD(hd, sino, si_next);
+
+	return (sino);
+}
+
+/*
+ * Lookup a block number in the hash and allocate a suj_blk if it does
+ * not exist.
+ */
+static struct suj_blk *
+blk_lookup(ufs2_daddr_t blk, int creat)
+{
+	struct suj_blk *sblk;
+	struct suj_cg *sc;
+	struct blkhd *hd;
+
+	sc = cg_lookup(dtog(fs, blk));
+	if (sc->sc_lastblk && sc->sc_lastblk->sb_blk == blk)
+		return (sc->sc_lastblk);
+	hd = &sc->sc_blkhash[SUJ_HASH(fragstoblks(fs, blk))];
+	LIST_FOREACH(sblk, hd, sb_next)
+		if (sblk->sb_blk == blk)
+			return (sblk);
+	if (creat == 0)
+		return (NULL);
+	sblk = errmalloc(sizeof(*sblk));
+	bzero(sblk, sizeof(*sblk));
+	sblk->sb_blk = blk;
+	TAILQ_INIT(&sblk->sb_recs);
+	LIST_INSERT_HEAD(hd, sblk, sb_next);
+
+	return (sblk);
+}
+
+static struct data_blk *
+dblk_lookup(ufs2_daddr_t blk)
+{
+	struct data_blk *dblk;
+	struct dblkhd *hd;
+
+	hd = &dbhash[SUJ_HASH(fragstoblks(fs, blk))];
+	if (lastblk && lastblk->db_blk == blk)
+		return (lastblk);
+	LIST_FOREACH(dblk, hd, db_next)
+		if (dblk->db_blk == blk)
+			return (dblk);
+	/*
+	 * The inode block wasn't located, allocate a new one.
+	 */
+	dblk = errmalloc(sizeof(*dblk));
+	bzero(dblk, sizeof(*dblk));
+	LIST_INSERT_HEAD(hd, dblk, db_next);
+	dblk->db_blk = blk;
+	return (dblk);
+}
+
+static uint8_t *
+dblk_read(ufs2_daddr_t blk, int size)
+{
+	struct data_blk *dblk;
+
+	dblk = dblk_lookup(blk);
+	/*
+	 * I doubt size mismatches can happen in practice but it is trivial
+	 * to handle.
+	 */
+	if (size != dblk->db_size) {
+		if (dblk->db_buf)
+			free(dblk->db_buf);
+		dblk->db_buf = errmalloc(size);
+		dblk->db_size = size;
+		if (bread(disk, fsbtodb(fs, blk), dblk->db_buf, size) == -1)
+			err(1, "Failed to read data block %jd", blk);
+	}
+	return (dblk->db_buf);
+}
+
+static void
+dblk_dirty(ufs2_daddr_t blk)
+{
+	struct data_blk *dblk;
+
+	dblk = dblk_lookup(blk);
+	dblk->db_dirty = 1;
+}
+
+static void
+dblk_write(void)
+{
+	struct data_blk *dblk;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++) {
+		LIST_FOREACH(dblk, &dbhash[i], db_next) {
+			if (dblk->db_dirty == 0 || dblk->db_size == 0)
+				continue;
+			if (bwrite(disk, fsbtodb(fs, dblk->db_blk),
+			    dblk->db_buf, dblk->db_size) == -1)
+				err(1, "Unable to write block %jd",
+				    dblk->db_blk);
+		}
+	}
+}
+
+static union dinode *
+ino_read(ino_t ino)
+{
+	struct ino_blk *iblk;
+	struct iblkhd *hd;
+	struct suj_cg *sc;
+	ufs2_daddr_t blk;
+	int off;
+
+	blk = ino_to_fsba(fs, ino);
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	iblk = sc->sc_lastiblk;
+	if (iblk && iblk->ib_blk == blk)
+		goto found;
+	hd = &sc->sc_iblkhash[SUJ_HASH(fragstoblks(fs, blk))];
+	LIST_FOREACH(iblk, hd, ib_next)
+		if (iblk->ib_blk == blk)
+			goto found;
+	/*
+	 * The inode block wasn't located, allocate a new one.
+	 */
+	iblk = errmalloc(sizeof(*iblk));
+	bzero(iblk, sizeof(*iblk));
+	iblk->ib_buf = errmalloc(fs->fs_bsize);
+	iblk->ib_blk = blk;
+	LIST_INSERT_HEAD(hd, iblk, ib_next);
+	if (bread(disk, fsbtodb(fs, blk), iblk->ib_buf, fs->fs_bsize) == -1)
+		err(1, "Failed to read inode block %jd", blk);
+found:
+	sc->sc_lastiblk = iblk;
+	off = ino_to_fsbo(fs, ino);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		return (union dinode *)&((struct ufs1_dinode *)iblk->ib_buf)[off];
+	else
+		return (union dinode *)&((struct ufs2_dinode *)iblk->ib_buf)[off];
+}
+
+static void
+ino_dirty(ino_t ino)
+{
+	struct ino_blk *iblk;
+	struct iblkhd *hd;
+	struct suj_cg *sc;
+	ufs2_daddr_t blk;
+
+	blk = ino_to_fsba(fs, ino);
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	iblk = sc->sc_lastiblk;
+	if (iblk && iblk->ib_blk == blk) {
+		iblk->ib_dirty = 1;
+		return;
+	}
+	hd = &sc->sc_iblkhash[SUJ_HASH(fragstoblks(fs, blk))];
+	LIST_FOREACH(iblk, hd, ib_next) {
+		if (iblk->ib_blk == blk) {
+			iblk->ib_dirty = 1;
+			return;
+		}
+	}
+	ino_read(ino);
+	ino_dirty(ino);
+}
+
+static void
+iblk_write(struct ino_blk *iblk)
+{
+
+	if (iblk->ib_dirty == 0)
+		return;
+	if (bwrite(disk, fsbtodb(fs, iblk->ib_blk), iblk->ib_buf,
+	    fs->fs_bsize) == -1)
+		err(1, "Failed to write inode block %jd", iblk->ib_blk);
+}
+
+static int
+blk_overlaps(struct jblkrec *brec, ufs2_daddr_t start, int frags)
+{
+	ufs2_daddr_t bstart;
+	ufs2_daddr_t bend;
+	ufs2_daddr_t end;
+
+	end = start + frags;
+	bstart = brec->jb_blkno + brec->jb_oldfrags;
+	bend = bstart + brec->jb_frags;
+	if (start < bend && end > bstart)
+		return (1);
+	return (0);
+}
+
+static int
+blk_equals(struct jblkrec *brec, ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t start,
+    int frags)
+{
+
+	if (brec->jb_ino != ino || brec->jb_lbn != lbn)
+		return (0);
+	if (brec->jb_blkno + brec->jb_oldfrags != start)
+		return (0);
+	if (brec->jb_frags != frags)
+		return (0);
+	return (1);
+}
+
+static void
+blk_setmask(struct jblkrec *brec, int *mask)
+{
+	int i;
+
+	for (i = brec->jb_oldfrags; i < brec->jb_oldfrags + brec->jb_frags; i++)
+		*mask |= 1 << i;
+}
+
+/*
+ * Determine whether a given block has been reallocated to a new location.
+ * Returns a mask of overlapping bits if any frags have been reused or
+ * zero if the block has not been re-used and the contents can be trusted.
+ * 
+ * This is used to ensure that an orphaned pointer due to truncate is safe
+ * to be freed.  The mask value can be used to free partial blocks.
+ */
+static int
+blk_freemask(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags)
+{
+	struct suj_blk *sblk;
+	struct suj_rec *srec;
+	struct jblkrec *brec;
+	int mask;
+	int off;
+
+	/*
+	 * To be certain we're not freeing a reallocated block we lookup
+	 * this block in the blk hash and see if there is an allocation
+	 * journal record that overlaps with any fragments in the block
+	 * we're concerned with.  If any fragments have ben reallocated
+	 * the block has already been freed and re-used for another purpose.
+	 */
+	mask = 0;
+	sblk = blk_lookup(blknum(fs, blk), 0);
+	if (sblk == NULL)
+		return (0);
+	off = blk - sblk->sb_blk;
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		brec = (struct jblkrec *)srec->sr_rec;
+		/*
+		 * If the block overlaps but does not match
+		 * exactly it's a new allocation.  If it matches
+		 * exactly this record refers to the current
+		 * location.
+		 */ 
+		if (blk_overlaps(brec, blk, frags) == 0)
+			continue;
+		if (blk_equals(brec, ino, lbn, blk, frags) == 1)
+			mask = 0;
+		else
+			blk_setmask(brec, &mask);
+	}
+	if (debug)
+		printf("blk_freemask: blk %jd sblk %jd off %d mask 0x%X\n",
+		    blk, sblk->sb_blk, off, mask);
+	return (mask >> off);
+}
+
+/*
+ * Determine whether it is safe to follow an indirect.  It is not safe
+ * if any part of the indirect has been reallocated or the last journal
+ * entry was an allocation.  Just allocated indirects may not have valid
+ * pointers yet and all of their children will have their own records.
+ * It is also not safe to follow an indirect if the cg bitmap has been
+ * cleared as a new allocation may write to the block prior to the journal
+ * being written.
+ * 
+ * Returns 1 if it's safe to follow the indirect and 0 otherwise.
+ */
+static int
+blk_isindir(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn)
+{
+	struct suj_blk *sblk;
+	struct jblkrec *brec;
+
+	sblk = blk_lookup(blk, 0);
+	if (sblk == NULL)
+		return (1);
+	if (TAILQ_EMPTY(&sblk->sb_recs))
+		return (1);
+	brec = (struct jblkrec *)TAILQ_LAST(&sblk->sb_recs, srechd)->sr_rec;
+	if (blk_equals(brec, ino, lbn, blk, fs->fs_frag))
+		if (brec->jb_op == JOP_FREEBLK)
+			return (!blk_isfree(blk));
+	return (0);
+}
+
+/*
+ * Clear an inode from the cg bitmap.  If the inode was already clear return
+ * 0 so the caller knows it does not have to check the inode contents.
+ */
+static int
+ino_free(ino_t ino, int mode)
+{
+	struct suj_cg *sc;
+	uint8_t *inosused;
+	struct cg *cgp;
+	int cg;
+
+	cg = ino_to_cg(fs, ino);
+	ino = ino % fs->fs_ipg;
+	sc = cg_lookup(cg);
+	cgp = sc->sc_cgp;
+	inosused = cg_inosused(cgp);
+	/*
+	 * The bitmap may never have made it to the disk so we have to
+	 * conditionally clear.  We can avoid writing the cg in this case.
+	 */
+	if (isclr(inosused, ino))
+		return (0);
+	freeinos++;
+	clrbit(inosused, ino);
+	if (ino < cgp->cg_irotor)
+		cgp->cg_irotor = ino;
+	cgp->cg_cs.cs_nifree++;
+	if ((mode & IFMT) == IFDIR) {
+		freedir++;
+		cgp->cg_cs.cs_ndir--;
+	}
+	sc->sc_dirty = 1;
+
+	return (1);
+}
+
+/*
+ * Free 'frags' frags starting at filesystem block 'bno' skipping any frags
+ * set in the mask.
+ */
+static void
+blk_free(ufs2_daddr_t bno, int mask, int frags)
+{
+	ufs1_daddr_t fragno, cgbno;
+	struct suj_cg *sc;
+	struct cg *cgp;
+	int i, cg;
+	uint8_t *blksfree;
+
+	if (debug)
+		printf("Freeing %d frags at blk %jd\n", frags, bno);
+	cg = dtog(fs, bno);
+	sc = cg_lookup(cg);
+	cgp = sc->sc_cgp;
+	cgbno = dtogd(fs, bno);
+	blksfree = cg_blksfree(cgp);
+
+	/*
+	 * If it's not allocated we only wrote the journal entry
+	 * and never the bitmaps.  Here we unconditionally clear and
+	 * resolve the cg summary later.
+	 */
+	if (frags == fs->fs_frag && mask == 0) {
+		fragno = fragstoblks(fs, cgbno);
+		ffs_setblock(fs, blksfree, fragno);
+		freeblocks++;
+	} else {
+		/*
+		 * deallocate the fragment
+		 */
+		for (i = 0; i < frags; i++)
+			if ((mask & (1 << i)) == 0 && isclr(blksfree, cgbno +i)) {
+				freefrags++;
+				setbit(blksfree, cgbno + i);
+			}
+	}
+	sc->sc_dirty = 1;
+}
+
+/*
+ * Returns 1 if the whole block starting at 'bno' is marked free and 0
+ * otherwise.
+ */
+static int
+blk_isfree(ufs2_daddr_t bno)
+{
+	struct suj_cg *sc;
+
+	sc = cg_lookup(dtog(fs, bno));
+	return ffs_isblock(fs, cg_blksfree(sc->sc_cgp), dtogd(fs, bno));
+}
+
+/*
+ * Fetch an indirect block to find the block at a given lbn.  The lbn
+ * may be negative to fetch a specific indirect block pointer or positive
+ * to fetch a specific block.
+ */
+static ufs2_daddr_t
+indir_blkatoff(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t cur, ufs_lbn_t lbn)
+{
+	ufs2_daddr_t *bap2;
+	ufs2_daddr_t *bap1;
+	ufs_lbn_t lbnadd;
+	ufs_lbn_t base;
+	int level;
+	int i;
+
+	if (blk == 0)
+		return (0);
+	level = lbn_level(cur);
+	if (level == -1)
+		errx(1, "Invalid indir lbn %jd", lbn);
+	if (level == 0 && lbn < 0)
+		errx(1, "Invalid lbn %jd", lbn);
+	bap2 = (void *)dblk_read(blk, fs->fs_bsize);
+	bap1 = (void *)bap2;
+	lbnadd = 1;
+	base = -(cur + level);
+	for (i = level; i > 0; i--)
+		lbnadd *= NINDIR(fs);
+	if (lbn > 0) 
+		i = (lbn - base) / lbnadd;
+	else
+		i = (-lbn - base) / lbnadd;
+	if (i < 0 || i >= NINDIR(fs))
+		errx(1, "Invalid indirect index %d produced by lbn %jd",
+		    i, lbn);
+	if (level == 0)
+		cur = base + (i * lbnadd);
+	else
+		cur = -(base + (i * lbnadd)) - (level - 1);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		blk = bap1[i];
+	else
+		blk = bap2[i];
+	if (cur == lbn)
+		return (blk);
+	if (level == 0) {
+		abort();
+		errx(1, "Invalid lbn %jd at level 0", lbn);
+	}
+	return indir_blkatoff(blk, ino, cur, lbn);
+}
+
+/*
+ * Finds the disk block address at the specified lbn within the inode
+ * specified by ip.  This follows the whole tree and honors di_size and
+ * di_extsize so it is a true test of reachability.  The lbn may be
+ * negative if an extattr or indirect block is requested.
+ */
+static ufs2_daddr_t
+ino_blkatoff(union dinode *ip, ino_t ino, ufs_lbn_t lbn, int *frags)
+{
+	ufs_lbn_t tmpval;
+	ufs_lbn_t cur;
+	ufs_lbn_t next;
+	int i;
+
+	/*
+	 * Handle extattr blocks first.
+	 */
+	if (lbn < 0 && lbn >= -NXADDR) {
+		lbn = -1 - lbn;
+		if (lbn > lblkno(fs, ip->dp2.di_extsize - 1))
+			return (0);
+		*frags = numfrags(fs, sblksize(fs, ip->dp2.di_extsize, lbn));
+		return (ip->dp2.di_extb[lbn]);
+	}
+	/*
+	 * Now direct and indirect.
+	 */
+	if (DIP(ip, di_mode) == IFLNK &&
+	    DIP(ip, di_size) < fs->fs_maxsymlinklen)
+		return (0);
+	if (lbn >= 0 && lbn < NDADDR) {
+		*frags = numfrags(fs, sblksize(fs, DIP(ip, di_size), lbn));
+		return (DIP(ip, di_db[lbn]));
+	}
+	*frags = fs->fs_frag;
+
+	for (i = 0, tmpval = NINDIR(fs), cur = NDADDR; i < NIADDR; i++,
+	    tmpval *= NINDIR(fs), cur = next) {
+		next = cur + tmpval;
+		if (lbn == -cur - i)
+			return (DIP(ip, di_ib[i]));
+		/*
+		 * Determine whether the lbn in question is within this tree.
+		 */
+		if (lbn < 0 && -lbn >= next)
+			continue;
+		if (lbn > 0 && lbn >= next)
+			continue;
+		return indir_blkatoff(DIP(ip, di_ib[i]), ino, -cur - i, lbn);
+	}
+	errx(1, "lbn %jd not in ino", lbn);
+}
+
+/*
+ * Determine whether a block exists at a particular lbn in an inode.
+ * Returns 1 if found, 0 if not.  lbn may be negative for indirects
+ * or ext blocks.
+ */
+static int
+blk_isat(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int *frags)
+{
+	union dinode *ip;
+	ufs2_daddr_t nblk;
+
+	ip = ino_read(ino);
+
+	if (DIP(ip, di_nlink) == 0 || DIP(ip, di_mode) == 0)
+		return (0);
+	nblk = ino_blkatoff(ip, ino, lbn, frags);
+
+	return (nblk == blk);
+}
+
+/*
+ * Determines whether a pointer to an inode exists within a directory
+ * at a specified offset.  Returns the mode of the found entry.
+ */
+static int
+ino_isat(ino_t parent, off_t diroff, ino_t child, int *mode, int *isdot)
+{
+	union dinode *dip;
+	struct direct *dp;
+	ufs2_daddr_t blk;
+	uint8_t *block;
+	ufs_lbn_t lbn;
+	int blksize;
+	int frags;
+	int dpoff;
+	int doff;
+
+	*isdot = 0;
+	dip = ino_read(parent);
+	*mode = DIP(dip, di_mode);
+	if ((*mode & IFMT) != IFDIR) {
+		if (debug) {
+			/*
+			 * This can happen if the parent inode
+			 * was reallocated.
+			 */
+			if (*mode != 0)
+				printf("Directory %d has bad mode %o\n",
+				    parent, *mode);
+			else
+				printf("Directory %d zero inode\n", parent);
+		}
+		return (0);
+	}
+	lbn = lblkno(fs, diroff);
+	doff = blkoff(fs, diroff);
+	blksize = sblksize(fs, DIP(dip, di_size), lbn);
+	if (diroff + DIRECTSIZ(1) > DIP(dip, di_size) || doff >= blksize) {
+		if (debug)
+			printf("ino %d absent from %d due to offset %jd"
+			    " exceeding size %jd\n",
+			    child, parent, diroff, DIP(dip, di_size));
+		return (0);
+	}
+	blk = ino_blkatoff(dip, parent, lbn, &frags);
+	if (blk <= 0) {
+		if (debug)
+			printf("Sparse directory %d", parent);
+		return (0);
+	}
+	block = dblk_read(blk, blksize);
+	/*
+	 * Walk through the records from the start of the block to be
+	 * certain we hit a valid record and not some junk in the middle
+	 * of a file name.  Stop when we reach or pass the expected offset.
+	 */
+	dpoff = (doff / DIRBLKSIZ) * DIRBLKSIZ;
+	do {
+		dp = (struct direct *)&block[dpoff];
+		if (dpoff == doff)
+			break;
+		if (dp->d_reclen == 0)
+			break;
+		dpoff += dp->d_reclen;
+	} while (dpoff <= doff);
+	if (dpoff > fs->fs_bsize)
+		errx(1, "Corrupt directory block in dir ino %d", parent);
+	/* Not found. */
+	if (dpoff != doff) {
+		if (debug)
+			printf("ino %d not found in %d, lbn %jd, dpoff %d\n",
+			    child, parent, lbn, dpoff);
+		return (0);
+	}
+	/*
+	 * We found the item in question.  Record the mode and whether it's
+	 * a . or .. link for the caller.
+	 */
+	if (dp->d_ino == child) {
+		if (child == parent)
+			*isdot = 1;
+		else if (dp->d_namlen == 2 &&
+		    dp->d_name[0] == '.' && dp->d_name[1] == '.')
+			*isdot = 1;
+		*mode = DTTOIF(dp->d_type);
+		return (1);
+	}
+	if (debug)
+		printf("ino %d doesn't match dirent ino %d in parent %d\n",
+		    child, dp->d_ino, parent);
+	return (0);
+}
+
+#define	VISIT_INDIR	0x0001
+#define	VISIT_EXT	0x0002
+#define	VISIT_ROOT	0x0004	/* Operation came via root & valid pointers. */
+
+/*
+ * Read an indirect level which may or may not be linked into an inode.
+ */
+static void
+indir_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, uint64_t *frags,
+    ino_visitor visitor, int flags)
+{
+	ufs2_daddr_t *bap2;
+	ufs1_daddr_t *bap1;
+	ufs_lbn_t lbnadd;
+	ufs2_daddr_t nblk;
+	ufs_lbn_t nlbn;
+	int level;
+	int i;
+
+	/*
+	 * Don't visit indirect blocks with contents we can't trust.  This
+	 * should only happen when indir_visit() is called to complete a
+	 * truncate that never finished and not when a pointer is found via
+	 * an inode.
+	 */
+	if (blk == 0)
+		return;
+	level = lbn_level(lbn);
+	if (level == -1)
+		errx(1, "Invalid level for lbn %jd", lbn);
+	if ((flags & VISIT_ROOT) == 0 && blk_isindir(blk, ino, lbn) == 0) {
+		if (debug)
+			printf("blk %jd ino %d lbn %jd(%d) is not indir.\n",
+			    blk, ino, lbn, level);
+		goto out;
+	}
+	lbnadd = 1;
+	for (i = level; i > 0; i--)
+		lbnadd *= NINDIR(fs);
+	bap1 = (void *)dblk_read(blk, fs->fs_bsize);
+	bap2 = (void *)bap1;
+	for (i = 0; i < NINDIR(fs); i++) {
+		if (fs->fs_magic == FS_UFS1_MAGIC)
+			nblk = *bap1++;
+		else
+			nblk = *bap2++;
+		if (nblk == 0)
+			continue;
+		if (level == 0) {
+			nlbn = -lbn + i * lbnadd;
+			(*frags) += fs->fs_frag;
+			visitor(ino, nlbn, nblk, fs->fs_frag);
+		} else {
+			nlbn = (lbn + 1) - (i * lbnadd);
+			indir_visit(ino, nlbn, nblk, frags, visitor, flags);
+		}
+	}
+out:
+	if (flags & VISIT_INDIR) {
+		(*frags) += fs->fs_frag;
+		visitor(ino, lbn, blk, fs->fs_frag);
+	}
+}
+
+/*
+ * Visit each block in an inode as specified by 'flags' and call a
+ * callback function.  The callback may inspect or free blocks.  The
+ * count of frags found according to the size in the file is returned.
+ * This is not valid for sparse files but may be used to determine
+ * the correct di_blocks for a file.
+ */
+static uint64_t
+ino_visit(union dinode *ip, ino_t ino, ino_visitor visitor, int flags)
+{
+	ufs_lbn_t nextlbn;
+	ufs_lbn_t tmpval;
+	ufs_lbn_t lbn;
+	uint64_t size;
+	uint64_t fragcnt;
+	int mode;
+	int frags;
+	int i;
+
+	size = DIP(ip, di_size);
+	mode = DIP(ip, di_mode) & IFMT;
+	fragcnt = 0;
+	if ((flags & VISIT_EXT) &&
+	    fs->fs_magic == FS_UFS2_MAGIC && ip->dp2.di_extsize) {
+		for (i = 0; i < NXADDR; i++) {
+			if (ip->dp2.di_extb[i] == 0)
+				continue;
+			frags = sblksize(fs, ip->dp2.di_extsize, i);
+			frags = numfrags(fs, frags);
+			fragcnt += frags;
+			visitor(ino, -1 - i, ip->dp2.di_extb[i], frags);
+		}
+	}
+	/* Skip datablocks for short links and devices. */
+	if (mode == IFBLK || mode == IFCHR ||
+	    (mode == IFLNK && size < fs->fs_maxsymlinklen))
+		return (fragcnt);
+	for (i = 0; i < NDADDR; i++) {
+		if (DIP(ip, di_db[i]) == 0)
+			continue;
+		frags = sblksize(fs, size, i);
+		frags = numfrags(fs, frags);
+		fragcnt += frags;
+		visitor(ino, i, DIP(ip, di_db[i]), frags);
+	}
+	/*
+	 * We know the following indirects are real as we're following
+	 * real pointers to them.
+	 */
+	flags |= VISIT_ROOT;
+	for (i = 0, tmpval = NINDIR(fs), lbn = NDADDR; i < NIADDR; i++,
+	    lbn = nextlbn) {
+		nextlbn = lbn + tmpval;
+		tmpval *= NINDIR(fs);
+		if (DIP(ip, di_ib[i]) == 0)
+			continue;
+		indir_visit(ino, -lbn - i, DIP(ip, di_ib[i]), &fragcnt, visitor,
+		    flags);
+	}
+	return (fragcnt);
+}
+
+/*
+ * Null visitor function used when we just want to count blocks and
+ * record the lbn.
+ */
+ufs_lbn_t visitlbn;
+static void
+null_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	if (lbn > 0)
+		visitlbn = lbn;
+}
+
+/*
+ * Recalculate di_blocks when we discover that a block allocation or
+ * free was not successfully completed.  The kernel does not roll this back
+ * because it would be too expensive to compute which indirects were
+ * reachable at the time the inode was written.
+ */
+static void
+ino_adjblks(struct suj_ino *sino)
+{
+	union dinode *ip;
+	uint64_t blocks;
+	uint64_t frags;
+	off_t isize;
+	off_t size;
+	ino_t ino;
+
+	ino = sino->si_ino;
+	ip = ino_read(ino);
+	/* No need to adjust zero'd inodes. */
+	if (DIP(ip, di_mode) == 0)
+		return;
+	/*
+	 * Visit all blocks and count them as well as recording the last
+	 * valid lbn in the file.  If the file size doesn't agree with the
+	 * last lbn we need to truncate to fix it.  Otherwise just adjust
+	 * the blocks count.
+	 */
+	visitlbn = 0;
+	frags = ino_visit(ip, ino, null_visit, VISIT_INDIR | VISIT_EXT);
+	blocks = fsbtodb(fs, frags);
+	/*
+	 * We assume the size and direct block list is kept coherent by
+	 * softdep.  For files that have extended into indirects we truncate
+	 * to the size in the inode or the maximum size permitted by
+	 * populated indirects.
+	 */
+	if (visitlbn >= NDADDR) {
+		isize = DIP(ip, di_size);
+		size = lblktosize(fs, visitlbn + 1);
+		if (isize > size)
+			isize = size;
+		/* Always truncate to free any unpopulated indirects. */
+		ino_trunc(sino->si_ino, isize);
+		return;
+	}
+	if (blocks == DIP(ip, di_blocks))
+		return;
+	if (debug)
+		printf("ino %d adjusting block count from %jd to %jd\n",
+		    ino, DIP(ip, di_blocks), blocks);
+	DIP_SET(ip, di_blocks, blocks);
+	ino_dirty(ino);
+}
+
+static void
+blk_free_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	int mask;
+
+	mask = blk_freemask(blk, ino, lbn, frags);
+	if (debug)
+		printf("blk %jd freemask 0x%X\n", blk, mask);
+	blk_free(blk, mask, frags);
+}
+
+/*
+ * Free a block or tree of blocks that was previously rooted in ino at
+ * the given lbn.  If the lbn is an indirect all children are freed
+ * recursively.
+ */
+static void
+blk_free_lbn(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags, int follow)
+{
+	uint64_t resid;
+	int mask;
+
+	mask = blk_freemask(blk, ino, lbn, frags);
+	if (debug)
+		printf("blk %jd freemask 0x%X\n", blk, mask);
+	resid = 0;
+	if (lbn <= -NDADDR && follow && mask == 0)
+		indir_visit(ino, lbn, blk, &resid, blk_free_visit, VISIT_INDIR);
+	else
+		blk_free(blk, mask, frags);
+}
+
+static void
+ino_setskip(struct suj_ino *sino, ino_t parent)
+{
+	int isdot;
+	int mode;
+
+	if (ino_isat(sino->si_ino, DOTDOT_OFFSET, parent, &mode, &isdot))
+		sino->si_skipparent = 1;
+}
+
+/*
+ * Free the children of a directory when the directory is discarded.
+ */
+static void
+ino_free_children(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	struct suj_ino *sino;
+	struct suj_rec *srec;
+	struct jrefrec *rrec;
+	struct direct *dp;
+	off_t diroff;
+	uint8_t *block;
+	int skipparent;
+	int isparent;
+	int dpoff;
+	int size;
+
+	sino = ino_lookup(ino, 0);
+	if (sino)
+		skipparent = sino->si_skipparent;
+	else
+		skipparent = 0;
+	size = lfragtosize(fs, frags);
+	block = dblk_read(blk, size);
+	dp = (struct direct *)&block[0];
+	for (dpoff = 0; dpoff < size && dp->d_reclen; dpoff += dp->d_reclen) {
+		dp = (struct direct *)&block[dpoff];
+		if (dp->d_ino == 0 || dp->d_ino == WINO)
+			continue;
+		if (dp->d_namlen == 1 && dp->d_name[0] == '.')
+			continue;
+		isparent = dp->d_namlen == 2 && dp->d_name[0] == '.' &&
+		    dp->d_name[1] == '.';
+		if (isparent && skipparent == 1)
+			continue;
+		if (debug)
+			printf("Directory %d removing ino %d name %s\n",
+			    ino, dp->d_ino, dp->d_name);
+		/*
+		 * Lookup this inode to see if we have a record for it.
+		 * If not, we've already adjusted it assuming this path
+		 * was valid and we have to adjust once more.
+		 */
+		sino = ino_lookup(dp->d_ino, 0);
+		if (sino == NULL || sino->si_hasrecs == 0) {
+			ino_decr(ino);
+			continue;
+		}
+		/*
+		 * Use ino_adjust() so if we lose the last non-dot reference
+		 * to a directory it can be discarded.
+		 */
+		if (sino->si_linkadj) {
+			sino->si_nlink--;
+			if (isparent)
+				sino->si_dotlinks--;
+			ino_adjust(sino);
+		}
+		/*
+		 * Tell any child directories we've already removed their
+		 * parent.  Don't try to adjust our link down again.
+		 */
+		if (isparent == 0)
+			ino_setskip(sino, ino);
+		/*
+		 * If we haven't yet processed this inode we need to make
+		 * sure we will successfully discover the lost path.  If not
+		 * use nlinkadj to remember.
+		 */
+		diroff = lblktosize(fs, lbn) + dpoff;
+		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+			rrec = (struct jrefrec *)srec->sr_rec;
+			if (rrec->jr_parent == ino &&
+			    rrec->jr_diroff == diroff)
+				break;
+		}
+		if (srec == NULL)
+			sino->si_nlinkadj++;
+	}
+}
+
+/*
+ * Reclaim an inode, freeing all blocks and decrementing all children's
+ * link counts.  Free the inode back to the cg.
+ */
+static void
+ino_reclaim(union dinode *ip, ino_t ino, int mode)
+{
+	uint32_t gen;
+
+	if (ino == ROOTINO)
+		errx(1, "Attempting to free ROOTINO");
+	if (debug)
+		printf("Truncating and freeing ino %d, nlink %d, mode %o\n",
+		    ino, DIP(ip, di_nlink), DIP(ip, di_mode));
+
+	/* We are freeing an inode or directory. */
+	if ((DIP(ip, di_mode) & IFMT) == IFDIR)
+		ino_visit(ip, ino, ino_free_children, 0);
+	DIP_SET(ip, di_nlink, 0);
+	ino_visit(ip, ino, blk_free_visit, VISIT_EXT | VISIT_INDIR);
+	/* Here we have to clear the inode and release any blocks it holds. */
+	gen = DIP(ip, di_gen);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		bzero(ip, sizeof(struct ufs1_dinode));
+	else
+		bzero(ip, sizeof(struct ufs2_dinode));
+	DIP_SET(ip, di_gen, gen);
+	ino_dirty(ino);
+	ino_free(ino, mode);
+	return;
+}
+
+/*
+ * Adjust an inode's link count down by one when a directory goes away.
+ */
+static void
+ino_decr(ino_t ino)
+{
+	union dinode *ip;
+	int reqlink;
+	int nlink;
+	int mode;
+
+	ip = ino_read(ino);
+	nlink = DIP(ip, di_nlink);
+	mode = DIP(ip, di_mode);
+	if (nlink < 1)
+		errx(1, "Inode %d link count %d invalid", ino, nlink);
+	if (mode == 0)
+		errx(1, "Inode %d has a link of %d with 0 mode.", ino, nlink);
+	nlink--;
+	if ((mode & IFMT) == IFDIR)
+		reqlink = 2;
+	else
+		reqlink = 1;
+	if (nlink < reqlink) {
+		if (debug)
+			printf("ino %d not enough links to live %d < %d\n",
+			    ino, nlink, reqlink);
+		ino_reclaim(ip, ino, mode);
+		return;
+	}
+	DIP_SET(ip, di_nlink, nlink);
+	ino_dirty(ino);
+}
+
+/*
+ * Adjust the inode link count to 'nlink'.  If the count reaches zero
+ * free it.
+ */
+static void
+ino_adjust(struct suj_ino *sino)
+{
+	struct jrefrec *rrec;
+	struct suj_rec *srec;
+	struct suj_ino *stmp;
+	union dinode *ip;
+	nlink_t nlink;
+	int reqlink;
+	int mode;
+	ino_t ino;
+
+	nlink = sino->si_nlink;
+	ino = sino->si_ino;
+	/*
+	 * If it's a directory with no real names pointing to it go ahead
+	 * and truncate it.  This will free any children.
+	 */
+	if ((sino->si_mode & IFMT) == IFDIR &&
+	    nlink - sino->si_dotlinks == 0) {
+		sino->si_nlink = nlink = 0;
+		/*
+		 * Mark any .. links so they know not to free this inode
+		 * when they are removed.
+		 */
+		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+			rrec = (struct jrefrec *)srec->sr_rec;
+			if (rrec->jr_diroff == DOTDOT_OFFSET) {
+				stmp = ino_lookup(rrec->jr_parent, 0);
+				if (stmp)
+					ino_setskip(stmp, ino);
+			}
+		}
+	}
+	ip = ino_read(ino);
+	mode = DIP(ip, di_mode) & IFMT;
+	if (nlink > LINK_MAX)
+		errx(1,
+		    "ino %d nlink manipulation error, new link %d, old link %d",
+		    ino, nlink, DIP(ip, di_nlink));
+	if (debug)
+		printf("Adjusting ino %d, nlink %d, old link %d lastmode %o\n",
+		    ino, nlink, DIP(ip, di_nlink), sino->si_mode);
+	if (mode == 0) {
+		if (debug)
+			printf("ino %d, zero inode freeing bitmap\n", ino);
+		ino_free(ino, sino->si_mode);
+		return;
+	}
+	/* XXX Should be an assert? */
+	if (mode != sino->si_mode && debug)
+		printf("ino %d, mode %o != %o\n", ino, mode, sino->si_mode);
+	if ((mode & IFMT) == IFDIR)
+		reqlink = 2;
+	else
+		reqlink = 1;
+	/* If the inode doesn't have enough links to live, free it. */
+	if (nlink < reqlink) {
+		if (debug)
+			printf("ino %d not enough links to live %d < %d\n",
+			    ino, nlink, reqlink);
+		ino_reclaim(ip, ino, mode);
+		return;
+	}
+	/* If required write the updated link count. */
+	if (DIP(ip, di_nlink) == nlink) {
+		if (debug)
+			printf("ino %d, link matches, skipping.\n", ino);
+		return;
+	}
+	DIP_SET(ip, di_nlink, nlink);
+	ino_dirty(ino);
+}
+
+/*
+ * Truncate some or all blocks in an indirect, freeing any that are required
+ * and zeroing the indirect.
+ */
+static void
+indir_trunc(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, ufs_lbn_t lastlbn)
+{
+	ufs2_daddr_t *bap2;
+	ufs1_daddr_t *bap1;
+	ufs_lbn_t lbnadd;
+	ufs2_daddr_t nblk;
+	ufs_lbn_t next;
+	ufs_lbn_t nlbn;
+	int dirty;
+	int level;
+	int i;
+
+	if (blk == 0)
+		return;
+	dirty = 0;
+	level = lbn_level(lbn);
+	if (level == -1)
+		errx(1, "Invalid level for lbn %jd", lbn);
+	lbnadd = 1;
+	for (i = level; i > 0; i--)
+		lbnadd *= NINDIR(fs);
+	bap1 = (void *)dblk_read(blk, fs->fs_bsize);
+	bap2 = (void *)bap1;
+	for (i = 0; i < NINDIR(fs); i++) {
+		if (fs->fs_magic == FS_UFS1_MAGIC)
+			nblk = *bap1++;
+		else
+			nblk = *bap2++;
+		if (nblk == 0)
+			continue;
+		if (level != 0) {
+			nlbn = (lbn + 1) - (i * lbnadd);
+			/*
+			 * Calculate the lbn of the next indirect to
+			 * determine if any of this indirect must be
+			 * reclaimed.
+			 */
+			next = -(lbn + level) + ((i+1) * lbnadd);
+			if (next <= lastlbn)
+				continue;
+			indir_trunc(ino, nlbn, nblk, lastlbn);
+			/* If all of this indirect was reclaimed, free it. */
+			nlbn = next - lbnadd;
+			if (nlbn < lastlbn)
+				continue;
+		} else {
+			nlbn = -lbn + i * lbnadd;
+			if (nlbn < lastlbn)
+				continue;
+		}
+		dirty = 1;
+		blk_free(nblk, 0, fs->fs_frag);
+		if (fs->fs_magic == FS_UFS1_MAGIC)
+			*(bap1 - 1) = 0;
+		else
+			*(bap2 - 1) = 0;
+	}
+	if (dirty)
+		dblk_dirty(blk);
+}
+
+/*
+ * Truncate an inode to the minimum of the given size or the last populated
+ * block after any over size have been discarded.  The kernel would allocate
+ * the last block in the file but fsck does not and neither do we.  This
+ * code never extends files, only shrinks them.
+ */
+static void
+ino_trunc(ino_t ino, off_t size)
+{
+	union dinode *ip;
+	ufs2_daddr_t bn;
+	uint64_t totalfrags;
+	ufs_lbn_t nextlbn;
+	ufs_lbn_t lastlbn;
+	ufs_lbn_t tmpval;
+	ufs_lbn_t lbn;
+	ufs_lbn_t i;
+	int frags;
+	off_t cursize;
+	off_t off;
+	int mode;
+
+	ip = ino_read(ino);
+	mode = DIP(ip, di_mode) & IFMT;
+	cursize = DIP(ip, di_size);
+	if (debug)
+		printf("Truncating ino %d, mode %o to size %jd from size %jd\n",
+		    ino, mode, size, cursize);
+
+	/* Skip datablocks for short links and devices. */
+	if (mode == 0 || mode == IFBLK || mode == IFCHR ||
+	    (mode == IFLNK && cursize < fs->fs_maxsymlinklen))
+		return;
+	/* Don't extend. */
+	if (size > cursize)
+		size = cursize;
+	lastlbn = lblkno(fs, blkroundup(fs, size));
+	for (i = lastlbn; i < NDADDR; i++) {
+		if (DIP(ip, di_db[i]) == 0)
+			continue;
+		frags = sblksize(fs, cursize, i);
+		frags = numfrags(fs, frags);
+		blk_free(DIP(ip, di_db[i]), 0, frags);
+		DIP_SET(ip, di_db[i], 0);
+	}
+	/*
+	 * Follow indirect blocks, freeing anything required.
+	 */
+	for (i = 0, tmpval = NINDIR(fs), lbn = NDADDR; i < NIADDR; i++,
+	    lbn = nextlbn) {
+		nextlbn = lbn + tmpval;
+		tmpval *= NINDIR(fs);
+		/* If we're not freeing any in this indirect range skip it. */
+		if (lastlbn >= nextlbn)
+			continue;
+		if (DIP(ip, di_ib[i]) == 0)
+			continue;
+		indir_trunc(ino, -lbn - i, DIP(ip, di_ib[i]), lastlbn);
+		/* If we freed everything in this indirect free the indir. */
+		if (lastlbn > lbn)
+			continue;
+		blk_free(DIP(ip, di_ib[i]), 0, frags);
+		DIP_SET(ip, di_ib[i], 0);
+	}
+	ino_dirty(ino);
+	/*
+	 * Now that we've freed any whole blocks that exceed the desired
+	 * truncation size, figure out how many blocks remain and what the
+	 * last populated lbn is.  We will set the size to this last lbn
+	 * rather than worrying about allocating the final lbn as the kernel
+	 * would've done.  This is consistent with normal fsck behavior.
+	 */ 
+	visitlbn = 0;
+	totalfrags = ino_visit(ip, ino, null_visit, VISIT_INDIR | VISIT_EXT);
+	if (size > lblktosize(fs, visitlbn + 1))
+		size = lblktosize(fs, visitlbn + 1);
+	/*
+	 * If we're truncating direct blocks we have to adjust frags
+	 * accordingly.
+	 */
+	if (visitlbn < NDADDR && totalfrags) {
+		long oldspace, newspace;
+
+		bn = DIP(ip, di_db[visitlbn]);
+		if (bn == 0)
+			errx(1, "Bad blk at ino %d lbn %jd\n", ino, visitlbn);
+		oldspace = sblksize(fs, cursize, visitlbn);
+		newspace = sblksize(fs, size, visitlbn);
+		if (oldspace != newspace) {
+			bn += numfrags(fs, newspace);
+			frags = numfrags(fs, oldspace - newspace);
+			blk_free(bn, 0, frags);
+			totalfrags -= frags;
+		}
+	}
+	DIP_SET(ip, di_blocks, fsbtodb(fs, totalfrags));
+	DIP_SET(ip, di_size, size);
+	/*
+	 * If we've truncated into the middle of a block or frag we have
+	 * to zero it here.  Otherwise the file could extend into
+	 * uninitialized space later.
+	 */
+	off = blkoff(fs, size);
+	if (off) {
+		uint8_t *buf;
+		long clrsize;
+
+		bn = ino_blkatoff(ip, ino, visitlbn, &frags);
+		if (bn == 0)
+			errx(1, "Block missing from ino %d at lbn %jd\n",
+			    ino, visitlbn);
+		clrsize = frags * fs->fs_fsize;
+		buf = dblk_read(bn, clrsize);
+		clrsize -= off;
+		buf += off;
+		bzero(buf, clrsize);
+		dblk_dirty(bn);
+	}
+	return;
+}
+
+/*
+ * Process records available for one inode and determine whether the
+ * link count is correct or needs adjusting.
+ */
+static void
+ino_check(struct suj_ino *sino)
+{
+	struct suj_rec *srec;
+	struct jrefrec *rrec;
+	nlink_t dotlinks;
+	int newlinks;
+	int removes;
+	int nlink;
+	ino_t ino;
+	int isdot;
+	int isat;
+	int mode;
+
+	if (sino->si_hasrecs == 0)
+		return;
+	ino = sino->si_ino;
+	rrec = (struct jrefrec *)TAILQ_FIRST(&sino->si_recs)->sr_rec;
+	nlink = rrec->jr_nlink;
+	newlinks = 0;
+	dotlinks = 0;
+	removes = sino->si_nlinkadj;
+	TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+		rrec = (struct jrefrec *)srec->sr_rec;
+		isat = ino_isat(rrec->jr_parent, rrec->jr_diroff, 
+		    rrec->jr_ino, &mode, &isdot);
+		if (isat && (mode & IFMT) != (rrec->jr_mode & IFMT))
+			errx(1, "Inode mode/directory type mismatch %o != %o",
+			    mode, rrec->jr_mode);
+		if (debug)
+			printf("jrefrec: op %d ino %d, nlink %d, parent %d, "
+			    "diroff %jd, mode %o, isat %d, isdot %d\n",
+			    rrec->jr_op, rrec->jr_ino, rrec->jr_nlink,
+			    rrec->jr_parent, rrec->jr_diroff, rrec->jr_mode,
+			    isat, isdot);
+		mode = rrec->jr_mode & IFMT;
+		if (rrec->jr_op == JOP_REMREF)
+			removes++;
+		newlinks += isat;
+		if (isdot)
+			dotlinks += isat;
+	}
+	/*
+	 * The number of links that remain are the starting link count
+	 * subtracted by the total number of removes with the total
+	 * links discovered back in.  An incomplete remove thus
+	 * makes no change to the link count but an add increases
+	 * by one.
+	 */
+	if (debug)
+		printf("ino %d nlink %d newlinks %d removes %d dotlinks %d\n",
+		    ino, nlink, newlinks, removes, dotlinks);
+	nlink += newlinks;
+	nlink -= removes;
+	sino->si_linkadj = 1;
+	sino->si_nlink = nlink;
+	sino->si_dotlinks = dotlinks;
+	sino->si_mode = mode;
+	ino_adjust(sino);
+}
+
+/*
+ * Process records available for one block and determine whether it is
+ * still allocated and whether the owning inode needs to be updated or
+ * a free completed.
+ */
+static void
+blk_check(struct suj_blk *sblk)
+{
+	struct suj_rec *srec;
+	struct jblkrec *brec;
+	struct suj_ino *sino;
+	ufs2_daddr_t blk;
+	int mask;
+	int frags;
+	int isat;
+
+	/*
+	 * Each suj_blk actually contains records for any fragments in that
+	 * block.  As a result we must evaluate each record individually.
+	 */
+	sino = NULL;
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		brec = (struct jblkrec *)srec->sr_rec;
+		frags = brec->jb_frags;
+		blk = brec->jb_blkno + brec->jb_oldfrags;
+		isat = blk_isat(brec->jb_ino, brec->jb_lbn, blk, &frags);
+		if (sino == NULL || sino->si_ino != brec->jb_ino) {
+			sino = ino_lookup(brec->jb_ino, 1);
+			sino->si_blkadj = 1;
+		}
+		if (debug)
+			printf("op %d blk %jd ino %d lbn %jd frags %d isat %d (%d)\n",
+			    brec->jb_op, blk, brec->jb_ino, brec->jb_lbn,
+			    brec->jb_frags, isat, frags);
+		/*
+		 * If we found the block at this address we still have to
+		 * determine if we need to free the tail end that was
+		 * added by adding contiguous fragments from the same block.
+		 */
+		if (isat == 1) {
+			if (frags == brec->jb_frags)
+				continue;
+			mask = blk_freemask(blk, brec->jb_ino, brec->jb_lbn,
+			    brec->jb_frags);
+			mask >>= frags;
+			blk += frags;
+			frags = brec->jb_frags - frags;
+			blk_free(blk, mask, frags);
+			continue;
+		}
+		/*
+	 	 * The block wasn't found, attempt to free it.  It won't be
+		 * freed if it was actually reallocated.  If this was an
+		 * allocation we don't want to follow indirects as they
+		 * may not be written yet.  Any children of the indirect will
+		 * have their own records.  If it's a free we need to
+		 * recursively free children.
+		 */
+		blk_free_lbn(blk, brec->jb_ino, brec->jb_lbn, brec->jb_frags,
+		    brec->jb_op == JOP_FREEBLK);
+	}
+}
+
+/*
+ * Walk the list of inode records for this cg and resolve moved and duplicate
+ * inode references now that we have a complete picture.
+ */
+static void
+cg_build(struct suj_cg *sc)
+{
+	struct suj_ino *sino;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(sino, &sc->sc_inohash[i], si_next)
+			ino_build(sino);
+}
+
+/*
+ * Handle inodes requiring truncation.  This must be done prior to
+ * looking up any inodes in directories.
+ */
+static void
+cg_trunc(struct suj_cg *sc)
+{
+	struct suj_ino *sino;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(sino, &sc->sc_inohash[i], si_next)
+			if (sino->si_trunc) {
+				ino_trunc(sino->si_ino,
+				    sino->si_trunc->jt_size);
+				sino->si_trunc = NULL;
+			}
+}
+
+/*
+ * Free any partially allocated blocks and then resolve inode block
+ * counts.
+ */
+static void
+cg_check_blk(struct suj_cg *sc)
+{
+	struct suj_ino *sino;
+	struct suj_blk *sblk;
+	int i;
+
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(sblk, &sc->sc_blkhash[i], sb_next)
+			blk_check(sblk);
+	/*
+	 * Now that we've freed blocks which are not referenced we
+	 * make a second pass over all inodes to adjust their block
+	 * counts.
+	 */
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(sino, &sc->sc_inohash[i], si_next)
+			if (sino->si_blkadj)
+				ino_adjblks(sino);
+}
+
+/*
+ * Walk the list of inode records for this cg, recovering any
+ * changes which were not complete at the time of crash.
+ */
+static void
+cg_check_ino(struct suj_cg *sc)
+{
+	struct suj_ino *sino;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(sino, &sc->sc_inohash[i], si_next)
+			ino_check(sino);
+}
+
+/*
+ * Write a potentially dirty cg.  Recalculate the summary information and
+ * update the superblock summary.
+ */
+static void
+cg_write(struct suj_cg *sc)
+{
+	ufs1_daddr_t fragno, cgbno, maxbno;
+	u_int8_t *blksfree;
+	struct cg *cgp;
+	int blk;
+	int i;
+
+	if (sc->sc_dirty == 0)
+		return;
+	/*
+	 * Fix the frag and cluster summary.
+	 */
+	cgp = sc->sc_cgp;
+	cgp->cg_cs.cs_nbfree = 0;
+	cgp->cg_cs.cs_nffree = 0;
+	bzero(&cgp->cg_frsum, sizeof(cgp->cg_frsum));
+	maxbno = fragstoblks(fs, fs->fs_fpg);
+	if (fs->fs_contigsumsize > 0) {
+		for (i = 1; i <= fs->fs_contigsumsize; i++)
+			cg_clustersum(cgp)[i] = 0;
+		bzero(cg_clustersfree(cgp), howmany(maxbno, CHAR_BIT));
+	}
+	blksfree = cg_blksfree(cgp);
+	for (cgbno = 0; cgbno < maxbno; cgbno++) {
+		if (ffs_isfreeblock(fs, blksfree, cgbno))
+			continue;
+		if (ffs_isblock(fs, blksfree, cgbno)) {
+			ffs_clusteracct(fs, cgp, cgbno, 1);
+			cgp->cg_cs.cs_nbfree++;
+			continue;
+		}
+		fragno = blkstofrags(fs, cgbno);
+		blk = blkmap(fs, blksfree, fragno);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, 1);
+		for (i = 0; i < fs->fs_frag; i++)
+			if (isset(blksfree, fragno + i))
+				cgp->cg_cs.cs_nffree++;
+	}
+	/*
+	 * Update the superblock cg summary from our now correct values
+	 * before writing the block.
+	 */
+	fs->fs_cs(fs, sc->sc_cgx) = cgp->cg_cs;
+	if (bwrite(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
+	    fs->fs_bsize) == -1)
+		err(1, "Unable to write cylinder group %d", sc->sc_cgx);
+}
+
+/*
+ * Write out any modified inodes.
+ */
+static void
+cg_write_inos(struct suj_cg *sc)
+{
+	struct ino_blk *iblk;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(iblk, &sc->sc_iblkhash[i], ib_next)
+			if (iblk->ib_dirty)
+				iblk_write(iblk);
+}
+
+static void
+cg_apply(void (*apply)(struct suj_cg *))
+{
+	struct suj_cg *scg;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(scg, &cghash[i], sc_next)
+			apply(scg);
+}
+
+/*
+ * Process the unlinked but referenced file list.  Freeing all inodes.
+ */
+static void
+ino_unlinked(void)
+{
+	union dinode *ip;
+	uint16_t mode;
+	ino_t inon;
+	ino_t ino;
+
+	ino = fs->fs_sujfree;
+	fs->fs_sujfree = 0;
+	while (ino != 0) {
+		ip = ino_read(ino);
+		mode = DIP(ip, di_mode) & IFMT;
+		inon = DIP(ip, di_freelink);
+		DIP_SET(ip, di_freelink, 0);
+		/*
+		 * XXX Should this be an errx?
+		 */
+		if (DIP(ip, di_nlink) == 0) {
+			if (debug)
+				printf("Freeing unlinked ino %d mode %o\n",
+				    ino, mode);
+			ino_reclaim(ip, ino, mode);
+		} else if (debug)
+			printf("Skipping ino %d mode %o with link %d\n",
+			    ino, mode, DIP(ip, di_nlink));
+		ino = inon;
+	}
+}
+
+/*
+ * Append a new record to the list of records requiring processing.
+ */
+static void
+ino_append(union jrec *rec)
+{
+	struct jrefrec *refrec;
+	struct jmvrec *mvrec;
+	struct suj_ino *sino;
+	struct suj_rec *srec;
+
+	mvrec = &rec->rec_jmvrec;
+	refrec = &rec->rec_jrefrec;
+	if (debug && mvrec->jm_op == JOP_MVREF)
+		printf("ino move: ino %d, parent %d, diroff %jd, oldoff %jd\n", 
+		    mvrec->jm_ino, mvrec->jm_parent, mvrec->jm_newoff,
+		    mvrec->jm_oldoff);
+	else if (debug &&
+	    (refrec->jr_op == JOP_ADDREF || refrec->jr_op == JOP_REMREF))
+		printf("ino ref: op %d, ino %d, nlink %d, "
+		    "parent %d, diroff %jd\n", 
+		    refrec->jr_op, refrec->jr_ino, refrec->jr_nlink,
+		    refrec->jr_parent, refrec->jr_diroff);
+	/*
+	 * Lookup the ino and clear truncate if one is found.  Partial
+	 * truncates are always done synchronously so if we discover
+	 * an operation that requires a lock the truncation has completed
+	 * and can be discarded.
+	 */
+	sino = ino_lookup(((struct jrefrec *)rec)->jr_ino, 1);
+	sino->si_trunc = NULL;
+	sino->si_hasrecs = 1;
+	srec = errmalloc(sizeof(*srec));
+	srec->sr_rec = rec;
+	TAILQ_INSERT_TAIL(&sino->si_newrecs, srec, sr_next);
+}
+
+/*
+ * Add a reference adjustment to the sino list and eliminate dups.  The
+ * primary loop in ino_build_ref() checks for dups but new ones may be
+ * created as a result of offset adjustments.
+ */
+static void
+ino_add_ref(struct suj_ino *sino, struct suj_rec *srec)
+{
+	struct jrefrec *refrec;
+	struct suj_rec *srn;
+	struct jrefrec *rrn;
+
+	refrec = (struct jrefrec *)srec->sr_rec;
+	/*
+	 * We walk backwards so that the oldest link count is preserved.  If
+	 * an add record conflicts with a remove keep the remove.  Redundant
+	 * removes are eliminated in ino_build_ref.  Otherwise we keep the
+	 * oldest record at a given location.
+	 */
+	for (srn = TAILQ_LAST(&sino->si_recs, srechd); srn;
+	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
+		rrn = (struct jrefrec *)srn->sr_rec;
+		if (rrn->jr_parent != refrec->jr_parent ||
+		    rrn->jr_diroff != refrec->jr_diroff)
+			continue;
+		if (rrn->jr_op == JOP_REMREF || refrec->jr_op == JOP_ADDREF) {
+			rrn->jr_mode = refrec->jr_mode;
+			return;
+		}
+		/*
+		 * Adding a remove.
+		 *
+		 * Replace the record in place with the old nlink in case
+		 * we replace the head of the list.  Abandon srec as a dup.
+		 */
+		refrec->jr_nlink = rrn->jr_nlink;
+		srn->sr_rec = srec->sr_rec;
+		return;
+	}
+	TAILQ_INSERT_TAIL(&sino->si_recs, srec, sr_next);
+}
+
+/*
+ * Create a duplicate of a reference at a previous location.
+ */
+static void
+ino_dup_ref(struct suj_ino *sino, struct jrefrec *refrec, off_t diroff)
+{
+	struct jrefrec *rrn;
+	struct suj_rec *srn;
+
+	rrn = errmalloc(sizeof(*refrec));
+	*rrn = *refrec;
+	rrn->jr_op = JOP_ADDREF;
+	rrn->jr_diroff = diroff;
+	srn = errmalloc(sizeof(*srn));
+	srn->sr_rec = (union jrec *)rrn;
+	ino_add_ref(sino, srn);
+}
+
+/*
+ * Add a reference to the list at all known locations.  We follow the offset
+ * changes for a single instance and create duplicate add refs at each so
+ * that we can tolerate any version of the directory block.  Eliminate
+ * removes which collide with adds that are seen in the journal.  They should
+ * not adjust the link count down.
+ */
+static void
+ino_build_ref(struct suj_ino *sino, struct suj_rec *srec)
+{
+	struct jrefrec *refrec;
+	struct jmvrec *mvrec;
+	struct suj_rec *srp;
+	struct suj_rec *srn;
+	struct jrefrec *rrn;
+	off_t diroff;
+
+	refrec = (struct jrefrec *)srec->sr_rec;
+	/*
+	 * Search for a mvrec that matches this offset.  Whether it's an add
+	 * or a remove we can delete the mvref after creating a dup record in
+	 * the old location.
+	 */
+	if (!TAILQ_EMPTY(&sino->si_movs)) {
+		diroff = refrec->jr_diroff;
+		for (srn = TAILQ_LAST(&sino->si_movs, srechd); srn; srn = srp) {
+			srp = TAILQ_PREV(srn, srechd, sr_next);
+			mvrec = (struct jmvrec *)srn->sr_rec;
+			if (mvrec->jm_parent != refrec->jr_parent ||
+			    mvrec->jm_newoff != diroff)
+				continue;
+			diroff = mvrec->jm_oldoff;
+			TAILQ_REMOVE(&sino->si_movs, srn, sr_next);
+			ino_dup_ref(sino, refrec, diroff);
+		}
+	}
+	/*
+	 * If a remove wasn't eliminated by an earlier add just append it to
+	 * the list.
+	 */
+	if (refrec->jr_op == JOP_REMREF) {
+		ino_add_ref(sino, srec);
+		return;
+	}
+	/*
+	 * Walk the list of records waiting to be added to the list.  We
+	 * must check for moves that apply to our current offset and remove
+	 * them from the list.  Remove any duplicates to eliminate removes
+	 * with corresponding adds.
+	 */
+	TAILQ_FOREACH_SAFE(srn, &sino->si_newrecs, sr_next, srp) {
+		switch (srn->sr_rec->rec_jrefrec.jr_op) {
+		case JOP_ADDREF:
+			/*
+			 * This should actually be an error we should
+			 * have a remove for every add journaled.
+			 */
+			rrn = (struct jrefrec *)srn->sr_rec;
+			if (rrn->jr_parent != refrec->jr_parent ||
+			    rrn->jr_diroff != refrec->jr_diroff)
+				break;
+			TAILQ_REMOVE(&sino->si_newrecs, srn, sr_next);
+			break;
+		case JOP_REMREF:
+			/*
+			 * Once we remove the current iteration of the
+			 * record at this address we're done.
+			 */
+			rrn = (struct jrefrec *)srn->sr_rec;
+			if (rrn->jr_parent != refrec->jr_parent ||
+			    rrn->jr_diroff != refrec->jr_diroff)
+				break;
+			TAILQ_REMOVE(&sino->si_newrecs, srn, sr_next);
+			ino_add_ref(sino, srec);
+			return;
+		case JOP_MVREF:
+			/*
+			 * Update our diroff based on any moves that match
+			 * and remove the move.
+			 */
+			mvrec = (struct jmvrec *)srn->sr_rec;
+			if (mvrec->jm_parent != refrec->jr_parent ||
+			    mvrec->jm_oldoff != refrec->jr_diroff)
+				break;
+			ino_dup_ref(sino, refrec, mvrec->jm_oldoff);
+			refrec->jr_diroff = mvrec->jm_newoff;
+			TAILQ_REMOVE(&sino->si_newrecs, srn, sr_next);
+			break;
+		default:
+			errx(1, "ino_build_ref: Unknown op %d",
+			    srn->sr_rec->rec_jrefrec.jr_op);
+		}
+	}
+	ino_add_ref(sino, srec);
+}
+
+/*
+ * Walk the list of new records and add them in-order resolving any
+ * dups and adjusted offsets.
+ */
+static void
+ino_build(struct suj_ino *sino)
+{
+	struct suj_rec *srec;
+
+	while ((srec = TAILQ_FIRST(&sino->si_newrecs)) != NULL) {
+		TAILQ_REMOVE(&sino->si_newrecs, srec, sr_next);
+		switch (srec->sr_rec->rec_jrefrec.jr_op) {
+		case JOP_ADDREF:
+		case JOP_REMREF:
+			ino_build_ref(sino, srec);
+			break;
+		case JOP_MVREF:
+			/*
+			 * Add this mvrec to the queue of pending mvs.
+			 */
+			TAILQ_INSERT_TAIL(&sino->si_movs, srec, sr_next);
+			break;
+		default:
+			errx(1, "ino_build: Unknown op %d",
+			    srec->sr_rec->rec_jrefrec.jr_op);
+		}
+	}
+	if (TAILQ_EMPTY(&sino->si_recs))
+		sino->si_hasrecs = 0;
+}
+
+/*
+ * Modify journal records so they refer to the base block number
+ * and a start and end frag range.  This is to facilitate the discovery
+ * of overlapping fragment allocations.
+ */
+static void
+blk_build(struct jblkrec *blkrec)
+{
+	struct suj_rec *srec;
+	struct suj_blk *sblk;
+	struct jblkrec *blkrn;
+	struct suj_ino *sino;
+	ufs2_daddr_t blk;
+	off_t foff;
+	int frag;
+
+	if (debug)
+		printf("blk_build: op %d blkno %jd frags %d oldfrags %d "
+		    "ino %d lbn %jd\n",
+		    blkrec->jb_op, blkrec->jb_blkno, blkrec->jb_frags,
+		    blkrec->jb_oldfrags, blkrec->jb_ino, blkrec->jb_lbn);
+
+	/*
+	 * Look up the inode and clear the truncate if any lbns after the
+	 * truncate lbn are freed or allocated.
+	 */
+	sino = ino_lookup(blkrec->jb_ino, 0);
+	if (sino && sino->si_trunc) {
+		foff = lblktosize(fs, blkrec->jb_lbn);
+		foff += lfragtosize(fs, blkrec->jb_frags);
+		if (foff > sino->si_trunc->jt_size)
+			sino->si_trunc = NULL;
+	}
+	blk = blknum(fs, blkrec->jb_blkno);
+	frag = fragnum(fs, blkrec->jb_blkno);
+	sblk = blk_lookup(blk, 1);
+	/*
+	 * Rewrite the record using oldfrags to indicate the offset into
+	 * the block.  Leave jb_frags as the actual allocated count.
+	 */
+	blkrec->jb_blkno -= frag;
+	blkrec->jb_oldfrags = frag;
+	if (blkrec->jb_oldfrags + blkrec->jb_frags > fs->fs_frag)
+		errx(1, "Invalid fragment count %d oldfrags %d",
+		    blkrec->jb_frags, frag);
+	/*
+	 * Detect dups.  If we detect a dup we always discard the oldest
+	 * record as it is superseded by the new record.  This speeds up
+	 * later stages but also eliminates free records which are used
+	 * to indicate that the contents of indirects can be trusted.
+	 */
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		blkrn = (struct jblkrec *)srec->sr_rec;
+		if (blkrn->jb_ino != blkrec->jb_ino ||
+		    blkrn->jb_lbn != blkrec->jb_lbn ||
+		    blkrn->jb_blkno != blkrec->jb_blkno ||
+		    blkrn->jb_frags != blkrec->jb_frags ||
+		    blkrn->jb_oldfrags != blkrec->jb_oldfrags)
+			continue;
+		if (debug)
+			printf("Removed dup.\n");
+		/* Discard the free which is a dup with an alloc. */
+		if (blkrec->jb_op == JOP_FREEBLK)
+			return;
+		TAILQ_REMOVE(&sblk->sb_recs, srec, sr_next);
+		free(srec);
+		break;
+	}
+	srec = errmalloc(sizeof(*srec));
+	srec->sr_rec = (union jrec *)blkrec;
+	TAILQ_INSERT_TAIL(&sblk->sb_recs, srec, sr_next);
+}
+
+static void
+ino_build_trunc(struct jtrncrec *rec)
+{
+	struct suj_ino *sino;
+
+	if (debug)
+		printf("ino_build_trunc: ino %d, size %jd\n",
+		    rec->jt_ino, rec->jt_size);
+	sino = ino_lookup(rec->jt_ino, 1);
+	sino->si_trunc = rec;
+}
+
+/*
+ * Build up tables of the operations we need to recover.
+ */
+static void
+suj_build(void)
+{
+	struct suj_seg *seg;
+	union jrec *rec;
+	int off;
+	int i;
+
+	TAILQ_FOREACH(seg, &allsegs, ss_next) {
+		if (debug)
+			printf("seg %jd has %d records, oldseq %jd.\n",
+			    seg->ss_rec.jsr_seq, seg->ss_rec.jsr_cnt,
+			    seg->ss_rec.jsr_oldest);
+		off = 0;
+		rec = (union jrec *)seg->ss_blk;
+		for (i = 0; i < seg->ss_rec.jsr_cnt; off += JREC_SIZE, rec++) {
+			/* skip the segrec. */
+			if ((off % DEV_BSIZE) == 0)
+				continue;
+			switch (rec->rec_jrefrec.jr_op) {
+			case JOP_ADDREF:
+			case JOP_REMREF:
+			case JOP_MVREF:
+				ino_append(rec);
+				break;
+			case JOP_NEWBLK:
+			case JOP_FREEBLK:
+				blk_build((struct jblkrec *)rec);
+				break;
+			case JOP_TRUNC:
+				ino_build_trunc((struct jtrncrec *)rec);
+				break;
+			default:
+				errx(1, "Unknown journal operation %d (%d)",
+				    rec->rec_jrefrec.jr_op, off);
+			}
+			i++;
+		}
+	}
+}
+
+/*
+ * Prune the journal segments to those we care about based on the
+ * oldest sequence in the newest segment.  Order the segment list
+ * based on sequence number.
+ */
+static void
+suj_prune(void)
+{
+	struct suj_seg *seg;
+	struct suj_seg *segn;
+	uint64_t newseq;
+	int discard;
+
+	if (debug)
+		printf("Pruning up to %jd\n", oldseq);
+	/* First free the expired segments. */
+	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
+		if (seg->ss_rec.jsr_seq >= oldseq)
+			continue;
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		free(seg->ss_blk);
+		free(seg);
+	}
+	/* Next ensure that segments are ordered properly. */
+	seg = TAILQ_FIRST(&allsegs);
+	if (seg == NULL) {
+		if (debug)
+			printf("Empty journal\n");
+		return;
+	}
+	newseq = seg->ss_rec.jsr_seq;
+	for (;;) {
+		seg = TAILQ_LAST(&allsegs, seghd);
+		if (seg->ss_rec.jsr_seq >= newseq)
+			break;
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		TAILQ_INSERT_HEAD(&allsegs, seg, ss_next);
+		newseq = seg->ss_rec.jsr_seq;
+		
+	}
+	if (newseq != oldseq)
+		errx(1, "Journal file sequence mismatch %jd != %jd",
+		    newseq, oldseq);
+	/*
+	 * The kernel may asynchronously write segments which can create
+	 * gaps in the sequence space.  Throw away any segments after the
+	 * gap as the kernel guarantees only those that are contiguously
+	 * reachable are marked as completed.
+	 */
+	discard = 0;
+	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
+		if (!discard && newseq++ == seg->ss_rec.jsr_seq) {
+			jrecs += seg->ss_rec.jsr_cnt;
+			jbytes += seg->ss_rec.jsr_blocks * DEV_BSIZE;
+			continue;
+		}
+		discard = 1;
+		if (debug)
+			printf("Journal order mismatch %jd != %jd pruning\n",
+			    newseq-1, seg->ss_rec.jsr_seq);
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		free(seg->ss_blk);
+		free(seg);
+	}
+	if (debug)
+		printf("Processing journal segments from %jd to %jd\n",
+		    oldseq, newseq-1);
+}
+
+/*
+ * Verify the journal inode before attempting to read records.
+ */
+static int
+suj_verifyino(union dinode *ip)
+{
+
+	if (DIP(ip, di_nlink) != 1) {
+		printf("Invalid link count %d for journal inode %d\n",
+		    DIP(ip, di_nlink), sujino);
+		return (-1);
+	}
+
+	if ((DIP(ip, di_flags) & (SF_IMMUTABLE | SF_NOUNLINK)) !=
+	    (SF_IMMUTABLE | SF_NOUNLINK)) {
+		printf("Invalid flags 0x%X for journal inode %d\n",
+		    DIP(ip, di_flags), sujino);
+		return (-1);
+	}
+
+	if (DIP(ip, di_mode) != (IFREG | IREAD)) {
+		printf("Invalid mode %o for journal inode %d\n",
+		    DIP(ip, di_mode), sujino);
+		return (-1);
+	}
+
+	if (DIP(ip, di_size) < SUJ_MIN || DIP(ip, di_size) > SUJ_MAX) {
+		printf("Invalid size %jd for journal inode %d\n",
+		    DIP(ip, di_size), sujino);
+		return (-1);
+	}
+
+	if (DIP(ip, di_modrev) != fs->fs_mtime) {
+		printf("Journal timestamp does not match fs mount time\n");
+		return (-1);
+	}
+
+	return (0);
+}
+
+struct jblocks {
+	struct jextent *jb_extent;	/* Extent array. */
+	int		jb_avail;	/* Available extents. */
+	int		jb_used;	/* Last used extent. */
+	int		jb_head;	/* Allocator head. */
+	int		jb_off;		/* Allocator extent offset. */
+};
+struct jextent {
+	ufs2_daddr_t	je_daddr;	/* Disk block address. */
+	int		je_blocks;	/* Disk block count. */
+};
+
+struct jblocks *suj_jblocks;
+
+static struct jblocks *
+jblocks_create(void)
+{
+	struct jblocks *jblocks;
+	int size;
+
+	jblocks = errmalloc(sizeof(*jblocks));
+	jblocks->jb_avail = 10;
+	jblocks->jb_used = 0;
+	jblocks->jb_head = 0;
+	jblocks->jb_off = 0;
+	size = sizeof(struct jextent) * jblocks->jb_avail;
+	jblocks->jb_extent = errmalloc(size);
+	bzero(jblocks->jb_extent, size);
+
+	return (jblocks);
+}
+
+/*
+ * Return the next available disk block and the amount of contiguous
+ * free space it contains.
+ */
+static ufs2_daddr_t
+jblocks_next(struct jblocks *jblocks, int bytes, int *actual)
+{
+	struct jextent *jext;
+	ufs2_daddr_t daddr;
+	int freecnt;
+	int blocks;
+
+	blocks = bytes / DEV_BSIZE;
+	jext = &jblocks->jb_extent[jblocks->jb_head];
+	freecnt = jext->je_blocks - jblocks->jb_off;
+	if (freecnt == 0) {
+		jblocks->jb_off = 0;
+		if (++jblocks->jb_head > jblocks->jb_used)
+			return (0);
+		jext = &jblocks->jb_extent[jblocks->jb_head];
+		freecnt = jext->je_blocks;
+	}
+	if (freecnt > blocks)
+		freecnt = blocks;
+	*actual = freecnt * DEV_BSIZE;
+	daddr = jext->je_daddr + jblocks->jb_off;
+
+	return (daddr);
+}
+
+/*
+ * Advance the allocation head by a specified number of bytes, consuming
+ * one journal segment.
+ */
+static void
+jblocks_advance(struct jblocks *jblocks, int bytes)
+{
+
+	jblocks->jb_off += bytes / DEV_BSIZE;
+}
+
+static void
+jblocks_destroy(struct jblocks *jblocks)
+{
+
+	free(jblocks->jb_extent);
+	free(jblocks);
+}
+
+static void
+jblocks_add(struct jblocks *jblocks, ufs2_daddr_t daddr, int blocks)
+{
+	struct jextent *jext;
+	int size;
+
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	/* Adding the first block. */
+	if (jext->je_daddr == 0) {
+		jext->je_daddr = daddr;
+		jext->je_blocks = blocks;
+		return;
+	}
+	/* Extending the last extent. */
+	if (jext->je_daddr + jext->je_blocks == daddr) {
+		jext->je_blocks += blocks;
+		return;
+	}
+	/* Adding a new extent. */
+	if (++jblocks->jb_used == jblocks->jb_avail) {
+		jblocks->jb_avail *= 2;
+		size = sizeof(struct jextent) * jblocks->jb_avail;
+		jext = errmalloc(size);
+		bzero(jext, size);
+		bcopy(jblocks->jb_extent, jext,
+		    sizeof(struct jextent) * jblocks->jb_used);
+		free(jblocks->jb_extent);
+		jblocks->jb_extent = jext;
+	}
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	jext->je_daddr = daddr;
+	jext->je_blocks = blocks;
+
+	return;
+}
+
+/*
+ * Add a file block from the journal to the extent map.  We can't read
+ * each file block individually because the kernel treats it as a circular
+ * buffer and segments may span mutliple contiguous blocks.
+ */
+static void
+suj_add_block(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+
+	jblocks_add(suj_jblocks, fsbtodb(fs, blk), fsbtodb(fs, frags));
+}
+
+static void
+suj_read(void)
+{
+	uint8_t block[1 * 1024 * 1024];
+	struct suj_seg *seg;
+	struct jsegrec *recn;
+	struct jsegrec *rec;
+	ufs2_daddr_t blk;
+	int readsize;
+	int blocks;
+	int recsize;
+	int size;
+	int i;
+
+	/*
+	 * Read records until we exhaust the journal space.  If we find
+	 * an invalid record we start searching for a valid segment header
+	 * at the next block.  This is because we don't have a head/tail
+	 * pointer and must recover the information indirectly.  At the gap
+	 * between the head and tail we won't necessarily have a valid
+	 * segment.
+	 */
+restart:
+	for (;;) {
+		size = sizeof(block);
+		blk = jblocks_next(suj_jblocks, size, &readsize);
+		if (blk == 0)
+			return;
+		size = readsize;
+		/*
+		 * Read 1MB at a time and scan for records within this block.
+		 */
+		if (bread(disk, blk, &block, size) == -1)
+			err(1, "Error reading journal block %jd",
+			    (intmax_t)blk);
+		for (rec = (void *)block; size; size -= recsize,
+		    rec = (struct jsegrec *)((uintptr_t)rec + recsize)) {
+			recsize = DEV_BSIZE;
+			if (rec->jsr_time != fs->fs_mtime) {
+				if (debug)
+					printf("Rec time %jd != fs mtime %jd\n",
+					    rec->jsr_time, fs->fs_mtime);
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			if (rec->jsr_cnt == 0) {
+				if (debug)
+					printf("Found illegal count %d\n",
+					    rec->jsr_cnt);
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			blocks = rec->jsr_blocks;
+			recsize = blocks * DEV_BSIZE;
+			if (recsize > size) {
+				/*
+				 * We may just have run out of buffer, restart
+				 * the loop to re-read from this spot.
+				 */
+				if (size < fs->fs_bsize &&
+				    size != readsize &&
+				    recsize <= fs->fs_bsize)
+					goto restart;
+				if (debug)
+					printf("Found invalid segsize %d > %d\n",
+					    recsize, size);
+				recsize = DEV_BSIZE;
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			/*
+			 * Verify that all blocks in the segment are present.
+			 */
+			for (i = 1; i < blocks; i++) {
+				recn = (void *)
+				    ((uintptr_t)rec) + i * DEV_BSIZE;
+				if (recn->jsr_seq == rec->jsr_seq &&
+				    recn->jsr_time == rec->jsr_time) 
+					continue;
+				if (debug)
+					printf("Incomplete record %jd (%d)\n",
+					    rec->jsr_seq, i);
+				recsize = i * DEV_BSIZE;
+				jblocks_advance(suj_jblocks, recsize);
+				goto restart;
+			}
+			seg = errmalloc(sizeof(*seg));
+			seg->ss_blk = errmalloc(recsize);
+			seg->ss_rec = *rec;
+			bcopy((void *)rec, seg->ss_blk, recsize);
+			if (rec->jsr_oldest > oldseq)
+				oldseq = rec->jsr_oldest;
+			TAILQ_INSERT_TAIL(&allsegs, seg, ss_next);
+			jblocks_advance(suj_jblocks, recsize);
+		}
+	}
+}
+
+/*
+ * Search a directory block for the SUJ_FILE.
+ */
+static void
+suj_find(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	char block[MAXBSIZE];
+	struct direct *dp;
+	int bytes;
+	int off;
+
+	if (sujino)
+		return;
+	bytes = lfragtosize(fs, frags);
+	if (bread(disk, fsbtodb(fs, blk), block, bytes) <= 0)
+		err(1, "Failed to read ROOTINO directory block %jd", blk);
+	for (off = 0; off < bytes; off += dp->d_reclen) {
+		dp = (struct direct *)&block[off];
+		if (dp->d_reclen == 0)
+			break;
+		if (dp->d_ino == 0)
+			continue;
+		if (dp->d_namlen != strlen(SUJ_FILE))
+			continue;
+		if (bcmp(dp->d_name, SUJ_FILE, dp->d_namlen) != 0)
+			continue;
+		sujino = dp->d_ino;
+		return;
+	}
+}
+
+/*
+ * Orchestrate the verification of a filesystem via the softupdates journal.
+ */
+int
+suj_check(const char *filesys)
+{
+	union dinode *jip;
+	union dinode *ip;
+	uint64_t blocks;
+
+	opendisk(filesys);
+	TAILQ_INIT(&allsegs);
+	/*
+	 * Find the journal inode.
+	 */
+	ip = ino_read(ROOTINO);
+	sujino = 0;
+	ino_visit(ip, ROOTINO, suj_find, 0);
+	if (sujino == 0)
+		errx(1, "Journal inode removed.  Use tunefs to re-create.");
+	/*
+	 * Fetch the journal inode and verify it.
+	 */
+	jip = ino_read(sujino);
+	printf("** SU+J Recovering %s\n", filesys);
+	if (suj_verifyino(jip) != 0)
+		return (-1);
+	/*
+	 * Build a list of journal blocks in jblocks before parsing the
+	 * available journal blocks in with suj_read().
+	 */
+	printf("** Reading %jd byte journal from inode %d.\n",
+	    DIP(jip, di_size), sujino);
+	suj_jblocks = jblocks_create();
+	blocks = ino_visit(jip, sujino, suj_add_block, 0);
+	if (blocks != numfrags(fs, DIP(jip, di_size)))
+		errx(1, "Sparse journal inode %d.\n", sujino);
+	suj_read();
+	jblocks_destroy(suj_jblocks);
+	suj_jblocks = NULL;
+	if (preen || reply("RECOVER")) {
+		printf("** Building recovery table.\n");
+		suj_prune();
+		suj_build();
+		cg_apply(cg_build);
+		printf("** Resolving unreferenced inode list.\n");
+		ino_unlinked();
+		printf("** Processing journal entries.\n");
+		cg_apply(cg_trunc);
+		cg_apply(cg_check_blk);
+		cg_apply(cg_check_ino);
+	}
+	if (preen == 0 && reply("WRITE CHANGES") == 0)
+		return (0);
+	/*
+	 * To remain idempotent with partial truncations the free bitmaps
+	 * must be written followed by indirect blocks and lastly inode
+	 * blocks.  This preserves access to the modified pointers until
+	 * they are freed.
+	 */
+	cg_apply(cg_write);
+	dblk_write();
+	cg_apply(cg_write_inos);
+	/* Write back superblock. */
+	closedisk(filesys);
+	printf("** %jd journal records in %jd bytes for %.2f%% utilization\n",
+	    jrecs, jbytes, ((float)jrecs / (float)(jbytes / JREC_SIZE)) * 100);
+	printf("** Freed %jd inodes (%jd dirs) %jd blocks, and %jd frags.\n",
+	    freeinos, freedir, freeblocks, freefrags);
+
+	return (0);
+}
+/*-
+ * Copyright (c) 2009 Jeffrey W. Roberson <jeff@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/disklabel.h>
+#include <sys/mount.h>
+#include <sys/stat.h>
+
+#include <ufs/ufs/ufsmount.h>
+#include <ufs/ufs/dinode.h>
+#include <ufs/ufs/dir.h>
+#include <ufs/ffs/fs.h>
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <stdint.h>
+#include <libufs.h>
+#include <strings.h>
+#include <err.h>
+#include <assert.h>
+
+#include "fsck.h"
+
+static void	ino_decr(ino_t);
+
+#define	SUJ_HASHSIZE	128
+#define	SUJ_HASHMASK	(SUJ_HASHSIZE - 1)
+#define	SUJ_HASH(x)	((x * 2654435761) & SUJ_HASHMASK)
+
+struct suj_seg {
+	TAILQ_ENTRY(suj_seg) ss_next;
+	struct jsegrec	ss_rec;
+	uint8_t		*ss_blk;
+};
+
+struct suj_rec {
+	TAILQ_ENTRY(suj_rec) sr_next;
+	union jrec	*sr_rec;
+};
+TAILQ_HEAD(srechd, suj_rec);
+
+struct suj_ino {
+	LIST_ENTRY(suj_ino)	si_next;
+	struct srechd		si_recs;
+	struct srechd		si_movs;
+	ino_t			si_ino;
+	int			si_nlinkadj;
+	int			si_skipparent;
+	int			si_linkadj;
+	int			si_hasrecs;
+	int			si_blkadj;
+};
+LIST_HEAD(inohd, suj_ino);
+
+struct suj_blk {
+	LIST_ENTRY(suj_blk)	sb_next;
+	struct srechd		sb_recs;
+	ufs2_daddr_t		sb_blk;
+};
+LIST_HEAD(blkhd, suj_blk);
+
+struct data_blk {
+	LIST_ENTRY(data_blk)	db_next;
+	uint8_t			*db_buf;
+	ufs2_daddr_t		db_blk;
+	int			db_size;
+};
+
+struct ino_blk {
+	LIST_ENTRY(ino_blk)	ib_next;
+	uint8_t			*ib_buf;
+	int			ib_dirty;
+	ufs2_daddr_t		ib_blk;
+};
+LIST_HEAD(iblkhd, ino_blk);
+
+struct suj_cg {
+	LIST_ENTRY(suj_cg)	sc_next;
+	struct blkhd		sc_blkhash[SUJ_HASHSIZE];
+	struct inohd		sc_inohash[SUJ_HASHSIZE];
+	struct iblkhd		sc_iblkhash[SUJ_HASHSIZE];
+	struct ino_blk		*sc_lastiblk;
+	uint8_t			*sc_cgbuf;
+	struct cg		*sc_cgp;
+	int			sc_dirty;
+	int			sc_cgx;
+};
+
+LIST_HEAD(cghd, suj_cg) cghash[SUJ_HASHSIZE];
+LIST_HEAD(dblkhd, data_blk) dbhash[SUJ_HASHSIZE];
+
+TAILQ_HEAD(seghd, suj_seg) allsegs;
+uint64_t oldseq;
+static struct uufsd *disk = NULL;
+static struct fs *fs = NULL;
+
+/*
+ * Summary statistics.
+ */
+uint64_t freefrags;
+uint64_t freeblocks;
+uint64_t freeinos;
+uint64_t freedir;
+uint64_t jbytes;
+uint64_t jrecs;
+
+typedef void (*ino_visitor)(ino_t, ufs_lbn_t, ufs2_daddr_t, int);
+
+static void *
+errmalloc(size_t n)
+{
+	void *a;
+
+	a = malloc(n);
+	if (a == NULL)
+		errx(1, "malloc(%zu)", n);
+	return (a);
+}
+
+/*
+ * Open the given provider, load superblock.
+ */
+static void
+opendisk(const char *devnam)
+{
+	if (disk != NULL)
+		return;
+	disk = malloc(sizeof(*disk));
+	if (disk == NULL)
+		errx(1, "malloc(%zu)", sizeof(*disk));
+	if (ufs_disk_fillout(disk, devnam) == -1) {
+		err(1, "ufs_disk_fillout(%s) failed: %s", devnam,
+		    disk->d_error);
+	}
+	fs = &disk->d_fs;
+	/*
+	 * Setup a few things so reply() can work.
+	 */
+	bcopy(fs, &sblock, sizeof(sblock));
+	fsreadfd = disk->d_fd;
+	fswritefd = disk->d_fd;
+}
+
+/*
+ * Mark file system as clean, write the super-block back, close the disk.
+ */
+static void
+closedisk(const char *devnam)
+{
+	struct csum *cgsum;
+	int i;
+
+	/*
+	 * Recompute the fs summary info from correct cs summaries.
+	 */
+	bzero(&fs->fs_cstotal, sizeof(struct csum_total));
+	for (i = 0; i < fs->fs_ncg; i++) {
+		cgsum = &fs->fs_cs(fs, i);
+		fs->fs_cstotal.cs_nffree += cgsum->cs_nffree;
+		fs->fs_cstotal.cs_nbfree += cgsum->cs_nbfree;
+		fs->fs_cstotal.cs_nifree += cgsum->cs_nifree;
+		fs->fs_cstotal.cs_ndir += cgsum->cs_ndir;
+	}
+	/* XXX Don't set clean for now, we don't trust the journal. */
+	/* fs->fs_clean = 1; */
+	fs->fs_time = time(NULL);
+	fs->fs_mtime = time(NULL);
+	if (sbwrite(disk, 0) == -1)
+		err(1, "sbwrite(%s)", devnam);
+	if (ufs_disk_close(disk) == -1)
+		err(1, "ufs_disk_close(%s)", devnam);
+	free(disk);
+	disk = NULL;
+	fs = NULL;
+	fsreadfd = -1;
+	fswritefd = -1;
+}
+
+/*
+ * Lookup a cg by number in the hash so we can keep track of which cgs
+ * need stats rebuilt.
+ */
+static struct suj_cg *
+cg_lookup(int cgx)
+{
+	struct cghd *hd;
+	struct suj_cg *sc;
+
+	if (cgx < 0 || cgx >= fs->fs_ncg) {
+		abort();
+		errx(1, "Bad cg number %d", cgx);
+	}
+	hd = &cghash[SUJ_HASH(cgx)];
+	LIST_FOREACH(sc, hd, sc_next)
+		if (sc->sc_cgx == cgx)
+			return (sc);
+	sc = errmalloc(sizeof(*sc));
+	bzero(sc, sizeof(*sc));
+	sc->sc_cgbuf = errmalloc(fs->fs_bsize);
+	sc->sc_cgp = (struct cg *)sc->sc_cgbuf;
+	sc->sc_cgx = cgx;
+	LIST_INSERT_HEAD(hd, sc, sc_next);
+	if (bread(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
+	    fs->fs_bsize) == -1)
+		err(1, "Unable to read cylinder group %d", sc->sc_cgx);
+
+	return (sc);
+}
+
+/*
+ * Lookup an inode number in the hash and allocate a suj_ino if it does
+ * not exist.
+ */
+static struct suj_ino *
+ino_lookup(ino_t ino, int creat)
+{
+	struct suj_ino *sino;
+	struct inohd *hd;
+	struct suj_cg *sc;
+
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	hd = &sc->sc_inohash[SUJ_HASH(ino)];
+	LIST_FOREACH(sino, hd, si_next)
+		if (sino->si_ino == ino)
+			return (sino);
+	if (creat == 0)
+		return (NULL);
+	sino = errmalloc(sizeof(*sino));
+	bzero(sino, sizeof(*sino));
+	sino->si_ino = ino;
+	sino->si_nlinkadj = 0;
+	TAILQ_INIT(&sino->si_recs);
+	TAILQ_INIT(&sino->si_movs);
+	LIST_INSERT_HEAD(hd, sino, si_next);
+
+	return (sino);
+}
+
+/*
+ * Lookup a block number in the hash and allocate a suj_blk if it does
+ * not exist.
+ */
+static struct suj_blk *
+blk_lookup(ufs2_daddr_t blk, int creat)
+{
+	struct suj_blk *sblk;
+	struct suj_cg *sc;
+	struct blkhd *hd;
+
+	sc = cg_lookup(dtog(fs, blk));
+	hd = &sc->sc_blkhash[SUJ_HASH(blk)];
+	LIST_FOREACH(sblk, hd, sb_next)
+		if (sblk->sb_blk == blk)
+			return (sblk);
+	if (creat == 0)
+		return (NULL);
+	sblk = errmalloc(sizeof(*sblk));
+	bzero(sblk, sizeof(*sblk));
+	sblk->sb_blk = blk;
+	TAILQ_INIT(&sblk->sb_recs);
+	LIST_INSERT_HEAD(hd, sblk, sb_next);
+
+	return (sblk);
+}
+
+static uint8_t *
+dblk_read(ufs2_daddr_t blk, int size)
+{
+	struct data_blk *dblk;
+	struct dblkhd *hd;
+
+	hd = &dbhash[SUJ_HASH(blk)];
+	LIST_FOREACH(dblk, hd, db_next)
+		if (dblk->db_blk == blk)
+			goto found;
+	/*
+	 * The inode block wasn't located, allocate a new one.
+	 */
+	dblk = errmalloc(sizeof(*dblk));
+	bzero(dblk, sizeof(*dblk));
+	LIST_INSERT_HEAD(hd, dblk, db_next);
+	dblk->db_blk = blk;
+found:
+	/*
+	 * I doubt size mismatches can happen in practice but it is trivial
+	 * to handle.
+	 */
+	if (size != dblk->db_size) {
+		if (dblk->db_buf)
+			free(dblk->db_buf);
+		dblk->db_buf = errmalloc(size);
+		dblk->db_size = size;
+		if (bread(disk, fsbtodb(fs, blk), dblk->db_buf, size) == -1)
+			err(1, "Failed to read data block %jd", blk);
+	}
+	return (dblk->db_buf);
+}
+
+static union dinode *
+ino_read(ino_t ino)
+{
+	struct ino_blk *iblk;
+	struct iblkhd *hd;
+	struct suj_cg *sc;
+	ufs2_daddr_t blk;
+	int off;
+
+	blk = ino_to_fsba(fs, ino);
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	hd = &sc->sc_iblkhash[SUJ_HASH(blk)];
+	LIST_FOREACH(iblk, hd, ib_next)
+		if (iblk->ib_blk == blk)
+			goto found;
+	/*
+	 * The inode block wasn't located, allocate a new one.
+	 */
+	iblk = errmalloc(sizeof(*iblk));
+	bzero(iblk, sizeof(*iblk));
+	iblk->ib_buf = errmalloc(fs->fs_bsize);
+	iblk->ib_blk = blk;
+	LIST_INSERT_HEAD(hd, iblk, ib_next);
+	if (bread(disk, fsbtodb(fs, blk), iblk->ib_buf, fs->fs_bsize) == -1)
+		err(1, "Failed to read inode block %jd", blk);
+found:
+	sc->sc_lastiblk = iblk;
+	off = ino_to_fsbo(fs, ino);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		return (union dinode *)&((struct ufs1_dinode *)iblk->ib_buf)[off];
+	else
+		return (union dinode *)&((struct ufs2_dinode *)iblk->ib_buf)[off];
+}
+
+static void
+ino_dirty(ino_t ino)
+{
+	struct ino_blk *iblk;
+	struct iblkhd *hd;
+	struct suj_cg *sc;
+	ufs2_daddr_t blk;
+
+	blk = ino_to_fsba(fs, ino);
+	sc = cg_lookup(ino_to_cg(fs, ino));
+	iblk = sc->sc_lastiblk;
+	if (iblk && iblk->ib_blk == blk) {
+		iblk->ib_dirty = 1;
+		return;
+	}
+	hd = &sc->sc_iblkhash[SUJ_HASH(blk)];
+	LIST_FOREACH(iblk, hd, ib_next) {
+		if (iblk->ib_blk == blk) {
+			iblk->ib_dirty = 1;
+			return;
+		}
+	}
+	ino_read(ino);
+	ino_dirty(ino);
+}
+
+static void
+iblk_write(struct ino_blk *iblk)
+{
+
+	if (iblk->ib_dirty == 0)
+		return;
+	if (bwrite(disk, fsbtodb(fs, iblk->ib_blk), iblk->ib_buf,
+	    fs->fs_bsize) == -1)
+		err(1, "Failed to write inode block %jd", iblk->ib_blk);
+}
+
+/*
+ * Return 1 if the inode was free and 0 if it is allocated.
+ */
+static int
+ino_isfree(ino_t ino)
+{
+	struct suj_cg *sc;
+	uint8_t *inosused;
+	struct cg *cgp;
+	int cg;
+
+	cg = ino_to_cg(fs, ino);
+	ino = ino % fs->fs_ipg;
+	sc = cg_lookup(cg);
+	cgp = sc->sc_cgp;
+	inosused = cg_inosused(cgp);
+	return isclr(inosused, ino);
+}
+
+static int
+blk_overlaps(struct jblkrec *brec, ufs2_daddr_t start, int frags)
+{
+	ufs2_daddr_t bstart;
+	ufs2_daddr_t bend;
+	ufs2_daddr_t end;
+
+	end = start + frags;
+	bstart = brec->jb_blkno + brec->jb_oldfrags;
+	bend = bstart + brec->jb_frags;
+	if (start < bend && end > bstart)
+		return (1);
+	return (0);
+}
+
+static int
+blk_equals(struct jblkrec *brec, ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t start,
+    int frags)
+{
+
+	if (brec->jb_ino != ino || brec->jb_lbn != lbn)
+		return (0);
+	if (brec->jb_blkno + brec->jb_oldfrags != start)
+		return (0);
+	if (brec->jb_frags != frags)
+		return (0);
+	return (1);
+}
+
+static void
+blk_setmask(struct jblkrec *brec, int *mask)
+{
+	int i;
+
+	for (i = brec->jb_oldfrags; i < brec->jb_oldfrags + brec->jb_frags; i++)
+		*mask |= 1 << i;
+}
+
+/*
+ * Determine whether a given block has been reallocated to a new location.
+ * Returns a mask of overlapping bits if any frags have been reused or
+ * zero if the block has not been re-used and the contents can be trusted.
+ * 
+ * This is used to ensure that an orphaned pointer due to truncate is safe
+ * to be freed.  The mask value can be used to free partial blocks.
+ */
+static int
+blk_isfree(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags)
+{
+	struct suj_blk *sblk;
+	struct suj_rec *srec;
+	struct jblkrec *brec;
+	int mask;
+	int off;
+
+	/*
+	 * To be certain we're not freeing a reallocated block we lookup
+	 * this block in the blk hash and see if there is an allocation
+	 * journal record that overlaps with any fragments in the block
+	 * we're concerned with.  If any fragments have ben reallocated
+	 * the block has already been freed and re-used for another purpose.
+	 */
+	mask = 0;
+	sblk = blk_lookup(blknum(fs, blk), 0);
+	if (sblk == NULL)
+		return (0);
+	off = blk - sblk->sb_blk;
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		brec = (struct jblkrec *)srec->sr_rec;
+		/*
+		 * If the block overlaps but does not match
+		 * exactly it's a new allocation.  If it matches
+		 * exactly this record refers to the current
+		 * location.
+		 */ 
+		if (blk_overlaps(brec, blk, frags) == 0)
+			continue;
+		if (blk_equals(brec, ino, lbn, blk, frags) == 1)
+			mask = 0;
+		else
+			blk_setmask(brec, &mask);
+	}
+	if (debug)
+		printf("blk_isfree: blk %jd sblk %jd off %d mask 0x%X\n",
+		    blk, sblk->sb_blk, off, mask);
+	return (mask >> off);
+}
+
+/*
+ * Determine whether it is safe to follow an indirect.  It is not safe
+ * if any part of the indirect has been reallocated or the last journal
+ * entry was an allocation.  Just allocated indirects may not have valid
+ * pointers yet and all of their children will have their own records.
+ * 
+ * Returns 1 if it's safe to follow the indirect and 0 otherwise.
+ */
+static int
+blk_isindir(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn)
+{
+	struct suj_blk *sblk;
+	struct jblkrec *brec;
+
+	sblk = blk_lookup(blk, 0);
+	if (sblk == NULL)
+		return (1);
+	if (TAILQ_EMPTY(&sblk->sb_recs))
+		return (1);
+	brec = (struct jblkrec *)TAILQ_LAST(&sblk->sb_recs, srechd)->sr_rec;
+	if (blk_equals(brec, ino, lbn, blk, fs->fs_frag))
+		if (brec->jb_op == JOP_FREEBLK)
+			return (1);
+	return (0);
+}
+
+/*
+ * Clear an inode from the cg bitmap.  If the inode was already clear return
+ * 0 so the caller knows it does not have to check the inode contents.
+ */
+static int
+ino_free(ino_t ino, int mode)
+{
+	struct suj_cg *sc;
+	uint8_t *inosused;
+	struct cg *cgp;
+	int cg;
+
+	cg = ino_to_cg(fs, ino);
+	ino = ino % fs->fs_ipg;
+	sc = cg_lookup(cg);
+	cgp = sc->sc_cgp;
+	inosused = cg_inosused(cgp);
+	/*
+	 * The bitmap may never have made it to the disk so we have to
+	 * conditionally clear.  We can avoid writing the cg in this case.
+	 */
+	if (isclr(inosused, ino))
+		return (0);
+	freeinos++;
+	clrbit(inosused, ino);
+	if (ino < cgp->cg_irotor)
+		cgp->cg_irotor = ino;
+	cgp->cg_cs.cs_nifree++;
+	if ((mode & IFMT) == IFDIR) {
+		freedir++;
+		cgp->cg_cs.cs_ndir--;
+	}
+	sc->sc_dirty = 1;
+
+	return (1);
+}
+
+/*
+ * Free 'frags' frags starting at filesystem block 'bno' skipping any frags
+ * set in the mask.
+ */
+static void
+blk_free(ufs2_daddr_t bno, int mask, int frags)
+{
+	ufs1_daddr_t fragno, cgbno;
+	struct suj_cg *sc;
+	struct cg *cgp;
+	int i, cg;
+	uint8_t *blksfree;
+
+	if (debug)
+		printf("Freeing %d frags at blk %jd\n", frags, bno);
+	cg = dtog(fs, bno);
+	sc = cg_lookup(cg);
+	cgp = sc->sc_cgp;
+	cgbno = dtogd(fs, bno);
+	blksfree = cg_blksfree(cgp);
+
+	/*
+	 * If it's not allocated we only wrote the journal entry
+	 * and never the bitmaps.  Here we unconditionally clear and
+	 * resolve the cg summary later.
+	 */
+	if (frags == fs->fs_frag && mask == 0) {
+		fragno = fragstoblks(fs, cgbno);
+		ffs_setblock(fs, blksfree, fragno);
+		freeblocks++;
+	} else {
+		/*
+		 * deallocate the fragment
+		 */
+		for (i = 0; i < frags; i++)
+			if ((mask & (1 << i)) == 0 && isclr(blksfree, cgbno +i)) {
+				freefrags++;
+				setbit(blksfree, cgbno + i);
+			}
+	}
+	sc->sc_dirty = 1;
+}
+
+/*
+ * Fetch an indirect block to find the block at a given lbn.  The lbn
+ * may be negative to fetch a specific indirect block pointer or positive
+ * to fetch a specific block.
+ */
+static ufs2_daddr_t
+indir_blkatoff(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t cur, ufs_lbn_t lbn, int level)
+{
+	ufs2_daddr_t *bap2;
+	ufs2_daddr_t *bap1;
+	ufs_lbn_t lbnadd;
+	ufs_lbn_t base;
+	int i;
+
+	if (blk == 0)
+		return (0);
+	if (cur == lbn)
+		return (blk);
+	if (level == 0 && lbn < 0) {
+		abort();
+		errx(1, "Invalid lbn %jd", lbn);
+	}
+	bap2 = (void *)dblk_read(blk, fs->fs_bsize);
+	bap1 = (void *)bap2;
+	lbnadd = 1;
+	base = -(cur + level);
+	for (i = level; i > 0; i--)
+		lbnadd *= NINDIR(fs);
+	if (lbn > 0) 
+		i = (lbn - base) / lbnadd;
+	else
+		i = (-lbn - base) / lbnadd;
+	if (i < 0 || i >= NINDIR(fs)) {
+		abort();
+		errx(1, "Invalid indirect index %d produced by lbn %jd",
+		    i, lbn);
+	}
+	if (level == 0)
+		cur = base + (i * lbnadd);
+	else
+		cur = -(base + (i * lbnadd)) - (level - 1);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		blk = bap1[i];
+	else
+		blk = bap2[i];
+	if (cur == lbn)
+		return (blk);
+	if (level == 0) {
+		abort();
+		errx(1, "Invalid lbn %jd at level 0", lbn);
+	}
+	return indir_blkatoff(blk, ino, cur, lbn, level - 1);
+}
+
+/*
+ * Finds the disk block address at the specified lbn within the inode
+ * specified by ip.  This follows the whole tree and honors di_size and
+ * di_extsize so it is a true test of reachability.  The lbn may be
+ * negative if an extattr or indirect block is requested.
+ */
+static ufs2_daddr_t
+ino_blkatoff(union dinode *ip, ino_t ino, ufs_lbn_t lbn, int *frags)
+{
+	ufs_lbn_t tmpval;
+	ufs_lbn_t cur;
+	ufs_lbn_t next;
+	int i;
+
+	/*
+	 * Handle extattr blocks first.
+	 */
+	if (lbn < 0 && lbn >= -NXADDR) {
+		lbn = -1 - lbn;
+		if (lbn > lblkno(fs, ip->dp2.di_extsize - 1))
+			return (0);
+		*frags = numfrags(fs, sblksize(fs, ip->dp2.di_extsize, lbn));
+		return (ip->dp2.di_extb[lbn]);
+	}
+	/*
+	 * And now direct and indirect.  Verify that the lbn does not
+	 * exceed the size required to store the file by asking for
+	 * the lbn of the last byte.  These blocks should be 0 anyway
+	 * so this simply saves the traversal.
+	 */
+	if (lbn > 0 && lbn > lblkno(fs, DIP(ip, di_size) - 1))
+		return (0);
+	if (lbn < 0 && -lbn > lblkno(fs, DIP(ip, di_size) - 1))
+		return (0);
+	if (lbn >= 0 && lbn < NDADDR) {
+		*frags = numfrags(fs, sblksize(fs, DIP(ip, di_size), lbn));
+		return (DIP(ip, di_db[lbn]));
+	}
+	*frags = fs->fs_frag;
+
+	for (i = 0, tmpval = NINDIR(fs), cur = NDADDR; i < NIADDR; i++,
+	    tmpval *= NINDIR(fs), cur = next) {
+		next = cur + tmpval;
+		if (lbn == -cur)
+			return (DIP(ip, di_ib[i]));
+		/*
+		 * Determine whether the lbn in question is within this tree.
+		 */
+		if (lbn < 0 && -lbn >= next)
+			continue;
+		if (lbn > 0 && lbn >= next)
+			continue;
+
+		return indir_blkatoff(DIP(ip, di_ib[i]), ino, -cur - i, lbn, i);
+	}
+	errx(1, "lbn %jd not in ino", lbn);
+}
+
+/*
+ * Determine whether a block exists at a particular lbn in an inode.
+ * Returns 1 if found, 0 if not.  lbn may be negative for indirects
+ * or ext blocks.
+ */
+static int
+blk_isat(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int *frags)
+{
+	union dinode *ip;
+	ufs2_daddr_t nblk;
+
+	ip = ino_read(ino);
+
+	if (DIP(ip, di_nlink) == 0 || DIP(ip, di_mode) == 0)
+		return (0);
+	nblk = ino_blkatoff(ip, ino, lbn, frags);
+
+	return (nblk == blk);
+}
+
+/*
+ * Determines whether a pointer to an inode exists within a directory
+ * at a specified offset.  Returns the mode of the found entry.
+ */
+static int
+ino_isat(ino_t parent, off_t diroff, ino_t child, int *mode, int *isdot)
+{
+	union dinode *dip;
+	struct direct *dp;
+	ufs2_daddr_t blk;
+	uint8_t *block;
+	ufs_lbn_t lbn;
+	int blksize;
+	int frags;
+	int dpoff;
+	int doff;
+
+	*isdot = 0;
+	dip = ino_read(parent);
+	*mode = DIP(dip, di_mode);
+	if ((*mode & IFMT) != IFDIR) {
+		if (debug) {
+			/* This can happen if the parent inode was reallocated. */
+			if (*mode != 0)
+				printf("Directory %d has bad mode %o\n",
+				    parent, *mode);
+			else
+				printf("Directory %d zero inode\n", parent);
+		}
+		return (0);
+	}
+	lbn = lblkno(fs, diroff);
+	doff = blkoff(fs, diroff);
+	blksize = sblksize(fs, DIP(dip, di_size), lbn);
+	if (diroff + DIRECTSIZ(1) > DIP(dip, di_size) || doff >= blksize) {
+		if (debug)
+			printf("ino %d absent from %d due to offset %jd"
+			    " exceeding size %jd\n",
+			    child, parent, diroff, DIP(dip, di_size));
+		return (0);
+	}
+	blk = ino_blkatoff(dip, parent, lbn, &frags);
+	if (blk <= 0) {
+		if (debug)
+			printf("Sparse directory %d", parent);
+		return (0);
+	}
+	block = dblk_read(blk, blksize);
+	/*
+	 * Walk through the records from the start of the block to be
+	 * certain we hit a valid record and not some junk in the middle
+	 * of a file name.  Stop when we reach or pass the expected offset.
+	 */
+	dpoff = 0;
+	do {
+		dp = (struct direct *)&block[dpoff];
+		if (dpoff == doff)
+			break;
+		if (dp->d_reclen == 0)
+			break;
+		dpoff += dp->d_reclen;
+	} while (dpoff <= doff);
+	if (dpoff > fs->fs_bsize)
+		errx(1, "Corrupt directory block in dir inode %d", parent);
+	/* Not found. */
+	if (dpoff != doff) {
+		if (debug)
+			printf("ino %d not found in %d, lbn %jd, dpoff %d\n",
+			    child, parent, lbn, dpoff);
+		return (0);
+	}
+	/*
+	 * We found the item in question.  Record the mode and whether it's
+	 * a . or .. link for the caller.
+	 */
+	if (dp->d_ino == child) {
+		if (child == parent)
+			*isdot = 1;
+		else if (dp->d_namlen == 2 &&
+		    dp->d_name[0] == '.' && dp->d_name[1] == '.')
+			*isdot = 1;
+		*mode = DTTOIF(dp->d_type);
+		return (1);
+	}
+	if (debug)
+		printf("ino %d doesn't match dirent ino %d in parent %d\n",
+		    child, dp->d_ino, parent);
+	return (0);
+}
+
+#define	VISIT_INDIR	0x0001
+#define	VISIT_EXT	0x0002
+
+/*
+ * Read an indirect level which may or may not be linked into an inode.
+ */
+static void
+indir_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, uint64_t *frags,
+    ino_visitor visitor, int flags)
+{
+	ufs2_daddr_t *bap2;
+	ufs1_daddr_t *bap1;
+	ufs_lbn_t lbnadd;
+	ufs2_daddr_t nblk;
+	ufs_lbn_t nlbn;
+	int level;
+	int i;
+
+	/*
+	 * Don't visit indirect blocks with contents we can't trust.  This
+	 * should only happen when indir_visit() is called to complete a
+	 * truncate that never finished and not when a pointer is found via
+	 * an inode.
+	 */
+	if (blk == 0)
+		return;
+	if (blk_isindir(blk, ino, lbn) == 0) {
+		if (debug)
+			printf("blk %jd ino %d lbn %jd is not indir.\n",
+			    blk, ino, lbn);
+		goto out;
+	}
+	level = lbn_level(lbn);
+	if (level == -1) {
+		abort();
+		errx(1, "Invalid level for lbn %jd", lbn);
+	}
+	lbnadd = 1;
+	for (i = level; i > 0; i--)
+		lbnadd *= NINDIR(fs);
+	bap1 = (void *)dblk_read(blk, fs->fs_bsize);
+	bap2 = (void *)bap1;
+	for (i = 0; i < NINDIR(fs); i++) {
+		if (fs->fs_magic == FS_UFS1_MAGIC)
+			nblk = *bap1++;
+		else
+			nblk = *bap2++;
+		if (nblk == 0)
+			continue;
+		if (level == 0) {
+			nlbn = -lbn + i * lbnadd;
+			(*frags) += fs->fs_frag;
+			visitor(ino, nlbn, nblk, fs->fs_frag);
+		} else {
+			nlbn = (lbn + 1) - (i * lbnadd);
+			indir_visit(ino, nlbn, nblk, frags, visitor, flags);
+		}
+	}
+out:
+	if (flags & VISIT_INDIR) {
+		(*frags) += fs->fs_frag;
+		visitor(ino, lbn, blk, fs->fs_frag);
+	}
+}
+
+/*
+ * Visit each block in an inode as specified by 'flags' and call a
+ * callback function.  The callback may inspect or free blocks.  The
+ * count of frags found according to the size in the file is returned.
+ * This is not valid for sparse files but may be used to determine
+ * the correct di_blocks for a file.
+ */
+static uint64_t
+ino_visit(union dinode *ip, ino_t ino, ino_visitor visitor, int flags)
+{
+	ufs_lbn_t tmpval;
+	ufs_lbn_t lbn;
+	uint64_t size;
+	uint64_t fragcnt;
+	int mode;
+	int frags;
+	int i;
+
+	size = DIP(ip, di_size);
+	mode = DIP(ip, di_mode) & IFMT;
+	fragcnt = 0;
+	if ((flags & VISIT_EXT) &&
+	    fs->fs_magic == FS_UFS2_MAGIC && ip->dp2.di_extsize) {
+		for (i = 0; i < NXADDR; i++) {
+			if (ip->dp2.di_extb[i] == 0)
+				continue;
+			frags = sblksize(fs, ip->dp2.di_extsize, i);
+			frags = numfrags(fs, frags);
+			fragcnt += frags;
+			visitor(ino, -1 - i, ip->dp2.di_extb[i], frags);
+		}
+	}
+	/* Skip datablocks for short links and devices. */
+	if (mode == IFBLK || mode == IFCHR ||
+	    (mode == IFLNK && size < fs->fs_maxsymlinklen))
+		return (fragcnt);
+	for (i = 0; i < NDADDR; i++) {
+		if (DIP(ip, di_db[i]) == 0)
+			continue;
+		frags = sblksize(fs, size, i);
+		frags = numfrags(fs, frags);
+		fragcnt += frags;
+		visitor(ino, i, DIP(ip, di_db[i]), frags);
+	}
+	for (i = 0, tmpval = NINDIR(fs), lbn = NDADDR; i < NIADDR; i++,
+	    tmpval *= NINDIR(fs), lbn += tmpval) {
+		if (DIP(ip, di_ib[i]) == 0)
+			continue;
+		indir_visit(ino, -lbn - i, DIP(ip, di_ib[i]), &fragcnt, visitor,
+		    flags);
+	}
+	return (fragcnt);
+}
+
+/*
+ * Null visitor function used when we just want to count blocks.
+ */
+static void
+null_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+}
+
+/*
+ * Recalculate di_blocks when we discover that a block allocation or
+ * free was not successfully completed.  The kernel does not roll this back
+ * because it would be too expensive to compute which indirects were
+ * reachable at the time the inode was written.
+ */
+static void
+ino_adjblks(ino_t ino)
+{
+	struct suj_ino *sino;
+	union dinode *ip;
+	uint64_t blocks;
+	uint64_t frags;
+
+	sino = ino_lookup(ino, 1);
+	if (sino->si_blkadj)
+		return;
+	sino->si_blkadj = 1;
+	ip = ino_read(ino);
+	/* No need to adjust zero'd inodes. */
+	if (DIP(ip, di_mode) == 0)
+		return;
+	frags = ino_visit(ip, ino, null_visit, VISIT_INDIR | VISIT_EXT);
+	blocks = fsbtodb(fs, frags);
+	if (blocks == DIP(ip, di_blocks))
+		return;
+	if (debug)
+		printf("ino %d adjusting block count from %jd to %jd\n",
+		    ino, DIP(ip, di_blocks), blocks);
+	DIP_SET(ip, di_blocks, blocks);
+	ino_dirty(ino);
+}
+
+static void
+blk_free_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	int mask;
+
+	mask = blk_isfree(blk, ino, lbn, frags);
+	if (debug)
+		printf("blk %jd freemask 0x%X\n", blk, mask);
+	blk_free(blk, mask, frags);
+}
+
+/*
+ * Free a block or tree of blocks that was previously rooted in ino at
+ * the given lbn.  If the lbn is an indirect all children are freed
+ * recursively.
+ */
+static void
+blk_free_lbn(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags, int follow)
+{
+	uint64_t resid;
+	int mask;
+
+	mask = blk_isfree(blk, ino, lbn, frags);
+	if (debug)
+		printf("blk %jd freemask 0x%X\n", blk, mask);
+	resid = 0;
+	if (lbn <= -NDADDR && follow && mask == 0)
+		indir_visit(ino, lbn, blk, &resid, blk_free_visit, VISIT_INDIR);
+	else
+		blk_free(blk, mask, frags);
+}
+
+static void
+ino_free_children(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+	struct suj_ino *sino;
+	struct suj_rec *srec;
+	struct jrefrec *rrec;
+	struct direct *dp;
+	off_t diroff;
+	uint8_t *block;
+	int skipparent;
+	int isparent;
+	int dpoff;
+	int size;
+
+	sino = ino_lookup(ino, 0);
+	if (sino)
+		skipparent = sino->si_skipparent;
+	else
+		skipparent = 0;
+	size = lfragtosize(fs, frags);
+	block = dblk_read(blk, size);
+	dp = (struct direct *)&block[0];
+	for (dpoff = 0; dpoff < size && dp->d_reclen; dpoff += dp->d_reclen) {
+		dp = (struct direct *)&block[dpoff];
+		if (dp->d_ino == 0 || dp->d_ino == WINO)
+			continue;
+		if (dp->d_namlen == 1 && dp->d_name[0] == '.')
+			continue;
+		isparent = dp->d_namlen == 2 && dp->d_name[0] == '.' &&
+		    dp->d_name[1] == '.';
+		if (isparent && skipparent == 1)
+			continue;
+		if (debug)
+			printf("Directory %d removing inode %d name %s\n",
+			    ino, dp->d_ino, dp->d_name);
+		/*
+		 * Lookup this inode to see if we have a record for it.
+		 * If not, we've already adjusted it assuming this path
+		 * was valid and we have to adjust once more.
+		 */
+		sino = ino_lookup(dp->d_ino, 0);
+		if (sino == NULL || sino->si_linkadj || sino->si_hasrecs == 0) {
+			ino_decr(dp->d_ino);
+			continue;
+		}
+		/*
+		 * Tell any child directories we've already removed their
+		 * parent.  Don't try to adjust our link down again.
+		 */
+		if (isparent == 0)
+			sino->si_skipparent = 1;
+		/*
+		 * If we haven't yet processed this inode we need to make
+		 * sure we will successfully discover the lost path.  If not
+		 * use nlinkadj to remember.
+		 */
+		diroff = lblktosize(fs, lbn) + dpoff;
+		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+			rrec = (struct jrefrec *)srec->sr_rec;
+			if (rrec->jr_parent == ino &&
+			    rrec->jr_diroff == diroff)
+				break;
+		}
+		if (srec == NULL)
+			sino->si_nlinkadj--;
+	}
+}
+
+/*
+ * Truncate an inode, freeing all blocks and decrementing all children's
+ * link counts.  Free the inode back to the cg.
+ */
+static void
+ino_truncate(union dinode *ip, ino_t ino, int mode)
+{
+	uint32_t gen;
+
+	if (ino == ROOTINO)
+		errx(1, "Attempting to free ROOTINO");
+	if (debug)
+		printf("Truncating and freeing ino %d, nlink %d, mode %o\n",
+		    ino, DIP(ip, di_nlink), DIP(ip, di_mode));
+
+	/* We are freeing an inode or directory. */
+	if ((DIP(ip, di_mode) & IFMT) == IFDIR)
+		ino_visit(ip, ino, ino_free_children, 0);
+	DIP_SET(ip, di_nlink, 0);
+	ino_visit(ip, ino, blk_free_visit, VISIT_EXT | VISIT_INDIR);
+	/* Here we have to clear the inode and release any blocks it holds. */
+	gen = DIP(ip, di_gen);
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		bzero(ip, sizeof(struct ufs1_dinode));
+	else
+		bzero(ip, sizeof(struct ufs2_dinode));
+	DIP_SET(ip, di_gen, gen);
+	ino_dirty(ino);
+	ino_free(ino, mode);
+	return;
+}
+
+/*
+ * Adjust an inode's link count down by one when a directory goes away.
+ */
+static void
+ino_decr(ino_t ino)
+{
+	union dinode *ip;
+	int reqlink;
+	int nlink;
+	int mode;
+
+	ip = ino_read(ino);
+	nlink = DIP(ip, di_nlink);
+	mode = DIP(ip, di_mode);
+	if (nlink < 1)
+		errx(1, "Inode %d link count %d invalid", ino, nlink);
+	if (mode == 0)
+		errx(1, "Inode %d has a link of %d with 0 mode.", ino, nlink);
+	nlink--;
+	if ((mode & IFMT) == IFDIR)
+		reqlink = 2;
+	else
+		reqlink = 1;
+	if (nlink < reqlink) {
+		if (debug)
+			printf("ino %d not enough links to live %d < %d\n",
+			    ino, nlink, reqlink);
+		ino_truncate(ip, ino, mode);
+		return;
+	}
+	DIP_SET(ip, di_nlink, nlink);
+	ino_dirty(ino);
+}
+
+/*
+ * Adjust the inode link count to 'nlink'.  If the count reaches zero
+ * free it.
+ */
+static void
+ino_adjust(ino_t ino, int lastmode, nlink_t nlink)
+{
+	union dinode *ip;
+	int reqlink;
+	int mode;
+
+	ip = ino_read(ino);
+	mode = DIP(ip, di_mode) & IFMT;
+	if (nlink > LINK_MAX)
+		errx(1,
+		    "ino %d nlink manipulation error, new link %d, old link %d",
+		    ino, nlink, DIP(ip, di_nlink));
+	if (debug)
+		printf("Adjusting ino %d, nlink %d, old link %d lastmode %o\n",
+		    ino, nlink, DIP(ip, di_nlink), lastmode);
+	if (mode == 0) {
+		if (debug)
+			printf("ino %d, zero inode freeing bitmap\n", ino);
+		ino_free(ino, lastmode);
+		return;
+	}
+	/* XXX Should be an assert? */
+	if (mode != lastmode && debug)
+		printf("ino %d, mode %o != %o\n", ino, mode, lastmode);
+	if ((mode & IFMT) == IFDIR)
+		reqlink = 2;
+	else
+		reqlink = 1;
+	/* If the inode doesn't have enough links to live, free it. */
+	if (nlink < reqlink) {
+		if (debug)
+			printf("ino %d not enough links to live %d < %d\n",
+			    ino, nlink, reqlink);
+		ino_truncate(ip, ino, mode);
+		return;
+	}
+	/* If required write the updated link count. */
+	if (DIP(ip, di_nlink) == nlink) {
+		if (debug)
+			printf("ino %d, link matches, skipping.\n", ino);
+		return;
+	}
+	DIP_SET(ip, di_nlink, nlink);
+	ino_dirty(ino);
+}
+
+#define	DOTDOT_OFFSET	DIRECTSIZ(1)
+
+/*
+ * Process records available for one inode and determine whether the
+ * link count is correct or needs adjusting.
+ *
+ * XXX Failed to fix zero length directory.  Shouldn't .. have been mising?
+ */
+static void
+ino_check(struct suj_ino *sino)
+{
+	struct suj_rec *srec;
+	struct jrefrec *rrec;
+	struct suj_ino *stmp;
+	nlink_t dotlinks;
+	int newlinks;
+	int removes;
+	int nlink;
+	ino_t ino;
+	int isdot;
+	int isat;
+	int mode;
+
+	if (sino->si_hasrecs == 0)
+		return;
+	ino = sino->si_ino;
+	/*
+	 * XXX ino_isfree currently is skipping initialized inodes
+	 * that are unreferenced.
+	 */
+	if (0 && ino_isfree(ino))
+		return;
+	rrec = (struct jrefrec *)TAILQ_FIRST(&sino->si_recs)->sr_rec;
+	nlink = rrec->jr_nlink;
+	newlinks = sino->si_nlinkadj;
+	dotlinks = 0;
+	removes = 0;
+	TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+		rrec = (struct jrefrec *)srec->sr_rec;
+		isat = ino_isat(rrec->jr_parent, rrec->jr_diroff, 
+		    rrec->jr_ino, &mode, &isdot);
+		if (isat && (mode & IFMT) != (rrec->jr_mode & IFMT))
+			errx(1, "Inode mode/directory type mismatch %o != %o",
+			    mode, rrec->jr_mode);
+		if (debug)
+			printf("jrefrec: op %d ino %d, nlink %d, parent %d, "
+			    "diroff %jd, mode %o, isat %d, isdot %d\n",
+			    rrec->jr_op, rrec->jr_ino, rrec->jr_nlink,
+			    rrec->jr_parent, rrec->jr_diroff, rrec->jr_mode,
+			    isat, isdot);
+		mode = rrec->jr_mode & IFMT;
+		if (rrec->jr_op == JOP_REMREF)
+			removes++;
+		newlinks += isat;
+		if (isdot)
+			dotlinks += isat;
+	}
+	/*
+	 * The number of links that remain are the starting link count
+	 * subtracted by the total number of removes with the total
+	 * links discovered back in.  An incomplete remove thus
+	 * makes no change to the link count but an add increases
+	 * by one.
+	 */
+	nlink += newlinks;
+	nlink -= removes;
+	/*
+	 * If it's a directory with no real names pointing to it go ahead
+	 * and truncate it.  This will free any children.
+	 */
+	if ((mode & IFMT) == IFDIR && nlink - dotlinks == 0) {
+		nlink = 0;
+		/*
+		 * Mark any .. links so they know not to free this inode
+		 * when they are removed.
+		 */
+		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
+			rrec = (struct jrefrec *)srec->sr_rec;
+			if (rrec->jr_diroff == DOTDOT_OFFSET) {
+				stmp = ino_lookup(rrec->jr_parent, 0);
+				if (stmp)
+					stmp->si_skipparent = 1;
+			}
+		}
+	}
+	sino->si_linkadj = 1;
+	ino_adjust(ino, mode, nlink);
+}
+
+/*
+ * Process records available for one block and determine whether it is
+ * still allocated and whether the owning inode needs to be updated or
+ * a free completed.
+ */
+static void
+blk_check(struct suj_blk *sblk)
+{
+	struct suj_rec *srec;
+	struct jblkrec *brec;
+	ufs2_daddr_t blk;
+	int mask;
+	int frags;
+	int isat;
+
+	/*
+	 * Each suj_blk actually contains records for any fragments in that
+	 * block.  As a result we must evaluate each record individually.
+	 */
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		brec = (struct jblkrec *)srec->sr_rec;
+		frags = brec->jb_frags;
+		blk = brec->jb_blkno + brec->jb_oldfrags;
+		isat = blk_isat(brec->jb_ino, brec->jb_lbn, blk, &frags);
+		if (debug)
+			printf("op %d blk %jd ino %d lbn %jd frags %d isat %d (%d)\n",
+			    brec->jb_op, blk, brec->jb_ino, brec->jb_lbn,
+			    brec->jb_frags, isat, frags);
+		/*
+		 * If we found the block at this address we still have to
+		 * determine if we need to free the tail end that was
+		 * added by adding contiguous fragments from the same block.
+		 */
+		if (isat == 1) {
+			if (frags == brec->jb_frags)
+				continue;
+			mask = blk_isfree(blk, brec->jb_ino, brec->jb_lbn,
+			    brec->jb_frags);
+			mask >>= frags;
+			blk += frags;
+			frags = brec->jb_frags - frags;
+			blk_free(blk, mask, frags);
+			ino_adjblks(brec->jb_ino);
+			continue;
+		}
+		/*
+	 	 * The block wasn't found, attempt to free it.  It won't be
+		 * freed if it was actually reallocated.  If this was an
+		 * allocation we don't want to follow indirects as they
+		 * may not be written yet.  Any children of the indirect will
+		 * have their own records.  If it's a free we need to
+		 * recursively free children.
+		 */
+		blk_free_lbn(blk, brec->jb_ino, brec->jb_lbn, brec->jb_frags,
+		    brec->jb_op == JOP_FREEBLK);
+		ino_adjblks(brec->jb_ino);
+	}
+}
+
+/*
+ * Walk the list of inode and block records for this cg, recovering any
+ * changes which were not complete at the time of crash.
+ */
+static void
+cg_check(struct suj_cg *sc)
+{
+	struct suj_blk *nextb;
+	struct suj_ino *nexti;
+	struct suj_ino *sino;
+	struct suj_blk *sblk;
+	int i;
+
+	if (debug)
+		printf("Recovering cg %d\n", sc->sc_cgx);
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH_SAFE(sino, &sc->sc_inohash[i], si_next, nexti)
+			ino_check(sino);
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH_SAFE(sblk, &sc->sc_blkhash[i], sb_next, nextb)
+			blk_check(sblk);
+}
+
+/*
+ * Write a potentially dirty cg.  All inodes must be written before the
+ * cg maps are so that an allocated inode is never marked free, even if
+ * we crash during fsck.
+ */
+static void
+cg_write(struct suj_cg *sc)
+{
+	struct ino_blk *iblk;
+	ufs1_daddr_t fragno, cgbno, maxbno;
+	u_int8_t *blksfree;
+	struct cg *cgp;
+	int blk;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(iblk, &sc->sc_iblkhash[i], ib_next)
+			iblk_write(iblk);
+	if (sc->sc_dirty == 0)
+		return;
+	/*
+	 * Fix the frag and cluster summary.
+	 */
+	cgp = sc->sc_cgp;
+	cgp->cg_cs.cs_nbfree = 0;
+	cgp->cg_cs.cs_nffree = 0;
+	bzero(&cgp->cg_frsum, sizeof(cgp->cg_frsum));
+	maxbno = fragstoblks(fs, fs->fs_fpg);
+	if (fs->fs_contigsumsize > 0) {
+		for (i = 1; i <= fs->fs_contigsumsize; i++)
+			cg_clustersum(cgp)[i] = 0;
+		bzero(cg_clustersfree(cgp), howmany(maxbno, CHAR_BIT));
+	}
+	blksfree = cg_blksfree(cgp);
+	for (cgbno = 0; cgbno < maxbno; cgbno++) {
+		if (ffs_isfreeblock(fs, blksfree, cgbno))
+			continue;
+		if (ffs_isblock(fs, blksfree, cgbno)) {
+			ffs_clusteracct(fs, cgp, cgbno, 1);
+			cgp->cg_cs.cs_nbfree++;
+			continue;
+		}
+		fragno = blkstofrags(fs, cgbno);
+		blk = blkmap(fs, blksfree, fragno);
+		ffs_fragacct(fs, blk, cgp->cg_frsum, 1);
+		for (i = 0; i < fs->fs_frag; i++)
+			if (isset(blksfree, fragno + i))
+				cgp->cg_cs.cs_nffree++;
+	}
+	/*
+	 * Update the superblock cg summary from our now correct values
+	 * before writing the block.
+	 */
+	fs->fs_cs(fs, sc->sc_cgx) = cgp->cg_cs;
+	if (bwrite(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
+	    fs->fs_bsize) == -1)
+		err(1, "Unable to write cylinder group %d", sc->sc_cgx);
+}
+
+static void
+cg_apply(void (*apply)(struct suj_cg *))
+{
+	struct suj_cg *scg;
+	int i;
+
+	for (i = 0; i < SUJ_HASHSIZE; i++)
+		LIST_FOREACH(scg, &cghash[i], sc_next)
+			apply(scg);
+}
+
+/*
+ * Process the unlinked but referenced file list.  Freeing all inodes.
+ */
+static void
+ino_unlinked(void)
+{
+	union dinode *ip;
+	uint16_t mode;
+	ino_t inon;
+	ino_t ino;
+
+	ino = fs->fs_sujfree;
+	fs->fs_sujfree = 0;
+	while (ino != 0) {
+		ip = ino_read(ino);
+		mode = DIP(ip, di_mode) & IFMT;
+		inon = DIP(ip, di_freelink);
+		DIP_SET(ip, di_freelink, 0);
+		/*
+		 * XXX Should this be an errx?
+		 */
+		if (DIP(ip, di_nlink) == 0) {
+			if (debug)
+				printf("Freeing unlinked ino %d mode %o\n",
+				    ino, mode);
+			ino_truncate(ip, ino, mode);
+		} else if (debug)
+			printf("Skipping ino %d mode %o with link %d\n",
+			    ino, mode, DIP(ip, di_nlink));
+		ino = inon;
+	}
+}
+
+/*
+ * If we see two ops for the same inode to the same parent at the same
+ * offset we could miscount the link with ino_isat() returning twice.
+ * Keep only the first record because it has the valid link count but keep
+ * the mode from the final op as that should be the correct mode in case
+ * it changed.
+ */
+static void
+suj_build_ino(struct jrefrec *refrec)
+{
+	struct jmvrec *mvrec;
+	struct suj_rec *srec;
+	struct suj_ino *sino;
+	struct suj_rec *srn;
+	struct jrefrec *rrn;
+
+	if (debug)
+		printf("suj_build_ino: op %d, ino %d, nlink %d, parent %d, diroff %jd\n", 
+		    refrec->jr_op, refrec->jr_ino, refrec->jr_nlink, refrec->jr_parent,
+		    refrec->jr_diroff);
+	sino = ino_lookup(refrec->jr_ino, 1);
+	/*
+	 * Search for a mvrec that matches this offset.  Whether it's an add
+	 * or a remove we can delete the mvref.  It no longer applies to this
+	 * location.
+	 *
+	 * For removes, we have to find the original offset so we can create
+	 * a remove that matches the earlier add so it can be abandoned
+	 * if necessary.  We create an add in the new location so we can
+	 * tolerate the directory block as it existed before or after
+	 * the move.
+	 */
+	if (!TAILQ_EMPTY(&sino->si_movs)) {
+		for (srn = TAILQ_LAST(&sino->si_movs, srechd); srn;
+		    srn = TAILQ_PREV(srn, srechd, sr_next)) {
+			mvrec = (struct jmvrec *)srn->sr_rec;
+			if (mvrec->jm_parent != refrec->jr_parent ||
+			    mvrec->jm_newoff != refrec->jr_diroff)
+				continue;
+			TAILQ_REMOVE(&sino->si_movs, srn, sr_next);
+			if (refrec->jr_op == JOP_REMREF) {
+				rrn = errmalloc(sizeof(*refrec));
+				*rrn = *refrec;
+				rrn->jr_op = JOP_ADDREF;
+				suj_build_ino(rrn);
+				refrec->jr_diroff = mvrec->jm_oldoff;
+			}
+		}
+	}
+	/*
+	 * We walk backwards so that adds and removes are evaluated in the
+	 * correct order.
+	 */
+	for (srn = TAILQ_LAST(&sino->si_recs, srechd); srn;
+	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
+		rrn = (struct jrefrec *)srn->sr_rec;
+		if (rrn->jr_parent != refrec->jr_parent ||
+		    rrn->jr_diroff != refrec->jr_diroff)
+			continue;
+		if (debug)
+			printf("Discarding dup.\n");
+		rrn->jr_mode = refrec->jr_mode;
+		return;
+	}
+	sino->si_hasrecs = 1;
+	srec = errmalloc(sizeof(*srec));
+	srec->sr_rec = (union jrec *)refrec;
+	TAILQ_INSERT_TAIL(&sino->si_recs, srec, sr_next);
+}
+
+/*
+ * Apply a move record to an inode.  We must search for adds that preceed us
+ * and add duplicates because we won't know which location to search first.
+ * Then we add movs to a queue that is maintained until the moved location
+ * is removed.  If a single record is moved multiple times we only maintain
+ * one copy that contains the original and final diroffs.
+ */
+static void
+suj_move_ino(struct jmvrec *mvrec)
+{
+	struct jrefrec *refrec;
+	struct suj_ino *sino;
+	struct suj_rec *srec;
+	struct jmvrec *mvrn;
+	struct suj_rec *srn;
+	struct jrefrec *rrn;
+
+	if (debug)
+		printf("suj_move_ino: ino %d, parent %d, diroff %jd, oldoff %jd\n", 
+		    mvrec->jm_ino, mvrec->jm_parent, mvrec->jm_newoff,
+		    mvrec->jm_oldoff);
+	sino = ino_lookup(mvrec->jm_ino, 0);
+	if (sino == NULL)
+		return;
+	/*
+	 * We walk backwards so we only evaluate the most recent record at
+	 * this offset.
+	 */
+	for (srn = TAILQ_LAST(&sino->si_recs, srechd); srn;
+	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
+		rrn = (struct jrefrec *)srn->sr_rec;
+		if (rrn->jr_op != JOP_ADDREF)
+			continue;
+		if (rrn->jr_parent != mvrec->jm_parent ||
+		    rrn->jr_diroff != mvrec->jm_oldoff)
+			continue;
+		/*
+		 * When an entry is moved we don't know whether the write
+		 * to move has completed yet.  To resolve this we create
+		 * a new add dependency in the new location as if it were added
+		 * twice.  Only one will succeed.
+		 */
+		refrec = errmalloc(sizeof(*refrec));
+		refrec->jr_op = JOP_ADDREF;
+		refrec->jr_ino = mvrec->jm_ino;
+		refrec->jr_parent = mvrec->jm_parent;
+		refrec->jr_diroff = mvrec->jm_newoff;
+		refrec->jr_mode = rrn->jr_mode;
+		refrec->jr_nlink = rrn->jr_nlink;
+		suj_build_ino(refrec);
+		break;
+	}
+	/*
+	 * Add this mvrec to the queue of pending mvs.
+	 */
+	for (srn = TAILQ_LAST(&sino->si_movs, srechd); srn;
+	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
+		mvrn = (struct jmvrec *)srn->sr_rec;
+		if (mvrn->jm_parent != mvrec->jm_parent ||
+		    mvrn->jm_newoff != mvrec->jm_oldoff)
+			continue;
+		mvrn->jm_newoff = mvrec->jm_newoff;
+		return;
+	}
+	srec = errmalloc(sizeof(*srec));
+	srec->sr_rec = (union jrec *)mvrec;
+	TAILQ_INSERT_TAIL(&sino->si_movs, srec, sr_next);
+}
+
+/*
+ * Modify journal records so they refer to the base block number
+ * and a start and end frag range.  This is to facilitate the discovery
+ * of overlapping fragment allocations.
+ */
+static void
+suj_build_blk(struct jblkrec *blkrec)
+{
+	struct suj_rec *srec;
+	struct suj_blk *sblk;
+	struct jblkrec *blkrn;
+	ufs2_daddr_t blk;
+	int frag;
+
+	if (debug)
+		printf("suj_build_blk: op %d blkno %jd frags %d oldfrags %d "
+		    "ino %d lbn %jd\n",
+		    blkrec->jb_op, blkrec->jb_blkno, blkrec->jb_frags,
+		    blkrec->jb_oldfrags, blkrec->jb_ino, blkrec->jb_lbn);
+	blk = blknum(fs, blkrec->jb_blkno);
+	frag = fragnum(fs, blkrec->jb_blkno);
+	sblk = blk_lookup(blk, 1);
+	/*
+	 * Rewrite the record using oldfrags to indicate the offset into
+	 * the block.  Leave jb_frags as the actual allocated count.
+	 */
+	blkrec->jb_blkno -= frag;
+	blkrec->jb_oldfrags = frag;
+	if (blkrec->jb_oldfrags + blkrec->jb_frags > fs->fs_frag)
+		errx(1, "Invalid fragment count %d oldfrags %d",
+		    blkrec->jb_frags, frag);
+	/*
+	 * Detect dups.  If we detect a dup we always discard the oldest
+	 * record as it is superseded by the new record.  This speeds up
+	 * later stages but also eliminates free records which are used
+	 * to indicate that the contents of indirects can be trusted.
+	 */
+	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
+		blkrn = (struct jblkrec *)srec->sr_rec;
+		if (blkrn->jb_ino != blkrec->jb_ino ||
+		    blkrn->jb_lbn != blkrec->jb_lbn ||
+		    blkrn->jb_blkno != blkrec->jb_blkno ||
+		    blkrn->jb_frags != blkrec->jb_frags ||
+		    blkrn->jb_oldfrags != blkrec->jb_oldfrags)
+			continue;
+		if (debug)
+			printf("Removed dup.\n");
+		/* Discard the free which is a dup with an alloc. */
+		if (blkrec->jb_op == JOP_FREEBLK)
+			return;
+		TAILQ_REMOVE(&sblk->sb_recs, srec, sr_next);
+		free(srec);
+		break;
+	}
+	srec = errmalloc(sizeof(*srec));
+	srec->sr_rec = (union jrec *)blkrec;
+	TAILQ_INSERT_TAIL(&sblk->sb_recs, srec, sr_next);
+}
+
+/*
+ * Build up tables of the operations we need to recover.
+ */
+static void
+suj_build(void)
+{
+	struct suj_seg *seg;
+	union jrec *rec;
+	int i;
+
+	TAILQ_FOREACH(seg, &allsegs, ss_next) {
+		rec = (union jrec *)seg->ss_blk;
+		rec++;	/* skip the segrec. */
+		if (debug)
+			printf("seg %jd has %d records, oldseq %jd.\n",
+			    seg->ss_rec.jsr_seq, seg->ss_rec.jsr_cnt,
+			    seg->ss_rec.jsr_oldest);
+		for (i = 0; i < seg->ss_rec.jsr_cnt; i++, rec++) {
+			switch (rec->rec_jrefrec.jr_op) {
+			case JOP_ADDREF:
+			case JOP_REMREF:
+				suj_build_ino((struct jrefrec *)rec);
+				break;
+			case JOP_MVREF:
+				suj_move_ino((struct jmvrec *)rec);
+				break;
+			case JOP_NEWBLK:
+			case JOP_FREEBLK:
+				suj_build_blk((struct jblkrec *)rec);
+				break;
+			default:
+				errx(1, "Unknown journal operation %d (%d)",
+				    rec->rec_jrefrec.jr_op, i);
+			}
+		}
+	}
+}
+
+/*
+ * Prune the journal segments to those we care about based on the
+ * oldest sequence in the newest segment.  Order the segment list
+ * based on sequence number.
+ */
+static void
+suj_prune(void)
+{
+	struct suj_seg *seg;
+	struct suj_seg *segn;
+	uint64_t newseq;
+	int discard;
+
+	if (debug)
+		printf("Pruning up to %jd\n", oldseq);
+	/* First free the expired segments. */
+	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
+		if (seg->ss_rec.jsr_seq >= oldseq)
+			continue;
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		free(seg->ss_blk);
+		free(seg);
+	}
+	/* Next ensure that segments are ordered properly. */
+	seg = TAILQ_FIRST(&allsegs);
+	if (seg == NULL) {
+		if (debug)
+			printf("Empty journal\n");
+		return;
+	}
+	newseq = seg->ss_rec.jsr_seq;
+	for (;;) {
+		seg = TAILQ_LAST(&allsegs, seghd);
+		if (seg->ss_rec.jsr_seq >= newseq)
+			break;
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		TAILQ_INSERT_HEAD(&allsegs, seg, ss_next);
+		newseq = seg->ss_rec.jsr_seq;
+		
+	}
+	if (newseq != oldseq)
+		errx(1, "Journal file sequence mismatch %jd != %jd",
+		    newseq, oldseq);
+	/*
+	 * The kernel may asynchronously write segments which can create
+	 * gaps in the sequence space.  Throw away any segments after the
+	 * gap as the kernel guarantees only those that are contiguously
+	 * reachable are marked as completed.
+	 */
+	discard = 0;
+	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
+		if (!discard && newseq++ == seg->ss_rec.jsr_seq)
+			continue;
+		discard = 1;
+		if (debug)
+			printf("Journal order mismatch %jd != %jd pruning\n",
+			    newseq-1, seg->ss_rec.jsr_seq);
+		TAILQ_REMOVE(&allsegs, seg, ss_next);
+		free(seg->ss_blk);
+		free(seg);
+	}
+	if (debug)
+		printf("Processing journal segments from %jd to %jd\n",
+		    oldseq, newseq-1);
+}
+
+/*
+ * Verify the journal inode before attempting to read records.
+ */
+static void
+suj_verifyino(union dinode *ip)
+{
+
+	if (DIP(ip, di_nlink) != 1)
+		errx(1, "Invalid link count %d for journal inode %d",
+		    DIP(ip, di_nlink), fs->fs_sujournal);
+
+	if (DIP(ip, di_mode) != IFREG)
+		errx(1, "Invalid mode %d for journal inode %d",
+		    DIP(ip, di_mode), fs->fs_sujournal);
+
+	if (DIP(ip, di_size) < SUJ_MIN || DIP(ip, di_size) > SUJ_MAX)
+		errx(1, "Invalid size %jd for journal inode %d",
+		    DIP(ip, di_size), fs->fs_sujournal);
+
+	if (DIP(ip, di_modrev) != fs->fs_mtime)
+		errx(1, "Journal timestamp does not match fs mount time");
+	/* XXX Add further checks. */
+}
+
+struct jblocks {
+	struct jextent *jb_extent;	/* Extent array. */
+	int		jb_avail;	/* Available extents. */
+	int		jb_used;	/* Last used extent. */
+	int		jb_head;	/* Allocator head. */
+	int		jb_off;		/* Allocator extent offset. */
+};
+struct jextent {
+	ufs2_daddr_t	je_daddr;	/* Disk block address. */
+	int		je_blocks;	/* Disk block count. */
+};
+
+struct jblocks *suj_jblocks;
+
+static struct jblocks *
+jblocks_create(void)
+{
+	struct jblocks *jblocks;
+	int size;
+
+	jblocks = errmalloc(sizeof(*jblocks));
+	jblocks->jb_avail = 10;
+	jblocks->jb_used = 0;
+	jblocks->jb_head = 0;
+	jblocks->jb_off = 0;
+	size = sizeof(struct jextent) * jblocks->jb_avail;
+	jblocks->jb_extent = errmalloc(size);
+	bzero(jblocks->jb_extent, size);
+
+	return (jblocks);
+}
+
+/*
+ * Return the next available disk block and the amount of contiguous
+ * free space it contains.
+ */
+static ufs2_daddr_t
+jblocks_next(struct jblocks *jblocks, int bytes, int *actual)
+{
+	struct jextent *jext;
+	ufs2_daddr_t daddr;
+	int freecnt;
+	int blocks;
+
+	blocks = bytes / DEV_BSIZE;
+	jext = &jblocks->jb_extent[jblocks->jb_head];
+	freecnt = jext->je_blocks - jblocks->jb_off;
+	if (freecnt == 0) {
+		jblocks->jb_off = 0;
+		if (++jblocks->jb_head > jblocks->jb_used)
+			return (0);
+		jext = &jblocks->jb_extent[jblocks->jb_head];
+		freecnt = jext->je_blocks;
+	}
+	if (freecnt > blocks)
+		freecnt = blocks;
+	*actual = freecnt * DEV_BSIZE;
+	daddr = jext->je_daddr + jblocks->jb_off;
+
+	return (daddr);
+}
+
+/*
+ * Advance the allocation head by a specified number of bytes, consuming
+ * one journal segment.
+ */
+static void
+jblocks_advance(struct jblocks *jblocks, int bytes)
+{
+
+	jblocks->jb_off += bytes / DEV_BSIZE;
+}
+
+static void
+jblocks_destroy(struct jblocks *jblocks)
+{
+
+	free(jblocks->jb_extent);
+	free(jblocks);
+}
+
+static void
+jblocks_add(struct jblocks *jblocks, ufs2_daddr_t daddr, int blocks)
+{
+	struct jextent *jext;
+	int size;
+
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	/* Adding the first block. */
+	if (jext->je_daddr == 0) {
+		jext->je_daddr = daddr;
+		jext->je_blocks = blocks;
+		return;
+	}
+	/* Extending the last extent. */
+	if (jext->je_daddr + jext->je_blocks == daddr) {
+		jext->je_blocks += blocks;
+		return;
+	}
+	/* Adding a new extent. */
+	if (++jblocks->jb_used == jblocks->jb_avail) {
+		jblocks->jb_avail *= 2;
+		size = sizeof(struct jextent) * jblocks->jb_avail;
+		jext = errmalloc(size);
+		bzero(jext, size);
+		bcopy(jblocks->jb_extent, jext,
+		    sizeof(struct jextent) * jblocks->jb_used);
+		free(jblocks->jb_extent);
+		jblocks->jb_extent = jext;
+	}
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	jext->je_daddr = daddr;
+	jext->je_blocks = blocks;
+
+	return;
+}
+
+/*
+ * Add a file block from the journal to the extent map.  We can't read
+ * each file block individually because the kernel treats it as a circular
+ * buffer and segments may span mutliple contiguous blocks.
+ */
+static void
+suj_add_block(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
+{
+
+	jblocks_add(suj_jblocks, fsbtodb(fs, blk), fsbtodb(fs, frags));
+}
+
+static void
+suj_read(void)
+{
+	uint8_t block[1 * 1024 * 1024];
+	struct suj_seg *seg;
+	struct jsegrec *rec;
+	ufs2_daddr_t blk;
+	int recsize;
+	int size;
+
+	/*
+	 * Read records until we exhaust the journal space.  If we find
+	 * an invalid record we start searching for a valid segment header
+	 * at the next block.  This is because we don't have a head/tail
+	 * pointer and must recover the information indirectly.  At the gap
+	 * between the head and tail we won't necessarily have a valid
+	 * segment.
+	 */
+	for (;;) {
+		size = sizeof(block);
+		blk = jblocks_next(suj_jblocks, size, &size);
+		if (blk == 0)
+			return;
+		/*
+		 * Read 1MB at a time and scan for records within this block.
+		 */
+		if (bread(disk, blk, &block, size) == -1)
+			err(1, "Error reading journal block %jd",
+			    (intmax_t)blk);
+		for (rec = (void *)block; size; size -= recsize,
+		    rec = (struct jsegrec *)((uintptr_t)rec + recsize)) {
+			recsize = DEV_BSIZE;
+			if (rec->jsr_time != fs->fs_mtime) {
+				if (debug)
+					printf("Rec time %jd != fs mtime %jd\n",
+					    rec->jsr_time, fs->fs_mtime);
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			if (rec->jsr_cnt == 0) {
+				if (debug)
+					printf("Found illegal count %d\n",
+					    rec->jsr_cnt);
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			recsize = roundup2((rec->jsr_cnt + 1) * JREC_SIZE,
+			    DEV_BSIZE);
+			if (recsize > size) {
+				/*
+				 * We may just have run out of buffer, restart
+				 * the loop to re-read from this spot.
+				 */
+				if (size < fs->fs_bsize && 
+				    recsize <= fs->fs_bsize) {
+					recsize = size;
+					continue;
+				}
+				if (debug)
+					printf("Found invalid segsize %d > %d\n",
+					    recsize, size);
+				recsize = DEV_BSIZE;
+				jblocks_advance(suj_jblocks, recsize);
+				continue;
+			}
+			seg = errmalloc(sizeof(*seg));
+			seg->ss_blk = errmalloc(recsize);
+			seg->ss_rec = *rec;
+			bcopy((void *)rec, seg->ss_blk, recsize);
+			if (rec->jsr_oldest > oldseq)
+				oldseq = rec->jsr_oldest;
+			TAILQ_INSERT_TAIL(&allsegs, seg, ss_next);
+			jrecs += rec->jsr_cnt;
+			jbytes += recsize;
+			jblocks_advance(suj_jblocks, recsize);
+		}
+	}
+}
+
+/*
+ * Orchestrate the verification of a filesystem via the softupdates journal.
+ */
+void
+suj_check(const char *filesys)
+{
+	union dinode *jip;
+	uint64_t blocks;
+
+	opendisk(filesys);
+	TAILQ_INIT(&allsegs);
+	/*
+	 * Fetch the journal inode and verify it.
+	 */
+	jip = ino_read(fs->fs_sujournal);
+	printf("SU+J Checking %s\n", filesys);
+	suj_verifyino(jip);
+	/*
+	 * Build a list of journal blocks in jblocks before parsing the
+	 * available journal blocks in with suj_read().
+	 */
+	printf("Reading %jd byte journal from inode %d.\n",
+	    DIP(jip, di_size), fs->fs_sujournal);
+	suj_jblocks = jblocks_create();
+	blocks = ino_visit(jip, fs->fs_sujournal, suj_add_block, 0);
+	if (blocks != numfrags(fs, DIP(jip, di_size)))
+		errx(1, "Sparse journal inode %d.\n", fs->fs_sujournal);
+	suj_read();
+	jblocks_destroy(suj_jblocks);
+	suj_jblocks = NULL;
+	if (reply("RECOVER")) {
+		printf("Building recovery table.\n");
+		suj_prune();
+		suj_build();
+		printf("Resolving unreferenced inode list.\n");
+		ino_unlinked();
+		printf("Processing journal entries.\n");
+		cg_apply(cg_check);
+	}
+	if (reply("WRITE CHANGES"))
+		cg_apply(cg_write);
+	printf("%jd journal records in %jd bytes for %.2f%% utilization\n",
+	    jrecs, jbytes, ((float)jrecs / (float)(jbytes / JREC_SIZE)) * 100);
+	printf("Freed %jd inodes (%jd directories) %jd blocks and %jd frags.\n",
+	    freeinos, freedir, freeblocks, freefrags);
+	/* Write back superblock. */
+	closedisk(filesys);
+}
diff --git a/sbin/fsdb/fsdb.c b/sbin/fsdb/fsdb.c
index f7354e86d9c..5622cbb2919 100644
--- a/sbin/fsdb/fsdb.c
+++ b/sbin/fsdb/fsdb.c
@@ -396,7 +396,8 @@ const char *typename[] = {
     "unregistered #13",
     "whiteout",
 };
-    
+
+int diroff; 
 int slot;
 
 int
@@ -404,9 +405,10 @@ scannames(struct inodesc *idesc)
 {
 	struct direct *dirp = idesc->id_dirp;
 
-	printf("slot %d ino %d reclen %d: %s, `%.*s'\n",
-	       slot++, dirp->d_ino, dirp->d_reclen, typename[dirp->d_type],
-	       dirp->d_namlen, dirp->d_name);
+	printf("slot %d off %d ino %d reclen %d: %s, `%.*s'\n",
+	       slot++, diroff, dirp->d_ino, dirp->d_reclen,
+	       typename[dirp->d_type], dirp->d_namlen, dirp->d_name);
+	diroff += dirp->d_reclen;
 	return (KEEPON);
 }
 
@@ -416,6 +418,7 @@ CMDFUNCSTART(ls)
     checkactivedir();			/* let it go on anyway */
 
     slot = 0;
+    diroff = 0;
     idesc.id_number = curinum;
     idesc.id_func = scannames;
     idesc.id_type = DATA;
diff --git a/sbin/fsdb/fsdbutil.c b/sbin/fsdb/fsdbutil.c
index d50c6c00d33..2c5710a80aa 100644
--- a/sbin/fsdb/fsdbutil.c
+++ b/sbin/fsdb/fsdbutil.c
@@ -52,7 +52,7 @@ static const char rcsid[] =
 #include "fsck.h"
 
 static int charsperline(void);
-static int printindir(ufs2_daddr_t blk, int level, char *bufp);
+static void printindir(ufs2_daddr_t blk, int level, char *bufp);
 static void printblocks(ino_t inum, union dinode *dp);
 
 char **
@@ -226,7 +226,7 @@ charsperline(void)
 /*
  * Recursively print a list of indirect blocks.
  */
-static int
+static void
 printindir(ufs2_daddr_t blk, int level, char *bufp)
 {
     struct bufarea buf, *bp;
@@ -234,6 +234,9 @@ printindir(ufs2_daddr_t blk, int level, char *bufp)
     int i, j, cpl, charssofar;
     ufs2_daddr_t blkno;
 
+    if (blk == 0)
+	return;
+    printf("%jd (%d) =>\n", (intmax_t)blk, level);
     if (level == 0) {
 	/* for the final indirect level, don't use the cache */
 	bp = &buf;
@@ -251,11 +254,8 @@ printindir(ufs2_daddr_t blk, int level, char *bufp)
 		blkno = bp->b_un.b_indir1[i];
 	else
 		blkno = bp->b_un.b_indir2[i];
-	if (blkno == 0) {
-	    if (level == 0)
-		putchar('\n');
-	    return 0;
-	}
+	if (blkno == 0)
+	    continue;
 	j = sprintf(tempbuf, "%jd", (intmax_t)blkno);
 	if (level == 0) {
 	    charssofar += j;
@@ -270,13 +270,14 @@ printindir(ufs2_daddr_t blk, int level, char *bufp)
 	    charssofar += 2;
 	} else {
 	    printf(" =>\n");
-	    if (printindir(blkno, level - 1, bufp) == 0)
-		return 0;
+	    printindir(blkno, level - 1, bufp);
+	    printf("\n");
+	    charssofar = 0;
 	}
     }
     if (level == 0)
 	putchar('\n');
-    return 1;
+    return;
 }
 
 
@@ -309,7 +310,7 @@ printblocks(ino_t inum, union dinode *dp)
 	}
     }
     putchar('\n');
-    if (DIP(dp, di_ib[0]) == 0)
+    if (ndb == 0)
 	return;
 
     bufp = malloc((unsigned int)sblock.fs_bsize);
@@ -317,8 +318,7 @@ printblocks(ino_t inum, union dinode *dp)
 	errx(EEXIT, "cannot allocate indirect block buffer");
     printf("Indirect blocks:\n");
     for (i = 0; i < NIADDR; i++)
-	if (printindir(DIP(dp, di_ib[i]), i, bufp) == 0)
-	    break;
+	printindir(DIP(dp, di_ib[i]), i, bufp);
     free(bufp);
 }
 
diff --git a/sbin/tunefs/tunefs.8 b/sbin/tunefs/tunefs.8
index 53e463cfc70..a883cd4fc61 100644
--- a/sbin/tunefs/tunefs.8
+++ b/sbin/tunefs/tunefs.8
@@ -28,7 +28,7 @@
 .\"     @(#)tunefs.8	8.2 (Berkeley) 12/11/93
 .\" $FreeBSD$
 .\"
-.Dd October 21, 2009
+.Dd March 6, 2010
 .Dt TUNEFS 8
 .Os
 .Sh NAME
@@ -40,6 +40,7 @@
 .Op Fl a Cm enable | disable
 .Op Fl e Ar maxbpg
 .Op Fl f Ar avgfilesize
+.Op Fl j Cm enable | disable
 .Op Fl J Cm enable | disable
 .Op Fl L Ar volname
 .Op Fl l Cm enable | disable
@@ -49,6 +50,7 @@
 .Op Fl o Cm space | time
 .Op Fl p
 .Op Fl s Ar avgfpdir
+.Op Fl S Ar size
 .Ar special | filesystem
 .Sh DESCRIPTION
 The
@@ -89,6 +91,8 @@ For file systems with exclusively large files,
 this parameter should be set higher.
 .It Fl f Ar avgfilesize
 Specify the expected average file size.
+.It Fl j Cm enable | disable
+Turn on/off soft updates journaling.
 .It Fl J Cm enable | disable
 Turn on/off gjournal flag.
 .It Fl L Ar volname
@@ -136,6 +140,9 @@ obtained from the
 utility.
 .It Fl s Ar avgfpdir
 Specify the expected number of files per directory.
+.It Fl S Ar size
+Specify the softdep journal size in bytes.
+The minimum is 4M.
 .El
 .Pp
 At least one of the above flags is required.
diff --git a/sbin/tunefs/tunefs.c b/sbin/tunefs/tunefs.c
index e4adb5267eb..a10b35daec2 100644
--- a/sbin/tunefs/tunefs.c
+++ b/sbin/tunefs/tunefs.c
@@ -52,6 +52,7 @@ __FBSDID("$FreeBSD$");
 #include <ufs/ufs/ufsmount.h>
 #include <ufs/ufs/dinode.h>
 #include <ufs/ffs/fs.h>
+#include <ufs/ufs/dir.h>
 
 #include <ctype.h>
 #include <err.h>
@@ -61,6 +62,7 @@ __FBSDID("$FreeBSD$");
 #include <paths.h>
 #include <stdio.h>
 #include <stdlib.h>
+#include <stdint.h>
 #include <string.h>
 #include <unistd.h>
 
@@ -72,16 +74,20 @@ struct uufsd disk;
 
 void usage(void);
 void printfs(void);
+int journal_alloc(int64_t size);
+void journal_clear(void);
+void sbdirty(void);
 
 int
 main(int argc, char *argv[])
 {
-	char *avalue, *Jvalue, *Lvalue, *lvalue, *Nvalue, *nvalue;
+	char *avalue, *jvalue, *Jvalue, *Lvalue, *lvalue, *Nvalue, *nvalue;
 	const char *special, *on;
 	const char *name;
 	int active;
-	int Aflag, aflag, eflag, evalue, fflag, fvalue, Jflag, Lflag, lflag;
-	int mflag, mvalue, Nflag, nflag, oflag, ovalue, pflag, sflag, svalue;
+	int Aflag, aflag, eflag, evalue, fflag, fvalue, jflag, Jflag, Lflag;
+	int lflag, mflag, mvalue, Nflag, nflag, oflag, ovalue, pflag, sflag;
+	int svalue, Sflag, Svalue;
 	int ch, found_arg, i;
 	const char *chg[2];
 	struct ufs_args args;
@@ -89,13 +95,13 @@ main(int argc, char *argv[])
 
 	if (argc < 3)
 		usage();
-	Aflag = aflag = eflag = fflag = Jflag = Lflag = lflag = mflag = 0;
-	Nflag = nflag = oflag = pflag = sflag = 0;
-	avalue = Jvalue = Lvalue = lvalue = Nvalue = nvalue = NULL;
-	evalue = fvalue = mvalue = ovalue = svalue = 0;
+	Aflag = aflag = eflag = fflag = jflag = Jflag = Lflag = lflag = 0;
+	mflag = Nflag = nflag = oflag = pflag = sflag = 0;
+	avalue = jvalue = Jvalue = Lvalue = lvalue = Nvalue = nvalue = NULL;
+	evalue = fvalue = mvalue = ovalue = svalue = Svalue = 0;
 	active = 0;
 	found_arg = 0;		/* At least one arg is required. */
-	while ((ch = getopt(argc, argv, "Aa:e:f:J:L:l:m:N:n:o:ps:")) != -1)
+	while ((ch = getopt(argc, argv, "Aa:e:f:j:J:L:l:m:N:n:o:ps:S:")) != -1)
 		switch (ch) {
 
 		case 'A':
@@ -135,6 +141,18 @@ main(int argc, char *argv[])
 			fflag = 1;
 			break;
 
+		case 'j':
+			found_arg = 1;
+			name = "softdep journaled file system";
+			jvalue = optarg;
+			if (strcmp(jvalue, "enable") &&
+			    strcmp(jvalue, "disable")) {
+				errx(10, "bad %s (options are %s)",
+				    name, "`enable' or `disable'");
+			}
+			jflag = 1;
+			break;
+
 		case 'J':
 			found_arg = 1;
 			name = "gjournaled file system";
@@ -240,6 +258,16 @@ main(int argc, char *argv[])
 			sflag = 1;
 			break;
 
+		case 'S':
+			found_arg = 1;
+			name = "Softdep Journal Size";
+			Svalue = atoi(optarg);
+			if (Svalue < SUJ_MIN)
+				errx(10, "%s must be >= %d (was %s)",
+				    name, SUJ_MIN, optarg);
+			Sflag = 1;
+			break;
+
 		default:
 			usage();
 		}
@@ -310,6 +338,33 @@ main(int argc, char *argv[])
 			sblock.fs_avgfilesize = fvalue;
 		}
 	}
+	if (jflag) {
+ 		name = "soft updates journaling";
+ 		if (strcmp(jvalue, "enable") == 0) {
+			if ((sblock.fs_flags & (FS_DOSOFTDEP | FS_SUJ)) ==
+			    (FS_DOSOFTDEP | FS_SUJ)) {
+				warnx("%s remains unchanged as enabled", name);
+			} else if (sblock.fs_clean == 0) {
+				warnx("%s cannot be enabled until fsck is run",
+				    name);
+			} else if (journal_alloc(Svalue) != 0) {
+				warnx("%s can not be enabled", name);
+			} else {
+ 				sblock.fs_flags |= FS_DOSOFTDEP | FS_SUJ;
+ 				warnx("%s set", name);
+			}
+ 		} else if (strcmp(jvalue, "disable") == 0) {
+			if ((~sblock.fs_flags & FS_SUJ) == FS_SUJ) {
+				warnx("%s remains unchanged as disabled", name);
+			} else {
+				journal_clear();
+ 				sblock.fs_flags &= ~(FS_DOSOFTDEP | FS_SUJ);
+				sblock.fs_sujfree = 0;
+ 				warnx("%s cleared, "
+				    "remove .sujournal to reclaim space", name);
+			}
+ 		}
+	}
 	if (Jflag) {
 		name = "gjournal";
 		if (strcmp(Jvalue, "enable") == 0) {
@@ -455,6 +510,500 @@ err:
 		err(12, "%s", special);
 }
 
+void
+sbdirty(void)
+{
+	disk.d_fs.fs_flags |= FS_UNCLEAN | FS_NEEDSFSCK;
+	disk.d_fs.fs_clean = 0;
+}
+
+int blocks;
+static char clrbuf[MAXBSIZE];
+
+static ufs2_daddr_t
+journal_balloc(void)
+{
+	ufs2_daddr_t blk;
+	struct cg *cgp;
+	int valid;
+	static int contig = 1;
+
+	cgp = &disk.d_cg;
+	for (;;) {
+		blk = cgballoc(&disk);
+		if (blk > 0)
+			break;
+		/*
+		 * If we failed to allocate a block from this cg, move to
+		 * the next.
+		 */
+		if (cgwrite(&disk) < 0) {
+			warn("Failed to write updated cg");
+			return (-1);
+		}
+		while ((valid = cgread(&disk)) == 1) {
+			/*
+			 * Try to minimize fragmentation by requiring a minimum
+			 * number of blocks present.
+			 */
+			if (cgp->cg_cs.cs_nbfree > blocks / 8)
+				break;
+			if (contig == 0 && cgp->cg_cs.cs_nbfree)
+				break;
+		}
+		if (valid)
+			continue;
+		/*
+		 * Try once through looking only for large contiguous regions
+		 * and again taking any space we can find.
+		 */
+		if (contig) {
+			contig = 0;
+			disk.d_ccg = 0;
+			warnx("Journal file fragmented.");
+			continue;
+		}
+		warnx("Failed to find sufficient free blocks for the journal");
+		return -1;
+	}
+	if (bwrite(&disk, fsbtodb(&sblock, blk), clrbuf,
+	    sblock.fs_bsize) <= 0) {
+		warn("Failed to initialize new block");
+		return -1;
+	}
+	return (blk);
+}
+
+/*
+ * Search a directory block for the SUJ_FILE.
+ */
+static ino_t
+dir_search(ufs2_daddr_t blk, int bytes)
+{
+	char block[MAXBSIZE];
+	struct direct *dp;
+	int off;
+
+	if (bread(&disk, fsbtodb(&sblock, blk), block, bytes) <= 0) {
+		warn("Failed to read dir block");
+		return (-1);
+	}
+	for (off = 0; off < bytes; off += dp->d_reclen) {
+		dp = (struct direct *)&block[off];
+		if (dp->d_reclen == 0)
+			break;
+		if (dp->d_ino == 0)
+			continue;
+		if (dp->d_namlen != strlen(SUJ_FILE))
+			continue;
+		if (bcmp(dp->d_name, SUJ_FILE, dp->d_namlen) != 0)
+			continue;
+		return (dp->d_ino);
+	}
+
+	return (0);
+}
+
+/*
+ * Search in the ROOTINO for the SUJ_FILE.  If it exists we can not enable
+ * journaling.
+ */
+static ino_t
+journal_findfile(void)
+{
+	struct ufs1_dinode *dp1;
+	struct ufs2_dinode *dp2;
+	ino_t ino;
+	int mode;
+	void *ip;
+	int i;
+
+	if (getino(&disk, &ip, ROOTINO, &mode) != 0) {
+		warn("Failed to get root inode");
+		return (-1);
+	}
+	dp2 = ip;
+	dp1 = ip;
+	if (sblock.fs_magic == FS_UFS1_MAGIC) {
+		if ((off_t)dp1->di_size >= lblktosize(&sblock, NDADDR)) {
+			warnx("ROOTINO extends beyond direct blocks.");
+			return (-1);
+		}
+		for (i = 0; i < NDADDR; i++) {
+			if (dp1->di_db[i] == 0)
+				break;
+			if ((ino = dir_search(dp1->di_db[i],
+			    sblksize(&sblock, (off_t)dp1->di_size, i))) != 0)
+				return (ino);
+		}
+	} else {
+		if ((off_t)dp1->di_size >= lblktosize(&sblock, NDADDR)) {
+			warnx("ROOTINO extends beyond direct blocks.");
+			return (-1);
+		}
+		for (i = 0; i < NDADDR; i++) {
+			if (dp2->di_db[i] == 0)
+				break;
+			if ((ino = dir_search(dp2->di_db[i],
+			    sblksize(&sblock, (off_t)dp2->di_size, i))) != 0)
+				return (ino);
+		}
+	}
+
+	return (0);
+}
+
+/*
+ * Insert the journal at inode 'ino' into directory blk 'blk' at the first
+ * free offset of 'off'.  DIRBLKSIZ blocks after off are initialized as
+ * empty.
+ */
+static int
+dir_insert(ufs2_daddr_t blk, off_t off, ino_t ino)
+{
+	struct direct *dp;
+	char block[MAXBSIZE];
+
+	if (bread(&disk, fsbtodb(&sblock, blk), block, sblock.fs_bsize) <= 0) {
+		warn("Failed to read dir block");
+		return (-1);
+	}
+	bzero(&block[off], sblock.fs_bsize - off);
+	dp = (struct direct *)&block[off];
+	dp->d_ino = ino;
+	dp->d_reclen = DIRBLKSIZ;
+	dp->d_type = DT_REG;
+	dp->d_namlen = strlen(SUJ_FILE);
+	bcopy(SUJ_FILE, &dp->d_name, strlen(SUJ_FILE));
+	off += DIRBLKSIZ;
+	for (; off < sblock.fs_bsize; off += DIRBLKSIZ) {
+		dp = (struct direct *)&block[off];
+		dp->d_ino = 0;
+		dp->d_reclen = DIRBLKSIZ;
+		dp->d_type = DT_UNKNOWN;
+	}
+	if (bwrite(&disk, fsbtodb(&sblock, blk), block, sblock.fs_bsize) <= 0) {
+		warn("Failed to write dir block");
+		return (-1);
+	}
+	return (0);
+}
+
+/*
+ * Extend a directory block in 'blk' by copying it to a full size block
+ * and inserting the new journal inode into .sujournal.
+ */
+static int
+dir_extend(ufs2_daddr_t blk, ufs2_daddr_t nblk, off_t size, ino_t ino)
+{
+	char block[MAXBSIZE];
+
+	if (bread(&disk, fsbtodb(&sblock, blk), block, size) <= 0) {
+		warn("Failed to read dir block");
+		return (-1);
+	}
+	if (bwrite(&disk, fsbtodb(&sblock, nblk), block, size) <= 0) {
+		warn("Failed to write dir block");
+		return (-1);
+	}
+
+	return dir_insert(nblk, size, ino);
+}
+
+/*
+ * Insert the journal file into the ROOTINO directory.  We always extend the
+ * last frag
+ */
+static int
+journal_insertfile(ino_t ino)
+{
+	struct ufs1_dinode *dp1;
+	struct ufs2_dinode *dp2;
+	void *ip;
+	ufs2_daddr_t nblk;
+	ufs2_daddr_t blk;
+	ufs_lbn_t lbn;
+	int size;
+	int mode;
+	int off;
+
+	if (getino(&disk, &ip, ROOTINO, &mode) != 0) {
+		warn("Failed to get root inode");
+		sbdirty();
+		return (-1);
+	}
+	dp2 = ip;
+	dp1 = ip;
+	blk = 0;
+	size = 0;
+	nblk = journal_balloc();
+	if (nblk <= 0)
+		return (-1);
+	/*
+	 * For simplicity sake we aways extend the ROOTINO into a new
+	 * directory block rather than searching for space and inserting
+	 * into an existing block.  However, if the rootino has frags
+	 * have to free them and extend the block.
+	 */
+	if (sblock.fs_magic == FS_UFS1_MAGIC) {
+		lbn = lblkno(&sblock, dp1->di_size);
+		off = blkoff(&sblock, dp1->di_size);
+		blk = dp1->di_db[lbn];
+		size = sblksize(&sblock, (off_t)dp1->di_size, lbn);
+	} else {
+		lbn = lblkno(&sblock, dp2->di_size);
+		off = blkoff(&sblock, dp2->di_size);
+		blk = dp2->di_db[lbn];
+		size = sblksize(&sblock, (off_t)dp2->di_size, lbn);
+	}
+	if (off != 0) {
+		if (dir_extend(blk, nblk, off, ino) == -1)
+			return (-1);
+	} else {
+		blk = 0;
+		if (dir_insert(nblk, 0, ino) == -1)
+			return (-1);
+	}
+	if (sblock.fs_magic == FS_UFS1_MAGIC) {
+		dp1->di_blocks += (sblock.fs_bsize - size) / DEV_BSIZE;
+		dp1->di_db[lbn] = nblk;
+		dp1->di_size = lblktosize(&sblock, lbn+1);
+	} else {
+		dp2->di_blocks += (sblock.fs_bsize - size) / DEV_BSIZE;
+		dp2->di_db[lbn] = nblk;
+		dp2->di_size = lblktosize(&sblock, lbn+1);
+	}
+	if (putino(&disk) < 0) {
+		warn("Failed to write root inode");
+		return (-1);
+	}
+	if (cgwrite(&disk) < 0) {
+		warn("Failed to write updated cg");
+		sbdirty();
+		return (-1);
+	}
+	if (blk) {
+		if (cgbfree(&disk, blk, size) < 0) {
+			warn("Failed to write cg");
+			return (-1);
+		}
+	}
+
+	return (0);
+}
+
+static int
+indir_fill(ufs2_daddr_t blk, int level, int *resid)
+{
+	char indirbuf[MAXBSIZE];
+	ufs1_daddr_t *bap1;
+	ufs2_daddr_t *bap2;
+	ufs2_daddr_t nblk;
+	int ncnt;
+	int cnt;
+	int i;
+
+	bzero(indirbuf, sizeof(indirbuf));
+	bap1 = (ufs1_daddr_t *)indirbuf;
+	bap2 = (void *)bap1;
+	cnt = 0;
+	for (i = 0; i < NINDIR(&sblock) && *resid != 0; i++) {
+		nblk = journal_balloc();
+		if (nblk <= 0)
+			return (-1);
+		cnt++;
+		if (sblock.fs_magic == FS_UFS1_MAGIC)
+			*bap1++ = nblk;
+		else
+			*bap2++ = nblk;
+		if (level != 0) {
+			ncnt = indir_fill(nblk, level - 1, resid);
+			if (ncnt <= 0)
+				return (-1);
+			cnt += ncnt;
+		} else 
+			(*resid)--;
+	}
+	if (bwrite(&disk, fsbtodb(&sblock, blk), indirbuf,
+	    sblock.fs_bsize) <= 0) {
+		warn("Failed to write indirect");
+		return (-1);
+	}
+	return (cnt);
+}
+
+/*
+ * Clear the flag bits so the journal can be removed.
+ */
+void
+journal_clear(void)
+{
+	struct ufs1_dinode *dp1;
+	struct ufs2_dinode *dp2;
+	ino_t ino;
+	int mode;
+	void *ip;
+
+	ino = journal_findfile();
+	if (ino == (ino_t)-1 || ino == 0) {
+		warnx("Journal file does not exist");
+		return;
+	}
+	printf("Clearing journal flags from inode %d\n", ino);
+	if (getino(&disk, &ip, ino, &mode) != 0) {
+		warn("Failed to get journal inode");
+		return;
+	}
+	dp2 = ip;
+	dp1 = ip;
+	if (sblock.fs_magic == FS_UFS1_MAGIC)
+		dp1->di_flags = 0;
+	else
+		dp2->di_flags = 0;
+	if (putino(&disk) < 0) {
+		warn("Failed to write journal inode");
+		return;
+	}
+}
+
+int
+journal_alloc(int64_t size)
+{
+	struct ufs1_dinode *dp1;
+	struct ufs2_dinode *dp2;
+	ufs2_daddr_t blk;
+	void *ip;
+	struct cg *cgp;
+	int resid;
+	ino_t ino;
+	int blks;
+	int mode;
+	int i;
+
+	cgp = &disk.d_cg;
+	ino = 0;
+
+	/*
+	 * If the journal file exists we can't allocate it.
+	 */
+	ino = journal_findfile();
+	if (ino == (ino_t)-1)
+		return (-1);
+	if (ino > 0) {
+		warnx("Journal file %s already exists, please remove.",
+		    SUJ_FILE);
+		return (-1);
+	}
+	/*
+	 * If the user didn't supply a size pick one based on the filesystem
+	 * size constrained with hardcoded MIN and MAX values.  We opt for
+	 * 1/1024th of the filesystem up to MAX but not exceeding one CG and
+	 * not less than the MIN.
+	 */
+	if (size == 0) {
+		size = (sblock.fs_size * sblock.fs_bsize) / 1024;
+		size = MIN(SUJ_MAX, size);
+		if (size / sblock.fs_fsize > sblock.fs_fpg)
+			size = sblock.fs_fpg * sblock.fs_fsize;
+		size = MAX(SUJ_MIN, size);
+	}
+	resid = blocks = size / sblock.fs_bsize;
+	if (sblock.fs_cstotal.cs_nbfree < blocks) {
+		warn("Insufficient free space for %jd byte journal", size);
+		return (-1);
+	}
+	/*
+	 * Find a cg with enough blocks to satisfy the journal
+	 * size.  Presently the journal does not span cgs.
+	 */
+	while (cgread(&disk) == 1) {
+		if (cgp->cg_cs.cs_nifree == 0)
+			continue;
+		ino = cgialloc(&disk);
+		if (ino <= 0)
+			break;
+		printf("Using inode %d in cg %d for %jd byte journal\n", 
+		    ino, cgp->cg_cgx, size);
+		if (getino(&disk, &ip, ino, &mode) != 0) {
+			warn("Failed to get allocated inode");
+			sbdirty();
+			goto out;
+		}
+		/*
+		 * We leave fields unrelated to the number of allocated
+		 * blocks and size uninitialized.  This causes legacy
+		 * fsck implementations to clear the inode.
+		 */
+		dp2 = ip;
+		dp1 = ip;
+		if (sblock.fs_magic == FS_UFS1_MAGIC) {
+			bzero(dp1, sizeof(*dp1));
+			dp1->di_size = size;
+			dp1->di_mode = IFREG | IREAD;
+			dp1->di_nlink = 1;
+			dp1->di_flags = SF_IMMUTABLE | SF_NOUNLINK | UF_NODUMP;
+		} else {
+			bzero(dp2, sizeof(*dp2));
+			dp2->di_size = size;
+			dp2->di_mode = IFREG | IREAD;
+			dp2->di_nlink = 1;
+			dp2->di_flags = SF_IMMUTABLE | SF_NOUNLINK | UF_NODUMP;
+		}
+		for (i = 0; i < NDADDR && resid; i++, resid--) {
+			blk = journal_balloc();
+			if (blk <= 0)
+				goto out;
+			if (sblock.fs_magic == FS_UFS1_MAGIC) {
+				dp1->di_db[i] = blk;
+				dp1->di_blocks++;
+			} else {
+				dp2->di_db[i] = blk;
+				dp2->di_blocks++;
+			}
+		}
+		for (i = 0; i < NIADDR && resid; i++) {
+			blk = journal_balloc();
+			if (blk <= 0)
+				goto out;
+			blks = indir_fill(blk, i, &resid) + 1;
+			if (blks <= 0) {
+				sbdirty();
+				goto out;
+			}
+			if (sblock.fs_magic == FS_UFS1_MAGIC) {
+				dp1->di_ib[i] = blk;
+				dp1->di_blocks += blks;
+			} else {
+				dp2->di_ib[i] = blk;
+				dp2->di_blocks += blks;
+			}
+		}
+		if (sblock.fs_magic == FS_UFS1_MAGIC)
+			dp1->di_blocks *= sblock.fs_bsize / disk.d_bsize;
+		else
+			dp2->di_blocks *= sblock.fs_bsize / disk.d_bsize;
+		if (putino(&disk) < 0) {
+			warn("Failed to write inode");
+			sbdirty();
+			return (-1);
+		}
+		if (cgwrite(&disk) < 0) {
+			warn("Failed to write updated cg");
+			sbdirty();
+			return (-1);
+		}
+		if (journal_insertfile(ino) < 0) {
+			sbdirty();
+			return (-1);
+		}
+		sblock.fs_sujfree = 0;
+		return (0);
+	}
+	warnx("Insufficient free space for the journal.");
+out:
+	return (-1);
+}
+
 void
 usage(void)
 {
@@ -477,6 +1026,8 @@ printfs(void)
 		(sblock.fs_flags & FS_MULTILABEL)? "enabled" : "disabled");
 	warnx("soft updates: (-n)                                 %s", 
 		(sblock.fs_flags & FS_DOSOFTDEP)? "enabled" : "disabled");
+	warnx("soft update journaling: (-j)                       %s", 
+		(sblock.fs_flags & FS_SUJ)? "enabled" : "disabled");
 	warnx("gjournal: (-J)                                     %s",
 		(sblock.fs_flags & FS_GJOURNAL)? "enabled" : "disabled");
 	warnx("maximum blocks per file in a cylinder group: (-e)  %d",
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 4e9cfc69957..156b676cf6b 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -215,6 +215,14 @@ SYSCTL_LONG(_vfs, OID_AUTO, notbufdflashes, CTLFLAG_RD, &notbufdflashes, 0,
  */
 static int bd_request;
 
+/*
+ * Request for the buf daemon to write more buffers than is indicated by
+ * lodirtybuf.  This may be necessary to push out excess dependencies or
+ * defragment the address space where a simple count of the number of dirty
+ * buffers is insufficient to characterize the demand for flushing them.
+ */
+static int bd_speedupreq;
+
 /*
  * This lock synchronizes access to bd_request.
  */
@@ -467,12 +475,20 @@ bd_wakeup(int dirtybuflevel)
  * bd_speedup - speedup the buffer cache flushing code
  */
 
-static __inline
 void
 bd_speedup(void)
 {
+	int needwake;
 
-	bd_wakeup(1);
+	mtx_lock(&bdlock);
+	needwake = 0;
+	if (bd_speedupreq == 0 || bd_request == 0)
+		needwake = 1;
+	bd_speedupreq = 1;
+	bd_request = 1;
+	if (needwake)
+		wakeup(&bd_request);
+	mtx_unlock(&bdlock);
 }
 
 /*
@@ -2120,6 +2136,7 @@ buf_do_flush(struct vnode *vp)
 static void
 buf_daemon()
 {
+	int lodirtysave;
 
 	/*
 	 * This process needs to be suspended prior to shutdown sync.
@@ -2137,7 +2154,11 @@ buf_daemon()
 		mtx_unlock(&bdlock);
 
 		kproc_suspend_check(bufdaemonproc);
-
+		lodirtysave = lodirtybuffers;
+		if (bd_speedupreq) {
+			lodirtybuffers = numdirtybuffers / 2;
+			bd_speedupreq = 0;
+		}
 		/*
 		 * Do the flush.  Limit the amount of in-transit I/O we
 		 * allow to build up, otherwise we would completely saturate
@@ -2149,6 +2170,7 @@ buf_daemon()
 				break;
 			uio_yield();
 		}
+		lodirtybuffers = lodirtysave;
 
 		/*
 		 * Only clear bd_request if we have reached our low water
diff --git a/sys/kern/vfs_subr.c b/sys/kern/vfs_subr.c
index 4810e769b15..ae182e00839 100644
--- a/sys/kern/vfs_subr.c
+++ b/sys/kern/vfs_subr.c
@@ -2815,6 +2815,7 @@ DB_SHOW_COMMAND(mount, db_show_mount)
 	MNT_FLAG(MNT_FORCE);
 	MNT_FLAG(MNT_SNAPSHOT);
 	MNT_FLAG(MNT_BYFSID);
+	MNT_FLAG(MNT_SOFTDEP);
 #undef MNT_FLAG
 	if (flags != 0) {
 		if (buf[0] != '\0')
diff --git a/sys/sys/buf.h b/sys/sys/buf.h
index 8f3b1b2ab3f..137f90f6d72 100644
--- a/sys/sys/buf.h
+++ b/sys/sys/buf.h
@@ -215,7 +215,7 @@ struct buf {
 #define	B_DIRTY		0x00200000	/* Needs writing later (in EXT2FS). */
 #define	B_RELBUF	0x00400000	/* Release VMIO buffer. */
 #define	B_00800000	0x00800000	/* Available flag. */
-#define	B_01000000	0x01000000	/* Available flag. */
+#define	B_NOCOPY	0x01000000	/* Don't copy-on-write this buf. */
 #define	B_NEEDSGIANT	0x02000000	/* Buffer's vnode needs giant. */
 #define	B_PAGING	0x04000000	/* volatile paging I/O -- bypass VMIO */
 #define B_MANAGED	0x08000000	/* Managed by FS. */
@@ -493,6 +493,7 @@ int	bufwait(struct buf *);
 int	bufwrite(struct buf *);
 void	bufdone(struct buf *);
 void	bufdone_finish(struct buf *);
+void	bd_speedup(void);
 
 int	cluster_read(struct vnode *, u_quad_t, daddr_t, long,
 	    struct ucred *, long, int, struct buf **);
diff --git a/sys/sys/mount.h b/sys/sys/mount.h
index b8264130a17..20dcf641c7b 100644
--- a/sys/sys/mount.h
+++ b/sys/sys/mount.h
@@ -275,7 +275,8 @@ void          __mnt_vnode_markerfree(struct vnode **mvp, struct mount *mp);
 			MNT_ROOTFS	| MNT_NOATIME	| MNT_NOCLUSTERR| \
 			MNT_NOCLUSTERW	| MNT_SUIDDIR	| MNT_SOFTDEP	| \
 			MNT_IGNORE	| MNT_EXPUBLIC	| MNT_NOSYMFOLLOW | \
-			MNT_GJOURNAL	| MNT_MULTILABEL | MNT_ACLS | MNT_NFS4ACLS)
+			MNT_GJOURNAL	| MNT_MULTILABEL | MNT_ACLS	| \
+			MNT_NFS4ACLS)
 
 /* Mask of flags that can be updated. */
 #define	MNT_UPDATEMASK (MNT_NOSUID	| MNT_NOEXEC	| \
@@ -324,6 +325,7 @@ void          __mnt_vnode_markerfree(struct vnode **mvp, struct mount *mp);
 #define	MNTK_REFEXPIRE	0x00000020	/* refcount expiring is happening */
 #define MNTK_EXTENDED_SHARED	0x00000040 /* Allow shared locking for more ops */
 #define	MNTK_SHARED_WRITES	0x00000080 /* Allow shared locking for writes */
+#define	MNTK_SUJ	0x00000100	/* Softdep journaling enabled */
 #define MNTK_UNMOUNT	0x01000000	/* unmount in progress */
 #define	MNTK_MWAIT	0x02000000	/* waiting for unmount to finish */
 #define	MNTK_SUSPEND	0x08000000	/* request write suspension */
diff --git a/sys/ufs/ffs/ffs_alloc.c b/sys/ufs/ffs/ffs_alloc.c
index 7bf11771972..b1f7ba0127f 100644
--- a/sys/ufs/ffs/ffs_alloc.c
+++ b/sys/ufs/ffs/ffs_alloc.c
@@ -94,24 +94,24 @@ __FBSDID("$FreeBSD$");
 #include <ufs/ffs/ffs_extern.h>
 
 typedef ufs2_daddr_t allocfcn_t(struct inode *ip, u_int cg, ufs2_daddr_t bpref,
-				  int size);
+				  int size, int rsize);
 
-static ufs2_daddr_t ffs_alloccg(struct inode *, u_int, ufs2_daddr_t, int);
+static ufs2_daddr_t ffs_alloccg(struct inode *, u_int, ufs2_daddr_t, int, int);
 static ufs2_daddr_t
-	      ffs_alloccgblk(struct inode *, struct buf *, ufs2_daddr_t);
+	      ffs_alloccgblk(struct inode *, struct buf *, ufs2_daddr_t, int);
 #ifdef INVARIANTS
 static int	ffs_checkblk(struct inode *, ufs2_daddr_t, long);
 #endif
-static ufs2_daddr_t ffs_clusteralloc(struct inode *, u_int, ufs2_daddr_t, int);
-static void	ffs_clusteracct(struct ufsmount *, struct fs *, struct cg *,
-		    ufs1_daddr_t, int);
+static ufs2_daddr_t ffs_clusteralloc(struct inode *, u_int, ufs2_daddr_t, int,
+		    int);
 static ino_t	ffs_dirpref(struct inode *);
 static ufs2_daddr_t ffs_fragextend(struct inode *, u_int, ufs2_daddr_t,
 		    int, int);
 static void	ffs_fserr(struct fs *, ino_t, char *);
 static ufs2_daddr_t	ffs_hashalloc
-		(struct inode *, u_int, ufs2_daddr_t, int, allocfcn_t *);
-static ufs2_daddr_t ffs_nodealloccg(struct inode *, u_int, ufs2_daddr_t, int);
+		(struct inode *, u_int, ufs2_daddr_t, int, int, allocfcn_t *);
+static ufs2_daddr_t ffs_nodealloccg(struct inode *, u_int, ufs2_daddr_t, int,
+		    int);
 static ufs1_daddr_t ffs_mapsearch(struct fs *, struct cg *, ufs2_daddr_t, int);
 static int	ffs_reallocblks_ufs1(struct vop_reallocblks_args *);
 static int	ffs_reallocblks_ufs2(struct vop_reallocblks_args *);
@@ -188,7 +188,7 @@ retry:
 		cg = ino_to_cg(fs, ip->i_number);
 	else
 		cg = dtog(fs, bpref);
-	bno = ffs_hashalloc(ip, cg, bpref, size, ffs_alloccg);
+	bno = ffs_hashalloc(ip, cg, bpref, size, size, ffs_alloccg);
 	if (bno > 0) {
 		delta = btodb(size);
 		if (ip->i_flag & IN_SPACECOUNTED) {
@@ -387,16 +387,12 @@ retry:
 		panic("ffs_realloccg: bad optim");
 		/* NOTREACHED */
 	}
-	bno = ffs_hashalloc(ip, cg, bpref, request, ffs_alloccg);
+	bno = ffs_hashalloc(ip, cg, bpref, request, nsize, ffs_alloccg);
 	if (bno > 0) {
 		bp->b_blkno = fsbtodb(fs, bno);
 		if (!DOINGSOFTDEP(vp))
 			ffs_blkfree(ump, fs, ip->i_devvp, bprev, (long)osize,
-			    ip->i_number);
-		if (nsize < request)
-			ffs_blkfree(ump, fs, ip->i_devvp,
-			    bno + numfrags(fs, nsize),
-			    (long)(request - nsize), ip->i_number);
+			    ip->i_number, NULL);
 		delta = btodb(nsize - osize);
 		if (ip->i_flag & IN_SPACECOUNTED) {
 			UFS_LOCK(ump);
@@ -487,6 +483,14 @@ ffs_reallocblks(ap)
 
 	if (doreallocblks == 0)
 		return (ENOSPC);
+	/*
+	 * We can't wait in softdep prealloc as it may fsync and recurse
+	 * here.  Instead we simply fail to reallocate blocks if this
+	 * rare condition arises.
+	 */
+	if (DOINGSOFTDEP(ap->a_vp))
+		if (softdep_prealloc(ap->a_vp, MNT_NOWAIT) != 0)
+			return (ENOSPC);
 	if (VTOI(ap->a_vp)->i_ump->um_fstype == UFS1)
 		return (ffs_reallocblks_ufs1(ap));
 	return (ffs_reallocblks_ufs2(ap));
@@ -587,7 +591,7 @@ ffs_reallocblks_ufs1(ap)
 	 * Search the block map looking for an allocation of the desired size.
 	 */
 	if ((newblk = ffs_hashalloc(ip, dtog(fs, pref), pref,
-	    len, ffs_clusteralloc)) == 0) {
+	    len, len, ffs_clusteralloc)) == 0) {
 		UFS_UNLOCK(ump);
 		goto fail;
 	}
@@ -673,7 +677,7 @@ ffs_reallocblks_ufs1(ap)
 		if (!DOINGSOFTDEP(vp))
 			ffs_blkfree(ump, fs, ip->i_devvp,
 			    dbtofsb(fs, buflist->bs_children[i]->b_blkno),
-			    fs->fs_bsize, ip->i_number);
+			    fs->fs_bsize, ip->i_number, NULL);
 		buflist->bs_children[i]->b_blkno = fsbtodb(fs, blkno);
 #ifdef INVARIANTS
 		if (!ffs_checkblk(ip,
@@ -795,7 +799,7 @@ ffs_reallocblks_ufs2(ap)
 	 * Search the block map looking for an allocation of the desired size.
 	 */
 	if ((newblk = ffs_hashalloc(ip, dtog(fs, pref), pref,
-	    len, ffs_clusteralloc)) == 0) {
+	    len, len, ffs_clusteralloc)) == 0) {
 		UFS_UNLOCK(ump);
 		goto fail;
 	}
@@ -881,7 +885,7 @@ ffs_reallocblks_ufs2(ap)
 		if (!DOINGSOFTDEP(vp))
 			ffs_blkfree(ump, fs, ip->i_devvp,
 			    dbtofsb(fs, buflist->bs_children[i]->b_blkno),
-			    fs->fs_bsize, ip->i_number);
+			    fs->fs_bsize, ip->i_number, NULL);
 		buflist->bs_children[i]->b_blkno = fsbtodb(fs, blkno);
 #ifdef INVARIANTS
 		if (!ffs_checkblk(ip,
@@ -969,7 +973,7 @@ ffs_valloc(pvp, mode, cred, vpp)
 		if (fs->fs_contigdirs[cg] > 0)
 			fs->fs_contigdirs[cg]--;
 	}
-	ino = (ino_t)ffs_hashalloc(pip, cg, ipref, mode,
+	ino = (ino_t)ffs_hashalloc(pip, cg, ipref, mode, 0,
 					(allocfcn_t *)ffs_nodealloccg);
 	if (ino == 0)
 		goto noinodes;
@@ -1278,11 +1282,12 @@ ffs_blkpref_ufs2(ip, lbn, indx, bap)
  */
 /*VARARGS5*/
 static ufs2_daddr_t
-ffs_hashalloc(ip, cg, pref, size, allocator)
+ffs_hashalloc(ip, cg, pref, size, rsize, allocator)
 	struct inode *ip;
 	u_int cg;
 	ufs2_daddr_t pref;
-	int size;	/* size for data blocks, mode for inodes */
+	int size;	/* Search size for data blocks, mode for inodes */
+	int rsize;	/* Real allocated size. */
 	allocfcn_t *allocator;
 {
 	struct fs *fs;
@@ -1298,7 +1303,7 @@ ffs_hashalloc(ip, cg, pref, size, allocator)
 	/*
 	 * 1: preferred cylinder group
 	 */
-	result = (*allocator)(ip, cg, pref, size);
+	result = (*allocator)(ip, cg, pref, size, rsize);
 	if (result)
 		return (result);
 	/*
@@ -1308,7 +1313,7 @@ ffs_hashalloc(ip, cg, pref, size, allocator)
 		cg += i;
 		if (cg >= fs->fs_ncg)
 			cg -= fs->fs_ncg;
-		result = (*allocator)(ip, cg, 0, size);
+		result = (*allocator)(ip, cg, 0, size, rsize);
 		if (result)
 			return (result);
 	}
@@ -1319,7 +1324,7 @@ ffs_hashalloc(ip, cg, pref, size, allocator)
 	 */
 	cg = (icg + 2) % fs->fs_ncg;
 	for (i = 2; i < fs->fs_ncg; i++) {
-		result = (*allocator)(ip, cg, 0, size);
+		result = (*allocator)(ip, cg, 0, size, rsize);
 		if (result)
 			return (result);
 		cg++;
@@ -1401,7 +1406,8 @@ ffs_fragextend(ip, cg, bprev, osize, nsize)
 	ACTIVECLEAR(fs, cg);
 	UFS_UNLOCK(ump);
 	if (DOINGSOFTDEP(ITOV(ip)))
-		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), bprev);
+		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), bprev,
+		    frags, numfrags(fs, osize));
 	bdwrite(bp);
 	return (bprev);
 
@@ -1419,11 +1425,12 @@ fail:
  * and if it is, allocate it.
  */
 static ufs2_daddr_t
-ffs_alloccg(ip, cg, bpref, size)
+ffs_alloccg(ip, cg, bpref, size, rsize)
 	struct inode *ip;
 	u_int cg;
 	ufs2_daddr_t bpref;
 	int size;
+	int rsize;
 {
 	struct fs *fs;
 	struct cg *cgp;
@@ -1451,7 +1458,7 @@ ffs_alloccg(ip, cg, bpref, size)
 	cgp->cg_old_time = cgp->cg_time = time_second;
 	if (size == fs->fs_bsize) {
 		UFS_LOCK(ump);
-		blkno = ffs_alloccgblk(ip, bp, bpref);
+		blkno = ffs_alloccgblk(ip, bp, bpref, rsize);
 		ACTIVECLEAR(fs, cg);
 		UFS_UNLOCK(ump);
 		bdwrite(bp);
@@ -1475,21 +1482,14 @@ ffs_alloccg(ip, cg, bpref, size)
 		if (cgp->cg_cs.cs_nbfree == 0)
 			goto fail;
 		UFS_LOCK(ump);
-		blkno = ffs_alloccgblk(ip, bp, bpref);
-		bno = dtogd(fs, blkno);
-		for (i = frags; i < fs->fs_frag; i++)
-			setbit(blksfree, bno + i);
-		i = fs->fs_frag - frags;
-		cgp->cg_cs.cs_nffree += i;
-		fs->fs_cstotal.cs_nffree += i;
-		fs->fs_cs(fs, cg).cs_nffree += i;
-		fs->fs_fmod = 1;
-		cgp->cg_frsum[i]++;
+		blkno = ffs_alloccgblk(ip, bp, bpref, rsize);
 		ACTIVECLEAR(fs, cg);
 		UFS_UNLOCK(ump);
 		bdwrite(bp);
 		return (blkno);
 	}
+	KASSERT(size == rsize,
+	    ("ffs_alloccg: size(%d) != rsize(%d)", size, rsize));
 	bno = ffs_mapsearch(fs, cgp, bpref, allocsiz);
 	if (bno < 0)
 		goto fail;
@@ -1507,7 +1507,7 @@ ffs_alloccg(ip, cg, bpref, size)
 	ACTIVECLEAR(fs, cg);
 	UFS_UNLOCK(ump);
 	if (DOINGSOFTDEP(ITOV(ip)))
-		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno);
+		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno, frags, 0);
 	bdwrite(bp);
 	return (blkno);
 
@@ -1529,10 +1529,11 @@ fail:
  * blocks may be fragmented by the routine that allocates them.
  */
 static ufs2_daddr_t
-ffs_alloccgblk(ip, bp, bpref)
+ffs_alloccgblk(ip, bp, bpref, size)
 	struct inode *ip;
 	struct buf *bp;
 	ufs2_daddr_t bpref;
+	int size;
 {
 	struct fs *fs;
 	struct cg *cgp;
@@ -1540,6 +1541,7 @@ ffs_alloccgblk(ip, bp, bpref)
 	ufs1_daddr_t bno;
 	ufs2_daddr_t blkno;
 	u_int8_t *blksfree;
+	int i;
 
 	fs = ip->i_fs;
 	ump = ip->i_ump;
@@ -1567,16 +1569,32 @@ ffs_alloccgblk(ip, bp, bpref)
 gotit:
 	blkno = fragstoblks(fs, bno);
 	ffs_clrblock(fs, blksfree, (long)blkno);
-	ffs_clusteracct(ump, fs, cgp, blkno, -1);
+	ffs_clusteracct(fs, cgp, blkno, -1);
 	cgp->cg_cs.cs_nbfree--;
 	fs->fs_cstotal.cs_nbfree--;
 	fs->fs_cs(fs, cgp->cg_cgx).cs_nbfree--;
 	fs->fs_fmod = 1;
 	blkno = cgbase(fs, cgp->cg_cgx) + bno;
+	/*
+	 * If the caller didn't want the whole block free the frags here.
+	 */
+	size = numfrags(fs, size);
+	if (size != fs->fs_frag) {
+		bno = dtogd(fs, blkno);
+		for (i = size; i < fs->fs_frag; i++)
+			setbit(blksfree, bno + i);
+		i = fs->fs_frag - size;
+		cgp->cg_cs.cs_nffree += i;
+		fs->fs_cstotal.cs_nffree += i;
+		fs->fs_cs(fs, cgp->cg_cgx).cs_nffree += i;
+		fs->fs_fmod = 1;
+		cgp->cg_frsum[i]++;
+	}
 	/* XXX Fixme. */
 	UFS_UNLOCK(ump);
 	if (DOINGSOFTDEP(ITOV(ip)))
-		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno);
+		softdep_setup_blkmapdep(bp, UFSTOVFS(ump), blkno,
+		    size, 0);
 	UFS_LOCK(ump);
 	return (blkno);
 }
@@ -1589,11 +1607,12 @@ gotit:
  * take the first one that we find following bpref.
  */
 static ufs2_daddr_t
-ffs_clusteralloc(ip, cg, bpref, len)
+ffs_clusteralloc(ip, cg, bpref, len, unused)
 	struct inode *ip;
 	u_int cg;
 	ufs2_daddr_t bpref;
 	int len;
+	int unused;
 {
 	struct fs *fs;
 	struct cg *cgp;
@@ -1689,7 +1708,7 @@ ffs_clusteralloc(ip, cg, bpref, len)
 	len = blkstofrags(fs, len);
 	UFS_LOCK(ump);
 	for (i = 0; i < len; i += fs->fs_frag)
-		if (ffs_alloccgblk(ip, bp, bno + i) != bno + i)
+		if (ffs_alloccgblk(ip, bp, bno + i, fs->fs_bsize) != bno + i)
 			panic("ffs_clusteralloc: lost block");
 	ACTIVECLEAR(fs, cg);
 	UFS_UNLOCK(ump);
@@ -1713,11 +1732,12 @@ fail:
  *      inode in the specified cylinder group.
  */
 static ufs2_daddr_t
-ffs_nodealloccg(ip, cg, ipref, mode)
+ffs_nodealloccg(ip, cg, ipref, mode, unused)
 	struct inode *ip;
 	u_int cg;
 	ufs2_daddr_t ipref;
 	int mode;
+	int unused;
 {
 	struct fs *fs;
 	struct cg *cgp;
@@ -1819,28 +1839,6 @@ gotit:
 	return ((ino_t)(cg * fs->fs_ipg + ipref));
 }
 
-/*
- * check if a block is free
- */
-static int
-ffs_isfreeblock(struct fs *fs, u_char *cp, ufs1_daddr_t h)
-{
-
-	switch ((int)fs->fs_frag) {
-	case 8:
-		return (cp[h] == 0);
-	case 4:
-		return ((cp[h >> 1] & (0x0f << ((h & 0x1) << 2))) == 0);
-	case 2:
-		return ((cp[h >> 2] & (0x03 << ((h & 0x3) << 1))) == 0);
-	case 1:
-		return ((cp[h >> 3] & (0x01 << (h & 0x7))) == 0);
-	default:
-		panic("ffs_isfreeblock");
-	}
-	return (0);
-}
-
 /*
  * Free a block or fragment.
  *
@@ -1849,14 +1847,16 @@ ffs_isfreeblock(struct fs *fs, u_char *cp, ufs1_daddr_t h)
  * block reassembly is checked.
  */
 void
-ffs_blkfree(ump, fs, devvp, bno, size, inum)
+ffs_blkfree(ump, fs, devvp, bno, size, inum, dephd)
 	struct ufsmount *ump;
 	struct fs *fs;
 	struct vnode *devvp;
 	ufs2_daddr_t bno;
 	long size;
 	ino_t inum;
+	struct workhead *dephd;
 {
+	struct mount *mp;
 	struct cg *cgp;
 	struct buf *bp;
 	ufs1_daddr_t fragno, cgbno;
@@ -1923,7 +1923,7 @@ ffs_blkfree(ump, fs, devvp, bno, size, inum)
 			panic("ffs_blkfree: freeing free block");
 		}
 		ffs_setblock(fs, blksfree, fragno);
-		ffs_clusteracct(ump, fs, cgp, fragno, 1);
+		ffs_clusteracct(fs, cgp, fragno, 1);
 		cgp->cg_cs.cs_nbfree++;
 		fs->fs_cstotal.cs_nbfree++;
 		fs->fs_cs(fs, cg).cs_nbfree++;
@@ -1963,7 +1963,7 @@ ffs_blkfree(ump, fs, devvp, bno, size, inum)
 			cgp->cg_cs.cs_nffree -= fs->fs_frag;
 			fs->fs_cstotal.cs_nffree -= fs->fs_frag;
 			fs->fs_cs(fs, cg).cs_nffree -= fs->fs_frag;
-			ffs_clusteracct(ump, fs, cgp, fragno, 1);
+			ffs_clusteracct(fs, cgp, fragno, 1);
 			cgp->cg_cs.cs_nbfree++;
 			fs->fs_cstotal.cs_nbfree++;
 			fs->fs_cs(fs, cg).cs_nbfree++;
@@ -1972,6 +1972,10 @@ ffs_blkfree(ump, fs, devvp, bno, size, inum)
 	fs->fs_fmod = 1;
 	ACTIVECLEAR(fs, cg);
 	UFS_UNLOCK(ump);
+	mp = UFSTOVFS(ump);
+	if (mp->mnt_flag & MNT_SOFTDEP && devvp->v_type != VREG)
+		softdep_setup_blkfree(UFSTOVFS(ump), bp, bno,
+		    numfrags(fs, size), dephd);
 	bdwrite(bp);
 }
 
@@ -2042,7 +2046,8 @@ ffs_vfree(pvp, ino, mode)
 		return (0);
 	}
 	ip = VTOI(pvp);
-	return (ffs_freefile(ip->i_ump, ip->i_fs, ip->i_devvp, ino, mode));
+	return (ffs_freefile(ip->i_ump, ip->i_fs, ip->i_devvp, ino, mode,
+	    NULL));
 }
 
 /*
@@ -2050,12 +2055,13 @@ ffs_vfree(pvp, ino, mode)
  * The specified inode is placed back in the free map.
  */
 int
-ffs_freefile(ump, fs, devvp, ino, mode)
+ffs_freefile(ump, fs, devvp, ino, mode, wkhd)
 	struct ufsmount *ump;
 	struct fs *fs;
 	struct vnode *devvp;
 	ino_t ino;
 	int mode;
+	struct workhead *wkhd;
 {
 	struct cg *cgp;
 	struct buf *bp;
@@ -2112,6 +2118,9 @@ ffs_freefile(ump, fs, devvp, ino, mode)
 	fs->fs_fmod = 1;
 	ACTIVECLEAR(fs, cg);
 	UFS_UNLOCK(ump);
+	if (UFSTOVFS(ump)->mnt_flag & MNT_SOFTDEP && devvp->v_type != VREG)
+		softdep_setup_inofree(UFSTOVFS(ump), bp,
+		    ino + cg * fs->fs_ipg, wkhd);
 	bdwrite(bp);
 	return (0);
 }
@@ -2225,101 +2234,6 @@ ffs_mapsearch(fs, cgp, bpref, allocsiz)
 	return (-1);
 }
 
-/*
- * Update the cluster map because of an allocation or free.
- *
- * Cnt == 1 means free; cnt == -1 means allocating.
- */
-void
-ffs_clusteracct(ump, fs, cgp, blkno, cnt)
-	struct ufsmount *ump;
-	struct fs *fs;
-	struct cg *cgp;
-	ufs1_daddr_t blkno;
-	int cnt;
-{
-	int32_t *sump;
-	int32_t *lp;
-	u_char *freemapp, *mapp;
-	int i, start, end, forw, back, map, bit;
-
-	mtx_assert(UFS_MTX(ump), MA_OWNED);
-
-	if (fs->fs_contigsumsize <= 0)
-		return;
-	freemapp = cg_clustersfree(cgp);
-	sump = cg_clustersum(cgp);
-	/*
-	 * Allocate or clear the actual block.
-	 */
-	if (cnt > 0)
-		setbit(freemapp, blkno);
-	else
-		clrbit(freemapp, blkno);
-	/*
-	 * Find the size of the cluster going forward.
-	 */
-	start = blkno + 1;
-	end = start + fs->fs_contigsumsize;
-	if (end >= cgp->cg_nclusterblks)
-		end = cgp->cg_nclusterblks;
-	mapp = &freemapp[start / NBBY];
-	map = *mapp++;
-	bit = 1 << (start % NBBY);
-	for (i = start; i < end; i++) {
-		if ((map & bit) == 0)
-			break;
-		if ((i & (NBBY - 1)) != (NBBY - 1)) {
-			bit <<= 1;
-		} else {
-			map = *mapp++;
-			bit = 1;
-		}
-	}
-	forw = i - start;
-	/*
-	 * Find the size of the cluster going backward.
-	 */
-	start = blkno - 1;
-	end = start - fs->fs_contigsumsize;
-	if (end < 0)
-		end = -1;
-	mapp = &freemapp[start / NBBY];
-	map = *mapp--;
-	bit = 1 << (start % NBBY);
-	for (i = start; i > end; i--) {
-		if ((map & bit) == 0)
-			break;
-		if ((i & (NBBY - 1)) != 0) {
-			bit >>= 1;
-		} else {
-			map = *mapp--;
-			bit = 1 << (NBBY - 1);
-		}
-	}
-	back = start - i;
-	/*
-	 * Account for old cluster and the possibly new forward and
-	 * back clusters.
-	 */
-	i = back + forw + 1;
-	if (i > fs->fs_contigsumsize)
-		i = fs->fs_contigsumsize;
-	sump[i] += cnt;
-	if (back > 0)
-		sump[back] -= cnt;
-	if (forw > 0)
-		sump[forw] -= cnt;
-	/*
-	 * Update cluster summary information.
-	 */
-	lp = &sump[fs->fs_contigsumsize];
-	for (i = fs->fs_contigsumsize; i > 0; i--)
-		if (*lp-- > 0)
-			break;
-	fs->fs_maxcluster[cgp->cg_cgx] = i;
-}
-
 /*
  * Fserr prints the name of a filesystem with an error diagnostic.
  *
@@ -2540,7 +2454,7 @@ sysctl_ffs_fsck(SYSCTL_HANDLER_ARGS)
 #endif /* DEBUG */
 		while (cmd.size > 0) {
 			if ((error = ffs_freefile(ump, fs, ump->um_devvp,
-			    cmd.value, filetype)))
+			    cmd.value, filetype, NULL)))
 				break;
 			cmd.size -= 1;
 			cmd.value += 1;
@@ -2568,7 +2482,7 @@ sysctl_ffs_fsck(SYSCTL_HANDLER_ARGS)
 			if (blksize > blkcnt)
 				blksize = blkcnt;
 			ffs_blkfree(ump, fs, ump->um_devvp, blkno,
-			    blksize * fs->fs_fsize, ROOTINO);
+			    blksize * fs->fs_fsize, ROOTINO, NULL);
 			blkno += blksize;
 			blkcnt -= blksize;
 			blksize = fs->fs_frag;
diff --git a/sys/ufs/ffs/ffs_balloc.c b/sys/ufs/ffs/ffs_balloc.c
index a12f96e60d0..6d5f27c1f30 100644
--- a/sys/ufs/ffs/ffs_balloc.c
+++ b/sys/ufs/ffs/ffs_balloc.c
@@ -120,6 +120,8 @@ ffs_balloc_ufs1(struct vnode *vp, off_t startoffset, int size,
 	if (lbn < 0)
 		return (EFBIG);
 
+	if (DOINGSOFTDEP(vp))
+		softdep_prealloc(vp, MNT_WAIT);
 	/*
 	 * If the next write will extend the file into a new block,
 	 * and the file is currently composed of a fragment
@@ -418,6 +420,8 @@ fail:
 	 * slow, running out of disk space is not expected to be a common
 	 * occurence. The error return from fsync is ignored as we already
 	 * have an error to return to the user.
+	 *
+	 * XXX Still have to journal the free below
 	 */
 	(void) ffs_syncvnode(vp, MNT_WAIT);
 	for (deallocated = 0, blkp = allociblk, lbns_remfree = lbns;
@@ -473,7 +477,7 @@ fail:
 	 */
 	for (blkp = allociblk; blkp < allocblk; blkp++) {
 		ffs_blkfree(ump, fs, ip->i_devvp, *blkp, fs->fs_bsize,
-		    ip->i_number);
+		    ip->i_number, NULL);
 	}
 	return (error);
 }
@@ -515,6 +519,9 @@ ffs_balloc_ufs2(struct vnode *vp, off_t startoffset, int size,
 	if (lbn < 0)
 		return (EFBIG);
 
+	if (DOINGSOFTDEP(vp))
+		softdep_prealloc(vp, MNT_WAIT);
+	
 	/*
 	 * Check for allocating external data.
 	 */
@@ -930,6 +937,8 @@ fail:
 	 * slow, running out of disk space is not expected to be a common
 	 * occurence. The error return from fsync is ignored as we already
 	 * have an error to return to the user.
+	 *
+	 * XXX Still have to journal the free below
 	 */
 	(void) ffs_syncvnode(vp, MNT_WAIT);
 	for (deallocated = 0, blkp = allociblk, lbns_remfree = lbns;
@@ -985,7 +994,7 @@ fail:
 	 */
 	for (blkp = allociblk; blkp < allocblk; blkp++) {
 		ffs_blkfree(ump, fs, ip->i_devvp, *blkp, fs->fs_bsize,
-		    ip->i_number);
+		    ip->i_number, NULL);
 	}
 	return (error);
 }
diff --git a/sys/ufs/ffs/ffs_extern.h b/sys/ufs/ffs/ffs_extern.h
index 7e32ced2ebe..7011623749b 100644
--- a/sys/ufs/ffs/ffs_extern.h
+++ b/sys/ufs/ffs/ffs_extern.h
@@ -47,6 +47,7 @@ struct ucred;
 struct vnode;
 struct vop_fsync_args;
 struct vop_reallocblks_args;
+struct workhead;
 
 int	ffs_alloc(struct inode *, ufs2_daddr_t, ufs2_daddr_t, int, int,
 	    struct ucred *, ufs2_daddr_t *);
@@ -56,20 +57,23 @@ int	ffs_balloc_ufs2(struct vnode *a_vp, off_t a_startoffset, int a_size,
             struct ucred *a_cred, int a_flags, struct buf **a_bpp);
 int	ffs_blkatoff(struct vnode *, off_t, char **, struct buf **);
 void	ffs_blkfree(struct ufsmount *, struct fs *, struct vnode *,
-	    ufs2_daddr_t, long, ino_t);
+	    ufs2_daddr_t, long, ino_t, struct workhead *);
 ufs2_daddr_t ffs_blkpref_ufs1(struct inode *, ufs_lbn_t, int, ufs1_daddr_t *);
 ufs2_daddr_t ffs_blkpref_ufs2(struct inode *, ufs_lbn_t, int, ufs2_daddr_t *);
 int	ffs_checkfreefile(struct fs *, struct vnode *, ino_t);
 void	ffs_clrblock(struct fs *, u_char *, ufs1_daddr_t);
+void	ffs_clusteracct(struct fs *, struct cg *, ufs1_daddr_t, int);
 void	ffs_bdflush(struct bufobj *, struct buf *);
 int	ffs_copyonwrite(struct vnode *, struct buf *);
 int	ffs_flushfiles(struct mount *, int, struct thread *);
 void	ffs_fragacct(struct fs *, int, int32_t [], int);
 int	ffs_freefile(struct ufsmount *, struct fs *, struct vnode *, ino_t,
-	    int);
+	    int, struct workhead *);
 int	ffs_isblock(struct fs *, u_char *, ufs1_daddr_t);
+int	ffs_isfreeblock(struct fs *, u_char *, ufs1_daddr_t);
 void	ffs_load_inode(struct buf *, struct inode *, struct fs *, ino_t);
 int	ffs_mountroot(void);
+void	ffs_oldfscompat_write(struct fs *, struct ufsmount *);
 int	ffs_reallocblks(struct vop_reallocblks_args *);
 int	ffs_realloccg(struct inode *, ufs2_daddr_t, ufs2_daddr_t,
 	    ufs2_daddr_t, int, int, int, struct ucred *, struct buf **);
@@ -103,12 +107,14 @@ extern struct vop_vector ffs_fifoops2;
 
 int	softdep_check_suspend(struct mount *, struct vnode *,
 	  int, int, int, int);
+int	softdep_complete_trunc(struct vnode *, void *);
 void	softdep_get_depcounts(struct mount *, int *, int *);
 void	softdep_initialize(void);
 void	softdep_uninitialize(void);
 int	softdep_mount(struct vnode *, struct mount *, struct fs *,
 	    struct ucred *);
-void	softdep_move_dependencies(struct buf *, struct buf *);
+void	softdep_unmount(struct mount *);
+int	softdep_move_dependencies(struct buf *, struct buf *);
 int	softdep_flushworklist(struct mount *, int *, struct thread *);
 int	softdep_flushfiles(struct mount *, int, struct thread *);
 void	softdep_update_inodeblock(struct inode *, struct buf *, int);
@@ -117,7 +123,8 @@ void	softdep_freefile(struct vnode *, ino_t, int);
 int	softdep_request_cleanup(struct fs *, struct vnode *);
 void	softdep_setup_freeblocks(struct inode *, off_t, int);
 void	softdep_setup_inomapdep(struct buf *, struct inode *, ino_t);
-void	softdep_setup_blkmapdep(struct buf *, struct mount *, ufs2_daddr_t);
+void	softdep_setup_blkmapdep(struct buf *, struct mount *, ufs2_daddr_t,
+	    int, int);
 void	softdep_setup_allocdirect(struct inode *, ufs_lbn_t, ufs2_daddr_t,
 	    ufs2_daddr_t, long, long, struct buf *);
 void	softdep_setup_allocext(struct inode *, ufs_lbn_t, ufs2_daddr_t,
@@ -126,11 +133,20 @@ void	softdep_setup_allocindir_meta(struct buf *, struct inode *,
 	    struct buf *, int, ufs2_daddr_t);
 void	softdep_setup_allocindir_page(struct inode *, ufs_lbn_t,
 	    struct buf *, int, ufs2_daddr_t, ufs2_daddr_t, struct buf *);
+void	softdep_setup_blkfree(struct mount *, struct buf *, ufs2_daddr_t, int,
+	    struct workhead *);
+void	softdep_setup_inofree(struct mount *, struct buf *, ino_t,
+	    struct workhead *);
+void	softdep_setup_sbupdate(struct ufsmount *, struct fs *, struct buf *);
+void 	*softdep_setup_trunc(struct vnode *vp, off_t length, int flags);
 void	softdep_fsync_mountdev(struct vnode *);
 int	softdep_sync_metadata(struct vnode *);
 int     softdep_process_worklist(struct mount *, int);
 int     softdep_fsync(struct vnode *);
 int	softdep_waitidle(struct mount *);
+int	softdep_prealloc(struct vnode *, int);
+int	softdep_journal_lookup(struct mount *, struct vnode **);
+
 
 int	ffs_rdonly(struct inode *);
 
diff --git a/sys/ufs/ffs/ffs_inode.c b/sys/ufs/ffs/ffs_inode.c
index b2f90673012..3b6983258b9 100644
--- a/sys/ufs/ffs/ffs_inode.c
+++ b/sys/ufs/ffs/ffs_inode.c
@@ -92,15 +92,6 @@ ffs_update(vp, waitfor)
 	fs = ip->i_fs;
 	if (fs->fs_ronly)
 		return (0);
-	/*
-	 * Ensure that uid and gid are correct. This is a temporary
-	 * fix until fsck has been changed to do the update.
-	 */
-	if (fs->fs_magic == FS_UFS1_MAGIC &&		/* XXX */
-	    fs->fs_old_inodefmt < FS_44INODEFMT) {	/* XXX */
-		ip->i_din1->di_ouid = ip->i_uid;	/* XXX */
-		ip->i_din1->di_ogid = ip->i_gid;	/* XXX */
-	}						/* XXX */
 	error = bread(ip->i_devvp, fsbtodb(fs, ino_to_fsba(fs, ip->i_number)),
 		(int)fs->fs_bsize, NOCRED, &bp);
 	if (error) {
@@ -160,6 +151,7 @@ ffs_truncate(vp, length, flags, cred, td)
 	ufs2_daddr_t bn, lbn, lastblock, lastiblock[NIADDR], indir_lbn[NIADDR];
 	ufs2_daddr_t oldblks[NDADDR + NIADDR], newblks[NDADDR + NIADDR];
 	ufs2_daddr_t count, blocksreleased = 0, datablocks;
+	void *cookie;
 	struct bufobj *bo;
 	struct fs *fs;
 	struct buf *bp;
@@ -173,11 +165,14 @@ ffs_truncate(vp, length, flags, cred, td)
 	fs = ip->i_fs;
 	ump = ip->i_ump;
 	bo = &vp->v_bufobj;
+	cookie = NULL;
 
 	ASSERT_VOP_LOCKED(vp, "ffs_truncate");
 
 	if (length < 0)
 		return (EINVAL);
+	if (length > fs->fs_maxfilesize)
+		return (EFBIG);
 	/*
 	 * Historically clients did not have to specify which data
 	 * they were truncating. So, if not specified, we assume
@@ -192,6 +187,7 @@ ffs_truncate(vp, length, flags, cred, td)
 	 * (e.g., the file is being unlinked), then pick it off with
 	 * soft updates below.
 	 */
+	allerror = 0;
 	needextclean = 0;
 	softdepslowdown = DOINGSOFTDEP(vp) && softdep_slowdown(vp);
 	extblocks = 0;
@@ -212,6 +208,8 @@ ffs_truncate(vp, length, flags, cred, td)
 				panic("ffs_truncate: partial trunc of extdata");
 			if ((error = ffs_syncvnode(vp, MNT_WAIT)) != 0)
 				return (error);
+			if (DOINGSUJ(vp))
+				cookie = softdep_setup_trunc(vp, length, flags);
 			osize = ip->i_din2->di_extsize;
 			ip->i_din2->di_blocks -= extblocks;
 #ifdef QUOTA
@@ -227,19 +225,19 @@ ffs_truncate(vp, length, flags, cred, td)
 			}
 			ip->i_flag |= IN_CHANGE;
 			if ((error = ffs_update(vp, 1)))
-				return (error);
+				goto out;
 			for (i = 0; i < NXADDR; i++) {
 				if (oldblks[i] == 0)
 					continue;
 				ffs_blkfree(ump, fs, ip->i_devvp, oldblks[i],
-				    sblksize(fs, osize, i), ip->i_number);
+				    sblksize(fs, osize, i), ip->i_number, NULL);
 			}
 		}
 	}
-	if ((flags & IO_NORMAL) == 0)
-		return (0);
-	if (length > fs->fs_maxfilesize)
-		return (EFBIG);
+	if ((flags & IO_NORMAL) == 0) {
+		error = 0;
+		goto out;
+	}
 	if (vp->v_type == VLNK &&
 	    (ip->i_size < vp->v_mount->mnt_maxsymlinklen ||
 	     datablocks == 0)) {
@@ -253,24 +251,52 @@ ffs_truncate(vp, length, flags, cred, td)
 		ip->i_flag |= IN_CHANGE | IN_UPDATE;
 		if (needextclean)
 			softdep_setup_freeblocks(ip, length, IO_EXT);
-		return (ffs_update(vp, 1));
+		error = ffs_update(vp, 1);
+		goto out;
 	}
 	if (ip->i_size == length) {
 		ip->i_flag |= IN_CHANGE | IN_UPDATE;
 		if (needextclean)
 			softdep_setup_freeblocks(ip, length, IO_EXT);
-		return (ffs_update(vp, 0));
+		error = ffs_update(vp, 0);
+		goto out;
 	}
 	if (fs->fs_ronly)
 		panic("ffs_truncate: read-only filesystem");
 #ifdef QUOTA
 	error = getinoquota(ip);
 	if (error)
-		return (error);
+		goto out;
 #endif
 	if ((ip->i_flags & SF_SNAPSHOT) != 0)
 		ffs_snapremove(vp);
 	vp->v_lasta = vp->v_clen = vp->v_cstart = vp->v_lastw = 0;
+	osize = ip->i_size;
+	/*
+	 * Lengthen the size of the file. We must ensure that the
+	 * last byte of the file is allocated. Since the smallest
+	 * value of osize is 0, length will be at least 1.
+	 */
+	if (osize < length) {
+		vnode_pager_setsize(vp, length);
+		flags |= BA_CLRBUF;
+		error = UFS_BALLOC(vp, length - 1, 1, cred, flags, &bp);
+		if (error) {
+			vnode_pager_setsize(vp, osize);
+			goto out;
+		}
+		ip->i_size = length;
+		DIP_SET(ip, i_size, length);
+		if (bp->b_bufsize == fs->fs_bsize)
+			bp->b_flags |= B_CLUSTEROK;
+		if (flags & IO_SYNC)
+			bwrite(bp);
+		else
+			bawrite(bp);
+		ip->i_flag |= IN_CHANGE | IN_UPDATE;
+		error = ffs_update(vp, 1);
+		goto out;
+	}
 	if (DOINGSOFTDEP(vp)) {
 		if (length > 0 || softdepslowdown) {
 			/*
@@ -283,11 +309,18 @@ ffs_truncate(vp, length, flags, cred, td)
 			 * so that it will have no data structures left.
 			 */
 			if ((error = ffs_syncvnode(vp, MNT_WAIT)) != 0)
-				return (error);
+				goto out;
 			UFS_LOCK(ump);
 			if (ip->i_flag & IN_SPACECOUNTED)
 				fs->fs_pendingblocks -= datablocks;
 			UFS_UNLOCK(ump);
+			/*
+			 * We have to journal the truncation before we change
+			 * any blocks so we don't leave the file partially
+			 * truncated.
+			 */
+			if (DOINGSUJ(vp) && cookie == NULL)
+				cookie = softdep_setup_trunc(vp, length, flags);
 		} else {
 #ifdef QUOTA
 			(void) chkdq(ip, -datablocks, NOCRED, 0);
@@ -301,34 +334,10 @@ ffs_truncate(vp, length, flags, cred, td)
 				    OFF_TO_IDX(lblktosize(fs, -extblocks)));
 			vnode_pager_setsize(vp, 0);
 			ip->i_flag |= IN_CHANGE | IN_UPDATE;
-			return (ffs_update(vp, 0));
+			error = ffs_update(vp, 0);
+			goto out;
 		}
 	}
-	osize = ip->i_size;
-	/*
-	 * Lengthen the size of the file. We must ensure that the
-	 * last byte of the file is allocated. Since the smallest
-	 * value of osize is 0, length will be at least 1.
-	 */
-	if (osize < length) {
-		vnode_pager_setsize(vp, length);
-		flags |= BA_CLRBUF;
-		error = UFS_BALLOC(vp, length - 1, 1, cred, flags, &bp);
-		if (error) {
-			vnode_pager_setsize(vp, osize);
-			return (error);
-		}
-		ip->i_size = length;
-		DIP_SET(ip, i_size, length);
-		if (bp->b_bufsize == fs->fs_bsize)
-			bp->b_flags |= B_CLUSTEROK;
-		if (flags & IO_SYNC)
-			bwrite(bp);
-		else
-			bawrite(bp);
-		ip->i_flag |= IN_CHANGE | IN_UPDATE;
-		return (ffs_update(vp, 1));
-	}
 	/*
 	 * Shorten the size of the file. If the file is not being
 	 * truncated to a block boundary, the contents of the
@@ -345,9 +354,8 @@ ffs_truncate(vp, length, flags, cred, td)
 		lbn = lblkno(fs, length);
 		flags |= BA_CLRBUF;
 		error = UFS_BALLOC(vp, length - 1, 1, cred, flags, &bp);
-		if (error) {
-			return (error);
-		}
+		if (error)
+			goto out;
 		/*
 		 * When we are doing soft updates and the UFS_BALLOC
 		 * above fills in a direct block hole with a full sized
@@ -359,7 +367,7 @@ ffs_truncate(vp, length, flags, cred, td)
 		if (DOINGSOFTDEP(vp) && lbn < NDADDR &&
 		    fragroundup(fs, blkoff(fs, length)) < fs->fs_bsize &&
 		    (error = ffs_syncvnode(vp, MNT_WAIT)) != 0)
-			return (error);
+			goto out;
 		ip->i_size = length;
 		DIP_SET(ip, i_size, length);
 		size = blksize(fs, ip, lbn);
@@ -405,7 +413,13 @@ ffs_truncate(vp, length, flags, cred, td)
 			DIP_SET(ip, i_db[i], 0);
 	}
 	ip->i_flag |= IN_CHANGE | IN_UPDATE;
-	allerror = ffs_update(vp, 1);
+	/*
+	 * When doing softupdate journaling we must preserve the size along
+	 * with the old pointers until they are freed or we might not
+	 * know how many fragments remain.
+	 */
+	if (!DOINGSUJ(vp))
+		allerror = ffs_update(vp, 1);
 	
 	/*
 	 * Having written the new inode to disk, save its new configuration
@@ -445,7 +459,7 @@ ffs_truncate(vp, length, flags, cred, td)
 			if (lastiblock[level] < 0) {
 				DIP_SET(ip, i_ib[level], 0);
 				ffs_blkfree(ump, fs, ip->i_devvp, bn,
-				    fs->fs_bsize, ip->i_number);
+				    fs->fs_bsize, ip->i_number, NULL);
 				blocksreleased += nblocks;
 			}
 		}
@@ -464,7 +478,8 @@ ffs_truncate(vp, length, flags, cred, td)
 			continue;
 		DIP_SET(ip, i_db[i], 0);
 		bsize = blksize(fs, ip, i);
-		ffs_blkfree(ump, fs, ip->i_devvp, bn, bsize, ip->i_number);
+		ffs_blkfree(ump, fs, ip->i_devvp, bn, bsize, ip->i_number,
+		    NULL);
 		blocksreleased += btodb(bsize);
 	}
 	if (lastblock < 0)
@@ -496,7 +511,7 @@ ffs_truncate(vp, length, flags, cred, td)
 			 */
 			bn += numfrags(fs, newspace);
 			ffs_blkfree(ump, fs, ip->i_devvp, bn,
-			    oldspace - newspace, ip->i_number);
+			    oldspace - newspace, ip->i_number, NULL);
 			blocksreleased += btodb(oldspace - newspace);
 		}
 	}
@@ -528,7 +543,14 @@ done:
 #ifdef QUOTA
 	(void) chkdq(ip, -blocksreleased, NOCRED, 0);
 #endif
-	return (allerror);
+	error = allerror;
+out:
+	if (cookie) {
+		allerror = softdep_complete_trunc(vp, cookie);
+		if (allerror != 0 && error == 0)
+			error = allerror;
+	}
+	return (error);
 }
 
 /*
@@ -638,7 +660,7 @@ ffs_indirtrunc(ip, lbn, dbn, lastbn, level, countp)
 			blocksreleased += blkcount;
 		}
 		ffs_blkfree(ip->i_ump, fs, ip->i_devvp, nb, fs->fs_bsize,
-		    ip->i_number);
+		    ip->i_number, NULL);
 		blocksreleased += nblocks;
 	}
 
diff --git a/sys/ufs/ffs/ffs_snapshot.c b/sys/ufs/ffs/ffs_snapshot.c
index b36cb58808b..11362cfbc75 100644
--- a/sys/ufs/ffs/ffs_snapshot.c
+++ b/sys/ufs/ffs/ffs_snapshot.c
@@ -142,7 +142,7 @@ MTX_SYSINIT(ffs_snapfree, &snapfree_lock, "snapdata free list", MTX_DEF);
 static int cgaccount(int, struct vnode *, struct buf *, int);
 static int expunge_ufs1(struct vnode *, struct inode *, struct fs *,
     int (*)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *, struct fs *,
-    ufs_lbn_t, int), int);
+    ufs_lbn_t, int), int, int);
 static int indiracct_ufs1(struct vnode *, struct vnode *, int,
     ufs1_daddr_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, struct fs *,
     int (*)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *, struct fs *,
@@ -155,7 +155,7 @@ static int mapacct_ufs1(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
     struct fs *, ufs_lbn_t, int);
 static int expunge_ufs2(struct vnode *, struct inode *, struct fs *,
     int (*)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *, struct fs *,
-    ufs_lbn_t, int), int);
+    ufs_lbn_t, int), int, int);
 static int indiracct_ufs2(struct vnode *, struct vnode *, int,
     ufs2_daddr_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, ufs_lbn_t, struct fs *,
     int (*)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *, struct fs *,
@@ -582,7 +582,8 @@ loop:
 			len = fragroundup(fs, blkoff(fs, xp->i_size));
 			if (len != 0 && len < fs->fs_bsize) {
 				ffs_blkfree(ump, copy_fs, vp,
-				    DIP(xp, i_db[loc]), len, xp->i_number);
+				    DIP(xp, i_db[loc]), len, xp->i_number,
+				    NULL);
 				blkno = DIP(xp, i_db[loc]);
 				DIP_SET(xp, i_db[loc], 0);
 			}
@@ -590,15 +591,15 @@ loop:
 		snaplistsize += 1;
 		if (xp->i_ump->um_fstype == UFS1)
 			error = expunge_ufs1(vp, xp, copy_fs, fullacct_ufs1,
-			    BLK_NOCOPY);
+			    BLK_NOCOPY, 1);
 		else
 			error = expunge_ufs2(vp, xp, copy_fs, fullacct_ufs2,
-			    BLK_NOCOPY);
+			    BLK_NOCOPY, 1);
 		if (blkno)
 			DIP_SET(xp, i_db[loc], blkno);
 		if (!error)
 			error = ffs_freefile(ump, copy_fs, vp, xp->i_number,
-			    xp->i_mode);
+			    xp->i_mode, NULL);
 		VOP_UNLOCK(xvp, 0);
 		vdrop(xvp);
 		if (error) {
@@ -611,6 +612,26 @@ loop:
 		MNT_ILOCK(mp);
 	}
 	MNT_IUNLOCK(mp);
+	/*
+	 * Erase the journal file from the snapshot.
+	 */
+	if (fs->fs_flags & FS_SUJ) {
+		error = softdep_journal_lookup(mp, &xvp);
+		if (error) {
+			free(copy_fs->fs_csp, M_UFSMNT);
+			bawrite(sbp);
+			sbp = NULL;
+			goto out1;
+		}
+		xp = VTOI(xvp);
+		if (xp->i_ump->um_fstype == UFS1)
+			error = expunge_ufs1(vp, xp, copy_fs, fullacct_ufs1,
+			    BLK_NOCOPY, 0);
+		else
+			error = expunge_ufs2(vp, xp, copy_fs, fullacct_ufs2,
+			    BLK_NOCOPY, 0);
+		vput(xvp);
+	}
 	/*
 	 * Acquire a lock on the snapdata structure, creating it if necessary.
 	 */
@@ -691,16 +712,16 @@ out1:
 			break;
 		if (xp->i_ump->um_fstype == UFS1)
 			error = expunge_ufs1(vp, xp, fs, snapacct_ufs1,
-			    BLK_SNAP);
+			    BLK_SNAP, 0);
 		else
 			error = expunge_ufs2(vp, xp, fs, snapacct_ufs2,
-			    BLK_SNAP);
+			    BLK_SNAP, 0);
 		if (error == 0 && xp->i_effnlink == 0) {
 			error = ffs_freefile(ump,
 					     copy_fs,
 					     vp,
 					     xp->i_number,
-					     xp->i_mode);
+					     xp->i_mode, NULL);
 		}
 		if (error) {
 			fs->fs_snapinum[snaploc] = 0;
@@ -719,9 +740,11 @@ out1:
 	 * the list of allocated blocks in i_snapblklist.
 	 */
 	if (ip->i_ump->um_fstype == UFS1)
-		error = expunge_ufs1(vp, ip, copy_fs, mapacct_ufs1, BLK_SNAP);
+		error = expunge_ufs1(vp, ip, copy_fs, mapacct_ufs1,
+		    BLK_SNAP, 0);
 	else
-		error = expunge_ufs2(vp, ip, copy_fs, mapacct_ufs2, BLK_SNAP);
+		error = expunge_ufs2(vp, ip, copy_fs, mapacct_ufs2,
+		    BLK_SNAP, 0);
 	if (error) {
 		fs->fs_snapinum[snaploc] = 0;
 		free(snapblklist, M_UFSMNT);
@@ -954,13 +977,14 @@ cgaccount(cg, vp, nbp, passno)
  * is reproduced once each for UFS1 and UFS2.
  */
 static int
-expunge_ufs1(snapvp, cancelip, fs, acctfunc, expungetype)
+expunge_ufs1(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
 	struct vnode *snapvp;
 	struct inode *cancelip;
 	struct fs *fs;
 	int (*acctfunc)(struct vnode *, ufs1_daddr_t *, ufs1_daddr_t *,
 	    struct fs *, ufs_lbn_t, int);
 	int expungetype;
+	int clearmode;
 {
 	int i, error, indiroff;
 	ufs_lbn_t lbn, rlbn;
@@ -1005,7 +1029,7 @@ expunge_ufs1(snapvp, cancelip, fs, acctfunc, expungetype)
 	 */
 	dip = (struct ufs1_dinode *)bp->b_data +
 	    ino_to_fsbo(fs, cancelip->i_number);
-	if (expungetype == BLK_NOCOPY || cancelip->i_effnlink == 0)
+	if (clearmode || cancelip->i_effnlink == 0)
 		dip->di_mode = 0;
 	dip->di_size = 0;
 	dip->di_blocks = 0;
@@ -1220,7 +1244,7 @@ mapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
 			*ip->i_snapblklist++ = lblkno;
 		if (blkno == BLK_SNAP)
 			blkno = blkstofrags(fs, lblkno);
-		ffs_blkfree(ip->i_ump, fs, vp, blkno, fs->fs_bsize, inum);
+		ffs_blkfree(ip->i_ump, fs, vp, blkno, fs->fs_bsize, inum, NULL);
 	}
 	return (0);
 }
@@ -1234,13 +1258,14 @@ mapacct_ufs1(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
  * is reproduced once each for UFS1 and UFS2.
  */
 static int
-expunge_ufs2(snapvp, cancelip, fs, acctfunc, expungetype)
+expunge_ufs2(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
 	struct vnode *snapvp;
 	struct inode *cancelip;
 	struct fs *fs;
 	int (*acctfunc)(struct vnode *, ufs2_daddr_t *, ufs2_daddr_t *,
 	    struct fs *, ufs_lbn_t, int);
 	int expungetype;
+	int clearmode;
 {
 	int i, error, indiroff;
 	ufs_lbn_t lbn, rlbn;
@@ -1285,7 +1310,7 @@ expunge_ufs2(snapvp, cancelip, fs, acctfunc, expungetype)
 	 */
 	dip = (struct ufs2_dinode *)bp->b_data +
 	    ino_to_fsbo(fs, cancelip->i_number);
-	if (expungetype == BLK_NOCOPY)
+	if (clearmode || cancelip->i_effnlink == 0)
 		dip->di_mode = 0;
 	dip->di_size = 0;
 	dip->di_blocks = 0;
@@ -1500,7 +1525,7 @@ mapacct_ufs2(vp, oldblkp, lastblkp, fs, lblkno, expungetype)
 			*ip->i_snapblklist++ = lblkno;
 		if (blkno == BLK_SNAP)
 			blkno = blkstofrags(fs, lblkno);
-		ffs_blkfree(ip->i_ump, fs, vp, blkno, fs->fs_bsize, inum);
+		ffs_blkfree(ip->i_ump, fs, vp, blkno, fs->fs_bsize, inum, NULL);
 	}
 	return (0);
 }
@@ -1657,6 +1682,13 @@ ffs_snapremove(vp)
 	ip->i_flags &= ~SF_SNAPSHOT;
 	DIP_SET(ip, i_flags, ip->i_flags);
 	ip->i_flag |= IN_CHANGE | IN_UPDATE;
+	/*
+	 * The dirtied indirects must be written out before
+	 * softdep_setup_freeblocks() is called.  Otherwise indir_trunc()
+	 * may find indirect pointers using the magic BLK_* values.
+	 */
+	if (DOINGSOFTDEP(vp))
+		ffs_syncvnode(vp, MNT_WAIT);
 #ifdef QUOTA
 	/*
 	 * Reenable disk quotas for ex-snapshot file.
diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index 4d652c114dd..30937051357 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -1,5 +1,7 @@
 /*-
- * Copyright 1998, 2000 Marshall Kirk McKusick. All Rights Reserved.
+ * Copyright 1998, 2000 Marshall Kirk McKusick.
+ * Copyright 2009, 2010 Jeffrey W. Roberson <jeff@FreeBSD.org>
+ * All rights reserved.
  *
  * The soft updates code is derived from the appendix of a University
  * of Michigan technical report (Gregory R. Ganger and Yale N. Patt,
@@ -23,17 +25,16 @@
  *    notice, this list of conditions and the following disclaimer in the
  *    documentation and/or other materials provided with the distribution.
  *
- * THIS SOFTWARE IS PROVIDED BY MARSHALL KIRK MCKUSICK ``AS IS'' AND ANY
- * EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
- * WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
- * DISCLAIMED.  IN NO EVENT SHALL MARSHALL KIRK MCKUSICK BE LIABLE FOR
- * ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHORS ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHORS BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
+ * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
+ * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
+ * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
+ * TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
+ * USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  *
  *	from: @(#)ffs_softdep.c	9.59 (McKusick) 6/21/00
  */
@@ -50,6 +51,7 @@ __FBSDID("$FreeBSD$");
 #ifndef DEBUG
 #define DEBUG
 #endif
+#define	SUJ_DEBUG
 
 #include <sys/param.h>
 #include <sys/kernel.h>
@@ -62,6 +64,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/malloc.h>
 #include <sys/mount.h>
 #include <sys/mutex.h>
+#include <sys/namei.h>
 #include <sys/proc.h>
 #include <sys/stat.h>
 #include <sys/sysctl.h>
@@ -130,10 +133,12 @@ softdep_setup_inomapdep(bp, ip, newinum)
 }
 
 void
-softdep_setup_blkmapdep(bp, mp, newblkno)
+softdep_setup_blkmapdep(bp, mp, newblkno, frags, oldfrags)
 	struct buf *bp;
 	struct mount *mp;
 	ufs2_daddr_t newblkno;
+	int frags;
+	int oldfrags;
 {
 
 	panic("softdep_setup_blkmapdep called");
@@ -403,31 +408,13 @@ softdep_get_depcounts(struct mount *mp,
  * These definitions need to be adapted to the system to which
  * this file is being ported.
  */
-/*
- * malloc types defined for the softdep system.
- */
-static MALLOC_DEFINE(M_PAGEDEP, "pagedep","File page dependencies");
-static MALLOC_DEFINE(M_INODEDEP, "inodedep","Inode dependencies");
-static MALLOC_DEFINE(M_NEWBLK, "newblk","New block allocation");
-static MALLOC_DEFINE(M_BMSAFEMAP, "bmsafemap","Block or frag allocated from cyl group map");
-static MALLOC_DEFINE(M_ALLOCDIRECT, "allocdirect","Block or frag dependency for an inode");
-static MALLOC_DEFINE(M_INDIRDEP, "indirdep","Indirect block dependencies");
-static MALLOC_DEFINE(M_ALLOCINDIR, "allocindir","Block dependency for an indirect block");
-static MALLOC_DEFINE(M_FREEFRAG, "freefrag","Previously used frag for an inode");
-static MALLOC_DEFINE(M_FREEBLKS, "freeblks","Blocks freed from an inode");
-static MALLOC_DEFINE(M_FREEFILE, "freefile","Inode deallocated");
-static MALLOC_DEFINE(M_DIRADD, "diradd","New directory entry");
-static MALLOC_DEFINE(M_MKDIR, "mkdir","New directory");
-static MALLOC_DEFINE(M_DIRREM, "dirrem","Directory entry deleted");
-static MALLOC_DEFINE(M_NEWDIRBLK, "newdirblk","Unclaimed new directory block");
-static MALLOC_DEFINE(M_SAVEDINO, "savedino","Saved inodes");
 
 #define M_SOFTDEP_FLAGS	(M_WAITOK | M_USE_RESERVE)
 
 #define	D_PAGEDEP	0
 #define	D_INODEDEP	1
-#define	D_NEWBLK	2
-#define	D_BMSAFEMAP	3
+#define	D_BMSAFEMAP	2
+#define	D_NEWBLK	3
 #define	D_ALLOCDIRECT	4
 #define	D_INDIRDEP	5
 #define	D_ALLOCINDIR	6
@@ -438,7 +425,67 @@ static MALLOC_DEFINE(M_SAVEDINO, "savedino","Saved inodes");
 #define	D_MKDIR		11
 #define	D_DIRREM	12
 #define	D_NEWDIRBLK	13
-#define	D_LAST		D_NEWDIRBLK
+#define	D_FREEWORK	14
+#define	D_FREEDEP	15
+#define	D_JADDREF	16
+#define	D_JREMREF	17
+#define	D_JMVREF	18
+#define	D_JNEWBLK	19
+#define	D_JFREEBLK	20
+#define	D_JFREEFRAG	21
+#define	D_JSEG		22
+#define	D_JSEGDEP	23
+#define	D_SBDEP		24
+#define	D_JTRUNC	25
+#define	D_LAST		D_JTRUNC
+
+unsigned long dep_current[D_LAST + 1];
+unsigned long dep_total[D_LAST + 1];
+
+
+SYSCTL_NODE(_debug, OID_AUTO, softdep, CTLFLAG_RW, 0, "soft updates stats");
+SYSCTL_NODE(_debug_softdep, OID_AUTO, total, CTLFLAG_RW, 0,
+    "total dependencies allocated");
+SYSCTL_NODE(_debug_softdep, OID_AUTO, current, CTLFLAG_RW, 0,
+    "current dependencies allocated");
+
+#define	SOFTDEP_TYPE(type, str, long)					\
+    static MALLOC_DEFINE(M_ ## type, #str, long);			\
+    SYSCTL_LONG(_debug_softdep_total, OID_AUTO, str, CTLFLAG_RD,	\
+	&dep_total[D_ ## type], 0, "");					\
+    SYSCTL_LONG(_debug_softdep_current, OID_AUTO, str, CTLFLAG_RD, 	\
+	&dep_current[D_ ## type], 0, "");
+
+SOFTDEP_TYPE(PAGEDEP, pagedep, "File page dependencies"); 
+SOFTDEP_TYPE(INODEDEP, inodedep, "Inode dependencies");
+SOFTDEP_TYPE(BMSAFEMAP, bmsafemap,
+    "Block or frag allocated from cyl group map");
+SOFTDEP_TYPE(NEWBLK, newblk, "New block or frag allocation dependency");
+SOFTDEP_TYPE(ALLOCDIRECT, allocdirect, "Block or frag dependency for an inode");
+SOFTDEP_TYPE(INDIRDEP, indirdep, "Indirect block dependencies");
+SOFTDEP_TYPE(ALLOCINDIR, allocindir, "Block dependency for an indirect block");
+SOFTDEP_TYPE(FREEFRAG, freefrag, "Previously used frag for an inode");
+SOFTDEP_TYPE(FREEBLKS, freeblks, "Blocks freed from an inode");
+SOFTDEP_TYPE(FREEFILE, freefile, "Inode deallocated");
+SOFTDEP_TYPE(DIRADD, diradd, "New directory entry");
+SOFTDEP_TYPE(MKDIR, mkdir, "New directory");
+SOFTDEP_TYPE(DIRREM, dirrem, "Directory entry deleted");
+SOFTDEP_TYPE(NEWDIRBLK, newdirblk, "Unclaimed new directory block");
+SOFTDEP_TYPE(FREEWORK, freework, "free an inode block");
+SOFTDEP_TYPE(FREEDEP, freedep, "track a block free");
+SOFTDEP_TYPE(JADDREF, jaddref, "Journal inode ref add");
+SOFTDEP_TYPE(JREMREF, jremref, "Journal inode ref remove");
+SOFTDEP_TYPE(JMVREF, jmvref, "Journal inode ref move");
+SOFTDEP_TYPE(JNEWBLK, jnewblk, "Journal new block");
+SOFTDEP_TYPE(JFREEBLK, jfreeblk, "Journal free block");
+SOFTDEP_TYPE(JFREEFRAG, jfreefrag, "Journal free frag");
+SOFTDEP_TYPE(JSEG, jseg, "Journal segment");
+SOFTDEP_TYPE(JSEGDEP, jsegdep, "Journal segment complete");
+SOFTDEP_TYPE(SBDEP, sbdep, "Superblock write dependency");
+SOFTDEP_TYPE(JTRUNC, jtrunc, "Journal inode truncation");
+
+static MALLOC_DEFINE(M_SAVEDINO, "savedino", "Saved inodes");
+static MALLOC_DEFINE(M_JBLOCKS, "jblocks", "Journal block locations");
 
 /* 
  * translate from workitem type to memory type
@@ -447,8 +494,8 @@ static MALLOC_DEFINE(M_SAVEDINO, "savedino","Saved inodes");
 static struct malloc_type *memtype[] = {
 	M_PAGEDEP,
 	M_INODEDEP,
-	M_NEWBLK,
 	M_BMSAFEMAP,
+	M_NEWBLK,
 	M_ALLOCDIRECT,
 	M_INDIRDEP,
 	M_ALLOCINDIR,
@@ -458,7 +505,19 @@ static struct malloc_type *memtype[] = {
 	M_DIRADD,
 	M_MKDIR,
 	M_DIRREM,
-	M_NEWDIRBLK
+	M_NEWDIRBLK,
+	M_FREEWORK,
+	M_FREEDEP,
+	M_JADDREF,
+	M_JREMREF,
+	M_JMVREF,
+	M_JNEWBLK,
+	M_JFREEBLK,
+	M_JFREEFRAG,
+	M_JSEG,
+	M_JSEGDEP,
+	M_SBDEP,
+	M_JTRUNC
 };
 
 #define DtoM(type) (memtype[type])
@@ -467,17 +526,21 @@ static struct malloc_type *memtype[] = {
  * Names of malloc types.
  */
 #define TYPENAME(type)  \
-	((unsigned)(type) < D_LAST ? memtype[type]->ks_shortdesc : "???")
+	((unsigned)(type) <= D_LAST ? memtype[type]->ks_shortdesc : "???")
 /*
  * End system adaptation definitions.
  */
 
+#define	DOTDOT_OFFSET	offsetof(struct dirtemplate, dotdot_ino)
+#define	DOT_OFFSET	offsetof(struct dirtemplate, dot_ino)
+
 /*
  * Forward declarations.
  */
 struct inodedep_hashhead;
 struct newblk_hashhead;
 struct pagedep_hashhead;
+struct bmsafemap_hashhead;
 
 /*
  * Internal function prototypes.
@@ -487,59 +550,172 @@ static	void drain_output(struct vnode *);
 static	struct buf *getdirtybuf(struct buf *, struct mtx *, int);
 static	void clear_remove(struct thread *);
 static	void clear_inodedeps(struct thread *);
+static	void unlinked_inodedep(struct mount *, struct inodedep *);
+static	void clear_unlinked_inodedep(struct inodedep *);
+static	struct inodedep *first_unlinked_inodedep(struct ufsmount *);
 static	int flush_pagedep_deps(struct vnode *, struct mount *,
 	    struct diraddhd *);
+static	void free_pagedep(struct pagedep *);
+static	int flush_newblk_dep(struct vnode *, struct mount *, ufs_lbn_t);
 static	int flush_inodedep_deps(struct mount *, ino_t);
 static	int flush_deplist(struct allocdirectlst *, int, int *);
 static	int handle_written_filepage(struct pagedep *, struct buf *);
+static	int handle_written_sbdep(struct sbdep *, struct buf *);
+static	void initiate_write_sbdep(struct sbdep *);
 static  void diradd_inode_written(struct diradd *, struct inodedep *);
+static	int handle_written_indirdep(struct indirdep *, struct buf *,
+	    struct buf**);
 static	int handle_written_inodeblock(struct inodedep *, struct buf *);
-static	void handle_allocdirect_partdone(struct allocdirect *);
+static	int handle_written_bmsafemap(struct bmsafemap *, struct buf *);
+static	void handle_written_jaddref(struct jaddref *);
+static	void handle_written_jremref(struct jremref *);
+static	void handle_written_jseg(struct jseg *, struct buf *);
+static	void handle_written_jnewblk(struct jnewblk *);
+static	void handle_written_jfreeblk(struct jfreeblk *);
+static	void handle_written_jfreefrag(struct jfreefrag *);
+static	void complete_jseg(struct jseg *);
+static	void jseg_write(struct fs *, struct jblocks *, struct jseg *,
+	    uint8_t *);
+static	void jaddref_write(struct jaddref *, struct jseg *, uint8_t *);
+static	void jremref_write(struct jremref *, struct jseg *, uint8_t *);
+static	void jmvref_write(struct jmvref *, struct jseg *, uint8_t *);
+static	void jtrunc_write(struct jtrunc *, struct jseg *, uint8_t *);
+static	void jnewblk_write(struct jnewblk *, struct jseg *, uint8_t *);
+static	void jfreeblk_write(struct jfreeblk *, struct jseg *, uint8_t *);
+static	void jfreefrag_write(struct jfreefrag *, struct jseg *, uint8_t *);
+static	inline void inoref_write(struct inoref *, struct jseg *,
+	    struct jrefrec *);
+static	void handle_allocdirect_partdone(struct allocdirect *,
+	    struct workhead *);
+static	void cancel_newblk(struct newblk *, struct workhead *);
+static	void indirdep_complete(struct indirdep *);
 static	void handle_allocindir_partdone(struct allocindir *);
 static	void initiate_write_filepage(struct pagedep *, struct buf *);
+static	void initiate_write_indirdep(struct indirdep*, struct buf *);
 static	void handle_written_mkdir(struct mkdir *, int);
+static	void initiate_write_bmsafemap(struct bmsafemap *, struct buf *);
 static	void initiate_write_inodeblock_ufs1(struct inodedep *, struct buf *);
 static	void initiate_write_inodeblock_ufs2(struct inodedep *, struct buf *);
 static	void handle_workitem_freefile(struct freefile *);
 static	void handle_workitem_remove(struct dirrem *, struct vnode *);
 static	struct dirrem *newdirrem(struct buf *, struct inode *,
 	    struct inode *, int, struct dirrem **);
-static	void free_diradd(struct diradd *);
-static	void free_allocindir(struct allocindir *, struct inodedep *);
+static	void cancel_indirdep(struct indirdep *, struct buf *, struct inodedep *,
+	    struct freeblks *);
+static	void free_indirdep(struct indirdep *);
+static	void free_diradd(struct diradd *, struct workhead *);
+static	void merge_diradd(struct inodedep *, struct diradd *);
+static	void complete_diradd(struct diradd *);
+static	struct diradd *diradd_lookup(struct pagedep *, int);
+static	struct jremref *cancel_diradd_dotdot(struct inode *, struct dirrem *,
+	    struct jremref *);
+static	struct jremref *cancel_mkdir_dotdot(struct inode *, struct dirrem *,
+	    struct jremref *);
+static	void cancel_diradd(struct diradd *, struct dirrem *, struct jremref *,
+	    struct jremref *, struct jremref *);
+static	void dirrem_journal(struct dirrem *, struct jremref *, struct jremref *,
+	    struct jremref *);
+static	void cancel_allocindir(struct allocindir *, struct inodedep *,
+	    struct freeblks *);
+static	void complete_mkdir(struct mkdir *);
 static	void free_newdirblk(struct newdirblk *);
-static	int indir_trunc(struct freeblks *, ufs2_daddr_t, int, ufs_lbn_t,
-	    ufs2_daddr_t *);
-static	void deallocate_dependencies(struct buf *, struct inodedep *);
-static	void free_allocdirect(struct allocdirectlst *,
-	    struct allocdirect *, int);
+static	void free_jremref(struct jremref *);
+static	void free_jaddref(struct jaddref *);
+static	void free_jsegdep(struct jsegdep *);
+static	void free_jseg(struct jseg *);
+static	void free_jnewblk(struct jnewblk *);
+static	void free_jfreeblk(struct jfreeblk *);
+static	void free_jfreefrag(struct jfreefrag *);
+static	void free_freedep(struct freedep *);
+static	void journal_jremref(struct dirrem *, struct jremref *,
+	    struct inodedep *);
+static	void cancel_jnewblk(struct jnewblk *, struct workhead *);
+static	int cancel_jaddref(struct jaddref *, struct inodedep *,
+	    struct workhead *);
+static	void cancel_jfreefrag(struct jfreefrag *);
+static	void indir_trunc(struct freework *, ufs2_daddr_t, ufs_lbn_t);
+static	int deallocate_dependencies(struct buf *, struct inodedep *,
+	    struct freeblks *);
+static	void free_newblk(struct newblk *);
+static	void cancel_allocdirect(struct allocdirectlst *,
+	    struct allocdirect *, struct freeblks *, int);
 static	int check_inode_unwritten(struct inodedep *);
 static	int free_inodedep(struct inodedep *);
+static	void freework_freeblock(struct freework *);
 static	void handle_workitem_freeblocks(struct freeblks *, int);
+static	void handle_complete_freeblocks(struct freeblks *);
+static	void handle_workitem_indirblk(struct freework *);
+static	void handle_written_freework(struct freework *);
 static	void merge_inode_lists(struct allocdirectlst *,struct allocdirectlst *);
 static	void setup_allocindir_phase2(struct buf *, struct inode *,
-	    struct allocindir *);
+	    struct inodedep *, struct allocindir *, ufs_lbn_t);
 static	struct allocindir *newallocindir(struct inode *, int, ufs2_daddr_t,
-	    ufs2_daddr_t);
+	    ufs2_daddr_t, ufs_lbn_t);
 static	void handle_workitem_freefrag(struct freefrag *);
-static	struct freefrag *newfreefrag(struct inode *, ufs2_daddr_t, long);
+static	struct freefrag *newfreefrag(struct inode *, ufs2_daddr_t, long,
+	    ufs_lbn_t);
 static	void allocdirect_merge(struct allocdirectlst *,
 	    struct allocdirect *, struct allocdirect *);
-static	struct bmsafemap *bmsafemap_lookup(struct mount *, struct buf *);
-static	int newblk_find(struct newblk_hashhead *, struct fs *, ufs2_daddr_t,
-	    struct newblk **);
-static	int newblk_lookup(struct fs *, ufs2_daddr_t, int, struct newblk **);
+static	struct freefrag *allocindir_merge(struct allocindir *,
+	    struct allocindir *);
+static	int bmsafemap_find(struct bmsafemap_hashhead *, struct mount *, int,
+	    struct bmsafemap **);
+static	struct bmsafemap *bmsafemap_lookup(struct mount *, struct buf *,
+	    int cg);
+static	int newblk_find(struct newblk_hashhead *, struct mount *, ufs2_daddr_t,
+	    int, struct newblk **);
+static	int newblk_lookup(struct mount *, ufs2_daddr_t, int, struct newblk **);
 static	int inodedep_find(struct inodedep_hashhead *, struct fs *, ino_t,
 	    struct inodedep **);
 static	int inodedep_lookup(struct mount *, ino_t, int, struct inodedep **);
-static	int pagedep_lookup(struct inode *, ufs_lbn_t, int, struct pagedep **);
+static	int pagedep_lookup(struct mount *, ino_t, ufs_lbn_t, int,
+	    struct pagedep **);
 static	int pagedep_find(struct pagedep_hashhead *, ino_t, ufs_lbn_t,
 	    struct mount *mp, int, struct pagedep **);
 static	void pause_timer(void *);
 static	int request_cleanup(struct mount *, int);
 static	int process_worklist_item(struct mount *, int);
-static	void add_to_worklist(struct worklist *);
+static	void process_removes(struct vnode *);
+static	void jwork_move(struct workhead *, struct workhead *);
+static	void add_to_worklist(struct worklist *, int);
+static	void remove_from_worklist(struct worklist *);
 static	void softdep_flush(void);
 static	int softdep_speedup(void);
+static	void worklist_speedup(void);
+static	int journal_mount(struct mount *, struct fs *, struct ucred *);
+static	void journal_unmount(struct mount *);
+static	int journal_space(struct ufsmount *, int);
+static	void journal_suspend(struct ufsmount *);
+static	void softdep_prelink(struct vnode *, struct vnode *);
+static	void add_to_journal(struct worklist *);
+static	void remove_from_journal(struct worklist *);
+static	void softdep_process_journal(struct mount *, int);
+static	struct jremref *newjremref(struct dirrem *, struct inode *,
+	    struct inode *ip, off_t, nlink_t);
+static	struct jaddref *newjaddref(struct inode *, ino_t, off_t, int16_t,
+	    uint16_t);
+static inline void newinoref(struct inoref *, ino_t, ino_t, off_t, nlink_t,
+	    uint16_t);
+static inline struct jsegdep *inoref_jseg(struct inoref *);
+static	struct jmvref *newjmvref(struct inode *, ino_t, off_t, off_t);
+static	struct jfreeblk *newjfreeblk(struct freeblks *, ufs_lbn_t,
+	    ufs2_daddr_t, int);
+static	struct jfreefrag *newjfreefrag(struct freefrag *, struct inode *,
+	    ufs2_daddr_t, long, ufs_lbn_t);
+static	struct freework *newfreework(struct freeblks *, struct freework *, 
+	    ufs_lbn_t, ufs2_daddr_t, int, int);
+static	void jwait(struct worklist *wk);
+static	struct inodedep *inodedep_lookup_ip(struct inode *);
+static	int bmsafemap_rollbacks(struct bmsafemap *);
+static	struct freefile *handle_bufwait(struct inodedep *, struct workhead *);
+static	void handle_jwork(struct workhead *);
+static	struct mkdir *setup_newdir(struct diradd *, ino_t, ino_t, struct buf *,
+	    struct mkdir **);
+static	struct jblocks *jblocks_create(void);
+static	ufs2_daddr_t jblocks_alloc(struct jblocks *, int, int *);
+static	void jblocks_free(struct jblocks *, struct mount *, int);
+static	void jblocks_destroy(struct jblocks *);
+static	void jblocks_add(struct jblocks *, ufs2_daddr_t, int);
 
 /*
  * Exported softdep operations.
@@ -572,39 +748,127 @@ MTX_SYSINIT(softdep_lock, &lk, "Softdep Lock", MTX_DEF);
 	(item)->wk_state &= ~ONWORKLIST;	\
 	LIST_REMOVE(item, wk_list);		\
 } while (0)
-#else /* DEBUG */
-static	void worklist_insert(struct workhead *, struct worklist *);
-static	void worklist_remove(struct worklist *);
+#define WORKLIST_INSERT_UNLOCKED	WORKLIST_INSERT
+#define WORKLIST_REMOVE_UNLOCKED	WORKLIST_REMOVE
 
-#define WORKLIST_INSERT(head, item) worklist_insert(head, item)
-#define WORKLIST_REMOVE(item) worklist_remove(item)
+#else /* DEBUG */
+static	void worklist_insert(struct workhead *, struct worklist *, int);
+static	void worklist_remove(struct worklist *, int);
+
+#define WORKLIST_INSERT(head, item) worklist_insert(head, item, 1)
+#define WORKLIST_INSERT_UNLOCKED(head, item) worklist_insert(head, item, 0)
+#define WORKLIST_REMOVE(item) worklist_remove(item, 1)
+#define WORKLIST_REMOVE_UNLOCKED(item) worklist_remove(item, 0)
 
 static void
-worklist_insert(head, item)
+worklist_insert(head, item, locked)
 	struct workhead *head;
 	struct worklist *item;
+	int locked;
 {
 
-	mtx_assert(&lk, MA_OWNED);
+	if (locked)
+		mtx_assert(&lk, MA_OWNED);
 	if (item->wk_state & ONWORKLIST)
-		panic("worklist_insert: already on list");
+		panic("worklist_insert: %p %s(0x%X) already on list",
+		    item, TYPENAME(item->wk_type), item->wk_state);
 	item->wk_state |= ONWORKLIST;
 	LIST_INSERT_HEAD(head, item, wk_list);
 }
 
 static void
-worklist_remove(item)
+worklist_remove(item, locked)
 	struct worklist *item;
+	int locked;
 {
 
-	mtx_assert(&lk, MA_OWNED);
+	if (locked)
+		mtx_assert(&lk, MA_OWNED);
 	if ((item->wk_state & ONWORKLIST) == 0)
-		panic("worklist_remove: not on list");
+		panic("worklist_remove: %p %s(0x%X) not on list",
+		    item, TYPENAME(item->wk_type), item->wk_state);
 	item->wk_state &= ~ONWORKLIST;
 	LIST_REMOVE(item, wk_list);
 }
 #endif /* DEBUG */
 
+/*
+ * Merge two jsegdeps keeping only the oldest one as newer references
+ * can't be discarded until after older references.
+ */
+static inline struct jsegdep *
+jsegdep_merge(struct jsegdep *one, struct jsegdep *two)
+{
+	struct jsegdep *swp;
+
+	if (two == NULL)
+		return (one);
+
+	if (one->jd_seg->js_seq > two->jd_seg->js_seq) {
+		swp = one;
+		one = two;
+		two = swp;
+	}
+	WORKLIST_REMOVE(&two->jd_list);
+	free_jsegdep(two);
+
+	return (one);
+}
+
+/*
+ * If two freedeps are compatible free one to reduce list size.
+ */
+static inline struct freedep *
+freedep_merge(struct freedep *one, struct freedep *two)
+{
+	if (two == NULL)
+		return (one);
+
+	if (one->fd_freework == two->fd_freework) {
+		WORKLIST_REMOVE(&two->fd_list);
+		free_freedep(two);
+	}
+	return (one);
+}
+
+/*
+ * Move journal work from one list to another.  Duplicate freedeps and
+ * jsegdeps are coalesced to keep the lists as small as possible.
+ */
+static void
+jwork_move(dst, src)
+	struct workhead *dst;
+	struct workhead *src;
+{
+	struct freedep *freedep;
+	struct jsegdep *jsegdep;
+	struct worklist *wkn;
+	struct worklist *wk;
+
+	KASSERT(dst != src,
+	    ("jwork_move: dst == src"));
+	freedep = NULL;
+	jsegdep = NULL;
+	LIST_FOREACH_SAFE(wk, dst, wk_list, wkn) {
+		if (wk->wk_type == D_JSEGDEP)
+			jsegdep = jsegdep_merge(WK_JSEGDEP(wk), jsegdep);
+		if (wk->wk_type == D_FREEDEP)
+			freedep = freedep_merge(WK_FREEDEP(wk), freedep);
+	}
+
+	mtx_assert(&lk, MA_OWNED);
+	while ((wk = LIST_FIRST(src)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		WORKLIST_INSERT(dst, wk);
+		if (wk->wk_type == D_JSEGDEP) {
+			jsegdep = jsegdep_merge(WK_JSEGDEP(wk), jsegdep);
+			continue;
+		}
+		if (wk->wk_type == D_FREEDEP)
+			freedep = freedep_merge(WK_FREEDEP(wk), freedep);
+	}
+}
+
 /*
  * Routines for tracking and managing workitems.
  */
@@ -623,13 +887,16 @@ workitem_free(item, type)
 
 #ifdef DEBUG
 	if (item->wk_state & ONWORKLIST)
-		panic("workitem_free: still on list");
+		panic("workitem_free: %s(0x%X) still on list",
+		    TYPENAME(item->wk_type), item->wk_state);
 	if (item->wk_type != type)
-		panic("workitem_free: type mismatch");
+		panic("workitem_free: type mismatch %s != %s",
+		    TYPENAME(item->wk_type), TYPENAME(type));
 #endif
 	ump = VFSTOUFS(item->wk_mp);
 	if (--ump->softdep_deps == 0 && ump->softdep_req)
 		wakeup(&ump->softdep_deps);
+	dep_current[type]--;
 	free(item, DtoM(type));
 }
 
@@ -643,6 +910,8 @@ workitem_alloc(item, type, mp)
 	item->wk_mp = mp;
 	item->wk_state = 0;
 	ACQUIRE_LOCK(&lk);
+	dep_current[type]++;
+	dep_total[type]++;
 	VFSTOUFS(mp)->softdep_deps++;
 	VFSTOUFS(mp)->softdep_accdeps++;
 	FREE_LOCK(&lk);
@@ -678,24 +947,66 @@ static int stat_indir_blk_ptrs;	/* bufs redirtied as indir ptrs not written */
 static int stat_inode_bitmap;	/* bufs redirtied as inode bitmap not written */
 static int stat_direct_blk_ptrs;/* bufs redirtied as direct ptrs not written */
 static int stat_dir_entry;	/* bufs redirtied as dir entry cannot write */
+static int stat_jaddref;	/* bufs redirtied as ino bitmap can not write */
+static int stat_jnewblk;	/* bufs redirtied as blk bitmap can not write */
+static int stat_journal_min;	/* Times hit journal min threshold */
+static int stat_journal_low;	/* Times hit journal low threshold */
+static int stat_journal_wait;	/* Times blocked in jwait(). */
+static int stat_jwait_filepage;	/* Times blocked in jwait() for filepage. */
+static int stat_jwait_freeblks;	/* Times blocked in jwait() for freeblks. */
+static int stat_jwait_inode;	/* Times blocked in jwait() for inodes. */
+static int stat_jwait_newblk;	/* Times blocked in jwait() for newblks. */
 
-SYSCTL_INT(_debug, OID_AUTO, max_softdeps, CTLFLAG_RW, &max_softdeps, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, tickdelay, CTLFLAG_RW, &tickdelay, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, maxindirdeps, CTLFLAG_RW, &maxindirdeps, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, worklist_push, CTLFLAG_RW, &stat_worklist_push, 0,"");
-SYSCTL_INT(_debug, OID_AUTO, blk_limit_push, CTLFLAG_RW, &stat_blk_limit_push, 0,"");
-SYSCTL_INT(_debug, OID_AUTO, ino_limit_push, CTLFLAG_RW, &stat_ino_limit_push, 0,"");
-SYSCTL_INT(_debug, OID_AUTO, blk_limit_hit, CTLFLAG_RW, &stat_blk_limit_hit, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, ino_limit_hit, CTLFLAG_RW, &stat_ino_limit_hit, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, sync_limit_hit, CTLFLAG_RW, &stat_sync_limit_hit, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, indir_blk_ptrs, CTLFLAG_RW, &stat_indir_blk_ptrs, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, inode_bitmap, CTLFLAG_RW, &stat_inode_bitmap, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, direct_blk_ptrs, CTLFLAG_RW, &stat_direct_blk_ptrs, 0, "");
-SYSCTL_INT(_debug, OID_AUTO, dir_entry, CTLFLAG_RW, &stat_dir_entry, 0, "");
-/* SYSCTL_INT(_debug, OID_AUTO, worklist_num, CTLFLAG_RD, &softdep_on_worklist, 0, ""); */
+SYSCTL_INT(_debug_softdep, OID_AUTO, max_softdeps, CTLFLAG_RW,
+    &max_softdeps, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, tickdelay, CTLFLAG_RW,
+    &tickdelay, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, maxindirdeps, CTLFLAG_RW,
+    &maxindirdeps, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, worklist_push, CTLFLAG_RW,
+    &stat_worklist_push, 0,"");
+SYSCTL_INT(_debug_softdep, OID_AUTO, blk_limit_push, CTLFLAG_RW,
+    &stat_blk_limit_push, 0,"");
+SYSCTL_INT(_debug_softdep, OID_AUTO, ino_limit_push, CTLFLAG_RW,
+    &stat_ino_limit_push, 0,"");
+SYSCTL_INT(_debug_softdep, OID_AUTO, blk_limit_hit, CTLFLAG_RW,
+    &stat_blk_limit_hit, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, ino_limit_hit, CTLFLAG_RW,
+    &stat_ino_limit_hit, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, sync_limit_hit, CTLFLAG_RW,
+    &stat_sync_limit_hit, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, indir_blk_ptrs, CTLFLAG_RW,
+    &stat_indir_blk_ptrs, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, inode_bitmap, CTLFLAG_RW,
+    &stat_inode_bitmap, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, direct_blk_ptrs, CTLFLAG_RW,
+    &stat_direct_blk_ptrs, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, dir_entry, CTLFLAG_RW,
+    &stat_dir_entry, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jaddref_rollback, CTLFLAG_RW,
+    &stat_jaddref, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jnewblk_rollback, CTLFLAG_RW,
+    &stat_jnewblk, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, journal_low, CTLFLAG_RW,
+    &stat_journal_low, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, journal_min, CTLFLAG_RW,
+    &stat_journal_min, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, journal_wait, CTLFLAG_RW,
+    &stat_journal_wait, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jwait_filepage, CTLFLAG_RW,
+    &stat_jwait_filepage, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jwait_freeblks, CTLFLAG_RW,
+    &stat_jwait_freeblks, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jwait_inode, CTLFLAG_RW,
+    &stat_jwait_inode, 0, "");
+SYSCTL_INT(_debug_softdep, OID_AUTO, jwait_newblk, CTLFLAG_RW,
+    &stat_jwait_newblk, 0, "");
 
 SYSCTL_DECL(_vfs_ffs);
 
+LIST_HEAD(bmsafemap_hashhead, bmsafemap) *bmsafemap_hashtbl;
+static u_long	bmsafemap_hash;	/* size of hash table - 1 */
+
 static int compute_summary_at_mount = 0;	/* Whether to recompute the summary at mount time */
 SYSCTL_INT(_vfs_ffs, OID_AUTO, compute_summary_at_mount, CTLFLAG_RW,
 	   &compute_summary_at_mount, 0, "Recompute summary at mount");
@@ -770,16 +1081,22 @@ softdep_flush(void)
 	}
 }
 
-static int
-softdep_speedup(void)
+static void
+worklist_speedup(void)
 {
-
 	mtx_assert(&lk, MA_OWNED);
 	if (req_pending == 0) {
 		req_pending = 1;
 		wakeup(&req_pending);
 	}
+}
 
+static int
+softdep_speedup(void)
+{
+
+	worklist_speedup();
+	bd_speedup();
 	return speedup_syncer();
 }
 
@@ -791,15 +1108,17 @@ softdep_speedup(void)
  * and does so in order from first to last.
  */
 static void
-add_to_worklist(wk)
+add_to_worklist(wk, nodelay)
 	struct worklist *wk;
+	int nodelay;
 {
 	struct ufsmount *ump;
 
 	mtx_assert(&lk, MA_OWNED);
 	ump = VFSTOUFS(wk->wk_mp);
 	if (wk->wk_state & ONWORKLIST)
-		panic("add_to_worklist: already on list");
+		panic("add_to_worklist: %s(0x%X) already on list",
+		    TYPENAME(wk->wk_type), wk->wk_state);
 	wk->wk_state |= ONWORKLIST;
 	if (LIST_EMPTY(&ump->softdep_workitem_pending))
 		LIST_INSERT_HEAD(&ump->softdep_workitem_pending, wk, wk_list);
@@ -807,6 +1126,30 @@ add_to_worklist(wk)
 		LIST_INSERT_AFTER(ump->softdep_worklist_tail, wk, wk_list);
 	ump->softdep_worklist_tail = wk;
 	ump->softdep_on_worklist += 1;
+	if (nodelay)
+		worklist_speedup();
+}
+
+/*
+ * Remove the item to be processed. If we are removing the last
+ * item on the list, we need to recalculate the tail pointer.
+ */
+static void
+remove_from_worklist(wk)
+	struct worklist *wk;
+{
+	struct ufsmount *ump;
+	struct worklist *wkend;
+
+	ump = VFSTOUFS(wk->wk_mp);
+	WORKLIST_REMOVE(wk);
+	if (wk == ump->softdep_worklist_tail) {
+		LIST_FOREACH(wkend, &ump->softdep_workitem_pending, wk_list)
+			if (LIST_NEXT(wkend, wk_list) == NULL)
+				break;
+		ump->softdep_worklist_tail = wkend;
+	}
+	ump->softdep_on_worklist -= 1;
 }
 
 /*
@@ -838,8 +1181,9 @@ softdep_process_worklist(mp, full)
 	ACQUIRE_LOCK(&lk);
 	loopcount = 1;
 	starttime = time_second;
+	softdep_process_journal(mp, full?MNT_WAIT:0);
 	while (ump->softdep_on_worklist > 0) {
-		if ((cnt = process_worklist_item(mp, 0)) == -1)
+		if ((cnt = process_worklist_item(mp, LK_NOWAIT)) == -1)
 			break;
 		else
 			matchcnt += cnt;
@@ -871,15 +1215,60 @@ softdep_process_worklist(mp, full)
 		 * second. Otherwise the other mountpoints may get
 		 * excessively backlogged.
 		 */
-		if (!full && starttime != time_second) {
-			matchcnt = -1;
+		if (!full && starttime != time_second)
 			break;
-		}
 	}
 	FREE_LOCK(&lk);
 	return (matchcnt);
 }
 
+/*
+ * Process all removes associated with a vnode if we are running out of
+ * journal space.  Any other process which attempts to flush these will
+ * be unable as we have the vnodes locked.
+ */
+static void
+process_removes(vp)
+	struct vnode *vp;
+{
+	struct inodedep *inodedep;
+	struct dirrem *dirrem;
+	struct mount *mp;
+	ino_t inum;
+
+	mtx_assert(&lk, MA_OWNED);
+
+	mp = vp->v_mount;
+	inum = VTOI(vp)->i_number;
+	for (;;) {
+		if (inodedep_lookup(mp, inum, 0, &inodedep) == 0)
+			return;
+		LIST_FOREACH(dirrem, &inodedep->id_dirremhd, dm_inonext)
+			if ((dirrem->dm_state & (COMPLETE | ONWORKLIST)) ==
+			    (COMPLETE | ONWORKLIST))
+				break;
+		if (dirrem == NULL)
+			return;
+		/*
+		 * If another thread is trying to lock this vnode it will
+		 * fail but we must wait for it to do so before we can
+		 * proceed.
+		 */
+		if (dirrem->dm_state & INPROGRESS) {
+			dirrem->dm_state |= IOWAITING;
+			msleep(&dirrem->dm_list, &lk, PVM, "pwrwait", 0);
+			continue;
+		}
+		remove_from_worklist(&dirrem->dm_list);
+		FREE_LOCK(&lk);
+		if (vn_start_secondary_write(NULL, &mp, V_NOWAIT))
+			panic("process_removes: suspended filesystem");
+		handle_workitem_remove(dirrem, vp);
+		vn_finished_secondary_write(mp);
+		ACQUIRE_LOCK(&lk);
+	}
+}
+
 /*
  * Process one item on the worklist.
  */
@@ -888,7 +1277,7 @@ process_worklist_item(mp, flags)
 	struct mount *mp;
 	int flags;
 {
-	struct worklist *wk, *wkend;
+	struct worklist *wk, *wkXXX;
 	struct ufsmount *ump;
 	struct vnode *vp;
 	int matchcnt = 0;
@@ -908,11 +1297,14 @@ process_worklist_item(mp, flags)
 	 * inodes, we have to skip over any dirrem requests whose
 	 * vnodes are resident and locked.
 	 */
-	ump = VFSTOUFS(mp);
 	vp = NULL;
+	ump = VFSTOUFS(mp);
 	LIST_FOREACH(wk, &ump->softdep_workitem_pending, wk_list) {
-		if (wk->wk_state & INPROGRESS)
+		if (wk->wk_state & INPROGRESS) {
+			wkXXX = wk;
 			continue;
+		}
+		wkXXX = wk;	/* Record the last valid wk pointer. */
 		if ((flags & LK_NOWAIT) == 0 || wk->wk_type != D_DIRREM)
 			break;
 		wk->wk_state |= INPROGRESS;
@@ -921,6 +1313,10 @@ process_worklist_item(mp, flags)
 		ffs_vgetf(mp, WK_DIRREM(wk)->dm_oldinum,
 		    LK_NOWAIT | LK_EXCLUSIVE, &vp, FFSV_FORCEINSMQ);
 		ACQUIRE_LOCK(&lk);
+		if (wk->wk_state & IOWAITING) {
+			wk->wk_state &= ~IOWAITING;
+			wakeup(wk);
+		}
 		wk->wk_state &= ~INPROGRESS;
 		ump->softdep_on_worklist_inprogress--;
 		if (vp != NULL)
@@ -928,21 +1324,7 @@ process_worklist_item(mp, flags)
 	}
 	if (wk == 0)
 		return (-1);
-	/*
-	 * Remove the item to be processed. If we are removing the last
-	 * item on the list, we need to recalculate the tail pointer.
-	 * As this happens rarely and usually when the list is short,
-	 * we just run down the list to find it rather than tracking it
-	 * in the above loop.
-	 */
-	WORKLIST_REMOVE(wk);
-	if (wk == ump->softdep_worklist_tail) {
-		LIST_FOREACH(wkend, &ump->softdep_workitem_pending, wk_list)
-			if (LIST_NEXT(wkend, wk_list) == NULL)
-				break;
-		ump->softdep_worklist_tail = wkend;
-	}
-	ump->softdep_on_worklist -= 1;
+	remove_from_worklist(wk);
 	FREE_LOCK(&lk);
 	if (vn_start_secondary_write(NULL, &mp, V_NOWAIT))
 		panic("process_worklist_item: suspended filesystem");
@@ -952,6 +1334,8 @@ process_worklist_item(mp, flags)
 	case D_DIRREM:
 		/* removal of a directory entry */
 		handle_workitem_remove(WK_DIRREM(wk), vp);
+		if (vp)
+			vput(vp);
 		break;
 
 	case D_FREEBLKS:
@@ -969,6 +1353,11 @@ process_worklist_item(mp, flags)
 		handle_workitem_freefile(WK_FREEFILE(wk));
 		break;
 
+	case D_FREEWORK:
+		/* Final block in an indirect was freed. */
+		handle_workitem_indirblk(WK_FREEWORK(wk));
+		break;
+
 	default:
 		panic("%s_process_worklist: Unknown type %s",
 		    "softdep", TYPENAME(wk->wk_type));
@@ -982,19 +1371,22 @@ process_worklist_item(mp, flags)
 /*
  * Move dependencies from one buffer to another.
  */
-void
+int
 softdep_move_dependencies(oldbp, newbp)
 	struct buf *oldbp;
 	struct buf *newbp;
 {
 	struct worklist *wk, *wktail;
+	int dirty;
 
-	if (!LIST_EMPTY(&newbp->b_dep))
-		panic("softdep_move_dependencies: need merge code");
-	wktail = 0;
+	dirty = 0;
+	wktail = NULL;
 	ACQUIRE_LOCK(&lk);
 	while ((wk = LIST_FIRST(&oldbp->b_dep)) != NULL) {
 		LIST_REMOVE(wk, wk_list);
+		if (wk->wk_type == D_BMSAFEMAP &&
+		    bmsafemap_rollbacks(WK_BMSAFEMAP(wk)))
+			dirty = 1;
 		if (wktail == 0)
 			LIST_INSERT_HEAD(&newbp->b_dep, wk, wk_list);
 		else
@@ -1002,6 +1394,8 @@ softdep_move_dependencies(oldbp, newbp)
 		wktail = wk;
 	}
 	FREE_LOCK(&lk);
+
+	return (dirty);
 }
 
 /*
@@ -1198,23 +1592,22 @@ pagedep_find(pagedephd, ino, lbn, mp, flags, pagedeppp)
  * This routine must be called with splbio interrupts blocked.
  */
 static int
-pagedep_lookup(ip, lbn, flags, pagedeppp)
-	struct inode *ip;
+pagedep_lookup(mp, ino, lbn, flags, pagedeppp)
+	struct mount *mp;
+	ino_t ino;
 	ufs_lbn_t lbn;
 	int flags;
 	struct pagedep **pagedeppp;
 {
 	struct pagedep *pagedep;
 	struct pagedep_hashhead *pagedephd;
-	struct mount *mp;
 	int ret;
 	int i;
 
 	mtx_assert(&lk, MA_OWNED);
-	mp = ITOV(ip)->v_mount;
-	pagedephd = PAGEDEP_HASH(mp, ip->i_number, lbn);
+	pagedephd = PAGEDEP_HASH(mp, ino, lbn);
 
-	ret = pagedep_find(pagedephd, ip->i_number, lbn, mp, flags, pagedeppp);
+	ret = pagedep_find(pagedephd, ino, lbn, mp, flags, pagedeppp);
 	if (*pagedeppp || (flags & DEPALLOC) == 0)
 		return (ret);
 	FREE_LOCK(&lk);
@@ -1222,12 +1615,12 @@ pagedep_lookup(ip, lbn, flags, pagedeppp)
 	    M_PAGEDEP, M_SOFTDEP_FLAGS|M_ZERO);
 	workitem_alloc(&pagedep->pd_list, D_PAGEDEP, mp);
 	ACQUIRE_LOCK(&lk);
-	ret = pagedep_find(pagedephd, ip->i_number, lbn, mp, flags, pagedeppp);
+	ret = pagedep_find(pagedephd, ino, lbn, mp, flags, pagedeppp);
 	if (*pagedeppp) {
 		WORKITEM_FREE(pagedep, D_PAGEDEP);
 		return (ret);
 	}
-	pagedep->pd_ino = ip->i_number;
+	pagedep->pd_ino = ino;
 	pagedep->pd_lbn = lbn;
 	LIST_INIT(&pagedep->pd_dirremhd);
 	LIST_INIT(&pagedep->pd_pendinghd);
@@ -1314,10 +1707,14 @@ inodedep_lookup(mp, inum, flags, inodedeppp)
 	inodedep->id_savedino1 = NULL;
 	inodedep->id_savedsize = -1;
 	inodedep->id_savedextsize = -1;
-	inodedep->id_buf = NULL;
+	inodedep->id_savednlink = -1;
+	inodedep->id_bmsafemap = NULL;
+	inodedep->id_mkdiradd = NULL;
+	LIST_INIT(&inodedep->id_dirremhd);
 	LIST_INIT(&inodedep->id_pendinghd);
 	LIST_INIT(&inodedep->id_inowait);
 	LIST_INIT(&inodedep->id_bufwait);
+	TAILQ_INIT(&inodedep->id_inoreflst);
 	TAILQ_INIT(&inodedep->id_inoupdt);
 	TAILQ_INIT(&inodedep->id_newinoupdt);
 	TAILQ_INIT(&inodedep->id_extupdt);
@@ -1336,17 +1733,29 @@ u_long	newblk_hash;		/* size of hash table - 1 */
 	(&newblk_hashtbl[((((register_t)(fs)) >> 13) + (inum)) & newblk_hash])
 
 static int
-newblk_find(newblkhd, fs, newblkno, newblkpp)
+newblk_find(newblkhd, mp, newblkno, flags, newblkpp)
 	struct newblk_hashhead *newblkhd;
-	struct fs *fs;
+	struct mount *mp;
 	ufs2_daddr_t newblkno;
+	int flags;
 	struct newblk **newblkpp;
 {
 	struct newblk *newblk;
 
-	LIST_FOREACH(newblk, newblkhd, nb_hash)
-		if (newblkno == newblk->nb_newblkno && fs == newblk->nb_fs)
-			break;
+	LIST_FOREACH(newblk, newblkhd, nb_hash) {
+		if (newblkno != newblk->nb_newblkno)
+			continue;
+		if (mp != newblk->nb_list.wk_mp)
+			continue;
+		/*
+		 * If we're creating a new dependency don't match those that
+		 * have already been converted to allocdirects.  This is for
+		 * a frag extend.
+		 */
+		if ((flags & DEPALLOC) && newblk->nb_list.wk_type != D_NEWBLK)
+			continue;
+		break;
+	}
 	if (newblk) {
 		*newblkpp = newblk;
 		return (1);
@@ -1361,8 +1770,8 @@ newblk_find(newblkhd, fs, newblkno, newblkpp)
  * Found or allocated entry is returned in newblkpp.
  */
 static int
-newblk_lookup(fs, newblkno, flags, newblkpp)
-	struct fs *fs;
+newblk_lookup(mp, newblkno, flags, newblkpp)
+	struct mount *mp;
 	ufs2_daddr_t newblkno;
 	int flags;
 	struct newblk **newblkpp;
@@ -1370,21 +1779,25 @@ newblk_lookup(fs, newblkno, flags, newblkpp)
 	struct newblk *newblk;
 	struct newblk_hashhead *newblkhd;
 
-	newblkhd = NEWBLK_HASH(fs, newblkno);
-	if (newblk_find(newblkhd, fs, newblkno, newblkpp))
+	newblkhd = NEWBLK_HASH(VFSTOUFS(mp)->um_fs, newblkno);
+	if (newblk_find(newblkhd, mp, newblkno, flags, newblkpp))
 		return (1);
 	if ((flags & DEPALLOC) == 0)
 		return (0);
 	FREE_LOCK(&lk);
-	newblk = malloc(sizeof(struct newblk),
-		M_NEWBLK, M_SOFTDEP_FLAGS);
+	newblk = malloc(sizeof(union allblk), M_NEWBLK,
+	    M_SOFTDEP_FLAGS | M_ZERO);
+	workitem_alloc(&newblk->nb_list, D_NEWBLK, mp);
 	ACQUIRE_LOCK(&lk);
-	if (newblk_find(newblkhd, fs, newblkno, newblkpp)) {
-		free(newblk, M_NEWBLK);
+	if (newblk_find(newblkhd, mp, newblkno, flags, newblkpp)) {
+		WORKITEM_FREE(newblk, D_NEWBLK);
 		return (1);
 	}
-	newblk->nb_state = 0;
-	newblk->nb_fs = fs;
+	newblk->nb_freefrag = NULL;
+	LIST_INIT(&newblk->nb_indirdeps);
+	LIST_INIT(&newblk->nb_newdirblk);
+	LIST_INIT(&newblk->nb_jwork);
+	newblk->nb_state = ATTACHED;
 	newblk->nb_newblkno = newblkno;
 	LIST_INSERT_HEAD(newblkhd, newblk, nb_hash);
 	*newblkpp = newblk;
@@ -1401,10 +1814,10 @@ softdep_initialize()
 
 	LIST_INIT(&mkdirlisthd);
 	max_softdeps = desiredvnodes * 4;
-	pagedep_hashtbl = hashinit(desiredvnodes / 5, M_PAGEDEP,
-	    &pagedep_hash);
+	pagedep_hashtbl = hashinit(desiredvnodes / 5, M_PAGEDEP, &pagedep_hash);
 	inodedep_hashtbl = hashinit(desiredvnodes, M_INODEDEP, &inodedep_hash);
-	newblk_hashtbl = hashinit(64, M_NEWBLK, &newblk_hash);
+	newblk_hashtbl = hashinit(desiredvnodes / 5,  M_NEWBLK, &newblk_hash);
+	bmsafemap_hashtbl = hashinit(1024, M_BMSAFEMAP, &bmsafemap_hash);
 
 	/* initialise bioops hack */
 	bioops.io_start = softdep_disk_io_initiation;
@@ -1428,6 +1841,7 @@ softdep_uninitialize()
 	hashdestroy(pagedep_hashtbl, M_PAGEDEP, pagedep_hash);
 	hashdestroy(inodedep_hashtbl, M_INODEDEP, inodedep_hash);
 	hashdestroy(newblk_hashtbl, M_NEWBLK, newblk_hash);
+	hashdestroy(bmsafemap_hashtbl, M_BMSAFEMAP, bmsafemap_hash);
 }
 
 /*
@@ -1457,9 +1871,16 @@ softdep_mount(devvp, mp, fs, cred)
 	MNT_IUNLOCK(mp);
 	ump = VFSTOUFS(mp);
 	LIST_INIT(&ump->softdep_workitem_pending);
+	LIST_INIT(&ump->softdep_journal_pending);
+	TAILQ_INIT(&ump->softdep_unlinked);
 	ump->softdep_worklist_tail = NULL;
 	ump->softdep_on_worklist = 0;
 	ump->softdep_deps = 0;
+	if ((fs->fs_flags & FS_SUJ) &&
+	    (error = journal_mount(mp, fs, cred)) != 0) {
+		printf("Failed to start journal: %d\n", error);
+		return (error);
+	}
 	/*
 	 * When doing soft updates, the counters in the
 	 * superblock may have gotten out of sync. Recomputation
@@ -1493,6 +1914,2019 @@ softdep_mount(devvp, mp, fs, cred)
 	return (0);
 }
 
+void
+softdep_unmount(mp)
+	struct mount *mp;
+{
+
+	if (mp->mnt_kern_flag & MNTK_SUJ)
+		journal_unmount(mp);
+}
+
+struct jblocks {
+	struct jseglst	jb_segs;	/* TAILQ of current segments. */
+	struct jseg	*jb_writeseg;	/* Next write to complete. */
+	struct jextent	*jb_extent;	/* Extent array. */
+	uint64_t	jb_nextseq;	/* Next sequence number. */
+	uint64_t	jb_oldestseq;	/* Oldest active sequence number. */
+	int		jb_avail;	/* Available extents. */
+	int		jb_used;	/* Last used extent. */
+	int		jb_head;	/* Allocator head. */
+	int		jb_off;		/* Allocator extent offset. */
+	int		jb_blocks;	/* Total disk blocks covered. */
+	int		jb_free;	/* Total disk blocks free. */
+	int		jb_min;		/* Minimum free space. */
+	int		jb_low;		/* Low on space. */
+	int		jb_age;		/* Insertion time of oldest rec. */
+	int		jb_suspended;	/* Did journal suspend writes? */
+};
+
+struct jextent {
+	ufs2_daddr_t	je_daddr;	/* Disk block address. */
+	int		je_blocks;	/* Disk block count. */
+};
+
+static struct jblocks *
+jblocks_create(void)
+{
+	struct jblocks *jblocks;
+
+	jblocks = malloc(sizeof(*jblocks), M_JBLOCKS, M_WAITOK | M_ZERO);
+	TAILQ_INIT(&jblocks->jb_segs);
+	jblocks->jb_avail = 10;
+	jblocks->jb_extent = malloc(sizeof(struct jextent) * jblocks->jb_avail,
+	    M_JBLOCKS, M_WAITOK | M_ZERO);
+
+	return (jblocks);
+}
+
+static ufs2_daddr_t
+jblocks_alloc(jblocks, bytes, actual)
+	struct jblocks *jblocks;
+	int bytes;
+	int *actual;
+{
+	ufs2_daddr_t daddr;
+	struct jextent *jext;
+	int freecnt;
+	int blocks;
+
+	blocks = bytes / DEV_BSIZE;
+	jext = &jblocks->jb_extent[jblocks->jb_head];
+	freecnt = jext->je_blocks - jblocks->jb_off;
+	if (freecnt == 0) {
+		jblocks->jb_off = 0;
+		if (++jblocks->jb_head > jblocks->jb_used)
+			jblocks->jb_head = 0;
+		jext = &jblocks->jb_extent[jblocks->jb_head];
+		freecnt = jext->je_blocks;
+	}
+	if (freecnt > blocks)
+		freecnt = blocks;
+	*actual = freecnt * DEV_BSIZE;
+	daddr = jext->je_daddr + jblocks->jb_off;
+	jblocks->jb_off += freecnt;
+	jblocks->jb_free -= freecnt;
+
+	return (daddr);
+}
+
+static void
+jblocks_free(jblocks, mp, bytes)
+	struct jblocks *jblocks;
+	struct mount *mp;
+	int bytes;
+{
+
+	jblocks->jb_free += bytes / DEV_BSIZE;
+	if (jblocks->jb_suspended)
+		worklist_speedup();
+	wakeup(jblocks);
+}
+
+static void
+jblocks_destroy(jblocks)
+	struct jblocks *jblocks;
+{
+
+	if (jblocks->jb_extent)
+		free(jblocks->jb_extent, M_JBLOCKS);
+	free(jblocks, M_JBLOCKS);
+}
+
+static void
+jblocks_add(jblocks, daddr, blocks)
+	struct jblocks *jblocks;
+	ufs2_daddr_t daddr;
+	int blocks;
+{
+	struct jextent *jext;
+
+	jblocks->jb_blocks += blocks;
+	jblocks->jb_free += blocks;
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	/* Adding the first block. */
+	if (jext->je_daddr == 0) {
+		jext->je_daddr = daddr;
+		jext->je_blocks = blocks;
+		return;
+	}
+	/* Extending the last extent. */
+	if (jext->je_daddr + jext->je_blocks == daddr) {
+		jext->je_blocks += blocks;
+		return;
+	}
+	/* Adding a new extent. */
+	if (++jblocks->jb_used == jblocks->jb_avail) {
+		jblocks->jb_avail *= 2;
+		jext = malloc(sizeof(struct jextent) * jblocks->jb_avail,
+		    M_JBLOCKS, M_WAITOK | M_ZERO);
+		memcpy(jext, jblocks->jb_extent,
+		    sizeof(struct jextent) * jblocks->jb_used);
+		free(jblocks->jb_extent, M_JBLOCKS);
+		jblocks->jb_extent = jext;
+	}
+	jext = &jblocks->jb_extent[jblocks->jb_used];
+	jext->je_daddr = daddr;
+	jext->je_blocks = blocks;
+	return;
+}
+
+int
+softdep_journal_lookup(mp, vpp)
+	struct mount *mp;
+	struct vnode **vpp;
+{
+	struct componentname cnp;
+	struct vnode *dvp;
+	ino_t sujournal;
+	int error;
+
+	error = VFS_VGET(mp, ROOTINO, LK_EXCLUSIVE, &dvp);
+	if (error)
+		return (error);
+	bzero(&cnp, sizeof(cnp));
+	cnp.cn_nameiop = LOOKUP;
+	cnp.cn_flags = ISLASTCN;
+	cnp.cn_thread = curthread;
+	cnp.cn_cred = curthread->td_ucred;
+	cnp.cn_pnbuf = SUJ_FILE;
+	cnp.cn_nameptr = SUJ_FILE;
+	cnp.cn_namelen = strlen(SUJ_FILE);
+	error = ufs_lookup_ino(dvp, NULL, &cnp, &sujournal);
+	vput(dvp);
+	if (error != 0)
+		return (error);
+	error = VFS_VGET(mp, sujournal, LK_EXCLUSIVE, vpp);
+	return (error);
+}
+
+/*
+ * Open and verify the journal file.
+ */
+static int
+journal_mount(mp, fs, cred)
+	struct mount *mp;
+	struct fs *fs;
+	struct ucred *cred;
+{
+	struct jblocks *jblocks;
+	struct vnode *vp;
+	struct inode *ip;
+	ufs2_daddr_t blkno;
+	int bcount;
+	int error;
+	int i;
+
+	mp->mnt_kern_flag |= MNTK_SUJ;
+	error = softdep_journal_lookup(mp, &vp);
+	if (error != 0) {
+		printf("Failed to find journal.  Use tunefs to create one\n");
+		return (error);
+	}
+	ip = VTOI(vp);
+	if (ip->i_size < SUJ_MIN) {
+		error = ENOSPC;
+		goto out;
+	}
+	bcount = lblkno(fs, ip->i_size);	/* Only use whole blocks. */
+	jblocks = jblocks_create();
+	for (i = 0; i < bcount; i++) {
+		error = ufs_bmaparray(vp, i, &blkno, NULL, NULL, NULL);
+		if (error)
+			break;
+		jblocks_add(jblocks, blkno, fsbtodb(fs, fs->fs_frag));
+	}
+	if (error) {
+		jblocks_destroy(jblocks);
+		goto out;
+	}
+	jblocks->jb_low = jblocks->jb_free / 3;	/* Reserve 33%. */
+	jblocks->jb_min = jblocks->jb_free / 10; /* Suspend at 10%. */
+	/*
+	 * Only validate the journal contents if the filesystem is clean,
+	 * otherwise we write the logs but they'll never be used.  If the
+	 * filesystem was still dirty when we mounted it the journal is
+	 * invalid and a new journal can only be valid if it starts from a
+	 * clean mount.
+	 */
+	if (fs->fs_clean) {
+		DIP_SET(ip, i_modrev, fs->fs_mtime);
+		ip->i_flags |= IN_MODIFIED;
+		ffs_update(vp, 1);
+	}
+	VFSTOUFS(mp)->softdep_jblocks = jblocks;
+out:
+	vput(vp);
+	return (error);
+}
+
+static void
+journal_unmount(mp)
+	struct mount *mp;
+{
+	struct ufsmount *ump;
+
+	ump = VFSTOUFS(mp);
+	if (ump->softdep_jblocks)
+		jblocks_destroy(ump->softdep_jblocks);
+	ump->softdep_jblocks = NULL;
+}
+
+/*
+ * Called when a journal record is ready to be written.  Space is allocated
+ * and the journal entry is created when the journal is flushed to stable
+ * store.
+ */
+static void
+add_to_journal(wk)
+	struct worklist *wk;
+{
+	struct ufsmount *ump;
+
+	mtx_assert(&lk, MA_OWNED);
+	ump = VFSTOUFS(wk->wk_mp);
+	if (wk->wk_state & ONWORKLIST)
+		panic("add_to_journal: %s(0x%X) already on list",
+		    TYPENAME(wk->wk_type), wk->wk_state);
+	wk->wk_state |= ONWORKLIST | DEPCOMPLETE;
+	if (LIST_EMPTY(&ump->softdep_journal_pending)) {
+		ump->softdep_jblocks->jb_age = ticks;
+		LIST_INSERT_HEAD(&ump->softdep_journal_pending, wk, wk_list);
+	} else
+		LIST_INSERT_AFTER(ump->softdep_journal_tail, wk, wk_list);
+	ump->softdep_journal_tail = wk;
+	ump->softdep_on_journal += 1;
+}
+
+/*
+ * Remove an arbitrary item for the journal worklist maintain the tail
+ * pointer.  This happens when a new operation obviates the need to
+ * journal an old operation.
+ */
+static void
+remove_from_journal(wk)
+	struct worklist *wk;
+{
+	struct ufsmount *ump;
+
+	mtx_assert(&lk, MA_OWNED);
+	ump = VFSTOUFS(wk->wk_mp);
+#ifdef DEBUG	/* XXX Expensive, temporary. */
+	{
+		struct worklist *wkn;
+
+		LIST_FOREACH(wkn, &ump->softdep_journal_pending, wk_list)
+			if (wkn == wk)
+				break;
+		if (wkn == NULL)
+			panic("remove_from_journal: %p is not in journal", wk);
+	}
+#endif
+	/*
+	 * We emulate a TAILQ to save space in most structures which do not
+	 * require TAILQ semantics.  Here we must update the tail position
+	 * when removing the tail which is not the final entry.
+	 */
+	if (ump->softdep_journal_tail == wk)
+		ump->softdep_journal_tail =
+		    (struct worklist *)wk->wk_list.le_prev;
+
+	WORKLIST_REMOVE(wk);
+	ump->softdep_on_journal -= 1;
+}
+
+/*
+ * Check for journal space as well as dependency limits so the prelink
+ * code can throttle both journaled and non-journaled filesystems.
+ * Threshold is 0 for low and 1 for min.
+ */
+static int
+journal_space(ump, thresh)
+	struct ufsmount *ump;
+	int thresh;
+{
+	struct jblocks *jblocks;
+	int avail;
+
+	/*
+	 * We use a tighter restriction here to prevent request_cleanup()
+	 * running in threads from running into locks we currently hold.
+	 */
+	if (num_inodedep > (max_softdeps / 10) * 9)
+		return (0);
+
+	jblocks = ump->softdep_jblocks;
+	if (jblocks == NULL)
+		return (1);
+	if (thresh)
+		thresh = jblocks->jb_min;
+	else
+		thresh = jblocks->jb_low;
+	avail = (ump->softdep_on_journal * JREC_SIZE) / DEV_BSIZE;
+	avail = jblocks->jb_free - avail;
+
+	return (avail > thresh);
+}
+
+static void
+journal_suspend(ump)
+	struct ufsmount *ump;
+{
+	struct jblocks *jblocks;
+	struct mount *mp;
+
+	mp = UFSTOVFS(ump);
+	jblocks = ump->softdep_jblocks;
+	MNT_ILOCK(mp);
+	if ((mp->mnt_kern_flag & MNTK_SUSPEND) == 0) {
+		stat_journal_min++;
+		mp->mnt_kern_flag |= MNTK_SUSPEND;
+		mp->mnt_susp_owner = FIRST_THREAD_IN_PROC(softdepproc);
+	}
+	jblocks->jb_suspended = 1;
+	MNT_IUNLOCK(mp);
+}
+
+/*
+ * Called before any allocation function to be certain that there is
+ * sufficient space in the journal prior to creating any new records.
+ * Since in the case of block allocation we may have multiple locked
+ * buffers at the time of the actual allocation we can not block
+ * when the journal records are created.  Doing so would create a deadlock
+ * if any of these buffers needed to be flushed to reclaim space.  Instead
+ * we require a sufficiently large amount of available space such that
+ * each thread in the system could have passed this allocation check and
+ * still have sufficient free space.  With 20% of a minimum journal size
+ * of 1MB we have 6553 records available.
+ */
+int
+softdep_prealloc(vp, waitok)
+	struct vnode *vp;
+	int waitok;
+{
+	struct ufsmount *ump;
+
+	if (DOINGSUJ(vp) == 0)
+		return (0);
+	ump = VFSTOUFS(vp->v_mount);
+	ACQUIRE_LOCK(&lk);
+	if (journal_space(ump, 0)) {
+		FREE_LOCK(&lk);
+		return (0);
+	}
+	stat_journal_low++;
+	FREE_LOCK(&lk);
+	if (waitok == MNT_NOWAIT)
+		return (ENOSPC);
+	/*
+	 * Attempt to sync this vnode once to flush any journal
+	 * work attached to it.
+	 */
+	ffs_syncvnode(vp, waitok);
+	ACQUIRE_LOCK(&lk);
+	process_removes(vp);
+	if (journal_space(ump, 0) == 0) {
+		softdep_speedup();
+		if (journal_space(ump, 1) == 0)
+			journal_suspend(ump);
+	}
+	FREE_LOCK(&lk);
+
+	return (0);
+}
+
+/*
+ * Before adjusting a link count on a vnode verify that we have sufficient
+ * journal space.  If not, process operations that depend on the currently
+ * locked pair of vnodes to try to flush space as the syncer, buf daemon,
+ * and softdep flush threads can not acquire these locks to reclaim space.
+ */
+static void
+softdep_prelink(dvp, vp)
+	struct vnode *dvp;
+	struct vnode *vp;
+{
+	struct ufsmount *ump;
+
+	ump = VFSTOUFS(dvp->v_mount);
+	mtx_assert(&lk, MA_OWNED);
+	if (journal_space(ump, 0))
+		return;
+	stat_journal_low++;
+	FREE_LOCK(&lk);
+	if (vp)
+		ffs_syncvnode(vp, MNT_NOWAIT);
+	ffs_syncvnode(dvp, MNT_WAIT);
+	ACQUIRE_LOCK(&lk);
+	/* Process vp before dvp as it may create .. removes. */
+	if (vp)
+		process_removes(vp);
+	process_removes(dvp);
+	softdep_speedup();
+	process_worklist_item(UFSTOVFS(ump), LK_NOWAIT);
+	process_worklist_item(UFSTOVFS(ump), LK_NOWAIT);
+	if (journal_space(ump, 0) == 0) {
+		softdep_speedup();
+		if (journal_space(ump, 1) == 0)
+			journal_suspend(ump);
+	}
+}
+
+static void
+jseg_write(fs, jblocks, jseg, data)
+	struct fs *fs;
+	struct jblocks *jblocks;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jsegrec *rec;
+
+	rec = (struct jsegrec *)data;
+	rec->jsr_seq = jseg->js_seq;
+	rec->jsr_oldest = jblocks->jb_oldestseq;
+	rec->jsr_cnt = jseg->js_cnt;
+	rec->jsr_blocks = jseg->js_size / DEV_BSIZE;
+	rec->jsr_crc = 0;
+	rec->jsr_time = fs->fs_mtime;
+}
+
+static inline void
+inoref_write(inoref, jseg, rec)
+	struct inoref *inoref;
+	struct jseg *jseg;
+	struct jrefrec *rec;
+{
+
+	inoref->if_jsegdep->jd_seg = jseg;
+	rec->jr_ino = inoref->if_ino;
+	rec->jr_parent = inoref->if_parent;
+	rec->jr_nlink = inoref->if_nlink;
+	rec->jr_mode = inoref->if_mode;
+	rec->jr_diroff = inoref->if_diroff;
+}
+
+static void
+jaddref_write(jaddref, jseg, data)
+	struct jaddref *jaddref;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jrefrec *rec;
+
+	rec = (struct jrefrec *)data;
+	rec->jr_op = JOP_ADDREF;
+	inoref_write(&jaddref->ja_ref, jseg, rec);
+}
+
+static void
+jremref_write(jremref, jseg, data)
+	struct jremref *jremref;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jrefrec *rec;
+
+	rec = (struct jrefrec *)data;
+	rec->jr_op = JOP_REMREF;
+	inoref_write(&jremref->jr_ref, jseg, rec);
+}
+
+static	void
+jmvref_write(jmvref, jseg, data)
+	struct jmvref *jmvref;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jmvrec *rec;
+
+	rec = (struct jmvrec *)data;
+	rec->jm_op = JOP_MVREF;
+	rec->jm_ino = jmvref->jm_ino;
+	rec->jm_parent = jmvref->jm_parent;
+	rec->jm_oldoff = jmvref->jm_oldoff;
+	rec->jm_newoff = jmvref->jm_newoff;
+}
+
+static void
+jnewblk_write(jnewblk, jseg, data)
+	struct jnewblk *jnewblk;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jblkrec *rec;
+
+	jnewblk->jn_jsegdep->jd_seg = jseg;
+	rec = (struct jblkrec *)data;
+	rec->jb_op = JOP_NEWBLK;
+	rec->jb_ino = jnewblk->jn_ino;
+	rec->jb_blkno = jnewblk->jn_blkno;
+	rec->jb_lbn = jnewblk->jn_lbn;
+	rec->jb_frags = jnewblk->jn_frags;
+	rec->jb_oldfrags = jnewblk->jn_oldfrags;
+}
+
+static void
+jfreeblk_write(jfreeblk, jseg, data)
+	struct jfreeblk *jfreeblk;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jblkrec *rec;
+
+	jfreeblk->jf_jsegdep->jd_seg = jseg;
+	rec = (struct jblkrec *)data;
+	rec->jb_op = JOP_FREEBLK;
+	rec->jb_ino = jfreeblk->jf_ino;
+	rec->jb_blkno = jfreeblk->jf_blkno;
+	rec->jb_lbn = jfreeblk->jf_lbn;
+	rec->jb_frags = jfreeblk->jf_frags;
+	rec->jb_oldfrags = 0;
+}
+
+static void
+jfreefrag_write(jfreefrag, jseg, data)
+	struct jfreefrag *jfreefrag;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jblkrec *rec;
+
+	jfreefrag->fr_jsegdep->jd_seg = jseg;
+	rec = (struct jblkrec *)data;
+	rec->jb_op = JOP_FREEBLK;
+	rec->jb_ino = jfreefrag->fr_ino;
+	rec->jb_blkno = jfreefrag->fr_blkno;
+	rec->jb_lbn = jfreefrag->fr_lbn;
+	rec->jb_frags = jfreefrag->fr_frags;
+	rec->jb_oldfrags = 0;
+}
+
+static void
+jtrunc_write(jtrunc, jseg, data)
+	struct jtrunc *jtrunc;
+	struct jseg *jseg;
+	uint8_t *data;
+{
+	struct jtrncrec *rec;
+
+	rec = (struct jtrncrec *)data;
+	rec->jt_op = JOP_TRUNC;
+	rec->jt_ino = jtrunc->jt_ino;
+	rec->jt_size = jtrunc->jt_size;
+	rec->jt_extsize = jtrunc->jt_extsize;
+}
+
+/*
+ * Flush some journal records to disk.
+ */
+static void
+softdep_process_journal(mp, flags)
+	struct mount *mp;
+	int flags;
+{
+	struct jblocks *jblocks;
+	struct ufsmount *ump;
+	struct worklist *wk;
+	struct jseg *jseg;
+	struct buf *bp;
+	uint8_t *data;
+	struct fs *fs;
+	int segwritten;
+	int jrecmin;	/* Minimum records per block. */
+	int jrecmax;	/* Maximum records per block. */
+	int size;
+	int cnt;
+	int off;
+
+	if ((mp->mnt_kern_flag & MNTK_SUJ) == 0)
+		return;
+	ump = VFSTOUFS(mp);
+	fs = ump->um_fs;
+	jblocks = ump->softdep_jblocks;
+	/*
+	 * We write anywhere between a disk block and fs block.  The upper
+	 * bound is picked to prevent buffer cache fragmentation and limit
+	 * processing time per I/O.
+	 */
+	jrecmin = (DEV_BSIZE / JREC_SIZE) - 1; /* -1 for seg header */
+	jrecmax = (fs->fs_bsize / DEV_BSIZE) * jrecmin;
+	segwritten = 0;
+	while ((cnt = ump->softdep_on_journal) != 0) {
+		/*
+		 * Create a new segment to hold as many as 'cnt' journal
+		 * entries and add them to the segment.  Notice cnt is
+		 * off by one to account for the space required by the
+		 * jsegrec.  If we don't have a full block to log skip it
+		 * unless we haven't written anything.
+		 */
+		cnt++;
+		if (cnt < jrecmax && segwritten)
+			break;
+		/*
+		 * Verify some free journal space.  softdep_prealloc() should
+	 	 * guarantee that we don't run out so this is indicative of
+		 * a problem with the flow control.  Try to recover
+		 * gracefully in any event.
+		 */
+		while (jblocks->jb_free == 0) {
+			if (flags != MNT_WAIT)
+				break;
+			printf("softdep: Out of journal space!\n");
+			softdep_speedup();
+			msleep(jblocks, &lk, PRIBIO, "jblocks", 1);
+		}
+		FREE_LOCK(&lk);
+		jseg = malloc(sizeof(*jseg), M_JSEG, M_SOFTDEP_FLAGS);
+		workitem_alloc(&jseg->js_list, D_JSEG, mp);
+		LIST_INIT(&jseg->js_entries);
+		jseg->js_state = ATTACHED;
+		jseg->js_jblocks = jblocks;
+		bp = geteblk(fs->fs_bsize, 0);
+		ACQUIRE_LOCK(&lk);
+		/*
+		 * If there was a race while we were allocating the block
+		 * and jseg the entry we care about was likely written.
+		 * We bail out in both the WAIT and NOWAIT case and assume
+		 * the caller will loop if the entry it cares about is
+		 * not written.
+		 */
+		if (ump->softdep_on_journal == 0 || jblocks->jb_free == 0) {
+			bp->b_flags |= B_INVAL | B_NOCACHE;
+			WORKITEM_FREE(jseg, D_JSEG);
+			FREE_LOCK(&lk);
+			brelse(bp);
+			ACQUIRE_LOCK(&lk);
+			break;
+		}
+		/*
+		 * Calculate the disk block size required for the available
+		 * records rounded to the min size.
+		 */
+		cnt = ump->softdep_on_journal;
+		if (cnt < jrecmax)
+			size = howmany(cnt, jrecmin) * DEV_BSIZE;
+		else
+			size = fs->fs_bsize;
+		/*
+		 * Allocate a disk block for this journal data and account
+		 * for truncation of the requested size if enough contiguous
+		 * space was not available.
+		 */
+		bp->b_blkno = jblocks_alloc(jblocks, size, &size);
+		bp->b_lblkno = bp->b_blkno;
+		bp->b_offset = bp->b_blkno * DEV_BSIZE;
+		bp->b_bcount = size;
+		bp->b_bufobj = &ump->um_devvp->v_bufobj;
+		bp->b_flags &= ~B_INVAL;
+		bp->b_flags |= B_VALIDSUSPWRT | B_NOCOPY;
+		/*
+		 * Initialize our jseg with cnt records.  Assign the next
+		 * sequence number to it and link it in-order.
+		 */
+		cnt = MIN(ump->softdep_on_journal,
+		    (size / DEV_BSIZE) * jrecmin);
+		jseg->js_buf = bp;
+		jseg->js_cnt = cnt;
+		jseg->js_refs = cnt + 1;	/* Self ref. */
+		jseg->js_size = size;
+		jseg->js_seq = jblocks->jb_nextseq++;
+		if (TAILQ_EMPTY(&jblocks->jb_segs))
+			jblocks->jb_oldestseq = jseg->js_seq;
+		TAILQ_INSERT_TAIL(&jblocks->jb_segs, jseg, js_next);
+		if (jblocks->jb_writeseg == NULL)
+			jblocks->jb_writeseg = jseg;
+		/*
+		 * Start filling in records from the pending list.
+		 */
+		data = bp->b_data;
+		off = 0;
+		while ((wk = LIST_FIRST(&ump->softdep_journal_pending))
+		    != NULL) {
+			/* Place a segment header on every device block. */
+			if ((off % DEV_BSIZE) == 0) {
+				jseg_write(fs, jblocks, jseg, data);
+				off += JREC_SIZE;
+				data = bp->b_data + off;
+			}
+			remove_from_journal(wk);
+			wk->wk_state |= IOSTARTED;
+			WORKLIST_INSERT(&jseg->js_entries, wk);
+			switch (wk->wk_type) {
+			case D_JADDREF:
+				jaddref_write(WK_JADDREF(wk), jseg, data);
+				break;
+			case D_JREMREF:
+				jremref_write(WK_JREMREF(wk), jseg, data);
+				break;
+			case D_JMVREF:
+				jmvref_write(WK_JMVREF(wk), jseg, data);
+				break;
+			case D_JNEWBLK:
+				jnewblk_write(WK_JNEWBLK(wk), jseg, data);
+				break;
+			case D_JFREEBLK:
+				jfreeblk_write(WK_JFREEBLK(wk), jseg, data);
+				break;
+			case D_JFREEFRAG:
+				jfreefrag_write(WK_JFREEFRAG(wk), jseg, data);
+				break;
+			case D_JTRUNC:
+				jtrunc_write(WK_JTRUNC(wk), jseg, data);
+				break;
+			default:
+				panic("process_journal: Unknown type %s",
+				    TYPENAME(wk->wk_type));
+				/* NOTREACHED */
+			}
+			if (--cnt == 0)
+				break;
+			off += JREC_SIZE;
+			data = bp->b_data + off;
+		}
+		/*
+		 * Write this one buffer and continue.
+		 */
+		WORKLIST_INSERT(&bp->b_dep, &jseg->js_list);
+		FREE_LOCK(&lk);
+		BO_LOCK(bp->b_bufobj);
+		bgetvp(ump->um_devvp, bp);
+		BO_UNLOCK(bp->b_bufobj);
+		if (flags == MNT_NOWAIT)
+			bawrite(bp);
+		else
+			bwrite(bp);
+		ACQUIRE_LOCK(&lk);
+	}
+	/*
+	 * If we've suspended the filesystem because we ran out of journal
+	 * space either try to sync it here to make some progress or
+	 * unsuspend it if we already have.
+	 */
+	if (flags == 0 && jblocks && jblocks->jb_suspended) {
+		if (journal_space(ump, jblocks->jb_min)) {
+			FREE_LOCK(&lk);
+			jblocks->jb_suspended = 0;
+			mp->mnt_susp_owner = curthread;
+			vfs_write_resume(mp);
+			ACQUIRE_LOCK(&lk);
+			return;
+		}
+		FREE_LOCK(&lk);
+		VFS_SYNC(mp, MNT_NOWAIT);
+		ffs_sbupdate(ump, MNT_WAIT, 0);
+		ACQUIRE_LOCK(&lk);
+	}
+}
+
+/*
+ * Complete a jseg, allowing all dependencies awaiting journal writes
+ * to proceed.  Each journal dependency also attaches a jsegdep to dependent
+ * structures so that the journal segment can be freed to reclaim space.
+ */
+static void
+complete_jseg(jseg)
+	struct jseg *jseg;
+{
+	struct worklist *wk;
+	struct jmvref *jmvref;
+	int waiting;
+	int i;
+
+	i = 0;
+	while ((wk = LIST_FIRST(&jseg->js_entries)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		waiting = wk->wk_state & IOWAITING;
+		wk->wk_state &= ~(IOSTARTED | IOWAITING);
+		wk->wk_state |= COMPLETE;
+		KASSERT(i < jseg->js_cnt,
+		    ("handle_written_jseg: overflow %d >= %d",
+		    i, jseg->js_cnt));
+		switch (wk->wk_type) {
+		case D_JADDREF:
+			handle_written_jaddref(WK_JADDREF(wk));
+			break;
+		case D_JREMREF:
+			handle_written_jremref(WK_JREMREF(wk));
+			break;
+		case D_JMVREF:
+			/* No jsegdep here. */
+			free_jseg(jseg);
+			jmvref = WK_JMVREF(wk);
+			LIST_REMOVE(jmvref, jm_deps);
+			free_pagedep(jmvref->jm_pagedep);
+			WORKITEM_FREE(jmvref, D_JMVREF);
+			break;
+		case D_JNEWBLK:
+			handle_written_jnewblk(WK_JNEWBLK(wk));
+			break;
+		case D_JFREEBLK:
+			handle_written_jfreeblk(WK_JFREEBLK(wk));
+			break;
+		case D_JFREEFRAG:
+			handle_written_jfreefrag(WK_JFREEFRAG(wk));
+			break;
+		case D_JTRUNC:
+			WK_JTRUNC(wk)->jt_jsegdep->jd_seg = jseg;
+			WORKITEM_FREE(wk, D_JTRUNC);
+			break;
+		default:
+			panic("handle_written_jseg: Unknown type %s",
+			    TYPENAME(wk->wk_type));
+			/* NOTREACHED */
+		}
+		if (waiting)
+			wakeup(wk);
+	}
+	/* Release the self reference so the structure may be freed. */
+	free_jseg(jseg);
+}
+
+/*
+ * Mark a jseg as DEPCOMPLETE and throw away the buffer.  Handle jseg
+ * completions in order only.
+ */
+static void
+handle_written_jseg(jseg, bp)
+	struct jseg *jseg;
+	struct buf *bp;
+{
+	struct jblocks *jblocks;
+	struct jseg *jsegn;
+
+	if (jseg->js_refs == 0)
+		panic("handle_written_jseg: No self-reference on %p", jseg);
+	jseg->js_state |= DEPCOMPLETE;
+	/*
+	 * We'll never need this buffer again, set flags so it will be
+	 * discarded.
+	 */
+	bp->b_flags |= B_INVAL | B_NOCACHE;
+	jblocks = jseg->js_jblocks;
+	/*
+	 * Don't allow out of order completions.  If this isn't the first
+	 * block wait for it to write before we're done.
+	 */
+	if (jseg != jblocks->jb_writeseg)
+		return;
+	/* Iterate through available jsegs processing their entries. */
+	do {
+		jsegn = TAILQ_NEXT(jseg, js_next);
+		complete_jseg(jseg);
+		jseg = jsegn;
+	} while (jseg && jseg->js_state & DEPCOMPLETE);
+	jblocks->jb_writeseg = jseg;
+}
+
+static inline struct jsegdep *
+inoref_jseg(inoref)
+	struct inoref *inoref;
+{
+	struct jsegdep *jsegdep;
+
+	jsegdep = inoref->if_jsegdep;
+	inoref->if_jsegdep = NULL;
+
+	return (jsegdep);
+}
+
+/*
+ * Called once a jremref has made it to stable store.  The jremref is marked
+ * complete and we attempt to free it.  Any pagedeps writes sleeping waiting
+ * for the jremref to complete will be awoken by free_jremref.
+ */
+static void
+handle_written_jremref(jremref)
+	struct jremref *jremref;
+{
+	struct inodedep *inodedep;
+	struct jsegdep *jsegdep;
+	struct dirrem *dirrem;
+
+	/* Grab the jsegdep. */
+	jsegdep = inoref_jseg(&jremref->jr_ref);
+	/*
+	 * Remove us from the inoref list.
+	 */
+	if (inodedep_lookup(jremref->jr_list.wk_mp, jremref->jr_ref.if_ino,
+	    0, &inodedep) == 0)
+		panic("handle_written_jremref: Lost inodedep");
+	TAILQ_REMOVE(&inodedep->id_inoreflst, &jremref->jr_ref, if_deps);
+	/*
+	 * Complete the dirrem.
+	 */
+	dirrem = jremref->jr_dirrem;
+	jremref->jr_dirrem = NULL;
+	LIST_REMOVE(jremref, jr_deps);
+	jsegdep->jd_state |= jremref->jr_state & MKDIR_PARENT;
+	WORKLIST_INSERT(&dirrem->dm_jwork, &jsegdep->jd_list);
+	if (LIST_EMPTY(&dirrem->dm_jremrefhd) &&
+	    (dirrem->dm_state & COMPLETE) != 0)
+		add_to_worklist(&dirrem->dm_list, 0);
+	free_jremref(jremref);
+}
+
+/*
+ * Called once a jaddref has made it to stable store.  The dependency is
+ * marked complete and any dependent structures are added to the inode
+ * bufwait list to be completed as soon as it is written.  If a bitmap write
+ * depends on this entry we move the inode into the inodedephd of the
+ * bmsafemap dependency and attempt to remove the jaddref from the bmsafemap.
+ */
+static void
+handle_written_jaddref(jaddref)
+	struct jaddref *jaddref;
+{
+	struct jsegdep *jsegdep;
+	struct inodedep *inodedep;
+	struct diradd *diradd;
+	struct mkdir *mkdir;
+
+	/* Grab the jsegdep. */
+	jsegdep = inoref_jseg(&jaddref->ja_ref);
+	mkdir = NULL;
+	diradd = NULL;
+	if (inodedep_lookup(jaddref->ja_list.wk_mp, jaddref->ja_ino,
+	    0, &inodedep) == 0)
+		panic("handle_written_jaddref: Lost inodedep.");
+	if (jaddref->ja_diradd == NULL)
+		panic("handle_written_jaddref: No dependency");
+	if (jaddref->ja_diradd->da_list.wk_type == D_DIRADD) {
+		diradd = jaddref->ja_diradd;
+		WORKLIST_INSERT(&inodedep->id_bufwait, &diradd->da_list);
+	} else if (jaddref->ja_state & MKDIR_PARENT) {
+		mkdir = jaddref->ja_mkdir;
+		WORKLIST_INSERT(&inodedep->id_bufwait, &mkdir->md_list);
+	} else if (jaddref->ja_state & MKDIR_BODY)
+		mkdir = jaddref->ja_mkdir;
+	else
+		panic("handle_written_jaddref: Unknown dependency %p",
+		    jaddref->ja_diradd);
+	jaddref->ja_diradd = NULL;	/* also clears ja_mkdir */
+	/*
+	 * Remove us from the inode list.
+	 */
+	TAILQ_REMOVE(&inodedep->id_inoreflst, &jaddref->ja_ref, if_deps);
+	/*
+	 * The mkdir may be waiting on the jaddref to clear before freeing.
+	 */
+	if (mkdir) {
+		KASSERT(mkdir->md_list.wk_type == D_MKDIR,
+		    ("handle_written_jaddref: Incorrect type for mkdir %s",
+		    TYPENAME(mkdir->md_list.wk_type)));
+		mkdir->md_jaddref = NULL;
+		diradd = mkdir->md_diradd;
+		mkdir->md_state |= DEPCOMPLETE;
+		complete_mkdir(mkdir);
+	}
+	WORKLIST_INSERT(&diradd->da_jwork, &jsegdep->jd_list);
+	if (jaddref->ja_state & NEWBLOCK) {
+		inodedep->id_state |= ONDEPLIST;
+		LIST_INSERT_HEAD(&inodedep->id_bmsafemap->sm_inodedephd,
+		    inodedep, id_deps);
+	}
+	free_jaddref(jaddref);
+}
+
+/*
+ * Called once a jnewblk journal is written.  The allocdirect or allocindir
+ * is placed in the bmsafemap to await notification of a written bitmap.
+ */
+static void
+handle_written_jnewblk(jnewblk)
+	struct jnewblk *jnewblk;
+{
+	struct bmsafemap *bmsafemap;
+	struct jsegdep *jsegdep;
+	struct newblk *newblk;
+
+	/* Grab the jsegdep. */
+	jsegdep = jnewblk->jn_jsegdep;
+	jnewblk->jn_jsegdep = NULL;
+	/*
+	 * Add the written block to the bmsafemap so it can be notified when
+	 * the bitmap is on disk.
+	 */
+	newblk = jnewblk->jn_newblk;
+	jnewblk->jn_newblk = NULL;
+	if (newblk == NULL) 
+		panic("handle_written_jnewblk: No dependency for the segdep.");
+
+	newblk->nb_jnewblk = NULL;
+	bmsafemap = newblk->nb_bmsafemap;
+	WORKLIST_INSERT(&newblk->nb_jwork, &jsegdep->jd_list);
+	newblk->nb_state |= ONDEPLIST;
+	LIST_INSERT_HEAD(&bmsafemap->sm_newblkhd, newblk, nb_deps);
+	free_jnewblk(jnewblk);
+}
+
+/*
+ * Cancel a jfreefrag that won't be needed, probably due to colliding with
+ * an in-flight allocation that has not yet been committed.  Divorce us
+ * from the freefrag and mark it DEPCOMPLETE so that it may be added
+ * to the worklist.
+ */
+static void
+cancel_jfreefrag(jfreefrag)
+	struct jfreefrag *jfreefrag;
+{
+	struct freefrag *freefrag;
+
+	if (jfreefrag->fr_jsegdep) {
+		free_jsegdep(jfreefrag->fr_jsegdep);
+		jfreefrag->fr_jsegdep = NULL;
+	}
+	freefrag = jfreefrag->fr_freefrag;
+	jfreefrag->fr_freefrag = NULL;
+	freefrag->ff_jfreefrag = NULL;
+	free_jfreefrag(jfreefrag);
+	freefrag->ff_state |= DEPCOMPLETE;
+}
+
+/*
+ * Free a jfreefrag when the parent freefrag is rendered obsolete.
+ */
+static void
+free_jfreefrag(jfreefrag)
+	struct jfreefrag *jfreefrag;
+{
+
+	if (jfreefrag->fr_state & IOSTARTED)
+		WORKLIST_REMOVE(&jfreefrag->fr_list);
+	else if (jfreefrag->fr_state & ONWORKLIST)
+		remove_from_journal(&jfreefrag->fr_list);
+	if (jfreefrag->fr_freefrag != NULL)
+		panic("free_jfreefrag:  Still attached to a freefrag.");
+	WORKITEM_FREE(jfreefrag, D_JFREEFRAG);
+}
+
+/*
+ * Called when the journal write for a jfreefrag completes.  The parent
+ * freefrag is added to the worklist if this completes its dependencies.
+ */
+static void
+handle_written_jfreefrag(jfreefrag)
+	struct jfreefrag *jfreefrag;
+{
+	struct jsegdep *jsegdep;
+	struct freefrag *freefrag;
+
+	/* Grab the jsegdep. */
+	jsegdep = jfreefrag->fr_jsegdep;
+	jfreefrag->fr_jsegdep = NULL;
+	freefrag = jfreefrag->fr_freefrag;
+	if (freefrag == NULL)
+		panic("handle_written_jfreefrag: No freefrag.");
+	freefrag->ff_state |= DEPCOMPLETE;
+	freefrag->ff_jfreefrag = NULL;
+	WORKLIST_INSERT(&freefrag->ff_jwork, &jsegdep->jd_list);
+	if ((freefrag->ff_state & ALLCOMPLETE) == ALLCOMPLETE)
+		add_to_worklist(&freefrag->ff_list, 0);
+	jfreefrag->fr_freefrag = NULL;
+	free_jfreefrag(jfreefrag);
+}
+
+/*
+ * Called when the journal write for a jfreeblk completes.  The jfreeblk
+ * is removed from the freeblks list of pending journal writes and the
+ * jsegdep is moved to the freeblks jwork to be completed when all blocks
+ * have been reclaimed.
+ */
+static void
+handle_written_jfreeblk(jfreeblk)
+	struct jfreeblk *jfreeblk;
+{
+	struct freeblks *freeblks;
+	struct jsegdep *jsegdep;
+
+	/* Grab the jsegdep. */
+	jsegdep = jfreeblk->jf_jsegdep;
+	jfreeblk->jf_jsegdep = NULL;
+	freeblks = jfreeblk->jf_freeblks;
+	LIST_REMOVE(jfreeblk, jf_deps);
+	WORKLIST_INSERT(&freeblks->fb_jwork, &jsegdep->jd_list);
+	/*
+	 * If the freeblks is all journaled, we can add it to the worklist.
+	 */
+	if (LIST_EMPTY(&freeblks->fb_jfreeblkhd) &&
+	    (freeblks->fb_state & ALLCOMPLETE) == ALLCOMPLETE) {
+		/* Remove from the b_dep that is waiting on this write. */
+		if (freeblks->fb_state & ONWORKLIST)
+			WORKLIST_REMOVE(&freeblks->fb_list);
+		add_to_worklist(&freeblks->fb_list, 1);
+	}
+
+	free_jfreeblk(jfreeblk);
+}
+
+static struct jsegdep *
+newjsegdep(struct worklist *wk)
+{
+	struct jsegdep *jsegdep;
+
+	jsegdep = malloc(sizeof(*jsegdep), M_JSEGDEP, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jsegdep->jd_list, D_JSEGDEP, wk->wk_mp);
+	jsegdep->jd_seg = NULL;
+
+	return (jsegdep);
+}
+
+static struct jmvref *
+newjmvref(dp, ino, oldoff, newoff)
+	struct inode *dp;
+	ino_t ino;
+	off_t oldoff;
+	off_t newoff;
+{
+	struct jmvref *jmvref;
+
+	jmvref = malloc(sizeof(*jmvref), M_JMVREF, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jmvref->jm_list, D_JMVREF, UFSTOVFS(dp->i_ump));
+	jmvref->jm_list.wk_state = ATTACHED | DEPCOMPLETE;
+	jmvref->jm_parent = dp->i_number;
+	jmvref->jm_ino = ino;
+	jmvref->jm_oldoff = oldoff;
+	jmvref->jm_newoff = newoff;
+
+	return (jmvref);
+}
+
+/*
+ * Allocate a new jremref that tracks the removal of ip from dp with the
+ * directory entry offset of diroff.  Mark the entry as ATTACHED and
+ * DEPCOMPLETE as we have all the information required for the journal write
+ * and the directory has already been removed from the buffer.  The caller
+ * is responsible for linking the jremref into the pagedep and adding it
+ * to the journal to write.  The MKDIR_PARENT flag is set if we're doing
+ * a DOTDOT addition so handle_workitem_remove() can properly assign
+ * the jsegdep when we're done.
+ */
+static struct jremref *
+newjremref(dirrem, dp, ip, diroff, nlink)
+	struct dirrem *dirrem;
+	struct inode *dp;
+	struct inode *ip;
+	off_t diroff;
+	nlink_t nlink;
+{
+	struct jremref *jremref;
+
+	jremref = malloc(sizeof(*jremref), M_JREMREF, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jremref->jr_list, D_JREMREF, UFSTOVFS(dp->i_ump));
+	jremref->jr_state = ATTACHED;
+	newinoref(&jremref->jr_ref, ip->i_number, dp->i_number, diroff,
+	   nlink, ip->i_mode);
+	jremref->jr_dirrem = dirrem;
+
+	return (jremref);
+}
+
+static inline void
+newinoref(inoref, ino, parent, diroff, nlink, mode)
+	struct inoref *inoref;
+	ino_t ino;
+	ino_t parent;
+	off_t diroff;
+	nlink_t nlink;
+	uint16_t mode;
+{
+
+	inoref->if_jsegdep = newjsegdep(&inoref->if_list);
+	inoref->if_diroff = diroff;
+	inoref->if_ino = ino;
+	inoref->if_parent = parent;
+	inoref->if_nlink = nlink;
+	inoref->if_mode = mode;
+}
+
+/*
+ * Allocate a new jaddref to track the addition of ino to dp at diroff.  The
+ * directory offset may not be known until later.  The caller is responsible
+ * adding the entry to the journal when this information is available.  nlink
+ * should be the link count prior to the addition and mode is only required
+ * to have the correct FMT.
+ */
+static struct jaddref *
+newjaddref(dp, ino, diroff, nlink, mode)
+	struct inode *dp;
+	ino_t ino;
+	off_t diroff;
+	int16_t nlink;
+	uint16_t mode;
+{
+	struct jaddref *jaddref;
+
+	jaddref = malloc(sizeof(*jaddref), M_JADDREF, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jaddref->ja_list, D_JADDREF, UFSTOVFS(dp->i_ump));
+	jaddref->ja_state = ATTACHED;
+	jaddref->ja_mkdir = NULL;
+	newinoref(&jaddref->ja_ref, ino, dp->i_number, diroff, nlink, mode);
+
+	return (jaddref);
+}
+
+/*
+ * Create a new free dependency for a freework.  The caller is responsible
+ * for adjusting the reference count when it has the lock held.  The freedep
+ * will track an outstanding bitmap write that will ultimately clear the
+ * freework to continue.
+ */
+static struct freedep *
+newfreedep(struct freework *freework)
+{
+	struct freedep *freedep;
+
+	freedep = malloc(sizeof(*freedep), M_FREEDEP, M_SOFTDEP_FLAGS);
+	workitem_alloc(&freedep->fd_list, D_FREEDEP, freework->fw_list.wk_mp);
+	freedep->fd_freework = freework;
+
+	return (freedep);
+}
+
+/*
+ * Free a freedep structure once the buffer it is linked to is written.  If
+ * this is the last reference to the freework schedule it for completion.
+ */
+static void
+free_freedep(freedep)
+	struct freedep *freedep;
+{
+
+	if (--freedep->fd_freework->fw_ref == 0)
+		add_to_worklist(&freedep->fd_freework->fw_list, 1);
+	WORKITEM_FREE(freedep, D_FREEDEP);
+}
+
+/*
+ * Allocate a new freework structure that may be a level in an indirect
+ * when parent is not NULL or a top level block when it is.  The top level
+ * freework structures are allocated without lk held and before the freeblks
+ * is visible outside of softdep_setup_freeblocks().
+ */
+static struct freework *
+newfreework(freeblks, parent, lbn, nb, frags, journal)
+	struct freeblks *freeblks;
+	struct freework *parent;
+	ufs_lbn_t lbn;
+	ufs2_daddr_t nb;
+	int frags;
+	int journal;
+{
+	struct freework *freework;
+
+	freework = malloc(sizeof(*freework), M_FREEWORK, M_SOFTDEP_FLAGS);
+	workitem_alloc(&freework->fw_list, D_FREEWORK, freeblks->fb_list.wk_mp);
+	freework->fw_freeblks = freeblks;
+	freework->fw_parent = parent;
+	freework->fw_lbn = lbn;
+	freework->fw_blkno = nb;
+	freework->fw_frags = frags;
+	freework->fw_ref = 0;
+	freework->fw_off = 0;
+	LIST_INIT(&freework->fw_jwork);
+
+	if (parent == NULL) {
+		WORKLIST_INSERT_UNLOCKED(&freeblks->fb_freeworkhd,
+		    &freework->fw_list);
+		freeblks->fb_ref++;
+	}
+	if (journal)
+		newjfreeblk(freeblks, lbn, nb, frags);
+
+	return (freework);
+}
+
+/*
+ * Allocate a new jfreeblk to journal top level block pointer when truncating
+ * a file.  The caller must add this to the worklist when lk is held.
+ */
+static struct jfreeblk *
+newjfreeblk(freeblks, lbn, blkno, frags)
+	struct freeblks *freeblks;
+	ufs_lbn_t lbn;
+	ufs2_daddr_t blkno;
+	int frags;
+{
+	struct jfreeblk *jfreeblk;
+
+	jfreeblk = malloc(sizeof(*jfreeblk), M_JFREEBLK, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jfreeblk->jf_list, D_JFREEBLK, freeblks->fb_list.wk_mp);
+	jfreeblk->jf_jsegdep = newjsegdep(&jfreeblk->jf_list);
+	jfreeblk->jf_state = ATTACHED | DEPCOMPLETE;
+	jfreeblk->jf_ino = freeblks->fb_previousinum;
+	jfreeblk->jf_lbn = lbn;
+	jfreeblk->jf_blkno = blkno;
+	jfreeblk->jf_frags = frags;
+	jfreeblk->jf_freeblks = freeblks;
+	LIST_INSERT_HEAD(&freeblks->fb_jfreeblkhd, jfreeblk, jf_deps);
+
+	return (jfreeblk);
+}
+
+static void move_newblock_dep(struct jaddref *, struct inodedep *);
+/*
+ * If we're canceling a new bitmap we have to search for another ref
+ * to move into the bmsafemap dep.  This might be better expressed
+ * with another structure.
+ */
+static void
+move_newblock_dep(jaddref, inodedep)
+	struct jaddref *jaddref;
+	struct inodedep *inodedep;
+{
+	struct inoref *inoref;
+	struct jaddref *jaddrefn;
+
+	jaddrefn = NULL;
+	for (inoref = TAILQ_NEXT(&jaddref->ja_ref, if_deps); inoref;
+	    inoref = TAILQ_NEXT(inoref, if_deps)) {
+		if ((jaddref->ja_state & NEWBLOCK) &&
+		    inoref->if_list.wk_type == D_JADDREF) {
+			jaddrefn = (struct jaddref *)inoref;
+			break;
+		}
+	}
+	if (jaddrefn == NULL)
+		return;
+	jaddrefn->ja_state &= ~(ATTACHED | UNDONE);
+	jaddrefn->ja_state |= jaddref->ja_state &
+	    (ATTACHED | UNDONE | NEWBLOCK);
+	jaddref->ja_state &= ~(ATTACHED | UNDONE | NEWBLOCK);
+	jaddref->ja_state |= ATTACHED;
+	LIST_REMOVE(jaddref, ja_bmdeps);
+	LIST_INSERT_HEAD(&inodedep->id_bmsafemap->sm_jaddrefhd, jaddrefn,
+	    ja_bmdeps);
+}
+
+/*
+ * Cancel a jaddref either before it has been written or while it is being
+ * written.  This happens when a link is removed before the add reaches
+ * the disk.  The jaddref dependency is kept linked into the bmsafemap
+ * and inode to prevent the link count or bitmap from reaching the disk
+ * until handle_workitem_remove() re-adjusts the counts and bitmaps as
+ * required.
+ *
+ * Returns 1 if the canceled addref requires journaling of the remove and
+ * 0 otherwise.
+ */
+static int
+cancel_jaddref(jaddref, inodedep, wkhd)
+	struct jaddref *jaddref;
+	struct inodedep *inodedep;
+	struct workhead *wkhd;
+{
+	struct inoref *inoref;
+	struct jsegdep *jsegdep;
+	int needsj;
+
+	KASSERT((jaddref->ja_state & COMPLETE) == 0,
+	    ("cancel_jaddref: Canceling complete jaddref"));
+	if (jaddref->ja_state & (IOSTARTED | COMPLETE))
+		needsj = 1;
+	else
+		needsj = 0;
+	if (inodedep == NULL)
+		if (inodedep_lookup(jaddref->ja_list.wk_mp, jaddref->ja_ino,
+		    0, &inodedep) == 0)
+			panic("cancel_jaddref: Lost inodedep");
+	/*
+	 * We must adjust the nlink of any reference operation that follows
+	 * us so that it is consistent with the in-memory reference.  This
+	 * ensures that inode nlink rollbacks always have the correct link.
+	 */
+	if (needsj == 0)
+		for (inoref = TAILQ_NEXT(&jaddref->ja_ref, if_deps); inoref;
+		    inoref = TAILQ_NEXT(inoref, if_deps))
+			inoref->if_nlink--;
+	jsegdep = inoref_jseg(&jaddref->ja_ref);
+	if (jaddref->ja_state & NEWBLOCK)
+		move_newblock_dep(jaddref, inodedep);
+	if (jaddref->ja_state & IOWAITING) {
+		jaddref->ja_state &= ~IOWAITING;
+		wakeup(&jaddref->ja_list);
+	}
+	jaddref->ja_mkdir = NULL;
+	if (jaddref->ja_state & IOSTARTED) {
+		jaddref->ja_state &= ~IOSTARTED;
+		WORKLIST_REMOVE(&jaddref->ja_list);
+		WORKLIST_INSERT(wkhd, &jsegdep->jd_list);
+	} else {
+		free_jsegdep(jsegdep);
+		remove_from_journal(&jaddref->ja_list);
+	}
+	/*
+	 * Leave NEWBLOCK jaddrefs on the inodedep so handle_workitem_remove
+	 * can arrange for them to be freed with the bitmap.  Otherwise we
+	 * no longer need this addref attached to the inoreflst and it
+	 * will incorrectly adjust nlink if we leave it.
+	 */
+	if ((jaddref->ja_state & NEWBLOCK) == 0) {
+		TAILQ_REMOVE(&inodedep->id_inoreflst, &jaddref->ja_ref,
+		    if_deps);
+		jaddref->ja_state |= COMPLETE;
+		free_jaddref(jaddref);
+		return (needsj);
+	}
+	jaddref->ja_state |= GOINGAWAY;
+	/*
+	 * Leave the head of the list for jsegdeps for fast merging.
+	 */
+	if (LIST_FIRST(wkhd) != NULL) {
+		jaddref->ja_state |= ONWORKLIST;
+		LIST_INSERT_AFTER(LIST_FIRST(wkhd), &jaddref->ja_list, wk_list);
+	} else
+		WORKLIST_INSERT(wkhd, &jaddref->ja_list);
+
+	return (needsj);
+}
+
+/* 
+ * Attempt to free a jaddref structure when some work completes.  This
+ * should only succeed once the entry is written and all dependencies have
+ * been notified.
+ */
+static void
+free_jaddref(jaddref)
+	struct jaddref *jaddref;
+{
+
+	if ((jaddref->ja_state & ALLCOMPLETE) != ALLCOMPLETE)
+		return;
+	if (jaddref->ja_ref.if_jsegdep)
+		panic("free_jaddref: segdep attached to jaddref %p(0x%X)\n",
+		    jaddref, jaddref->ja_state);
+	if (jaddref->ja_state & NEWBLOCK)
+		LIST_REMOVE(jaddref, ja_bmdeps);
+	if (jaddref->ja_state & (IOSTARTED | ONWORKLIST))
+		panic("free_jaddref: Bad state %p(0x%X)",
+		    jaddref, jaddref->ja_state);
+	if (jaddref->ja_mkdir != NULL)
+		panic("free_jaddref: Work pending, 0x%X\n", jaddref->ja_state);
+	WORKITEM_FREE(jaddref, D_JADDREF);
+}
+
+/*
+ * Free a jremref structure once it has been written or discarded.
+ */
+static void
+free_jremref(jremref)
+	struct jremref *jremref;
+{
+
+	if (jremref->jr_ref.if_jsegdep)
+		free_jsegdep(jremref->jr_ref.if_jsegdep);
+	if (jremref->jr_state & IOSTARTED)
+		panic("free_jremref: IO still pending");
+	WORKITEM_FREE(jremref, D_JREMREF);
+}
+
+/*
+ * Free a jnewblk structure.
+ */
+static void
+free_jnewblk(jnewblk)
+	struct jnewblk *jnewblk;
+{
+
+	if ((jnewblk->jn_state & ALLCOMPLETE) != ALLCOMPLETE)
+		return;
+	LIST_REMOVE(jnewblk, jn_deps);
+	if (jnewblk->jn_newblk != NULL)
+		panic("free_jnewblk: Dependency still attached.");
+	WORKITEM_FREE(jnewblk, D_JNEWBLK);
+}
+
+/*
+ * Cancel a jnewblk which has been superseded by a freeblk.  The jnewblk
+ * is kept linked into the bmsafemap until the free completes, thus
+ * preventing the modified state from ever reaching disk.  The free
+ * routine must pass this structure via ffs_blkfree() to
+ * softdep_setup_freeblks() so there is no race in releasing the space.
+ */
+static void
+cancel_jnewblk(jnewblk, wkhd)
+	struct jnewblk *jnewblk;
+	struct workhead *wkhd;
+{
+	struct jsegdep *jsegdep;
+
+	jsegdep = jnewblk->jn_jsegdep;
+	jnewblk->jn_jsegdep  = NULL;
+	free_jsegdep(jsegdep);
+	jnewblk->jn_newblk = NULL;
+	jnewblk->jn_state |= GOINGAWAY;
+	if (jnewblk->jn_state & IOSTARTED) {
+		jnewblk->jn_state &= ~IOSTARTED;
+		WORKLIST_REMOVE(&jnewblk->jn_list);
+	} else
+		remove_from_journal(&jnewblk->jn_list);
+	/*
+	 * Leave the head of the list for jsegdeps for fast merging.
+	 */
+	if (LIST_FIRST(wkhd) != NULL) {
+		jnewblk->jn_state |= ONWORKLIST;
+		LIST_INSERT_AFTER(LIST_FIRST(wkhd), &jnewblk->jn_list, wk_list);
+	} else
+		WORKLIST_INSERT(wkhd, &jnewblk->jn_list);
+	if (jnewblk->jn_state & IOWAITING) {
+		jnewblk->jn_state &= ~IOWAITING;
+		wakeup(&jnewblk->jn_list);
+	}
+}
+
+static void
+free_jfreeblk(jfreeblk)
+	struct jfreeblk *jfreeblk;
+{
+
+	WORKITEM_FREE(jfreeblk, D_JFREEBLK);
+}
+
+/*
+ * Release one reference to a jseg and free it if the count reaches 0.  This
+ * should eventually reclaim journal space as well.
+ */
+static void
+free_jseg(jseg)
+	struct jseg *jseg;
+{
+	struct jblocks *jblocks;
+
+	KASSERT(jseg->js_refs > 0,
+	    ("free_jseg: Invalid refcnt %d", jseg->js_refs));
+	if (--jseg->js_refs != 0)
+		return;
+	/*
+	 * Free only those jsegs which have none allocated before them to
+	 * preserve the journal space ordering.
+	 */
+	jblocks = jseg->js_jblocks;
+	while ((jseg = TAILQ_FIRST(&jblocks->jb_segs)) != NULL) {
+		jblocks->jb_oldestseq = jseg->js_seq;
+		if (jseg->js_refs != 0)
+			break;
+		TAILQ_REMOVE(&jblocks->jb_segs, jseg, js_next);
+		jblocks_free(jblocks, jseg->js_list.wk_mp, jseg->js_size);
+		KASSERT(LIST_EMPTY(&jseg->js_entries),
+		    ("free_jseg: Freed jseg has valid entries."));
+		WORKITEM_FREE(jseg, D_JSEG);
+	}
+}
+
+/*
+ * Release a jsegdep and decrement the jseg count.
+ */
+static void
+free_jsegdep(jsegdep)
+	struct jsegdep *jsegdep;
+{
+
+	if (jsegdep->jd_seg)
+		free_jseg(jsegdep->jd_seg);
+	WORKITEM_FREE(jsegdep, D_JSEGDEP);
+}
+
+/*
+ * Wait for a journal item to make it to disk.  Initiate journal processing
+ * if required.
+ */
+static void
+jwait(wk)
+	struct worklist *wk;
+{
+
+	stat_journal_wait++;
+	/*
+	 * If IO has not started we process the journal.  We can't mark the
+	 * worklist item as IOWAITING because we drop the lock while
+	 * processing the journal and the worklist entry may be freed after
+	 * this point.  The caller may call back in and re-issue the request.
+	 */
+	if ((wk->wk_state & IOSTARTED) == 0) {
+		softdep_process_journal(wk->wk_mp, MNT_WAIT);
+		return;
+	}
+	wk->wk_state |= IOWAITING;
+	msleep(wk, &lk, PRIBIO, "jwait", 0);
+}
+
+/*
+ * Lookup an inodedep based on an inode pointer and set the nlinkdelta as
+ * appropriate.  This is a convenience function to reduce duplicate code
+ * for the setup and revert functions below.
+ */
+static struct inodedep *
+inodedep_lookup_ip(ip)
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+
+	KASSERT(ip->i_nlink >= ip->i_effnlink,
+	    ("inodedep_lookup_ip: bad delta"));
+	(void) inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number,
+	    DEPALLOC, &inodedep);
+	inodedep->id_nlinkdelta = ip->i_nlink - ip->i_effnlink;
+
+	return (inodedep);
+}
+
+/*
+ * Create a journal entry that describes a truncate that we're about to
+ * perform.  The inode allocations and frees between here and the completion
+ * of the operation are done asynchronously and without journaling.  At
+ * the end of the operation the vnode is sync'd and the journal space
+ * is released.  Recovery will discover the partially completed truncate
+ * and complete it.
+ */
+void *
+softdep_setup_trunc(vp, length, flags)
+	struct vnode *vp;
+	off_t length;
+	int flags;
+{
+	struct jsegdep *jsegdep;
+	struct jtrunc *jtrunc;
+	struct ufsmount *ump;
+	struct inode *ip;
+
+	softdep_prealloc(vp, MNT_WAIT);
+	ip = VTOI(vp);
+	ump = VFSTOUFS(vp->v_mount);
+	jtrunc = malloc(sizeof(*jtrunc), M_JTRUNC, M_SOFTDEP_FLAGS);
+	workitem_alloc(&jtrunc->jt_list, D_JTRUNC, vp->v_mount);
+	jsegdep = jtrunc->jt_jsegdep = newjsegdep(&jtrunc->jt_list);
+	jtrunc->jt_ino = ip->i_number;
+	jtrunc->jt_extsize = 0;
+	jtrunc->jt_size = length;
+	if ((flags & IO_EXT) == 0 && ump->um_fstype == UFS2)
+		jtrunc->jt_extsize = ip->i_din2->di_extsize;
+	if ((flags & IO_NORMAL) == 0)
+		jtrunc->jt_size = DIP(ip, i_size);
+	ACQUIRE_LOCK(&lk);
+	add_to_journal(&jtrunc->jt_list);
+	while (jsegdep->jd_seg == NULL) {
+		stat_jwait_freeblks++;
+		jwait(&jtrunc->jt_list);
+	}
+	FREE_LOCK(&lk);
+
+	return (jsegdep);
+}
+
+/*
+ * After synchronous truncation is complete we free sync the vnode and
+ * release the jsegdep so the journal space can be freed.
+ */
+int
+softdep_complete_trunc(vp, cookie)
+	struct vnode *vp;
+	void *cookie;
+{
+	int error;
+	
+	error = ffs_syncvnode(vp, MNT_WAIT);
+	ACQUIRE_LOCK(&lk);
+	free_jsegdep((struct jsegdep *)cookie);
+	FREE_LOCK(&lk);
+
+	return (error);
+}
+
+/*
+ * Called prior to creating a new inode and linking it to a directory.  The
+ * jaddref structure must already be allocated by softdep_setup_inomapdep
+ * and it is discovered here so we can initialize the mode and update
+ * nlinkdelta.
+ */
+void
+softdep_setup_create(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	KASSERT(ip->i_nlink == 1,
+	    ("softdep_setup_create: Invalid link count."));
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(ip);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref != NULL && jaddref->ja_parent == dp->i_number,
+		    ("softdep_setup_create: No addref structure present."));
+		jaddref->ja_mode = ip->i_mode;
+	}
+	softdep_prelink(dvp, NULL);
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Create a jaddref structure to track the addition of a DOTDOT link when
+ * we are reparenting an inode as part of a rename.  This jaddref will be
+ * found by softdep_setup_directory_change.  Adjusts nlinkdelta for
+ * non-journaling softdep.
+ */
+void
+softdep_setup_dotdot_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+	struct vnode *vp;
+
+	dvp = ITOV(dp);
+	vp = ITOV(ip);
+	jaddref = NULL;
+	/*
+	 * We don't set MKDIR_PARENT as this is not tied to a mkdir and
+	 * is used as a normal link would be.
+	 */
+	if (DOINGSUJ(dvp))
+		jaddref = newjaddref(ip, dp->i_number, DOTDOT_OFFSET,
+		    dp->i_effnlink - 1, dp->i_mode);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(dp);
+	if (jaddref)
+		TAILQ_INSERT_TAIL(&inodedep->id_inoreflst, &jaddref->ja_ref,
+		    if_deps);
+	softdep_prelink(dvp, ITOV(ip));
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Create a jaddref structure to track a new link to an inode.  The directory
+ * offset is not known until softdep_setup_directory_add or
+ * softdep_setup_directory_change.  Adjusts nlinkdelta for non-journaling
+ * softdep.
+ */
+void
+softdep_setup_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	jaddref = NULL;
+	if (DOINGSUJ(dvp))
+		jaddref = newjaddref(dp, ip->i_number, 0, ip->i_effnlink - 1,
+		    ip->i_mode);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(ip);
+	if (jaddref)
+		TAILQ_INSERT_TAIL(&inodedep->id_inoreflst, &jaddref->ja_ref,
+		    if_deps);
+	softdep_prelink(dvp, ITOV(ip));
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to create the jaddref structures to track . and .. references as
+ * well as lookup and further initialize the incomplete jaddref created
+ * by softdep_setup_inomapdep when the inode was allocated.  Adjusts
+ * nlinkdelta for non-journaling softdep.
+ */
+void
+softdep_setup_mkdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *dotdotaddref;
+	struct jaddref *dotaddref;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	dotaddref = dotdotaddref = NULL;
+	if (DOINGSUJ(dvp)) {
+		dotaddref = newjaddref(ip, ip->i_number, DOT_OFFSET, 1,
+		    ip->i_mode);
+		dotaddref->ja_state |= MKDIR_BODY;
+		dotdotaddref = newjaddref(ip, dp->i_number, DOTDOT_OFFSET,
+		    dp->i_effnlink - 1, dp->i_mode);
+		dotdotaddref->ja_state |= MKDIR_PARENT;
+	}
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(ip);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref != NULL,
+		    ("softdep_setup_mkdir: No addref structure present."));
+		KASSERT(jaddref->ja_parent == dp->i_number, 
+		    ("softdep_setup_mkdir: bad parent %d",
+		    jaddref->ja_parent));
+		jaddref->ja_mode = ip->i_mode;
+		TAILQ_INSERT_BEFORE(&jaddref->ja_ref, &dotaddref->ja_ref,
+		    if_deps);
+	}
+	inodedep = inodedep_lookup_ip(dp);
+	if (DOINGSUJ(dvp))
+		TAILQ_INSERT_TAIL(&inodedep->id_inoreflst,
+		    &dotdotaddref->ja_ref, if_deps);
+	softdep_prelink(ITOV(dp), NULL);
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to track nlinkdelta of the inode and parent directories prior to
+ * unlinking a directory.
+ */
+void
+softdep_setup_rmdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	(void) inodedep_lookup_ip(ip);
+	(void) inodedep_lookup_ip(dp);
+	softdep_prelink(dvp, ITOV(ip));
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to track nlinkdelta of the inode and parent directories prior to
+ * unlink.
+ */
+void
+softdep_setup_unlink(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	(void) inodedep_lookup_ip(ip);
+	(void) inodedep_lookup_ip(dp);
+	softdep_prelink(dvp, ITOV(ip));
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to release the journal structures created by a failed non-directory
+ * creation.  Adjusts nlinkdelta for non-journaling softdep.
+ */
+void
+softdep_revert_create(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(ip);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == dp->i_number,
+		    ("softdep_revert_create: addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+	}
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to release the journal structures created by a failed dotdot link
+ * creation.  Adjusts nlinkdelta for non-journaling softdep.
+ */
+void
+softdep_revert_dotdot_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(dp);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == ip->i_number,
+		    ("softdep_revert_dotdot_link: addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+	}
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to release the journal structures created by a failed link
+ * addition.  Adjusts nlinkdelta for non-journaling softdep.
+ */
+void
+softdep_revert_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(ip);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == dp->i_number,
+		    ("softdep_revert_link: addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+	}
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Called to release the journal structures created by a failed mkdir
+ * attempt.  Adjusts nlinkdelta for non-journaling softdep.
+ */
+void
+softdep_revert_mkdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct vnode *dvp;
+
+	dvp = ITOV(dp);
+
+	ACQUIRE_LOCK(&lk);
+	inodedep = inodedep_lookup_ip(dp);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == ip->i_number,
+		    ("softdep_revert_mkdir: dotdot addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+	}
+	inodedep = inodedep_lookup_ip(ip);
+	if (DOINGSUJ(dvp)) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == dp->i_number,
+		    ("softdep_revert_mkdir: addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref->ja_parent == ip->i_number,
+		    ("softdep_revert_mkdir: dot addref parent mismatch"));
+		cancel_jaddref(jaddref, inodedep, &inodedep->id_inowait);
+	}
+	FREE_LOCK(&lk);
+}
+
+/* 
+ * Called to correct nlinkdelta after a failed rmdir.
+ */
+void
+softdep_revert_rmdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	ACQUIRE_LOCK(&lk);
+	(void) inodedep_lookup_ip(ip);
+	(void) inodedep_lookup_ip(dp);
+	FREE_LOCK(&lk);
+}
+
 /*
  * Protecting the freemaps (or bitmaps).
  * 
@@ -1536,6 +3970,22 @@ softdep_setup_inomapdep(bp, ip, newinum)
 {
 	struct inodedep *inodedep;
 	struct bmsafemap *bmsafemap;
+	struct jaddref *jaddref;
+	struct mount *mp;
+	struct fs *fs;
+
+	mp = UFSTOVFS(ip->i_ump);
+	fs = ip->i_ump->um_fs;
+	jaddref = NULL;
+
+	/*
+	 * Allocate the journal reference add structure so that the bitmap
+	 * can be dependent on it.
+	 */
+	if (mp->mnt_kern_flag & MNTK_SUJ) {
+		jaddref = newjaddref(ip, newinum, 0, 0, 0);
+		jaddref->ja_state |= NEWBLOCK;
+	}
 
 	/*
 	 * Create a dependency for the newly allocated inode.
@@ -1544,14 +3994,20 @@ softdep_setup_inomapdep(bp, ip, newinum)
 	 * the cylinder group map from which it was allocated.
 	 */
 	ACQUIRE_LOCK(&lk);
-	if ((inodedep_lookup(UFSTOVFS(ip->i_ump), newinum, DEPALLOC|NODELAY,
-	    &inodedep)))
-		panic("softdep_setup_inomapdep: dependency for new inode "
-		    "already exists");
-	inodedep->id_buf = bp;
+	if ((inodedep_lookup(mp, newinum, DEPALLOC|NODELAY, &inodedep)))
+		panic("softdep_setup_inomapdep: dependency %p for new"
+		    "inode already exists", inodedep);
+	bmsafemap = bmsafemap_lookup(mp, bp, ino_to_cg(fs, newinum));
+	if (jaddref) {
+		LIST_INSERT_HEAD(&bmsafemap->sm_jaddrefhd, jaddref, ja_bmdeps);
+		TAILQ_INSERT_TAIL(&inodedep->id_inoreflst, &jaddref->ja_ref,
+		    if_deps);
+	} else {
+		inodedep->id_state |= ONDEPLIST;
+		LIST_INSERT_HEAD(&bmsafemap->sm_inodedephd, inodedep, id_deps);
+	}
+	inodedep->id_bmsafemap = bmsafemap;
 	inodedep->id_state &= ~DEPCOMPLETE;
-	bmsafemap = bmsafemap_lookup(inodedep->id_list.wk_mp, bp);
-	LIST_INSERT_HEAD(&bmsafemap->sm_inodedephd, inodedep, id_deps);
 	FREE_LOCK(&lk);
 }
 
@@ -1560,29 +4016,98 @@ softdep_setup_inomapdep(bp, ip, newinum)
  * allocate block or fragment.
  */
 void
-softdep_setup_blkmapdep(bp, mp, newblkno)
+softdep_setup_blkmapdep(bp, mp, newblkno, frags, oldfrags)
 	struct buf *bp;		/* buffer for cylgroup block with block map */
 	struct mount *mp;	/* filesystem doing allocation */
 	ufs2_daddr_t newblkno;	/* number of newly allocated block */
+	int frags;		/* Number of fragments. */
+	int oldfrags;		/* Previous number of fragments for extend. */
 {
 	struct newblk *newblk;
 	struct bmsafemap *bmsafemap;
+	struct jnewblk *jnewblk;
 	struct fs *fs;
 
 	fs = VFSTOUFS(mp)->um_fs;
+	jnewblk = NULL;
 	/*
 	 * Create a dependency for the newly allocated block.
 	 * Add it to the dependency list for the buffer holding
 	 * the cylinder group map from which it was allocated.
 	 */
+	if (mp->mnt_kern_flag & MNTK_SUJ) {
+		jnewblk = malloc(sizeof(*jnewblk), M_JNEWBLK, M_SOFTDEP_FLAGS);
+		workitem_alloc(&jnewblk->jn_list, D_JNEWBLK, mp);
+		jnewblk->jn_jsegdep = newjsegdep(&jnewblk->jn_list);
+		jnewblk->jn_state = ATTACHED;
+		jnewblk->jn_blkno = newblkno;
+		jnewblk->jn_frags = frags;
+		jnewblk->jn_oldfrags = oldfrags;
+#ifdef SUJ_DEBUG
+		{
+			struct cg *cgp;
+			uint8_t *blksfree;
+			long bno;
+			int i;
+	
+			cgp = (struct cg *)bp->b_data;
+			blksfree = cg_blksfree(cgp);
+			bno = dtogd(fs, jnewblk->jn_blkno);
+			for (i = jnewblk->jn_oldfrags; i < jnewblk->jn_frags;
+			    i++) {
+				if (isset(blksfree, bno + i))
+					panic("softdep_setup_blkmapdep: "
+					    "free fragment %d from %d-%d "
+					    "state 0x%X dep %p", i,
+					    jnewblk->jn_oldfrags,
+					    jnewblk->jn_frags,
+					    jnewblk->jn_state,
+					    jnewblk->jn_newblk);
+			}
+		}
+#endif
+	}
 	ACQUIRE_LOCK(&lk);
-	if (newblk_lookup(fs, newblkno, DEPALLOC, &newblk) != 0)
+	if (newblk_lookup(mp, newblkno, DEPALLOC, &newblk) != 0)
 		panic("softdep_setup_blkmapdep: found block");
-	newblk->nb_bmsafemap = bmsafemap = bmsafemap_lookup(mp, bp);
-	LIST_INSERT_HEAD(&bmsafemap->sm_newblkhd, newblk, nb_deps);
+	newblk->nb_bmsafemap = bmsafemap = bmsafemap_lookup(mp, bp,
+	    dtog(fs, newblkno));
+	if (jnewblk) {
+		jnewblk->jn_newblk = newblk;
+		LIST_INSERT_HEAD(&bmsafemap->sm_jnewblkhd, jnewblk, jn_deps);
+	} else {
+		newblk->nb_state |= ONDEPLIST;
+		LIST_INSERT_HEAD(&bmsafemap->sm_newblkhd, newblk, nb_deps);
+	}
+	newblk->nb_bmsafemap = bmsafemap;
+	newblk->nb_jnewblk = jnewblk;
 	FREE_LOCK(&lk);
 }
 
+#define	BMSAFEMAP_HASH(fs, cg) \
+      (&bmsafemap_hashtbl[((((register_t)(fs)) >> 13) + (cg)) & bmsafemap_hash])
+
+static int
+bmsafemap_find(bmsafemaphd, mp, cg, bmsafemapp)
+	struct bmsafemap_hashhead *bmsafemaphd;
+	struct mount *mp;
+	int cg;
+	struct bmsafemap **bmsafemapp;
+{
+	struct bmsafemap *bmsafemap;
+
+	LIST_FOREACH(bmsafemap, bmsafemaphd, sm_hash)
+		if (bmsafemap->sm_list.wk_mp == mp && bmsafemap->sm_cg == cg)
+			break;
+	if (bmsafemap) {
+		*bmsafemapp = bmsafemap;
+		return (1);
+	}
+	*bmsafemapp = NULL;
+
+	return (0);
+}
+
 /*
  * Find the bmsafemap associated with a cylinder group buffer.
  * If none exists, create one. The buffer must be locked when
@@ -1590,27 +4115,43 @@ softdep_setup_blkmapdep(bp, mp, newblkno)
  * splbio interrupts blocked.
  */
 static struct bmsafemap *
-bmsafemap_lookup(mp, bp)
+bmsafemap_lookup(mp, bp, cg)
 	struct mount *mp;
 	struct buf *bp;
+	int cg;
 {
-	struct bmsafemap *bmsafemap;
+	struct bmsafemap_hashhead *bmsafemaphd;
+	struct bmsafemap *bmsafemap, *collision;
 	struct worklist *wk;
+	struct fs *fs;
 
 	mtx_assert(&lk, MA_OWNED);
-	LIST_FOREACH(wk, &bp->b_dep, wk_list)
-		if (wk->wk_type == D_BMSAFEMAP)
-			return (WK_BMSAFEMAP(wk));
+	if (bp)
+		LIST_FOREACH(wk, &bp->b_dep, wk_list)
+			if (wk->wk_type == D_BMSAFEMAP)
+				return (WK_BMSAFEMAP(wk));
+	fs = VFSTOUFS(mp)->um_fs;
+	bmsafemaphd = BMSAFEMAP_HASH(fs, cg);
+	if (bmsafemap_find(bmsafemaphd, mp, cg, &bmsafemap) == 1)
+		return (bmsafemap);
 	FREE_LOCK(&lk);
 	bmsafemap = malloc(sizeof(struct bmsafemap),
 		M_BMSAFEMAP, M_SOFTDEP_FLAGS);
 	workitem_alloc(&bmsafemap->sm_list, D_BMSAFEMAP, mp);
 	bmsafemap->sm_buf = bp;
-	LIST_INIT(&bmsafemap->sm_allocdirecthd);
-	LIST_INIT(&bmsafemap->sm_allocindirhd);
 	LIST_INIT(&bmsafemap->sm_inodedephd);
+	LIST_INIT(&bmsafemap->sm_inodedepwr);
 	LIST_INIT(&bmsafemap->sm_newblkhd);
+	LIST_INIT(&bmsafemap->sm_newblkwr);
+	LIST_INIT(&bmsafemap->sm_jaddrefhd);
+	LIST_INIT(&bmsafemap->sm_jnewblkhd);
 	ACQUIRE_LOCK(&lk);
+	if (bmsafemap_find(bmsafemaphd, mp, cg, &collision) == 1) {
+		WORKITEM_FREE(bmsafemap, D_BMSAFEMAP);
+		return (collision);
+	}
+	bmsafemap->sm_cg = cg;
+	LIST_INSERT_HEAD(bmsafemaphd, bmsafemap, sm_hash);
 	WORKLIST_INSERT(&bp->b_dep, &bmsafemap->sm_list);
 	return (bmsafemap);
 }
@@ -1645,9 +4186,9 @@ bmsafemap_lookup(mp, bp)
  * unreferenced fragments.
  */ 
 void 
-softdep_setup_allocdirect(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
+softdep_setup_allocdirect(ip, off, newblkno, oldblkno, newsize, oldsize, bp)
 	struct inode *ip;	/* inode to which block is being added */
-	ufs_lbn_t lbn;		/* block pointer within inode */
+	ufs_lbn_t off;		/* block pointer within inode */
 	ufs2_daddr_t newblkno;	/* disk block number being added */
 	ufs2_daddr_t oldblkno;	/* previous block number, 0 unless frag */
 	long newsize;		/* size of new block */
@@ -1656,34 +4197,33 @@ softdep_setup_allocdirect(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
 {
 	struct allocdirect *adp, *oldadp;
 	struct allocdirectlst *adphead;
-	struct bmsafemap *bmsafemap;
+	struct freefrag *freefrag;
 	struct inodedep *inodedep;
 	struct pagedep *pagedep;
+	struct jnewblk *jnewblk;
 	struct newblk *newblk;
 	struct mount *mp;
+	ufs_lbn_t lbn;
 
+	lbn = bp->b_lblkno;
 	mp = UFSTOVFS(ip->i_ump);
-	adp = malloc(sizeof(struct allocdirect),
-		M_ALLOCDIRECT, M_SOFTDEP_FLAGS|M_ZERO);
-	workitem_alloc(&adp->ad_list, D_ALLOCDIRECT, mp);
-	adp->ad_lbn = lbn;
-	adp->ad_newblkno = newblkno;
-	adp->ad_oldblkno = oldblkno;
-	adp->ad_newsize = newsize;
-	adp->ad_oldsize = oldsize;
-	adp->ad_state = ATTACHED;
-	LIST_INIT(&adp->ad_newdirblk);
-	if (newblkno == oldblkno)
-		adp->ad_freefrag = NULL;
+	if (oldblkno && oldblkno != newblkno)
+		freefrag = newfreefrag(ip, oldblkno, oldsize, lbn);
 	else
-		adp->ad_freefrag = newfreefrag(ip, oldblkno, oldsize);
+		freefrag = NULL;
 
 	ACQUIRE_LOCK(&lk);
-	if (lbn >= NDADDR) {
+	if (off >= NDADDR) {
+		if (lbn > 0)
+			panic("softdep_setup_allocdirect: bad lbn %jd, off %jd",
+			    lbn, off);
 		/* allocating an indirect block */
 		if (oldblkno != 0)
 			panic("softdep_setup_allocdirect: non-zero indir");
 	} else {
+		if (off != lbn)
+			panic("softdep_setup_allocdirect: lbn %jd != off %jd",
+			    lbn, off);
 		/*
 		 * Allocating a direct block.
 		 *
@@ -1692,26 +4232,39 @@ softdep_setup_allocdirect(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
 		 * deletions.
 		 */
 		if ((ip->i_mode & IFMT) == IFDIR &&
-		    pagedep_lookup(ip, lbn, DEPALLOC, &pagedep) == 0)
+		    pagedep_lookup(mp, ip->i_number, off, DEPALLOC,
+		    &pagedep) == 0)
 			WORKLIST_INSERT(&bp->b_dep, &pagedep->pd_list);
 	}
-	if (newblk_lookup(ip->i_fs, newblkno, 0, &newblk) == 0)
+	if (newblk_lookup(mp, newblkno, 0, &newblk) == 0)
 		panic("softdep_setup_allocdirect: lost block");
-	if (newblk->nb_state == DEPCOMPLETE) {
-		adp->ad_state |= DEPCOMPLETE;
-		adp->ad_buf = NULL;
-	} else {
-		bmsafemap = newblk->nb_bmsafemap;
-		adp->ad_buf = bmsafemap->sm_buf;
-		LIST_REMOVE(newblk, nb_deps);
-		LIST_INSERT_HEAD(&bmsafemap->sm_allocdirecthd, adp, ad_deps);
-	}
-	LIST_REMOVE(newblk, nb_hash);
-	free(newblk, M_NEWBLK);
+	KASSERT(newblk->nb_list.wk_type == D_NEWBLK,
+	    ("softdep_setup_allocdirect: newblk already initialized"));
+	/*
+	 * Convert the newblk to an allocdirect.
+	 */
+	newblk->nb_list.wk_type = D_ALLOCDIRECT;
+	adp = (struct allocdirect *)newblk;
+	newblk->nb_freefrag = freefrag;
+	adp->ad_offset = off;
+	adp->ad_oldblkno = oldblkno;
+	adp->ad_newsize = newsize;
+	adp->ad_oldsize = oldsize;
 
+	/*
+	 * Finish initializing the journal.
+	 */
+	if ((jnewblk = newblk->nb_jnewblk) != NULL) {
+		jnewblk->jn_ino = ip->i_number;
+		jnewblk->jn_lbn = lbn;
+		add_to_journal(&jnewblk->jn_list);
+	}
+	if (freefrag && freefrag->ff_jfreefrag != NULL)
+		add_to_journal(&freefrag->ff_jfreefrag->fr_list);
 	inodedep_lookup(mp, ip->i_number, DEPALLOC | NODELAY, &inodedep);
 	adp->ad_inodedep = inodedep;
-	WORKLIST_INSERT(&bp->b_dep, &adp->ad_list);
+
+	WORKLIST_INSERT(&bp->b_dep, &newblk->nb_list);
 	/*
 	 * The list of allocdirects must be kept in sorted and ascending
 	 * order so that the rollback routines can quickly determine the
@@ -1726,24 +4279,25 @@ softdep_setup_allocdirect(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
 	 */
 	adphead = &inodedep->id_newinoupdt;
 	oldadp = TAILQ_LAST(adphead, allocdirectlst);
-	if (oldadp == NULL || oldadp->ad_lbn <= lbn) {
+	if (oldadp == NULL || oldadp->ad_offset <= off) {
 		/* insert at end of list */
 		TAILQ_INSERT_TAIL(adphead, adp, ad_next);
-		if (oldadp != NULL && oldadp->ad_lbn == lbn)
+		if (oldadp != NULL && oldadp->ad_offset == off)
 			allocdirect_merge(adphead, adp, oldadp);
 		FREE_LOCK(&lk);
 		return;
 	}
 	TAILQ_FOREACH(oldadp, adphead, ad_next) {
-		if (oldadp->ad_lbn >= lbn)
+		if (oldadp->ad_offset >= off)
 			break;
 	}
 	if (oldadp == NULL)
 		panic("softdep_setup_allocdirect: lost entry");
 	/* insert in middle of list */
 	TAILQ_INSERT_BEFORE(oldadp, adp, ad_next);
-	if (oldadp->ad_lbn == lbn)
+	if (oldadp->ad_offset == off)
 		allocdirect_merge(adphead, adp, oldadp);
+
 	FREE_LOCK(&lk);
 }
 
@@ -1761,10 +4315,11 @@ allocdirect_merge(adphead, newadp, oldadp)
 	struct freefrag *freefrag;
 	struct newdirblk *newdirblk;
 
+	freefrag = NULL;
 	mtx_assert(&lk, MA_OWNED);
 	if (newadp->ad_oldblkno != oldadp->ad_newblkno ||
 	    newadp->ad_oldsize != oldadp->ad_newsize ||
-	    newadp->ad_lbn >= NDADDR)
+	    newadp->ad_offset >= NDADDR)
 		panic("%s %jd != new %jd || old size %ld != new %ld",
 		    "allocdirect_merge: old blkno",
 		    (intmax_t)newadp->ad_oldblkno,
@@ -1779,7 +4334,7 @@ allocdirect_merge(adphead, newadp, oldadp)
 	 * This action is done by swapping the freefrag dependencies.
 	 * The new dependency gains the old one's freefrag, and the
 	 * old one gets the new one and then immediately puts it on
-	 * the worklist when it is freed by free_allocdirect. It is
+	 * the worklist when it is freed by free_newblk. It is
 	 * not possible to do this swap when the old dependency had a
 	 * non-zero size but no previous fragment to free. This condition
 	 * arises when the new block is an extension of the old block.
@@ -1788,8 +4343,8 @@ allocdirect_merge(adphead, newadp, oldadp)
 	 * the old dependency, so cannot legitimately be freed until the
 	 * conditions for the new dependency are fulfilled.
 	 */
+	freefrag = newadp->ad_freefrag;
 	if (oldadp->ad_freefrag != NULL || oldadp->ad_oldblkno == 0) {
-		freefrag = newadp->ad_freefrag;
 		newadp->ad_freefrag = oldadp->ad_freefrag;
 		oldadp->ad_freefrag = freefrag;
 	}
@@ -1804,32 +4359,118 @@ allocdirect_merge(adphead, newadp, oldadp)
 			panic("allocdirect_merge: extra newdirblk");
 		WORKLIST_INSERT(&newadp->ad_newdirblk, &newdirblk->db_list);
 	}
-	free_allocdirect(adphead, oldadp, 0);
+	TAILQ_REMOVE(adphead, oldadp, ad_next);
+	/*
+	 * We need to move any journal dependencies over to the freefrag
+	 * that releases this block if it exists.  Otherwise we are
+	 * extending an existing block and we'll wait until that is
+	 * complete to release the journal space and extend the
+	 * new journal to cover this old space as well.
+	 */
+	if (freefrag == NULL) {
+		struct jnewblk *jnewblk;
+		struct jnewblk *njnewblk;
+
+		if (oldadp->ad_newblkno != newadp->ad_newblkno)
+			panic("allocdirect_merge: %jd != %jd",
+			    oldadp->ad_newblkno, newadp->ad_newblkno);
+		jnewblk = oldadp->ad_block.nb_jnewblk;
+		cancel_newblk(&oldadp->ad_block, &newadp->ad_block.nb_jwork);
+		/*
+		 * We have an unwritten jnewblk, we need to merge the
+		 * frag bits with our own.  The newer adp's journal can not
+		 * be written prior to the old one so no need to check for
+		 * it here.
+		 */
+		if (jnewblk) {
+			njnewblk = newadp->ad_block.nb_jnewblk;
+			if (njnewblk == NULL)
+				panic("allocdirect_merge: No jnewblk");
+			if (jnewblk->jn_state & UNDONE) {
+				njnewblk->jn_state |= UNDONE | NEWBLOCK;
+				njnewblk->jn_state &= ~ATTACHED;
+				jnewblk->jn_state &= ~UNDONE;
+			}
+			njnewblk->jn_oldfrags = jnewblk->jn_oldfrags;
+			WORKLIST_REMOVE(&jnewblk->jn_list);
+			jnewblk->jn_state |= ATTACHED | COMPLETE;
+			free_jnewblk(jnewblk);
+		}
+	} else {
+		/*
+		 * We can skip journaling for this freefrag and just complete
+		 * any pending journal work for the allocdirect that is being
+		 * removed after the freefrag completes.
+		 */
+		if (freefrag->ff_jfreefrag)
+			cancel_jfreefrag(freefrag->ff_jfreefrag);
+		cancel_newblk(&oldadp->ad_block, &freefrag->ff_jwork);
+	}
+	free_newblk(&oldadp->ad_block);
 }
-		
+
 /*
- * Allocate a new freefrag structure if needed.
+ * Allocate a jfreefrag structure to journal a single block free.
  */
-static struct freefrag *
-newfreefrag(ip, blkno, size)
+static struct jfreefrag *
+newjfreefrag(freefrag, ip, blkno, size, lbn)
+	struct freefrag *freefrag;
 	struct inode *ip;
 	ufs2_daddr_t blkno;
 	long size;
+	ufs_lbn_t lbn;
+{
+	struct jfreefrag *jfreefrag;
+	struct fs *fs;
+
+	fs = ip->i_fs;
+	jfreefrag = malloc(sizeof(struct jfreefrag), M_JFREEFRAG,
+	    M_SOFTDEP_FLAGS);
+	workitem_alloc(&jfreefrag->fr_list, D_JFREEFRAG, UFSTOVFS(ip->i_ump));
+	jfreefrag->fr_jsegdep = newjsegdep(&jfreefrag->fr_list);
+	jfreefrag->fr_state = ATTACHED | DEPCOMPLETE;
+	jfreefrag->fr_ino = ip->i_number;
+	jfreefrag->fr_lbn = lbn;
+	jfreefrag->fr_blkno = blkno;
+	jfreefrag->fr_frags = numfrags(fs, size);
+	jfreefrag->fr_freefrag = freefrag;
+
+	return (jfreefrag);
+}
+
+/*
+ * Allocate a new freefrag structure.
+ */
+static struct freefrag *
+newfreefrag(ip, blkno, size, lbn)
+	struct inode *ip;
+	ufs2_daddr_t blkno;
+	long size;
+	ufs_lbn_t lbn;
 {
 	struct freefrag *freefrag;
 	struct fs *fs;
 
-	if (blkno == 0)
-		return (NULL);
 	fs = ip->i_fs;
 	if (fragnum(fs, blkno) + numfrags(fs, size) > fs->fs_frag)
 		panic("newfreefrag: frag size");
 	freefrag = malloc(sizeof(struct freefrag),
-		M_FREEFRAG, M_SOFTDEP_FLAGS);
+	    M_FREEFRAG, M_SOFTDEP_FLAGS);
 	workitem_alloc(&freefrag->ff_list, D_FREEFRAG, UFSTOVFS(ip->i_ump));
+	freefrag->ff_state = ATTACHED;
+	LIST_INIT(&freefrag->ff_jwork);
 	freefrag->ff_inum = ip->i_number;
 	freefrag->ff_blkno = blkno;
 	freefrag->ff_fragsize = size;
+
+	if (fs->fs_flags & FS_SUJ) {
+		freefrag->ff_jfreefrag =
+		    newjfreefrag(freefrag, ip, blkno, size, lbn);
+	} else {
+		freefrag->ff_state |= DEPCOMPLETE;
+		freefrag->ff_jfreefrag = NULL;
+	}
+
 	return (freefrag);
 }
 
@@ -1842,9 +4483,17 @@ handle_workitem_freefrag(freefrag)
 	struct freefrag *freefrag;
 {
 	struct ufsmount *ump = VFSTOUFS(freefrag->ff_list.wk_mp);
+	struct workhead wkhd;
 
+	/*
+	 * It would be illegal to add new completion items to the
+	 * freefrag after it was schedule to be done so it must be
+	 * safe to modify the list head here.
+	 */
+	LIST_INIT(&wkhd);
+	LIST_SWAP(&freefrag->ff_jwork, &wkhd, worklist, wk_list);
 	ffs_blkfree(ump, ump->um_fs, ump->um_devvp, freefrag->ff_blkno,
-	    freefrag->ff_fragsize, freefrag->ff_inum);
+	    freefrag->ff_fragsize, freefrag->ff_inum, &wkhd);
 	ACQUIRE_LOCK(&lk);
 	WORKITEM_FREE(freefrag, D_FREEFRAG);
 	FREE_LOCK(&lk);
@@ -1856,9 +4505,9 @@ handle_workitem_freefrag(freefrag)
  * See the description of softdep_setup_allocdirect above for details.
  */
 void 
-softdep_setup_allocext(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
+softdep_setup_allocext(ip, off, newblkno, oldblkno, newsize, oldsize, bp)
 	struct inode *ip;
-	ufs_lbn_t lbn;
+	ufs_lbn_t off;
 	ufs2_daddr_t newblkno;
 	ufs2_daddr_t oldblkno;
 	long newsize;
@@ -1867,50 +4516,55 @@ softdep_setup_allocext(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
 {
 	struct allocdirect *adp, *oldadp;
 	struct allocdirectlst *adphead;
-	struct bmsafemap *bmsafemap;
+	struct freefrag *freefrag;
 	struct inodedep *inodedep;
+	struct jnewblk *jnewblk;
 	struct newblk *newblk;
 	struct mount *mp;
+	ufs_lbn_t lbn;
 
+	if (off >= NXADDR)
+		panic("softdep_setup_allocext: lbn %lld > NXADDR",
+		    (long long)off);
+
+	lbn = bp->b_lblkno;
 	mp = UFSTOVFS(ip->i_ump);
-	adp = malloc(sizeof(struct allocdirect),
-		M_ALLOCDIRECT, M_SOFTDEP_FLAGS|M_ZERO);
-	workitem_alloc(&adp->ad_list, D_ALLOCDIRECT, mp);
-	adp->ad_lbn = lbn;
-	adp->ad_newblkno = newblkno;
+	if (oldblkno && oldblkno != newblkno)
+		freefrag = newfreefrag(ip, oldblkno, oldsize, lbn);
+	else
+		freefrag = NULL;
+
+	ACQUIRE_LOCK(&lk);
+	if (newblk_lookup(mp, newblkno, 0, &newblk) == 0)
+		panic("softdep_setup_allocext: lost block");
+	KASSERT(newblk->nb_list.wk_type == D_NEWBLK,
+	    ("softdep_setup_allocext: newblk already initialized"));
+	/*
+	 * Convert the newblk to an allocdirect.
+	 */
+	newblk->nb_list.wk_type = D_ALLOCDIRECT;
+	adp = (struct allocdirect *)newblk;
+	newblk->nb_freefrag = freefrag;
+	adp->ad_offset = off;
 	adp->ad_oldblkno = oldblkno;
 	adp->ad_newsize = newsize;
 	adp->ad_oldsize = oldsize;
-	adp->ad_state = ATTACHED | EXTDATA;
-	LIST_INIT(&adp->ad_newdirblk);
-	if (newblkno == oldblkno)
-		adp->ad_freefrag = NULL;
-	else
-		adp->ad_freefrag = newfreefrag(ip, oldblkno, oldsize);
-
-	ACQUIRE_LOCK(&lk);
-	if (newblk_lookup(ip->i_fs, newblkno, 0, &newblk) == 0)
-		panic("softdep_setup_allocext: lost block");
+	adp->ad_state |=  EXTDATA;
 
+	/*
+	 * Finish initializing the journal.
+	 */
+	if ((jnewblk = newblk->nb_jnewblk) != NULL) {
+		jnewblk->jn_ino = ip->i_number;
+		jnewblk->jn_lbn = lbn;
+		add_to_journal(&jnewblk->jn_list);
+	}
+	if (freefrag && freefrag->ff_jfreefrag != NULL)
+		add_to_journal(&freefrag->ff_jfreefrag->fr_list);
 	inodedep_lookup(mp, ip->i_number, DEPALLOC | NODELAY, &inodedep);
 	adp->ad_inodedep = inodedep;
 
-	if (newblk->nb_state == DEPCOMPLETE) {
-		adp->ad_state |= DEPCOMPLETE;
-		adp->ad_buf = NULL;
-	} else {
-		bmsafemap = newblk->nb_bmsafemap;
-		adp->ad_buf = bmsafemap->sm_buf;
-		LIST_REMOVE(newblk, nb_deps);
-		LIST_INSERT_HEAD(&bmsafemap->sm_allocdirecthd, adp, ad_deps);
-	}
-	LIST_REMOVE(newblk, nb_hash);
-	free(newblk, M_NEWBLK);
-
-	WORKLIST_INSERT(&bp->b_dep, &adp->ad_list);
-	if (lbn >= NXADDR)
-		panic("softdep_setup_allocext: lbn %lld > NXADDR",
-		    (long long)lbn);
+	WORKLIST_INSERT(&bp->b_dep, &newblk->nb_list);
 	/*
 	 * The list of allocdirects must be kept in sorted and ascending
 	 * order so that the rollback routines can quickly determine the
@@ -1925,23 +4579,23 @@ softdep_setup_allocext(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
 	 */
 	adphead = &inodedep->id_newextupdt;
 	oldadp = TAILQ_LAST(adphead, allocdirectlst);
-	if (oldadp == NULL || oldadp->ad_lbn <= lbn) {
+	if (oldadp == NULL || oldadp->ad_offset <= off) {
 		/* insert at end of list */
 		TAILQ_INSERT_TAIL(adphead, adp, ad_next);
-		if (oldadp != NULL && oldadp->ad_lbn == lbn)
+		if (oldadp != NULL && oldadp->ad_offset == off)
 			allocdirect_merge(adphead, adp, oldadp);
 		FREE_LOCK(&lk);
 		return;
 	}
 	TAILQ_FOREACH(oldadp, adphead, ad_next) {
-		if (oldadp->ad_lbn >= lbn)
+		if (oldadp->ad_offset >= off)
 			break;
 	}
 	if (oldadp == NULL)
 		panic("softdep_setup_allocext: lost entry");
 	/* insert in middle of list */
 	TAILQ_INSERT_BEFORE(oldadp, adp, ad_next);
-	if (oldadp->ad_lbn == lbn)
+	if (oldadp->ad_offset == off)
 		allocdirect_merge(adphead, adp, oldadp);
 	FREE_LOCK(&lk);
 }
@@ -1975,22 +4629,39 @@ softdep_setup_allocext(ip, lbn, newblkno, oldblkno, newsize, oldsize, bp)
  * Allocate a new allocindir structure.
  */
 static struct allocindir *
-newallocindir(ip, ptrno, newblkno, oldblkno)
+newallocindir(ip, ptrno, newblkno, oldblkno, lbn)
 	struct inode *ip;	/* inode for file being extended */
 	int ptrno;		/* offset of pointer in indirect block */
 	ufs2_daddr_t newblkno;	/* disk block number being added */
 	ufs2_daddr_t oldblkno;	/* previous block number, 0 if none */
+	ufs_lbn_t lbn;
 {
+	struct newblk *newblk;
 	struct allocindir *aip;
+	struct freefrag *freefrag;
+	struct jnewblk *jnewblk;
 
-	aip = malloc(sizeof(struct allocindir),
-		M_ALLOCINDIR, M_SOFTDEP_FLAGS|M_ZERO);
-	workitem_alloc(&aip->ai_list, D_ALLOCINDIR, UFSTOVFS(ip->i_ump));
-	aip->ai_state = ATTACHED;
+	if (oldblkno)
+		freefrag = newfreefrag(ip, oldblkno, ip->i_fs->fs_bsize, lbn);
+	else
+		freefrag = NULL;
+	ACQUIRE_LOCK(&lk);
+	if (newblk_lookup(UFSTOVFS(ip->i_ump), newblkno, 0, &newblk) == 0)
+		panic("new_allocindir: lost block");
+	KASSERT(newblk->nb_list.wk_type == D_NEWBLK,
+	    ("newallocindir: newblk already initialized"));
+	newblk->nb_list.wk_type = D_ALLOCINDIR;
+	newblk->nb_freefrag = freefrag;
+	aip = (struct allocindir *)newblk;
 	aip->ai_offset = ptrno;
-	aip->ai_newblkno = newblkno;
 	aip->ai_oldblkno = oldblkno;
-	aip->ai_freefrag = newfreefrag(ip, oldblkno, ip->i_fs->fs_bsize);
+	if ((jnewblk = newblk->nb_jnewblk) != NULL) {
+		jnewblk->jn_ino = ip->i_number;
+		jnewblk->jn_lbn = lbn;
+		add_to_journal(&jnewblk->jn_list);
+	}
+	if (freefrag && freefrag->ff_jfreefrag != NULL)
+		add_to_journal(&freefrag->ff_jfreefrag->fr_list);
 	return (aip);
 }
 
@@ -2008,22 +4679,28 @@ softdep_setup_allocindir_page(ip, lbn, bp, ptrno, newblkno, oldblkno, nbp)
 	ufs2_daddr_t oldblkno;	/* previous block number, 0 if none */
 	struct buf *nbp;	/* buffer holding allocated page */
 {
+	struct inodedep *inodedep;
 	struct allocindir *aip;
 	struct pagedep *pagedep;
+	struct mount *mp;
 
+	if (lbn != nbp->b_lblkno)
+		panic("softdep_setup_allocindir_page: lbn %jd != lblkno %jd",
+		    lbn, bp->b_lblkno);
 	ASSERT_VOP_LOCKED(ITOV(ip), "softdep_setup_allocindir_page");
-	aip = newallocindir(ip, ptrno, newblkno, oldblkno);
-	ACQUIRE_LOCK(&lk);
+	mp = UFSTOVFS(ip->i_ump);
+	aip = newallocindir(ip, ptrno, newblkno, oldblkno, lbn);
+	(void) inodedep_lookup(mp, ip->i_number, DEPALLOC, &inodedep);
 	/*
 	 * If we are allocating a directory page, then we must
 	 * allocate an associated pagedep to track additions and
 	 * deletions.
 	 */
 	if ((ip->i_mode & IFMT) == IFDIR &&
-	    pagedep_lookup(ip, lbn, DEPALLOC, &pagedep) == 0)
+	    pagedep_lookup(mp, ip->i_number, lbn, DEPALLOC, &pagedep) == 0)
 		WORKLIST_INSERT(&nbp->b_dep, &pagedep->pd_list);
-	WORKLIST_INSERT(&nbp->b_dep, &aip->ai_list);
-	setup_allocindir_phase2(bp, ip, aip);
+	WORKLIST_INSERT(&nbp->b_dep, &aip->ai_block.nb_list);
+	setup_allocindir_phase2(bp, ip, inodedep, aip, lbn);
 	FREE_LOCK(&lk);
 }
 
@@ -2038,15 +4715,40 @@ softdep_setup_allocindir_meta(nbp, ip, bp, ptrno, newblkno)
 	struct buf *bp;		/* indirect block referencing allocated block */
 	int ptrno;		/* offset of pointer in indirect block */
 	ufs2_daddr_t newblkno;	/* disk block number being added */
+{
+	struct inodedep *inodedep;
+	struct allocindir *aip;
+	ufs_lbn_t lbn;
+
+	lbn = nbp->b_lblkno;
+	ASSERT_VOP_LOCKED(ITOV(ip), "softdep_setup_allocindir_meta");
+	aip = newallocindir(ip, ptrno, newblkno, 0, lbn);
+	inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, DEPALLOC, &inodedep);
+	WORKLIST_INSERT(&nbp->b_dep, &aip->ai_block.nb_list);
+	setup_allocindir_phase2(bp, ip, inodedep, aip, lbn);
+	FREE_LOCK(&lk);
+}
+
+static void
+indirdep_complete(indirdep)
+	struct indirdep *indirdep;
 {
 	struct allocindir *aip;
 
-	ASSERT_VOP_LOCKED(ITOV(ip), "softdep_setup_allocindir_meta");
-	aip = newallocindir(ip, ptrno, newblkno, 0);
-	ACQUIRE_LOCK(&lk);
-	WORKLIST_INSERT(&nbp->b_dep, &aip->ai_list);
-	setup_allocindir_phase2(bp, ip, aip);
-	FREE_LOCK(&lk);
+	LIST_REMOVE(indirdep, ir_next);
+	indirdep->ir_state &= ~ONDEPLIST;
+
+	while ((aip = LIST_FIRST(&indirdep->ir_completehd)) != NULL) {
+		LIST_REMOVE(aip, ai_next);
+		free_newblk(&aip->ai_block);
+	}
+	/*
+	 * If this indirdep is not attached to a buf it was simply waiting
+	 * on completion to clear completehd.  free_indirdep() asserts
+	 * that nothing is dangling.
+	 */
+	if ((indirdep->ir_state & ONWORKLIST) == 0)
+		free_indirdep(indirdep);
 }
 
 /*
@@ -2054,23 +4756,28 @@ softdep_setup_allocindir_meta(nbp, ip, bp, ptrno, newblkno)
  * by one of the two routines above.
  */
 static void 
-setup_allocindir_phase2(bp, ip, aip)
+setup_allocindir_phase2(bp, ip, inodedep, aip, lbn)
 	struct buf *bp;		/* in-memory copy of the indirect block */
 	struct inode *ip;	/* inode for file being extended */
+	struct inodedep *inodedep; /* Inodedep for ip */
 	struct allocindir *aip;	/* allocindir allocated by the above routines */
+	ufs_lbn_t lbn;		/* Logical block number for this block. */
 {
 	struct worklist *wk;
+	struct fs *fs;
+	struct newblk *newblk;
 	struct indirdep *indirdep, *newindirdep;
-	struct bmsafemap *bmsafemap;
 	struct allocindir *oldaip;
 	struct freefrag *freefrag;
-	struct newblk *newblk;
+	struct mount *mp;
 	ufs2_daddr_t blkno;
 
+	mp = UFSTOVFS(ip->i_ump);
+	fs = ip->i_fs;
 	mtx_assert(&lk, MA_OWNED);
 	if (bp->b_lblkno >= 0)
 		panic("setup_allocindir_phase2: not indir blk");
-	for (indirdep = NULL, newindirdep = NULL; ; ) {
+	for (freefrag = NULL, indirdep = NULL, newindirdep = NULL; ; ) {
 		LIST_FOREACH(wk, &bp->b_dep, wk_list) {
 			if (wk->wk_type != D_INDIRDEP)
 				continue;
@@ -2079,49 +4786,41 @@ setup_allocindir_phase2(bp, ip, aip)
 		}
 		if (indirdep == NULL && newindirdep) {
 			indirdep = newindirdep;
-			WORKLIST_INSERT(&bp->b_dep, &indirdep->ir_list);
 			newindirdep = NULL;
+			WORKLIST_INSERT(&bp->b_dep, &indirdep->ir_list);
+			if (newblk_lookup(mp, dbtofsb(fs, bp->b_blkno), 0,
+			    &newblk)) {
+				indirdep->ir_state |= ONDEPLIST;
+				LIST_INSERT_HEAD(&newblk->nb_indirdeps,
+				    indirdep, ir_next);
+			} else
+				indirdep->ir_state |= DEPCOMPLETE;
 		}
 		if (indirdep) {
-			if (newblk_lookup(ip->i_fs, aip->ai_newblkno, 0,
-			    &newblk) == 0)
-				panic("setup_allocindir: lost block");
-			if (newblk->nb_state == DEPCOMPLETE) {
-				aip->ai_state |= DEPCOMPLETE;
-				aip->ai_buf = NULL;
-			} else {
-				bmsafemap = newblk->nb_bmsafemap;
-				aip->ai_buf = bmsafemap->sm_buf;
-				LIST_REMOVE(newblk, nb_deps);
-				LIST_INSERT_HEAD(&bmsafemap->sm_allocindirhd,
-				    aip, ai_deps);
-			}
-			LIST_REMOVE(newblk, nb_hash);
-			free(newblk, M_NEWBLK);
 			aip->ai_indirdep = indirdep;
 			/*
 			 * Check to see if there is an existing dependency
 			 * for this block. If there is, merge the old
-			 * dependency into the new one.
+			 * dependency into the new one.  This happens
+			 * as a result of reallocblk only.
 			 */
 			if (aip->ai_oldblkno == 0)
 				oldaip = NULL;
 			else
 
-				LIST_FOREACH(oldaip, &indirdep->ir_deplisthd, ai_next)
+				LIST_FOREACH(oldaip, &indirdep->ir_deplisthd,
+				    ai_next)
 					if (oldaip->ai_offset == aip->ai_offset)
 						break;
-			freefrag = NULL;
-			if (oldaip != NULL) {
-				if (oldaip->ai_newblkno != aip->ai_oldblkno)
-					panic("setup_allocindir_phase2: blkno");
-				aip->ai_oldblkno = oldaip->ai_oldblkno;
-				freefrag = aip->ai_freefrag;
-				aip->ai_freefrag = oldaip->ai_freefrag;
-				oldaip->ai_freefrag = NULL;
-				free_allocindir(oldaip, NULL);
-			}
+			if (oldaip != NULL)
+				freefrag = allocindir_merge(aip, oldaip);
 			LIST_INSERT_HEAD(&indirdep->ir_deplisthd, aip, ai_next);
+			KASSERT(aip->ai_offset >= 0 &&
+			    aip->ai_offset < NINDIR(ip->i_ump->um_fs),
+			    ("setup_allocindir_phase2: Bad offset %d",
+			    aip->ai_offset));
+			KASSERT(indirdep->ir_savebp != NULL,
+			    ("setup_allocindir_phase2 NULL ir_savebp"));
 			if (ip->i_ump->um_fstype == UFS1)
 				((ufs1_daddr_t *)indirdep->ir_savebp->b_data)
 				    [aip->ai_offset] = aip->ai_oldblkno;
@@ -2148,13 +4847,16 @@ setup_allocindir_phase2(bp, ip, aip)
 		}
 		newindirdep = malloc(sizeof(struct indirdep),
 			M_INDIRDEP, M_SOFTDEP_FLAGS);
-		workitem_alloc(&newindirdep->ir_list, D_INDIRDEP,
-		    UFSTOVFS(ip->i_ump));
+		workitem_alloc(&newindirdep->ir_list, D_INDIRDEP, mp);
 		newindirdep->ir_state = ATTACHED;
 		if (ip->i_ump->um_fstype == UFS1)
 			newindirdep->ir_state |= UFS1FMT;
+		newindirdep->ir_saveddata = NULL;
 		LIST_INIT(&newindirdep->ir_deplisthd);
 		LIST_INIT(&newindirdep->ir_donehd);
+		LIST_INIT(&newindirdep->ir_writehd);
+		LIST_INIT(&newindirdep->ir_completehd);
+		LIST_INIT(&newindirdep->ir_jwork);
 		if (bp->b_blkno == bp->b_lblkno) {
 			ufs_bmaparray(bp->b_vp, bp->b_lblkno, &blkno, bp,
 			    NULL, NULL);
@@ -2168,6 +4870,51 @@ setup_allocindir_phase2(bp, ip, aip)
 	}
 }
 
+/*
+ * Merge two allocindirs which refer to the same block.  Move newblock
+ * dependencies and setup the freefrags appropriately.
+ */
+static struct freefrag *
+allocindir_merge(aip, oldaip)
+	struct allocindir *aip;
+	struct allocindir *oldaip;
+{
+	struct newdirblk *newdirblk;
+	struct freefrag *freefrag;
+	struct worklist *wk;
+
+	if (oldaip->ai_newblkno != aip->ai_oldblkno)
+		panic("allocindir_merge: blkno");
+	aip->ai_oldblkno = oldaip->ai_oldblkno;
+	freefrag = aip->ai_freefrag;
+	aip->ai_freefrag = oldaip->ai_freefrag;
+	oldaip->ai_freefrag = NULL;
+	KASSERT(freefrag != NULL, ("setup_allocindir_phase2: No freefrag"));
+	/*
+	 * If we are tracking a new directory-block allocation,
+	 * move it from the old allocindir to the new allocindir.
+	 */
+	if ((wk = LIST_FIRST(&oldaip->ai_newdirblk)) != NULL) {
+		newdirblk = WK_NEWDIRBLK(wk);
+		WORKLIST_REMOVE(&newdirblk->db_list);
+		if (!LIST_EMPTY(&oldaip->ai_newdirblk))
+			panic("allocindir_merge: extra newdirblk");
+		WORKLIST_INSERT(&aip->ai_newdirblk, &newdirblk->db_list);
+	}
+	/*
+	 * We can skip journaling for this freefrag and just complete
+	 * any pending journal work for the allocindir that is being
+	 * removed after the freefrag completes.
+	 */
+	if (freefrag->ff_jfreefrag)
+		cancel_jfreefrag(freefrag->ff_jfreefrag);
+	LIST_REMOVE(oldaip, ai_next);
+	cancel_newblk(&oldaip->ai_block, &freefrag->ff_jwork);
+	free_newblk(&oldaip->ai_block);
+
+	return (freefrag);
+}
+
 /*
  * Block de-allocation dependencies.
  * 
@@ -2203,9 +4950,12 @@ softdep_setup_freeblocks(ip, length, flags)
 	off_t length;		/* The new length for the file */
 	int flags;		/* IO_EXT and/or IO_NORMAL */
 {
+	struct ufs1_dinode *dp1;
+	struct ufs2_dinode *dp2;
 	struct freeblks *freeblks;
 	struct inodedep *inodedep;
 	struct allocdirect *adp;
+	struct jfreeblk *jfreeblk;
 	struct bufobj *bo;
 	struct vnode *vp;
 	struct buf *bp;
@@ -2213,6 +4963,13 @@ softdep_setup_freeblocks(ip, length, flags)
 	ufs2_daddr_t extblocks, datablocks;
 	struct mount *mp;
 	int i, delay, error;
+	ufs2_daddr_t blkno;
+	ufs_lbn_t tmpval;
+	ufs_lbn_t lbn;
+	long oldextsize;
+	long oldsize;
+	int frags;
+	int needj;
 
 	fs = ip->i_fs;
 	mp = UFSTOVFS(ip->i_ump);
@@ -2221,32 +4978,53 @@ softdep_setup_freeblocks(ip, length, flags)
 	freeblks = malloc(sizeof(struct freeblks),
 		M_FREEBLKS, M_SOFTDEP_FLAGS|M_ZERO);
 	workitem_alloc(&freeblks->fb_list, D_FREEBLKS, mp);
+	LIST_INIT(&freeblks->fb_jfreeblkhd);
+	LIST_INIT(&freeblks->fb_jwork);
 	freeblks->fb_state = ATTACHED;
 	freeblks->fb_uid = ip->i_uid;
 	freeblks->fb_previousinum = ip->i_number;
 	freeblks->fb_devvp = ip->i_devvp;
+	freeblks->fb_chkcnt = 0;
 	ACQUIRE_LOCK(&lk);
+	/*
+	 * If we're truncating a removed file that will never be written
+	 * we don't need to journal the block frees.  The canceled journals
+	 * for the allocations will suffice.
+	 */
+	inodedep_lookup(mp, ip->i_number, DEPALLOC, &inodedep);
+	if ((inodedep->id_state & (UNLINKED | DEPCOMPLETE)) == UNLINKED ||
+	    (fs->fs_flags & FS_SUJ) == 0)
+		needj = 0;
+	else
+		needj = 1;
 	num_freeblkdep++;
 	FREE_LOCK(&lk);
 	extblocks = 0;
 	if (fs->fs_magic == FS_UFS2_MAGIC)
 		extblocks = btodb(fragroundup(fs, ip->i_din2->di_extsize));
 	datablocks = DIP(ip, i_blocks) - extblocks;
-	if ((flags & IO_NORMAL) == 0) {
-		freeblks->fb_oldsize = 0;
-		freeblks->fb_chkcnt = 0;
-	} else {
-		freeblks->fb_oldsize = ip->i_size;
+	if ((flags & IO_NORMAL) != 0) {
+		oldsize = ip->i_size;
 		ip->i_size = 0;
 		DIP_SET(ip, i_size, 0);
 		freeblks->fb_chkcnt = datablocks;
 		for (i = 0; i < NDADDR; i++) {
-			freeblks->fb_dblks[i] = DIP(ip, i_db[i]);
+			blkno = DIP(ip, i_db[i]);
 			DIP_SET(ip, i_db[i], 0);
+			if (blkno == 0)
+				continue;
+			frags = sblksize(fs, oldsize, i);
+			frags = numfrags(fs, frags);
+			newfreework(freeblks, NULL, i, blkno, frags, needj);
 		}
-		for (i = 0; i < NIADDR; i++) {
-			freeblks->fb_iblks[i] = DIP(ip, i_ib[i]);
+		for (i = 0, tmpval = NINDIR(fs), lbn = NDADDR; i < NIADDR;
+		    i++, tmpval *= NINDIR(fs)) {
+			blkno = DIP(ip, i_ib[i]);
 			DIP_SET(ip, i_ib[i], 0);
+			if (blkno) 
+				newfreework(freeblks, NULL, -lbn - i, blkno,
+				    fs->fs_frag, needj);
+			lbn += tmpval;
 		}
 		/*
 		 * If the file was removed, then the space being freed was
@@ -2259,17 +5037,23 @@ softdep_setup_freeblocks(ip, length, flags)
 			UFS_UNLOCK(ip->i_ump);
 		}
 	}
-	if ((flags & IO_EXT) == 0) {
-		freeblks->fb_oldextsize = 0;
-	} else {
-		freeblks->fb_oldextsize = ip->i_din2->di_extsize;
+	if ((flags & IO_EXT) != 0) {
+		oldextsize = ip->i_din2->di_extsize;
 		ip->i_din2->di_extsize = 0;
 		freeblks->fb_chkcnt += extblocks;
 		for (i = 0; i < NXADDR; i++) {
-			freeblks->fb_eblks[i] = ip->i_din2->di_extb[i];
+			blkno = ip->i_din2->di_extb[i];
 			ip->i_din2->di_extb[i] = 0;
+			if (blkno == 0)
+				continue;
+			frags = sblksize(fs, oldextsize, i);
+			frags = numfrags(fs, frags);
+			newfreework(freeblks, NULL, -1 - i, blkno, frags,
+			    needj);
 		}
 	}
+	if (LIST_EMPTY(&freeblks->fb_jfreeblkhd))
+		needj = 0;
 	DIP_SET(ip, i_blocks, DIP(ip, i_blocks) - freeblks->fb_chkcnt);
 	/*
 	 * Push the zero'ed inode to to its disk buffer so that we are free
@@ -2282,12 +5066,17 @@ softdep_setup_freeblocks(ip, length, flags)
 		brelse(bp);
 		softdep_error("softdep_setup_freeblocks", error);
 	}
-	if (ip->i_ump->um_fstype == UFS1)
-		*((struct ufs1_dinode *)bp->b_data +
-		    ino_to_fsbo(fs, ip->i_number)) = *ip->i_din1;
-	else
-		*((struct ufs2_dinode *)bp->b_data +
-		    ino_to_fsbo(fs, ip->i_number)) = *ip->i_din2;
+	if (ip->i_ump->um_fstype == UFS1) {
+		dp1 = ((struct ufs1_dinode *)bp->b_data +
+		    ino_to_fsbo(fs, ip->i_number));
+		ip->i_din1->di_freelink = dp1->di_freelink;
+		*dp1 = *ip->i_din1;
+	} else {
+		dp2 = ((struct ufs2_dinode *)bp->b_data +
+		    ino_to_fsbo(fs, ip->i_number));
+		ip->i_din2->di_freelink = dp2->di_freelink;
+		*dp2 = *ip->i_din2;
+	}
 	/*
 	 * Find and eliminate any inode dependencies.
 	 */
@@ -2304,7 +5093,9 @@ softdep_setup_freeblocks(ip, length, flags)
 	 */
 	delay = (inodedep->id_state & DEPCOMPLETE);
 	if (delay)
-		WORKLIST_INSERT(&inodedep->id_bufwait, &freeblks->fb_list);
+		WORKLIST_INSERT(&bp->b_dep, &freeblks->fb_list);
+	else if (needj)
+		freeblks->fb_state |= DEPCOMPLETE | COMPLETE;
 	/*
 	 * Because the file length has been truncated to zero, any
 	 * pending block allocation dependency structures associated
@@ -2318,14 +5109,19 @@ softdep_setup_freeblocks(ip, length, flags)
 		merge_inode_lists(&inodedep->id_newinoupdt,
 		    &inodedep->id_inoupdt);
 		while ((adp = TAILQ_FIRST(&inodedep->id_inoupdt)) != 0)
-			free_allocdirect(&inodedep->id_inoupdt, adp, delay);
+			cancel_allocdirect(&inodedep->id_inoupdt, adp,
+			    freeblks, delay);
 	}
 	if (flags & IO_EXT) {
 		merge_inode_lists(&inodedep->id_newextupdt,
 		    &inodedep->id_extupdt);
 		while ((adp = TAILQ_FIRST(&inodedep->id_extupdt)) != 0)
-			free_allocdirect(&inodedep->id_extupdt, adp, delay);
+			cancel_allocdirect(&inodedep->id_extupdt, adp,
+			    freeblks, delay);
 	}
+	LIST_FOREACH(jfreeblk, &freeblks->fb_jfreeblkhd, jf_deps)
+		add_to_journal(&jfreeblk->jf_list);
+
 	FREE_LOCK(&lk);
 	bdwrite(bp);
 	/*
@@ -2349,9 +5145,9 @@ restart:
 		BO_UNLOCK(bo);
 		ACQUIRE_LOCK(&lk);
 		(void) inodedep_lookup(mp, ip->i_number, 0, &inodedep);
-		deallocate_dependencies(bp, inodedep);
+		if (deallocate_dependencies(bp, inodedep, freeblks))
+			bp->b_flags |= B_INVAL | B_NOCACHE;
 		FREE_LOCK(&lk);
-		bp->b_flags |= B_INVAL | B_NOCACHE;
 		brelse(bp);
 		BO_LOCK(bo);
 		goto restart;
@@ -2361,7 +5157,7 @@ restart:
 	if (inodedep_lookup(mp, ip->i_number, 0, &inodedep) != 0)
 		(void) free_inodedep(inodedep);
 
-	if(delay) {
+	if (delay) {
 		freeblks->fb_state |= DEPCOMPLETE;
 		/*
 		 * If the inode with zeroed block pointers is now on disk
@@ -2371,16 +5167,16 @@ restart:
 		 * the request here than in the !delay case.
 		 */  
 		if ((freeblks->fb_state & ALLCOMPLETE) == ALLCOMPLETE)
-			add_to_worklist(&freeblks->fb_list);
+			add_to_worklist(&freeblks->fb_list, 1);
 	}
 
 	FREE_LOCK(&lk);
 	/*
-	 * If the inode has never been written to disk (delay == 0),
-	 * then we can process the freeblks now that we have deleted
-	 * the dependencies.
+	 * If the inode has never been written to disk (delay == 0) and
+	 * we're not waiting on any journal writes, then we can process the
+	 * freeblks now that we have deleted the dependencies.
 	 */
-	if (!delay)
+	if (!delay && !needj)
 		handle_workitem_freeblocks(freeblks, 0);
 }
 
@@ -2389,19 +5185,23 @@ restart:
  * be reallocated to a new vnode. The buffer must be locked, thus,
  * no I/O completion operations can occur while we are manipulating
  * its associated dependencies. The mutex is held so that other I/O's
- * associated with related dependencies do not occur.
+ * associated with related dependencies do not occur.  Returns 1 if
+ * all dependencies were cleared, 0 otherwise.
  */
-static void
-deallocate_dependencies(bp, inodedep)
+static int
+deallocate_dependencies(bp, inodedep, freeblks)
 	struct buf *bp;
 	struct inodedep *inodedep;
+	struct freeblks *freeblks;
 {
 	struct worklist *wk;
 	struct indirdep *indirdep;
+	struct newdirblk *newdirblk;
 	struct allocindir *aip;
 	struct pagedep *pagedep;
+	struct jremref *jremref;
+	struct jmvref *jmvref;
 	struct dirrem *dirrem;
-	struct diradd *dap;
 	int i;
 
 	mtx_assert(&lk, MA_OWNED);
@@ -2410,47 +5210,24 @@ deallocate_dependencies(bp, inodedep)
 
 		case D_INDIRDEP:
 			indirdep = WK_INDIRDEP(wk);
-			/*
-			 * None of the indirect pointers will ever be visible,
-			 * so they can simply be tossed. GOINGAWAY ensures
-			 * that allocated pointers will be saved in the buffer
-			 * cache until they are freed. Note that they will
-			 * only be able to be found by their physical address
-			 * since the inode mapping the logical address will
-			 * be gone. The save buffer used for the safe copy
-			 * was allocated in setup_allocindir_phase2 using
-			 * the physical address so it could be used for this
-			 * purpose. Hence we swap the safe copy with the real
-			 * copy, allowing the safe copy to be freed and holding
-			 * on to the real copy for later use in indir_trunc.
-			 */
-			if (indirdep->ir_state & GOINGAWAY)
-				panic("deallocate_dependencies: already gone");
-			indirdep->ir_state |= GOINGAWAY;
-			VFSTOUFS(bp->b_vp->v_mount)->um_numindirdeps += 1;
-			while ((aip = LIST_FIRST(&indirdep->ir_deplisthd)) != 0)
-				free_allocindir(aip, inodedep);
 			if (bp->b_lblkno >= 0 ||
 			    bp->b_blkno != indirdep->ir_savebp->b_lblkno)
 				panic("deallocate_dependencies: not indir");
-			bcopy(bp->b_data, indirdep->ir_savebp->b_data,
-			    bp->b_bcount);
-			WORKLIST_REMOVE(wk);
-			WORKLIST_INSERT(&indirdep->ir_savebp->b_dep, wk);
+			cancel_indirdep(indirdep, bp, inodedep, freeblks);
 			continue;
 
 		case D_PAGEDEP:
 			pagedep = WK_PAGEDEP(wk);
 			/*
-			 * None of the directory additions will ever be
-			 * visible, so they can simply be tossed.
+			 * There should be no directory add dependencies present
+			 * as the directory could not be truncated until all
+			 * children were removed.
 			 */
+			KASSERT(LIST_FIRST(&pagedep->pd_pendinghd) == NULL,
+			    ("deallocate_dependencies: pendinghd != NULL"));
 			for (i = 0; i < DAHASHSZ; i++)
-				while ((dap =
-				    LIST_FIRST(&pagedep->pd_diraddhd[i])))
-					free_diradd(dap);
-			while ((dap = LIST_FIRST(&pagedep->pd_pendinghd)) != 0)
-				free_diradd(dap);
+				KASSERT(LIST_FIRST(&pagedep->pd_diraddhd[i]) == NULL,
+				    ("deallocate_dependencies: diraddhd != NULL"));
 			/*
 			 * Copy any directory remove dependencies to the list
 			 * to be processed after the zero'ed inode is written.
@@ -2458,28 +5235,40 @@ deallocate_dependencies(bp, inodedep)
 			 * can be dumped directly onto the work list.
 			 */
 			LIST_FOREACH(dirrem, &pagedep->pd_dirremhd, dm_next) {
+				/*
+				 * If there are any dirrems we wait for
+				 * the journal write to complete and
+				 * then restart the buf scan as the lock
+				 * has been dropped.
+				 */
+				while ((jremref =
+				    LIST_FIRST(&dirrem->dm_jremrefhd))
+				    != NULL) {
+					stat_jwait_filepage++;
+					jwait(&jremref->jr_list);
+					return (0);
+				}
 				LIST_REMOVE(dirrem, dm_next);
 				dirrem->dm_dirinum = pagedep->pd_ino;
 				if (inodedep == NULL ||
 				    (inodedep->id_state & ALLCOMPLETE) ==
-				     ALLCOMPLETE)
-					add_to_worklist(&dirrem->dm_list);
-				else
+				     ALLCOMPLETE) {
+					dirrem->dm_state |= COMPLETE;
+					add_to_worklist(&dirrem->dm_list, 0);
+				} else
 					WORKLIST_INSERT(&inodedep->id_bufwait,
 					    &dirrem->dm_list);
 			}
 			if ((pagedep->pd_state & NEWBLOCK) != 0) {
-				LIST_FOREACH(wk, &inodedep->id_bufwait, wk_list)
-					if (wk->wk_type == D_NEWDIRBLK &&
-					    WK_NEWDIRBLK(wk)->db_pagedep ==
-					      pagedep)
-						break;
-				if (wk != NULL) {
-					WORKLIST_REMOVE(wk);
-					free_newdirblk(WK_NEWDIRBLK(wk));
-				} else
-					panic("deallocate_dependencies: "
-					      "lost pagedep");
+				newdirblk = pagedep->pd_newdirblk;
+				WORKLIST_REMOVE(&newdirblk->db_list);
+				free_newdirblk(newdirblk);
+			}
+			while ((jmvref = LIST_FIRST(&pagedep->pd_jmvrefhd))
+			    != NULL) {
+				stat_jwait_filepage++;
+				jwait(&jmvref->jm_list);
+				return (0);
 			}
 			WORKLIST_REMOVE(&pagedep->pd_list);
 			LIST_REMOVE(pagedep, pd_hash);
@@ -2487,7 +5276,8 @@ deallocate_dependencies(bp, inodedep)
 			continue;
 
 		case D_ALLOCINDIR:
-			free_allocindir(WK_ALLOCINDIR(wk), inodedep);
+			aip = WK_ALLOCINDIR(wk);
+			cancel_allocindir(aip, inodedep, freeblks);
 			continue;
 
 		case D_ALLOCDIRECT:
@@ -2502,46 +5292,155 @@ deallocate_dependencies(bp, inodedep)
 			/* NOTREACHED */
 		}
 	}
+
+	return (1);
 }
 
 /*
- * Free an allocdirect. Generate a new freefrag work request if appropriate.
- * This routine must be called with splbio interrupts blocked.
+ * An allocdirect is being canceled due to a truncate.  We must make sure
+ * the journal entry is released in concert with the blkfree that releases
+ * the storage.  Completed journal entries must not be released until the
+ * space is no longer pointed to by the inode or in the bitmap.
  */
 static void
-free_allocdirect(adphead, adp, delay)
+cancel_allocdirect(adphead, adp, freeblks, delay)
 	struct allocdirectlst *adphead;
 	struct allocdirect *adp;
+	struct freeblks *freeblks;
 	int delay;
 {
+	struct freework *freework;
+	struct newblk *newblk;
+	struct worklist *wk;
+	ufs_lbn_t lbn;
+
+	TAILQ_REMOVE(adphead, adp, ad_next);
+	newblk = (struct newblk *)adp;
+	/*
+	 * If the journal hasn't been written the jnewblk must be passed
+	 * to the call to ffs_freeblk that reclaims the space.  We accomplish
+	 * this by linking the journal dependency into the freework to be
+	 * freed when freework_freeblock() is called.  If the journal has
+	 * been written we can simply reclaim the journal space when the
+	 * freeblks work is complete.
+	 */
+	if (newblk->nb_jnewblk == NULL) {
+		cancel_newblk(newblk, &freeblks->fb_jwork);
+		goto found;
+	}
+	lbn = newblk->nb_jnewblk->jn_lbn;
+	/*
+	 * Find the correct freework structure so it releases the canceled
+	 * journal when the bitmap is cleared.  This preserves rollback
+	 * until the allocation is reverted.
+	 */
+	LIST_FOREACH(wk, &freeblks->fb_freeworkhd, wk_list) {
+		freework = WK_FREEWORK(wk);
+		if (freework->fw_lbn != lbn)
+			continue;
+		cancel_newblk(newblk, &freework->fw_jwork);
+		goto found;
+	}
+	panic("cancel_allocdirect: Freework not found for lbn %jd\n", lbn);
+found:
+	if (delay)
+		WORKLIST_INSERT(&adp->ad_inodedep->id_bufwait,
+		    &newblk->nb_list);
+	else
+		free_newblk(newblk);
+	return;
+}
+
+
+static void
+cancel_newblk(newblk, wkhd)
+	struct newblk *newblk;
+	struct workhead *wkhd;
+{
+	struct indirdep *indirdep;
+	struct allocindir *aip;
+
+	while ((indirdep = LIST_FIRST(&newblk->nb_indirdeps)) != NULL) {
+		indirdep->ir_state &= ~ONDEPLIST;
+		LIST_REMOVE(indirdep, ir_next);
+		/*
+		 * If an indirdep is not on the buf worklist we need to
+		 * free it here as deallocate_dependencies() will never
+		 * find it.  These pointers were never visible on disk and
+		 * can be discarded immediately.
+		 */
+		while ((aip = LIST_FIRST(&indirdep->ir_completehd)) != NULL) {
+			LIST_REMOVE(aip, ai_next);
+			cancel_newblk(&aip->ai_block, wkhd);
+			free_newblk(&aip->ai_block);
+		}
+		/*
+		 * If this indirdep is not attached to a buf it was simply
+		 * waiting on completion to clear completehd.  free_indirdep()
+		 * asserts that nothing is dangling.
+		 */
+		if ((indirdep->ir_state & ONWORKLIST) == 0)
+			free_indirdep(indirdep);
+	}
+	if (newblk->nb_state & ONDEPLIST) {
+		newblk->nb_state &= ~ONDEPLIST;
+		LIST_REMOVE(newblk, nb_deps);
+	}
+	if (newblk->nb_state & ONWORKLIST)
+		WORKLIST_REMOVE(&newblk->nb_list);
+	/*
+	 * If the journal entry hasn't been written we hold onto the dep
+	 * until it is safe to free along with the other journal work.
+	 */
+	if (newblk->nb_jnewblk != NULL) {
+		cancel_jnewblk(newblk->nb_jnewblk, wkhd);
+		newblk->nb_jnewblk = NULL;
+	}
+	if (!LIST_EMPTY(&newblk->nb_jwork))
+		jwork_move(wkhd, &newblk->nb_jwork);
+}
+
+/*
+ * Free a newblk. Generate a new freefrag work request if appropriate.
+ * This must be called after the inode pointer and any direct block pointers
+ * are valid or fully removed via truncate or frag extension.
+ */
+static void
+free_newblk(newblk)
+	struct newblk *newblk;
+{
+	struct indirdep *indirdep;
 	struct newdirblk *newdirblk;
+	struct freefrag *freefrag;
 	struct worklist *wk;
 
 	mtx_assert(&lk, MA_OWNED);
-	if ((adp->ad_state & DEPCOMPLETE) == 0)
-		LIST_REMOVE(adp, ad_deps);
-	TAILQ_REMOVE(adphead, adp, ad_next);
-	if ((adp->ad_state & COMPLETE) == 0)
-		WORKLIST_REMOVE(&adp->ad_list);
-	if (adp->ad_freefrag != NULL) {
-		if (delay)
-			WORKLIST_INSERT(&adp->ad_inodedep->id_bufwait,
-			    &adp->ad_freefrag->ff_list);
-		else
-			add_to_worklist(&adp->ad_freefrag->ff_list);
+	if (newblk->nb_state & ONDEPLIST)
+		LIST_REMOVE(newblk, nb_deps);
+	if (newblk->nb_state & ONWORKLIST)
+		WORKLIST_REMOVE(&newblk->nb_list);
+	LIST_REMOVE(newblk, nb_hash);
+	if ((freefrag = newblk->nb_freefrag) != NULL) {
+		freefrag->ff_state |= COMPLETE;
+		if ((freefrag->ff_state & ALLCOMPLETE) == ALLCOMPLETE)
+			add_to_worklist(&freefrag->ff_list, 0);
 	}
-	if ((wk = LIST_FIRST(&adp->ad_newdirblk)) != NULL) {
+	if ((wk = LIST_FIRST(&newblk->nb_newdirblk)) != NULL) {
 		newdirblk = WK_NEWDIRBLK(wk);
 		WORKLIST_REMOVE(&newdirblk->db_list);
-		if (!LIST_EMPTY(&adp->ad_newdirblk))
-			panic("free_allocdirect: extra newdirblk");
-		if (delay)
-			WORKLIST_INSERT(&adp->ad_inodedep->id_bufwait,
-			    &newdirblk->db_list);
-		else
-			free_newdirblk(newdirblk);
+		if (!LIST_EMPTY(&newblk->nb_newdirblk))
+			panic("free_newblk: extra newdirblk");
+		free_newdirblk(newdirblk);
 	}
-	WORKITEM_FREE(adp, D_ALLOCDIRECT);
+	while ((indirdep = LIST_FIRST(&newblk->nb_indirdeps)) != NULL) {
+		indirdep->ir_state |= DEPCOMPLETE;
+		indirdep_complete(indirdep);
+	}
+	KASSERT(newblk->nb_jnewblk == NULL,
+	    ("free_newblk; jnewblk %p still attached", newblk->nb_jnewblk));
+	handle_jwork(&newblk->nb_jwork);
+	newblk->nb_list.wk_type = D_NEWBLK;
+	WORKITEM_FREE(newblk, D_NEWBLK);
 }
 
 /*
@@ -2554,6 +5453,7 @@ free_newdirblk(newdirblk)
 {
 	struct pagedep *pagedep;
 	struct diradd *dap;
+	struct worklist *wk;
 	int i;
 
 	mtx_assert(&lk, MA_OWNED);
@@ -2571,17 +5471,25 @@ free_newdirblk(newdirblk)
 	pagedep->pd_state &= ~NEWBLOCK;
 	if ((pagedep->pd_state & ONWORKLIST) == 0)
 		while ((dap = LIST_FIRST(&pagedep->pd_pendinghd)) != NULL)
-			free_diradd(dap);
+			free_diradd(dap, NULL);
 	/*
 	 * If no dependencies remain, the pagedep will be freed.
 	 */
 	for (i = 0; i < DAHASHSZ; i++)
 		if (!LIST_EMPTY(&pagedep->pd_diraddhd[i]))
 			break;
-	if (i == DAHASHSZ && (pagedep->pd_state & ONWORKLIST) == 0) {
+	if (i == DAHASHSZ && (pagedep->pd_state & ONWORKLIST) == 0 &&
+	    LIST_EMPTY(&pagedep->pd_jmvrefhd)) {
+		KASSERT(LIST_FIRST(&pagedep->pd_dirremhd) == NULL,
+		    ("free_newdirblk: Freeing non-free pagedep %p", pagedep));
 		LIST_REMOVE(pagedep, pd_hash);
 		WORKITEM_FREE(pagedep, D_PAGEDEP);
 	}
+	/* Should only ever be one item in the list. */
+	while ((wk = LIST_FIRST(&newdirblk->db_mkdir)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		handle_written_mkdir(WK_MKDIR(wk), MKDIR_BODY);
+	}
 	WORKITEM_FREE(newdirblk, D_NEWDIRBLK);
 }
 
@@ -2608,6 +5516,7 @@ softdep_freefile(pvp, ino, mode)
 	freefile->fx_mode = mode;
 	freefile->fx_oldinum = ino;
 	freefile->fx_devvp = ip->i_devvp;
+	LIST_INIT(&freefile->fx_jwork);
 	if ((ip->i_flag & IN_SPACECOUNTED) == 0) {
 		UFS_LOCK(ip->i_ump);
 		ip->i_fs->fs_pendinginodes += 1;
@@ -2618,11 +5527,34 @@ softdep_freefile(pvp, ino, mode)
 	 * If the inodedep does not exist, then the zero'ed inode has
 	 * been written to disk. If the allocated inode has never been
 	 * written to disk, then the on-disk inode is zero'ed. In either
-	 * case we can free the file immediately.
+	 * case we can free the file immediately.  If the journal was
+	 * canceled before being written the inode will never make it to
+	 * disk and we must send the canceled journal entrys to
+	 * ffs_freefile() to be cleared in conjunction with the bitmap.
+	 * Any blocks waiting on the inode to write can be safely freed
+	 * here as it will never been written.
 	 */
 	ACQUIRE_LOCK(&lk);
-	if (inodedep_lookup(pvp->v_mount, ino, 0, &inodedep) == 0 ||
-	    check_inode_unwritten(inodedep)) {
+	inodedep_lookup(pvp->v_mount, ino, 0, &inodedep);
+	/*
+	 * Remove this inode from the unlinked list and set
+	 * GOINGAWAY as appropriate to indicate that this inode
+	 * will never be written.
+	 */
+	if (inodedep && inodedep->id_state & UNLINKED) {
+		/*
+		 * Save the journal work to be freed with the bitmap
+		 * before we clear UNLINKED.  Otherwise it can be lost
+		 * if the inode block is written.
+		 */
+		handle_bufwait(inodedep, &freefile->fx_jwork);
+		clear_unlinked_inodedep(inodedep);
+		/* Re-acquire inodedep as we've dropped lk. */
+		inodedep_lookup(pvp->v_mount, ino, 0, &inodedep);
+		if (inodedep && (inodedep->id_state & DEPCOMPLETE) == 0)
+			inodedep->id_state |= GOINGAWAY;
+	}
+	if (inodedep == NULL || check_inode_unwritten(inodedep)) {
 		FREE_LOCK(&lk);
 		handle_workitem_freefile(freefile);
 		return;
@@ -2654,7 +5586,8 @@ check_inode_unwritten(inodedep)
 {
 
 	mtx_assert(&lk, MA_OWNED);
-	if ((inodedep->id_state & DEPCOMPLETE) != 0 ||
+
+	if ((inodedep->id_state & (DEPCOMPLETE | UNLINKED)) != 0 ||
 	    !LIST_EMPTY(&inodedep->id_pendinghd) ||
 	    !LIST_EMPTY(&inodedep->id_bufwait) ||
 	    !LIST_EMPTY(&inodedep->id_inowait) ||
@@ -2662,9 +5595,9 @@ check_inode_unwritten(inodedep)
 	    !TAILQ_EMPTY(&inodedep->id_newinoupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_extupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_newextupdt) ||
+	    inodedep->id_mkdiradd != NULL || 
 	    inodedep->id_nlinkdelta != 0)
 		return (0);
-
 	/*
 	 * Another process might be in initiate_write_inodeblock_ufs[12]
 	 * trying to allocate memory without holding "Softdep Lock".
@@ -2673,9 +5606,11 @@ check_inode_unwritten(inodedep)
 	    inodedep->id_savedino1 == NULL)
 		return (0);
 
+	if (inodedep->id_state & ONDEPLIST)
+		LIST_REMOVE(inodedep, id_deps);
+	inodedep->id_state &= ~ONDEPLIST;
 	inodedep->id_state |= ALLCOMPLETE;
-	LIST_REMOVE(inodedep, id_deps);
-	inodedep->id_buf = NULL;
+	inodedep->id_bmsafemap = NULL;
 	if (inodedep->id_state & ONWORKLIST)
 		WORKLIST_REMOVE(&inodedep->id_list);
 	if (inodedep->id_savedino1 != NULL) {
@@ -2696,23 +5631,149 @@ free_inodedep(inodedep)
 {
 
 	mtx_assert(&lk, MA_OWNED);
-	if ((inodedep->id_state & ONWORKLIST) != 0 ||
+	if ((inodedep->id_state & (ONWORKLIST | UNLINKED)) != 0 ||
 	    (inodedep->id_state & ALLCOMPLETE) != ALLCOMPLETE ||
+	    !LIST_EMPTY(&inodedep->id_dirremhd) ||
 	    !LIST_EMPTY(&inodedep->id_pendinghd) ||
 	    !LIST_EMPTY(&inodedep->id_bufwait) ||
 	    !LIST_EMPTY(&inodedep->id_inowait) ||
+	    !TAILQ_EMPTY(&inodedep->id_inoreflst) ||
 	    !TAILQ_EMPTY(&inodedep->id_inoupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_newinoupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_extupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_newextupdt) ||
-	    inodedep->id_nlinkdelta != 0 || inodedep->id_savedino1 != NULL)
+	    inodedep->id_mkdiradd != NULL ||
+	    inodedep->id_nlinkdelta != 0 ||
+	    inodedep->id_savedino1 != NULL)
 		return (0);
+	if (inodedep->id_state & ONDEPLIST)
+		LIST_REMOVE(inodedep, id_deps);
 	LIST_REMOVE(inodedep, id_hash);
 	WORKITEM_FREE(inodedep, D_INODEDEP);
 	num_inodedep -= 1;
 	return (1);
 }
 
+/*
+ * Free the block referenced by a freework structure.  The parent freeblks
+ * structure is released and completed when the final cg bitmap reaches
+ * the disk.  This routine may be freeing a jnewblk which never made it to
+ * disk in which case we do not have to wait as the operation is undone
+ * in memory immediately.
+ */
+static void
+freework_freeblock(freework)
+	struct freework *freework;
+{
+	struct freeblks *freeblks;
+	struct ufsmount *ump;
+	struct workhead wkhd;
+	struct fs *fs;
+	int complete;
+	int pending;
+	int bsize;
+	int needj;
+
+	freeblks = freework->fw_freeblks;
+	ump = VFSTOUFS(freeblks->fb_list.wk_mp);
+	fs = ump->um_fs;
+	needj = freeblks->fb_list.wk_mp->mnt_kern_flag & MNTK_SUJ;
+	complete = 0;
+	LIST_INIT(&wkhd);
+	/*
+	 * If we are canceling an existing jnewblk pass it to the free
+	 * routine, otherwise pass the freeblk which will ultimately
+	 * release the freeblks.  If we're not journaling, we can just
+	 * free the freeblks immediately.
+	 */
+	if (!LIST_EMPTY(&freework->fw_jwork)) {
+		LIST_SWAP(&wkhd, &freework->fw_jwork, worklist, wk_list);
+		complete = 1;
+	} else if (needj)
+		WORKLIST_INSERT_UNLOCKED(&wkhd, &freework->fw_list);
+	bsize = lfragtosize(fs, freework->fw_frags);
+	pending = btodb(bsize);
+	ACQUIRE_LOCK(&lk);
+	freeblks->fb_chkcnt -= pending;
+	FREE_LOCK(&lk);
+	/*
+	 * extattr blocks don't show up in pending blocks.  XXX why?
+	 */
+	if (freework->fw_lbn >= 0 || freework->fw_lbn <= -NDADDR) {
+		UFS_LOCK(ump);
+		fs->fs_pendingblocks -= pending;
+		UFS_UNLOCK(ump);
+	}
+	ffs_blkfree(ump, fs, freeblks->fb_devvp, freework->fw_blkno,
+	    bsize, freeblks->fb_previousinum, &wkhd);
+	if (complete == 0 && needj)
+		return;
+	/*
+	 * The jnewblk will be discarded and the bits in the map never
+	 * made it to disk.  We can immediately free the freeblk.
+	 */
+	ACQUIRE_LOCK(&lk);
+	handle_written_freework(freework);
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Start, continue, or finish the process of freeing an indirect block tree.
+ * The free operation may be paused at any point with fw_off containing the
+ * offset to restart from.  This enables us to implement some flow control
+ * for large truncates which may fan out and generate a huge number of
+ * dependencies.
+ */
+static void
+handle_workitem_indirblk(freework)
+	struct freework *freework;
+{
+	struct freeblks *freeblks;
+	struct ufsmount *ump;
+	struct fs *fs;
+
+
+	freeblks = freework->fw_freeblks;
+	ump = VFSTOUFS(freeblks->fb_list.wk_mp);
+	fs = ump->um_fs;
+	if (freework->fw_off == NINDIR(fs))
+		freework_freeblock(freework);
+	else
+		indir_trunc(freework, fsbtodb(fs, freework->fw_blkno),
+		    freework->fw_lbn);
+}
+
+/*
+ * Called when a freework structure attached to a cg buf is written.  The
+ * ref on either the parent or the freeblks structure is released and
+ * either may be added to the worklist if it is the final ref.
+ */
+static void
+handle_written_freework(freework)
+	struct freework *freework;
+{
+	struct freeblks *freeblks;
+	struct freework *parent;
+
+	freeblks = freework->fw_freeblks;
+	parent = freework->fw_parent;
+	if (parent) {
+		if (--parent->fw_ref != 0)
+			parent = NULL;
+		freeblks = NULL;
+	} else if (--freeblks->fb_ref != 0)
+		freeblks = NULL;
+	WORKITEM_FREE(freework, D_FREEWORK);
+	/*
+	 * Don't delay these block frees or it takes an intolerable amount
+	 * of time to process truncates and free their journal entries.
+	 */
+	if (freeblks)
+		add_to_worklist(&freeblks->fb_list, 1);
+	if (parent)
+		add_to_worklist(&parent->fw_list, 1);
+}
+
 /*
  * This workitem routine performs the block de-allocation.
  * The workitem is added to the pending list after the updated
@@ -2725,100 +5786,80 @@ static void
 handle_workitem_freeblocks(freeblks, flags)
 	struct freeblks *freeblks;
 	int flags;
+{
+	struct freework *freework;
+	struct worklist *wk;
+
+	KASSERT(LIST_EMPTY(&freeblks->fb_jfreeblkhd),
+	    ("handle_workitem_freeblocks: Journal entries not written."));
+	if (LIST_EMPTY(&freeblks->fb_freeworkhd)) {
+		handle_complete_freeblocks(freeblks);
+		return;
+	}
+	freeblks->fb_ref++;
+	while ((wk = LIST_FIRST(&freeblks->fb_freeworkhd)) != NULL) {
+		KASSERT(wk->wk_type == D_FREEWORK,
+		    ("handle_workitem_freeblocks: Unknown type %s",
+		    TYPENAME(wk->wk_type)));
+		WORKLIST_REMOVE_UNLOCKED(wk);
+		freework = WK_FREEWORK(wk);
+		if (freework->fw_lbn <= -NDADDR)
+			handle_workitem_indirblk(freework);
+		else
+			freework_freeblock(freework);
+	}
+	ACQUIRE_LOCK(&lk);
+	if (--freeblks->fb_ref != 0)
+		freeblks = NULL;
+	FREE_LOCK(&lk);
+	if (freeblks)
+		handle_complete_freeblocks(freeblks);
+}
+
+/*
+ * Once all of the freework workitems are complete we can retire the
+ * freeblocks dependency and any journal work awaiting completion.  This
+ * can not be called until all other dependencies are stable on disk.
+ */
+static void
+handle_complete_freeblocks(freeblks)
+	struct freeblks *freeblks;
 {
 	struct inode *ip;
 	struct vnode *vp;
 	struct fs *fs;
 	struct ufsmount *ump;
-	int i, nblocks, level, bsize;
-	ufs2_daddr_t bn, blocksreleased = 0;
-	int error, allerror = 0;
-	ufs_lbn_t baselbns[NIADDR], tmpval;
-	int fs_pendingblocks;
+	int flags;
 
 	ump = VFSTOUFS(freeblks->fb_list.wk_mp);
 	fs = ump->um_fs;
-	fs_pendingblocks = 0;
-	tmpval = 1;
-	baselbns[0] = NDADDR;
-	for (i = 1; i < NIADDR; i++) {
-		tmpval *= NINDIR(fs);
-		baselbns[i] = baselbns[i - 1] + tmpval;
-	}
-	nblocks = btodb(fs->fs_bsize);
-	blocksreleased = 0;
-	/*
-	 * Release all extended attribute blocks or frags.
-	 */
-	if (freeblks->fb_oldextsize > 0) {
-		for (i = (NXADDR - 1); i >= 0; i--) {
-			if ((bn = freeblks->fb_eblks[i]) == 0)
-				continue;
-			bsize = sblksize(fs, freeblks->fb_oldextsize, i);
-			ffs_blkfree(ump, fs, freeblks->fb_devvp, bn, bsize,
-			    freeblks->fb_previousinum);
-			blocksreleased += btodb(bsize);
-		}
-	}
-	/*
-	 * Release all data blocks or frags.
-	 */
-	if (freeblks->fb_oldsize > 0) {
-		/*
-		 * Indirect blocks first.
-		 */
-		for (level = (NIADDR - 1); level >= 0; level--) {
-			if ((bn = freeblks->fb_iblks[level]) == 0)
-				continue;
-			if ((error = indir_trunc(freeblks, fsbtodb(fs, bn),
-			    level, baselbns[level], &blocksreleased)) != 0)
-				allerror = error;
-			ffs_blkfree(ump, fs, freeblks->fb_devvp, bn,
-			    fs->fs_bsize, freeblks->fb_previousinum);
-			fs_pendingblocks += nblocks;
-			blocksreleased += nblocks;
-		}
-		/*
-		 * All direct blocks or frags.
-		 */
-		for (i = (NDADDR - 1); i >= 0; i--) {
-			if ((bn = freeblks->fb_dblks[i]) == 0)
-				continue;
-			bsize = sblksize(fs, freeblks->fb_oldsize, i);
-			ffs_blkfree(ump, fs, freeblks->fb_devvp, bn, bsize,
-			    freeblks->fb_previousinum);
-			fs_pendingblocks += btodb(bsize);
-			blocksreleased += btodb(bsize);
-		}
-	}
-	UFS_LOCK(ump);
-	fs->fs_pendingblocks -= fs_pendingblocks;
-	UFS_UNLOCK(ump);
+	flags = LK_NOWAIT;
+
 	/*
 	 * If we still have not finished background cleanup, then check
 	 * to see if the block count needs to be adjusted.
 	 */
-	if (freeblks->fb_chkcnt != blocksreleased &&
-	    (fs->fs_flags & FS_UNCLEAN) != 0 &&
+	if (freeblks->fb_chkcnt != 0 && (fs->fs_flags & FS_UNCLEAN) != 0 &&
 	    ffs_vgetf(freeblks->fb_list.wk_mp, freeblks->fb_previousinum,
-		(flags & LK_NOWAIT) | LK_EXCLUSIVE, &vp, FFSV_FORCEINSMQ)
-	    == 0) {
+	    (flags & LK_NOWAIT) | LK_EXCLUSIVE, &vp, FFSV_FORCEINSMQ) == 0) {
 		ip = VTOI(vp);
-		DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + \
-		    freeblks->fb_chkcnt - blocksreleased);
+		DIP_SET(ip, i_blocks, DIP(ip, i_blocks) + freeblks->fb_chkcnt);
 		ip->i_flag |= IN_CHANGE;
 		vput(vp);
 	}
 
 #ifdef INVARIANTS
-	if (freeblks->fb_chkcnt != blocksreleased &&
+	if (freeblks->fb_chkcnt != 0 && 
 	    ((fs->fs_flags & FS_UNCLEAN) == 0 || (flags & LK_NOWAIT) != 0))
 		printf("handle_workitem_freeblocks: block count\n");
-	if (allerror)
-		softdep_error("handle_workitem_freeblks", allerror);
 #endif /* INVARIANTS */
 
 	ACQUIRE_LOCK(&lk);
+	/*
+	 * All of the freeblock deps must be complete prior to this call
+	 * so it's now safe to complete earlier outstanding journal entries.
+	 */
+	handle_jwork(&freeblks->fb_jwork);
 	WORKITEM_FREE(freeblks, D_FREEBLKS);
 	num_freeblkdep--;
 	FREE_LOCK(&lk);
@@ -2830,29 +5871,42 @@ handle_workitem_freeblocks(freeblks, flags)
  * and recursive calls to indirtrunc must be used to cleanse other indirect
  * blocks.
  */
-static int
-indir_trunc(freeblks, dbn, level, lbn, countp)
-	struct freeblks *freeblks;
+static void
+indir_trunc(freework, dbn, lbn)
+	struct freework *freework;
 	ufs2_daddr_t dbn;
-	int level;
 	ufs_lbn_t lbn;
-	ufs2_daddr_t *countp;
 {
+	struct freework *nfreework;
+	struct workhead wkhd;
+	struct jnewblk *jnewblk;
+	struct freeblks *freeblks;
 	struct buf *bp;
 	struct fs *fs;
+	struct worklist *wkn;
 	struct worklist *wk;
 	struct indirdep *indirdep;
 	struct ufsmount *ump;
 	ufs1_daddr_t *bap1 = 0;
-	ufs2_daddr_t nb, *bap2 = 0;
+	ufs2_daddr_t nb, nnb, *bap2 = 0;
 	ufs_lbn_t lbnadd;
 	int i, nblocks, ufs1fmt;
-	int error, allerror = 0;
 	int fs_pendingblocks;
+	int freedeps;
+	int needj;
+	int level;
+	int cnt;
 
+	LIST_INIT(&wkhd);
+	level = lbn_level(lbn);
+	if (level == -1)
+		panic("indir_trunc: Invalid lbn %jd\n", lbn);
+	freeblks = freework->fw_freeblks;
 	ump = VFSTOUFS(freeblks->fb_list.wk_mp);
 	fs = ump->um_fs;
 	fs_pendingblocks = 0;
+	freedeps = 0;
+	needj = UFSTOVFS(ump)->mnt_kern_flag & MNTK_SUJ;
 	lbnadd = 1;
 	for (i = level; i > 0; i--)
 		lbnadd *= NINDIR(fs);
@@ -2877,13 +5931,14 @@ indir_trunc(freeblks, dbn, level, lbn, countp)
 	ACQUIRE_LOCK(&lk);
 	if (bp != NULL && (wk = LIST_FIRST(&bp->b_dep)) != NULL) {
 		if (wk->wk_type != D_INDIRDEP ||
-		    (indirdep = WK_INDIRDEP(wk))->ir_savebp != bp ||
-		    (indirdep->ir_state & GOINGAWAY) == 0)
-			panic("indir_trunc: lost indirdep");
-		WORKLIST_REMOVE(wk);
-		WORKITEM_FREE(indirdep, D_INDIRDEP);
+		    (wk->wk_state & GOINGAWAY) == 0)
+			panic("indir_trunc: lost indirdep %p", wk);
+		indirdep = WK_INDIRDEP(wk);
+		LIST_SWAP(&wkhd, &indirdep->ir_jwork, worklist, wk_list);
+		free_indirdep(indirdep);
 		if (!LIST_EMPTY(&bp->b_dep))
-			panic("indir_trunc: dangling dep");
+			panic("indir_trunc: dangling dep %p",
+			    LIST_FIRST(&bp->b_dep));
 		ump->um_numindirdeps -= 1;
 		FREE_LOCK(&lk);
 	} else {
@@ -2892,11 +5947,10 @@ indir_trunc(freeblks, dbn, level, lbn, countp)
 			brelse(bp);
 #endif
 		FREE_LOCK(&lk);
-		error = bread(freeblks->fb_devvp, dbn, (int)fs->fs_bsize,
-		    NOCRED, &bp);
-		if (error) {
+		if (bread(freeblks->fb_devvp, dbn, (int)fs->fs_bsize,
+		    NOCRED, &bp) != 0) {
 			brelse(bp);
-			return (error);
+			return;
 		}
 	}
 	/*
@@ -2909,57 +5963,264 @@ indir_trunc(freeblks, dbn, level, lbn, countp)
 		ufs1fmt = 0;
 		bap2 = (ufs2_daddr_t *)bp->b_data;
 	}
-	nblocks = btodb(fs->fs_bsize);
-	for (i = NINDIR(fs) - 1; i >= 0; i--) {
-		if (ufs1fmt)
-			nb = bap1[i];
+	/*
+	 * Reclaim indirect blocks which never made it to disk.
+	 */
+	cnt = 0;
+	LIST_FOREACH_SAFE(wk, &wkhd, wk_list, wkn) {
+		struct workhead freewk;
+		if (wk->wk_type != D_JNEWBLK)
+			continue;
+		WORKLIST_REMOVE_UNLOCKED(wk);
+		LIST_INIT(&freewk);
+		WORKLIST_INSERT_UNLOCKED(&freewk, wk);
+		jnewblk = WK_JNEWBLK(wk);
+		if (jnewblk->jn_lbn > 0)
+			i = (jnewblk->jn_lbn - -lbn) / lbnadd;
 		else
+			i = (jnewblk->jn_lbn - (lbn + 1)) / lbnadd;
+		KASSERT(i >= 0 && i < NINDIR(fs),
+		    ("indir_trunc: Index out of range %d parent %jd lbn %jd",
+		    i, lbn, jnewblk->jn_lbn));
+		/* Clear the pointer so it isn't found below. */
+		if (ufs1fmt) {
+			nb = bap1[i];
+			bap1[i] = 0;
+		} else {
 			nb = bap2[i];
+			bap2[i] = 0;
+		}
+		KASSERT(nb == jnewblk->jn_blkno,
+		    ("indir_trunc: Block mismatch %jd != %jd",
+		    nb, jnewblk->jn_blkno));
+		ffs_blkfree(ump, fs, freeblks->fb_devvp, jnewblk->jn_blkno,
+		    fs->fs_bsize, freeblks->fb_previousinum, &freewk);
+		cnt++;
+	}
+	ACQUIRE_LOCK(&lk);
+	if (needj)
+		freework->fw_ref += NINDIR(fs) + 1;
+	/* Any remaining journal work can be completed with freeblks. */
+	jwork_move(&freeblks->fb_jwork, &wkhd);
+	FREE_LOCK(&lk);
+	nblocks = btodb(fs->fs_bsize);
+	if (ufs1fmt)
+		nb = bap1[0];
+	else
+		nb = bap2[0];
+	nfreework = freework;
+	/*
+	 * Reclaim on disk blocks.
+	 */
+	for (i = freework->fw_off; i < NINDIR(fs); i++, nb = nnb) {
+		if (i != NINDIR(fs) - 1) {
+			if (ufs1fmt)
+				nnb = bap1[i+1];
+			else
+				nnb = bap2[i+1];
+		} else
+			nnb = 0;
 		if (nb == 0)
 			continue;
+		cnt++;
 		if (level != 0) {
-			if ((error = indir_trunc(freeblks, fsbtodb(fs, nb),
-			     level - 1, lbn + (i * lbnadd), countp)) != 0)
-				allerror = error;
+			ufs_lbn_t nlbn;
+
+			nlbn = (lbn + 1) - (i * lbnadd);
+			if (needj != 0) {
+				nfreework = newfreework(freeblks, freework,
+				    nlbn, nb, fs->fs_frag, 0);
+				freedeps++;
+			}
+			indir_trunc(nfreework, fsbtodb(fs, nb), nlbn);
+		} else {
+			struct freedep *freedep;
+
+			/*
+			 * Attempt to aggregate freedep dependencies for
+			 * all blocks being released to the same CG.
+			 */
+			LIST_INIT(&wkhd);
+			if (needj != 0 &&
+			    (nnb == 0 || (dtog(fs, nb) != dtog(fs, nnb)))) {
+				freedep = newfreedep(freework);
+				WORKLIST_INSERT_UNLOCKED(&wkhd,
+				    &freedep->fd_list);
+				freedeps++;
+			}
+			ffs_blkfree(ump, fs, freeblks->fb_devvp, nb,
+			    fs->fs_bsize, freeblks->fb_previousinum, &wkhd);
 		}
-		ffs_blkfree(ump, fs, freeblks->fb_devvp, nb, fs->fs_bsize,
-		    freeblks->fb_previousinum);
-		fs_pendingblocks += nblocks;
-		*countp += nblocks;
 	}
-	UFS_LOCK(ump);
-	fs->fs_pendingblocks -= fs_pendingblocks;
-	UFS_UNLOCK(ump);
+	if (level == 0)
+		fs_pendingblocks = (nblocks * cnt);
+	/*
+	 * If we're not journaling we can free the indirect now.  Otherwise
+	 * setup the ref counts and offset so this indirect can be completed
+	 * when its children are free.
+	 */
+	if (needj == 0) {
+		fs_pendingblocks += nblocks;
+		dbn = dbtofsb(fs, dbn);
+		ffs_blkfree(ump, fs, freeblks->fb_devvp, dbn, fs->fs_bsize,
+		    freeblks->fb_previousinum, NULL);
+		ACQUIRE_LOCK(&lk);
+		freeblks->fb_chkcnt -= fs_pendingblocks;
+		if (freework->fw_blkno == dbn)
+			handle_written_freework(freework);
+		FREE_LOCK(&lk);
+		freework = NULL;
+	} else {
+		ACQUIRE_LOCK(&lk);
+		freework->fw_off = i;
+		freework->fw_ref += freedeps;
+		freework->fw_ref -= NINDIR(fs) + 1;
+		if (freework->fw_ref != 0)
+			freework = NULL;
+		freeblks->fb_chkcnt -= fs_pendingblocks;
+		FREE_LOCK(&lk);
+	}
+	if (fs_pendingblocks) {
+		UFS_LOCK(ump);
+		fs->fs_pendingblocks -= fs_pendingblocks;
+		UFS_UNLOCK(ump);
+	}
 	bp->b_flags |= B_INVAL | B_NOCACHE;
 	brelse(bp);
-	return (allerror);
+	if (freework)
+		handle_workitem_indirblk(freework);
+	return;
 }
 
 /*
- * Free an allocindir.
- * This routine must be called with splbio interrupts blocked.
+ * Cancel an allocindir when it is removed via truncation.
  */
 static void
-free_allocindir(aip, inodedep)
+cancel_allocindir(aip, inodedep, freeblks)
 	struct allocindir *aip;
 	struct inodedep *inodedep;
+	struct freeblks *freeblks;
 {
-	struct freefrag *freefrag;
+	struct newblk *newblk;
 
-	mtx_assert(&lk, MA_OWNED);
-	if ((aip->ai_state & DEPCOMPLETE) == 0)
-		LIST_REMOVE(aip, ai_deps);
-	if (aip->ai_state & ONWORKLIST)
-		WORKLIST_REMOVE(&aip->ai_list);
+	/*
+	 * If the journal hasn't been written the jnewblk must be passed
+	 * to the call to ffs_freeblk that reclaims the space.  We accomplish
+	 * this by linking the journal dependency into the indirdep to be
+	 * freed when indir_trunc() is called.  If the journal has already
+	 * been written we can simply reclaim the journal space when the
+	 * freeblks work is complete.
+	 */
 	LIST_REMOVE(aip, ai_next);
-	if ((freefrag = aip->ai_freefrag) != NULL) {
-		if (inodedep == NULL)
-			add_to_worklist(&freefrag->ff_list);
-		else
-			WORKLIST_INSERT(&inodedep->id_bufwait,
-			    &freefrag->ff_list);
+	newblk = (struct newblk *)aip;
+	if (newblk->nb_jnewblk == NULL)
+		cancel_newblk(newblk, &freeblks->fb_jwork);
+	else
+		cancel_newblk(newblk, &aip->ai_indirdep->ir_jwork);
+	if (inodedep && inodedep->id_state & DEPCOMPLETE)
+		WORKLIST_INSERT(&inodedep->id_bufwait, &newblk->nb_list);
+	else
+		free_newblk(newblk);
+}
+
+/*
+ * Create the mkdir dependencies for . and .. in a new directory.  Link them
+ * in to a newdirblk so any subsequent additions are tracked properly.  The
+ * caller is responsible for adding the mkdir1 dependency to the journal
+ * and updating id_mkdiradd.  This function returns with lk held.
+ */
+static struct mkdir *
+setup_newdir(dap, newinum, dinum, newdirbp, mkdirp)
+	struct diradd *dap;
+	ino_t newinum;
+	ino_t dinum;
+	struct buf *newdirbp;
+	struct mkdir **mkdirp;
+{
+	struct newblk *newblk;
+	struct pagedep *pagedep;
+	struct inodedep *inodedep;
+	struct newdirblk *newdirblk = 0;
+	struct mkdir *mkdir1, *mkdir2;
+	struct worklist *wk;
+	struct jaddref *jaddref;
+	struct mount *mp;
+
+	mp = dap->da_list.wk_mp;
+	newdirblk = malloc(sizeof(struct newdirblk), M_NEWDIRBLK,
+	    M_SOFTDEP_FLAGS);
+	workitem_alloc(&newdirblk->db_list, D_NEWDIRBLK, mp);
+	LIST_INIT(&newdirblk->db_mkdir);
+	mkdir1 = malloc(sizeof(struct mkdir), M_MKDIR, M_SOFTDEP_FLAGS);
+	workitem_alloc(&mkdir1->md_list, D_MKDIR, mp);
+	mkdir1->md_state = ATTACHED | MKDIR_BODY;
+	mkdir1->md_diradd = dap;
+	mkdir1->md_jaddref = NULL;
+	mkdir2 = malloc(sizeof(struct mkdir), M_MKDIR, M_SOFTDEP_FLAGS);
+	workitem_alloc(&mkdir2->md_list, D_MKDIR, mp);
+	mkdir2->md_state = ATTACHED | MKDIR_PARENT;
+	mkdir2->md_diradd = dap;
+	mkdir2->md_jaddref = NULL;
+	if ((mp->mnt_kern_flag & MNTK_SUJ) == 0) {
+		mkdir1->md_state |= DEPCOMPLETE;
+		mkdir2->md_state |= DEPCOMPLETE;
 	}
-	WORKITEM_FREE(aip, D_ALLOCINDIR);
+	/*
+	 * Dependency on "." and ".." being written to disk.
+	 */
+	mkdir1->md_buf = newdirbp;
+	ACQUIRE_LOCK(&lk);
+	LIST_INSERT_HEAD(&mkdirlisthd, mkdir1, md_mkdirs);
+	/*
+	 * We must link the pagedep, allocdirect, and newdirblk for
+	 * the initial file page so the pointer to the new directory
+	 * is not written until the directory contents are live and
+	 * any subsequent additions are not marked live until the
+	 * block is reachable via the inode.
+	 */
+	if (pagedep_lookup(mp, newinum, 0, 0, &pagedep) == 0)
+		panic("setup_newdir: lost pagedep");
+	LIST_FOREACH(wk, &newdirbp->b_dep, wk_list)
+		if (wk->wk_type == D_ALLOCDIRECT)
+			break;
+	if (wk == NULL)
+		panic("setup_newdir: lost allocdirect");
+	newblk = WK_NEWBLK(wk);
+	pagedep->pd_state |= NEWBLOCK;
+	pagedep->pd_newdirblk = newdirblk;
+	newdirblk->db_pagedep = pagedep;
+	WORKLIST_INSERT(&newblk->nb_newdirblk, &newdirblk->db_list);
+	WORKLIST_INSERT(&newdirblk->db_mkdir, &mkdir1->md_list);
+	/*
+	 * Look up the inodedep for the parent directory so that we
+	 * can link mkdir2 into the pending dotdot jaddref or
+	 * the inode write if there is none.  If the inode is
+	 * ALLCOMPLETE and no jaddref is present all dependencies have
+	 * been satisfied and mkdir2 can be freed.
+	 */
+	inodedep_lookup(mp, dinum, 0, &inodedep);
+	if (mp->mnt_kern_flag & MNTK_SUJ) {
+		if (inodedep == NULL)
+			panic("setup_newdir: Lost parent.");
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref != NULL && jaddref->ja_parent == newinum &&
+		    (jaddref->ja_state & MKDIR_PARENT),
+		    ("setup_newdir: bad dotdot jaddref %p", jaddref));
+		LIST_INSERT_HEAD(&mkdirlisthd, mkdir2, md_mkdirs);
+		mkdir2->md_jaddref = jaddref;
+		jaddref->ja_mkdir = mkdir2;
+	} else if (inodedep == NULL ||
+	    (inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE) {
+		dap->da_state &= ~MKDIR_PARENT;
+		WORKITEM_FREE(mkdir2, D_MKDIR);
+	} else {
+		LIST_INSERT_HEAD(&mkdirlisthd, mkdir2, md_mkdirs);
+		WORKLIST_INSERT(&inodedep->id_bufwait,&mkdir2->md_list);
+	}
+	*mkdirp = mkdir2;
+
+	return (mkdir1);
 }
 
 /*
@@ -2998,12 +6259,14 @@ softdep_setup_directory_add(bp, dp, diroffset, newinum, newdirbp, isnewblk)
 	ufs_lbn_t lbn;		/* block in directory containing new entry */
 	struct fs *fs;
 	struct diradd *dap;
-	struct allocdirect *adp;
+	struct newblk *newblk;
 	struct pagedep *pagedep;
 	struct inodedep *inodedep;
 	struct newdirblk *newdirblk = 0;
 	struct mkdir *mkdir1, *mkdir2;
+	struct jaddref *jaddref;
 	struct mount *mp;
+	int isindir;
 
 	/*
 	 * Whiteouts have no dependencies.
@@ -3013,6 +6276,8 @@ softdep_setup_directory_add(bp, dp, diroffset, newinum, newdirbp, isnewblk)
 			bdwrite(newdirbp);
 		return (0);
 	}
+	jaddref = NULL;
+	mkdir1 = mkdir2 = NULL;
 	mp = UFSTOVFS(dp->i_ump);
 	fs = dp->i_fs;
 	lbn = lblkno(fs, diroffset);
@@ -3023,111 +6288,123 @@ softdep_setup_directory_add(bp, dp, diroffset, newinum, newdirbp, isnewblk)
 	dap->da_offset = offset;
 	dap->da_newinum = newinum;
 	dap->da_state = ATTACHED;
-	if (isnewblk && lbn < NDADDR && fragoff(fs, diroffset) == 0) {
+	LIST_INIT(&dap->da_jwork);
+	isindir = bp->b_lblkno >= NDADDR;
+	if (isnewblk &&
+	    (isindir ? blkoff(fs, diroffset) : fragoff(fs, diroffset)) == 0) {
 		newdirblk = malloc(sizeof(struct newdirblk),
 		    M_NEWDIRBLK, M_SOFTDEP_FLAGS);
 		workitem_alloc(&newdirblk->db_list, D_NEWDIRBLK, mp);
+		LIST_INIT(&newdirblk->db_mkdir);
 	}
+	/*
+	 * If we're creating a new directory setup the dependencies and set
+	 * the dap state to wait for them.  Otherwise it's COMPLETE and
+	 * we can move on.
+	 */
 	if (newdirbp == NULL) {
 		dap->da_state |= DEPCOMPLETE;
 		ACQUIRE_LOCK(&lk);
 	} else {
 		dap->da_state |= MKDIR_BODY | MKDIR_PARENT;
-		mkdir1 = malloc(sizeof(struct mkdir), M_MKDIR,
-		    M_SOFTDEP_FLAGS);
-		workitem_alloc(&mkdir1->md_list, D_MKDIR, mp);
-		mkdir1->md_state = MKDIR_BODY;
-		mkdir1->md_diradd = dap;
-		mkdir2 = malloc(sizeof(struct mkdir), M_MKDIR,
-		    M_SOFTDEP_FLAGS);
-		workitem_alloc(&mkdir2->md_list, D_MKDIR, mp);
-		mkdir2->md_state = MKDIR_PARENT;
-		mkdir2->md_diradd = dap;
-		/*
-		 * Dependency on "." and ".." being written to disk.
-		 */
-		mkdir1->md_buf = newdirbp;
-		ACQUIRE_LOCK(&lk);
-		LIST_INSERT_HEAD(&mkdirlisthd, mkdir1, md_mkdirs);
-		WORKLIST_INSERT(&newdirbp->b_dep, &mkdir1->md_list);
-		FREE_LOCK(&lk);
-		bdwrite(newdirbp);
-		/*
-		 * Dependency on link count increase for parent directory
-		 */
-		ACQUIRE_LOCK(&lk);
-		if (inodedep_lookup(mp, dp->i_number, 0, &inodedep) == 0
-		    || (inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE) {
-			dap->da_state &= ~MKDIR_PARENT;
-			WORKITEM_FREE(mkdir2, D_MKDIR);
-		} else {
-			LIST_INSERT_HEAD(&mkdirlisthd, mkdir2, md_mkdirs);
-			WORKLIST_INSERT(&inodedep->id_bufwait,&mkdir2->md_list);
-		}
+		mkdir1 = setup_newdir(dap, newinum, dp->i_number, newdirbp,
+		    &mkdir2);
 	}
 	/*
 	 * Link into parent directory pagedep to await its being written.
 	 */
-	if (pagedep_lookup(dp, lbn, DEPALLOC, &pagedep) == 0)
+	if (pagedep_lookup(mp, dp->i_number, lbn, DEPALLOC, &pagedep) == 0)
 		WORKLIST_INSERT(&bp->b_dep, &pagedep->pd_list);
+#ifdef DEBUG
+	if (diradd_lookup(pagedep, offset) != NULL)
+		panic("softdep_setup_directory_add: %p already at off %d\n",
+		    diradd_lookup(pagedep, offset), offset);
+#endif
 	dap->da_pagedep = pagedep;
 	LIST_INSERT_HEAD(&pagedep->pd_diraddhd[DIRADDHASH(offset)], dap,
 	    da_pdlist);
+	inodedep_lookup(mp, newinum, DEPALLOC, &inodedep);
 	/*
-	 * Link into its inodedep. Put it on the id_bufwait list if the inode
-	 * is not yet written. If it is written, do the post-inode write
-	 * processing to put it on the id_pendinghd list.
+	 * If we're journaling, link the diradd into the jaddref so it
+	 * may be completed after the journal entry is written.  Otherwise,
+	 * link the diradd into its inodedep.  If the inode is not yet
+	 * written place it on the bufwait list, otherwise do the post-inode
+	 * write processing to put it on the id_pendinghd list.
 	 */
-	(void) inodedep_lookup(mp, newinum, DEPALLOC, &inodedep);
-	if ((inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE)
+	if (mp->mnt_kern_flag & MNTK_SUJ) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref != NULL && jaddref->ja_parent == dp->i_number,
+		    ("softdep_setup_directory_add: bad jaddref %p", jaddref));
+		jaddref->ja_diroff = diroffset;
+		jaddref->ja_diradd = dap;
+		add_to_journal(&jaddref->ja_list);
+	} else if ((inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE)
 		diradd_inode_written(dap, inodedep);
 	else
 		WORKLIST_INSERT(&inodedep->id_bufwait, &dap->da_list);
-	if (isnewblk) {
+	/*
+	 * Add the journal entries for . and .. links now that the primary
+	 * link is written.
+	 */
+	if (mkdir1 != NULL && mp->mnt_kern_flag & MNTK_SUJ) {
+		jaddref = (struct jaddref *)TAILQ_PREV(&jaddref->ja_ref,
+		    inoreflst, if_deps);
+		KASSERT(jaddref != NULL &&
+		    jaddref->ja_ino == jaddref->ja_parent &&
+		    (jaddref->ja_state & MKDIR_BODY),
+		    ("softdep_setup_directory_add: bad dot jaddref %p",
+		    jaddref));
+		mkdir1->md_jaddref = jaddref;
+		jaddref->ja_mkdir = mkdir1;
 		/*
-		 * Directories growing into indirect blocks are rare
-		 * enough and the frequency of new block allocation
-		 * in those cases even more rare, that we choose not
-		 * to bother tracking them. Rather we simply force the
-		 * new directory entry to disk.
+		 * It is important that the dotdot journal entry
+		 * is added prior to the dot entry since dot writes
+		 * both the dot and dotdot links.  These both must
+		 * be added after the primary link for the journal
+		 * to remain consistent.
 		 */
-		if (lbn >= NDADDR) {
-			FREE_LOCK(&lk);
-			/*
-			 * We only have a new allocation when at the
-			 * beginning of a new block, not when we are
-			 * expanding into an existing block.
-			 */
-			if (blkoff(fs, diroffset) == 0)
-				return (1);
-			return (0);
-		}
+		add_to_journal(&mkdir2->md_jaddref->ja_list);
+		add_to_journal(&jaddref->ja_list);
+	}
+	/*
+	 * If we are adding a new directory remember this diradd so that if
+	 * we rename it we can keep the dot and dotdot dependencies.  If
+	 * we are adding a new name for an inode that has a mkdiradd we
+	 * must be in rename and we have to move the dot and dotdot
+	 * dependencies to this new name.  The old name is being orphaned
+	 * soon.
+	 */
+	if (mkdir1 != NULL) {
+		if (inodedep->id_mkdiradd != NULL)
+			panic("softdep_setup_directory_add: Existing mkdir");
+		inodedep->id_mkdiradd = dap;
+	} else if (inodedep->id_mkdiradd)
+		merge_diradd(inodedep, dap);
+	if (newdirblk) {
 		/*
-		 * We only have a new allocation when at the beginning
-		 * of a new fragment, not when we are expanding into an
-		 * existing fragment. Also, there is nothing to do if we
-		 * are already tracking this block.
+		 * There is nothing to do if we are already tracking
+		 * this block.
 		 */
-		if (fragoff(fs, diroffset) != 0) {
-			FREE_LOCK(&lk);
-			return (0);
-		}
 		if ((pagedep->pd_state & NEWBLOCK) != 0) {
 			WORKITEM_FREE(newdirblk, D_NEWDIRBLK);
 			FREE_LOCK(&lk);
 			return (0);
 		}
-		/*
-		 * Find our associated allocdirect and have it track us.
-		 */
-		if (inodedep_lookup(mp, dp->i_number, 0, &inodedep) == 0)
-			panic("softdep_setup_directory_add: lost inodedep");
-		adp = TAILQ_LAST(&inodedep->id_newinoupdt, allocdirectlst);
-		if (adp == NULL || adp->ad_lbn != lbn)
+		if (newblk_lookup(mp, dbtofsb(fs, bp->b_blkno), 0, &newblk)
+		    == 0)
 			panic("softdep_setup_directory_add: lost entry");
+		WORKLIST_INSERT(&newblk->nb_newdirblk, &newdirblk->db_list);
 		pagedep->pd_state |= NEWBLOCK;
+		pagedep->pd_newdirblk = newdirblk;
 		newdirblk->db_pagedep = pagedep;
-		WORKLIST_INSERT(&adp->ad_newdirblk, &newdirblk->db_list);
+		FREE_LOCK(&lk);
+		/*
+		 * If we extended into an indirect signal direnter to sync.
+		 */
+		if (isindir)
+			return (1);
+		return (0);
 	}
 	FREE_LOCK(&lk);
 	return (0);
@@ -3141,7 +6418,8 @@ softdep_setup_directory_add(bp, dp, diroffset, newinum, newdirbp, isnewblk)
  * occur while the move is in progress.
  */
 void 
-softdep_change_directoryentry_offset(dp, base, oldloc, newloc, entrysize)
+softdep_change_directoryentry_offset(bp, dp, base, oldloc, newloc, entrysize)
+	struct buf *bp;		/* Buffer holding directory block. */
 	struct inode *dp;	/* inode for directory */
 	caddr_t base;		/* address of dp->i_offset */
 	caddr_t oldloc;		/* address of old directory location */
@@ -3150,49 +6428,214 @@ softdep_change_directoryentry_offset(dp, base, oldloc, newloc, entrysize)
 {
 	int offset, oldoffset, newoffset;
 	struct pagedep *pagedep;
+	struct jmvref *jmvref;
 	struct diradd *dap;
+	struct direct *de;
+	struct mount *mp;
 	ufs_lbn_t lbn;
+	int flags;
 
-	ACQUIRE_LOCK(&lk);
+	mp = UFSTOVFS(dp->i_ump);
+	de = (struct direct *)oldloc;
+	jmvref = NULL;
+	flags = 0;
+	/*
+	 * Moves are always journaled as it would be too complex to
+	 * determine if any affected adds or removes are present in the
+	 * journal.
+	 */
+	if (mp->mnt_kern_flag & MNTK_SUJ)  {
+		flags = DEPALLOC;
+		jmvref = newjmvref(dp, de->d_ino,
+		    dp->i_offset + (oldloc - base),
+		    dp->i_offset + (newloc - base));
+	}
 	lbn = lblkno(dp->i_fs, dp->i_offset);
 	offset = blkoff(dp->i_fs, dp->i_offset);
-	if (pagedep_lookup(dp, lbn, 0, &pagedep) == 0)
-		goto done;
 	oldoffset = offset + (oldloc - base);
 	newoffset = offset + (newloc - base);
-
-	LIST_FOREACH(dap, &pagedep->pd_diraddhd[DIRADDHASH(oldoffset)], da_pdlist) {
-		if (dap->da_offset != oldoffset)
-			continue;
-		dap->da_offset = newoffset;
-		if (DIRADDHASH(newoffset) == DIRADDHASH(oldoffset))
-			break;
-		LIST_REMOVE(dap, da_pdlist);
-		LIST_INSERT_HEAD(&pagedep->pd_diraddhd[DIRADDHASH(newoffset)],
-		    dap, da_pdlist);
-		break;
+	ACQUIRE_LOCK(&lk);
+	if (pagedep_lookup(mp, dp->i_number, lbn, flags, &pagedep) == 0) {
+		if (pagedep)
+			WORKLIST_INSERT(&bp->b_dep, &pagedep->pd_list);
+		goto done;
 	}
-	if (dap == NULL) {
-
-		LIST_FOREACH(dap, &pagedep->pd_pendinghd, da_pdlist) {
-			if (dap->da_offset == oldoffset) {
-				dap->da_offset = newoffset;
-				break;
-			}
+	dap = diradd_lookup(pagedep, oldoffset);
+	if (dap) {
+		dap->da_offset = newoffset;
+		newoffset = DIRADDHASH(newoffset);
+		oldoffset = DIRADDHASH(oldoffset);
+		if ((dap->da_state & ALLCOMPLETE) != ALLCOMPLETE &&
+		    newoffset != oldoffset) {
+			LIST_REMOVE(dap, da_pdlist);
+			LIST_INSERT_HEAD(&pagedep->pd_diraddhd[newoffset],
+			    dap, da_pdlist);
 		}
 	}
 done:
+	if (jmvref) {
+		jmvref->jm_pagedep = pagedep;
+		LIST_INSERT_HEAD(&pagedep->pd_jmvrefhd, jmvref, jm_deps);
+		add_to_journal(&jmvref->jm_list);
+	}
 	bcopy(oldloc, newloc, entrysize);
 	FREE_LOCK(&lk);
 }
 
+/*
+ * Move the mkdir dependencies and journal work from one diradd to another
+ * when renaming a directory.  The new name must depend on the mkdir deps
+ * completing as the old name did.  Directories can only have one valid link
+ * at a time so one must be canonical.
+ */
+static void
+merge_diradd(inodedep, newdap)
+	struct inodedep *inodedep;
+	struct diradd *newdap;
+{
+	struct diradd *olddap;
+	struct mkdir *mkdir, *nextmd;
+	short state;
+
+	olddap = inodedep->id_mkdiradd;
+	inodedep->id_mkdiradd = newdap;
+	if ((olddap->da_state & (MKDIR_PARENT | MKDIR_BODY)) != 0) {
+		newdap->da_state &= ~DEPCOMPLETE;
+		for (mkdir = LIST_FIRST(&mkdirlisthd); mkdir; mkdir = nextmd) {
+			nextmd = LIST_NEXT(mkdir, md_mkdirs);
+			if (mkdir->md_diradd != olddap)
+				continue;
+			mkdir->md_diradd = newdap;
+			state = mkdir->md_state & (MKDIR_PARENT | MKDIR_BODY);
+			newdap->da_state |= state;
+			olddap->da_state &= ~state;
+			if ((olddap->da_state &
+			    (MKDIR_PARENT | MKDIR_BODY)) == 0)
+				break;
+		}
+		if ((olddap->da_state & (MKDIR_PARENT | MKDIR_BODY)) != 0)
+			panic("merge_diradd: unfound ref");
+	}
+	/*
+	 * Any mkdir related journal items are not safe to be freed until
+	 * the new name is stable.
+	 */
+	jwork_move(&newdap->da_jwork, &olddap->da_jwork);
+	olddap->da_state |= DEPCOMPLETE;
+	complete_diradd(olddap);
+}
+
+/*
+ * Move the diradd to the pending list when all diradd dependencies are
+ * complete.
+ */
+static void
+complete_diradd(dap)
+	struct diradd *dap;
+{
+	struct pagedep *pagedep;
+
+	if ((dap->da_state & ALLCOMPLETE) == ALLCOMPLETE) {
+		if (dap->da_state & DIRCHG)
+			pagedep = dap->da_previous->dm_pagedep;
+		else
+			pagedep = dap->da_pagedep;
+		LIST_REMOVE(dap, da_pdlist);
+		LIST_INSERT_HEAD(&pagedep->pd_pendinghd, dap, da_pdlist);
+	}
+}
+
+/*
+ * Cancel a diradd when a dirrem overlaps with it.  We must cancel the journal
+ * add entries and conditonally journal the remove.
+ */
+static void
+cancel_diradd(dap, dirrem, jremref, dotremref, dotdotremref)
+	struct diradd *dap;
+	struct dirrem *dirrem;
+	struct jremref *jremref;
+	struct jremref *dotremref;
+	struct jremref *dotdotremref;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct inoref *inoref;
+	struct mkdir *mkdir;
+
+	/*
+	 * If no remove references were allocated we're on a non-journaled
+	 * filesystem and can skip the cancel step.
+	 */
+	if (jremref == NULL) {
+		free_diradd(dap, NULL);
+		return;
+	}
+	/*
+	 * Cancel the primary name an free it if it does not require
+	 * journaling.
+	 */
+	if (inodedep_lookup(dap->da_list.wk_mp, dap->da_newinum,
+	    0, &inodedep) != 0) {
+		/* Abort the addref that reference this diradd.  */
+		TAILQ_FOREACH(inoref, &inodedep->id_inoreflst, if_deps) {
+			if (inoref->if_list.wk_type != D_JADDREF)
+				continue;
+			jaddref = (struct jaddref *)inoref;
+			if (jaddref->ja_diradd != dap)
+				continue;
+			if (cancel_jaddref(jaddref, inodedep,
+			    &dirrem->dm_jwork) == 0) {
+				free_jremref(jremref);
+				jremref = NULL;
+			}
+			break;
+		}
+	}
+	/*
+	 * Cancel subordinate names and free them if they do not require
+	 * journaling.
+	 */
+	if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) != 0) {
+		LIST_FOREACH(mkdir, &mkdirlisthd, md_mkdirs) {
+			if (mkdir->md_diradd != dap)
+				continue;
+			if ((jaddref = mkdir->md_jaddref) == NULL)
+				continue;
+			mkdir->md_jaddref = NULL;
+			if (mkdir->md_state & MKDIR_PARENT) {
+				if (cancel_jaddref(jaddref, NULL,
+				    &dirrem->dm_jwork) == 0) {
+					free_jremref(dotdotremref);
+					dotdotremref = NULL;
+				}
+			} else {
+				if (cancel_jaddref(jaddref, inodedep,
+				    &dirrem->dm_jwork) == 0) {
+					free_jremref(dotremref);
+					dotremref = NULL;
+				}
+			}
+		}
+	}
+
+	if (jremref)
+		journal_jremref(dirrem, jremref, inodedep);
+	if (dotremref)
+		journal_jremref(dirrem, dotremref, inodedep);
+	if (dotdotremref)
+		journal_jremref(dirrem, dotdotremref, NULL);
+	jwork_move(&dirrem->dm_jwork, &dap->da_jwork);
+	free_diradd(dap, &dirrem->dm_jwork);
+}
+
 /*
  * Free a diradd dependency structure. This routine must be called
  * with splbio interrupts blocked.
  */
 static void
-free_diradd(dap)
+free_diradd(dap, wkhd)
 	struct diradd *dap;
+	struct workhead *wkhd;
 {
 	struct dirrem *dirrem;
 	struct pagedep *pagedep;
@@ -3200,32 +6643,48 @@ free_diradd(dap)
 	struct mkdir *mkdir, *nextmd;
 
 	mtx_assert(&lk, MA_OWNED);
-	WORKLIST_REMOVE(&dap->da_list);
 	LIST_REMOVE(dap, da_pdlist);
+	if (dap->da_state & ONWORKLIST)
+		WORKLIST_REMOVE(&dap->da_list);
 	if ((dap->da_state & DIRCHG) == 0) {
 		pagedep = dap->da_pagedep;
 	} else {
 		dirrem = dap->da_previous;
 		pagedep = dirrem->dm_pagedep;
 		dirrem->dm_dirinum = pagedep->pd_ino;
-		add_to_worklist(&dirrem->dm_list);
+		dirrem->dm_state |= COMPLETE;
+		if (LIST_EMPTY(&dirrem->dm_jremrefhd))
+			add_to_worklist(&dirrem->dm_list, 0);
 	}
 	if (inodedep_lookup(pagedep->pd_list.wk_mp, dap->da_newinum,
 	    0, &inodedep) != 0)
-		(void) free_inodedep(inodedep);
+		if (inodedep->id_mkdiradd == dap)
+			inodedep->id_mkdiradd = NULL;
 	if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) != 0) {
 		for (mkdir = LIST_FIRST(&mkdirlisthd); mkdir; mkdir = nextmd) {
 			nextmd = LIST_NEXT(mkdir, md_mkdirs);
 			if (mkdir->md_diradd != dap)
 				continue;
-			dap->da_state &= ~mkdir->md_state;
-			WORKLIST_REMOVE(&mkdir->md_list);
+			dap->da_state &=
+			    ~(mkdir->md_state & (MKDIR_PARENT | MKDIR_BODY));
 			LIST_REMOVE(mkdir, md_mkdirs);
+			if (mkdir->md_state & ONWORKLIST)
+				WORKLIST_REMOVE(&mkdir->md_list);
+			if (mkdir->md_jaddref != NULL)
+				panic("free_diradd: Unexpected jaddref");
 			WORKITEM_FREE(mkdir, D_MKDIR);
+			if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) == 0)
+				break;
 		}
 		if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) != 0)
 			panic("free_diradd: unfound ref");
 	}
+	if (inodedep)
+		free_inodedep(inodedep);
+	/*
+	 * Free any journal segments waiting for the directory write.
+	 */
+	handle_jwork(&dap->da_jwork);
 	WORKITEM_FREE(dap, D_DIRADD);
 }
 
@@ -3254,11 +6713,24 @@ softdep_setup_remove(bp, dp, ip, isrmdir)
 	int isrmdir;		/* indicates if doing RMDIR */
 {
 	struct dirrem *dirrem, *prevdirrem;
+	struct inodedep *inodedep;
+	int direct;
 
 	/*
-	 * Allocate a new dirrem if appropriate and ACQUIRE_LOCK.
+	 * Allocate a new dirrem if appropriate and ACQUIRE_LOCK.  We want
+	 * newdirrem() to setup the full directory remove which requires
+	 * isrmdir > 1.
 	 */
-	dirrem = newdirrem(bp, dp, ip, isrmdir, &prevdirrem);
+	dirrem = newdirrem(bp, dp, ip, isrmdir?2:0, &prevdirrem);
+	/*
+	 * Add the dirrem to the inodedep's pending remove list for quick
+	 * discovery later.
+	 */
+	if (inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, 0,
+	    &inodedep) == 0)
+		panic("softdep_setup_remove: Lost inodedep.");
+	dirrem->dm_state |= ONDEPLIST;
+	LIST_INSERT_HEAD(&inodedep->id_dirremhd, dirrem, dm_inonext);
 
 	/*
 	 * If the COMPLETE flag is clear, then there were no active
@@ -3280,11 +6752,148 @@ softdep_setup_remove(bp, dp, ip, isrmdir)
 			LIST_INSERT_HEAD(&dirrem->dm_pagedep->pd_dirremhd,
 			    prevdirrem, dm_next);
 		dirrem->dm_dirinum = dirrem->dm_pagedep->pd_ino;
+		direct = LIST_EMPTY(&dirrem->dm_jremrefhd);
 		FREE_LOCK(&lk);
-		handle_workitem_remove(dirrem, NULL);
+		if (direct)
+			handle_workitem_remove(dirrem, NULL);
 	}
 }
 
+/*
+ * Check for an entry matching 'offset' on both the pd_dirraddhd list and the
+ * pd_pendinghd list of a pagedep.
+ */
+static struct diradd *
+diradd_lookup(pagedep, offset)
+	struct pagedep *pagedep;
+	int offset;
+{
+	struct diradd *dap;
+
+	LIST_FOREACH(dap, &pagedep->pd_diraddhd[DIRADDHASH(offset)], da_pdlist)
+		if (dap->da_offset == offset)
+			return (dap);
+	LIST_FOREACH(dap, &pagedep->pd_pendinghd, da_pdlist)
+		if (dap->da_offset == offset)
+			return (dap);
+	return (NULL);
+}
+
+/*
+ * Search for a .. diradd dependency in a directory that is being removed.
+ * If the directory was renamed to a new parent we have a diradd rather
+ * than a mkdir for the .. entry.  We need to cancel it now before
+ * it is found in truncate().
+ */
+static struct jremref *
+cancel_diradd_dotdot(ip, dirrem, jremref)
+	struct inode *ip;
+	struct dirrem *dirrem;
+	struct jremref *jremref;
+{
+	struct pagedep *pagedep;
+	struct diradd *dap;
+	struct worklist *wk;
+
+	if (pagedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, 0, 0,
+	    &pagedep) == 0)
+		return (jremref);
+	dap = diradd_lookup(pagedep, DOTDOT_OFFSET);
+	if (dap == NULL)
+		return (jremref);
+	cancel_diradd(dap, dirrem, jremref, NULL, NULL);
+	/*
+	 * Mark any journal work as belonging to the parent so it is freed
+	 * with the .. reference.
+	 */
+	LIST_FOREACH(wk, &dirrem->dm_jwork, wk_list)
+		wk->wk_state |= MKDIR_PARENT;
+	return (NULL);
+}
+
+/*
+ * Cancel the MKDIR_PARENT mkdir component of a diradd when we're going to
+ * replace it with a dirrem/diradd pair as a result of re-parenting a
+ * directory.  This ensures that we don't simultaneously have a mkdir and
+ * a diradd for the same .. entry.
+ */
+static struct jremref *
+cancel_mkdir_dotdot(ip, dirrem, jremref)
+	struct inode *ip;
+	struct dirrem *dirrem;
+	struct jremref *jremref;
+{
+	struct inodedep *inodedep;
+	struct jaddref *jaddref;
+	struct mkdir *mkdir;
+	struct diradd *dap;
+
+	if (inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, 0,
+	    &inodedep) == 0)
+		panic("cancel_mkdir_dotdot: Lost inodedep");
+	dap = inodedep->id_mkdiradd;
+	if (dap == NULL || (dap->da_state & MKDIR_PARENT) == 0)
+		return (jremref);
+	for (mkdir = LIST_FIRST(&mkdirlisthd); mkdir;
+	    mkdir = LIST_NEXT(mkdir, md_mkdirs))
+		if (mkdir->md_diradd == dap && mkdir->md_state & MKDIR_PARENT)
+			break;
+	if (mkdir == NULL)
+		panic("cancel_mkdir_dotdot: Unable to find mkdir\n");
+	if ((jaddref = mkdir->md_jaddref) != NULL) {
+		mkdir->md_jaddref = NULL;
+		jaddref->ja_state &= ~MKDIR_PARENT;
+		if (inodedep_lookup(UFSTOVFS(ip->i_ump), jaddref->ja_ino, 0,
+		    &inodedep) == 0)
+			panic("cancel_mkdir_dotdot: Lost parent inodedep");
+		if (cancel_jaddref(jaddref, inodedep, &dirrem->dm_jwork)) {
+			journal_jremref(dirrem, jremref, inodedep);
+			jremref = NULL;
+		}
+	}
+	if (mkdir->md_state & ONWORKLIST)
+		WORKLIST_REMOVE(&mkdir->md_list);
+	mkdir->md_state |= ALLCOMPLETE;
+	complete_mkdir(mkdir);
+	return (jremref);
+}
+
+static void
+journal_jremref(dirrem, jremref, inodedep)
+	struct dirrem *dirrem;
+	struct jremref *jremref;
+	struct inodedep *inodedep;
+{
+
+	if (inodedep == NULL)
+		if (inodedep_lookup(jremref->jr_list.wk_mp,
+		    jremref->jr_ref.if_ino, 0, &inodedep) == 0)
+			panic("journal_jremref: Lost inodedep");
+	LIST_INSERT_HEAD(&dirrem->dm_jremrefhd, jremref, jr_deps);
+	TAILQ_INSERT_TAIL(&inodedep->id_inoreflst, &jremref->jr_ref, if_deps);
+	add_to_journal(&jremref->jr_list);
+}
+
+static void
+dirrem_journal(dirrem, jremref, dotremref, dotdotremref)
+	struct dirrem *dirrem;
+	struct jremref *jremref;
+	struct jremref *dotremref;
+	struct jremref *dotdotremref;
+{
+	struct inodedep *inodedep;
+
+
+	if (inodedep_lookup(jremref->jr_list.wk_mp, jremref->jr_ref.if_ino, 0,
+	    &inodedep) == 0)
+		panic("dirrem_journal: Lost inodedep");
+	journal_jremref(dirrem, jremref, inodedep);
+	if (dotremref)
+		journal_jremref(dirrem, dotremref, inodedep);
+	if (dotdotremref)
+		journal_jremref(dirrem, dotdotremref, NULL);
+}
+
 /*
  * Allocate a new dirrem if appropriate and return it along with
  * its associated pagedep. Called without a lock, returns with lock.
@@ -3303,12 +6912,17 @@ newdirrem(bp, dp, ip, isrmdir, prevdirremp)
 	struct diradd *dap;
 	struct dirrem *dirrem;
 	struct pagedep *pagedep;
+	struct jremref *jremref;
+	struct jremref *dotremref;
+	struct jremref *dotdotremref;
+	struct vnode *dvp;
 
 	/*
 	 * Whiteouts have no deletion dependencies.
 	 */
 	if (ip == NULL)
 		panic("newdirrem: whiteout");
+	dvp = ITOV(dp);
 	/*
 	 * If we are over our limit, try to improve the situation.
 	 * Limiting the number of dirrem structures will also limit
@@ -3321,34 +6935,75 @@ newdirrem(bp, dp, ip, isrmdir, prevdirremp)
 	FREE_LOCK(&lk);
 	dirrem = malloc(sizeof(struct dirrem),
 		M_DIRREM, M_SOFTDEP_FLAGS|M_ZERO);
-	workitem_alloc(&dirrem->dm_list, D_DIRREM, ITOV(dp)->v_mount);
+	workitem_alloc(&dirrem->dm_list, D_DIRREM, dvp->v_mount);
+	LIST_INIT(&dirrem->dm_jremrefhd);
+	LIST_INIT(&dirrem->dm_jwork);
 	dirrem->dm_state = isrmdir ? RMDIR : 0;
 	dirrem->dm_oldinum = ip->i_number;
 	*prevdirremp = NULL;
-
+	/*
+	 * Allocate remove reference structures to track journal write
+	 * dependencies.  We will always have one for the link and
+	 * when doing directories we will always have one more for dot.
+	 * When renaming a directory we skip the dotdot link change so
+	 * this is not needed.
+	 */
+	jremref = dotremref = dotdotremref = NULL;
+	if (DOINGSUJ(dvp)) {
+		if (isrmdir) {
+			jremref = newjremref(dirrem, dp, ip, dp->i_offset,
+			    ip->i_effnlink + 2);
+			dotremref = newjremref(dirrem, ip, ip, DOT_OFFSET,
+			    ip->i_effnlink + 1);
+		} else
+			jremref = newjremref(dirrem, dp, ip, dp->i_offset,
+			    ip->i_effnlink + 1);
+		if (isrmdir > 1) {
+			dotdotremref = newjremref(dirrem, ip, dp, DOTDOT_OFFSET,
+			    dp->i_effnlink + 1);
+			dotdotremref->jr_state |= MKDIR_PARENT;
+		}
+	}
 	ACQUIRE_LOCK(&lk);
 	lbn = lblkno(dp->i_fs, dp->i_offset);
 	offset = blkoff(dp->i_fs, dp->i_offset);
-	if (pagedep_lookup(dp, lbn, DEPALLOC, &pagedep) == 0)
+	if (pagedep_lookup(UFSTOVFS(dp->i_ump), dp->i_number, lbn, DEPALLOC,
+	    &pagedep) == 0)
 		WORKLIST_INSERT(&bp->b_dep, &pagedep->pd_list);
 	dirrem->dm_pagedep = pagedep;
+	/*
+	 * If we're renaming a .. link to a new directory, cancel any
+	 * existing MKDIR_PARENT mkdir.  If it has already been canceled
+	 * the jremref is preserved for any potential diradd in this
+	 * location.  This can not coincide with a rmdir.
+	 */
+	if (dp->i_offset == DOTDOT_OFFSET) {
+		if (isrmdir)
+			panic("newdirrem: .. directory change during remove?");
+		jremref = cancel_mkdir_dotdot(dp, dirrem, jremref);
+	}
+	/*
+	 * If we're removing a directory search for the .. dependency now and
+	 * cancel it.  Any pending journal work will be added to the dirrem
+	 * to be completed when the workitem remove completes.
+	 */
+	if (isrmdir > 1)
+		dotdotremref = cancel_diradd_dotdot(ip, dirrem, dotdotremref);
 	/*
 	 * Check for a diradd dependency for the same directory entry.
 	 * If present, then both dependencies become obsolete and can
-	 * be de-allocated. Check for an entry on both the pd_dirraddhd
-	 * list and the pd_pendinghd list.
+	 * be de-allocated.
 	 */
-
-	LIST_FOREACH(dap, &pagedep->pd_diraddhd[DIRADDHASH(offset)], da_pdlist)
-		if (dap->da_offset == offset)
-			break;
+	dap = diradd_lookup(pagedep, offset);
 	if (dap == NULL) {
-
-		LIST_FOREACH(dap, &pagedep->pd_pendinghd, da_pdlist)
-			if (dap->da_offset == offset)
-				break;
-		if (dap == NULL)
-			return (dirrem);
+		/*
+		 * Link the jremref structures into the dirrem so they are
+		 * written prior to the pagedep.
+		 */
+		if (jremref)
+			dirrem_journal(dirrem, jremref, dotremref,
+			    dotdotremref);
+		return (dirrem);
 	}
 	/*
 	 * Must be ATTACHED at this point.
@@ -3373,7 +7028,17 @@ newdirrem(bp, dp, ip, isrmdir, prevdirremp)
 	 * Mark it COMPLETE so we can delete its inode immediately.
 	 */
 	dirrem->dm_state |= COMPLETE;
-	free_diradd(dap);
+	cancel_diradd(dap, dirrem, jremref, dotremref, dotdotremref);
+#ifdef SUJ_DEBUG
+	if (isrmdir == 0) {
+		struct worklist *wk;
+
+		LIST_FOREACH(wk, &dirrem->dm_jwork, wk_list)
+			if (wk->wk_state & (MKDIR_BODY | MKDIR_PARENT))
+				panic("bad wk %p (0x%X)\n", wk, wk->wk_state);
+	}
+#endif
+
 	return (dirrem);
 }
 
@@ -3407,6 +7072,7 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 	struct dirrem *dirrem, *prevdirrem;
 	struct pagedep *pagedep;
 	struct inodedep *inodedep;
+	struct jaddref *jaddref;
 	struct mount *mp;
 
 	offset = blkoff(dp->i_fs, dp->i_offset);
@@ -3422,6 +7088,7 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 		dap->da_state = DIRCHG | ATTACHED | DEPCOMPLETE;
 		dap->da_offset = offset;
 		dap->da_newinum = newinum;
+		LIST_INIT(&dap->da_jwork);
 	}
 
 	/*
@@ -3454,11 +7121,21 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 			    dm_next);
 		} else {
 			dirrem->dm_dirinum = pagedep->pd_ino;
-			add_to_worklist(&dirrem->dm_list);
+			if (LIST_EMPTY(&dirrem->dm_jremrefhd))
+				add_to_worklist(&dirrem->dm_list, 0);
 		}
 		FREE_LOCK(&lk);
 		return;
 	}
+	/*
+	 * Add the dirrem to the inodedep's pending remove list for quick
+	 * discovery later.  A valid nlinkdelta ensures that this lookup
+	 * will not fail.
+	 */
+	if (inodedep_lookup(mp, ip->i_number, 0, &inodedep) == 0)
+		panic("softdep_setup_directory_change: Lost inodedep.");
+	dirrem->dm_state |= ONDEPLIST;
+	LIST_INSERT_HEAD(&inodedep->id_dirremhd, dirrem, dm_inonext);
 
 	/*
 	 * If the COMPLETE flag is clear, then there were no active
@@ -3483,15 +7160,29 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 			dap->da_pagedep = pagedep;
 		}
 		dirrem->dm_dirinum = pagedep->pd_ino;
-		add_to_worklist(&dirrem->dm_list);
+		if (LIST_EMPTY(&dirrem->dm_jremrefhd))
+			add_to_worklist(&dirrem->dm_list, 0);
 	}
 	/*
-	 * Link into its inodedep. Put it on the id_bufwait list if the inode
+	 * Lookup the jaddref for this journal entry.  We must finish
+	 * initializing it and make the diradd write dependent on it.
+	 * If we're not journaling Put it on the id_bufwait list if the inode
 	 * is not yet written. If it is written, do the post-inode write
 	 * processing to put it on the id_pendinghd list.
 	 */
-	if (inodedep_lookup(mp, newinum, DEPALLOC, &inodedep) == 0 ||
-	    (inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE) {
+	inodedep_lookup(mp, newinum, DEPALLOC, &inodedep);
+	if (mp->mnt_kern_flag & MNTK_SUJ) {
+		jaddref = (struct jaddref *)TAILQ_LAST(&inodedep->id_inoreflst,
+		    inoreflst);
+		KASSERT(jaddref != NULL && jaddref->ja_parent == dp->i_number,
+		    ("softdep_setup_directory_change: bad jaddref %p",
+		    jaddref));
+		jaddref->ja_diroff = dp->i_offset;
+		jaddref->ja_diradd = dap;
+		LIST_INSERT_HEAD(&pagedep->pd_diraddhd[DIRADDHASH(offset)],
+		    dap, da_pdlist);
+		add_to_journal(&jaddref->ja_list);
+	} else if ((inodedep->id_state & ALLCOMPLETE) == ALLCOMPLETE) {
 		dap->da_state |= COMPLETE;
 		LIST_INSERT_HEAD(&pagedep->pd_pendinghd, dap, da_pdlist);
 		WORKLIST_INSERT(&inodedep->id_pendinghd, &dap->da_list);
@@ -3500,6 +7191,13 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 		    dap, da_pdlist);
 		WORKLIST_INSERT(&inodedep->id_bufwait, &dap->da_list);
 	}
+	/*
+	 * If we're making a new name for a directory that has not been
+	 * committed when need to move the dot and dotdot references to
+	 * this new name.
+	 */
+	if (inodedep->id_mkdiradd && dp->i_offset != DOTDOT_OFFSET)
+		merge_diradd(inodedep, dap);
 	FREE_LOCK(&lk);
 }
 
@@ -3516,8 +7214,7 @@ softdep_change_linkcnt(ip)
 	struct inodedep *inodedep;
 
 	ACQUIRE_LOCK(&lk);
-	(void) inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number,
-	    DEPALLOC, &inodedep);
+	inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, DEPALLOC, &inodedep);
 	if (ip->i_nlink < ip->i_effnlink)
 		panic("softdep_change_linkcnt: bad delta");
 	inodedep->id_nlinkdelta = ip->i_nlink - ip->i_effnlink;
@@ -3573,6 +7270,305 @@ softdep_releasefile(ip)
 	ip->i_flag |= IN_SPACECOUNTED;
 }
 
+/*
+ * Attach a sbdep dependency to the superblock buf so that we can keep
+ * track of the head of the linked list of referenced but unlinked inodes.
+ */
+void
+softdep_setup_sbupdate(ump, fs, bp)
+	struct ufsmount *ump;
+	struct fs *fs;
+	struct buf *bp;
+{
+	struct sbdep *sbdep;
+	struct worklist *wk;
+
+	if ((fs->fs_flags & FS_SUJ) == 0)
+		return;
+	LIST_FOREACH(wk, &bp->b_dep, wk_list)
+		if (wk->wk_type == D_SBDEP)
+			break;
+	if (wk != NULL)
+		return;
+	sbdep = malloc(sizeof(struct sbdep), M_SBDEP, M_SOFTDEP_FLAGS);
+	workitem_alloc(&sbdep->sb_list, D_SBDEP, UFSTOVFS(ump));
+	sbdep->sb_fs = fs;
+	sbdep->sb_ump = ump;
+	ACQUIRE_LOCK(&lk);
+	WORKLIST_INSERT(&bp->b_dep, &sbdep->sb_list);
+	FREE_LOCK(&lk);
+}
+
+/*
+ * Return the first unlinked inodedep which is ready to be the head of the
+ * list.  The inodedep and all those after it must have valid next pointers.
+ */
+static struct inodedep *
+first_unlinked_inodedep(ump)
+	struct ufsmount *ump;
+{
+	struct inodedep *inodedep;
+	struct inodedep *idp;
+
+	for (inodedep = TAILQ_LAST(&ump->softdep_unlinked, inodedeplst);
+	    inodedep; inodedep = idp) {
+		if ((inodedep->id_state & UNLINKNEXT) == 0)
+			return (NULL);
+		idp = TAILQ_PREV(inodedep, inodedeplst, id_unlinked);
+		if (idp == NULL || (idp->id_state & UNLINKNEXT) == 0)
+			break;
+		if ((inodedep->id_state & UNLINKPREV) == 0)
+			panic("first_unlinked_inodedep: prev != next");
+	}
+	if (inodedep == NULL)
+		return (NULL);
+
+	return (inodedep);
+}
+
+/*
+ * Set the sujfree unlinked head pointer prior to writing a superblock.
+ */
+static void
+initiate_write_sbdep(sbdep)
+	struct sbdep *sbdep;
+{
+	struct inodedep *inodedep;
+	struct fs *bpfs;
+	struct fs *fs;
+
+	bpfs = sbdep->sb_fs;
+	fs = sbdep->sb_ump->um_fs;
+	inodedep = first_unlinked_inodedep(sbdep->sb_ump);
+	if (inodedep) {
+		fs->fs_sujfree = inodedep->id_ino;
+		inodedep->id_state |= UNLINKPREV;
+	} else
+		fs->fs_sujfree = 0;
+	bpfs->fs_sujfree = fs->fs_sujfree;
+}
+
+/*
+ * After a superblock is written determine whether it must be written again
+ * due to a changing unlinked list head.
+ */
+static int
+handle_written_sbdep(sbdep, bp)
+	struct sbdep *sbdep;
+	struct buf *bp;
+{
+	struct inodedep *inodedep;
+	struct mount *mp;
+	struct fs *fs;
+
+	fs = sbdep->sb_fs;
+	mp = UFSTOVFS(sbdep->sb_ump);
+	inodedep = first_unlinked_inodedep(sbdep->sb_ump);
+	if ((inodedep && fs->fs_sujfree != inodedep->id_ino) ||
+	    (inodedep == NULL && fs->fs_sujfree != 0)) {
+		bdirty(bp);
+		return (1);
+	}
+	WORKITEM_FREE(sbdep, D_SBDEP);
+	if (fs->fs_sujfree == 0)
+		return (0);
+	if (inodedep_lookup(mp, fs->fs_sujfree, 0, &inodedep) == 0)
+		panic("handle_written_sbdep: lost inodedep");
+	/*
+	 * Now that we have a record of this indode in stable store allow it
+	 * to be written to free up pending work.  Inodes may see a lot of
+	 * write activity after they are unlinked which we must not hold up.
+	 */
+	for (; inodedep != NULL; inodedep = TAILQ_NEXT(inodedep, id_unlinked)) {
+		if ((inodedep->id_state & UNLINKLINKS) != UNLINKLINKS)
+			panic("handle_written_sbdep: Bad inodedep %p (0x%X)",
+			    inodedep, inodedep->id_state);
+		if (inodedep->id_state & UNLINKONLIST)
+			break;
+		inodedep->id_state |= DEPCOMPLETE | UNLINKONLIST;
+	}
+
+	return (0);
+}
+
+/*
+ * Mark an inodedep has unlinked and insert it into the in-memory unlinked
+ * list.
+ */
+static void
+unlinked_inodedep(mp, inodedep)
+	struct mount *mp;
+	struct inodedep *inodedep;
+{
+	struct ufsmount *ump;
+
+	if ((mp->mnt_kern_flag & MNTK_SUJ) == 0)
+		return;
+	ump = VFSTOUFS(mp);
+	ump->um_fs->fs_fmod = 1;
+	inodedep->id_state |= UNLINKED;
+	TAILQ_INSERT_HEAD(&ump->softdep_unlinked, inodedep, id_unlinked);
+}
+
+/*
+ * Remove an inodedep from the unlinked inodedep list.  This may require
+ * disk writes if the inode has made it that far.
+ */
+static void
+clear_unlinked_inodedep(inodedep)
+	struct inodedep *inodedep;
+{
+	struct ufsmount *ump;
+	struct inodedep *idp;
+	struct inodedep *idn;
+	struct fs *fs;
+	struct buf *bp;
+	ino_t ino;
+	ino_t nino;
+	ino_t pino;
+	int error;
+
+	ump = VFSTOUFS(inodedep->id_list.wk_mp);
+	fs = ump->um_fs;
+	ino = inodedep->id_ino;
+	error = 0;
+	for (;;) {
+		/*
+		 * If nothing has yet been written simply remove us from
+		 * the in memory list and return.  This is the most common
+		 * case where handle_workitem_remove() loses the final
+		 * reference.
+		 */
+		if ((inodedep->id_state & UNLINKLINKS) == 0)
+			break;
+		/*
+		 * If we have a NEXT pointer and no PREV pointer we can simply
+		 * clear NEXT's PREV and remove ourselves from the list.  Be
+		 * careful not to clear PREV if the superblock points at
+		 * next as well.
+		 */
+		idn = TAILQ_NEXT(inodedep, id_unlinked);
+		if ((inodedep->id_state & UNLINKLINKS) == UNLINKNEXT) {
+			if (idn && fs->fs_sujfree != idn->id_ino)
+				idn->id_state &= ~UNLINKPREV;
+			break;
+		}
+		/*
+		 * Here we have an inodedep which is actually linked into
+		 * the list.  We must remove it by forcing a write to the
+		 * link before us, whether it be the superblock or an inode.
+		 * Unfortunately the list may change while we're waiting
+		 * on the buf lock for either resource so we must loop until
+		 * we lock. the right one.  If both the superblock and an
+		 * inode point to this inode we must clear the inode first
+		 * followed by the superblock.
+		 */
+		idp = TAILQ_PREV(inodedep, inodedeplst, id_unlinked);
+		pino = 0;
+		if (idp && (idp->id_state & UNLINKNEXT))
+			pino = idp->id_ino;
+		FREE_LOCK(&lk);
+		if (pino == 0)
+			bp = getblk(ump->um_devvp, btodb(fs->fs_sblockloc),
+			    (int)fs->fs_sbsize, 0, 0, 0);
+		else
+			error = bread(ump->um_devvp,
+			    fsbtodb(fs, ino_to_fsba(fs, pino)),
+			    (int)fs->fs_bsize, NOCRED, &bp);
+		ACQUIRE_LOCK(&lk);
+		if (error)
+			break;
+		/* If the list has changed restart the loop. */
+		idp = TAILQ_PREV(inodedep, inodedeplst, id_unlinked);
+		nino = 0;
+		if (idp && (idp->id_state & UNLINKNEXT))
+			nino = idp->id_ino;
+		if (nino != pino ||
+		    (inodedep->id_state & UNLINKPREV) != UNLINKPREV) {
+			FREE_LOCK(&lk);
+			brelse(bp);
+			ACQUIRE_LOCK(&lk);
+			continue;
+		}
+		/*
+		 * Remove us from the in memory list.  After this we cannot
+		 * access the inodedep.
+		 */
+		idn = TAILQ_NEXT(inodedep, id_unlinked);
+		inodedep->id_state &= ~(UNLINKED | UNLINKLINKS);
+		TAILQ_REMOVE(&ump->softdep_unlinked, inodedep, id_unlinked);
+		/*
+		 * Determine the next inode number.
+		 */
+		nino = 0;
+		if (idn) {
+			/*
+			 * If next isn't on the list we can just clear prev's
+			 * state and schedule it to be fixed later.  No need
+			 * to synchronously write if we're not in the real
+			 * list.
+			 */
+			if ((idn->id_state & UNLINKPREV) == 0 && pino != 0) {
+				idp->id_state &= ~UNLINKNEXT;
+				if ((idp->id_state & ONWORKLIST) == 0)
+					WORKLIST_INSERT(&bp->b_dep,
+					    &idp->id_list);
+				FREE_LOCK(&lk);
+				bawrite(bp);
+				ACQUIRE_LOCK(&lk);
+				return;
+			}
+			nino = idn->id_ino;
+		}
+		FREE_LOCK(&lk);
+		/*
+		 * The predecessor's next pointer is manually updated here
+		 * so that the NEXT flag is never cleared for an element
+		 * that is in the list.
+		 */
+		if (pino == 0) {
+			bcopy((caddr_t)fs, bp->b_data, (u_int)fs->fs_sbsize);
+			ffs_oldfscompat_write((struct fs *)bp->b_data, ump);
+			softdep_setup_sbupdate(ump, (struct fs *)bp->b_data,
+			    bp);
+		} else if (fs->fs_magic == FS_UFS1_MAGIC)
+			((struct ufs1_dinode *)bp->b_data +
+			    ino_to_fsbo(fs, pino))->di_freelink = nino;
+		else
+			((struct ufs2_dinode *)bp->b_data +
+			    ino_to_fsbo(fs, pino))->di_freelink = nino;
+		/*
+		 * If the bwrite fails we have no recourse to recover.  The
+		 * filesystem is corrupted already.
+		 */
+		bwrite(bp);
+		ACQUIRE_LOCK(&lk);
+		/*
+		 * If the superblock pointer still needs to be cleared force
+		 * a write here.
+		 */
+		if (fs->fs_sujfree == ino) {
+			FREE_LOCK(&lk);
+			bp = getblk(ump->um_devvp, btodb(fs->fs_sblockloc),
+			    (int)fs->fs_sbsize, 0, 0, 0);
+			bcopy((caddr_t)fs, bp->b_data, (u_int)fs->fs_sbsize);
+			ffs_oldfscompat_write((struct fs *)bp->b_data, ump);
+			softdep_setup_sbupdate(ump, (struct fs *)bp->b_data,
+			    bp);
+			bwrite(bp);
+			ACQUIRE_LOCK(&lk);
+		}
+		if (fs->fs_sujfree != ino)
+			return;
+		panic("clear_unlinked_inodedep: Failed to clear free head");
+	}
+	if (inodedep->id_ino == fs->fs_sujfree)
+		panic("clear_unlinked_inodedep: Freeing head of free list");
+	inodedep->id_state &= ~(UNLINKED | UNLINKLINKS);
+	TAILQ_REMOVE(&ump->softdep_unlinked, inodedep, id_unlinked);
+	return;
+}
+
 /*
  * This workitem decrements the inode's link count.
  * If the link count reaches zero, the file is removed.
@@ -3584,22 +7580,54 @@ handle_workitem_remove(dirrem, xp)
 {
 	struct thread *td = curthread;
 	struct inodedep *inodedep;
+	struct workhead dotdotwk;
+	struct worklist *wk;
+	struct ufsmount *ump;
+	struct mount *mp;
 	struct vnode *vp;
 	struct inode *ip;
 	ino_t oldinum;
 	int error;
 
+	if (dirrem->dm_state & ONWORKLIST)
+		panic("handle_workitem_remove: dirrem %p still on worklist",
+		    dirrem);
+	oldinum = dirrem->dm_oldinum;
+	mp = dirrem->dm_list.wk_mp;
+	ump = VFSTOUFS(mp);
 	if ((vp = xp) == NULL &&
-	    (error = ffs_vgetf(dirrem->dm_list.wk_mp,
-		    dirrem->dm_oldinum, LK_EXCLUSIVE, &vp, FFSV_FORCEINSMQ)) != 0) {
+	    (error = ffs_vgetf(mp, oldinum, LK_EXCLUSIVE, &vp,
+	    FFSV_FORCEINSMQ)) != 0) {
 		softdep_error("handle_workitem_remove: vget", error);
 		return;
 	}
 	ip = VTOI(vp);
 	ACQUIRE_LOCK(&lk);
-	if ((inodedep_lookup(dirrem->dm_list.wk_mp,
-	    dirrem->dm_oldinum, 0, &inodedep)) == 0)
+	if ((inodedep_lookup(mp, oldinum, 0, &inodedep)) == 0)
 		panic("handle_workitem_remove: lost inodedep");
+	if (dirrem->dm_state & ONDEPLIST)
+		LIST_REMOVE(dirrem, dm_inonext);
+	KASSERT(LIST_EMPTY(&dirrem->dm_jremrefhd),
+	    ("handle_workitem_remove:  Journal entries not written."));
+
+	/*
+	 * Move all dependencies waiting on the remove to complete
+	 * from the dirrem to the inode inowait list to be completed
+	 * after the inode has been updated and written to disk.  Any
+	 * marked MKDIR_PARENT are saved to be completed when the .. ref
+	 * is removed.
+	 */
+	LIST_INIT(&dotdotwk);
+	while ((wk = LIST_FIRST(&dirrem->dm_jwork)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		if (wk->wk_state & MKDIR_PARENT) {
+			wk->wk_state &= ~MKDIR_PARENT;
+			WORKLIST_INSERT(&dotdotwk, wk);
+			continue;
+		}
+		WORKLIST_INSERT(&inodedep->id_inowait, wk);
+	}
+	LIST_SWAP(&dirrem->dm_jwork, &dotdotwk, worklist, wk_list);
 	/*
 	 * Normal file deletion.
 	 */
@@ -3609,12 +7637,16 @@ handle_workitem_remove(dirrem, xp)
 		ip->i_flag |= IN_CHANGE;
 		if (ip->i_nlink < ip->i_effnlink)
 			panic("handle_workitem_remove: bad file delta");
+		if (ip->i_nlink == 0) 
+			unlinked_inodedep(mp, inodedep);
 		inodedep->id_nlinkdelta = ip->i_nlink - ip->i_effnlink;
 		num_dirrem -= 1;
+		KASSERT(LIST_EMPTY(&dirrem->dm_jwork),
+		    ("handle_workitem_remove: worklist not empty. %s",
+		    TYPENAME(LIST_FIRST(&dirrem->dm_jwork)->wk_type)));
 		WORKITEM_FREE(dirrem, D_DIRREM);
 		FREE_LOCK(&lk);
-		vput(vp);
-		return;
+		goto out;
 	}
 	/*
 	 * Directory deletion. Decrement reference count for both the
@@ -3628,6 +7660,8 @@ handle_workitem_remove(dirrem, xp)
 	ip->i_flag |= IN_CHANGE;
 	if (ip->i_nlink < ip->i_effnlink)
 		panic("handle_workitem_remove: bad dir delta");
+	if (ip->i_nlink == 0)
+		unlinked_inodedep(mp, inodedep);
 	inodedep->id_nlinkdelta = ip->i_nlink - ip->i_effnlink;
 	FREE_LOCK(&lk);
 	if ((error = ffs_truncate(vp, (off_t)0, 0, td->td_ucred, td)) != 0)
@@ -3639,36 +7673,47 @@ handle_workitem_remove(dirrem, xp)
 	 * directory should not change. Thus we skip the followup dirrem.
 	 */
 	if (dirrem->dm_state & DIRCHG) {
+		KASSERT(LIST_EMPTY(&dirrem->dm_jwork),
+		    ("handle_workitem_remove: DIRCHG and worklist not empty."));
 		num_dirrem -= 1;
 		WORKITEM_FREE(dirrem, D_DIRREM);
 		FREE_LOCK(&lk);
-		vput(vp);
-		return;
+		goto out;
 	}
-	/*
-	 * If the inodedep does not exist, then the zero'ed inode has
-	 * been written to disk. If the allocated inode has never been
-	 * written to disk, then the on-disk inode is zero'ed. In either
-	 * case we can remove the file immediately.
-	 */
-	dirrem->dm_state = 0;
-	oldinum = dirrem->dm_oldinum;
+	dirrem->dm_state = ONDEPLIST;
 	dirrem->dm_oldinum = dirrem->dm_dirinum;
-	if (inodedep_lookup(dirrem->dm_list.wk_mp, oldinum,
-	    0, &inodedep) == 0 || check_inode_unwritten(inodedep)) {
+	/*
+	 * Place the dirrem on the parent's diremhd list.
+	 */
+	if (inodedep_lookup(mp, dirrem->dm_oldinum, 0, &inodedep) == 0)
+		panic("handle_workitem_remove: lost dir inodedep");
+	LIST_INSERT_HEAD(&inodedep->id_dirremhd, dirrem, dm_inonext);
+	/*
+	 * If the allocated inode has never been written to disk, then
+	 * the on-disk inode is zero'ed and we can remove the file
+	 * immediately.  When journaling if the inode has been marked
+	 * unlinked and not DEPCOMPLETE we know it can never be written.
+	 */
+	inodedep_lookup(mp, oldinum, 0, &inodedep);
+	if (inodedep == NULL ||
+	    (inodedep->id_state & (DEPCOMPLETE | UNLINKED)) == UNLINKED ||
+	    check_inode_unwritten(inodedep)) {
 		if (xp != NULL)
-			add_to_worklist(&dirrem->dm_list);
+			add_to_worklist(&dirrem->dm_list, 0);
 		FREE_LOCK(&lk);
-		vput(vp);
-		if (xp == NULL)
+		if (xp == NULL) {
+			vput(vp);
 			handle_workitem_remove(dirrem, NULL);
+		}
 		return;
 	}
 	WORKLIST_INSERT(&inodedep->id_inowait, &dirrem->dm_list);
 	FREE_LOCK(&lk);
 	ip->i_flag |= IN_CHANGE;
+out:
 	ffs_update(vp, 0);
-	vput(vp);
+	if (xp == NULL)
+		vput(vp);
 }
 
 /*
@@ -3689,6 +7734,7 @@ static void
 handle_workitem_freefile(freefile)
 	struct freefile *freefile;
 {
+	struct workhead wkhd;
 	struct fs *fs;
 	struct inodedep *idp;
 	struct ufsmount *ump;
@@ -3701,13 +7747,15 @@ handle_workitem_freefile(freefile)
 	error = inodedep_lookup(UFSTOVFS(ump), freefile->fx_oldinum, 0, &idp);
 	FREE_LOCK(&lk);
 	if (error)
-		panic("handle_workitem_freefile: inodedep survived");
+		panic("handle_workitem_freefile: inodedep %p survived", idp);
 #endif
 	UFS_LOCK(ump);
 	fs->fs_pendinginodes -= 1;
 	UFS_UNLOCK(ump);
+	LIST_INIT(&wkhd);
+	LIST_SWAP(&freefile->fx_jwork, &wkhd, worklist, wk_list);
 	if ((error = ffs_freefile(ump, fs, freefile->fx_devvp,
-	    freefile->fx_oldinum, freefile->fx_mode)) != 0)
+	    freefile->fx_oldinum, freefile->fx_mode, &wkhd)) != 0)
 		softdep_error("handle_workitem_freefile", error);
 	ACQUIRE_LOCK(&lk);
 	WORKITEM_FREE(freefile, D_FREEFILE);
@@ -3757,8 +7805,10 @@ softdep_disk_io_initiation(bp)
 {
 	struct worklist *wk;
 	struct worklist marker;
-	struct indirdep *indirdep;
 	struct inodedep *inodedep;
+	struct freeblks *freeblks;
+	struct jfreeblk *jfreeblk;
+	struct newblk *newblk;
 
 	/*
 	 * We only care about write operations. There should never
@@ -3767,6 +7817,10 @@ softdep_disk_io_initiation(bp)
 	if (bp->b_iocmd != BIO_WRITE)
 		panic("softdep_disk_io_initiation: not write");
 
+	if (bp->b_vflags & BV_BKGRDINPROG)
+		panic("softdep_disk_io_initiation: Writing buffer with "
+		    "background write in progress: %p", bp);
+
 	marker.wk_type = D_LAST + 1;	/* Not a normal workitem */
 	PHOLD(curproc);			/* Don't swap out kernel stack */
 
@@ -3792,46 +7846,58 @@ softdep_disk_io_initiation(bp)
 			continue;
 
 		case D_INDIRDEP:
-			indirdep = WK_INDIRDEP(wk);
-			if (indirdep->ir_state & GOINGAWAY)
-				panic("disk_io_initiation: indirdep gone");
-			/*
-			 * If there are no remaining dependencies, this
-			 * will be writing the real pointers, so the
-			 * dependency can be freed.
-			 */
-			if (LIST_EMPTY(&indirdep->ir_deplisthd)) {
-				struct buf *bp;
+			initiate_write_indirdep(WK_INDIRDEP(wk), bp);
+			continue;
 
-				bp = indirdep->ir_savebp;
-				bp->b_flags |= B_INVAL | B_NOCACHE;
-				/* inline expand WORKLIST_REMOVE(wk); */
-				wk->wk_state &= ~ONWORKLIST;
-				LIST_REMOVE(wk, wk_list);
-				WORKITEM_FREE(indirdep, D_INDIRDEP);
-				FREE_LOCK(&lk);
-				brelse(bp);
-				ACQUIRE_LOCK(&lk);
-				continue;
-			}
+		case D_BMSAFEMAP:
+			initiate_write_bmsafemap(WK_BMSAFEMAP(wk), bp);
+			continue;
+
+		case D_JSEG:
+			WK_JSEG(wk)->js_buf = NULL;
+			continue;
+
+		case D_FREEBLKS:
+			freeblks = WK_FREEBLKS(wk);
+			jfreeblk = LIST_FIRST(&freeblks->fb_jfreeblkhd);
 			/*
-			 * Replace up-to-date version with safe version.
+			 * We have to wait for the jfreeblks to be journaled
+			 * before we can write an inodeblock with updated
+			 * pointers.  Be careful to arrange the marker so
+			 * we revisit the jfreeblk if it's not removed by
+			 * the first jwait().
 			 */
-			FREE_LOCK(&lk);
-			indirdep->ir_saveddata = malloc(bp->b_bcount,
-			    M_INDIRDEP, M_SOFTDEP_FLAGS);
-			ACQUIRE_LOCK(&lk);
-			indirdep->ir_state &= ~ATTACHED;
-			indirdep->ir_state |= UNDONE;
-			bcopy(bp->b_data, indirdep->ir_saveddata, bp->b_bcount);
-			bcopy(indirdep->ir_savebp->b_data, bp->b_data,
-			    bp->b_bcount);
+			if (jfreeblk != NULL) {
+				LIST_REMOVE(&marker, wk_list);
+				LIST_INSERT_BEFORE(wk, &marker, wk_list);
+				jwait(&jfreeblk->jf_list);
+			}
+			continue;
+		case D_ALLOCDIRECT:
+		case D_ALLOCINDIR:
+			/*
+			 * We have to wait for the jnewblk to be journaled
+			 * before we can write to a block otherwise the
+			 * contents may be confused with an earlier file
+			 * at recovery time.  Handle the marker as described
+			 * above.
+			 */
+			newblk = WK_NEWBLK(wk);
+			if (newblk->nb_jnewblk != NULL) {
+				LIST_REMOVE(&marker, wk_list);
+				LIST_INSERT_BEFORE(wk, &marker, wk_list);
+				jwait(&newblk->nb_jnewblk->jn_list);
+			}
+			continue;
+
+		case D_SBDEP:
+			initiate_write_sbdep(WK_SBDEP(wk));
 			continue;
 
 		case D_MKDIR:
-		case D_BMSAFEMAP:
-		case D_ALLOCDIRECT:
-		case D_ALLOCINDIR:
+		case D_FREEWORK:
+		case D_FREEDEP:
+		case D_JSEGDEP:
 			continue;
 
 		default:
@@ -3855,6 +7921,9 @@ initiate_write_filepage(pagedep, bp)
 	struct pagedep *pagedep;
 	struct buf *bp;
 {
+	struct jremref *jremref;
+	struct jmvref *jmvref;
+	struct dirrem *dirrem;
 	struct diradd *dap;
 	struct direct *ep;
 	int i;
@@ -3869,6 +7938,22 @@ initiate_write_filepage(pagedep, bp)
 		return;
 	}
 	pagedep->pd_state |= IOSTARTED;
+	/*
+	 * Wait for all journal remove dependencies to hit the disk.
+	 * We can not allow any potentially conflicting directory adds
+	 * to be visible before removes and rollback is too difficult.
+	 * lk may be dropped and re-acquired, however we hold the buf
+	 * locked so the dependency can not go away.
+	 */
+	LIST_FOREACH(dirrem, &pagedep->pd_dirremhd, dm_next)
+		while ((jremref = LIST_FIRST(&dirrem->dm_jremrefhd)) != NULL) {
+			stat_jwait_filepage++;
+			jwait(&jremref->jr_list);
+		}
+	while ((jmvref = LIST_FIRST(&pagedep->pd_jmvrefhd)) != NULL) {
+		stat_jwait_filepage++;
+		jwait(&jmvref->jm_list);
+	}
 	for (i = 0; i < DAHASHSZ; i++) {
 		LIST_FOREACH(dap, &pagedep->pd_diraddhd[i], da_pdlist) {
 			ep = (struct direct *)
@@ -3905,6 +7990,7 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	struct allocdirect *adp, *lastadp;
 	struct ufs1_dinode *dp;
 	struct ufs1_dinode *sip;
+	struct inoref *inoref;
 	struct fs *fs;
 	ufs_lbn_t i;
 #ifdef INVARIANTS
@@ -3918,6 +8004,17 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	fs = inodedep->id_fs;
 	dp = (struct ufs1_dinode *)bp->b_data +
 	    ino_to_fsbo(fs, inodedep->id_ino);
+
+	/*
+	 * If we're on the unlinked list but have not yet written our
+	 * next pointer initialize it here.
+	 */
+	if ((inodedep->id_state & (UNLINKED | UNLINKNEXT)) == UNLINKED) {
+		struct inodedep *inon;
+
+		inon = TAILQ_NEXT(inodedep, id_unlinked);
+		dp->di_freelink = inon ? inon->id_ino : 0;
+	}
 	/*
 	 * If the bitmap is not yet written, then the allocated
 	 * inode cannot be written to disk.
@@ -3933,6 +8030,7 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 		*inodedep->id_savedino1 = *dp;
 		bzero((caddr_t)dp, sizeof(struct ufs1_dinode));
 		dp->di_gen = inodedep->id_savedino1->di_gen;
+		dp->di_freelink = inodedep->id_savedino1->di_freelink;
 		return;
 	}
 	/*
@@ -3940,32 +8038,40 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	 */
 	inodedep->id_savedsize = dp->di_size;
 	inodedep->id_savedextsize = 0;
-	if (TAILQ_EMPTY(&inodedep->id_inoupdt))
+	inodedep->id_savednlink = dp->di_nlink;
+	if (TAILQ_EMPTY(&inodedep->id_inoupdt) &&
+	    TAILQ_EMPTY(&inodedep->id_inoreflst))
 		return;
+	/*
+	 * Revert the link count to that of the first unwritten journal entry.
+	 */
+	inoref = TAILQ_FIRST(&inodedep->id_inoreflst);
+	if (inoref)
+		dp->di_nlink = inoref->if_nlink;
 	/*
 	 * Set the dependencies to busy.
 	 */
 	for (deplist = 0, adp = TAILQ_FIRST(&inodedep->id_inoupdt); adp;
 	     adp = TAILQ_NEXT(adp, ad_next)) {
 #ifdef INVARIANTS
-		if (deplist != 0 && prevlbn >= adp->ad_lbn)
+		if (deplist != 0 && prevlbn >= adp->ad_offset)
 			panic("softdep_write_inodeblock: lbn order");
-		prevlbn = adp->ad_lbn;
-		if (adp->ad_lbn < NDADDR &&
-		    dp->di_db[adp->ad_lbn] != adp->ad_newblkno)
+		prevlbn = adp->ad_offset;
+		if (adp->ad_offset < NDADDR &&
+		    dp->di_db[adp->ad_offset] != adp->ad_newblkno)
 			panic("%s: direct pointer #%jd mismatch %d != %jd",
 			    "softdep_write_inodeblock",
-			    (intmax_t)adp->ad_lbn,
-			    dp->di_db[adp->ad_lbn],
+			    (intmax_t)adp->ad_offset,
+			    dp->di_db[adp->ad_offset],
 			    (intmax_t)adp->ad_newblkno);
-		if (adp->ad_lbn >= NDADDR &&
-		    dp->di_ib[adp->ad_lbn - NDADDR] != adp->ad_newblkno)
+		if (adp->ad_offset >= NDADDR &&
+		    dp->di_ib[adp->ad_offset - NDADDR] != adp->ad_newblkno)
 			panic("%s: indirect pointer #%jd mismatch %d != %jd",
 			    "softdep_write_inodeblock",
-			    (intmax_t)adp->ad_lbn - NDADDR,
-			    dp->di_ib[adp->ad_lbn - NDADDR],
+			    (intmax_t)adp->ad_offset - NDADDR,
+			    dp->di_ib[adp->ad_offset - NDADDR],
 			    (intmax_t)adp->ad_newblkno);
-		deplist |= 1 << adp->ad_lbn;
+		deplist |= 1 << adp->ad_offset;
 		if ((adp->ad_state & ATTACHED) == 0)
 			panic("softdep_write_inodeblock: Unknown state 0x%x",
 			    adp->ad_state);
@@ -3981,14 +8087,14 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	 */
 	for (lastadp = NULL, adp = TAILQ_FIRST(&inodedep->id_inoupdt); adp;
 	     lastadp = adp, adp = TAILQ_NEXT(adp, ad_next)) {
-		if (adp->ad_lbn >= NDADDR)
+		if (adp->ad_offset >= NDADDR)
 			break;
-		dp->di_db[adp->ad_lbn] = adp->ad_oldblkno;
+		dp->di_db[adp->ad_offset] = adp->ad_oldblkno;
 		/* keep going until hitting a rollback to a frag */
 		if (adp->ad_oldsize == 0 || adp->ad_oldsize == fs->fs_bsize)
 			continue;
-		dp->di_size = fs->fs_bsize * adp->ad_lbn + adp->ad_oldsize;
-		for (i = adp->ad_lbn + 1; i < NDADDR; i++) {
+		dp->di_size = fs->fs_bsize * adp->ad_offset + adp->ad_oldsize;
+		for (i = adp->ad_offset + 1; i < NDADDR; i++) {
 #ifdef INVARIANTS
 			if (dp->di_db[i] != 0 && (deplist & (1 << i)) == 0)
 				panic("softdep_write_inodeblock: lost dep1");
@@ -4012,8 +8118,8 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	 * we already checked for fragments in the loop above.
 	 */
 	if (lastadp != NULL &&
-	    dp->di_size <= (lastadp->ad_lbn + 1) * fs->fs_bsize) {
-		for (i = lastadp->ad_lbn; i >= 0; i--)
+	    dp->di_size <= (lastadp->ad_offset + 1) * fs->fs_bsize) {
+		for (i = lastadp->ad_offset; i >= 0; i--)
 			if (dp->di_db[i] != 0)
 				break;
 		dp->di_size = (i + 1) * fs->fs_bsize;
@@ -4030,7 +8136,7 @@ initiate_write_inodeblock_ufs1(inodedep, bp)
 	 * postpone fsck, we are stuck with this argument.
 	 */
 	for (; adp; adp = TAILQ_NEXT(adp, ad_next))
-		dp->di_ib[adp->ad_lbn - NDADDR] = 0;
+		dp->di_ib[adp->ad_offset - NDADDR] = 0;
 }
 		
 /*
@@ -4051,6 +8157,7 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	struct allocdirect *adp, *lastadp;
 	struct ufs2_dinode *dp;
 	struct ufs2_dinode *sip;
+	struct inoref *inoref;
 	struct fs *fs;
 	ufs_lbn_t i;
 #ifdef INVARIANTS
@@ -4064,6 +8171,29 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	fs = inodedep->id_fs;
 	dp = (struct ufs2_dinode *)bp->b_data +
 	    ino_to_fsbo(fs, inodedep->id_ino);
+
+	/*
+	 * If we're on the unlinked list but have not yet written our
+	 * next pointer initialize it here.
+	 */
+	if ((inodedep->id_state & (UNLINKED | UNLINKNEXT)) == UNLINKED) {
+		struct inodedep *inon;
+
+		inon = TAILQ_NEXT(inodedep, id_unlinked);
+		dp->di_freelink = inon ? inon->id_ino : 0;
+	}
+	if ((inodedep->id_state & (UNLINKED | UNLINKNEXT)) ==
+	    (UNLINKED | UNLINKNEXT)) {
+		struct inodedep *inon;
+		ino_t freelink;
+
+		inon = TAILQ_NEXT(inodedep, id_unlinked);
+		freelink = inon ? inon->id_ino : 0;
+		if (freelink != dp->di_freelink)
+			panic("ino %p(0x%X) %d, %d != %d",
+			    inodedep, inodedep->id_state, inodedep->id_ino,
+			    freelink, dp->di_freelink);
+	}
 	/*
 	 * If the bitmap is not yet written, then the allocated
 	 * inode cannot be written to disk.
@@ -4079,6 +8209,7 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 		*inodedep->id_savedino2 = *dp;
 		bzero((caddr_t)dp, sizeof(struct ufs2_dinode));
 		dp->di_gen = inodedep->id_savedino2->di_gen;
+		dp->di_freelink = inodedep->id_savedino2->di_freelink;
 		return;
 	}
 	/*
@@ -4086,25 +8217,34 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 */
 	inodedep->id_savedsize = dp->di_size;
 	inodedep->id_savedextsize = dp->di_extsize;
+	inodedep->id_savednlink = dp->di_nlink;
 	if (TAILQ_EMPTY(&inodedep->id_inoupdt) &&
-	    TAILQ_EMPTY(&inodedep->id_extupdt))
+	    TAILQ_EMPTY(&inodedep->id_extupdt) &&
+	    TAILQ_EMPTY(&inodedep->id_inoreflst))
 		return;
+	/*
+	 * Revert the link count to that of the first unwritten journal entry.
+	 */
+	inoref = TAILQ_FIRST(&inodedep->id_inoreflst);
+	if (inoref)
+		dp->di_nlink = inoref->if_nlink;
+
 	/*
 	 * Set the ext data dependencies to busy.
 	 */
 	for (deplist = 0, adp = TAILQ_FIRST(&inodedep->id_extupdt); adp;
 	     adp = TAILQ_NEXT(adp, ad_next)) {
 #ifdef INVARIANTS
-		if (deplist != 0 && prevlbn >= adp->ad_lbn)
+		if (deplist != 0 && prevlbn >= adp->ad_offset)
 			panic("softdep_write_inodeblock: lbn order");
-		prevlbn = adp->ad_lbn;
-		if (dp->di_extb[adp->ad_lbn] != adp->ad_newblkno)
+		prevlbn = adp->ad_offset;
+		if (dp->di_extb[adp->ad_offset] != adp->ad_newblkno)
 			panic("%s: direct pointer #%jd mismatch %jd != %jd",
 			    "softdep_write_inodeblock",
-			    (intmax_t)adp->ad_lbn,
-			    (intmax_t)dp->di_extb[adp->ad_lbn],
+			    (intmax_t)adp->ad_offset,
+			    (intmax_t)dp->di_extb[adp->ad_offset],
 			    (intmax_t)adp->ad_newblkno);
-		deplist |= 1 << adp->ad_lbn;
+		deplist |= 1 << adp->ad_offset;
 		if ((adp->ad_state & ATTACHED) == 0)
 			panic("softdep_write_inodeblock: Unknown state 0x%x",
 			    adp->ad_state);
@@ -4120,12 +8260,12 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 */
 	for (lastadp = NULL, adp = TAILQ_FIRST(&inodedep->id_extupdt); adp;
 	     lastadp = adp, adp = TAILQ_NEXT(adp, ad_next)) {
-		dp->di_extb[adp->ad_lbn] = adp->ad_oldblkno;
+		dp->di_extb[adp->ad_offset] = adp->ad_oldblkno;
 		/* keep going until hitting a rollback to a frag */
 		if (adp->ad_oldsize == 0 || adp->ad_oldsize == fs->fs_bsize)
 			continue;
-		dp->di_extsize = fs->fs_bsize * adp->ad_lbn + adp->ad_oldsize;
-		for (i = adp->ad_lbn + 1; i < NXADDR; i++) {
+		dp->di_extsize = fs->fs_bsize * adp->ad_offset + adp->ad_oldsize;
+		for (i = adp->ad_offset + 1; i < NXADDR; i++) {
 #ifdef INVARIANTS
 			if (dp->di_extb[i] != 0 && (deplist & (1 << i)) == 0)
 				panic("softdep_write_inodeblock: lost dep1");
@@ -4142,8 +8282,8 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 * we already checked for fragments in the loop above.
 	 */
 	if (lastadp != NULL &&
-	    dp->di_extsize <= (lastadp->ad_lbn + 1) * fs->fs_bsize) {
-		for (i = lastadp->ad_lbn; i >= 0; i--)
+	    dp->di_extsize <= (lastadp->ad_offset + 1) * fs->fs_bsize) {
+		for (i = lastadp->ad_offset; i >= 0; i--)
 			if (dp->di_extb[i] != 0)
 				break;
 		dp->di_extsize = (i + 1) * fs->fs_bsize;
@@ -4154,24 +8294,24 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	for (deplist = 0, adp = TAILQ_FIRST(&inodedep->id_inoupdt); adp;
 	     adp = TAILQ_NEXT(adp, ad_next)) {
 #ifdef INVARIANTS
-		if (deplist != 0 && prevlbn >= adp->ad_lbn)
+		if (deplist != 0 && prevlbn >= adp->ad_offset)
 			panic("softdep_write_inodeblock: lbn order");
-		prevlbn = adp->ad_lbn;
-		if (adp->ad_lbn < NDADDR &&
-		    dp->di_db[adp->ad_lbn] != adp->ad_newblkno)
+		prevlbn = adp->ad_offset;
+		if (adp->ad_offset < NDADDR &&
+		    dp->di_db[adp->ad_offset] != adp->ad_newblkno)
 			panic("%s: direct pointer #%jd mismatch %jd != %jd",
 			    "softdep_write_inodeblock",
-			    (intmax_t)adp->ad_lbn,
-			    (intmax_t)dp->di_db[adp->ad_lbn],
+			    (intmax_t)adp->ad_offset,
+			    (intmax_t)dp->di_db[adp->ad_offset],
 			    (intmax_t)adp->ad_newblkno);
-		if (adp->ad_lbn >= NDADDR &&
-		    dp->di_ib[adp->ad_lbn - NDADDR] != adp->ad_newblkno)
+		if (adp->ad_offset >= NDADDR &&
+		    dp->di_ib[adp->ad_offset - NDADDR] != adp->ad_newblkno)
 			panic("%s indirect pointer #%jd mismatch %jd != %jd",
 			    "softdep_write_inodeblock:",
-			    (intmax_t)adp->ad_lbn - NDADDR,
-			    (intmax_t)dp->di_ib[adp->ad_lbn - NDADDR],
+			    (intmax_t)adp->ad_offset - NDADDR,
+			    (intmax_t)dp->di_ib[adp->ad_offset - NDADDR],
 			    (intmax_t)adp->ad_newblkno);
-		deplist |= 1 << adp->ad_lbn;
+		deplist |= 1 << adp->ad_offset;
 		if ((adp->ad_state & ATTACHED) == 0)
 			panic("softdep_write_inodeblock: Unknown state 0x%x",
 			    adp->ad_state);
@@ -4187,14 +8327,14 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 */
 	for (lastadp = NULL, adp = TAILQ_FIRST(&inodedep->id_inoupdt); adp;
 	     lastadp = adp, adp = TAILQ_NEXT(adp, ad_next)) {
-		if (adp->ad_lbn >= NDADDR)
+		if (adp->ad_offset >= NDADDR)
 			break;
-		dp->di_db[adp->ad_lbn] = adp->ad_oldblkno;
+		dp->di_db[adp->ad_offset] = adp->ad_oldblkno;
 		/* keep going until hitting a rollback to a frag */
 		if (adp->ad_oldsize == 0 || adp->ad_oldsize == fs->fs_bsize)
 			continue;
-		dp->di_size = fs->fs_bsize * adp->ad_lbn + adp->ad_oldsize;
-		for (i = adp->ad_lbn + 1; i < NDADDR; i++) {
+		dp->di_size = fs->fs_bsize * adp->ad_offset + adp->ad_oldsize;
+		for (i = adp->ad_offset + 1; i < NDADDR; i++) {
 #ifdef INVARIANTS
 			if (dp->di_db[i] != 0 && (deplist & (1 << i)) == 0)
 				panic("softdep_write_inodeblock: lost dep2");
@@ -4218,8 +8358,8 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 * we already checked for fragments in the loop above.
 	 */
 	if (lastadp != NULL &&
-	    dp->di_size <= (lastadp->ad_lbn + 1) * fs->fs_bsize) {
-		for (i = lastadp->ad_lbn; i >= 0; i--)
+	    dp->di_size <= (lastadp->ad_offset + 1) * fs->fs_bsize) {
+		for (i = lastadp->ad_offset; i >= 0; i--)
 			if (dp->di_db[i] != 0)
 				break;
 		dp->di_size = (i + 1) * fs->fs_bsize;
@@ -4236,7 +8376,355 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
 	 * postpone fsck, we are stuck with this argument.
 	 */
 	for (; adp; adp = TAILQ_NEXT(adp, ad_next))
-		dp->di_ib[adp->ad_lbn - NDADDR] = 0;
+		dp->di_ib[adp->ad_offset - NDADDR] = 0;
+}
+
+/*
+ * Cancel an indirdep as a result of truncation.  Release all of the
+ * children allocindirs and place their journal work on the appropriate
+ * list.
+ */
+static void
+cancel_indirdep(indirdep, bp, inodedep, freeblks)
+	struct indirdep *indirdep;
+	struct buf *bp;
+	struct inodedep *inodedep;
+	struct freeblks *freeblks;
+{
+	struct allocindir *aip;
+
+	/*
+	 * None of the indirect pointers will ever be visible,
+	 * so they can simply be tossed. GOINGAWAY ensures
+	 * that allocated pointers will be saved in the buffer
+	 * cache until they are freed. Note that they will
+	 * only be able to be found by their physical address
+	 * since the inode mapping the logical address will
+	 * be gone. The save buffer used for the safe copy
+	 * was allocated in setup_allocindir_phase2 using
+	 * the physical address so it could be used for this
+	 * purpose. Hence we swap the safe copy with the real
+	 * copy, allowing the safe copy to be freed and holding
+	 * on to the real copy for later use in indir_trunc.
+	 */
+	if (indirdep->ir_state & GOINGAWAY)
+		panic("cancel_indirdep: already gone");
+	if (indirdep->ir_state & ONDEPLIST) {
+		indirdep->ir_state &= ~ONDEPLIST;
+		LIST_REMOVE(indirdep, ir_next);
+	}
+	indirdep->ir_state |= GOINGAWAY;
+	VFSTOUFS(indirdep->ir_list.wk_mp)->um_numindirdeps += 1;
+	while ((aip = LIST_FIRST(&indirdep->ir_deplisthd)) != 0)
+		cancel_allocindir(aip, inodedep, freeblks);
+	while ((aip = LIST_FIRST(&indirdep->ir_donehd)) != 0)
+		cancel_allocindir(aip, inodedep, freeblks);
+	while ((aip = LIST_FIRST(&indirdep->ir_writehd)) != 0)
+		cancel_allocindir(aip, inodedep, freeblks);
+	while ((aip = LIST_FIRST(&indirdep->ir_completehd)) != 0)
+		cancel_allocindir(aip, inodedep, freeblks);
+	bcopy(bp->b_data, indirdep->ir_savebp->b_data, bp->b_bcount);
+	WORKLIST_REMOVE(&indirdep->ir_list);
+	WORKLIST_INSERT(&indirdep->ir_savebp->b_dep, &indirdep->ir_list);
+	indirdep->ir_savebp = NULL;
+}
+
+/*
+ * Free an indirdep once it no longer has new pointers to track.
+ */
+static void
+free_indirdep(indirdep)
+	struct indirdep *indirdep;
+{
+
+	KASSERT(LIST_EMPTY(&indirdep->ir_jwork),
+	    ("free_indirdep: Journal work not empty."));
+	KASSERT(LIST_EMPTY(&indirdep->ir_completehd),
+	    ("free_indirdep: Complete head not empty."));
+	KASSERT(LIST_EMPTY(&indirdep->ir_writehd),
+	    ("free_indirdep: write head not empty."));
+	KASSERT(LIST_EMPTY(&indirdep->ir_donehd),
+	    ("free_indirdep: done head not empty."));
+	KASSERT(LIST_EMPTY(&indirdep->ir_deplisthd),
+	    ("free_indirdep: deplist head not empty."));
+	KASSERT(indirdep->ir_savebp == NULL,
+	    ("free_indirdep: %p ir_savebp != NULL", indirdep));
+	KASSERT((indirdep->ir_state & ONDEPLIST) == 0,
+	    ("free_indirdep: %p still on deplist.", indirdep));
+	if (indirdep->ir_state & ONWORKLIST)
+		WORKLIST_REMOVE(&indirdep->ir_list);
+	WORKITEM_FREE(indirdep, D_INDIRDEP);
+}
+
+/*
+ * Called before a write to an indirdep.  This routine is responsible for
+ * rolling back pointers to a safe state which includes only those
+ * allocindirs which have been completed.
+ */
+static void
+initiate_write_indirdep(indirdep, bp)
+	struct indirdep *indirdep;
+	struct buf *bp;
+{
+
+	if (indirdep->ir_state & GOINGAWAY)
+		panic("disk_io_initiation: indirdep gone");
+
+	/*
+	 * If there are no remaining dependencies, this will be writing
+	 * the real pointers.
+	 */
+	if (LIST_EMPTY(&indirdep->ir_deplisthd))
+		return;
+	/*
+	 * Replace up-to-date version with safe version.
+	 */
+	FREE_LOCK(&lk);
+	indirdep->ir_saveddata = malloc(bp->b_bcount, M_INDIRDEP,
+	    M_SOFTDEP_FLAGS);
+	ACQUIRE_LOCK(&lk);
+	indirdep->ir_state &= ~ATTACHED;
+	indirdep->ir_state |= UNDONE;
+	bcopy(bp->b_data, indirdep->ir_saveddata, bp->b_bcount);
+	bcopy(indirdep->ir_savebp->b_data, bp->b_data,
+	    bp->b_bcount);
+}
+
+/*
+ * Called when an inode has been cleared in a cg bitmap.  This finally
+ * eliminates any canceled jaddrefs
+ */
+void
+softdep_setup_inofree(mp, bp, ino, wkhd)
+	struct mount *mp;
+	struct buf *bp;
+	ino_t ino;
+	struct workhead *wkhd;
+{
+	struct worklist *wk, *wkn;
+	struct inodedep *inodedep;
+	uint8_t *inosused;
+	struct cg *cgp;
+	struct fs *fs;
+
+	ACQUIRE_LOCK(&lk);
+	fs = VFSTOUFS(mp)->um_fs;
+	cgp = (struct cg *)bp->b_data;
+	inosused = cg_inosused(cgp);
+	if (isset(inosused, ino % fs->fs_ipg))
+		panic("softdep_setup_inofree: inode %d not freed.", ino);
+	if (inodedep_lookup(mp, ino, 0, &inodedep))
+		panic("softdep_setup_inofree: ino %d has existing inodedep %p",
+		    ino, inodedep);
+	if (wkhd) {
+		LIST_FOREACH_SAFE(wk, wkhd, wk_list, wkn) {
+			if (wk->wk_type != D_JADDREF)
+				continue;
+			WORKLIST_REMOVE(wk);
+			/*
+			 * We can free immediately even if the jaddref
+			 * isn't attached in a background write as now
+			 * the bitmaps are reconciled.
+		 	 */
+			wk->wk_state |= COMPLETE | ATTACHED;
+			free_jaddref(WK_JADDREF(wk));
+		}
+		jwork_move(&bp->b_dep, wkhd);
+	}
+	FREE_LOCK(&lk);
+}
+
+
+/*
+ * Called via ffs_blkfree() after a set of frags has been cleared from a cg
+ * map.  Any dependencies waiting for the write to clear are added to the
+ * buf's list and any jnewblks that are being canceled are discarded
+ * immediately.
+ */
+void
+softdep_setup_blkfree(mp, bp, blkno, frags, wkhd)
+	struct mount *mp;
+	struct buf *bp;
+	ufs2_daddr_t blkno;
+	int frags;
+	struct workhead *wkhd;
+{
+	struct jnewblk *jnewblk;
+	struct worklist *wk, *wkn;
+#ifdef SUJ_DEBUG
+	struct bmsafemap *bmsafemap;
+	struct fs *fs;
+	uint8_t *blksfree;
+	struct cg *cgp;
+	ufs2_daddr_t jstart;
+	ufs2_daddr_t jend;
+	ufs2_daddr_t end;
+	long bno;
+	int i;
+#endif
+
+	ACQUIRE_LOCK(&lk);
+	/*
+	 * Detach any jnewblks which have been canceled.  They must linger
+	 * until the bitmap is cleared again by ffs_blkfree() to prevent
+	 * an unjournaled allocation from hitting the disk.
+	 */
+	if (wkhd) {
+		LIST_FOREACH_SAFE(wk, wkhd, wk_list, wkn) {
+			if (wk->wk_type != D_JNEWBLK)
+				continue;
+			jnewblk = WK_JNEWBLK(wk);
+			KASSERT(jnewblk->jn_state & GOINGAWAY,
+			    ("softdep_setup_blkfree: jnewblk not canceled."));
+			WORKLIST_REMOVE(wk);
+#ifdef SUJ_DEBUG
+			/*
+			 * Assert that this block is free in the bitmap
+			 * before we discard the jnewblk.
+			 */
+			fs = VFSTOUFS(mp)->um_fs;
+			cgp = (struct cg *)bp->b_data;
+			blksfree = cg_blksfree(cgp);
+			bno = dtogd(fs, jnewblk->jn_blkno);
+			for (i = jnewblk->jn_oldfrags;
+			    i < jnewblk->jn_frags; i++) {
+				if (isset(blksfree, bno + i))
+					continue;
+				panic("softdep_setup_blkfree: not free");
+			}
+#endif
+			/*
+			 * Even if it's not attached we can free immediately
+			 * as the new bitmap is correct.
+			 */
+			wk->wk_state |= COMPLETE | ATTACHED;
+			free_jnewblk(jnewblk);
+		}
+		/*
+		 * The buf must be locked by the caller otherwise these could
+		 * be added while it's being written and the write would
+		 * complete them before they made it to disk.
+		 */
+		jwork_move(&bp->b_dep, wkhd);
+	}
+
+#ifdef SUJ_DEBUG
+	/*
+	 * Assert that we are not freeing a block which has an outstanding
+	 * allocation dependency.
+	 */
+	fs = VFSTOUFS(mp)->um_fs;
+	bmsafemap = bmsafemap_lookup(mp, bp, dtog(fs, blkno));
+	end = blkno + frags;
+	LIST_FOREACH(jnewblk, &bmsafemap->sm_jnewblkhd, jn_deps) {
+		/*
+		 * Don't match against blocks that will be freed when the
+		 * background write is done.
+		 */
+		if ((jnewblk->jn_state & (ATTACHED | COMPLETE | DEPCOMPLETE)) ==
+		    (COMPLETE | DEPCOMPLETE))
+			continue;
+		jstart = jnewblk->jn_blkno + jnewblk->jn_oldfrags;
+		jend = jnewblk->jn_blkno + jnewblk->jn_frags;
+		if ((blkno >= jstart && blkno < jend) ||
+		    (end > jstart && end <= jend)) {
+			printf("state 0x%X %jd - %d %d dep %p\n",
+			    jnewblk->jn_state, jnewblk->jn_blkno,
+			    jnewblk->jn_oldfrags, jnewblk->jn_frags,
+			    jnewblk->jn_newblk);
+			panic("softdep_setup_blkfree: "
+			    "%jd-%jd(%d) overlaps with %jd-%jd",
+			    blkno, end, frags, jstart, jend);
+		}
+	}
+#endif
+	FREE_LOCK(&lk);
+}
+
+static void 
+initiate_write_bmsafemap(bmsafemap, bp)
+	struct bmsafemap *bmsafemap;
+	struct buf *bp;			/* The cg block. */
+{
+	struct jaddref *jaddref;
+	struct jnewblk *jnewblk;
+	uint8_t *inosused;
+	uint8_t *blksfree;
+	struct cg *cgp;
+	struct fs *fs;
+	int cleared;
+	ino_t ino;
+	long bno;
+	int i;
+
+	if (bmsafemap->sm_state & IOSTARTED)
+		panic("initiate_write_bmsafemap: Already started\n");
+	bmsafemap->sm_state |= IOSTARTED;
+	/*
+	 * Clear any inode allocations which are pending journal writes.
+	 */
+	if (LIST_FIRST(&bmsafemap->sm_jaddrefhd) != NULL) {
+		cgp = (struct cg *)bp->b_data;
+		fs = VFSTOUFS(bmsafemap->sm_list.wk_mp)->um_fs;
+		inosused = cg_inosused(cgp);
+		LIST_FOREACH(jaddref, &bmsafemap->sm_jaddrefhd, ja_bmdeps) {
+			ino = jaddref->ja_ino % fs->fs_ipg;
+			/*
+			 * If this is a background copy the inode may not
+			 * be marked used yet.
+			 */
+			if (isset(inosused, ino)) {
+				if ((jaddref->ja_mode & IFMT) == IFDIR)
+					cgp->cg_cs.cs_ndir--;
+				cgp->cg_cs.cs_nifree++;
+				clrbit(inosused, ino);
+				jaddref->ja_state &= ~ATTACHED;
+				jaddref->ja_state |= UNDONE;
+				stat_jaddref++;
+			} else if ((bp->b_xflags & BX_BKGRDMARKER) == 0)
+				panic("initiate_write_bmsafemap: inode %d "
+				    "marked free", jaddref->ja_ino);
+		}
+	}
+	/*
+	 * Clear any block allocations which are pending journal writes.
+	 */
+	if (LIST_FIRST(&bmsafemap->sm_jnewblkhd) != NULL) {
+		cgp = (struct cg *)bp->b_data;
+		fs = VFSTOUFS(bmsafemap->sm_list.wk_mp)->um_fs;
+		blksfree = cg_blksfree(cgp);
+		LIST_FOREACH(jnewblk, &bmsafemap->sm_jnewblkhd, jn_deps) {
+			bno = dtogd(fs, jnewblk->jn_blkno);
+			cleared = 0;
+			for (i = jnewblk->jn_oldfrags; i < jnewblk->jn_frags;
+			    i++) {
+				if (isclr(blksfree, bno + i)) {
+					cleared = 1;
+					setbit(blksfree, bno + i);
+				}
+			}
+			/*
+			 * We may not clear the block if it's a background
+			 * copy.  In that case there is no reason to detach
+			 * it.
+			 */
+			if (cleared) {
+				stat_jnewblk++;
+				jnewblk->jn_state &= ~ATTACHED;
+				jnewblk->jn_state |= UNDONE;
+			} else if ((bp->b_xflags & BX_BKGRDMARKER) == 0)
+				panic("initiate_write_bmsafemap: block %jd "
+				    "marked free", jnewblk->jn_blkno);
+		}
+	}
+	/*
+	 * Move allocation lists to the written lists so they can be
+	 * cleared once the block write is complete.
+	 */
+	LIST_SWAP(&bmsafemap->sm_inodedephd, &bmsafemap->sm_inodedepwr,
+	    inodedep, id_deps);
+	LIST_SWAP(&bmsafemap->sm_newblkhd, &bmsafemap->sm_newblkwr,
+	    newblk, nb_deps);
 }
 
 /*
@@ -4246,6 +8734,7 @@ initiate_write_inodeblock_ufs2(inodedep, bp)
  * a request completion).  It should be called early in this
  * procedure, before the block is made available to other
  * processes or other routines are called.
+ *
  */
 static void 
 softdep_disk_write_complete(bp)
@@ -4254,12 +8743,7 @@ softdep_disk_write_complete(bp)
 	struct worklist *wk;
 	struct worklist *owk;
 	struct workhead reattach;
-	struct newblk *newblk;
-	struct allocindir *aip;
-	struct allocdirect *adp;
-	struct indirdep *indirdep;
-	struct inodedep *inodedep;
-	struct bmsafemap *bmsafemap;
+	struct buf *sbp;
 
 	/*
 	 * If an error occurred while doing the write, then the data
@@ -4271,8 +8755,9 @@ softdep_disk_write_complete(bp)
 	/*
 	 * This lock must not be released anywhere in this code segment.
 	 */
-	ACQUIRE_LOCK(&lk);
+	sbp = NULL;
 	owk = NULL;
+	ACQUIRE_LOCK(&lk);
 	while ((wk = LIST_FIRST(&bp->b_dep)) != NULL) {
 		WORKLIST_REMOVE(wk);
 		if (wk == owk)
@@ -4291,33 +8776,8 @@ softdep_disk_write_complete(bp)
 			continue;
 
 		case D_BMSAFEMAP:
-			bmsafemap = WK_BMSAFEMAP(wk);
-			while ((newblk = LIST_FIRST(&bmsafemap->sm_newblkhd))) {
-				newblk->nb_state |= DEPCOMPLETE;
-				newblk->nb_bmsafemap = NULL;
-				LIST_REMOVE(newblk, nb_deps);
-			}
-			while ((adp =
-			   LIST_FIRST(&bmsafemap->sm_allocdirecthd))) {
-				adp->ad_state |= DEPCOMPLETE;
-				adp->ad_buf = NULL;
-				LIST_REMOVE(adp, ad_deps);
-				handle_allocdirect_partdone(adp);
-			}
-			while ((aip =
-			    LIST_FIRST(&bmsafemap->sm_allocindirhd))) {
-				aip->ai_state |= DEPCOMPLETE;
-				aip->ai_buf = NULL;
-				LIST_REMOVE(aip, ai_deps);
-				handle_allocindir_partdone(aip);
-			}
-			while ((inodedep =
-			     LIST_FIRST(&bmsafemap->sm_inodedephd)) != NULL) {
-				inodedep->id_state |= DEPCOMPLETE;
-				LIST_REMOVE(inodedep, id_deps);
-				inodedep->id_buf = NULL;
-			}
-			WORKITEM_FREE(bmsafemap, D_BMSAFEMAP);
+			if (handle_written_bmsafemap(WK_BMSAFEMAP(wk), bp))
+				WORKLIST_INSERT(&reattach, wk);
 			continue;
 
 		case D_MKDIR:
@@ -4325,35 +8785,45 @@ softdep_disk_write_complete(bp)
 			continue;
 
 		case D_ALLOCDIRECT:
-			adp = WK_ALLOCDIRECT(wk);
-			adp->ad_state |= COMPLETE;
-			handle_allocdirect_partdone(adp);
+			wk->wk_state |= COMPLETE;
+			handle_allocdirect_partdone(WK_ALLOCDIRECT(wk), NULL);
 			continue;
 
 		case D_ALLOCINDIR:
-			aip = WK_ALLOCINDIR(wk);
-			aip->ai_state |= COMPLETE;
-			handle_allocindir_partdone(aip);
+			wk->wk_state |= COMPLETE;
+			handle_allocindir_partdone(WK_ALLOCINDIR(wk));
 			continue;
 
 		case D_INDIRDEP:
-			indirdep = WK_INDIRDEP(wk);
-			if (indirdep->ir_state & GOINGAWAY)
-				panic("disk_write_complete: indirdep gone");
-			bcopy(indirdep->ir_saveddata, bp->b_data, bp->b_bcount);
-			free(indirdep->ir_saveddata, M_INDIRDEP);
-			indirdep->ir_saveddata = 0;
-			indirdep->ir_state &= ~UNDONE;
-			indirdep->ir_state |= ATTACHED;
-			while ((aip = LIST_FIRST(&indirdep->ir_donehd)) != 0) {
-				handle_allocindir_partdone(aip);
-				if (aip == LIST_FIRST(&indirdep->ir_donehd))
-					panic("disk_write_complete: not gone");
-			}
-			WORKLIST_INSERT(&reattach, wk);
-			if ((bp->b_flags & B_DELWRI) == 0)
-				stat_indir_blk_ptrs++;
-			bdirty(bp);
+			if (handle_written_indirdep(WK_INDIRDEP(wk), bp, &sbp))
+				WORKLIST_INSERT(&reattach, wk);
+			continue;
+
+		case D_FREEBLKS:
+			wk->wk_state |= COMPLETE;
+			if ((wk->wk_state & ALLCOMPLETE) == ALLCOMPLETE)
+				add_to_worklist(wk, 1);
+			continue;
+
+		case D_FREEWORK:
+			handle_written_freework(WK_FREEWORK(wk));
+			break;
+
+		case D_FREEDEP:
+			free_freedep(WK_FREEDEP(wk));
+			continue;
+
+		case D_JSEGDEP:
+			free_jsegdep(WK_JSEGDEP(wk));
+			continue;
+
+		case D_JSEG:
+			handle_written_jseg(WK_JSEG(wk), bp);
+			continue;
+
+		case D_SBDEP:
+			if (handle_written_sbdep(WK_SBDEP(wk), bp))
+				WORKLIST_INSERT(&reattach, wk);
 			continue;
 
 		default:
@@ -4370,6 +8840,8 @@ softdep_disk_write_complete(bp)
 		WORKLIST_INSERT(&bp->b_dep, wk);
 	}
 	FREE_LOCK(&lk);
+	if (sbp)
+		brelse(sbp);
 }
 
 /*
@@ -4378,18 +8850,17 @@ softdep_disk_write_complete(bp)
  * splbio interrupts blocked.
  */
 static void 
-handle_allocdirect_partdone(adp)
+handle_allocdirect_partdone(adp, wkhd)
 	struct allocdirect *adp;	/* the completed allocdirect */
+	struct workhead *wkhd;		/* Work to do when inode is writtne. */
 {
 	struct allocdirectlst *listhead;
 	struct allocdirect *listadp;
 	struct inodedep *inodedep;
-	long bsize, delay;
+	long bsize;
 
 	if ((adp->ad_state & ALLCOMPLETE) != ALLCOMPLETE)
 		return;
-	if (adp->ad_buf != NULL)
-		panic("handle_allocdirect_partdone: dangling dep");
 	/*
 	 * The on-disk inode cannot claim to be any larger than the last
 	 * fragment that has been written. Otherwise, the on-disk inode
@@ -4439,25 +8910,27 @@ handle_allocdirect_partdone(adp)
 		return;
 	}
 	/*
-	 * If we have found the just finished dependency, then free
+	 * If we have found the just finished dependency, then queue
 	 * it along with anything that follows it that is complete.
-	 * If the inode still has a bitmap dependency, then it has
-	 * never been written to disk, hence the on-disk inode cannot
-	 * reference the old fragment so we can free it without delay.
+	 * Since the pointer has not yet been written in the inode
+	 * as the dependency prevents it, place the allocdirect on the
+	 * bufwait list where it will be freed once the pointer is
+	 * valid.
 	 */
-	delay = (inodedep->id_state & DEPCOMPLETE);
+	if (wkhd == NULL)
+		wkhd = &inodedep->id_bufwait;
 	for (; adp; adp = listadp) {
 		listadp = TAILQ_NEXT(adp, ad_next);
 		if ((adp->ad_state & ALLCOMPLETE) != ALLCOMPLETE)
 			return;
-		free_allocdirect(listhead, adp, delay);
+		TAILQ_REMOVE(listhead, adp, ad_next);
+		WORKLIST_INSERT(wkhd, &adp->ad_block.nb_list);
 	}
 }
 
 /*
- * Called from within softdep_disk_write_complete above. Note that
- * this routine is always called from interrupt level with further
- * splbio interrupts blocked.
+ * Called from within softdep_disk_write_complete above.  This routine
+ * completes successfully written allocindirs.
  */
 static void
 handle_allocindir_partdone(aip)
@@ -4467,11 +8940,9 @@ handle_allocindir_partdone(aip)
 
 	if ((aip->ai_state & ALLCOMPLETE) != ALLCOMPLETE)
 		return;
-	if (aip->ai_buf != NULL)
-		panic("handle_allocindir_partdone: dangling dependency");
 	indirdep = aip->ai_indirdep;
+	LIST_REMOVE(aip, ai_next);
 	if (indirdep->ir_state & UNDONE) {
-		LIST_REMOVE(aip, ai_next);
 		LIST_INSERT_HEAD(&indirdep->ir_donehd, aip, ai_next);
 		return;
 	}
@@ -4481,12 +8952,129 @@ handle_allocindir_partdone(aip)
 	else
 		((ufs2_daddr_t *)indirdep->ir_savebp->b_data)[aip->ai_offset] =
 		    aip->ai_newblkno;
-	LIST_REMOVE(aip, ai_next);
-	if (aip->ai_freefrag != NULL)
-		add_to_worklist(&aip->ai_freefrag->ff_list);
-	WORKITEM_FREE(aip, D_ALLOCINDIR);
+	/*
+	 * Await the pointer write before freeing the allocindir.
+	 */
+	LIST_INSERT_HEAD(&indirdep->ir_writehd, aip, ai_next);
 }
 
+/*
+ * Release segments held on a jwork list.
+ */
+static void
+handle_jwork(wkhd)
+	struct workhead *wkhd;
+{
+	struct worklist *wk;
+
+	while ((wk = LIST_FIRST(wkhd)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		switch (wk->wk_type) {
+		case D_JSEGDEP:
+			free_jsegdep(WK_JSEGDEP(wk));
+			continue;
+		default:
+			panic("handle_jwork: Unknown type %s\n",
+			    TYPENAME(wk->wk_type));
+		}
+	}
+}
+
+/*
+ * Handle the bufwait list on an inode when it is safe to release items
+ * held there.  This normally happens after an inode block is written but
+ * may be delayed and handle later if there are pending journal items that
+ * are not yet safe to be released.
+ */
+static struct freefile *
+handle_bufwait(inodedep, refhd)
+	struct inodedep *inodedep;
+	struct workhead *refhd;
+{
+	struct jaddref *jaddref;
+	struct freefile *freefile;
+	struct worklist *wk;
+
+	freefile = NULL;
+	while ((wk = LIST_FIRST(&inodedep->id_bufwait)) != NULL) {
+		WORKLIST_REMOVE(wk);
+		switch (wk->wk_type) {
+		case D_FREEFILE:
+			/*
+			 * We defer adding freefile to the worklist
+			 * until all other additions have been made to
+			 * ensure that it will be done after all the
+			 * old blocks have been freed.
+			 */
+			if (freefile != NULL)
+				panic("handle_bufwait: freefile");
+			freefile = WK_FREEFILE(wk);
+			continue;
+
+		case D_MKDIR:
+			handle_written_mkdir(WK_MKDIR(wk), MKDIR_PARENT);
+			continue;
+
+		case D_DIRADD:
+			diradd_inode_written(WK_DIRADD(wk), inodedep);
+			continue;
+
+		case D_FREEFRAG:
+			wk->wk_state |= COMPLETE;
+			if ((wk->wk_state & ALLCOMPLETE) == ALLCOMPLETE)
+				add_to_worklist(wk, 0);
+			continue;
+
+		case D_DIRREM:
+			wk->wk_state |= COMPLETE;
+			add_to_worklist(wk, 0);
+			continue;
+
+		case D_ALLOCDIRECT:
+		case D_ALLOCINDIR:
+			free_newblk(WK_NEWBLK(wk));
+			continue;
+
+		case D_JNEWBLK:
+			wk->wk_state |= COMPLETE;
+			free_jnewblk(WK_JNEWBLK(wk));
+			continue;
+
+		/*
+		 * Save freed journal segments and add references on
+		 * the supplied list which will delay their release
+		 * until the cg bitmap is cleared on disk.
+		 */
+		case D_JSEGDEP:
+			if (refhd == NULL)
+				free_jsegdep(WK_JSEGDEP(wk));
+			else
+				WORKLIST_INSERT(refhd, wk);
+			continue;
+
+		case D_JADDREF:
+			jaddref = WK_JADDREF(wk);
+			TAILQ_REMOVE(&inodedep->id_inoreflst, &jaddref->ja_ref,
+			    if_deps);
+			/*
+			 * Transfer any jaddrefs to the list to be freed with
+			 * the bitmap if we're handling a removed file.
+			 */
+			if (refhd == NULL) {
+				wk->wk_state |= COMPLETE;
+				free_jaddref(jaddref);
+			} else
+				WORKLIST_INSERT(refhd, wk);
+			continue;
+
+		default:
+			panic("handle_bufwait: Unknown type %p(%s)",
+			    wk, TYPENAME(wk->wk_type));
+			/* NOTREACHED */
+		}
+	}
+	return (freefile);
+}
 /*
  * Called from within softdep_disk_write_complete above to restore
  * in-memory inode block contents to their most up-to-date state. Note
@@ -4498,12 +9086,17 @@ handle_written_inodeblock(inodedep, bp)
 	struct inodedep *inodedep;
 	struct buf *bp;		/* buffer containing the inode block */
 {
-	struct worklist *wk, *filefree;
+	struct freefile *freefile;
 	struct allocdirect *adp, *nextadp;
 	struct ufs1_dinode *dp1 = NULL;
 	struct ufs2_dinode *dp2 = NULL;
+	struct workhead wkhd;
 	int hadchanges, fstype;
+	ino_t freelink;
 
+	LIST_INIT(&wkhd);
+	hadchanges = 0;
+	freefile = NULL;
 	if ((inodedep->id_state & IOSTARTED) == 0)
 		panic("handle_written_inodeblock: not started");
 	inodedep->id_state &= ~IOSTARTED;
@@ -4511,11 +9104,32 @@ handle_written_inodeblock(inodedep, bp)
 		fstype = UFS1;
 		dp1 = (struct ufs1_dinode *)bp->b_data +
 		    ino_to_fsbo(inodedep->id_fs, inodedep->id_ino);
+		freelink = dp1->di_freelink;
 	} else {
 		fstype = UFS2;
 		dp2 = (struct ufs2_dinode *)bp->b_data +
 		    ino_to_fsbo(inodedep->id_fs, inodedep->id_ino);
+		freelink = dp2->di_freelink;
 	}
+	/*
+	 * If we wrote a valid freelink pointer during the last write
+	 * record it here.
+	 */
+	if ((inodedep->id_state & (UNLINKED | UNLINKNEXT)) == UNLINKED) {
+		struct inodedep *inon;
+
+		inon = TAILQ_NEXT(inodedep, id_unlinked);
+		if ((inon == NULL && freelink == 0) ||
+		    (inon && inon->id_ino == freelink)) {
+			if (inon)
+				inon->id_state |= UNLINKPREV;
+			inodedep->id_state |= UNLINKNEXT;
+		} else
+			hadchanges = 1;
+	}
+	/* Leave this inodeblock dirty until it's in the list. */
+	if ((inodedep->id_state & (UNLINKED | DEPCOMPLETE)) == UNLINKED)
+		hadchanges = 1;
 	/*
 	 * If we had to rollback the inode allocation because of
 	 * bitmaps being incomplete, then simply restore it.
@@ -4524,6 +9138,7 @@ handle_written_inodeblock(inodedep, bp)
 	 * corresponding updates written to disk.
 	 */
 	if (inodedep->id_savedino1 != NULL) {
+		hadchanges = 1;
 		if (fstype == UFS1)
 			*dp1 = *inodedep->id_savedino1;
 		else
@@ -4533,6 +9148,13 @@ handle_written_inodeblock(inodedep, bp)
 		if ((bp->b_flags & B_DELWRI) == 0)
 			stat_inode_bitmap++;
 		bdirty(bp);
+		/*
+		 * If the inode is clear here and GOINGAWAY it will never
+		 * be written.  Process the bufwait and clear any pending
+		 * work which may include the freefile.
+		 */
+		if (inodedep->id_state & GOINGAWAY)
+			goto bufwait;
 		return (1);
 	}
 	inodedep->id_state |= COMPLETE;
@@ -4540,50 +9162,49 @@ handle_written_inodeblock(inodedep, bp)
 	 * Roll forward anything that had to be rolled back before 
 	 * the inode could be updated.
 	 */
-	hadchanges = 0;
 	for (adp = TAILQ_FIRST(&inodedep->id_inoupdt); adp; adp = nextadp) {
 		nextadp = TAILQ_NEXT(adp, ad_next);
 		if (adp->ad_state & ATTACHED)
 			panic("handle_written_inodeblock: new entry");
 		if (fstype == UFS1) {
-			if (adp->ad_lbn < NDADDR) {
-				if (dp1->di_db[adp->ad_lbn]!=adp->ad_oldblkno)
+			if (adp->ad_offset < NDADDR) {
+				if (dp1->di_db[adp->ad_offset]!=adp->ad_oldblkno)
 					panic("%s %s #%jd mismatch %d != %jd",
 					    "handle_written_inodeblock:",
 					    "direct pointer",
-					    (intmax_t)adp->ad_lbn,
-					    dp1->di_db[adp->ad_lbn],
+					    (intmax_t)adp->ad_offset,
+					    dp1->di_db[adp->ad_offset],
 					    (intmax_t)adp->ad_oldblkno);
-				dp1->di_db[adp->ad_lbn] = adp->ad_newblkno;
+				dp1->di_db[adp->ad_offset] = adp->ad_newblkno;
 			} else {
-				if (dp1->di_ib[adp->ad_lbn - NDADDR] != 0)
+				if (dp1->di_ib[adp->ad_offset - NDADDR] != 0)
 					panic("%s: %s #%jd allocated as %d",
 					    "handle_written_inodeblock",
 					    "indirect pointer",
-					    (intmax_t)adp->ad_lbn - NDADDR,
-					    dp1->di_ib[adp->ad_lbn - NDADDR]);
-				dp1->di_ib[adp->ad_lbn - NDADDR] =
+					    (intmax_t)adp->ad_offset - NDADDR,
+					    dp1->di_ib[adp->ad_offset - NDADDR]);
+				dp1->di_ib[adp->ad_offset - NDADDR] =
 				    adp->ad_newblkno;
 			}
 		} else {
-			if (adp->ad_lbn < NDADDR) {
-				if (dp2->di_db[adp->ad_lbn]!=adp->ad_oldblkno)
+			if (adp->ad_offset < NDADDR) {
+				if (dp2->di_db[adp->ad_offset]!=adp->ad_oldblkno)
 					panic("%s: %s #%jd %s %jd != %jd",
 					    "handle_written_inodeblock",
 					    "direct pointer",
-					    (intmax_t)adp->ad_lbn, "mismatch",
-					    (intmax_t)dp2->di_db[adp->ad_lbn],
+					    (intmax_t)adp->ad_offset, "mismatch",
+					    (intmax_t)dp2->di_db[adp->ad_offset],
 					    (intmax_t)adp->ad_oldblkno);
-				dp2->di_db[adp->ad_lbn] = adp->ad_newblkno;
+				dp2->di_db[adp->ad_offset] = adp->ad_newblkno;
 			} else {
-				if (dp2->di_ib[adp->ad_lbn - NDADDR] != 0)
+				if (dp2->di_ib[adp->ad_offset - NDADDR] != 0)
 					panic("%s: %s #%jd allocated as %jd",
 					    "handle_written_inodeblock",
 					    "indirect pointer",
-					    (intmax_t)adp->ad_lbn - NDADDR,
+					    (intmax_t)adp->ad_offset - NDADDR,
 					    (intmax_t)
-					    dp2->di_ib[adp->ad_lbn - NDADDR]);
-				dp2->di_ib[adp->ad_lbn - NDADDR] =
+					    dp2->di_ib[adp->ad_offset - NDADDR]);
+				dp2->di_ib[adp->ad_offset - NDADDR] =
 				    adp->ad_newblkno;
 			}
 		}
@@ -4595,13 +9216,13 @@ handle_written_inodeblock(inodedep, bp)
 		nextadp = TAILQ_NEXT(adp, ad_next);
 		if (adp->ad_state & ATTACHED)
 			panic("handle_written_inodeblock: new entry");
-		if (dp2->di_extb[adp->ad_lbn] != adp->ad_oldblkno)
+		if (dp2->di_extb[adp->ad_offset] != adp->ad_oldblkno)
 			panic("%s: direct pointers #%jd %s %jd != %jd",
 			    "handle_written_inodeblock",
-			    (intmax_t)adp->ad_lbn, "mismatch",
-			    (intmax_t)dp2->di_extb[adp->ad_lbn],
+			    (intmax_t)adp->ad_offset, "mismatch",
+			    (intmax_t)dp2->di_extb[adp->ad_offset],
 			    (intmax_t)adp->ad_oldblkno);
-		dp2->di_extb[adp->ad_lbn] = adp->ad_newblkno;
+		dp2->di_extb[adp->ad_offset] = adp->ad_newblkno;
 		adp->ad_state &= ~UNDONE;
 		adp->ad_state |= ATTACHED;
 		hadchanges = 1;
@@ -4613,12 +9234,23 @@ handle_written_inodeblock(inodedep, bp)
 	 */
 	if (inodedep->id_savedsize == -1 || inodedep->id_savedextsize == -1)
 		panic("handle_written_inodeblock: bad size");
+	if (inodedep->id_savednlink > LINK_MAX)
+		panic("handle_written_inodeblock: Invalid link count "
+		    "%d for inodedep %p", inodedep->id_savednlink, inodedep);
 	if (fstype == UFS1) {
+		if (dp1->di_nlink != inodedep->id_savednlink) { 
+			dp1->di_nlink = inodedep->id_savednlink;
+			hadchanges = 1;
+		}
 		if (dp1->di_size != inodedep->id_savedsize) {
 			dp1->di_size = inodedep->id_savedsize;
 			hadchanges = 1;
 		}
 	} else {
+		if (dp2->di_nlink != inodedep->id_savednlink) { 
+			dp2->di_nlink = inodedep->id_savednlink;
+			hadchanges = 1;
+		}
 		if (dp2->di_size != inodedep->id_savedsize) {
 			dp2->di_size = inodedep->id_savedsize;
 			hadchanges = 1;
@@ -4630,6 +9262,7 @@ handle_written_inodeblock(inodedep, bp)
 	}
 	inodedep->id_savedsize = -1;
 	inodedep->id_savedextsize = -1;
+	inodedep->id_savednlink = -1;
 	/*
 	 * If there were any rollbacks in the inode block, then it must be
 	 * marked dirty so that its will eventually get written back in
@@ -4637,69 +9270,49 @@ handle_written_inodeblock(inodedep, bp)
 	 */
 	if (hadchanges)
 		bdirty(bp);
+bufwait:
 	/*
 	 * Process any allocdirects that completed during the update.
 	 */
 	if ((adp = TAILQ_FIRST(&inodedep->id_inoupdt)) != NULL)
-		handle_allocdirect_partdone(adp);
+		handle_allocdirect_partdone(adp, &wkhd);
 	if ((adp = TAILQ_FIRST(&inodedep->id_extupdt)) != NULL)
-		handle_allocdirect_partdone(adp);
+		handle_allocdirect_partdone(adp, &wkhd);
 	/*
 	 * Process deallocations that were held pending until the
 	 * inode had been written to disk. Freeing of the inode
 	 * is delayed until after all blocks have been freed to
 	 * avoid creation of new <vfsid, inum, lbn> triples
-	 * before the old ones have been deleted.
+	 * before the old ones have been deleted.  Completely
+	 * unlinked inodes are not processed until the unlinked
+	 * inode list is written or the last reference is removed.
 	 */
-	filefree = NULL;
-	while ((wk = LIST_FIRST(&inodedep->id_bufwait)) != NULL) {
-		WORKLIST_REMOVE(wk);
-		switch (wk->wk_type) {
-
-		case D_FREEFILE:
-			/*
-			 * We defer adding filefree to the worklist until
-			 * all other additions have been made to ensure
-			 * that it will be done after all the old blocks
-			 * have been freed.
-			 */
-			if (filefree != NULL)
-				panic("handle_written_inodeblock: filefree");
-			filefree = wk;
-			continue;
-
-		case D_MKDIR:
-			handle_written_mkdir(WK_MKDIR(wk), MKDIR_PARENT);
-			continue;
-
-		case D_DIRADD:
-			diradd_inode_written(WK_DIRADD(wk), inodedep);
-			continue;
-
-		case D_FREEBLKS:
-			wk->wk_state |= COMPLETE;
-			if ((wk->wk_state  & ALLCOMPLETE) != ALLCOMPLETE)
-				continue;
-			 /* -- fall through -- */
-		case D_FREEFRAG:
-		case D_DIRREM:
-			add_to_worklist(wk);
-			continue;
-
-		case D_NEWDIRBLK:
-			free_newdirblk(WK_NEWDIRBLK(wk));
-			continue;
-
-		default:
-			panic("handle_written_inodeblock: Unknown type %s",
-			    TYPENAME(wk->wk_type));
-			/* NOTREACHED */
+	if ((inodedep->id_state & (UNLINKED | UNLINKONLIST)) != UNLINKED) {
+		freefile = handle_bufwait(inodedep, NULL);
+		if (freefile && !LIST_EMPTY(&wkhd)) {
+			WORKLIST_INSERT(&wkhd, &freefile->fx_list);
+			freefile = NULL;
 		}
 	}
-	if (filefree != NULL) {
+	/*
+	 * Move rolled forward dependency completions to the bufwait list
+	 * now that those that were already written have been processed.
+	 */
+	if (!LIST_EMPTY(&wkhd) && hadchanges == 0)
+		panic("handle_written_inodeblock: bufwait but no changes");
+	jwork_move(&inodedep->id_bufwait, &wkhd);
+
+	if (freefile != NULL) {
+		/*
+		 * If the inode is goingaway it was never written.  Fake up
+		 * the state here so free_inodedep() can succeed.
+		 */
+		if (inodedep->id_state & GOINGAWAY)
+			inodedep->id_state |= COMPLETE | DEPCOMPLETE;
 		if (free_inodedep(inodedep) == 0)
-			panic("handle_written_inodeblock: live inodedep");
-		add_to_worklist(filefree);
+			panic("handle_written_inodeblock: live inodedep %p",
+			    inodedep);
+		add_to_worklist(&freefile->fx_list, 0);
 		return (0);
 	}
 
@@ -4707,12 +9320,101 @@ handle_written_inodeblock(inodedep, bp)
 	 * If no outstanding dependencies, free it.
 	 */
 	if (free_inodedep(inodedep) ||
-	    (TAILQ_FIRST(&inodedep->id_inoupdt) == 0 &&
-	     TAILQ_FIRST(&inodedep->id_extupdt) == 0))
+	    (TAILQ_FIRST(&inodedep->id_inoreflst) == 0 &&
+	     TAILQ_FIRST(&inodedep->id_inoupdt) == 0 &&
+	     TAILQ_FIRST(&inodedep->id_extupdt) == 0 &&
+	     LIST_FIRST(&inodedep->id_bufwait) == 0))
 		return (0);
 	return (hadchanges);
 }
 
+static int
+handle_written_indirdep(indirdep, bp, bpp)
+	struct indirdep *indirdep;
+	struct buf *bp;
+	struct buf **bpp;
+{
+	struct allocindir *aip;
+	int chgs;
+
+	if (indirdep->ir_state & GOINGAWAY)
+		panic("disk_write_complete: indirdep gone");
+	chgs = 0;
+	/*
+	 * If there were rollbacks revert them here.
+	 */
+	if (indirdep->ir_saveddata) {
+		bcopy(indirdep->ir_saveddata, bp->b_data, bp->b_bcount);
+		free(indirdep->ir_saveddata, M_INDIRDEP);
+		indirdep->ir_saveddata = 0;
+		chgs = 1;
+	}
+	indirdep->ir_state &= ~UNDONE;
+	indirdep->ir_state |= ATTACHED;
+	/*
+	 * Move allocindirs with written pointers to the completehd if
+	 * the the indirdep's pointer is not yet written.  Otherwise
+	 * free them here.
+	 */
+	while ((aip = LIST_FIRST(&indirdep->ir_writehd)) != 0) {
+		LIST_REMOVE(aip, ai_next);
+		if ((indirdep->ir_state & DEPCOMPLETE) == 0) {
+			LIST_INSERT_HEAD(&indirdep->ir_completehd, aip,
+			    ai_next);
+			continue;
+		}
+		free_newblk(&aip->ai_block);
+	}
+	/*
+	 * Move allocindirs that have finished dependency processing from
+	 * the done list to the write list after updating the pointers.
+	 */
+	while ((aip = LIST_FIRST(&indirdep->ir_donehd)) != 0) {
+		handle_allocindir_partdone(aip);
+		if (aip == LIST_FIRST(&indirdep->ir_donehd))
+			panic("disk_write_complete: not gone");
+		chgs = 1;
+	}
+	/*
+	 * If this indirdep has been detached from its newblk during
+	 * I/O we need to keep this dep attached to the buffer so
+	 * deallocate_dependencies can find it and properly resolve
+	 * any outstanding dependencies.
+	 */
+	if ((indirdep->ir_state & (ONDEPLIST | DEPCOMPLETE)) == 0)
+		chgs = 1;
+	if ((bp->b_flags & B_DELWRI) == 0)
+		stat_indir_blk_ptrs++;
+	/*
+	 * If there were no changes we can discard the savedbp and detach
+	 * ourselves from the buf.  We are only carrying completed pointers
+	 * in this case.
+	 */
+	if (chgs == 0) {
+		struct buf *sbp;
+
+		sbp = indirdep->ir_savebp;
+		sbp->b_flags |= B_INVAL | B_NOCACHE;
+		indirdep->ir_savebp = NULL;
+		if (*bpp != NULL)
+			panic("handle_written_indirdep: bp already exists.");
+		*bpp = sbp;
+	} else
+		bdirty(bp);
+	/*
+	 * If there are no fresh dependencies and none waiting on writes
+	 * we can free the indirdep. 
+	 */
+	if ((indirdep->ir_state & DEPCOMPLETE) && chgs == 0) {
+		if (indirdep->ir_state & ONDEPLIST)
+			LIST_REMOVE(indirdep, ir_next);
+		free_indirdep(indirdep);
+		return (0);
+	}
+
+	return (chgs);
+}
+
 /*
  * Process a diradd entry after its dependent inode has been written.
  * This routine must be called with splbio interrupts blocked.
@@ -4722,20 +9424,163 @@ diradd_inode_written(dap, inodedep)
 	struct diradd *dap;
 	struct inodedep *inodedep;
 {
-	struct pagedep *pagedep;
 
 	dap->da_state |= COMPLETE;
-	if ((dap->da_state & ALLCOMPLETE) == ALLCOMPLETE) {
-		if (dap->da_state & DIRCHG)
-			pagedep = dap->da_previous->dm_pagedep;
-		else
-			pagedep = dap->da_pagedep;
-		LIST_REMOVE(dap, da_pdlist);
-		LIST_INSERT_HEAD(&pagedep->pd_pendinghd, dap, da_pdlist);
-	}
+	complete_diradd(dap);
 	WORKLIST_INSERT(&inodedep->id_pendinghd, &dap->da_list);
 }
 
+/*
+ * Returns true if the bmsafemap will have rollbacks when written.  Must
+ * only be called with lk and the buf lock on the cg held.
+ */
+static int
+bmsafemap_rollbacks(bmsafemap)
+	struct bmsafemap *bmsafemap;
+{
+
+	return (!LIST_EMPTY(&bmsafemap->sm_jaddrefhd) | 
+	    !LIST_EMPTY(&bmsafemap->sm_jnewblkhd));
+}
+
+/*
+ * Complete a write to a bmsafemap structure.  Roll forward any bitmap
+ * changes if it's not a background write.  Set all written dependencies 
+ * to DEPCOMPLETE and free the structure if possible.
+ */
+static int
+handle_written_bmsafemap(bmsafemap, bp)
+	struct bmsafemap *bmsafemap;
+	struct buf *bp;
+{
+	struct newblk *newblk;
+	struct inodedep *inodedep;
+	struct jaddref *jaddref, *jatmp;
+	struct jnewblk *jnewblk, *jntmp;
+	uint8_t *inosused;
+	uint8_t *blksfree;
+	struct cg *cgp;
+	struct fs *fs;
+	ino_t ino;
+	long bno;
+	int chgs;
+	int i;
+
+	if ((bmsafemap->sm_state & IOSTARTED) == 0)
+		panic("initiate_write_bmsafemap: Not started\n");
+	chgs = 0;
+	bmsafemap->sm_state &= ~IOSTARTED;
+	/*
+	 * Restore unwritten inode allocation pending jaddref writes.
+	 */
+	if (!LIST_EMPTY(&bmsafemap->sm_jaddrefhd)) {
+		cgp = (struct cg *)bp->b_data;
+		fs = VFSTOUFS(bmsafemap->sm_list.wk_mp)->um_fs;
+		inosused = cg_inosused(cgp);
+		LIST_FOREACH_SAFE(jaddref, &bmsafemap->sm_jaddrefhd,
+		    ja_bmdeps, jatmp) {
+			if ((jaddref->ja_state & UNDONE) == 0)
+				continue;
+			ino = jaddref->ja_ino % fs->fs_ipg;
+			if (isset(inosused, ino))
+				panic("handle_written_bmsafemap: "
+				    "re-allocated inode");
+			if ((bp->b_xflags & BX_BKGRDMARKER) == 0) {
+				if ((jaddref->ja_mode & IFMT) == IFDIR)
+					cgp->cg_cs.cs_ndir++;
+				cgp->cg_cs.cs_nifree--;
+				setbit(inosused, ino);
+				chgs = 1;
+			}
+			jaddref->ja_state &= ~UNDONE;
+			jaddref->ja_state |= ATTACHED;
+			free_jaddref(jaddref);
+		}
+	}
+	/*
+	 * Restore any block allocations which are pending journal writes.
+	 */
+	if (LIST_FIRST(&bmsafemap->sm_jnewblkhd) != NULL) {
+		cgp = (struct cg *)bp->b_data;
+		fs = VFSTOUFS(bmsafemap->sm_list.wk_mp)->um_fs;
+		blksfree = cg_blksfree(cgp);
+		LIST_FOREACH_SAFE(jnewblk, &bmsafemap->sm_jnewblkhd, jn_deps,
+		    jntmp) {
+			if ((jnewblk->jn_state & UNDONE) == 0)
+				continue;
+			bno = dtogd(fs, jnewblk->jn_blkno);
+			for (i = jnewblk->jn_oldfrags; i < jnewblk->jn_frags;
+			    i++) {
+				if (bp->b_xflags & BX_BKGRDMARKER)
+					break;
+				if ((jnewblk->jn_state & NEWBLOCK) == 0 &&
+				    isclr(blksfree, bno + i))
+					panic("handle_written_bmsafemap: "
+					    "re-allocated fragment");
+				clrbit(blksfree, bno + i);
+				chgs = 1;
+			}
+			jnewblk->jn_state &= ~(UNDONE | NEWBLOCK);
+			jnewblk->jn_state |= ATTACHED;
+			free_jnewblk(jnewblk);
+		}
+	}
+	while ((newblk = LIST_FIRST(&bmsafemap->sm_newblkwr))) {
+		newblk->nb_state |= DEPCOMPLETE;
+		newblk->nb_state &= ~ONDEPLIST;
+		newblk->nb_bmsafemap = NULL;
+		LIST_REMOVE(newblk, nb_deps);
+		if (newblk->nb_list.wk_type == D_ALLOCDIRECT)
+			handle_allocdirect_partdone(
+			    WK_ALLOCDIRECT(&newblk->nb_list), NULL);
+		else if (newblk->nb_list.wk_type == D_ALLOCINDIR)
+			handle_allocindir_partdone(
+			    WK_ALLOCINDIR(&newblk->nb_list));
+		else if (newblk->nb_list.wk_type != D_NEWBLK)
+			panic("handle_written_bmsafemap: Unexpected type: %s",
+			    TYPENAME(newblk->nb_list.wk_type));
+	}
+	while ((inodedep = LIST_FIRST(&bmsafemap->sm_inodedepwr)) != NULL) {
+		inodedep->id_state |= DEPCOMPLETE;
+		inodedep->id_state &= ~ONDEPLIST;
+		LIST_REMOVE(inodedep, id_deps);
+		inodedep->id_bmsafemap = NULL;
+	}
+	if (LIST_EMPTY(&bmsafemap->sm_jaddrefhd) &&
+	    LIST_EMPTY(&bmsafemap->sm_jnewblkhd) &&
+	    LIST_EMPTY(&bmsafemap->sm_newblkhd) &&
+	    LIST_EMPTY(&bmsafemap->sm_inodedephd)) {
+		if (chgs)
+			bdirty(bp);
+		LIST_REMOVE(bmsafemap, sm_hash);
+		WORKITEM_FREE(bmsafemap, D_BMSAFEMAP);
+		return (0);
+	}
+	bdirty(bp);
+	return (1);
+}
+
+/*
+ * Try to free a mkdir dependency.
+ */
+static void
+complete_mkdir(mkdir)
+	struct mkdir *mkdir;
+{
+	struct diradd *dap;
+
+	if ((mkdir->md_state & ALLCOMPLETE) != ALLCOMPLETE)
+		return;
+	LIST_REMOVE(mkdir, md_mkdirs);
+	dap = mkdir->md_diradd;
+	dap->da_state &= ~(mkdir->md_state & (MKDIR_PARENT | MKDIR_BODY));
+	if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) == 0) {
+		dap->da_state |= DEPCOMPLETE;
+		complete_diradd(dap);
+	}
+	WORKITEM_FREE(mkdir, D_MKDIR);
+}
+
 /*
  * Handle the completion of a mkdir dependency.
  */
@@ -4744,25 +9589,32 @@ handle_written_mkdir(mkdir, type)
 	struct mkdir *mkdir;
 	int type;
 {
-	struct diradd *dap;
-	struct pagedep *pagedep;
 
-	if (mkdir->md_state != type)
+	if ((mkdir->md_state & (MKDIR_PARENT | MKDIR_BODY)) != type)
 		panic("handle_written_mkdir: bad type");
-	dap = mkdir->md_diradd;
-	dap->da_state &= ~type;
-	if ((dap->da_state & (MKDIR_PARENT | MKDIR_BODY)) == 0)
-		dap->da_state |= DEPCOMPLETE;
-	if ((dap->da_state & ALLCOMPLETE) == ALLCOMPLETE) {
-		if (dap->da_state & DIRCHG)
-			pagedep = dap->da_previous->dm_pagedep;
-		else
-			pagedep = dap->da_pagedep;
-		LIST_REMOVE(dap, da_pdlist);
-		LIST_INSERT_HEAD(&pagedep->pd_pendinghd, dap, da_pdlist);
-	}
-	LIST_REMOVE(mkdir, md_mkdirs);
-	WORKITEM_FREE(mkdir, D_MKDIR);
+	mkdir->md_state |= COMPLETE;
+	complete_mkdir(mkdir);
+}
+
+static void
+free_pagedep(pagedep)
+	struct pagedep *pagedep;
+{
+	int i;
+
+	if (pagedep->pd_state & (NEWBLOCK | ONWORKLIST))
+		return;
+	for (i = 0; i < DAHASHSZ; i++)
+		if (!LIST_EMPTY(&pagedep->pd_diraddhd[i]))
+			return;
+	if (!LIST_EMPTY(&pagedep->pd_jmvrefhd))
+		return;
+	if (!LIST_EMPTY(&pagedep->pd_dirremhd))
+		return;
+	if (!LIST_EMPTY(&pagedep->pd_pendinghd))
+		return;
+	LIST_REMOVE(pagedep, pd_hash);
+	WORKITEM_FREE(pagedep, D_PAGEDEP);
 }
 
 /*
@@ -4790,8 +9642,11 @@ handle_written_filepage(pagedep, bp)
 	 */
 	while ((dirrem = LIST_FIRST(&pagedep->pd_dirremhd)) != NULL) {
 		LIST_REMOVE(dirrem, dm_next);
+		dirrem->dm_state |= COMPLETE;
 		dirrem->dm_dirinum = pagedep->pd_ino;
-		add_to_worklist(&dirrem->dm_list);
+		KASSERT(LIST_EMPTY(&dirrem->dm_jremrefhd),
+		    ("handle_written_filepage: Journal entries not written."));
+		add_to_worklist(&dirrem->dm_list, 0);
 	}
 	/*
 	 * Free any directory additions that have been committed.
@@ -4800,7 +9655,7 @@ handle_written_filepage(pagedep, bp)
 	 */
 	if ((pagedep->pd_state & NEWBLOCK) == 0)
 		while ((dap = LIST_FIRST(&pagedep->pd_pendinghd)) != NULL)
-			free_diradd(dap);
+			free_diradd(dap, NULL);
 	/*
 	 * Uncommitted directory entries must be restored.
 	 */
@@ -4845,7 +9700,8 @@ handle_written_filepage(pagedep, bp)
 	 * Otherwise it will remain to track any new entries on
 	 * the page in case they are fsync'ed.
 	 */
-	if ((pagedep->pd_state & NEWBLOCK) == 0) {
+	if ((pagedep->pd_state & NEWBLOCK) == 0 &&
+	    LIST_EMPTY(&pagedep->pd_jmvrefhd)) {
 		LIST_REMOVE(pagedep, pd_hash);
 		WORKITEM_FREE(pagedep, D_PAGEDEP);
 	}
@@ -4880,8 +9736,8 @@ softdep_load_inodeblock(ip)
 	 */
 	ip->i_effnlink = ip->i_nlink;
 	ACQUIRE_LOCK(&lk);
-	if (inodedep_lookup(UFSTOVFS(ip->i_ump),
-	    ip->i_number, 0, &inodedep) == 0) {
+	if (inodedep_lookup(UFSTOVFS(ip->i_ump), ip->i_number, 0,
+	    &inodedep) == 0) {
 		FREE_LOCK(&lk);
 		return;
 	}
@@ -4908,11 +9764,26 @@ softdep_update_inodeblock(ip, bp, waitfor)
 	int waitfor;		/* nonzero => update must be allowed */
 {
 	struct inodedep *inodedep;
+	struct inoref *inoref;
 	struct worklist *wk;
 	struct mount *mp;
 	struct buf *ibp;
+	struct fs *fs;
 	int error;
 
+	mp = UFSTOVFS(ip->i_ump);
+	fs = ip->i_fs;
+	/*
+	 * Preserve the freelink that is on disk.  clear_unlinked_inodedep()
+	 * does not have access to the in-core ip so must write directly into
+	 * the inode block buffer when setting freelink.
+	 */
+	if (fs->fs_magic == FS_UFS1_MAGIC)
+		DIP_SET(ip, i_freelink, ((struct ufs1_dinode *)bp->b_data +
+		    ino_to_fsbo(fs, ip->i_number))->di_freelink);
+	else
+		DIP_SET(ip, i_freelink, ((struct ufs2_dinode *)bp->b_data +
+		    ino_to_fsbo(fs, ip->i_number))->di_freelink);
 	/*
 	 * If the effective link count is not equal to the actual link
 	 * count, then we must track the difference in an inodedep while
@@ -4920,8 +9791,8 @@ softdep_update_inodeblock(ip, bp, waitfor)
 	 * if there is no existing inodedep, then there are no dependencies
 	 * to track.
 	 */
-	mp = UFSTOVFS(ip->i_ump);
 	ACQUIRE_LOCK(&lk);
+again:
 	if (inodedep_lookup(mp, ip->i_number, 0, &inodedep) == 0) {
 		FREE_LOCK(&lk);
 		if (ip->i_effnlink != ip->i_nlink)
@@ -4930,6 +9801,20 @@ softdep_update_inodeblock(ip, bp, waitfor)
 	}
 	if (inodedep->id_nlinkdelta != ip->i_nlink - ip->i_effnlink)
 		panic("softdep_update_inodeblock: bad delta");
+	/*
+	 * If we're flushing all dependencies we must also move any waiting
+	 * for journal writes onto the bufwait list prior to I/O.
+	 */
+	if (waitfor) {
+		TAILQ_FOREACH(inoref, &inodedep->id_inoreflst, if_deps) {
+			if ((inoref->if_state & (DEPCOMPLETE | GOINGAWAY))
+			    == DEPCOMPLETE) {
+				stat_jwait_inode++;
+				jwait(&inoref->if_list);
+				goto again;
+			}
+		}
+	}
 	/*
 	 * Changes have been initiated. Anything depending on these
 	 * changes cannot occur until this inode has been written.
@@ -4945,10 +9830,12 @@ softdep_update_inodeblock(ip, bp, waitfor)
 	 */
 	merge_inode_lists(&inodedep->id_newinoupdt, &inodedep->id_inoupdt);
 	if (!TAILQ_EMPTY(&inodedep->id_inoupdt))
-		handle_allocdirect_partdone(TAILQ_FIRST(&inodedep->id_inoupdt));
+		handle_allocdirect_partdone(TAILQ_FIRST(&inodedep->id_inoupdt),
+		    NULL);
 	merge_inode_lists(&inodedep->id_newextupdt, &inodedep->id_extupdt);
 	if (!TAILQ_EMPTY(&inodedep->id_extupdt))
-		handle_allocdirect_partdone(TAILQ_FIRST(&inodedep->id_extupdt));
+		handle_allocdirect_partdone(TAILQ_FIRST(&inodedep->id_extupdt),
+		    NULL);
 	/*
 	 * Now that the inode has been pushed into the buffer, the
 	 * operations dependent on the inode being written to disk
@@ -4971,11 +9858,11 @@ softdep_update_inodeblock(ip, bp, waitfor)
 		return;
 	}
 retry:
-	if ((inodedep->id_state & DEPCOMPLETE) != 0) {
+	if ((inodedep->id_state & (DEPCOMPLETE | GOINGAWAY)) != 0) {
 		FREE_LOCK(&lk);
 		return;
 	}
-	ibp = inodedep->id_buf;
+	ibp = inodedep->id_bmsafemap->sm_buf;
 	ibp = getdirtybuf(ibp, &lk, MNT_WAIT);
 	if (ibp == NULL) {
 		/*
@@ -5007,13 +9894,13 @@ merge_inode_lists(newlisthead, oldlisthead)
 
 	newadp = TAILQ_FIRST(newlisthead);
 	for (listadp = TAILQ_FIRST(oldlisthead); listadp && newadp;) {
-		if (listadp->ad_lbn < newadp->ad_lbn) {
+		if (listadp->ad_offset < newadp->ad_offset) {
 			listadp = TAILQ_NEXT(listadp, ad_next);
 			continue;
 		}
 		TAILQ_REMOVE(newlisthead, newadp, ad_next);
 		TAILQ_INSERT_BEFORE(listadp, newadp, ad_next);
-		if (listadp->ad_lbn == newadp->ad_lbn) {
+		if (listadp->ad_offset == newadp->ad_offset) {
 			allocdirect_merge(oldlisthead, newadp,
 			    listadp);
 			listadp = newadp;
@@ -5036,6 +9923,7 @@ softdep_fsync(vp)
 {
 	struct inodedep *inodedep;
 	struct pagedep *pagedep;
+	struct inoref *inoref;
 	struct worklist *wk;
 	struct diradd *dap;
 	struct mount *mp;
@@ -5052,17 +9940,25 @@ softdep_fsync(vp)
 	fs = ip->i_fs;
 	mp = vp->v_mount;
 	ACQUIRE_LOCK(&lk);
+restart:
 	if (inodedep_lookup(mp, ip->i_number, 0, &inodedep) == 0) {
 		FREE_LOCK(&lk);
 		return (0);
 	}
+	TAILQ_FOREACH(inoref, &inodedep->id_inoreflst, if_deps) {
+		if ((inoref->if_state & (DEPCOMPLETE | GOINGAWAY))
+		    == DEPCOMPLETE) {
+			stat_jwait_inode++;
+			jwait(&inoref->if_list);
+			goto restart;
+		}
+	}
 	if (!LIST_EMPTY(&inodedep->id_inowait) ||
-	    !LIST_EMPTY(&inodedep->id_bufwait) ||
 	    !TAILQ_EMPTY(&inodedep->id_extupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_newextupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_inoupdt) ||
 	    !TAILQ_EMPTY(&inodedep->id_newinoupdt))
-		panic("softdep_fsync: pending ops");
+		panic("softdep_fsync: pending ops %p", inodedep);
 	for (error = 0, flushparent = 0; ; ) {
 		if ((wk = LIST_FIRST(&inodedep->id_pendinghd)) == NULL)
 			break;
@@ -5254,8 +10150,8 @@ int
 softdep_sync_metadata(struct vnode *vp)
 {
 	struct pagedep *pagedep;
-	struct allocdirect *adp;
 	struct allocindir *aip;
+	struct newblk *newblk;
 	struct buf *bp, *nbp;
 	struct worklist *wk;
 	struct bufobj *bo;
@@ -5319,27 +10215,16 @@ loop:
 		switch (wk->wk_type) {
 
 		case D_ALLOCDIRECT:
-			adp = WK_ALLOCDIRECT(wk);
-			if (adp->ad_state & DEPCOMPLETE)
-				continue;
-			nbp = adp->ad_buf;
-			nbp = getdirtybuf(nbp, &lk, waitfor);
-			if (nbp == NULL)
-				continue;
-			FREE_LOCK(&lk);
-			if (waitfor == MNT_NOWAIT) {
-				bawrite(nbp);
-			} else if ((error = bwrite(nbp)) != 0) {
-				break;
-			}
-			ACQUIRE_LOCK(&lk);
-			continue;
-
 		case D_ALLOCINDIR:
-			aip = WK_ALLOCINDIR(wk);
-			if (aip->ai_state & DEPCOMPLETE)
+			newblk = WK_NEWBLK(wk);
+			if (newblk->nb_jnewblk != NULL) {
+				stat_jwait_newblk++;
+				jwait(&newblk->nb_jnewblk->jn_list);
+				goto restart;
+			}
+			if (newblk->nb_state & DEPCOMPLETE)
 				continue;
-			nbp = aip->ai_buf;
+			nbp = newblk->nb_bmsafemap->sm_buf;
 			nbp = getdirtybuf(nbp, &lk, waitfor);
 			if (nbp == NULL)
 				continue;
@@ -5355,10 +10240,17 @@ loop:
 		case D_INDIRDEP:
 		restart:
 
-			LIST_FOREACH(aip, &WK_INDIRDEP(wk)->ir_deplisthd, ai_next) {
-				if (aip->ai_state & DEPCOMPLETE)
+			LIST_FOREACH(aip,
+			    &WK_INDIRDEP(wk)->ir_deplisthd, ai_next) {
+				newblk = (struct newblk *)aip;
+				if (newblk->nb_jnewblk != NULL) {
+					stat_jwait_newblk++;
+					jwait(&newblk->nb_jnewblk->jn_list);
+					goto restart;
+				}
+				if (newblk->nb_state & DEPCOMPLETE)
 					continue;
-				nbp = aip->ai_buf;
+				nbp = newblk->nb_bmsafemap->sm_buf;
 				nbp = getdirtybuf(nbp, &lk, MNT_WAIT);
 				if (nbp == NULL)
 					goto restart;
@@ -5371,14 +10263,6 @@ loop:
 			}
 			continue;
 
-		case D_INODEDEP:
-			if ((error = flush_inodedep_deps(wk->wk_mp,
-			    WK_INODEDEP(wk)->id_ino)) != 0) {
-				FREE_LOCK(&lk);
-				break;
-			}
-			continue;
-
 		case D_PAGEDEP:
 			/*
 			 * We are trying to sync a directory that may
@@ -5400,48 +10284,6 @@ loop:
 			}
 			continue;
 
-		case D_MKDIR:
-			/*
-			 * This case should never happen if the vnode has
-			 * been properly sync'ed. However, if this function
-			 * is used at a place where the vnode has not yet
-			 * been sync'ed, this dependency can show up. So,
-			 * rather than panic, just flush it.
-			 */
-			nbp = WK_MKDIR(wk)->md_buf;
-			nbp = getdirtybuf(nbp, &lk, waitfor);
-			if (nbp == NULL)
-				continue;
-			FREE_LOCK(&lk);
-			if (waitfor == MNT_NOWAIT) {
-				bawrite(nbp);
-			} else if ((error = bwrite(nbp)) != 0) {
-				break;
-			}
-			ACQUIRE_LOCK(&lk);
-			continue;
-
-		case D_BMSAFEMAP:
-			/*
-			 * This case should never happen if the vnode has
-			 * been properly sync'ed. However, if this function
-			 * is used at a place where the vnode has not yet
-			 * been sync'ed, this dependency can show up. So,
-			 * rather than panic, just flush it.
-			 */
-			nbp = WK_BMSAFEMAP(wk)->sm_buf;
-			nbp = getdirtybuf(nbp, &lk, waitfor);
-			if (nbp == NULL)
-				continue;
-			FREE_LOCK(&lk);
-			if (waitfor == MNT_NOWAIT) {
-				bawrite(nbp);
-			} else if ((error = bwrite(nbp)) != 0) {
-				break;
-			}
-			ACQUIRE_LOCK(&lk);
-			continue;
-
 		default:
 			panic("softdep_sync_metadata: Unknown type %s",
 			    TYPENAME(wk->wk_type));
@@ -5489,7 +10331,8 @@ loop:
 	BO_LOCK(bo);
 	drain_output(vp);
 	BO_UNLOCK(bo);
-	return (0);
+	return ffs_update(vp, 1);
+	/* return (0); */
 }
 
 /*
@@ -5502,6 +10345,7 @@ flush_inodedep_deps(mp, ino)
 	ino_t ino;
 {
 	struct inodedep *inodedep;
+	struct inoref *inoref;
 	int error, waitfor;
 
 	/*
@@ -5522,8 +10366,17 @@ flush_inodedep_deps(mp, ino)
 			return (error);
 		FREE_LOCK(&lk);
 		ACQUIRE_LOCK(&lk);
+restart:
 		if (inodedep_lookup(mp, ino, 0, &inodedep) == 0)
 			return (0);
+		TAILQ_FOREACH(inoref, &inodedep->id_inoreflst, if_deps) {
+			if ((inoref->if_state & (DEPCOMPLETE | GOINGAWAY))
+			    == DEPCOMPLETE) {
+				stat_jwait_inode++;
+				jwait(&inoref->if_list);
+				goto restart;
+			}
+		}
 		if (flush_deplist(&inodedep->id_inoupdt, waitfor, &error) ||
 		    flush_deplist(&inodedep->id_newinoupdt, waitfor, &error) ||
 		    flush_deplist(&inodedep->id_extupdt, waitfor, &error) ||
@@ -5555,13 +10408,20 @@ flush_deplist(listhead, waitfor, errorp)
 	int *errorp;
 {
 	struct allocdirect *adp;
+	struct newblk *newblk;
 	struct buf *bp;
 
 	mtx_assert(&lk, MA_OWNED);
 	TAILQ_FOREACH(adp, listhead, ad_next) {
-		if (adp->ad_state & DEPCOMPLETE)
+		newblk = (struct newblk *)adp;
+		if (newblk->nb_jnewblk != NULL) {
+			stat_jwait_newblk++;
+			jwait(&newblk->nb_jnewblk->jn_list);
+			return (1);
+		}
+		if (newblk->nb_state & DEPCOMPLETE)
 			continue;
-		bp = adp->ad_buf;
+		bp = newblk->nb_bmsafemap->sm_buf;
 		bp = getdirtybuf(bp, &lk, waitfor);
 		if (bp == NULL) {
 			if (waitfor == MNT_NOWAIT)
@@ -5581,6 +10441,101 @@ flush_deplist(listhead, waitfor, errorp)
 	return (0);
 }
 
+/*
+ * Flush dependencies associated with an allocdirect block.
+ */
+static int
+flush_newblk_dep(vp, mp, lbn)
+	struct vnode *vp;
+	struct mount *mp;
+	ufs_lbn_t lbn;
+{
+	struct newblk *newblk;
+	struct bufobj *bo;
+	struct inode *ip;
+	struct buf *bp;
+	ufs2_daddr_t blkno;
+	int error;
+
+	error = 0;
+	bo = &vp->v_bufobj;
+	ip = VTOI(vp);
+	blkno = DIP(ip, i_db[lbn]);
+	if (blkno == 0)
+		panic("flush_newblk_dep: Missing block");
+	ACQUIRE_LOCK(&lk);
+	/*
+	 * Loop until all dependencies related to this block are satisfied.
+	 * We must be careful to restart after each sleep in case a write
+	 * completes some part of this process for us.
+	 */
+	for (;;) {
+		if (newblk_lookup(mp, blkno, 0, &newblk) == 0) {
+			FREE_LOCK(&lk);
+			break;
+		}
+		if (newblk->nb_list.wk_type != D_ALLOCDIRECT)
+			panic("flush_newblk_deps: Bad newblk %p", newblk);
+		/*
+		 * Flush the journal.
+		 */
+		if (newblk->nb_jnewblk != NULL) {
+			stat_jwait_newblk++;
+			jwait(&newblk->nb_jnewblk->jn_list);
+			continue;
+		}
+		/*
+		 * Write the bitmap dependency.
+		 */
+		if ((newblk->nb_state & DEPCOMPLETE) == 0) {
+			bp = newblk->nb_bmsafemap->sm_buf;
+			bp = getdirtybuf(bp, &lk, MNT_WAIT);
+			if (bp == NULL)
+				continue;
+			FREE_LOCK(&lk);
+			error = bwrite(bp);
+			if (error)
+				break;
+			ACQUIRE_LOCK(&lk);
+			continue;
+		}
+		/*
+		 * Write the buffer.
+		 */
+		FREE_LOCK(&lk);
+		BO_LOCK(bo);
+		bp = gbincore(bo, lbn);
+		if (bp != NULL) {
+			error = BUF_LOCK(bp, LK_EXCLUSIVE | LK_SLEEPFAIL |
+			    LK_INTERLOCK, BO_MTX(bo));
+			if (error == ENOLCK) {
+				ACQUIRE_LOCK(&lk);
+				continue; /* Slept, retry */
+			}
+			if (error != 0)
+				break;	/* Failed */
+			if (bp->b_flags & B_DELWRI) {
+				bremfree(bp);
+				error = bwrite(bp);
+				if (error)
+					break;
+			} else
+				BUF_UNLOCK(bp);
+		} else
+			BO_UNLOCK(bo);
+		/*
+		 * We have to wait for the direct pointers to
+		 * point at the newdirblk before the dependency
+		 * will go away.
+		 */
+		error = ffs_update(vp, MNT_WAIT);
+		if (error)
+			break;
+		ACQUIRE_LOCK(&lk);
+	}
+	return (error);
+}
+
 /*
  * Eliminate a pagedep dependency by flushing out all its diradd dependencies.
  * Called with splbio blocked.
@@ -5592,16 +10547,16 @@ flush_pagedep_deps(pvp, mp, diraddhdp)
 	struct diraddhd *diraddhdp;
 {
 	struct inodedep *inodedep;
+	struct inoref *inoref;
 	struct ufsmount *ump;
 	struct diradd *dap;
 	struct vnode *vp;
-	struct bufobj *bo;
 	int error = 0;
 	struct buf *bp;
 	ino_t inum;
-	struct worklist *wk;
 
 	ump = VFSTOUFS(mp);
+restart:
 	while ((dap = LIST_FIRST(diraddhdp)) != NULL) {
 		/*
 		 * Flush ourselves if this directory entry
@@ -5609,7 +10564,7 @@ flush_pagedep_deps(pvp, mp, diraddhdp)
 		 */
 		if (dap->da_state & MKDIR_PARENT) {
 			FREE_LOCK(&lk);
-			if ((error = ffs_update(pvp, 1)) != 0)
+			if ((error = ffs_update(pvp, MNT_WAIT)) != 0)
 				break;
 			ACQUIRE_LOCK(&lk);
 			/*
@@ -5623,84 +10578,52 @@ flush_pagedep_deps(pvp, mp, diraddhdp)
 		/*
 		 * A newly allocated directory must have its "." and
 		 * ".." entries written out before its name can be
-		 * committed in its parent. We do not want or need
-		 * the full semantics of a synchronous ffs_syncvnode as
-		 * that may end up here again, once for each directory
-		 * level in the filesystem. Instead, we push the blocks
-		 * and wait for them to clear. We have to fsync twice
-		 * because the first call may choose to defer blocks
-		 * that still have dependencies, but deferral will
-		 * happen at most once.
+		 * committed in its parent. 
 		 */
 		inum = dap->da_newinum;
+		if (inodedep_lookup(UFSTOVFS(ump), inum, 0, &inodedep) == 0)
+			panic("flush_pagedep_deps: lost inode1");
+		/*
+		 * Wait for any pending journal adds to complete so we don't
+		 * cause rollbacks while syncing.
+		 */
+		TAILQ_FOREACH(inoref, &inodedep->id_inoreflst, if_deps) {
+			if ((inoref->if_state & (DEPCOMPLETE | GOINGAWAY))
+			    == DEPCOMPLETE) {
+				stat_jwait_inode++;
+				jwait(&inoref->if_list);
+				goto restart;
+			}
+		}
 		if (dap->da_state & MKDIR_BODY) {
 			FREE_LOCK(&lk);
 			if ((error = ffs_vgetf(mp, inum, LK_EXCLUSIVE, &vp,
 			    FFSV_FORCEINSMQ)))
 				break;
-			if ((error=ffs_syncvnode(vp, MNT_NOWAIT)) ||
-			    (error=ffs_syncvnode(vp, MNT_NOWAIT))) {
-				vput(vp);
-				break;
-			}
-			bo = &vp->v_bufobj;
-			BO_LOCK(bo);
-			drain_output(vp);
+			error = flush_newblk_dep(vp, mp, 0);
 			/*
-			 * If first block is still dirty with a D_MKDIR
-			 * dependency then it needs to be written now.
+			 * If we still have the dependency we might need to
+			 * update the vnode to sync the new link count to
+			 * disk.
 			 */
-			for (;;) {
-				error = 0;
-				bp = gbincore(bo, 0);
-				if (bp == NULL)
-					break;	/* First block not present */
-				error = BUF_LOCK(bp,
-						 LK_EXCLUSIVE |
-						 LK_SLEEPFAIL |
-						 LK_INTERLOCK,
-						 BO_MTX(bo));
-				BO_LOCK(bo);
-				if (error == ENOLCK)
-					continue;	/* Slept, retry */
-				if (error != 0)
-					break;		/* Failed */
-				if ((bp->b_flags & B_DELWRI) == 0) {
-					BUF_UNLOCK(bp);
-					break;	/* Buffer not dirty */
-				}
-				for (wk = LIST_FIRST(&bp->b_dep);
-				     wk != NULL;
-				     wk = LIST_NEXT(wk, wk_list))
-					if (wk->wk_type == D_MKDIR)
-						break;
-				if (wk == NULL)
-					BUF_UNLOCK(bp);	/* Dependency gone */
-				else {
-					/*
-					 * D_MKDIR dependency remains,
-					 * must write buffer to stable
-					 * storage.
-					 */
-					BO_UNLOCK(bo);
-					bremfree(bp);
-					error = bwrite(bp);
-					BO_LOCK(bo);
-				}
-				break;
-			}
-			BO_UNLOCK(bo);
+			if (error == 0 && dap == LIST_FIRST(diraddhdp))
+				error = ffs_update(vp, MNT_WAIT);
 			vput(vp);
 			if (error != 0)
-				break;	/* Flushing of first block failed */
+				break;
 			ACQUIRE_LOCK(&lk);
 			/*
 			 * If that cleared dependencies, go on to next.
 			 */
 			if (dap != LIST_FIRST(diraddhdp))
 				continue;
-			if (dap->da_state & MKDIR_BODY)
-				panic("flush_pagedep_deps: MKDIR_BODY");
+			if (dap->da_state & MKDIR_BODY) {
+				inodedep_lookup(UFSTOVFS(ump), inum, 0,
+				    &inodedep);
+				panic("flush_pagedep_deps: MKDIR_BODY "
+				    "inodedep %p dap %p vp %p",
+				    inodedep, dap, vp);
+			}
 		}
 		/*
 		 * Flush the inode on which the directory entry depends.
@@ -5719,8 +10642,8 @@ retry:
 		 * If the inode still has bitmap dependencies,
 		 * push them to disk.
 		 */
-		if ((inodedep->id_state & DEPCOMPLETE) == 0) {
-			bp = inodedep->id_buf;
+		if ((inodedep->id_state & (DEPCOMPLETE | GOINGAWAY)) == 0) {
+			bp = inodedep->id_bmsafemap->sm_buf;
 			bp = getdirtybuf(bp, &lk, MNT_WAIT);
 			if (bp == NULL)
 				goto retry;
@@ -5733,24 +10656,29 @@ retry:
 		}
 		/*
 		 * If the inode is still sitting in a buffer waiting
-		 * to be written, push it to disk.
+		 * to be written or waiting for the link count to be
+		 * adjusted update it here to flush it to disk.
 		 */
-		FREE_LOCK(&lk);
-		if ((error = bread(ump->um_devvp,
-		    fsbtodb(ump->um_fs, ino_to_fsba(ump->um_fs, inum)),
-		    (int)ump->um_fs->fs_bsize, NOCRED, &bp)) != 0) {
-			brelse(bp);
-			break;
+		if (dap == LIST_FIRST(diraddhdp)) {
+			FREE_LOCK(&lk);
+			if ((error = ffs_vgetf(mp, inum, LK_EXCLUSIVE, &vp,
+			    FFSV_FORCEINSMQ)))
+				break;
+			error = ffs_update(vp, MNT_WAIT);
+			vput(vp);
+			if (error)
+				break;
+			ACQUIRE_LOCK(&lk);
 		}
-		if ((error = bwrite(bp)) != 0)
-			break;
-		ACQUIRE_LOCK(&lk);
 		/*
 		 * If we have failed to get rid of all the dependencies
 		 * then something is seriously wrong.
 		 */
-		if (dap == LIST_FIRST(diraddhdp))
-			panic("flush_pagedep_deps: flush failed");
+		if (dap == LIST_FIRST(diraddhdp)) {
+			inodedep_lookup(UFSTOVFS(ump), inum, 0, &inodedep);
+			panic("flush_pagedep_deps: failed to flush " 
+			    "inodedep %p ino %d dap %p", inodedep, inum, dap);
+		}
 	}
 	if (error)
 		ACQUIRE_LOCK(&lk);
@@ -5828,6 +10756,7 @@ softdep_request_cleanup(fs, vp)
 			return (0);
 		UFS_UNLOCK(ump);
 		ACQUIRE_LOCK(&lk);
+		process_removes(vp);
 		if (ump->softdep_on_worklist > 0 &&
 		    process_worklist_item(UFSTOVFS(ump), LK_NOWAIT) != -1) {
 			stat_worklist_push += 1;
@@ -6100,10 +11029,15 @@ softdep_count_dependencies(bp, wantcount)
 	int wantcount;
 {
 	struct worklist *wk;
+	struct bmsafemap *bmsafemap;
 	struct inodedep *inodedep;
 	struct indirdep *indirdep;
+	struct freeblks *freeblks;
 	struct allocindir *aip;
 	struct pagedep *pagedep;
+	struct dirrem *dirrem;
+	struct newblk *newblk;
+	struct mkdir *mkdir;
 	struct diradd *dap;
 	int i, retval;
 
@@ -6132,6 +11066,12 @@ softdep_count_dependencies(bp, wantcount)
 				if (!wantcount)
 					goto out;
 			}
+			if (TAILQ_FIRST(&inodedep->id_inoreflst)) {
+				/* Add reference dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
 			continue;
 
 		case D_INDIRDEP:
@@ -6147,6 +11087,14 @@ softdep_count_dependencies(bp, wantcount)
 
 		case D_PAGEDEP:
 			pagedep = WK_PAGEDEP(wk);
+			LIST_FOREACH(dirrem, &pagedep->pd_dirremhd, dm_next) {
+				if (LIST_FIRST(&dirrem->dm_jremrefhd)) {
+					/* Journal remove ref dependency. */
+					retval += 1;
+					if (!wantcount)
+						goto out;
+				}
+			}
 			for (i = 0; i < DAHASHSZ; i++) {
 
 				LIST_FOREACH(dap, &pagedep->pd_diraddhd[i], da_pdlist) {
@@ -6159,14 +11107,62 @@ softdep_count_dependencies(bp, wantcount)
 			continue;
 
 		case D_BMSAFEMAP:
+			bmsafemap = WK_BMSAFEMAP(wk);
+			if (LIST_FIRST(&bmsafemap->sm_jaddrefhd)) {
+				/* Add reference dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
+			if (LIST_FIRST(&bmsafemap->sm_jnewblkhd)) {
+				/* Allocate block dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
+			continue;
+
+		case D_FREEBLKS:
+			freeblks = WK_FREEBLKS(wk);
+			if (LIST_FIRST(&freeblks->fb_jfreeblkhd)) {
+				/* Freeblk journal dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
+			continue;
+
 		case D_ALLOCDIRECT:
 		case D_ALLOCINDIR:
+			newblk = WK_NEWBLK(wk);
+			if (newblk->nb_jnewblk) {
+				/* Journal allocate dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
+			continue;
+
 		case D_MKDIR:
+			mkdir = WK_MKDIR(wk);
+			if (mkdir->md_jaddref) {
+				/* Journal reference dependency. */
+				retval += 1;
+				if (!wantcount)
+					goto out;
+			}
+			continue;
+
+		case D_FREEWORK:
+		case D_FREEDEP:
+		case D_JSEGDEP:
+		case D_JSEG:
+		case D_SBDEP:
 			/* never a dependency on these blocks */
 			continue;
 
 		default:
-			panic("softdep_check_for_rollback: Unexpected type %s",
+			panic("softdep_count_dependencies: Unexpected type %s",
 			    TYPENAME(wk->wk_type));
 			/* NOTREACHED */
 		}
@@ -6382,6 +11378,45 @@ softdep_error(func, error)
 
 #ifdef DDB
 
+static void
+inodedep_print(struct inodedep *inodedep, int verbose)
+{
+	db_printf("%p fs %p st %x ino %jd inoblk %jd delta %d nlink %d"
+	    " saveino %p\n",
+	    inodedep, inodedep->id_fs, inodedep->id_state,
+	    (intmax_t)inodedep->id_ino,
+	    (intmax_t)fsbtodb(inodedep->id_fs,
+	    ino_to_fsba(inodedep->id_fs, inodedep->id_ino)),
+	    inodedep->id_nlinkdelta, inodedep->id_savednlink,
+	    inodedep->id_savedino1);
+
+	if (verbose == 0)
+		return;
+
+	db_printf("\tpendinghd %p, bufwait %p, inowait %p, inoreflst %p, "
+	    "mkdiradd %p\n",
+	    LIST_FIRST(&inodedep->id_pendinghd),
+	    LIST_FIRST(&inodedep->id_bufwait),
+	    LIST_FIRST(&inodedep->id_inowait),
+	    TAILQ_FIRST(&inodedep->id_inoreflst),
+	    inodedep->id_mkdiradd);
+	db_printf("\tinoupdt %p, newinoupdt %p, extupdt %p, newextupdt %p\n",
+	    TAILQ_FIRST(&inodedep->id_inoupdt),
+	    TAILQ_FIRST(&inodedep->id_newinoupdt),
+	    TAILQ_FIRST(&inodedep->id_extupdt),
+	    TAILQ_FIRST(&inodedep->id_newextupdt));
+}
+
+DB_SHOW_COMMAND(inodedep, db_show_inodedep)
+{
+
+	if (have_addr == 0) {
+		db_printf("Address required\n");
+		return;
+	}
+	inodedep_print((struct inodedep*)addr, 1);
+}
+
 DB_SHOW_COMMAND(inodedeps, db_show_inodedeps)
 {
 	struct inodedep_hashhead *inodedephd;
@@ -6395,15 +11430,62 @@ DB_SHOW_COMMAND(inodedeps, db_show_inodedeps)
 		LIST_FOREACH(inodedep, inodedephd, id_hash) {
 			if (fs != NULL && fs != inodedep->id_fs)
 				continue;
-			db_printf("%p fs %p st %x ino %jd inoblk %jd\n",
-			    inodedep, inodedep->id_fs, inodedep->id_state,
-			    (intmax_t)inodedep->id_ino,
-			    (intmax_t)fsbtodb(inodedep->id_fs,
-			    ino_to_fsba(inodedep->id_fs, inodedep->id_ino)));
+			inodedep_print(inodedep, 0);
 		}
 	}
 }
 
+DB_SHOW_COMMAND(worklist, db_show_worklist)
+{
+	struct worklist *wk;
+
+	if (have_addr == 0) {
+		db_printf("Address required\n");
+		return;
+	}
+	wk = (struct worklist *)addr;
+	printf("worklist: %p type %s state 0x%X\n",
+	    wk, TYPENAME(wk->wk_type), wk->wk_state);
+}
+
+DB_SHOW_COMMAND(workhead, db_show_workhead)
+{
+	struct workhead *wkhd;
+	struct worklist *wk;
+	int i;
+
+	if (have_addr == 0) {
+		db_printf("Address required\n");
+		return;
+	}
+	wkhd = (struct workhead *)addr;
+	wk = LIST_FIRST(wkhd);
+	for (i = 0; i < 100 && wk != NULL; i++, wk = LIST_NEXT(wk, wk_list))
+		db_printf("worklist: %p type %s state 0x%X",
+		    wk, TYPENAME(wk->wk_type), wk->wk_state);
+	if (i == 100)
+		db_printf("workhead overflow");
+	printf("\n");
+}
+
+
+DB_SHOW_COMMAND(mkdirs, db_show_mkdirs)
+{
+	struct jaddref *jaddref;
+	struct diradd *diradd;
+	struct mkdir *mkdir;
+
+	LIST_FOREACH(mkdir, &mkdirlisthd, md_mkdirs) {
+		diradd = mkdir->md_diradd;
+		db_printf("mkdir: %p state 0x%X dap %p state 0x%X",
+		    mkdir, mkdir->md_state, diradd, diradd->da_state);
+		if ((jaddref = mkdir->md_jaddref) != NULL)
+			db_printf(" jaddref %p jaddref state 0x%X",
+			    jaddref, jaddref->ja_state);
+		db_printf("\n");
+	}
+}
+
 #endif /* DDB */
 
 #endif /* SOFTUPDATES */
diff --git a/sys/ufs/ffs/ffs_subr.c b/sys/ufs/ffs/ffs_subr.c
index e34bc1372a2..e2460a36be2 100644
--- a/sys/ufs/ffs/ffs_subr.c
+++ b/sys/ufs/ffs/ffs_subr.c
@@ -37,7 +37,6 @@ __FBSDID("$FreeBSD$");
 #ifndef _KERNEL
 #include <ufs/ufs/dinode.h>
 #include <ufs/ffs/fs.h>
-#include "fsck.h"
 #else
 #include <sys/systm.h>
 #include <sys/lock.h>
@@ -223,7 +222,38 @@ ffs_isblock(fs, cp, h)
 		mask = 0x01 << (h & 0x7);
 		return ((cp[h >> 3] & mask) == mask);
 	default:
+#ifdef _KERNEL
 		panic("ffs_isblock");
+#endif
+		break;
+	}
+	return (0);
+}
+
+/*
+ * check if a block is free
+ */
+int
+ffs_isfreeblock(fs, cp, h)
+	struct fs *fs;
+	u_char *cp;
+	ufs1_daddr_t h;
+{
+ 
+	switch ((int)fs->fs_frag) {
+	case 8:
+		return (cp[h] == 0);
+	case 4:
+		return ((cp[h >> 1] & (0x0f << ((h & 0x1) << 2))) == 0);
+	case 2:
+		return ((cp[h >> 2] & (0x03 << ((h & 0x3) << 1))) == 0);
+	case 1:
+		return ((cp[h >> 3] & (0x01 << (h & 0x7))) == 0);
+	default:
+#ifdef _KERNEL
+		panic("ffs_isfreeblock");
+#endif
+		break;
 	}
 	return (0);
 }
@@ -252,7 +282,10 @@ ffs_clrblock(fs, cp, h)
 		cp[h >> 3] &= ~(0x01 << (h & 0x7));
 		return;
 	default:
+#ifdef _KERNEL
 		panic("ffs_clrblock");
+#endif
+		break;
 	}
 }
 
@@ -281,6 +314,101 @@ ffs_setblock(fs, cp, h)
 		cp[h >> 3] |= (0x01 << (h & 0x7));
 		return;
 	default:
+#ifdef _KERNEL
 		panic("ffs_setblock");
+#endif
+		break;
 	}
 }
+
+/*
+ * Update the cluster map because of an allocation or free.
+ *
+ * Cnt == 1 means free; cnt == -1 means allocating.
+ */
+void
+ffs_clusteracct(fs, cgp, blkno, cnt)
+	struct fs *fs;
+	struct cg *cgp;
+	ufs1_daddr_t blkno;
+	int cnt;
+{
+	int32_t *sump;
+	int32_t *lp;
+	u_char *freemapp, *mapp;
+	int i, start, end, forw, back, map, bit;
+
+	if (fs->fs_contigsumsize <= 0)
+		return;
+	freemapp = cg_clustersfree(cgp);
+	sump = cg_clustersum(cgp);
+	/*
+	 * Allocate or clear the actual block.
+	 */
+	if (cnt > 0)
+		setbit(freemapp, blkno);
+	else
+		clrbit(freemapp, blkno);
+	/*
+	 * Find the size of the cluster going forward.
+	 */
+	start = blkno + 1;
+	end = start + fs->fs_contigsumsize;
+	if (end >= cgp->cg_nclusterblks)
+		end = cgp->cg_nclusterblks;
+	mapp = &freemapp[start / NBBY];
+	map = *mapp++;
+	bit = 1 << (start % NBBY);
+	for (i = start; i < end; i++) {
+		if ((map & bit) == 0)
+			break;
+		if ((i & (NBBY - 1)) != (NBBY - 1)) {
+			bit <<= 1;
+		} else {
+			map = *mapp++;
+			bit = 1;
+		}
+	}
+	forw = i - start;
+	/*
+	 * Find the size of the cluster going backward.
+	 */
+	start = blkno - 1;
+	end = start - fs->fs_contigsumsize;
+	if (end < 0)
+		end = -1;
+	mapp = &freemapp[start / NBBY];
+	map = *mapp--;
+	bit = 1 << (start % NBBY);
+	for (i = start; i > end; i--) {
+		if ((map & bit) == 0)
+			break;
+		if ((i & (NBBY - 1)) != 0) {
+			bit >>= 1;
+		} else {
+			map = *mapp--;
+			bit = 1 << (NBBY - 1);
+		}
+	}
+	back = start - i;
+	/*
+	 * Account for old cluster and the possibly new forward and
+	 * back clusters.
+	 */
+	i = back + forw + 1;
+	if (i > fs->fs_contigsumsize)
+		i = fs->fs_contigsumsize;
+	sump[i] += cnt;
+	if (back > 0)
+		sump[back] -= cnt;
+	if (forw > 0)
+		sump[forw] -= cnt;
+	/*
+	 * Update cluster summary information.
+	 */
+	lp = &sump[fs->fs_contigsumsize];
+	for (i = fs->fs_contigsumsize; i > 0; i--)
+		if (*lp-- > 0)
+			break;
+	fs->fs_maxcluster[cgp->cg_cgx] = i;
+}
diff --git a/sys/ufs/ffs/ffs_vfsops.c b/sys/ufs/ffs/ffs_vfsops.c
index 656c03666c5..e4033686324 100644
--- a/sys/ufs/ffs/ffs_vfsops.c
+++ b/sys/ufs/ffs/ffs_vfsops.c
@@ -79,7 +79,6 @@ static int	ffs_reload(struct mount *, struct thread *);
 static int	ffs_mountfs(struct vnode *, struct mount *, struct thread *);
 static void	ffs_oldfscompat_read(struct fs *, struct ufsmount *,
 		    ufs2_daddr_t);
-static void	ffs_oldfscompat_write(struct fs *, struct ufsmount *);
 static void	ffs_ifree(struct ufsmount *ump, struct inode *ip);
 static vfs_init_t ffs_init;
 static vfs_uninit_t ffs_uninit;
@@ -299,7 +298,8 @@ ffs_mount(struct mount *mp)
 			if (fs->fs_clean == 0) {
 				fs->fs_flags |= FS_UNCLEAN;
 				if ((mp->mnt_flag & MNT_FORCE) ||
-				    ((fs->fs_flags & FS_NEEDSFSCK) == 0 &&
+				    ((fs->fs_flags &
+				     (FS_SUJ | FS_NEEDSFSCK)) == 0 &&
 				     (fs->fs_flags & FS_DOSOFTDEP))) {
 					printf("WARNING: %s was not %s\n",
 					   fs->fs_fsmnt, "properly dismounted");
@@ -307,6 +307,9 @@ ffs_mount(struct mount *mp)
 					printf(
 "WARNING: R/W mount of %s denied.  Filesystem is not clean - run fsck\n",
 					    fs->fs_fsmnt);
+					if (fs->fs_flags & FS_SUJ)
+						printf(
+"WARNING: Forced mount will invalidated journal contents\n");
 					return (EPERM);
 				}
 			}
@@ -330,17 +333,18 @@ ffs_mount(struct mount *mp)
 			MNT_ILOCK(mp);
 			mp->mnt_flag &= ~MNT_RDONLY;
 			MNT_IUNLOCK(mp);
-			fs->fs_clean = 0;
-			if ((error = ffs_sbupdate(ump, MNT_WAIT, 0)) != 0) {
-				vn_finished_write(mp);
-				return (error);
-			}
+			fs->fs_mtime = time_second;
 			/* check to see if we need to start softdep */
 			if ((fs->fs_flags & FS_DOSOFTDEP) &&
 			    (error = softdep_mount(devvp, mp, fs, td->td_ucred))){
 				vn_finished_write(mp);
 				return (error);
 			}
+			fs->fs_clean = 0;
+			if ((error = ffs_sbupdate(ump, MNT_WAIT, 0)) != 0) {
+				vn_finished_write(mp);
+				return (error);
+			}
 			if (fs->fs_snapinum[0] != 0)
 				ffs_snapshot_mount(mp);
 			vn_finished_write(mp);
@@ -705,7 +709,7 @@ ffs_mountfs(devvp, mp, td)
 	if (fs->fs_clean == 0) {
 		fs->fs_flags |= FS_UNCLEAN;
 		if (ronly || (mp->mnt_flag & MNT_FORCE) ||
-		    ((fs->fs_flags & FS_NEEDSFSCK) == 0 &&
+		    ((fs->fs_flags & (FS_SUJ | FS_NEEDSFSCK)) == 0 &&
 		     (fs->fs_flags & FS_DOSOFTDEP))) {
 			printf(
 "WARNING: %s was not properly dismounted\n",
@@ -714,6 +718,9 @@ ffs_mountfs(devvp, mp, td)
 			printf(
 "WARNING: R/W mount of %s denied.  Filesystem is not clean - run fsck\n",
 			    fs->fs_fsmnt);
+			if (fs->fs_flags & FS_SUJ)
+				printf(
+"WARNING: Forced mount will invalidated journal contents\n");
 			error = EPERM;
 			goto out;
 		}
@@ -896,6 +903,7 @@ ffs_mountfs(devvp, mp, td)
 	 */
 	bzero(fs->fs_fsmnt, MAXMNTLEN);
 	strlcpy(fs->fs_fsmnt, mp->mnt_stat.f_mntonname, MAXMNTLEN);
+	mp->mnt_stat.f_iosize = fs->fs_bsize;
 
 	if( mp->mnt_flag & MNT_ROOTFS) {
 		/*
@@ -907,6 +915,7 @@ ffs_mountfs(devvp, mp, td)
 	}
 
 	if (ronly == 0) {
+		fs->fs_mtime = time_second;
 		if ((fs->fs_flags & FS_DOSOFTDEP) &&
 		    (error = softdep_mount(devvp, mp, fs, cred)) != 0) {
 			free(fs->fs_csp, M_UFSMNT);
@@ -937,7 +946,6 @@ ffs_mountfs(devvp, mp, td)
 	 * This would all happen while the filesystem was busy/not
 	 * available, so would effectively be "atomic".
 	 */
-	mp->mnt_stat.f_iosize = fs->fs_bsize;
 	(void) ufs_extattr_autostart(mp, td);
 #endif /* !UFS_EXTATTR_AUTOSTART */
 #endif /* !UFS_EXTATTR */
@@ -1037,7 +1045,7 @@ ffs_oldfscompat_read(fs, ump, sblockloc)
  * XXX - Parts get retired eventually.
  * Unfortunately new bits get added.
  */
-static void
+void
 ffs_oldfscompat_write(fs, ump)
 	struct fs *fs;
 	struct ufsmount *ump;
@@ -1132,6 +1140,7 @@ ffs_unmount(mp, mntflags)
 		fs->fs_pendinginodes = 0;
 	}
 	UFS_UNLOCK(ump);
+	softdep_unmount(mp);
 	if (fs->fs_ronly == 0) {
 		fs->fs_clean = fs->fs_flags & (FS_UNCLEAN|FS_NEEDSFSCK) ? 0 : 1;
 		error = ffs_sbupdate(ump, MNT_WAIT, 0);
@@ -1573,16 +1582,6 @@ ffs_vgetf(mp, ino, flags, vpp, ffs_flags)
 			DIP_SET(ip, i_gen, ip->i_gen);
 		}
 	}
-	/*
-	 * Ensure that uid and gid are correct. This is a temporary
-	 * fix until fsck has been changed to do the update.
-	 */
-	if (fs->fs_magic == FS_UFS1_MAGIC &&		/* XXX */
-	    fs->fs_old_inodefmt < FS_44INODEFMT) {	/* XXX */
-		ip->i_uid = ip->i_din1->di_ouid;	/* XXX */
-		ip->i_gid = ip->i_din1->di_ogid;	/* XXX */
-	}						/* XXX */
-
 #ifdef MAC
 	if ((mp->mnt_flag & MNT_MULTILABEL) && ip->i_mode) {
 		/*
@@ -1726,6 +1725,8 @@ ffs_sbupdate(mp, waitfor, suspended)
 	}
 	fs->fs_fmod = 0;
 	fs->fs_time = time_second;
+	if (fs->fs_flags & FS_DOSOFTDEP)
+		softdep_setup_sbupdate(mp, (struct fs *)bp->b_data, bp);
 	bcopy((caddr_t)fs, bp->b_data, (u_int)fs->fs_sbsize);
 	ffs_oldfscompat_write((struct fs *)bp->b_data, mp);
 	if (suspended)
@@ -1867,9 +1868,6 @@ ffs_bufwrite(struct buf *bp)
 	}
 	BO_UNLOCK(bp->b_bufobj);
 
-	/* Mark the buffer clean */
-	bundirty(bp);
-
 	/*
 	 * If this buffer is marked for background writing and we
 	 * do not have to wait for it, make a copy and write the
@@ -1910,9 +1908,16 @@ ffs_bufwrite(struct buf *bp)
 		newbp->b_flags &= ~B_INVAL;
 
 #ifdef SOFTUPDATES
-		/* move over the dependencies */
-		if (!LIST_EMPTY(&bp->b_dep))
-			softdep_move_dependencies(bp, newbp);
+		/*
+		 * Move over the dependencies.  If there are rollbacks,
+		 * leave the parent buffer dirtied as it will need to
+		 * be written again.
+		 */
+		if (LIST_EMPTY(&bp->b_dep) ||
+		    softdep_move_dependencies(bp, newbp) == 0)
+			bundirty(bp);
+#else
+		bundirty(bp);
 #endif 
 
 		/*
@@ -1925,7 +1930,10 @@ ffs_bufwrite(struct buf *bp)
 		 */
 		bqrelse(bp);
 		bp = newbp;
-	}
+	} else
+		/* Mark the buffer clean */
+		bundirty(bp);
+
 
 	/* Let the normal bufwrite do the rest for us */
 normal_write:
@@ -1939,6 +1947,7 @@ ffs_geom_strategy(struct bufobj *bo, struct buf *bp)
 	struct vnode *vp;
 	int error;
 	struct buf *tbp;
+	int nocopy;
 
 	vp = bo->__bo_vnode;
 	if (bp->b_iocmd == BIO_WRITE) {
@@ -1946,8 +1955,9 @@ ffs_geom_strategy(struct bufobj *bo, struct buf *bp)
 		    bp->b_vp != NULL && bp->b_vp->v_mount != NULL &&
 		    (bp->b_vp->v_mount->mnt_kern_flag & MNTK_SUSPENDED) != 0)
 			panic("ffs_geom_strategy: bad I/O");
-		bp->b_flags &= ~B_VALIDSUSPWRT;
-		if ((vp->v_vflag & VV_COPYONWRITE) &&
+		nocopy = bp->b_flags & B_NOCOPY;
+		bp->b_flags &= ~(B_VALIDSUSPWRT | B_NOCOPY);
+		if ((vp->v_vflag & VV_COPYONWRITE) && nocopy == 0 &&
 		    vp->v_rdev->si_snapdata != NULL) {
 			if ((bp->b_flags & B_CLUSTER) != 0) {
 				runningbufwakeup(bp);
diff --git a/sys/ufs/ffs/ffs_vnops.c b/sys/ufs/ffs/ffs_vnops.c
index 464a7613e16..e6617cbcdfa 100644
--- a/sys/ufs/ffs/ffs_vnops.c
+++ b/sys/ufs/ffs/ffs_vnops.c
@@ -225,6 +225,7 @@ ffs_syncvnode(struct vnode *vp, int waitfor)
 	wait = (waitfor == MNT_WAIT);
 	lbn = lblkno(ip->i_fs, (ip->i_size + ip->i_fs->fs_bsize - 1));
 	bo = &vp->v_bufobj;
+	ip->i_flag &= ~IN_NEEDSYNC;
 
 	/*
 	 * Flush all dirty buffers associated with a vnode.
diff --git a/sys/ufs/ffs/fs.h b/sys/ufs/ffs/fs.h
index 5452e2be6de..e863b961c62 100644
--- a/sys/ufs/ffs/fs.h
+++ b/sys/ufs/ffs/fs.h
@@ -340,7 +340,9 @@ struct fs {
 	u_int32_t fs_avgfilesize;	/* expected average file size */
 	u_int32_t fs_avgfpdir;		/* expected # of files per directory */
 	int32_t	 fs_save_cgsize;	/* save real cg size to use fs_bsize */
-	int32_t	 fs_sparecon32[26];	/* reserved for future constants */
+	ufs_time_t fs_mtime;		/* Last mount or fsck time. */
+	int32_t  fs_sujfree;		/* SUJ free list */
+	int32_t	 fs_sparecon32[23];	/* reserved for future constants */
 	int32_t  fs_flags;		/* see FS_ flags below */
 	int32_t	 fs_contigsumsize;	/* size of cluster summary array */ 
 	int32_t	 fs_maxsymlinklen;	/* max length of an internal symlink */
@@ -408,12 +410,13 @@ CTASSERT(sizeof(struct fs) == 1376);
 #define FS_UNCLEAN	0x0001	/* filesystem not clean at mount */
 #define FS_DOSOFTDEP	0x0002	/* filesystem using soft dependencies */
 #define FS_NEEDSFSCK	0x0004	/* filesystem needs sync fsck before mount */
-#define FS_INDEXDIRS	0x0008	/* kernel supports indexed directories */
+#define	FS_SUJ       	0x0008	/* Filesystem using softupdate journal */
 #define FS_ACLS		0x0010	/* file system has POSIX.1e ACLs enabled */
 #define FS_MULTILABEL	0x0020	/* file system is MAC multi-label */
 #define FS_GJOURNAL	0x0040	/* gjournaled file system */
 #define FS_FLAGS_UPDATED 0x0080	/* flags have been moved to new location */
 #define FS_NFS4ACLS	0x0100	/* file system has NFSv4 ACLs enabled */
+#define FS_INDEXDIRS	0x0200	/* kernel supports indexed directories */
 
 /*
  * Macros to access bits in the fs_active array.
@@ -603,7 +606,31 @@ struct cg {
 	  ? (fs)->fs_bsize \
 	  : (fragroundup(fs, blkoff(fs, (size)))))
 
-
+/*
+ * Indirect lbns are aligned on NDADDR addresses where single indirects
+ * are the negated address of the lowest lbn reachable, double indirects
+ * are this lbn - 1 and triple indirects are this lbn - 2.  This yields
+ * an unusual bit order to determine level.
+ */
+static inline int
+lbn_level(ufs_lbn_t lbn)
+{
+	if (lbn >= 0)
+		return 0;
+	switch (lbn & 0x3) {
+	case 0:
+		return (0);
+	case 1:
+		break;
+	case 2:
+		return (2);
+	case 3:
+		return (1);
+	default:
+		break;
+	}
+	return (-1);
+}
 /*
  * Number of inodes in a secondary storage block/fragment.
  */
@@ -615,6 +642,108 @@ struct cg {
  */
 #define	NINDIR(fs)	((fs)->fs_nindir)
 
+/*
+ * Softdep journal record format.
+ */
+
+#define	JOP_ADDREF	1	/* Add a reference to an inode. */
+#define	JOP_REMREF	2	/* Remove a reference from an inode. */
+#define	JOP_NEWBLK	3	/* Allocate a block. */
+#define	JOP_FREEBLK	4	/* Free a block or a tree of blocks. */
+#define	JOP_MVREF	5	/* Move a reference from one off to another. */
+#define	JOP_TRUNC	6	/* Partial truncation record. */
+
+#define	JREC_SIZE	32	/* Record and segment header size. */
+
+#define	SUJ_MIN		(4 * 1024 * 1024)	/* Minimum journal size */
+#define	SUJ_MAX		(32 * 1024 * 1024)	/* Maximum journal size */
+#define	SUJ_FILE	".sujournal"		/* Journal file name */
+
+/*
+ * Size of the segment record header.  There is at most one for each disk
+ * block n the journal.  The segment header is followed by an array of
+ * records.  fsck depends on the first element in each record being 'op'
+ * and the second being 'ino'.  Segments may span multiple disk blocks but
+ * the header is present on each.
+ */
+struct jsegrec {
+	uint64_t	jsr_seq;	/* Our sequence number */
+	uint64_t	jsr_oldest;	/* Oldest valid sequence number */
+	uint16_t	jsr_cnt;	/* Count of valid records */
+	uint16_t	jsr_blocks;	/* Count of DEV_BSIZE blocks. */
+	uint32_t	jsr_crc;	/* 32bit crc of the valid space */
+	ufs_time_t	jsr_time;	/* timestamp for mount instance */
+};
+
+/*
+ * Reference record.  Records a single link count modification.
+ */
+struct jrefrec {
+	uint32_t	jr_op;
+	ino_t		jr_ino;
+	ino_t		jr_parent;
+	uint16_t	jr_nlink;
+	uint16_t	jr_mode;
+	off_t		jr_diroff;
+	uint64_t	jr_unused;
+};
+
+/*
+ * Move record.  Records a reference moving within a directory block.  The
+ * nlink is unchanged but we must search both locations.
+ */
+struct jmvrec {
+	uint32_t	jm_op;
+	ino_t		jm_ino;
+	ino_t		jm_parent;
+	uint16_t	jm_unused;
+	off_t		jm_oldoff;
+	off_t		jm_newoff;
+};
+
+/*
+ * Block record.  A set of frags or tree of blocks starting at an indirect are
+ * freed or a set of frags are allocated.
+ */
+struct jblkrec {
+	uint32_t	jb_op;
+	uint32_t	jb_ino;
+	ufs2_daddr_t	jb_blkno;
+	ufs_lbn_t	jb_lbn;
+	uint16_t	jb_frags;
+	uint16_t	jb_oldfrags;
+	uint32_t	jb_unused;
+};
+
+/*
+ * Truncation record.  Records a partial truncation so that it may be
+ * completed later.
+ */
+struct jtrncrec {
+	uint32_t	jt_op;
+	uint32_t	jt_ino;
+	off_t		jt_size;
+	uint32_t	jt_extsize;
+	uint32_t	jt_pad[3];
+};
+
+union jrec {
+	struct jsegrec	rec_jsegrec;
+	struct jrefrec	rec_jrefrec;
+	struct jmvrec	rec_jmvrec;
+	struct jblkrec	rec_jblkrec;
+	struct jtrncrec	rec_jtrncrec;
+};
+
+#ifdef CTASSERT
+CTASSERT(sizeof(struct jsegrec) == JREC_SIZE);
+CTASSERT(sizeof(struct jrefrec) == JREC_SIZE);
+CTASSERT(sizeof(struct jmvrec) == JREC_SIZE);
+CTASSERT(sizeof(struct jblkrec) == JREC_SIZE);
+CTASSERT(sizeof(struct jtrncrec) == JREC_SIZE);
+CTASSERT(sizeof(union jrec) == JREC_SIZE);
+#endif
+
 extern int inside[], around[];
 extern u_char *fragtbl[];
 
diff --git a/sys/ufs/ffs/softdep.h b/sys/ufs/ffs/softdep.h
index b00183bcfd2..5d8a8691b17 100644
--- a/sys/ufs/ffs/softdep.h
+++ b/sys/ufs/ffs/softdep.h
@@ -94,22 +94,29 @@
  * The ONWORKLIST flag shows whether the structure is currently linked
  * onto a worklist.
  */
-#define	ATTACHED	0x0001
-#define	UNDONE		0x0002
-#define	COMPLETE	0x0004
-#define	DEPCOMPLETE	0x0008
-#define	MKDIR_PARENT	0x0010	/* diradd & mkdir only */
-#define	MKDIR_BODY	0x0020	/* diradd & mkdir only */
-#define	RMDIR		0x0040	/* dirrem only */
-#define	DIRCHG		0x0080	/* diradd & dirrem only */
-#define	GOINGAWAY	0x0100	/* indirdep only */
-#define	IOSTARTED	0x0200	/* inodedep & pagedep only */
-#define	SPACECOUNTED	0x0400	/* inodedep only */
-#define	NEWBLOCK	0x0800	/* pagedep only */
-#define	INPROGRESS	0x1000	/* dirrem, freeblks, freefrag, freefile only */
-#define	UFS1FMT		0x2000	/* indirdep only */
-#define	EXTDATA		0x4000	/* allocdirect only */
-#define ONWORKLIST	0x8000
+#define	ATTACHED	0x000001
+#define	UNDONE		0x000002
+#define	COMPLETE	0x000004
+#define	DEPCOMPLETE	0x000008
+#define	MKDIR_PARENT	0x000010 /* diradd, mkdir, jaddref, jsegdep only */
+#define	MKDIR_BODY	0x000020 /* diradd, mkdir, jaddref only */
+#define	RMDIR		0x000040 /* dirrem only */
+#define	DIRCHG		0x000080 /* diradd, dirrem only */
+#define	GOINGAWAY	0x000100 /* indirdep, jremref only */
+#define	IOSTARTED	0x000200 /* inodedep, pagedep, bmsafemap only */
+#define	SPACECOUNTED	0x000400 /* inodedep only */
+#define	NEWBLOCK	0x000800 /* pagedep, jaddref only */
+#define	INPROGRESS	0x001000 /* dirrem, freeblks, freefrag, freefile only */
+#define	UFS1FMT		0x002000 /* indirdep only */
+#define	EXTDATA		0x004000 /* allocdirect only */
+#define ONWORKLIST	0x008000
+#define	IOWAITING	0x010000 /* Thread is waiting for IO to complete. */
+#define	ONDEPLIST	0x020000 /* Structure is on a dependency list. */
+#define	UNLINKED	0x040000 /* inodedep has been unlinked. */
+#define	UNLINKNEXT	0x080000 /* inodedep has valid di_freelink */
+#define	UNLINKPREV	0x100000 /* inodedep is pointed at in the unlink list */
+#define	UNLINKONLIST	0x200000 /* inodedep is in the unlinked list on disk */
+#define	UNLINKLINKS	(UNLINKNEXT | UNLINKPREV)
 
 #define	ALLCOMPLETE	(ATTACHED | COMPLETE | DEPCOMPLETE)
 
@@ -135,25 +142,38 @@
  * and the macros below changed to use it.
  */
 struct worklist {
-	struct mount		*wk_mp;		/* Mount we live in */
 	LIST_ENTRY(worklist)	wk_list;	/* list of work requests */
-	unsigned short		wk_type;	/* type of request */
-	unsigned short		wk_state;	/* state flags */
+	struct mount		*wk_mp;		/* Mount we live in */
+	unsigned int		wk_type:8,	/* type of request */
+				wk_state:24;	/* state flags */
 };
 #define WK_DATA(wk) ((void *)(wk))
 #define WK_PAGEDEP(wk) ((struct pagedep *)(wk))
 #define WK_INODEDEP(wk) ((struct inodedep *)(wk))
 #define WK_BMSAFEMAP(wk) ((struct bmsafemap *)(wk))
+#define	WK_NEWBLK(wk)  ((struct newblk *)(wk))
 #define WK_ALLOCDIRECT(wk) ((struct allocdirect *)(wk))
 #define WK_INDIRDEP(wk) ((struct indirdep *)(wk))
 #define WK_ALLOCINDIR(wk) ((struct allocindir *)(wk))
 #define WK_FREEFRAG(wk) ((struct freefrag *)(wk))
 #define WK_FREEBLKS(wk) ((struct freeblks *)(wk))
+#define WK_FREEWORK(wk) ((struct freework *)(wk))
 #define WK_FREEFILE(wk) ((struct freefile *)(wk))
 #define WK_DIRADD(wk) ((struct diradd *)(wk))
 #define WK_MKDIR(wk) ((struct mkdir *)(wk))
 #define WK_DIRREM(wk) ((struct dirrem *)(wk))
 #define WK_NEWDIRBLK(wk) ((struct newdirblk *)(wk))
+#define	WK_JADDREF(wk) ((struct jaddref *)(wk))
+#define	WK_JREMREF(wk) ((struct jremref *)(wk))
+#define	WK_JMVREF(wk) ((struct jmvref *)(wk))
+#define	WK_JSEGDEP(wk) ((struct jsegdep *)(wk))
+#define	WK_JSEG(wk) ((struct jseg *)(wk))
+#define	WK_JNEWBLK(wk) ((struct jnewblk *)(wk))
+#define	WK_JFREEBLK(wk) ((struct jfreeblk *)(wk))
+#define	WK_FREEDEP(wk) ((struct freedep *)(wk))
+#define	WK_JFREEFRAG(wk) ((struct jfreefrag *)(wk))
+#define	WK_SBDEP(wk) ((struct sbdep *)wk)
+#define	WK_JTRUNC(wk) ((struct jtrunc *)(wk))
 
 /*
  * Various types of lists
@@ -165,6 +185,15 @@ LIST_HEAD(inodedephd, inodedep);
 LIST_HEAD(allocindirhd, allocindir);
 LIST_HEAD(allocdirecthd, allocdirect);
 TAILQ_HEAD(allocdirectlst, allocdirect);
+LIST_HEAD(indirdephd, indirdep);
+LIST_HEAD(jaddrefhd, jaddref);
+LIST_HEAD(jremrefhd, jremref);
+LIST_HEAD(jmvrefhd, jmvref);
+LIST_HEAD(jnewblkhd, jnewblk);
+LIST_HEAD(jfreeblkhd, jfreeblk);
+LIST_HEAD(freeworkhd, freework);
+TAILQ_HEAD(jseglst, jseg);
+TAILQ_HEAD(inoreflst, inoref);
 
 /*
  * The "pagedep" structure tracks the various dependencies related to
@@ -192,9 +221,11 @@ struct pagedep {
 	LIST_ENTRY(pagedep) pd_hash;	/* hashed lookup */
 	ino_t	pd_ino;			/* associated file */
 	ufs_lbn_t pd_lbn;		/* block within file */
+	struct	newdirblk *pd_newdirblk; /* associated newdirblk if NEWBLOCK */
 	struct	dirremhd pd_dirremhd;	/* dirrem's waiting for page */
 	struct	diraddhd pd_diraddhd[DAHASHSZ]; /* diradd dir entry updates */
 	struct	diraddhd pd_pendinghd;	/* directory entries awaiting write */
+	struct	jmvrefhd pd_jmvrefhd;	/* Dependent journal writes. */
 };
 
 /*
@@ -248,13 +279,18 @@ struct inodedep {
 	struct	worklist id_list;	/* buffer holding inode block */
 #	define	id_state id_list.wk_state /* inode dependency state */
 	LIST_ENTRY(inodedep) id_hash;	/* hashed lookup */
+	TAILQ_ENTRY(inodedep) id_unlinked;	/* Unlinked but ref'd inodes */
 	struct	fs *id_fs;		/* associated filesystem */
 	ino_t	id_ino;			/* dependent inode */
 	nlink_t	id_nlinkdelta;		/* saved effective link count */
+	nlink_t	id_savednlink;		/* Link saved during rollback */
 	LIST_ENTRY(inodedep) id_deps;	/* bmsafemap's list of inodedep's */
-	struct	buf *id_buf;		/* related bmsafemap (if pending) */
+	struct	bmsafemap *id_bmsafemap; /* related bmsafemap (if pending) */
+	struct	diradd *id_mkdiradd;	/* diradd for a mkdir. */
+	struct	inoreflst id_inoreflst;	/* Inode reference adjustments. */
 	long	id_savedextsize;	/* ext size saved during rollback */
 	off_t	id_savedsize;		/* file size saved during rollback */
+	struct	dirremhd id_dirremhd;	/* Removals pending. */
 	struct	workhead id_pendinghd;	/* entries awaiting directory write */
 	struct	workhead id_bufwait;	/* operations after inode written */
 	struct	workhead id_inowait;	/* operations waiting inode update */
@@ -270,23 +306,6 @@ struct inodedep {
 #define id_savedino1 id_un.idu_savedino1
 #define id_savedino2 id_un.idu_savedino2
 
-/*
- * A "newblk" structure is attached to a bmsafemap structure when a block
- * or fragment is allocated from a cylinder group. Its state is set to
- * DEPCOMPLETE when its cylinder group map is written. It is consumed by
- * an associated allocdirect or allocindir allocation which will attach
- * themselves to the bmsafemap structure if the newblk's DEPCOMPLETE flag
- * is not set (i.e., its cylinder group map has not been written).
- */ 
-struct newblk {
-	LIST_ENTRY(newblk) nb_hash;	/* hashed lookup */
-	struct	fs *nb_fs;		/* associated filesystem */
-	int	nb_state;		/* state of bitmap dependency */
-	ufs2_daddr_t nb_newblkno;	/* allocated block number */
-	LIST_ENTRY(newblk) nb_deps;	/* bmsafemap's list of newblk's */
-	struct	bmsafemap *nb_bmsafemap; /* associated bmsafemap */
-};
-
 /*
  * A "bmsafemap" structure maintains a list of dependency structures
  * that depend on the update of a particular cylinder group map.
@@ -299,11 +318,41 @@ struct newblk {
  */
 struct bmsafemap {
 	struct	worklist sm_list;	/* cylgrp buffer */
+#	define	sm_state sm_list.wk_state
+	int	sm_cg;
+	LIST_ENTRY(bmsafemap) sm_hash;	/* Hash links. */
 	struct	buf *sm_buf;		/* associated buffer */
 	struct	allocdirecthd sm_allocdirecthd; /* allocdirect deps */
+	struct	allocdirecthd sm_allocdirectwr; /* writing allocdirect deps */
 	struct	allocindirhd sm_allocindirhd; /* allocindir deps */
+	struct	allocindirhd sm_allocindirwr; /* writing allocindir deps */
 	struct	inodedephd sm_inodedephd; /* inodedep deps */
+	struct	inodedephd sm_inodedepwr; /* writing inodedep deps */
 	struct	newblkhd sm_newblkhd;	/* newblk deps */
+	struct	newblkhd sm_newblkwr;	/* writing newblk deps */
+	struct	jaddrefhd sm_jaddrefhd;	/* Pending inode allocations. */
+	struct	jnewblkhd sm_jnewblkhd;	/* Pending block allocations. */
+};
+
+/*
+ * A "newblk" structure is attached to a bmsafemap structure when a block
+ * or fragment is allocated from a cylinder group. Its state is set to
+ * DEPCOMPLETE when its cylinder group map is written. It is converted to
+ * an allocdirect or allocindir allocation once the allocator calls the
+ * appropriate setup function.
+ */ 
+struct newblk {
+	struct	worklist nb_list;
+#	define	nb_state nb_list.wk_state
+	LIST_ENTRY(newblk) nb_hash;	/* hashed lookup */
+	LIST_ENTRY(newblk) nb_deps; /* bmsafemap's list of newblks */
+	struct	jnewblk *nb_jnewblk;	/* New block journal entry. */
+	struct	bmsafemap *nb_bmsafemap;/* cylgrp dep (if pending) */
+	struct	freefrag *nb_freefrag;	/* fragment to be freed (if any) */
+	struct	indirdephd nb_indirdeps; /* Children indirect blocks. */
+	struct	workhead nb_newdirblk;	/* dir block to notify when written */
+	struct	workhead nb_jwork;	/* Journal work pending. */
+	ufs2_daddr_t	nb_newblkno;	/* new value of block pointer */
 };
 
 /*
@@ -334,20 +383,18 @@ struct bmsafemap {
  * and inodedep->id_pendinghd lists.
  */
 struct allocdirect {
-	struct	worklist ad_list;	/* buffer holding block */
-#	define	ad_state ad_list.wk_state /* block pointer state */
+	struct	newblk ad_block;	/* Common block logic */
+#	define	ad_state ad_block.nb_list.wk_state /* block pointer state */
 	TAILQ_ENTRY(allocdirect) ad_next; /* inodedep's list of allocdirect's */
-	ufs_lbn_t ad_lbn;		/* block within file */
-	ufs2_daddr_t ad_newblkno;	/* new value of block pointer */
-	ufs2_daddr_t ad_oldblkno;	/* old value of block pointer */
-	long	ad_newsize;		/* size of new block */
-	long	ad_oldsize;		/* size of old block */
-	LIST_ENTRY(allocdirect) ad_deps; /* bmsafemap's list of allocdirect's */
-	struct	buf *ad_buf;		/* cylgrp buffer (if pending) */
 	struct	inodedep *ad_inodedep;	/* associated inodedep */
-	struct	freefrag *ad_freefrag;	/* fragment to be freed (if any) */
-	struct	workhead ad_newdirblk;	/* dir block to notify when written */
+	ufs2_daddr_t	ad_oldblkno;	/* old value of block pointer */
+	int		ad_offset;	/* Pointer offset in parent. */
+	long		ad_newsize;	/* size of new block */
+	long		ad_oldsize;	/* size of old block */
 };
+#define	ad_newblkno	ad_block.nb_newblkno
+#define	ad_freefrag	ad_block.nb_freefrag
+#define	ad_newdirblk	ad_block.nb_newdirblk
 
 /*
  * A single "indirdep" structure manages all allocation dependencies for
@@ -369,10 +416,14 @@ struct allocdirect {
 struct indirdep {
 	struct	worklist ir_list;	/* buffer holding indirect block */
 #	define	ir_state ir_list.wk_state /* indirect block pointer state */
-	caddr_t ir_saveddata;		/* buffer cache contents */
+	LIST_ENTRY(indirdep) ir_next;	/* alloc{direct,indir} list */
+	caddr_t	ir_saveddata;		/* buffer cache contents */
 	struct	buf *ir_savebp;		/* buffer holding safe copy */
+	struct	allocindirhd ir_completehd; /* waiting for indirdep complete */
+	struct	allocindirhd ir_writehd; /* Waiting for the pointer write. */
 	struct	allocindirhd ir_donehd;	/* done waiting to update safecopy */
 	struct	allocindirhd ir_deplisthd; /* allocindir deps for this block */
+	struct	workhead ir_jwork;	/* Journal work pending. */
 };
 
 /*
@@ -389,16 +440,25 @@ struct indirdep {
  * can then be freed as it is no longer applicable.
  */
 struct allocindir {
-	struct	worklist ai_list;	/* buffer holding indirect block */
-#	define	ai_state ai_list.wk_state /* indirect block pointer state */
+	struct	newblk ai_block;	/* Common block area */
+#	define	ai_state ai_block.nb_list.wk_state /* indirect pointer state */
 	LIST_ENTRY(allocindir) ai_next;	/* indirdep's list of allocindir's */
-	int	ai_offset;		/* pointer offset in indirect block */
-	ufs2_daddr_t ai_newblkno;	/* new block pointer value */
-	ufs2_daddr_t ai_oldblkno;	/* old block pointer value */
-	struct	freefrag *ai_freefrag;	/* block to be freed when complete */
 	struct	indirdep *ai_indirdep;	/* address of associated indirdep */
-	LIST_ENTRY(allocindir) ai_deps;	/* bmsafemap's list of allocindir's */
-	struct	buf *ai_buf;		/* cylgrp buffer (if pending) */
+	ufs2_daddr_t	ai_oldblkno;	/* old value of block pointer */
+	int		ai_offset;	/* Pointer offset in parent. */
+};
+#define	ai_newblkno	ai_block.nb_newblkno
+#define	ai_freefrag	ai_block.nb_freefrag
+#define	ai_newdirblk	ai_block.nb_newdirblk
+
+/*
+ * The allblk union is used to size the newblk structure on allocation so
+ * that it may be any one of three types.
+ */
+union allblk {
+	struct	allocindir ab_allocindir;
+	struct	allocdirect ab_allocdirect;
+	struct	newblk	ab_newblk;
 };
 
 /*
@@ -406,14 +466,13 @@ struct allocindir {
  * allocated fragment is replaced with a larger fragment, rather than extended.
  * The "freefrag" structure is constructed and attached when the replacement
  * block is first allocated. It is processed after the inode claiming the
- * bigger block that replaces it has been written to disk. Note that the
- * ff_state field is is used to store the uid, so may lose data. However,
- * the uid is used only in printing an error message, so is not critical.
- * Keeping it in a short keeps the data structure down to 32 bytes.
+ * bigger block that replaces it has been written to disk.
  */
 struct freefrag {
 	struct	worklist ff_list;	/* id_inowait or delayed worklist */
-#	define	ff_state ff_list.wk_state /* owning user; should be uid_t */
+#	define	ff_state ff_list.wk_state
+	struct	jfreefrag *ff_jfreefrag; /* Associated journal entry. */
+	struct	workhead ff_jwork;	/* Journal work pending. */
 	ufs2_daddr_t ff_blkno;		/* fragment physical block number */
 	long	ff_fragsize;		/* size of fragment being deleted */
 	ino_t	ff_inum;		/* owning inode number */
@@ -423,20 +482,57 @@ struct freefrag {
  * A "freeblks" structure is attached to an "inodedep" when the
  * corresponding file's length is reduced to zero. It records all
  * the information needed to free the blocks of a file after its
- * zero'ed inode has been written to disk.
+ * zero'ed inode has been written to disk.  The actual work is done
+ * by child freework structures which are responsible for individual
+ * inode pointers while freeblks is responsible for retiring the
+ * entire operation when it is complete and holding common members.
  */
 struct freeblks {
 	struct	worklist fb_list;	/* id_inowait or delayed worklist */
 #	define	fb_state fb_list.wk_state /* inode and dirty block state */
+	struct	jfreeblkhd fb_jfreeblkhd; /* Journal entries pending */
+	struct	workhead fb_freeworkhd;	/* Work items pending */
+	struct	workhead fb_jwork;	/* Journal work pending */
 	ino_t	fb_previousinum;	/* inode of previous owner of blocks */
 	uid_t	fb_uid;			/* uid of previous owner of blocks */
 	struct	vnode *fb_devvp;	/* filesystem device vnode */
-	long	fb_oldextsize;		/* previous ext data size */
-	off_t	fb_oldsize;		/* previous file size */
 	ufs2_daddr_t fb_chkcnt;		/* used to check cnt of blks released */
-	ufs2_daddr_t fb_dblks[NDADDR];	/* direct blk ptrs to deallocate */
-	ufs2_daddr_t fb_iblks[NIADDR];	/* indirect blk ptrs to deallocate */
-	ufs2_daddr_t fb_eblks[NXADDR];	/* indirect blk ptrs to deallocate */
+	int	fb_ref;			/* Children outstanding. */
+};
+
+/*
+ * A "freework" structure handles the release of a tree of blocks or a single
+ * block.  Each indirect block in a tree is allocated its own freework
+ * structure so that the indrect block may be freed only when all of its
+ * children are freed.  In this way we enforce the rule that an allocated
+ * block must have a valid path to a root that is journaled.  Each child
+ * block acquires a reference and when the ref hits zero the parent ref
+ * is decremented.  If there is no parent the freeblks ref is decremented.
+ */
+struct freework {
+	struct	worklist fw_list;
+#	define	fw_state fw_list.wk_state
+	LIST_ENTRY(freework) fw_next;		/* Queue for freeblksk. */
+	struct	freeblks *fw_freeblks;		/* Root of operation. */
+	struct	freework *fw_parent;		/* Parent indirect. */
+	ufs2_daddr_t	 fw_blkno;		/* Our block #. */
+	ufs_lbn_t	 fw_lbn;		/* Original lbn before free. */
+	int		 fw_frags;		/* Number of frags. */
+	int		 fw_ref;		/* Number of children out. */
+	int		 fw_off;		/* Current working position. */
+	struct	workhead fw_jwork;		/* Journal work pending. */
+};
+
+/*
+ * A "freedep" structure is allocated to track the completion of a bitmap
+ * write for a freework.  One freedep may cover many freed blocks so long
+ * as they reside in the same cylinder group.  When the cg is written
+ * the freedep decrements the ref on the freework which may permit it
+ * to be freed as well.
+ */
+struct freedep {
+	struct	worklist fd_list;
+	struct	freework *fd_freework;	/* Parent freework. */
 };
 
 /*
@@ -450,6 +546,7 @@ struct freefile {
 	mode_t	fx_mode;		/* mode of inode */
 	ino_t	fx_oldinum;		/* inum of the unlinked file */
 	struct	vnode *fx_devvp;	/* filesystem device vnode */
+	struct	workhead fx_jwork;	/* journal work pending. */
 };
 
 /*
@@ -482,12 +579,11 @@ struct freefile {
  * than zero.
  *
  * The overlaying of da_pagedep and da_previous is done to keep the
- * structure down to 32 bytes in size on a 32-bit machine. If a
- * da_previous entry is present, the pointer to its pagedep is available
- * in the associated dirrem entry. If the DIRCHG flag is set, the
- * da_previous entry is valid; if not set the da_pagedep entry is valid.
- * The DIRCHG flag never changes; it is set when the structure is created
- * if appropriate and is never cleared.
+ * structure down. If a da_previous entry is present, the pointer to its
+ * pagedep is available in the associated dirrem entry. If the DIRCHG flag
+ * is set, the da_previous entry is valid; if not set the da_pagedep entry
+ * is valid. The DIRCHG flag never changes; it is set when the structure
+ * is created if appropriate and is never cleared.
  */
 struct diradd {
 	struct	worklist da_list;	/* id_inowait or id_pendinghd list */
@@ -499,6 +595,7 @@ struct diradd {
 	struct	dirrem *dau_previous;	/* entry being replaced in dir change */
 	struct	pagedep *dau_pagedep;	/* pagedep dependency for addition */
 	} da_un;
+	struct workhead da_jwork;	/* Journal work awaiting completion. */
 };
 #define da_previous da_un.dau_previous
 #define da_pagedep da_un.dau_pagedep
@@ -525,12 +622,13 @@ struct diradd {
  * mkdir structures that reference it. The deletion would be faster if the
  * diradd structure were simply augmented to have two pointers that referenced
  * the associated mkdir's. However, this would increase the size of the diradd
- * structure from 32 to 64-bits to speed a very infrequent operation.
+ * structure to speed a very infrequent operation.
  */
 struct mkdir {
 	struct	worklist md_list;	/* id_inowait or buffer holding dir */
 #	define	md_state md_list.wk_state /* type: MKDIR_PARENT or MKDIR_BODY */
 	struct	diradd *md_diradd;	/* associated diradd */
+	struct	jaddref *md_jaddref;	/* dependent jaddref. */
 	struct	buf *md_buf;		/* MKDIR_BODY: buffer holding dir */
 	LIST_ENTRY(mkdir) md_mkdirs;	/* list of all mkdirs */
 };
@@ -542,20 +640,19 @@ LIST_HEAD(mkdirlist, mkdir) mkdirlisthd;
  * list of the pagedep for the directory page that contains the entry.
  * It is processed after the directory page with the deleted entry has
  * been written to disk.
- *
- * The overlaying of dm_pagedep and dm_dirinum is done to keep the
- * structure down to 32 bytes in size on a 32-bit machine. It works
- * because they are never used concurrently.
  */
 struct dirrem {
 	struct	worklist dm_list;	/* delayed worklist */
 #	define	dm_state dm_list.wk_state /* state of the old directory entry */
 	LIST_ENTRY(dirrem) dm_next;	/* pagedep's list of dirrem's */
+	LIST_ENTRY(dirrem) dm_inonext;	/* inodedep's list of dirrem's */
+	struct	jremrefhd dm_jremrefhd;	/* Pending remove reference deps. */
 	ino_t	dm_oldinum;		/* inum of the removed dir entry */
 	union {
 	struct	pagedep *dmu_pagedep;	/* pagedep dependency for remove */
 	ino_t	dmu_dirinum;		/* parent inode number (for rmdir) */
 	} dm_un;
+	struct workhead dm_jwork;	/* Journal work awaiting completion. */
 };
 #define dm_pagedep dm_un.dmu_pagedep
 #define dm_dirinum dm_un.dmu_dirinum
@@ -577,9 +674,200 @@ struct dirrem {
  * blocks using a similar scheme with the allocindir structures. Rather
  * than adding this level of complexity, we simply write those newly 
  * allocated indirect blocks synchronously as such allocations are rare.
+ * In the case of a new directory the . and .. links are tracked with
+ * a mkdir rather than a pagedep.  In this case we track the mkdir
+ * so it can be released when it is written.  A workhead is used
+ * to simplify canceling a mkdir that is removed by a subsequent dirrem.
  */
 struct newdirblk {
 	struct	worklist db_list;	/* id_inowait or pg_newdirblk */
 #	define	db_state db_list.wk_state /* unused */
 	struct	pagedep *db_pagedep;	/* associated pagedep */
+	struct	workhead db_mkdir;
+};
+
+/*
+ * The inoref structure holds the elements common to jaddref and jremref
+ * so they may easily be queued in-order on the inodedep.
+ */
+struct inoref {
+	struct	worklist if_list;
+#	define	if_state if_list.wk_state
+	TAILQ_ENTRY(inoref) if_deps;	/* Links for inodedep. */
+	struct	jsegdep	*if_jsegdep;
+	off_t		if_diroff;	/* Directory offset. */
+	ino_t		if_ino;		/* Inode number. */
+	ino_t		if_parent;	/* Parent inode number. */
+	nlink_t		if_nlink;	/* nlink before addition. */
+	uint16_t	if_mode;	/* File mode, needed for IFMT. */
+};
+
+/*
+ * A "jaddref" structure tracks a new reference (link count) on an inode
+ * and prevents the link count increase and bitmap allocation until a
+ * journal entry can be written.  Once the journal entry is written,
+ * the inode is put on the pendinghd of the bmsafemap and a diradd or
+ * mkdir entry is placed on the bufwait list of the inode.  The DEPCOMPLETE
+ * flag is used to indicate that all of the required information for writing
+ * the journal entry is present.  MKDIR_BODY and MKDIR_PARENT are used to
+ * differentiate . and .. links from regular file names.  NEWBLOCK indicates
+ * a bitmap is still pending.  If a new reference is canceled by a delete
+ * prior to writing the journal the jaddref write is canceled and the
+ * structure persists to prevent any disk-visible changes until it is
+ * ultimately released when the file is freed or the link is dropped again.
+ */
+struct jaddref {
+	struct	inoref	ja_ref;
+#	define	ja_list	ja_ref.if_list	/* Journal pending or jseg entries. */
+#	define	ja_state ja_ref.if_list.wk_state
+	LIST_ENTRY(jaddref) ja_bmdeps;	/* Links for bmsafemap. */
+	union {
+		struct	diradd	*jau_diradd;	/* Pending diradd. */
+		struct	mkdir	*jau_mkdir;	/* MKDIR_{PARENT,BODY} */
+	} ja_un;
+};
+#define	ja_diradd	ja_un.jau_diradd
+#define	ja_mkdir	ja_un.jau_mkdir
+#define	ja_diroff	ja_ref.if_diroff
+#define	ja_ino		ja_ref.if_ino
+#define	ja_parent	ja_ref.if_parent
+#define	ja_mode		ja_ref.if_mode
+
+/*
+ * A "jremref" structure tracks a removed reference (unlink) on an
+ * inode and prevents the directory remove from proceeding until the
+ * journal entry is written.  Once the journal has been written the remove
+ * may proceed as normal. 
+ */
+struct jremref {
+	struct	inoref	jr_ref;
+#	define	jr_list	jr_ref.if_list	/* Journal pending or jseg entries. */
+#	define	jr_state jr_ref.if_list.wk_state
+	LIST_ENTRY(jremref) jr_deps;	/* Links for pagdep. */
+	struct	dirrem	*jr_dirrem;	/* Back pointer to dirrem. */
+};
+
+struct jmvref {
+	struct	worklist jm_list;
+	LIST_ENTRY(jmvref) jm_deps;
+	struct pagedep	*jm_pagedep;
+	ino_t		jm_parent;
+	ino_t		jm_ino;
+	off_t		jm_oldoff;
+	off_t		jm_newoff;
+};
+
+/*
+ * A "jnewblk" structure tracks a newly allocated block or fragment and
+ * prevents the direct or indirect block pointer as well as the cg bitmap
+ * from being written until it is logged.  After it is logged the jsegdep
+ * is attached to the allocdirect or allocindir until the operation is
+ * completed or reverted.  If the operation is reverted prior to the journal
+ * write the jnewblk structure is maintained to prevent the bitmaps from
+ * reaching the disk.  Ultimately the jnewblk structure will be passed
+ * to the free routine as the in memory cg is modified back to the free
+ * state at which time it can be released.
+ */
+struct jnewblk {
+	struct	worklist jn_list;
+#	define	jn_state jn_list.wk_state
+	struct	jsegdep	*jn_jsegdep;
+	LIST_ENTRY(jnewblk) jn_deps;		/* All jnewblks on bmsafemap */
+	struct	newblk	*jn_newblk;
+	ino_t		jn_ino;
+	ufs_lbn_t	jn_lbn;
+	ufs2_daddr_t	jn_blkno;
+	int		jn_oldfrags;
+	int		jn_frags;
+};
+
+/*
+ * A "jfreeblk" structure tracks the journal write for freeing a block
+ * or tree of blocks.  The block pointer must not be cleared in the inode
+ * or indirect prior to the jfreeblk being written.
+ */
+struct jfreeblk {
+	struct	worklist jf_list;
+#	define	jf_state jf_list.wk_state
+	struct	jsegdep	*jf_jsegdep;
+	struct freeblks	*jf_freeblks;
+	LIST_ENTRY(jfreeblk) jf_deps;
+	ino_t		jf_ino;
+	ufs_lbn_t	jf_lbn;
+	ufs2_daddr_t	jf_blkno;
+	int		jf_frags;
+};
+
+/*
+ * A "jfreefrag" tracks the freeing of a single block when a fragment is
+ * extended or an indirect page is replaced.  It is not part of a larger
+ * freeblks operation.
+ */
+struct jfreefrag {
+	struct	worklist fr_list;
+#	define	fr_state fr_list.wk_state
+	struct	jsegdep	*fr_jsegdep;
+	struct freefrag	*fr_freefrag;
+	ino_t		fr_ino;
+	ufs_lbn_t	fr_lbn;
+	ufs2_daddr_t	fr_blkno;
+	int		fr_frags;
+};
+
+/*
+ * A "jtrunc" journals the intent to truncate an inode to a non-zero
+ * value.  This is done synchronously prior to the synchronous partial
+ * truncation process.  The jsegdep is not released until the truncation
+ * is complete and the truncated inode is fsync'd.
+ */
+struct jtrunc {
+	struct	worklist jt_list;
+	struct	jsegdep	*jt_jsegdep;
+	ino_t		 jt_ino;
+	off_t		 jt_size;
+	int		 jt_extsize;
+};
+
+/*
+ * A "jsegdep" structure tracks a single reference to a written journal
+ * segment so the journal space can be reclaimed when all dependencies
+ * have been written.
+ */
+struct jsegdep {
+	struct	worklist jd_list;
+#	define	jd_state jd_list.wk_state
+	struct	jseg	*jd_seg;
+};
+
+/*
+ * A "jseg" structure contains all of the journal records written in a
+ * single disk write.  jaddref and jremref structures are linked into
+ * js_entries so thay may be completed when the write completes.  The
+ * js_deps array contains as many entries as there are ref counts to
+ * reduce the number of allocations required per journal write to one.
+ */
+struct jseg {
+	struct	worklist js_list;	/* b_deps link for journal */
+#	define	js_state js_list.wk_state
+	struct	workhead js_entries;	/* Entries awaiting write */
+	TAILQ_ENTRY(jseg) js_next;
+	struct	jblocks *js_jblocks;	/* Back pointer to block/seg list */
+	struct	buf *js_buf;		/* Buffer while unwritten */
+	uint64_t js_seq;
+	int	js_size;		/* Allocated size in bytes */
+	int	js_cnt;			/* Total items allocated */
+	int	js_refs;		/* Count of items pending completion */
+};
+
+/*
+ * A 'sbdep' structure tracks the head of the free inode list and
+ * superblock writes.  This makes sure the superblock is always pointing at
+ * the first possible unlinked inode for the suj recovery process.  If a
+ * block write completes and we discover a new head is available the buf
+ * is dirtied and the dep is kept.
+ */
+struct sbdep {
+	struct	worklist sb_list;	/* b_dep linkage */
+	struct	fs	*sb_fs;		/* Filesystem pointer within buf. */
+	struct	ufsmount *sb_ump;
 };
diff --git a/sys/ufs/ufs/dinode.h b/sys/ufs/ufs/dinode.h
index 7f9e7c56496..c75257c8e62 100644
--- a/sys/ufs/ufs/dinode.h
+++ b/sys/ufs/ufs/dinode.h
@@ -146,7 +146,8 @@ struct ufs2_dinode {
 	ufs2_daddr_t	di_db[NDADDR];	/* 112: Direct disk blocks. */
 	ufs2_daddr_t	di_ib[NIADDR];	/* 208: Indirect disk blocks. */
 	u_int64_t	di_modrev;	/* 232: i_modrev for NFSv4 */
-	int64_t		di_spare[2];	/* 240: Reserved; currently unused */
+	ino_t		di_freelink;	/* 240: SUJ: Next unlinked inode. */
+	uint32_t	di_spare[3];	/* 244: Reserved; currently unused */
 };
 
 /*
@@ -167,9 +168,7 @@ struct ufs2_dinode {
 struct ufs1_dinode {
 	u_int16_t	di_mode;	/*   0: IFMT, permissions; see below. */
 	int16_t		di_nlink;	/*   2: File link count. */
-	union {
-		u_int16_t oldids[2];	/*   4: Ffs: old user and group ids. */
-	} di_u;
+	ino_t		di_freelink;	/*   4: SUJ: Next unlinked inode. */
 	u_int64_t	di_size;	/*   8: File byte count. */
 	int32_t		di_atime;	/*  16: Last access time. */
 	int32_t		di_atimensec;	/*  20: Last access time. */
@@ -186,7 +185,5 @@ struct ufs1_dinode {
 	u_int32_t	di_gid;		/* 116: File group. */
 	u_int64_t	di_modrev;	/* 120: i_modrev for NFSv4 */
 };
-#define	di_ogid		di_u.oldids[1]
-#define	di_ouid		di_u.oldids[0]
 
 #endif /* _UFS_UFS_DINODE_H_ */
diff --git a/sys/ufs/ufs/inode.h b/sys/ufs/ufs/inode.h
index 565580e6046..295b12975e2 100644
--- a/sys/ufs/ufs/inode.h
+++ b/sys/ufs/ufs/inode.h
@@ -120,7 +120,7 @@ struct inode {
 #define	IN_CHANGE	0x0002		/* Inode change time update request. */
 #define	IN_UPDATE	0x0004		/* Modification time update request. */
 #define	IN_MODIFIED	0x0008		/* Inode has been modified. */
-#define	IN_RENAME	0x0010		/* Inode is being renamed. */
+#define	IN_NEEDSYNC	0x0010		/* Inode requires fsync. */
 #define	IN_LAZYMOD	0x0040		/* Modified, but don't write yet. */
 #define	IN_SPACECOUNTED	0x0080		/* Blocks to be freed in free count. */
 #define	IN_LAZYACCESS	0x0100		/* Process IN_ACCESS after the
@@ -175,6 +175,7 @@ struct indir {
 /* Determine if soft dependencies are being done */
 #define DOINGSOFTDEP(vp)	((vp)->v_mount->mnt_flag & MNT_SOFTDEP)
 #define DOINGASYNC(vp)		((vp)->v_mount->mnt_kern_flag & MNTK_ASYNC)
+#define DOINGSUJ(vp)		((vp)->v_mount->mnt_kern_flag & MNTK_SUJ)
 
 /* This overlays the fid structure (see mount.h). */
 struct ufid {
diff --git a/sys/ufs/ufs/ufs_dirhash.c b/sys/ufs/ufs/ufs_dirhash.c
index c85fdc8980f..d7c1d0ddb82 100644
--- a/sys/ufs/ufs/ufs_dirhash.c
+++ b/sys/ufs/ufs/ufs_dirhash.c
@@ -68,8 +68,6 @@ __FBSDID("$FreeBSD$");
 
 static MALLOC_DEFINE(M_DIRHASH, "ufs_dirhash", "UFS directory hash tables");
 
-static SYSCTL_NODE(_vfs, OID_AUTO, ufs, CTLFLAG_RD, 0, "UFS filesystem");
-
 static int ufs_mindirhashsize = DIRBLKSIZ * 5;
 SYSCTL_INT(_vfs_ufs, OID_AUTO, dirhash_minsize, CTLFLAG_RW,
     &ufs_mindirhashsize,
diff --git a/sys/ufs/ufs/ufs_extern.h b/sys/ufs/ufs/ufs_extern.h
index b2e4a975730..6658b663fb1 100644
--- a/sys/ufs/ufs/ufs_extern.h
+++ b/sys/ufs/ufs/ufs_extern.h
@@ -57,7 +57,7 @@ int	 ufs_bmap(struct vop_bmap_args *);
 int	 ufs_bmaparray(struct vnode *, ufs2_daddr_t, ufs2_daddr_t *,
 	    struct buf *, int *, int *);
 int	 ufs_fhtovp(struct mount *, struct ufid *, struct vnode **);
-int	 ufs_checkpath(ino_t, struct inode *, struct ucred *);
+int	 ufs_checkpath(ino_t, ino_t, struct inode *, struct ucred *, ino_t *);
 void	 ufs_dirbad(struct inode *, doff_t, char *);
 int	 ufs_dirbadentry(struct vnode *, struct direct *, int);
 int	 ufs_dirempty(struct inode *, ino_t, struct ucred *);
@@ -66,9 +66,11 @@ int	 ufs_extwrite(struct vop_write_args *);
 void	 ufs_makedirentry(struct inode *, struct componentname *,
 	    struct direct *);
 int	 ufs_direnter(struct vnode *, struct vnode *, struct direct *,
-	    struct componentname *, struct buf *);
+	    struct componentname *, struct buf *, int);
 int	 ufs_dirremove(struct vnode *, struct inode *, int, int);
 int	 ufs_dirrewrite(struct inode *, struct inode *, ino_t, int, int);
+int	 ufs_lookup_ino(struct vnode *, struct vnode **, struct componentname *,
+	    ino_t *);
 int	 ufs_getlbns(struct vnode *, ufs2_daddr_t, struct indir *, int *);
 int	 ufs_inactive(struct vop_inactive_args *);
 int	 ufs_init(struct vfsconf *);
@@ -81,19 +83,33 @@ vfs_root_t ufs_root;
 int	 ufs_uninit(struct vfsconf *);
 int	 ufs_vinit(struct mount *, struct vop_vector *, struct vnode **);
 
+#include <sys/sysctl.h>
+SYSCTL_DECL(_vfs_ufs);
+
 /*
  * Soft update function prototypes.
  */
 int	softdep_setup_directory_add(struct buf *, struct inode *, off_t,
 	    ino_t, struct buf *, int);
-void	softdep_change_directoryentry_offset(struct inode *, caddr_t,
-	    caddr_t, caddr_t, int);
+void	softdep_change_directoryentry_offset(struct buf *, struct inode *,
+	    caddr_t, caddr_t, caddr_t, int);
 void	softdep_setup_remove(struct buf *,struct inode *, struct inode *, int);
 void	softdep_setup_directory_change(struct buf *, struct inode *,
 	    struct inode *, ino_t, int);
 void	softdep_change_linkcnt(struct inode *);
 void	softdep_releasefile(struct inode *);
 int	softdep_slowdown(struct vnode *);
+void	softdep_setup_create(struct inode *, struct inode *);
+void	softdep_setup_dotdot_link(struct inode *, struct inode *);
+void	softdep_setup_link(struct inode *, struct inode *);
+void	softdep_setup_mkdir(struct inode *, struct inode *);
+void	softdep_setup_rmdir(struct inode *, struct inode *);
+void	softdep_setup_unlink(struct inode *, struct inode *);
+void	softdep_revert_create(struct inode *, struct inode *);
+void	softdep_revert_dotdot_link(struct inode *, struct inode *);
+void	softdep_revert_link(struct inode *, struct inode *);
+void	softdep_revert_mkdir(struct inode *, struct inode *);
+void	softdep_revert_rmdir(struct inode *, struct inode *);
 
 /*
  * Flags to low-level allocation routines.  The low 16-bits are reserved
diff --git a/sys/ufs/ufs/ufs_lookup.c b/sys/ufs/ufs/ufs_lookup.c
index ab71cf6df55..0030c5264bd 100644
--- a/sys/ufs/ufs/ufs_lookup.c
+++ b/sys/ufs/ufs/ufs_lookup.c
@@ -77,9 +77,6 @@ SYSCTL_INT(_debug, OID_AUTO, dircheck, CTLFLAG_RW, &dirchk, 0, "");
 /* true if old FS format...*/
 #define OFSFMT(vp)	((vp)->v_mount->mnt_maxsymlinklen <= 0)
 
-static int ufs_lookup_(struct vnode *, struct vnode **, struct componentname *,
-    ino_t *);
-
 static int
 ufs_delete_denied(struct vnode *vdp, struct vnode *tdp, struct ucred *cred,
     struct thread *td)
@@ -189,11 +186,11 @@ ufs_lookup(ap)
 	} */ *ap;
 {
 
-	return (ufs_lookup_(ap->a_dvp, ap->a_vpp, ap->a_cnp, NULL));
+	return (ufs_lookup_ino(ap->a_dvp, ap->a_vpp, ap->a_cnp, NULL));
 }
 
-static int
-ufs_lookup_(struct vnode *vdp, struct vnode **vpp, struct componentname *cnp,
+int
+ufs_lookup_ino(struct vnode *vdp, struct vnode **vpp, struct componentname *cnp,
     ino_t *dd_ino)
 {
 	struct inode *dp;		/* inode for directory being searched */
@@ -524,6 +521,8 @@ notfound:
 	return (ENOENT);
 
 found:
+	if (dd_ino != NULL)
+		*dd_ino = ino;
 	if (numdirpasses == 2)
 		nchstats.ncs_pass2++;
 	/*
@@ -546,11 +545,6 @@ found:
 	if ((flags & ISLASTCN) && nameiop == LOOKUP)
 		dp->i_diroff = i_offset &~ (DIRBLKSIZ - 1);
 
-	if (dd_ino != NULL) {
-		*dd_ino = ino;
-		return (0);
-	}
-
 	/*
 	 * If deleting, and at end of pathname, return
 	 * parameters which can be used to remove file.
@@ -558,17 +552,6 @@ found:
 	if (nameiop == DELETE && (flags & ISLASTCN)) {
 		if (flags & LOCKPARENT)
 			ASSERT_VOP_ELOCKED(vdp, __FUNCTION__);
-		if ((error = VFS_VGET(vdp->v_mount, ino,
-		    LK_EXCLUSIVE, &tdp)) != 0)
-			return (error);
-
-		error = ufs_delete_denied(vdp, tdp, cred, cnp->cn_thread);
-		if (error) {
-			vput(tdp);
-			return (error);
-		}
-
-
 		/*
 		 * Return pointer to current entry in dp->i_offset,
 		 * and distance past previous entry (if there
@@ -585,6 +568,16 @@ found:
 			dp->i_count = 0;
 		else
 			dp->i_count = dp->i_offset - prevoff;
+		if (dd_ino != NULL)
+			return (0);
+		if ((error = VFS_VGET(vdp->v_mount, ino,
+		    LK_EXCLUSIVE, &tdp)) != 0)
+			return (error);
+		error = ufs_delete_denied(vdp, tdp, cred, cnp->cn_thread);
+		if (error) {
+			vput(tdp);
+			return (error);
+		}
 		if (dp->i_number == ino) {
 			VREF(vdp);
 			*vpp = vdp;
@@ -616,6 +609,8 @@ found:
 		dp->i_offset = i_offset;
 		if (dp->i_number == ino)
 			return (EISDIR);
+		if (dd_ino != NULL)
+			return (0);
 		if ((error = VFS_VGET(vdp->v_mount, ino,
 		    LK_EXCLUSIVE, &tdp)) != 0)
 			return (error);
@@ -650,6 +645,8 @@ found:
 		cnp->cn_flags |= SAVENAME;
 		return (0);
 	}
+	if (dd_ino != NULL)
+		return (0);
 
 	/*
 	 * Step through the translation in the name.  We do not `vput' the
@@ -681,7 +678,7 @@ found:
 		 * to the inode we looked up before vdp lock was
 		 * dropped.
 		 */
-		error = ufs_lookup_(pdp, NULL, cnp, &ino1);
+		error = ufs_lookup_ino(pdp, NULL, cnp, &ino1);
 		if (error) {
 			vput(tdp);
 			return (error);
@@ -833,12 +830,13 @@ ufs_makedirentry(ip, cnp, newdirp)
  * soft dependency code).
  */
 int
-ufs_direnter(dvp, tvp, dirp, cnp, newdirbp)
+ufs_direnter(dvp, tvp, dirp, cnp, newdirbp, isrename)
 	struct vnode *dvp;
 	struct vnode *tvp;
 	struct direct *dirp;
 	struct componentname *cnp;
 	struct buf *newdirbp;
+	int isrename;
 {
 	struct ucred *cr;
 	struct thread *td;
@@ -911,22 +909,28 @@ ufs_direnter(dvp, tvp, dirp, cnp, newdirbp)
 				blkoff += DIRBLKSIZ;
 			}
 			if (softdep_setup_directory_add(bp, dp, dp->i_offset,
-			    dirp->d_ino, newdirbp, 1) == 0) {
-				bdwrite(bp);
+			    dirp->d_ino, newdirbp, 1))
+				dp->i_flag |= IN_NEEDSYNC;
+			if (newdirbp)
+				bdwrite(newdirbp);
+			bdwrite(bp);
+			if ((dp->i_flag & IN_NEEDSYNC) == 0)
 				return (UFS_UPDATE(dvp, 0));
-			}
-			/* We have just allocated a directory block in an
-			 * indirect block. Rather than tracking when it gets
-			 * claimed by the inode, we simply do a VOP_FSYNC
-			 * now to ensure that it is there (in case the user
-			 * does a future fsync). Note that we have to unlock
-			 * the inode for the entry that we just entered, as
-			 * the VOP_FSYNC may need to lock other inodes which
-			 * can lead to deadlock if we also hold a lock on
-			 * the newly entered node.
+			/*
+			 * We have just allocated a directory block in an
+			 * indirect block.  We must prevent holes in the
+			 * directory created if directory entries are
+			 * written out of order.  To accomplish this we
+			 * fsync when we extend a directory into indirects.
+			 * During rename it's not safe to drop the tvp lock
+			 * so sync must be delayed until it is.
+			 *
+			 * This synchronous step could be removed if fsck and
+			 * the kernel were taught to fill in sparse
+			 * directories rather than panic.
 			 */
-			if ((error = bwrite(bp)))
-				return (error);
+			if (isrename)
+				return (0);
 			if (tvp != NULL)
 				VOP_UNLOCK(tvp, 0);
 			error = VOP_FSYNC(dvp, MNT_WAIT, td);
@@ -1015,7 +1019,7 @@ ufs_direnter(dvp, tvp, dirp, cnp, newdirbp)
 			    dp->i_offset + ((char *)ep - dirbuf));
 #endif
 		if (DOINGSOFTDEP(dvp))
-			softdep_change_directoryentry_offset(dp, dirbuf,
+			softdep_change_directoryentry_offset(bp, dp, dirbuf,
 			    (caddr_t)nep, (caddr_t)ep, dsize); 
 		else
 			bcopy((caddr_t)nep, (caddr_t)ep, dsize);
@@ -1067,6 +1071,8 @@ ufs_direnter(dvp, tvp, dirp, cnp, newdirbp)
 		(void) softdep_setup_directory_add(bp, dp,
 		    dp->i_offset + (caddr_t)ep - dirbuf,
 		    dirp->d_ino, newdirbp, 0);
+		if (newdirbp != NULL)
+			bdwrite(newdirbp);
 		bdwrite(bp);
 	} else {
 		if (DOINGASYNC(dvp)) {
@@ -1084,7 +1090,8 @@ ufs_direnter(dvp, tvp, dirp, cnp, newdirbp)
 	 * lock other inodes which can lead to deadlock if we also hold a
 	 * lock on the newly entered node.
 	 */
-	if (error == 0 && dp->i_endoff && dp->i_endoff < dp->i_size) {
+	if (isrename == 0 && error == 0 &&
+	    dp->i_endoff && dp->i_endoff < dp->i_size) {
 		if (tvp != NULL)
 			VOP_UNLOCK(tvp, 0);
 #ifdef UFS_DIRHASH
@@ -1125,6 +1132,19 @@ ufs_dirremove(dvp, ip, flags, isrmdir)
 
 	dp = VTOI(dvp);
 
+	/*
+	 * Adjust the link count early so softdep can block if necessary.
+	 */
+	if (ip) {
+		ip->i_effnlink--;
+		if (DOINGSOFTDEP(dvp)) {
+			softdep_setup_unlink(dp, ip);
+		} else {
+			ip->i_nlink--;
+			DIP_SET(ip, i_nlink, ip->i_nlink);
+			ip->i_flag |= IN_CHANGE;
+		}
+	}
 	if (flags & DOWHITEOUT) {
 		/*
 		 * Whiteout entry: set d_ino to WINO.
@@ -1154,6 +1174,9 @@ ufs_dirremove(dvp, ip, flags, isrmdir)
 	if (dp->i_dirhash != NULL)
 		ufsdirhash_remove(dp, rep, dp->i_offset);
 #endif
+	if (ip && rep->d_ino != ip->i_number)
+		panic("ufs_dirremove: ip %d does not match dirent ino %d\n",
+		    ip->i_number, rep->d_ino);
 	if (dp->i_count == 0) {
 		/*
 		 * First entry in block: set d_ino to zero.
@@ -1172,31 +1195,20 @@ ufs_dirremove(dvp, ip, flags, isrmdir)
 		    dp->i_offset & ~(DIRBLKSIZ - 1));
 #endif
 out:
+	error = 0;
 	if (DOINGSOFTDEP(dvp)) {
-		if (ip) {
-			ip->i_effnlink--;
-			softdep_change_linkcnt(ip);
+		if (ip)
 			softdep_setup_remove(bp, dp, ip, isrmdir);
-		}
-		if (softdep_slowdown(dvp)) {
+		if (softdep_slowdown(dvp))
 			error = bwrite(bp);
-		} else {
+		else
 			bdwrite(bp);
-			error = 0;
-		}
 	} else {
-		if (ip) {
-			ip->i_effnlink--;
-			ip->i_nlink--;
-			DIP_SET(ip, i_nlink, ip->i_nlink);
-			ip->i_flag |= IN_CHANGE;
-		}
 		if (flags & DOWHITEOUT)
 			error = bwrite(bp);
-		else if (DOINGASYNC(dvp) && dp->i_count != 0) {
+		else if (DOINGASYNC(dvp) && dp->i_count != 0)
 			bdwrite(bp);
-			error = 0;
-		} else
+		else
 			error = bwrite(bp);
 	}
 	dp->i_flag |= IN_CHANGE | IN_UPDATE;
@@ -1229,6 +1241,19 @@ ufs_dirrewrite(dp, oip, newinum, newtype, isrmdir)
 	struct vnode *vdp = ITOV(dp);
 	int error;
 
+	/*
+	 * Drop the link before we lock the buf so softdep can block if
+	 * necessary.
+	 */
+	oip->i_effnlink--;
+	if (DOINGSOFTDEP(vdp)) {
+		softdep_setup_unlink(dp, oip);
+	} else {
+		oip->i_nlink--;
+		DIP_SET(oip, i_nlink, oip->i_nlink);
+		oip->i_flag |= IN_CHANGE;
+	}
+
 	error = UFS_BLKATOFF(vdp, (off_t)dp->i_offset, (char **)&ep, &bp);
 	if (error)
 		return (error);
@@ -1240,15 +1265,10 @@ ufs_dirrewrite(dp, oip, newinum, newtype, isrmdir)
 	ep->d_ino = newinum;
 	if (!OFSFMT(vdp))
 		ep->d_type = newtype;
-	oip->i_effnlink--;
 	if (DOINGSOFTDEP(vdp)) {
-		softdep_change_linkcnt(oip);
 		softdep_setup_directory_change(bp, dp, oip, newinum, isrmdir);
 		bdwrite(bp);
 	} else {
-		oip->i_nlink--;
-		DIP_SET(oip, i_nlink, oip->i_nlink);
-		oip->i_flag |= IN_CHANGE;
 		if (DOINGASYNC(vdp)) {
 			bdwrite(bp);
 			error = 0;
@@ -1363,25 +1383,25 @@ ufs_dir_dd_ino(struct vnode *vp, struct ucred *cred, ino_t *dd_ino)
 
 /*
  * Check if source directory is in the path of the target directory.
- * Target is supplied locked, source is unlocked.
- * The target is always vput before returning.
  */
 int
-ufs_checkpath(ino_t source_ino, struct inode *target, struct ucred *cred)
+ufs_checkpath(ino_t source_ino, ino_t parent_ino, struct inode *target, struct ucred *cred, ino_t *wait_ino)
 {
-	struct vnode *vp, *vp1;
+	struct mount *mp;
+	struct vnode *tvp, *vp, *vp1;
 	int error;
 	ino_t dd_ino;
 
-	vp = ITOV(target);
-	if (target->i_number == source_ino) {
-		error = EEXIST;
-		goto out;
-	}
-	error = 0;
+	vp = tvp = ITOV(target);
+	mp = vp->v_mount;
+	*wait_ino = 0;
+	if (target->i_number == source_ino)
+		return (EEXIST);
+	if (target->i_number == parent_ino)
+		return (0);
 	if (target->i_number == ROOTINO)
-		goto out;
-
+		return (0);
+	error = 0;
 	for (;;) {
 		error = ufs_dir_dd_ino(vp, cred, &dd_ino);
 		if (error != 0)
@@ -1392,9 +1412,13 @@ ufs_checkpath(ino_t source_ino, struct inode *target, struct ucred *cred)
 		}
 		if (dd_ino == ROOTINO)
 			break;
-		error = vn_vget_ino(vp, dd_ino, LK_EXCLUSIVE, &vp1);
-		if (error != 0)
+		if (dd_ino == parent_ino)
 			break;
+		error = VFS_VGET(mp, dd_ino, LK_SHARED | LK_NOWAIT, &vp1);
+		if (error != 0) {
+			*wait_ino = dd_ino;
+			break;
+		}
 		/* Recheck that ".." still points to vp1 after relock of vp */
 		error = ufs_dir_dd_ino(vp, cred, &dd_ino);
 		if (error != 0) {
@@ -1406,14 +1430,14 @@ ufs_checkpath(ino_t source_ino, struct inode *target, struct ucred *cred)
 			vput(vp1);
 			continue;
 		}
-		vput(vp);
+		if (vp != tvp)
+			vput(vp);
 		vp = vp1;
 	}
 
-out:
 	if (error == ENOTDIR)
-		printf("checkpath: .. not a directory\n");
-	if (vp != NULL)
+		panic("checkpath: .. not a directory\n");
+	if (vp != tvp)
 		vput(vp);
 	return (error);
 }
diff --git a/sys/ufs/ufs/ufs_vnops.c b/sys/ufs/ufs/ufs_vnops.c
index 9d4d93dbc8f..f8d45cfceb8 100644
--- a/sys/ufs/ufs/ufs_vnops.c
+++ b/sys/ufs/ufs/ufs_vnops.c
@@ -114,6 +114,8 @@ static vop_close_t	ufsfifo_close;
 static vop_kqfilter_t	ufsfifo_kqfilter;
 static vop_pathconf_t	ufsfifo_pathconf;
 
+SYSCTL_NODE(_vfs, OID_AUTO, ufs, CTLFLAG_RD, 0, "UFS filesystem");
+
 /*
  * A virgin directory (no blushing please).
  */
@@ -974,6 +976,9 @@ ufs_link(ap)
 		error = EXDEV;
 		goto out;
 	}
+	if (VTOI(tdvp)->i_effnlink < 2)
+		panic("ufs_link: Bad link count %d on parent",
+		    VTOI(tdvp)->i_effnlink);
 	ip = VTOI(vp);
 	if ((nlink_t)ip->i_nlink >= LINK_MAX) {
 		error = EMLINK;
@@ -988,11 +993,11 @@ ufs_link(ap)
 	DIP_SET(ip, i_nlink, ip->i_nlink);
 	ip->i_flag |= IN_CHANGE;
 	if (DOINGSOFTDEP(vp))
-		softdep_change_linkcnt(ip);
+		softdep_setup_link(VTOI(tdvp), ip);
 	error = UFS_UPDATE(vp, !(DOINGSOFTDEP(vp) | DOINGASYNC(vp)));
 	if (!error) {
 		ufs_makedirentry(ip, cnp, &newdir);
-		error = ufs_direnter(tdvp, vp, &newdir, cnp, NULL);
+		error = ufs_direnter(tdvp, vp, &newdir, cnp, NULL, 0);
 	}
 
 	if (error) {
@@ -1001,7 +1006,7 @@ ufs_link(ap)
 		DIP_SET(ip, i_nlink, ip->i_nlink);
 		ip->i_flag |= IN_CHANGE;
 		if (DOINGSOFTDEP(vp))
-			softdep_change_linkcnt(ip);
+			softdep_revert_link(VTOI(tdvp), ip);
 	}
 out:
 	return (error);
@@ -1043,7 +1048,7 @@ ufs_whiteout(ap)
 		newdir.d_namlen = cnp->cn_namelen;
 		bcopy(cnp->cn_nameptr, newdir.d_name, (unsigned)cnp->cn_namelen + 1);
 		newdir.d_type = DT_WHT;
-		error = ufs_direnter(dvp, NULL, &newdir, cnp, NULL);
+		error = ufs_direnter(dvp, NULL, &newdir, cnp, NULL, 0);
 		break;
 
 	case DELETE:
@@ -1062,6 +1067,11 @@ ufs_whiteout(ap)
 	return (error);
 }
 
+static volatile int rename_restarts;
+SYSCTL_INT(_vfs_ufs, OID_AUTO, rename_restarts, CTLFLAG_RD,
+    __DEVOLATILE(int *, &rename_restarts), 0,
+    "Times rename had to restart due to lock contention");
+
 /*
  * Rename system call.
  * 	rename("foo", "bar");
@@ -1101,111 +1111,183 @@ ufs_rename(ap)
 	struct vnode *tdvp = ap->a_tdvp;
 	struct vnode *fvp = ap->a_fvp;
 	struct vnode *fdvp = ap->a_fdvp;
+	struct vnode *nvp;
 	struct componentname *tcnp = ap->a_tcnp;
 	struct componentname *fcnp = ap->a_fcnp;
 	struct thread *td = fcnp->cn_thread;
-	struct inode *ip, *xp, *dp;
+	struct inode *fip, *tip, *tdp, *fdp;
 	struct direct newdir;
-	int doingdirectory = 0, oldparent = 0, newparent = 0;
+	off_t endoff;
+	int doingdirectory, newparent;
 	int error = 0, ioflag;
-	ino_t fvp_ino;
+	struct mount *mp;
+	ino_t ino;
 
 #ifdef INVARIANTS
 	if ((tcnp->cn_flags & HASBUF) == 0 ||
 	    (fcnp->cn_flags & HASBUF) == 0)
 		panic("ufs_rename: no name");
 #endif
+	endoff = 0;
+	mp = tdvp->v_mount;
+	VOP_UNLOCK(tdvp, 0);
+	if (tvp && tvp != tdvp)
+		VOP_UNLOCK(tvp, 0);
 	/*
 	 * Check for cross-device rename.
 	 */
 	if ((fvp->v_mount != tdvp->v_mount) ||
 	    (tvp && (fvp->v_mount != tvp->v_mount))) {
 		error = EXDEV;
-abortit:
-		if (tdvp == tvp)
-			vrele(tdvp);
-		else
-			vput(tdvp);
-		if (tvp)
-			vput(tvp);
-		vrele(fdvp);
-		vrele(fvp);
-		return (error);
+		mp = NULL;
+		goto releout;
 	}
-
+	error = vfs_busy(mp, 0);
+	if (error) {
+		mp = NULL;
+		goto releout;
+	}
+relock:
+	/* 
+	 * We need to acquire 2 to 4 locks depending on whether tvp is NULL
+	 * and fdvp and tdvp are the same directory.  Subsequently we need
+	 * to double-check all paths and in the directory rename case we
+	 * need to verify that we are not creating a directory loop.  To
+	 * handle this we acquire all but fdvp using non-blocking
+	 * acquisitions.  If we fail to acquire any lock in the path we will
+	 * drop all held locks, acquire the new lock in a blocking fashion,
+	 * and then release it and restart the rename.  This acquire/release
+	 * step ensures that we do not spin on a lock waiting for release.
+	 */
+	error = vn_lock(fdvp, LK_EXCLUSIVE);
+	if (error)
+		goto releout;
+	if (vn_lock(tdvp, LK_EXCLUSIVE | LK_NOWAIT) != 0) {
+		VOP_UNLOCK(fdvp, 0);
+		error = vn_lock(tdvp, LK_EXCLUSIVE);
+		if (error)
+			goto releout;
+		VOP_UNLOCK(tdvp, 0);
+		atomic_add_int(&rename_restarts, 1);
+		goto relock;
+	}
+	/*
+	 * Re-resolve fvp to be certain it still exists and fetch the
+	 * correct vnode.
+	 */
+	error = ufs_lookup_ino(fdvp, NULL, fcnp, &ino);
+	if (error) {
+		VOP_UNLOCK(fdvp, 0);
+		VOP_UNLOCK(tdvp, 0);
+		goto releout;
+	}
+	error = VFS_VGET(mp, ino, LK_EXCLUSIVE | LK_NOWAIT, &nvp);
+	if (error) {
+		VOP_UNLOCK(fdvp, 0);
+		VOP_UNLOCK(tdvp, 0);
+		if (error != EBUSY)
+			goto releout;
+		error = VFS_VGET(mp, ino, LK_EXCLUSIVE, &nvp);
+		if (error != 0)
+			goto releout;
+		VOP_UNLOCK(nvp, 0);
+		vrele(fvp);
+		fvp = nvp;
+		atomic_add_int(&rename_restarts, 1);
+		goto relock;
+	}
+	vrele(fvp);
+	fvp = nvp;
+	/*
+	 * Re-resolve tvp and acquire the vnode lock if present.
+	 */
+	error = ufs_lookup_ino(tdvp, NULL, tcnp, &ino);
+	if (error != 0 && error != EJUSTRETURN) {
+		VOP_UNLOCK(fdvp, 0);
+		VOP_UNLOCK(tdvp, 0);
+		VOP_UNLOCK(fvp, 0);
+		goto releout;
+	}
+	/*
+	 * If tvp disappeared we just carry on.
+	 */
+	if (error == EJUSTRETURN && tvp != NULL) {
+		vrele(tvp);
+		tvp = NULL;
+	}
+	/*
+	 * Get the tvp ino if the lookup succeeded.  We may have to restart
+	 * if the non-blocking acquire fails.
+	 */
+	if (error == 0) {
+		nvp = NULL;
+		error = VFS_VGET(mp, ino, LK_EXCLUSIVE | LK_NOWAIT, &nvp);
+		if (tvp)
+			vrele(tvp);
+		tvp = nvp;
+		if (error) {
+			VOP_UNLOCK(fdvp, 0);
+			VOP_UNLOCK(tdvp, 0);
+			VOP_UNLOCK(fvp, 0);
+			if (error != EBUSY)
+				goto releout;
+			error = VFS_VGET(mp, ino, LK_EXCLUSIVE, &nvp);
+			if (error != 0)
+				goto releout;
+			VOP_UNLOCK(nvp, 0);
+			atomic_add_int(&rename_restarts, 1);
+			goto relock;
+		}
+	}
+	fdp = VTOI(fdvp);
+	fip = VTOI(fvp);
+	tdp = VTOI(tdvp);
+	tip = NULL;
+	if (tvp)
+		tip = VTOI(tvp);
 	if (tvp && ((VTOI(tvp)->i_flags & (NOUNLINK | IMMUTABLE | APPEND)) ||
 	    (VTOI(tdvp)->i_flags & APPEND))) {
 		error = EPERM;
-		goto abortit;
+		goto unlockout;
 	}
-
 	/*
 	 * Renaming a file to itself has no effect.  The upper layers should
-	 * not call us in that case.  Temporarily just warn if they do.
+	 * not call us in that case.  However, things could change after
+	 * we drop the locks above.
 	 */
 	if (fvp == tvp) {
-		printf("ufs_rename: fvp == tvp (can't happen)\n");
 		error = 0;
-		goto abortit;
+		goto unlockout;
 	}
-
-	if ((error = vn_lock(fvp, LK_EXCLUSIVE)) != 0)
-		goto abortit;
-	dp = VTOI(fdvp);
-	ip = VTOI(fvp);
-	if (ip->i_nlink >= LINK_MAX) {
-		VOP_UNLOCK(fvp, 0);
+	doingdirectory = 0;
+	newparent = 0;
+	ino = fip->i_number;
+	if (fip->i_nlink >= LINK_MAX) {
 		error = EMLINK;
-		goto abortit;
+		goto unlockout;
 	}
-	if ((ip->i_flags & (NOUNLINK | IMMUTABLE | APPEND))
-	    || (dp->i_flags & APPEND)) {
-		VOP_UNLOCK(fvp, 0);
+	if ((fip->i_flags & (NOUNLINK | IMMUTABLE | APPEND))
+	    || (fdp->i_flags & APPEND)) {
 		error = EPERM;
-		goto abortit;
+		goto unlockout;
 	}
-	if ((ip->i_mode & IFMT) == IFDIR) {
+	if ((fip->i_mode & IFMT) == IFDIR) {
 		/*
 		 * Avoid ".", "..", and aliases of "." for obvious reasons.
 		 */
 		if ((fcnp->cn_namelen == 1 && fcnp->cn_nameptr[0] == '.') ||
-		    dp == ip || (fcnp->cn_flags | tcnp->cn_flags) & ISDOTDOT ||
-		    (ip->i_flag & IN_RENAME)) {
-			VOP_UNLOCK(fvp, 0);
+		    fdp == fip ||
+		    (fcnp->cn_flags | tcnp->cn_flags) & ISDOTDOT) {
 			error = EINVAL;
-			goto abortit;
+			goto unlockout;
 		}
-		ip->i_flag |= IN_RENAME;
-		oldparent = dp->i_number;
+		if (fdp->i_number != tdp->i_number)
+			newparent = tdp->i_number;
 		doingdirectory = 1;
 	}
-	vrele(fdvp);
-
-	/*
-	 * When the target exists, both the directory
-	 * and target vnodes are returned locked.
-	 */
-	dp = VTOI(tdvp);
-	xp = NULL;
-	if (tvp)
-		xp = VTOI(tvp);
-
-	/*
-	 * 1) Bump link count while we're moving stuff
-	 *    around.  If we crash somewhere before
-	 *    completing our work, the link count
-	 *    may be wrong, but correctable.
-	 */
-	ip->i_effnlink++;
-	ip->i_nlink++;
-	DIP_SET(ip, i_nlink, ip->i_nlink);
-	ip->i_flag |= IN_CHANGE;
-	if (DOINGSOFTDEP(fvp))
-		softdep_change_linkcnt(ip);
-	if ((error = UFS_UPDATE(fvp, !(DOINGSOFTDEP(fvp) |
-				       DOINGASYNC(fvp)))) != 0) {
-		VOP_UNLOCK(fvp, 0);
-		goto bad;
+	if (fvp->v_mountedhere != NULL || (tvp && tvp->v_mountedhere != NULL)) {
+		error = EXDEV;
+		goto unlockout;
 	}
 
 	/*
@@ -1214,35 +1296,55 @@ abortit:
 	 * directory hierarchy above the target, as this would
 	 * orphan everything below the source directory. Also
 	 * the user must have write permission in the source so
-	 * as to be able to change "..". We must repeat the call
-	 * to namei, as the parent directory is unlocked by the
-	 * call to checkpath().
+	 * as to be able to change "..".
 	 */
-	error = VOP_ACCESS(fvp, VWRITE, tcnp->cn_cred, tcnp->cn_thread);
-	fvp_ino = ip->i_number;
-	VOP_UNLOCK(fvp, 0);
-	if (oldparent != dp->i_number)
-		newparent = dp->i_number;
 	if (doingdirectory && newparent) {
-		if (error)	/* write access check above */
-			goto bad;
-		if (xp != NULL)
-			vput(tvp);
-		error = ufs_checkpath(fvp_ino, dp, tcnp->cn_cred);
+		error = VOP_ACCESS(fvp, VWRITE, tcnp->cn_cred, tcnp->cn_thread);
 		if (error)
-			goto out;
+			goto unlockout;
+		error = ufs_checkpath(ino, fdp->i_number, tdp, tcnp->cn_cred,
+		    &ino);
+		/*
+		 * We encountered a lock that we have to wait for.  Unlock
+		 * everything else and VGET before restarting.
+		 */
+		if (ino) {
+			VOP_UNLOCK(fdvp, 0);
+			VOP_UNLOCK(fvp, 0);
+			VOP_UNLOCK(tdvp, 0);
+			if (tvp)
+				VOP_UNLOCK(tvp, 0);
+			error = VFS_VGET(mp, ino, LK_SHARED, &nvp);
+			if (error == 0)
+				vput(nvp);
+			atomic_add_int(&rename_restarts, 1);
+			goto relock;
+		}
+		if (error)
+			goto unlockout;
 		if ((tcnp->cn_flags & SAVESTART) == 0)
 			panic("ufs_rename: lost to startdir");
-		VREF(tdvp);
-		error = relookup(tdvp, &tvp, tcnp);
-		if (error)
-			goto out;
-		vrele(tdvp);
-		dp = VTOI(tdvp);
-		xp = NULL;
-		if (tvp)
-			xp = VTOI(tvp);
 	}
+	if (fip->i_effnlink == 0 || fdp->i_effnlink == 0 ||
+	    tdp->i_effnlink == 0)
+		panic("Bad effnlink fip %p, fdp %p, tdp %p", fip, fdp, tdp);
+
+	/*
+	 * 1) Bump link count while we're moving stuff
+	 *    around.  If we crash somewhere before
+	 *    completing our work, the link count
+	 *    may be wrong, but correctable.
+	 */
+	fip->i_effnlink++;
+	fip->i_nlink++;
+	DIP_SET(fip, i_nlink, fip->i_nlink);
+	fip->i_flag |= IN_CHANGE;
+	if (DOINGSOFTDEP(fvp))
+		softdep_setup_link(tdp, fip);
+	error = UFS_UPDATE(fvp, !(DOINGSOFTDEP(fvp) | DOINGASYNC(fvp)));
+	if (error)
+		goto bad;
+
 	/*
 	 * 2) If target doesn't exist, link the target
 	 *    to the source and unlink the source.
@@ -1250,52 +1352,37 @@ abortit:
 	 *    entry to reference the source inode and
 	 *    expunge the original entry's existence.
 	 */
-	if (xp == NULL) {
-		if (dp->i_dev != ip->i_dev)
+	if (tip == NULL) {
+		if (tdp->i_dev != fip->i_dev)
 			panic("ufs_rename: EXDEV");
-		/*
-		 * Account for ".." in new directory.
-		 * When source and destination have the same
-		 * parent we don't fool with the link count.
-		 */
 		if (doingdirectory && newparent) {
-			if ((nlink_t)dp->i_nlink >= LINK_MAX) {
+			/*
+			 * Account for ".." in new directory.
+			 * When source and destination have the same
+			 * parent we don't adjust the link count.  The
+			 * actual link modification is completed when
+			 * .. is rewritten below.
+			 */
+			if ((nlink_t)tdp->i_nlink >= LINK_MAX) {
 				error = EMLINK;
 				goto bad;
 			}
-			dp->i_effnlink++;
-			dp->i_nlink++;
-			DIP_SET(dp, i_nlink, dp->i_nlink);
-			dp->i_flag |= IN_CHANGE;
-			if (DOINGSOFTDEP(tdvp))
-				softdep_change_linkcnt(dp);
-			error = UFS_UPDATE(tdvp, !(DOINGSOFTDEP(tdvp) |
-						   DOINGASYNC(tdvp)));
-			if (error)
-				goto bad;
 		}
-		ufs_makedirentry(ip, tcnp, &newdir);
-		error = ufs_direnter(tdvp, NULL, &newdir, tcnp, NULL);
-		if (error) {
-			if (doingdirectory && newparent) {
-				dp->i_effnlink--;
-				dp->i_nlink--;
-				DIP_SET(dp, i_nlink, dp->i_nlink);
-				dp->i_flag |= IN_CHANGE;
-				if (DOINGSOFTDEP(tdvp))
-					softdep_change_linkcnt(dp);
-				(void)UFS_UPDATE(tdvp, 1);
-			}
+		ufs_makedirentry(fip, tcnp, &newdir);
+		error = ufs_direnter(tdvp, NULL, &newdir, tcnp, NULL, 1);
+		if (error)
 			goto bad;
-		}
-		vput(tdvp);
+		/* Setup tdvp for directory compaction if needed. */
+		if (tdp->i_count && tdp->i_endoff &&
+		    tdp->i_endoff < tdp->i_size)
+			endoff = tdp->i_endoff;
 	} else {
-		if (xp->i_dev != dp->i_dev || xp->i_dev != ip->i_dev)
+		if (tip->i_dev != tdp->i_dev || tip->i_dev != fip->i_dev)
 			panic("ufs_rename: EXDEV");
 		/*
 		 * Short circuit rename(foo, foo).
 		 */
-		if (xp->i_number == ip->i_number)
+		if (tip->i_number == fip->i_number)
 			panic("ufs_rename: same file");
 		/*
 		 * If the parent directory is "sticky", then the caller
@@ -1303,7 +1390,7 @@ abortit:
 		 * destination of the rename.  This implements append-only
 		 * directories.
 		 */
-		if ((dp->i_mode & S_ISTXT) &&
+		if ((tdp->i_mode & S_ISTXT) &&
 		    VOP_ACCESS(tdvp, VADMIN, tcnp->cn_cred, td) &&
 		    VOP_ACCESS(tvp, VADMIN, tcnp->cn_cred, td)) {
 			error = EPERM;
@@ -1314,9 +1401,9 @@ abortit:
 		 * to it. Also, ensure source and target are compatible
 		 * (both directories, or both not directories).
 		 */
-		if ((xp->i_mode&IFMT) == IFDIR) {
-			if ((xp->i_effnlink > 2) ||
-			    !ufs_dirempty(xp, dp->i_number, tcnp->cn_cred)) {
+		if ((tip->i_mode & IFMT) == IFDIR) {
+			if ((tip->i_effnlink > 2) ||
+			    !ufs_dirempty(tip, tdp->i_number, tcnp->cn_cred)) {
 				error = ENOTEMPTY;
 				goto bad;
 			}
@@ -1329,20 +1416,30 @@ abortit:
 			error = EISDIR;
 			goto bad;
 		}
-		error = ufs_dirrewrite(dp, xp, ip->i_number,
-		    IFTODT(ip->i_mode),
-		    (doingdirectory && newparent) ? newparent : doingdirectory);
-		if (error)
-			goto bad;
 		if (doingdirectory) {
 			if (!newparent) {
-				dp->i_effnlink--;
+				tdp->i_effnlink--;
 				if (DOINGSOFTDEP(tdvp))
-					softdep_change_linkcnt(dp);
+					softdep_change_linkcnt(tdp);
 			}
-			xp->i_effnlink--;
+			tip->i_effnlink--;
 			if (DOINGSOFTDEP(tvp))
-				softdep_change_linkcnt(xp);
+				softdep_change_linkcnt(tip);
+		}
+		error = ufs_dirrewrite(tdp, tip, fip->i_number,
+		    IFTODT(fip->i_mode),
+		    (doingdirectory && newparent) ? newparent : doingdirectory);
+		if (error) {
+			if (doingdirectory) {
+				if (!newparent) {
+					tdp->i_effnlink++;
+					if (DOINGSOFTDEP(tdvp))
+						softdep_change_linkcnt(tdp);
+				}
+				tip->i_effnlink++;
+				if (DOINGSOFTDEP(tvp))
+					softdep_change_linkcnt(tip);
+			}
 		}
 		if (doingdirectory && !DOINGSOFTDEP(tvp)) {
 			/*
@@ -1357,115 +1454,107 @@ abortit:
 			 * them now.
 			 */
 			if (!newparent) {
-				dp->i_nlink--;
-				DIP_SET(dp, i_nlink, dp->i_nlink);
-				dp->i_flag |= IN_CHANGE;
+				tdp->i_nlink--;
+				DIP_SET(tdp, i_nlink, tdp->i_nlink);
+				tdp->i_flag |= IN_CHANGE;
 			}
-			xp->i_nlink--;
-			DIP_SET(xp, i_nlink, xp->i_nlink);
-			xp->i_flag |= IN_CHANGE;
+			tip->i_nlink--;
+			DIP_SET(tip, i_nlink, tip->i_nlink);
+			tip->i_flag |= IN_CHANGE;
 			ioflag = IO_NORMAL;
 			if (!DOINGASYNC(tvp))
 				ioflag |= IO_SYNC;
+			/* Don't go to bad here as the new link exists. */
 			if ((error = UFS_TRUNCATE(tvp, (off_t)0, ioflag,
 			    tcnp->cn_cred, tcnp->cn_thread)) != 0)
-				goto bad;
+				goto unlockout;
 		}
-		vput(tdvp);
-		vput(tvp);
-		xp = NULL;
 	}
 
 	/*
-	 * 3) Unlink the source.
+	 * 3) Unlink the source.  We have to resolve the path again to
+	 * fixup the directory offset and count for ufs_dirremove.
 	 */
-	fcnp->cn_flags &= ~MODMASK;
-	fcnp->cn_flags |= LOCKPARENT | LOCKLEAF;
-	if ((fcnp->cn_flags & SAVESTART) == 0)
-		panic("ufs_rename: lost from startdir");
-	VREF(fdvp);
-	error = relookup(fdvp, &fvp, fcnp);
-	if (error == 0)
-		vrele(fdvp);
-	if (fvp != NULL) {
-		xp = VTOI(fvp);
-		dp = VTOI(fdvp);
-	} else {
-		/*
-		 * From name has disappeared.  IN_RENAME is not sufficient
-		 * to protect against directory races due to timing windows,
-		 * so we have to remove the panic.  XXX the only real way
-		 * to solve this issue is at a much higher level.  By the
-		 * time we hit ufs_rename() it's too late.
-		 */
-#if 0
-		if (doingdirectory)
-			panic("ufs_rename: lost dir entry");
-#endif
-		vrele(ap->a_fvp);
-		return (0);
+	if (fdvp == tdvp) {
+		error = ufs_lookup_ino(fdvp, NULL, fcnp, &ino);
+		if (error)
+			panic("ufs_rename: from entry went away!");
+		if (ino != fip->i_number)
+			panic("ufs_rename: ino mismatch %d != %d\n", ino,
+			    fip->i_number);
 	}
 	/*
-	 * Ensure that the directory entry still exists and has not
-	 * changed while the new name has been entered. If the source is
-	 * a file then the entry may have been unlinked or renamed. In
-	 * either case there is no further work to be done. If the source
-	 * is a directory then it cannot have been rmdir'ed; the IN_RENAME
-	 * flag ensures that it cannot be moved by another rename or removed
-	 * by a rmdir.
+	 * If the source is a directory with a
+	 * new parent, the link count of the old
+	 * parent directory must be decremented
+	 * and ".." set to point to the new parent.
 	 */
-	if (xp != ip) {
+	if (doingdirectory && newparent) {
 		/*
-		 * From name resolves to a different inode.  IN_RENAME is
-		 * not sufficient protection against timing window races
-		 * so we can't panic here.  XXX the only real way
-		 * to solve this issue is at a much higher level.  By the
-		 * time we hit ufs_rename() it's too late.
+		 * If tip exists we simply use its link, otherwise we must
+		 * add a new one.
 		 */
-#if 0
-		if (doingdirectory)
-			panic("ufs_rename: lost dir entry");
-#endif
-	} else {
-		/*
-		 * If the source is a directory with a
-		 * new parent, the link count of the old
-		 * parent directory must be decremented
-		 * and ".." set to point to the new parent.
-		 */
-		if (doingdirectory && newparent) {
-			xp->i_offset = mastertemplate.dot_reclen;
-			ufs_dirrewrite(xp, dp, newparent, DT_DIR, 0);
-			cache_purge(fdvp);
+		if (tip == NULL) {
+			tdp->i_effnlink++;
+			tdp->i_nlink++;
+			DIP_SET(tdp, i_nlink, tdp->i_nlink);
+			tdp->i_flag |= IN_CHANGE;
+			if (DOINGSOFTDEP(tdvp))
+				softdep_setup_dotdot_link(tdp, fip);
+			error = UFS_UPDATE(tdvp, !(DOINGSOFTDEP(tdvp) |
+						   DOINGASYNC(tdvp)));
+			/* Don't go to bad here as the new link exists. */
+			if (error)
+				goto unlockout;
 		}
-		error = ufs_dirremove(fdvp, xp, fcnp->cn_flags, 0);
-		xp->i_flag &= ~IN_RENAME;
+		fip->i_offset = mastertemplate.dot_reclen;
+		ufs_dirrewrite(fip, fdp, newparent, DT_DIR, 0);
+		cache_purge(fdvp);
 	}
-	if (dp)
-		vput(fdvp);
-	if (xp)
-		vput(fvp);
-	vrele(ap->a_fvp);
+	error = ufs_dirremove(fdvp, fip, fcnp->cn_flags, 0);
+
+unlockout:
+	vput(fdvp);
+	vput(fvp);
+	if (tvp)
+		vput(tvp);
+	/*
+	 * If compaction or fsync was requested do it now that other locks
+	 * are no longer needed.
+	 */
+	if (error == 0 && endoff != 0) {
+#ifdef UFS_DIRHASH
+		if (tdp->i_dirhash != NULL)
+			ufsdirhash_dirtrunc(tdp, endoff);
+#endif
+		UFS_TRUNCATE(tdvp, endoff, IO_NORMAL | IO_SYNC, tcnp->cn_cred,
+		    td);
+	}
+	if (error == 0 && tdp->i_flag & IN_NEEDSYNC)
+		error = VOP_FSYNC(tdvp, MNT_WAIT, td);
+	vput(tdvp);
+	if (mp)
+		vfs_unbusy(mp);
 	return (error);
 
 bad:
-	if (xp)
-		vput(ITOV(xp));
-	vput(ITOV(dp));
-out:
-	if (doingdirectory)
-		ip->i_flag &= ~IN_RENAME;
-	if (vn_lock(fvp, LK_EXCLUSIVE) == 0) {
-		ip->i_effnlink--;
-		ip->i_nlink--;
-		DIP_SET(ip, i_nlink, ip->i_nlink);
-		ip->i_flag |= IN_CHANGE;
-		ip->i_flag &= ~IN_RENAME;
-		if (DOINGSOFTDEP(fvp))
-			softdep_change_linkcnt(ip);
-		vput(fvp);
-	} else
-		vrele(fvp);
+	fip->i_effnlink--;
+	fip->i_nlink--;
+	DIP_SET(fip, i_nlink, fip->i_nlink);
+	fip->i_flag |= IN_CHANGE;
+	if (DOINGSOFTDEP(fvp))
+		softdep_revert_link(tdp, fip);
+	goto unlockout;
+
+releout:
+	vrele(fdvp);
+	vrele(fvp);
+	vrele(tdvp);
+	if (tvp)
+		vrele(tvp);
+	if (mp)
+		vfs_unbusy(mp);
+
 	return (error);
 }
 
@@ -1767,8 +1856,7 @@ ufs_mkdir(ap)
 	ip->i_effnlink = 2;
 	ip->i_nlink = 2;
 	DIP_SET(ip, i_nlink, 2);
-	if (DOINGSOFTDEP(tvp))
-		softdep_change_linkcnt(ip);
+
 	if (cnp->cn_flags & ISWHITEOUT) {
 		ip->i_flags |= UF_OPAQUE;
 		DIP_SET(ip, i_flags, ip->i_flags);
@@ -1784,8 +1872,8 @@ ufs_mkdir(ap)
 	DIP_SET(dp, i_nlink, dp->i_nlink);
 	dp->i_flag |= IN_CHANGE;
 	if (DOINGSOFTDEP(dvp))
-		softdep_change_linkcnt(dp);
-	error = UFS_UPDATE(tvp, !(DOINGSOFTDEP(dvp) | DOINGASYNC(dvp)));
+		softdep_setup_mkdir(dp, ip);
+	error = UFS_UPDATE(dvp, !(DOINGSOFTDEP(dvp) | DOINGASYNC(dvp)));
 	if (error)
 		goto bad;
 #ifdef MAC
@@ -1863,7 +1951,7 @@ ufs_mkdir(ap)
 	else if (!DOINGSOFTDEP(dvp) && ((error = bwrite(bp))))
 		goto bad;
 	ufs_makedirentry(ip, cnp, &newdir);
-	error = ufs_direnter(dvp, tvp, &newdir, cnp, bp);
+	error = ufs_direnter(dvp, tvp, &newdir, cnp, bp, 0);
 	
 bad:
 	if (error == 0) {
@@ -1873,8 +1961,6 @@ bad:
 		dp->i_nlink--;
 		DIP_SET(dp, i_nlink, dp->i_nlink);
 		dp->i_flag |= IN_CHANGE;
-		if (DOINGSOFTDEP(dvp))
-			softdep_change_linkcnt(dp);
 		/*
 		 * No need to do an explicit VOP_TRUNCATE here, vrele will
 		 * do this for us because we set the link count to 0.
@@ -1884,7 +1970,8 @@ bad:
 		DIP_SET(ip, i_nlink, 0);
 		ip->i_flag |= IN_CHANGE;
 		if (DOINGSOFTDEP(tvp))
-			softdep_change_linkcnt(ip);
+			softdep_revert_mkdir(dp, ip);
+
 		vput(tvp);
 	}
 out:
@@ -1920,10 +2007,13 @@ ufs_rmdir(ap)
 	 * tries to remove a locally mounted on directory).
 	 */
 	error = 0;
-	if ((ip->i_flag & IN_RENAME) || ip->i_effnlink < 2) {
+	if (ip->i_effnlink < 2) {
 		error = EINVAL;
 		goto out;
 	}
+	if (dp->i_effnlink < 3)
+		panic("ufs_dirrem: Bad link count %d on parent",
+		    dp->i_effnlink);
 	if (!ufs_dirempty(ip, dp->i_number, cnp->cn_cred)) {
 		error = ENOTEMPTY;
 		goto out;
@@ -1947,18 +2037,14 @@ ufs_rmdir(ap)
 	 */
 	dp->i_effnlink--;
 	ip->i_effnlink--;
-	if (DOINGSOFTDEP(vp)) {
-		softdep_change_linkcnt(dp);
-		softdep_change_linkcnt(ip);
-	}
+	if (DOINGSOFTDEP(vp))
+		softdep_setup_rmdir(dp, ip);
 	error = ufs_dirremove(dvp, ip, cnp->cn_flags, 1);
 	if (error) {
 		dp->i_effnlink++;
 		ip->i_effnlink++;
-		if (DOINGSOFTDEP(vp)) {
-			softdep_change_linkcnt(dp);
-			softdep_change_linkcnt(ip);
-		}
+		if (DOINGSOFTDEP(vp))
+			softdep_revert_rmdir(dp, ip);
 		goto out;
 	}
 	cache_purge(dvp);
@@ -2464,6 +2550,9 @@ ufs_makeinode(mode, dvp, vpp, cnp)
 	if ((mode & IFMT) == 0)
 		mode |= IFREG;
 
+	if (VTOI(dvp)->i_effnlink < 2)
+		panic("ufs_makeinode: Bad link count %d on parent",
+		    VTOI(dvp)->i_effnlink);
 	error = UFS_VALLOC(dvp, mode, cnp->cn_cred, &tvp);
 	if (error)
 		return (error);
@@ -2539,7 +2628,7 @@ ufs_makeinode(mode, dvp, vpp, cnp)
 	ip->i_nlink = 1;
 	DIP_SET(ip, i_nlink, 1);
 	if (DOINGSOFTDEP(tvp))
-		softdep_change_linkcnt(ip);
+		softdep_setup_create(VTOI(dvp), ip);
 	if ((ip->i_mode & ISGID) && !groupmember(ip->i_gid, cnp->cn_cred) &&
 	    priv_check_cred(cnp->cn_cred, PRIV_VFS_SETGID, 0)) {
 		ip->i_mode &= ~ISGID;
@@ -2579,7 +2668,7 @@ ufs_makeinode(mode, dvp, vpp, cnp)
 	}
 #endif /* !UFS_ACL */
 	ufs_makedirentry(ip, cnp, &newdir);
-	error = ufs_direnter(dvp, tvp, &newdir, cnp, NULL);
+	error = ufs_direnter(dvp, tvp, &newdir, cnp, NULL, 0);
 	if (error)
 		goto bad;
 	*vpp = tvp;
@@ -2595,7 +2684,7 @@ bad:
 	DIP_SET(ip, i_nlink, 0);
 	ip->i_flag |= IN_CHANGE;
 	if (DOINGSOFTDEP(tvp))
-		softdep_change_linkcnt(ip);
+		softdep_revert_create(VTOI(dvp), ip);
 	vput(tvp);
 	return (error);
 }
diff --git a/sys/ufs/ufs/ufsmount.h b/sys/ufs/ufs/ufsmount.h
index 126867baa2e..3f356b1d521 100644
--- a/sys/ufs/ufs/ufsmount.h
+++ b/sys/ufs/ufs/ufsmount.h
@@ -57,6 +57,10 @@ struct ucred;
 struct uio;
 struct vnode;
 struct ufs_extattr_per_mount;
+struct jblocks;
+struct inodedep;
+
+TAILQ_HEAD(inodedeplst, inodedep);
 
 /* This structure describes the UFS specific mount structure data. */
 struct ufsmount {
@@ -75,6 +79,11 @@ struct ufsmount {
 	long	um_numindirdeps;		/* outstanding indirdeps */
 	struct	workhead softdep_workitem_pending; /* softdep work queue */
 	struct	worklist *softdep_worklist_tail; /* Tail pointer for above */
+	struct	workhead softdep_journal_pending; /* journal work queue */
+	struct	worklist *softdep_journal_tail;	/* Tail pointer for above */
+	struct	jblocks *softdep_jblocks;	/* Journal block information */
+	struct	inodedeplst softdep_unlinked; /* Unlinked inodes */
+	int	softdep_on_journal;		/* Items on the journal list */
 	int	softdep_on_worklist;		/* Items on the worklist */
 	int	softdep_on_worklist_inprogress;	/* Busy items on worklist */
 	int	softdep_deps;			/* Total dependency count */
diff --git a/usr.sbin/makefs/ffs/ffs_bswap.c b/usr.sbin/makefs/ffs/ffs_bswap.c
index d560884f637..18ace03123f 100644
--- a/usr.sbin/makefs/ffs/ffs_bswap.c
+++ b/usr.sbin/makefs/ffs/ffs_bswap.c
@@ -136,8 +136,6 @@ ffs_dinode1_swap(struct ufs1_dinode *o, struct ufs1_dinode *n)
 
 	n->di_mode = bswap16(o->di_mode);
 	n->di_nlink = bswap16(o->di_nlink);
-	n->di_u.oldids[0] = bswap16(o->di_u.oldids[0]);
-	n->di_u.oldids[1] = bswap16(o->di_u.oldids[1]);
 	n->di_size = bswap64(o->di_size);
 	n->di_atime = bswap32(o->di_atime);
 	n->di_atimensec = bswap32(o->di_atimensec);

From a8750f2dca95ef256490ef95fa1d1ead2a88a162 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sat, 24 Apr 2010 07:36:33 +0000
Subject: [PATCH 226/532] Fix build for UFS without SOFTUPDATES.

---
 sys/ufs/ffs/ffs_softdep.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index 30937051357..4a659f9de7b 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -232,7 +232,8 @@ softdep_setup_directory_add(bp, dp, diroffset, newinum, newdirbp, isnewblk)
 }
 
 void 
-softdep_change_directoryentry_offset(dp, base, oldloc, newloc, entrysize)
+softdep_change_directoryentry_offset(bp, dp, base, oldloc, newloc, entrysize)
+	struct buf *bp;
 	struct inode *dp;
 	caddr_t base;
 	caddr_t oldloc;

From d00690aec50fc0edda9272b6566685f036f2f86d Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sat, 24 Apr 2010 07:54:49 +0000
Subject: [PATCH 227/532] Protect fsck.h from being included twice.

---
 sbin/fsck_ffs/fsck.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/sbin/fsck_ffs/fsck.h b/sbin/fsck_ffs/fsck.h
index 6dad6b7318b..08f9ef5790b 100644
--- a/sbin/fsck_ffs/fsck.h
+++ b/sbin/fsck_ffs/fsck.h
@@ -60,6 +60,9 @@
  * $FreeBSD$
  */
 
+#ifndef _FSCK_H_
+#define	_FSCK_H_
+
 #include <unistd.h>
 #include <stdlib.h>
 #include <stdio.h>
@@ -389,3 +392,5 @@ void		setinodebuf(ino_t);
 int		setup(char *dev);
 void		gjournal_check(const char *filesys);
 int		suj_check(const char *filesys);
+
+#endif	/* !_FSCK_H_ */

From 132f1ed5859f8f28fe76567003a7be213d719c4b Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Sat, 24 Apr 2010 07:58:59 +0000
Subject: [PATCH 228/532] suj.c seems to contain two versions of the code.
 Remove the one that doesn't compile.

---
 sbin/fsck_ffs/suj.c | 2065 -------------------------------------------
 1 file changed, 2065 deletions(-)

diff --git a/sbin/fsck_ffs/suj.c b/sbin/fsck_ffs/suj.c
index 84608f164e4..f0240bdcaeb 100644
--- a/sbin/fsck_ffs/suj.c
+++ b/sbin/fsck_ffs/suj.c
@@ -2632,2068 +2632,3 @@ suj_check(const char *filesys)
 
 	return (0);
 }
-/*-
- * Copyright (c) 2009 Jeffrey W. Roberson <jeff@FreeBSD.org>
- * All rights reserved.
- *
- * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions
- * are met:
- * 1. Redistributions of source code must retain the above copyright
- *    notice, this list of conditions and the following disclaimer.
- * 2. Redistributions in binary form must reproduce the above copyright
- *    notice, this list of conditions and the following disclaimer in the
- *    documentation and/or other materials provided with the distribution.
- *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
- * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
- * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
- * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
- * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
- * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
- * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
- * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
- * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
- * SUCH DAMAGE.
- */
-
-#include <sys/cdefs.h>
-__FBSDID("$FreeBSD$");
-
-#include <sys/param.h>
-#include <sys/disklabel.h>
-#include <sys/mount.h>
-#include <sys/stat.h>
-
-#include <ufs/ufs/ufsmount.h>
-#include <ufs/ufs/dinode.h>
-#include <ufs/ufs/dir.h>
-#include <ufs/ffs/fs.h>
-
-#include <stdio.h>
-#include <stdlib.h>
-#include <stdint.h>
-#include <libufs.h>
-#include <strings.h>
-#include <err.h>
-#include <assert.h>
-
-#include "fsck.h"
-
-static void	ino_decr(ino_t);
-
-#define	SUJ_HASHSIZE	128
-#define	SUJ_HASHMASK	(SUJ_HASHSIZE - 1)
-#define	SUJ_HASH(x)	((x * 2654435761) & SUJ_HASHMASK)
-
-struct suj_seg {
-	TAILQ_ENTRY(suj_seg) ss_next;
-	struct jsegrec	ss_rec;
-	uint8_t		*ss_blk;
-};
-
-struct suj_rec {
-	TAILQ_ENTRY(suj_rec) sr_next;
-	union jrec	*sr_rec;
-};
-TAILQ_HEAD(srechd, suj_rec);
-
-struct suj_ino {
-	LIST_ENTRY(suj_ino)	si_next;
-	struct srechd		si_recs;
-	struct srechd		si_movs;
-	ino_t			si_ino;
-	int			si_nlinkadj;
-	int			si_skipparent;
-	int			si_linkadj;
-	int			si_hasrecs;
-	int			si_blkadj;
-};
-LIST_HEAD(inohd, suj_ino);
-
-struct suj_blk {
-	LIST_ENTRY(suj_blk)	sb_next;
-	struct srechd		sb_recs;
-	ufs2_daddr_t		sb_blk;
-};
-LIST_HEAD(blkhd, suj_blk);
-
-struct data_blk {
-	LIST_ENTRY(data_blk)	db_next;
-	uint8_t			*db_buf;
-	ufs2_daddr_t		db_blk;
-	int			db_size;
-};
-
-struct ino_blk {
-	LIST_ENTRY(ino_blk)	ib_next;
-	uint8_t			*ib_buf;
-	int			ib_dirty;
-	ufs2_daddr_t		ib_blk;
-};
-LIST_HEAD(iblkhd, ino_blk);
-
-struct suj_cg {
-	LIST_ENTRY(suj_cg)	sc_next;
-	struct blkhd		sc_blkhash[SUJ_HASHSIZE];
-	struct inohd		sc_inohash[SUJ_HASHSIZE];
-	struct iblkhd		sc_iblkhash[SUJ_HASHSIZE];
-	struct ino_blk		*sc_lastiblk;
-	uint8_t			*sc_cgbuf;
-	struct cg		*sc_cgp;
-	int			sc_dirty;
-	int			sc_cgx;
-};
-
-LIST_HEAD(cghd, suj_cg) cghash[SUJ_HASHSIZE];
-LIST_HEAD(dblkhd, data_blk) dbhash[SUJ_HASHSIZE];
-
-TAILQ_HEAD(seghd, suj_seg) allsegs;
-uint64_t oldseq;
-static struct uufsd *disk = NULL;
-static struct fs *fs = NULL;
-
-/*
- * Summary statistics.
- */
-uint64_t freefrags;
-uint64_t freeblocks;
-uint64_t freeinos;
-uint64_t freedir;
-uint64_t jbytes;
-uint64_t jrecs;
-
-typedef void (*ino_visitor)(ino_t, ufs_lbn_t, ufs2_daddr_t, int);
-
-static void *
-errmalloc(size_t n)
-{
-	void *a;
-
-	a = malloc(n);
-	if (a == NULL)
-		errx(1, "malloc(%zu)", n);
-	return (a);
-}
-
-/*
- * Open the given provider, load superblock.
- */
-static void
-opendisk(const char *devnam)
-{
-	if (disk != NULL)
-		return;
-	disk = malloc(sizeof(*disk));
-	if (disk == NULL)
-		errx(1, "malloc(%zu)", sizeof(*disk));
-	if (ufs_disk_fillout(disk, devnam) == -1) {
-		err(1, "ufs_disk_fillout(%s) failed: %s", devnam,
-		    disk->d_error);
-	}
-	fs = &disk->d_fs;
-	/*
-	 * Setup a few things so reply() can work.
-	 */
-	bcopy(fs, &sblock, sizeof(sblock));
-	fsreadfd = disk->d_fd;
-	fswritefd = disk->d_fd;
-}
-
-/*
- * Mark file system as clean, write the super-block back, close the disk.
- */
-static void
-closedisk(const char *devnam)
-{
-	struct csum *cgsum;
-	int i;
-
-	/*
-	 * Recompute the fs summary info from correct cs summaries.
-	 */
-	bzero(&fs->fs_cstotal, sizeof(struct csum_total));
-	for (i = 0; i < fs->fs_ncg; i++) {
-		cgsum = &fs->fs_cs(fs, i);
-		fs->fs_cstotal.cs_nffree += cgsum->cs_nffree;
-		fs->fs_cstotal.cs_nbfree += cgsum->cs_nbfree;
-		fs->fs_cstotal.cs_nifree += cgsum->cs_nifree;
-		fs->fs_cstotal.cs_ndir += cgsum->cs_ndir;
-	}
-	/* XXX Don't set clean for now, we don't trust the journal. */
-	/* fs->fs_clean = 1; */
-	fs->fs_time = time(NULL);
-	fs->fs_mtime = time(NULL);
-	if (sbwrite(disk, 0) == -1)
-		err(1, "sbwrite(%s)", devnam);
-	if (ufs_disk_close(disk) == -1)
-		err(1, "ufs_disk_close(%s)", devnam);
-	free(disk);
-	disk = NULL;
-	fs = NULL;
-	fsreadfd = -1;
-	fswritefd = -1;
-}
-
-/*
- * Lookup a cg by number in the hash so we can keep track of which cgs
- * need stats rebuilt.
- */
-static struct suj_cg *
-cg_lookup(int cgx)
-{
-	struct cghd *hd;
-	struct suj_cg *sc;
-
-	if (cgx < 0 || cgx >= fs->fs_ncg) {
-		abort();
-		errx(1, "Bad cg number %d", cgx);
-	}
-	hd = &cghash[SUJ_HASH(cgx)];
-	LIST_FOREACH(sc, hd, sc_next)
-		if (sc->sc_cgx == cgx)
-			return (sc);
-	sc = errmalloc(sizeof(*sc));
-	bzero(sc, sizeof(*sc));
-	sc->sc_cgbuf = errmalloc(fs->fs_bsize);
-	sc->sc_cgp = (struct cg *)sc->sc_cgbuf;
-	sc->sc_cgx = cgx;
-	LIST_INSERT_HEAD(hd, sc, sc_next);
-	if (bread(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
-	    fs->fs_bsize) == -1)
-		err(1, "Unable to read cylinder group %d", sc->sc_cgx);
-
-	return (sc);
-}
-
-/*
- * Lookup an inode number in the hash and allocate a suj_ino if it does
- * not exist.
- */
-static struct suj_ino *
-ino_lookup(ino_t ino, int creat)
-{
-	struct suj_ino *sino;
-	struct inohd *hd;
-	struct suj_cg *sc;
-
-	sc = cg_lookup(ino_to_cg(fs, ino));
-	hd = &sc->sc_inohash[SUJ_HASH(ino)];
-	LIST_FOREACH(sino, hd, si_next)
-		if (sino->si_ino == ino)
-			return (sino);
-	if (creat == 0)
-		return (NULL);
-	sino = errmalloc(sizeof(*sino));
-	bzero(sino, sizeof(*sino));
-	sino->si_ino = ino;
-	sino->si_nlinkadj = 0;
-	TAILQ_INIT(&sino->si_recs);
-	TAILQ_INIT(&sino->si_movs);
-	LIST_INSERT_HEAD(hd, sino, si_next);
-
-	return (sino);
-}
-
-/*
- * Lookup a block number in the hash and allocate a suj_blk if it does
- * not exist.
- */
-static struct suj_blk *
-blk_lookup(ufs2_daddr_t blk, int creat)
-{
-	struct suj_blk *sblk;
-	struct suj_cg *sc;
-	struct blkhd *hd;
-
-	sc = cg_lookup(dtog(fs, blk));
-	hd = &sc->sc_blkhash[SUJ_HASH(blk)];
-	LIST_FOREACH(sblk, hd, sb_next)
-		if (sblk->sb_blk == blk)
-			return (sblk);
-	if (creat == 0)
-		return (NULL);
-	sblk = errmalloc(sizeof(*sblk));
-	bzero(sblk, sizeof(*sblk));
-	sblk->sb_blk = blk;
-	TAILQ_INIT(&sblk->sb_recs);
-	LIST_INSERT_HEAD(hd, sblk, sb_next);
-
-	return (sblk);
-}
-
-static uint8_t *
-dblk_read(ufs2_daddr_t blk, int size)
-{
-	struct data_blk *dblk;
-	struct dblkhd *hd;
-
-	hd = &dbhash[SUJ_HASH(blk)];
-	LIST_FOREACH(dblk, hd, db_next)
-		if (dblk->db_blk == blk)
-			goto found;
-	/*
-	 * The inode block wasn't located, allocate a new one.
-	 */
-	dblk = errmalloc(sizeof(*dblk));
-	bzero(dblk, sizeof(*dblk));
-	LIST_INSERT_HEAD(hd, dblk, db_next);
-	dblk->db_blk = blk;
-found:
-	/*
-	 * I doubt size mismatches can happen in practice but it is trivial
-	 * to handle.
-	 */
-	if (size != dblk->db_size) {
-		if (dblk->db_buf)
-			free(dblk->db_buf);
-		dblk->db_buf = errmalloc(size);
-		dblk->db_size = size;
-		if (bread(disk, fsbtodb(fs, blk), dblk->db_buf, size) == -1)
-			err(1, "Failed to read data block %jd", blk);
-	}
-	return (dblk->db_buf);
-}
-
-static union dinode *
-ino_read(ino_t ino)
-{
-	struct ino_blk *iblk;
-	struct iblkhd *hd;
-	struct suj_cg *sc;
-	ufs2_daddr_t blk;
-	int off;
-
-	blk = ino_to_fsba(fs, ino);
-	sc = cg_lookup(ino_to_cg(fs, ino));
-	hd = &sc->sc_iblkhash[SUJ_HASH(blk)];
-	LIST_FOREACH(iblk, hd, ib_next)
-		if (iblk->ib_blk == blk)
-			goto found;
-	/*
-	 * The inode block wasn't located, allocate a new one.
-	 */
-	iblk = errmalloc(sizeof(*iblk));
-	bzero(iblk, sizeof(*iblk));
-	iblk->ib_buf = errmalloc(fs->fs_bsize);
-	iblk->ib_blk = blk;
-	LIST_INSERT_HEAD(hd, iblk, ib_next);
-	if (bread(disk, fsbtodb(fs, blk), iblk->ib_buf, fs->fs_bsize) == -1)
-		err(1, "Failed to read inode block %jd", blk);
-found:
-	sc->sc_lastiblk = iblk;
-	off = ino_to_fsbo(fs, ino);
-	if (fs->fs_magic == FS_UFS1_MAGIC)
-		return (union dinode *)&((struct ufs1_dinode *)iblk->ib_buf)[off];
-	else
-		return (union dinode *)&((struct ufs2_dinode *)iblk->ib_buf)[off];
-}
-
-static void
-ino_dirty(ino_t ino)
-{
-	struct ino_blk *iblk;
-	struct iblkhd *hd;
-	struct suj_cg *sc;
-	ufs2_daddr_t blk;
-
-	blk = ino_to_fsba(fs, ino);
-	sc = cg_lookup(ino_to_cg(fs, ino));
-	iblk = sc->sc_lastiblk;
-	if (iblk && iblk->ib_blk == blk) {
-		iblk->ib_dirty = 1;
-		return;
-	}
-	hd = &sc->sc_iblkhash[SUJ_HASH(blk)];
-	LIST_FOREACH(iblk, hd, ib_next) {
-		if (iblk->ib_blk == blk) {
-			iblk->ib_dirty = 1;
-			return;
-		}
-	}
-	ino_read(ino);
-	ino_dirty(ino);
-}
-
-static void
-iblk_write(struct ino_blk *iblk)
-{
-
-	if (iblk->ib_dirty == 0)
-		return;
-	if (bwrite(disk, fsbtodb(fs, iblk->ib_blk), iblk->ib_buf,
-	    fs->fs_bsize) == -1)
-		err(1, "Failed to write inode block %jd", iblk->ib_blk);
-}
-
-/*
- * Return 1 if the inode was free and 0 if it is allocated.
- */
-static int
-ino_isfree(ino_t ino)
-{
-	struct suj_cg *sc;
-	uint8_t *inosused;
-	struct cg *cgp;
-	int cg;
-
-	cg = ino_to_cg(fs, ino);
-	ino = ino % fs->fs_ipg;
-	sc = cg_lookup(cg);
-	cgp = sc->sc_cgp;
-	inosused = cg_inosused(cgp);
-	return isclr(inosused, ino);
-}
-
-static int
-blk_overlaps(struct jblkrec *brec, ufs2_daddr_t start, int frags)
-{
-	ufs2_daddr_t bstart;
-	ufs2_daddr_t bend;
-	ufs2_daddr_t end;
-
-	end = start + frags;
-	bstart = brec->jb_blkno + brec->jb_oldfrags;
-	bend = bstart + brec->jb_frags;
-	if (start < bend && end > bstart)
-		return (1);
-	return (0);
-}
-
-static int
-blk_equals(struct jblkrec *brec, ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t start,
-    int frags)
-{
-
-	if (brec->jb_ino != ino || brec->jb_lbn != lbn)
-		return (0);
-	if (brec->jb_blkno + brec->jb_oldfrags != start)
-		return (0);
-	if (brec->jb_frags != frags)
-		return (0);
-	return (1);
-}
-
-static void
-blk_setmask(struct jblkrec *brec, int *mask)
-{
-	int i;
-
-	for (i = brec->jb_oldfrags; i < brec->jb_oldfrags + brec->jb_frags; i++)
-		*mask |= 1 << i;
-}
-
-/*
- * Determine whether a given block has been reallocated to a new location.
- * Returns a mask of overlapping bits if any frags have been reused or
- * zero if the block has not been re-used and the contents can be trusted.
- * 
- * This is used to ensure that an orphaned pointer due to truncate is safe
- * to be freed.  The mask value can be used to free partial blocks.
- */
-static int
-blk_isfree(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags)
-{
-	struct suj_blk *sblk;
-	struct suj_rec *srec;
-	struct jblkrec *brec;
-	int mask;
-	int off;
-
-	/*
-	 * To be certain we're not freeing a reallocated block we lookup
-	 * this block in the blk hash and see if there is an allocation
-	 * journal record that overlaps with any fragments in the block
-	 * we're concerned with.  If any fragments have ben reallocated
-	 * the block has already been freed and re-used for another purpose.
-	 */
-	mask = 0;
-	sblk = blk_lookup(blknum(fs, blk), 0);
-	if (sblk == NULL)
-		return (0);
-	off = blk - sblk->sb_blk;
-	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
-		brec = (struct jblkrec *)srec->sr_rec;
-		/*
-		 * If the block overlaps but does not match
-		 * exactly it's a new allocation.  If it matches
-		 * exactly this record refers to the current
-		 * location.
-		 */ 
-		if (blk_overlaps(brec, blk, frags) == 0)
-			continue;
-		if (blk_equals(brec, ino, lbn, blk, frags) == 1)
-			mask = 0;
-		else
-			blk_setmask(brec, &mask);
-	}
-	if (debug)
-		printf("blk_isfree: blk %jd sblk %jd off %d mask 0x%X\n",
-		    blk, sblk->sb_blk, off, mask);
-	return (mask >> off);
-}
-
-/*
- * Determine whether it is safe to follow an indirect.  It is not safe
- * if any part of the indirect has been reallocated or the last journal
- * entry was an allocation.  Just allocated indirects may not have valid
- * pointers yet and all of their children will have their own records.
- * 
- * Returns 1 if it's safe to follow the indirect and 0 otherwise.
- */
-static int
-blk_isindir(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn)
-{
-	struct suj_blk *sblk;
-	struct jblkrec *brec;
-
-	sblk = blk_lookup(blk, 0);
-	if (sblk == NULL)
-		return (1);
-	if (TAILQ_EMPTY(&sblk->sb_recs))
-		return (1);
-	brec = (struct jblkrec *)TAILQ_LAST(&sblk->sb_recs, srechd)->sr_rec;
-	if (blk_equals(brec, ino, lbn, blk, fs->fs_frag))
-		if (brec->jb_op == JOP_FREEBLK)
-			return (1);
-	return (0);
-}
-
-/*
- * Clear an inode from the cg bitmap.  If the inode was already clear return
- * 0 so the caller knows it does not have to check the inode contents.
- */
-static int
-ino_free(ino_t ino, int mode)
-{
-	struct suj_cg *sc;
-	uint8_t *inosused;
-	struct cg *cgp;
-	int cg;
-
-	cg = ino_to_cg(fs, ino);
-	ino = ino % fs->fs_ipg;
-	sc = cg_lookup(cg);
-	cgp = sc->sc_cgp;
-	inosused = cg_inosused(cgp);
-	/*
-	 * The bitmap may never have made it to the disk so we have to
-	 * conditionally clear.  We can avoid writing the cg in this case.
-	 */
-	if (isclr(inosused, ino))
-		return (0);
-	freeinos++;
-	clrbit(inosused, ino);
-	if (ino < cgp->cg_irotor)
-		cgp->cg_irotor = ino;
-	cgp->cg_cs.cs_nifree++;
-	if ((mode & IFMT) == IFDIR) {
-		freedir++;
-		cgp->cg_cs.cs_ndir--;
-	}
-	sc->sc_dirty = 1;
-
-	return (1);
-}
-
-/*
- * Free 'frags' frags starting at filesystem block 'bno' skipping any frags
- * set in the mask.
- */
-static void
-blk_free(ufs2_daddr_t bno, int mask, int frags)
-{
-	ufs1_daddr_t fragno, cgbno;
-	struct suj_cg *sc;
-	struct cg *cgp;
-	int i, cg;
-	uint8_t *blksfree;
-
-	if (debug)
-		printf("Freeing %d frags at blk %jd\n", frags, bno);
-	cg = dtog(fs, bno);
-	sc = cg_lookup(cg);
-	cgp = sc->sc_cgp;
-	cgbno = dtogd(fs, bno);
-	blksfree = cg_blksfree(cgp);
-
-	/*
-	 * If it's not allocated we only wrote the journal entry
-	 * and never the bitmaps.  Here we unconditionally clear and
-	 * resolve the cg summary later.
-	 */
-	if (frags == fs->fs_frag && mask == 0) {
-		fragno = fragstoblks(fs, cgbno);
-		ffs_setblock(fs, blksfree, fragno);
-		freeblocks++;
-	} else {
-		/*
-		 * deallocate the fragment
-		 */
-		for (i = 0; i < frags; i++)
-			if ((mask & (1 << i)) == 0 && isclr(blksfree, cgbno +i)) {
-				freefrags++;
-				setbit(blksfree, cgbno + i);
-			}
-	}
-	sc->sc_dirty = 1;
-}
-
-/*
- * Fetch an indirect block to find the block at a given lbn.  The lbn
- * may be negative to fetch a specific indirect block pointer or positive
- * to fetch a specific block.
- */
-static ufs2_daddr_t
-indir_blkatoff(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t cur, ufs_lbn_t lbn, int level)
-{
-	ufs2_daddr_t *bap2;
-	ufs2_daddr_t *bap1;
-	ufs_lbn_t lbnadd;
-	ufs_lbn_t base;
-	int i;
-
-	if (blk == 0)
-		return (0);
-	if (cur == lbn)
-		return (blk);
-	if (level == 0 && lbn < 0) {
-		abort();
-		errx(1, "Invalid lbn %jd", lbn);
-	}
-	bap2 = (void *)dblk_read(blk, fs->fs_bsize);
-	bap1 = (void *)bap2;
-	lbnadd = 1;
-	base = -(cur + level);
-	for (i = level; i > 0; i--)
-		lbnadd *= NINDIR(fs);
-	if (lbn > 0) 
-		i = (lbn - base) / lbnadd;
-	else
-		i = (-lbn - base) / lbnadd;
-	if (i < 0 || i >= NINDIR(fs)) {
-		abort();
-		errx(1, "Invalid indirect index %d produced by lbn %jd",
-		    i, lbn);
-	}
-	if (level == 0)
-		cur = base + (i * lbnadd);
-	else
-		cur = -(base + (i * lbnadd)) - (level - 1);
-	if (fs->fs_magic == FS_UFS1_MAGIC)
-		blk = bap1[i];
-	else
-		blk = bap2[i];
-	if (cur == lbn)
-		return (blk);
-	if (level == 0) {
-		abort();
-		errx(1, "Invalid lbn %jd at level 0", lbn);
-	}
-	return indir_blkatoff(blk, ino, cur, lbn, level - 1);
-}
-
-/*
- * Finds the disk block address at the specified lbn within the inode
- * specified by ip.  This follows the whole tree and honors di_size and
- * di_extsize so it is a true test of reachability.  The lbn may be
- * negative if an extattr or indirect block is requested.
- */
-static ufs2_daddr_t
-ino_blkatoff(union dinode *ip, ino_t ino, ufs_lbn_t lbn, int *frags)
-{
-	ufs_lbn_t tmpval;
-	ufs_lbn_t cur;
-	ufs_lbn_t next;
-	int i;
-
-	/*
-	 * Handle extattr blocks first.
-	 */
-	if (lbn < 0 && lbn >= -NXADDR) {
-		lbn = -1 - lbn;
-		if (lbn > lblkno(fs, ip->dp2.di_extsize - 1))
-			return (0);
-		*frags = numfrags(fs, sblksize(fs, ip->dp2.di_extsize, lbn));
-		return (ip->dp2.di_extb[lbn]);
-	}
-	/*
-	 * And now direct and indirect.  Verify that the lbn does not
-	 * exceed the size required to store the file by asking for
-	 * the lbn of the last byte.  These blocks should be 0 anyway
-	 * so this simply saves the traversal.
-	 */
-	if (lbn > 0 && lbn > lblkno(fs, DIP(ip, di_size) - 1))
-		return (0);
-	if (lbn < 0 && -lbn > lblkno(fs, DIP(ip, di_size) - 1))
-		return (0);
-	if (lbn >= 0 && lbn < NDADDR) {
-		*frags = numfrags(fs, sblksize(fs, DIP(ip, di_size), lbn));
-		return (DIP(ip, di_db[lbn]));
-	}
-	*frags = fs->fs_frag;
-
-	for (i = 0, tmpval = NINDIR(fs), cur = NDADDR; i < NIADDR; i++,
-	    tmpval *= NINDIR(fs), cur = next) {
-		next = cur + tmpval;
-		if (lbn == -cur)
-			return (DIP(ip, di_ib[i]));
-		/*
-		 * Determine whether the lbn in question is within this tree.
-		 */
-		if (lbn < 0 && -lbn >= next)
-			continue;
-		if (lbn > 0 && lbn >= next)
-			continue;
-
-		return indir_blkatoff(DIP(ip, di_ib[i]), ino, -cur - i, lbn, i);
-	}
-	errx(1, "lbn %jd not in ino", lbn);
-}
-
-/*
- * Determine whether a block exists at a particular lbn in an inode.
- * Returns 1 if found, 0 if not.  lbn may be negative for indirects
- * or ext blocks.
- */
-static int
-blk_isat(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int *frags)
-{
-	union dinode *ip;
-	ufs2_daddr_t nblk;
-
-	ip = ino_read(ino);
-
-	if (DIP(ip, di_nlink) == 0 || DIP(ip, di_mode) == 0)
-		return (0);
-	nblk = ino_blkatoff(ip, ino, lbn, frags);
-
-	return (nblk == blk);
-}
-
-/*
- * Determines whether a pointer to an inode exists within a directory
- * at a specified offset.  Returns the mode of the found entry.
- */
-static int
-ino_isat(ino_t parent, off_t diroff, ino_t child, int *mode, int *isdot)
-{
-	union dinode *dip;
-	struct direct *dp;
-	ufs2_daddr_t blk;
-	uint8_t *block;
-	ufs_lbn_t lbn;
-	int blksize;
-	int frags;
-	int dpoff;
-	int doff;
-
-	*isdot = 0;
-	dip = ino_read(parent);
-	*mode = DIP(dip, di_mode);
-	if ((*mode & IFMT) != IFDIR) {
-		if (debug) {
-			/* This can happen if the parent inode was reallocated. */
-			if (*mode != 0)
-				printf("Directory %d has bad mode %o\n",
-				    parent, *mode);
-			else
-				printf("Directory %d zero inode\n", parent);
-		}
-		return (0);
-	}
-	lbn = lblkno(fs, diroff);
-	doff = blkoff(fs, diroff);
-	blksize = sblksize(fs, DIP(dip, di_size), lbn);
-	if (diroff + DIRECTSIZ(1) > DIP(dip, di_size) || doff >= blksize) {
-		if (debug)
-			printf("ino %d absent from %d due to offset %jd"
-			    " exceeding size %jd\n",
-			    child, parent, diroff, DIP(dip, di_size));
-		return (0);
-	}
-	blk = ino_blkatoff(dip, parent, lbn, &frags);
-	if (blk <= 0) {
-		if (debug)
-			printf("Sparse directory %d", parent);
-		return (0);
-	}
-	block = dblk_read(blk, blksize);
-	/*
-	 * Walk through the records from the start of the block to be
-	 * certain we hit a valid record and not some junk in the middle
-	 * of a file name.  Stop when we reach or pass the expected offset.
-	 */
-	dpoff = 0;
-	do {
-		dp = (struct direct *)&block[dpoff];
-		if (dpoff == doff)
-			break;
-		if (dp->d_reclen == 0)
-			break;
-		dpoff += dp->d_reclen;
-	} while (dpoff <= doff);
-	if (dpoff > fs->fs_bsize)
-		errx(1, "Corrupt directory block in dir inode %d", parent);
-	/* Not found. */
-	if (dpoff != doff) {
-		if (debug)
-			printf("ino %d not found in %d, lbn %jd, dpoff %d\n",
-			    child, parent, lbn, dpoff);
-		return (0);
-	}
-	/*
-	 * We found the item in question.  Record the mode and whether it's
-	 * a . or .. link for the caller.
-	 */
-	if (dp->d_ino == child) {
-		if (child == parent)
-			*isdot = 1;
-		else if (dp->d_namlen == 2 &&
-		    dp->d_name[0] == '.' && dp->d_name[1] == '.')
-			*isdot = 1;
-		*mode = DTTOIF(dp->d_type);
-		return (1);
-	}
-	if (debug)
-		printf("ino %d doesn't match dirent ino %d in parent %d\n",
-		    child, dp->d_ino, parent);
-	return (0);
-}
-
-#define	VISIT_INDIR	0x0001
-#define	VISIT_EXT	0x0002
-
-/*
- * Read an indirect level which may or may not be linked into an inode.
- */
-static void
-indir_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, uint64_t *frags,
-    ino_visitor visitor, int flags)
-{
-	ufs2_daddr_t *bap2;
-	ufs1_daddr_t *bap1;
-	ufs_lbn_t lbnadd;
-	ufs2_daddr_t nblk;
-	ufs_lbn_t nlbn;
-	int level;
-	int i;
-
-	/*
-	 * Don't visit indirect blocks with contents we can't trust.  This
-	 * should only happen when indir_visit() is called to complete a
-	 * truncate that never finished and not when a pointer is found via
-	 * an inode.
-	 */
-	if (blk == 0)
-		return;
-	if (blk_isindir(blk, ino, lbn) == 0) {
-		if (debug)
-			printf("blk %jd ino %d lbn %jd is not indir.\n",
-			    blk, ino, lbn);
-		goto out;
-	}
-	level = lbn_level(lbn);
-	if (level == -1) {
-		abort();
-		errx(1, "Invalid level for lbn %jd", lbn);
-	}
-	lbnadd = 1;
-	for (i = level; i > 0; i--)
-		lbnadd *= NINDIR(fs);
-	bap1 = (void *)dblk_read(blk, fs->fs_bsize);
-	bap2 = (void *)bap1;
-	for (i = 0; i < NINDIR(fs); i++) {
-		if (fs->fs_magic == FS_UFS1_MAGIC)
-			nblk = *bap1++;
-		else
-			nblk = *bap2++;
-		if (nblk == 0)
-			continue;
-		if (level == 0) {
-			nlbn = -lbn + i * lbnadd;
-			(*frags) += fs->fs_frag;
-			visitor(ino, nlbn, nblk, fs->fs_frag);
-		} else {
-			nlbn = (lbn + 1) - (i * lbnadd);
-			indir_visit(ino, nlbn, nblk, frags, visitor, flags);
-		}
-	}
-out:
-	if (flags & VISIT_INDIR) {
-		(*frags) += fs->fs_frag;
-		visitor(ino, lbn, blk, fs->fs_frag);
-	}
-}
-
-/*
- * Visit each block in an inode as specified by 'flags' and call a
- * callback function.  The callback may inspect or free blocks.  The
- * count of frags found according to the size in the file is returned.
- * This is not valid for sparse files but may be used to determine
- * the correct di_blocks for a file.
- */
-static uint64_t
-ino_visit(union dinode *ip, ino_t ino, ino_visitor visitor, int flags)
-{
-	ufs_lbn_t tmpval;
-	ufs_lbn_t lbn;
-	uint64_t size;
-	uint64_t fragcnt;
-	int mode;
-	int frags;
-	int i;
-
-	size = DIP(ip, di_size);
-	mode = DIP(ip, di_mode) & IFMT;
-	fragcnt = 0;
-	if ((flags & VISIT_EXT) &&
-	    fs->fs_magic == FS_UFS2_MAGIC && ip->dp2.di_extsize) {
-		for (i = 0; i < NXADDR; i++) {
-			if (ip->dp2.di_extb[i] == 0)
-				continue;
-			frags = sblksize(fs, ip->dp2.di_extsize, i);
-			frags = numfrags(fs, frags);
-			fragcnt += frags;
-			visitor(ino, -1 - i, ip->dp2.di_extb[i], frags);
-		}
-	}
-	/* Skip datablocks for short links and devices. */
-	if (mode == IFBLK || mode == IFCHR ||
-	    (mode == IFLNK && size < fs->fs_maxsymlinklen))
-		return (fragcnt);
-	for (i = 0; i < NDADDR; i++) {
-		if (DIP(ip, di_db[i]) == 0)
-			continue;
-		frags = sblksize(fs, size, i);
-		frags = numfrags(fs, frags);
-		fragcnt += frags;
-		visitor(ino, i, DIP(ip, di_db[i]), frags);
-	}
-	for (i = 0, tmpval = NINDIR(fs), lbn = NDADDR; i < NIADDR; i++,
-	    tmpval *= NINDIR(fs), lbn += tmpval) {
-		if (DIP(ip, di_ib[i]) == 0)
-			continue;
-		indir_visit(ino, -lbn - i, DIP(ip, di_ib[i]), &fragcnt, visitor,
-		    flags);
-	}
-	return (fragcnt);
-}
-
-/*
- * Null visitor function used when we just want to count blocks.
- */
-static void
-null_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
-{
-}
-
-/*
- * Recalculate di_blocks when we discover that a block allocation or
- * free was not successfully completed.  The kernel does not roll this back
- * because it would be too expensive to compute which indirects were
- * reachable at the time the inode was written.
- */
-static void
-ino_adjblks(ino_t ino)
-{
-	struct suj_ino *sino;
-	union dinode *ip;
-	uint64_t blocks;
-	uint64_t frags;
-
-	sino = ino_lookup(ino, 1);
-	if (sino->si_blkadj)
-		return;
-	sino->si_blkadj = 1;
-	ip = ino_read(ino);
-	/* No need to adjust zero'd inodes. */
-	if (DIP(ip, di_mode) == 0)
-		return;
-	frags = ino_visit(ip, ino, null_visit, VISIT_INDIR | VISIT_EXT);
-	blocks = fsbtodb(fs, frags);
-	if (blocks == DIP(ip, di_blocks))
-		return;
-	if (debug)
-		printf("ino %d adjusting block count from %jd to %jd\n",
-		    ino, DIP(ip, di_blocks), blocks);
-	DIP_SET(ip, di_blocks, blocks);
-	ino_dirty(ino);
-}
-
-static void
-blk_free_visit(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
-{
-	int mask;
-
-	mask = blk_isfree(blk, ino, lbn, frags);
-	if (debug)
-		printf("blk %jd freemask 0x%X\n", blk, mask);
-	blk_free(blk, mask, frags);
-}
-
-/*
- * Free a block or tree of blocks that was previously rooted in ino at
- * the given lbn.  If the lbn is an indirect all children are freed
- * recursively.
- */
-static void
-blk_free_lbn(ufs2_daddr_t blk, ino_t ino, ufs_lbn_t lbn, int frags, int follow)
-{
-	uint64_t resid;
-	int mask;
-
-	mask = blk_isfree(blk, ino, lbn, frags);
-	if (debug)
-		printf("blk %jd freemask 0x%X\n", blk, mask);
-	resid = 0;
-	if (lbn <= -NDADDR && follow && mask == 0)
-		indir_visit(ino, lbn, blk, &resid, blk_free_visit, VISIT_INDIR);
-	else
-		blk_free(blk, mask, frags);
-}
-
-static void
-ino_free_children(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
-{
-	struct suj_ino *sino;
-	struct suj_rec *srec;
-	struct jrefrec *rrec;
-	struct direct *dp;
-	off_t diroff;
-	uint8_t *block;
-	int skipparent;
-	int isparent;
-	int dpoff;
-	int size;
-
-	sino = ino_lookup(ino, 0);
-	if (sino)
-		skipparent = sino->si_skipparent;
-	else
-		skipparent = 0;
-	size = lfragtosize(fs, frags);
-	block = dblk_read(blk, size);
-	dp = (struct direct *)&block[0];
-	for (dpoff = 0; dpoff < size && dp->d_reclen; dpoff += dp->d_reclen) {
-		dp = (struct direct *)&block[dpoff];
-		if (dp->d_ino == 0 || dp->d_ino == WINO)
-			continue;
-		if (dp->d_namlen == 1 && dp->d_name[0] == '.')
-			continue;
-		isparent = dp->d_namlen == 2 && dp->d_name[0] == '.' &&
-		    dp->d_name[1] == '.';
-		if (isparent && skipparent == 1)
-			continue;
-		if (debug)
-			printf("Directory %d removing inode %d name %s\n",
-			    ino, dp->d_ino, dp->d_name);
-		/*
-		 * Lookup this inode to see if we have a record for it.
-		 * If not, we've already adjusted it assuming this path
-		 * was valid and we have to adjust once more.
-		 */
-		sino = ino_lookup(dp->d_ino, 0);
-		if (sino == NULL || sino->si_linkadj || sino->si_hasrecs == 0) {
-			ino_decr(dp->d_ino);
-			continue;
-		}
-		/*
-		 * Tell any child directories we've already removed their
-		 * parent.  Don't try to adjust our link down again.
-		 */
-		if (isparent == 0)
-			sino->si_skipparent = 1;
-		/*
-		 * If we haven't yet processed this inode we need to make
-		 * sure we will successfully discover the lost path.  If not
-		 * use nlinkadj to remember.
-		 */
-		diroff = lblktosize(fs, lbn) + dpoff;
-		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
-			rrec = (struct jrefrec *)srec->sr_rec;
-			if (rrec->jr_parent == ino &&
-			    rrec->jr_diroff == diroff)
-				break;
-		}
-		if (srec == NULL)
-			sino->si_nlinkadj--;
-	}
-}
-
-/*
- * Truncate an inode, freeing all blocks and decrementing all children's
- * link counts.  Free the inode back to the cg.
- */
-static void
-ino_truncate(union dinode *ip, ino_t ino, int mode)
-{
-	uint32_t gen;
-
-	if (ino == ROOTINO)
-		errx(1, "Attempting to free ROOTINO");
-	if (debug)
-		printf("Truncating and freeing ino %d, nlink %d, mode %o\n",
-		    ino, DIP(ip, di_nlink), DIP(ip, di_mode));
-
-	/* We are freeing an inode or directory. */
-	if ((DIP(ip, di_mode) & IFMT) == IFDIR)
-		ino_visit(ip, ino, ino_free_children, 0);
-	DIP_SET(ip, di_nlink, 0);
-	ino_visit(ip, ino, blk_free_visit, VISIT_EXT | VISIT_INDIR);
-	/* Here we have to clear the inode and release any blocks it holds. */
-	gen = DIP(ip, di_gen);
-	if (fs->fs_magic == FS_UFS1_MAGIC)
-		bzero(ip, sizeof(struct ufs1_dinode));
-	else
-		bzero(ip, sizeof(struct ufs2_dinode));
-	DIP_SET(ip, di_gen, gen);
-	ino_dirty(ino);
-	ino_free(ino, mode);
-	return;
-}
-
-/*
- * Adjust an inode's link count down by one when a directory goes away.
- */
-static void
-ino_decr(ino_t ino)
-{
-	union dinode *ip;
-	int reqlink;
-	int nlink;
-	int mode;
-
-	ip = ino_read(ino);
-	nlink = DIP(ip, di_nlink);
-	mode = DIP(ip, di_mode);
-	if (nlink < 1)
-		errx(1, "Inode %d link count %d invalid", ino, nlink);
-	if (mode == 0)
-		errx(1, "Inode %d has a link of %d with 0 mode.", ino, nlink);
-	nlink--;
-	if ((mode & IFMT) == IFDIR)
-		reqlink = 2;
-	else
-		reqlink = 1;
-	if (nlink < reqlink) {
-		if (debug)
-			printf("ino %d not enough links to live %d < %d\n",
-			    ino, nlink, reqlink);
-		ino_truncate(ip, ino, mode);
-		return;
-	}
-	DIP_SET(ip, di_nlink, nlink);
-	ino_dirty(ino);
-}
-
-/*
- * Adjust the inode link count to 'nlink'.  If the count reaches zero
- * free it.
- */
-static void
-ino_adjust(ino_t ino, int lastmode, nlink_t nlink)
-{
-	union dinode *ip;
-	int reqlink;
-	int mode;
-
-	ip = ino_read(ino);
-	mode = DIP(ip, di_mode) & IFMT;
-	if (nlink > LINK_MAX)
-		errx(1,
-		    "ino %d nlink manipulation error, new link %d, old link %d",
-		    ino, nlink, DIP(ip, di_nlink));
-	if (debug)
-		printf("Adjusting ino %d, nlink %d, old link %d lastmode %o\n",
-		    ino, nlink, DIP(ip, di_nlink), lastmode);
-	if (mode == 0) {
-		if (debug)
-			printf("ino %d, zero inode freeing bitmap\n", ino);
-		ino_free(ino, lastmode);
-		return;
-	}
-	/* XXX Should be an assert? */
-	if (mode != lastmode && debug)
-		printf("ino %d, mode %o != %o\n", ino, mode, lastmode);
-	if ((mode & IFMT) == IFDIR)
-		reqlink = 2;
-	else
-		reqlink = 1;
-	/* If the inode doesn't have enough links to live, free it. */
-	if (nlink < reqlink) {
-		if (debug)
-			printf("ino %d not enough links to live %d < %d\n",
-			    ino, nlink, reqlink);
-		ino_truncate(ip, ino, mode);
-		return;
-	}
-	/* If required write the updated link count. */
-	if (DIP(ip, di_nlink) == nlink) {
-		if (debug)
-			printf("ino %d, link matches, skipping.\n", ino);
-		return;
-	}
-	DIP_SET(ip, di_nlink, nlink);
-	ino_dirty(ino);
-}
-
-#define	DOTDOT_OFFSET	DIRECTSIZ(1)
-
-/*
- * Process records available for one inode and determine whether the
- * link count is correct or needs adjusting.
- *
- * XXX Failed to fix zero length directory.  Shouldn't .. have been mising?
- */
-static void
-ino_check(struct suj_ino *sino)
-{
-	struct suj_rec *srec;
-	struct jrefrec *rrec;
-	struct suj_ino *stmp;
-	nlink_t dotlinks;
-	int newlinks;
-	int removes;
-	int nlink;
-	ino_t ino;
-	int isdot;
-	int isat;
-	int mode;
-
-	if (sino->si_hasrecs == 0)
-		return;
-	ino = sino->si_ino;
-	/*
-	 * XXX ino_isfree currently is skipping initialized inodes
-	 * that are unreferenced.
-	 */
-	if (0 && ino_isfree(ino))
-		return;
-	rrec = (struct jrefrec *)TAILQ_FIRST(&sino->si_recs)->sr_rec;
-	nlink = rrec->jr_nlink;
-	newlinks = sino->si_nlinkadj;
-	dotlinks = 0;
-	removes = 0;
-	TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
-		rrec = (struct jrefrec *)srec->sr_rec;
-		isat = ino_isat(rrec->jr_parent, rrec->jr_diroff, 
-		    rrec->jr_ino, &mode, &isdot);
-		if (isat && (mode & IFMT) != (rrec->jr_mode & IFMT))
-			errx(1, "Inode mode/directory type mismatch %o != %o",
-			    mode, rrec->jr_mode);
-		if (debug)
-			printf("jrefrec: op %d ino %d, nlink %d, parent %d, "
-			    "diroff %jd, mode %o, isat %d, isdot %d\n",
-			    rrec->jr_op, rrec->jr_ino, rrec->jr_nlink,
-			    rrec->jr_parent, rrec->jr_diroff, rrec->jr_mode,
-			    isat, isdot);
-		mode = rrec->jr_mode & IFMT;
-		if (rrec->jr_op == JOP_REMREF)
-			removes++;
-		newlinks += isat;
-		if (isdot)
-			dotlinks += isat;
-	}
-	/*
-	 * The number of links that remain are the starting link count
-	 * subtracted by the total number of removes with the total
-	 * links discovered back in.  An incomplete remove thus
-	 * makes no change to the link count but an add increases
-	 * by one.
-	 */
-	nlink += newlinks;
-	nlink -= removes;
-	/*
-	 * If it's a directory with no real names pointing to it go ahead
-	 * and truncate it.  This will free any children.
-	 */
-	if ((mode & IFMT) == IFDIR && nlink - dotlinks == 0) {
-		nlink = 0;
-		/*
-		 * Mark any .. links so they know not to free this inode
-		 * when they are removed.
-		 */
-		TAILQ_FOREACH(srec, &sino->si_recs, sr_next) {
-			rrec = (struct jrefrec *)srec->sr_rec;
-			if (rrec->jr_diroff == DOTDOT_OFFSET) {
-				stmp = ino_lookup(rrec->jr_parent, 0);
-				if (stmp)
-					stmp->si_skipparent = 1;
-			}
-		}
-	}
-	sino->si_linkadj = 1;
-	ino_adjust(ino, mode, nlink);
-}
-
-/*
- * Process records available for one block and determine whether it is
- * still allocated and whether the owning inode needs to be updated or
- * a free completed.
- */
-static void
-blk_check(struct suj_blk *sblk)
-{
-	struct suj_rec *srec;
-	struct jblkrec *brec;
-	ufs2_daddr_t blk;
-	int mask;
-	int frags;
-	int isat;
-
-	/*
-	 * Each suj_blk actually contains records for any fragments in that
-	 * block.  As a result we must evaluate each record individually.
-	 */
-	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
-		brec = (struct jblkrec *)srec->sr_rec;
-		frags = brec->jb_frags;
-		blk = brec->jb_blkno + brec->jb_oldfrags;
-		isat = blk_isat(brec->jb_ino, brec->jb_lbn, blk, &frags);
-		if (debug)
-			printf("op %d blk %jd ino %d lbn %jd frags %d isat %d (%d)\n",
-			    brec->jb_op, blk, brec->jb_ino, brec->jb_lbn,
-			    brec->jb_frags, isat, frags);
-		/*
-		 * If we found the block at this address we still have to
-		 * determine if we need to free the tail end that was
-		 * added by adding contiguous fragments from the same block.
-		 */
-		if (isat == 1) {
-			if (frags == brec->jb_frags)
-				continue;
-			mask = blk_isfree(blk, brec->jb_ino, brec->jb_lbn,
-			    brec->jb_frags);
-			mask >>= frags;
-			blk += frags;
-			frags = brec->jb_frags - frags;
-			blk_free(blk, mask, frags);
-			ino_adjblks(brec->jb_ino);
-			continue;
-		}
-		/*
-	 	 * The block wasn't found, attempt to free it.  It won't be
-		 * freed if it was actually reallocated.  If this was an
-		 * allocation we don't want to follow indirects as they
-		 * may not be written yet.  Any children of the indirect will
-		 * have their own records.  If it's a free we need to
-		 * recursively free children.
-		 */
-		blk_free_lbn(blk, brec->jb_ino, brec->jb_lbn, brec->jb_frags,
-		    brec->jb_op == JOP_FREEBLK);
-		ino_adjblks(brec->jb_ino);
-	}
-}
-
-/*
- * Walk the list of inode and block records for this cg, recovering any
- * changes which were not complete at the time of crash.
- */
-static void
-cg_check(struct suj_cg *sc)
-{
-	struct suj_blk *nextb;
-	struct suj_ino *nexti;
-	struct suj_ino *sino;
-	struct suj_blk *sblk;
-	int i;
-
-	if (debug)
-		printf("Recovering cg %d\n", sc->sc_cgx);
-
-	for (i = 0; i < SUJ_HASHSIZE; i++)
-		LIST_FOREACH_SAFE(sino, &sc->sc_inohash[i], si_next, nexti)
-			ino_check(sino);
-
-	for (i = 0; i < SUJ_HASHSIZE; i++)
-		LIST_FOREACH_SAFE(sblk, &sc->sc_blkhash[i], sb_next, nextb)
-			blk_check(sblk);
-}
-
-/*
- * Write a potentially dirty cg.  All inodes must be written before the
- * cg maps are so that an allocated inode is never marked free, even if
- * we crash during fsck.
- */
-static void
-cg_write(struct suj_cg *sc)
-{
-	struct ino_blk *iblk;
-	ufs1_daddr_t fragno, cgbno, maxbno;
-	u_int8_t *blksfree;
-	struct cg *cgp;
-	int blk;
-	int i;
-
-	for (i = 0; i < SUJ_HASHSIZE; i++)
-		LIST_FOREACH(iblk, &sc->sc_iblkhash[i], ib_next)
-			iblk_write(iblk);
-	if (sc->sc_dirty == 0)
-		return;
-	/*
-	 * Fix the frag and cluster summary.
-	 */
-	cgp = sc->sc_cgp;
-	cgp->cg_cs.cs_nbfree = 0;
-	cgp->cg_cs.cs_nffree = 0;
-	bzero(&cgp->cg_frsum, sizeof(cgp->cg_frsum));
-	maxbno = fragstoblks(fs, fs->fs_fpg);
-	if (fs->fs_contigsumsize > 0) {
-		for (i = 1; i <= fs->fs_contigsumsize; i++)
-			cg_clustersum(cgp)[i] = 0;
-		bzero(cg_clustersfree(cgp), howmany(maxbno, CHAR_BIT));
-	}
-	blksfree = cg_blksfree(cgp);
-	for (cgbno = 0; cgbno < maxbno; cgbno++) {
-		if (ffs_isfreeblock(fs, blksfree, cgbno))
-			continue;
-		if (ffs_isblock(fs, blksfree, cgbno)) {
-			ffs_clusteracct(fs, cgp, cgbno, 1);
-			cgp->cg_cs.cs_nbfree++;
-			continue;
-		}
-		fragno = blkstofrags(fs, cgbno);
-		blk = blkmap(fs, blksfree, fragno);
-		ffs_fragacct(fs, blk, cgp->cg_frsum, 1);
-		for (i = 0; i < fs->fs_frag; i++)
-			if (isset(blksfree, fragno + i))
-				cgp->cg_cs.cs_nffree++;
-	}
-	/*
-	 * Update the superblock cg summary from our now correct values
-	 * before writing the block.
-	 */
-	fs->fs_cs(fs, sc->sc_cgx) = cgp->cg_cs;
-	if (bwrite(disk, fsbtodb(fs, cgtod(fs, sc->sc_cgx)), sc->sc_cgbuf,
-	    fs->fs_bsize) == -1)
-		err(1, "Unable to write cylinder group %d", sc->sc_cgx);
-}
-
-static void
-cg_apply(void (*apply)(struct suj_cg *))
-{
-	struct suj_cg *scg;
-	int i;
-
-	for (i = 0; i < SUJ_HASHSIZE; i++)
-		LIST_FOREACH(scg, &cghash[i], sc_next)
-			apply(scg);
-}
-
-/*
- * Process the unlinked but referenced file list.  Freeing all inodes.
- */
-static void
-ino_unlinked(void)
-{
-	union dinode *ip;
-	uint16_t mode;
-	ino_t inon;
-	ino_t ino;
-
-	ino = fs->fs_sujfree;
-	fs->fs_sujfree = 0;
-	while (ino != 0) {
-		ip = ino_read(ino);
-		mode = DIP(ip, di_mode) & IFMT;
-		inon = DIP(ip, di_freelink);
-		DIP_SET(ip, di_freelink, 0);
-		/*
-		 * XXX Should this be an errx?
-		 */
-		if (DIP(ip, di_nlink) == 0) {
-			if (debug)
-				printf("Freeing unlinked ino %d mode %o\n",
-				    ino, mode);
-			ino_truncate(ip, ino, mode);
-		} else if (debug)
-			printf("Skipping ino %d mode %o with link %d\n",
-			    ino, mode, DIP(ip, di_nlink));
-		ino = inon;
-	}
-}
-
-/*
- * If we see two ops for the same inode to the same parent at the same
- * offset we could miscount the link with ino_isat() returning twice.
- * Keep only the first record because it has the valid link count but keep
- * the mode from the final op as that should be the correct mode in case
- * it changed.
- */
-static void
-suj_build_ino(struct jrefrec *refrec)
-{
-	struct jmvrec *mvrec;
-	struct suj_rec *srec;
-	struct suj_ino *sino;
-	struct suj_rec *srn;
-	struct jrefrec *rrn;
-
-	if (debug)
-		printf("suj_build_ino: op %d, ino %d, nlink %d, parent %d, diroff %jd\n", 
-		    refrec->jr_op, refrec->jr_ino, refrec->jr_nlink, refrec->jr_parent,
-		    refrec->jr_diroff);
-	sino = ino_lookup(refrec->jr_ino, 1);
-	/*
-	 * Search for a mvrec that matches this offset.  Whether it's an add
-	 * or a remove we can delete the mvref.  It no longer applies to this
-	 * location.
-	 *
-	 * For removes, we have to find the original offset so we can create
-	 * a remove that matches the earlier add so it can be abandoned
-	 * if necessary.  We create an add in the new location so we can
-	 * tolerate the directory block as it existed before or after
-	 * the move.
-	 */
-	if (!TAILQ_EMPTY(&sino->si_movs)) {
-		for (srn = TAILQ_LAST(&sino->si_movs, srechd); srn;
-		    srn = TAILQ_PREV(srn, srechd, sr_next)) {
-			mvrec = (struct jmvrec *)srn->sr_rec;
-			if (mvrec->jm_parent != refrec->jr_parent ||
-			    mvrec->jm_newoff != refrec->jr_diroff)
-				continue;
-			TAILQ_REMOVE(&sino->si_movs, srn, sr_next);
-			if (refrec->jr_op == JOP_REMREF) {
-				rrn = errmalloc(sizeof(*refrec));
-				*rrn = *refrec;
-				rrn->jr_op = JOP_ADDREF;
-				suj_build_ino(rrn);
-				refrec->jr_diroff = mvrec->jm_oldoff;
-			}
-		}
-	}
-	/*
-	 * We walk backwards so that adds and removes are evaluated in the
-	 * correct order.
-	 */
-	for (srn = TAILQ_LAST(&sino->si_recs, srechd); srn;
-	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
-		rrn = (struct jrefrec *)srn->sr_rec;
-		if (rrn->jr_parent != refrec->jr_parent ||
-		    rrn->jr_diroff != refrec->jr_diroff)
-			continue;
-		if (debug)
-			printf("Discarding dup.\n");
-		rrn->jr_mode = refrec->jr_mode;
-		return;
-	}
-	sino->si_hasrecs = 1;
-	srec = errmalloc(sizeof(*srec));
-	srec->sr_rec = (union jrec *)refrec;
-	TAILQ_INSERT_TAIL(&sino->si_recs, srec, sr_next);
-}
-
-/*
- * Apply a move record to an inode.  We must search for adds that preceed us
- * and add duplicates because we won't know which location to search first.
- * Then we add movs to a queue that is maintained until the moved location
- * is removed.  If a single record is moved multiple times we only maintain
- * one copy that contains the original and final diroffs.
- */
-static void
-suj_move_ino(struct jmvrec *mvrec)
-{
-	struct jrefrec *refrec;
-	struct suj_ino *sino;
-	struct suj_rec *srec;
-	struct jmvrec *mvrn;
-	struct suj_rec *srn;
-	struct jrefrec *rrn;
-
-	if (debug)
-		printf("suj_move_ino: ino %d, parent %d, diroff %jd, oldoff %jd\n", 
-		    mvrec->jm_ino, mvrec->jm_parent, mvrec->jm_newoff,
-		    mvrec->jm_oldoff);
-	sino = ino_lookup(mvrec->jm_ino, 0);
-	if (sino == NULL)
-		return;
-	/*
-	 * We walk backwards so we only evaluate the most recent record at
-	 * this offset.
-	 */
-	for (srn = TAILQ_LAST(&sino->si_recs, srechd); srn;
-	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
-		rrn = (struct jrefrec *)srn->sr_rec;
-		if (rrn->jr_op != JOP_ADDREF)
-			continue;
-		if (rrn->jr_parent != mvrec->jm_parent ||
-		    rrn->jr_diroff != mvrec->jm_oldoff)
-			continue;
-		/*
-		 * When an entry is moved we don't know whether the write
-		 * to move has completed yet.  To resolve this we create
-		 * a new add dependency in the new location as if it were added
-		 * twice.  Only one will succeed.
-		 */
-		refrec = errmalloc(sizeof(*refrec));
-		refrec->jr_op = JOP_ADDREF;
-		refrec->jr_ino = mvrec->jm_ino;
-		refrec->jr_parent = mvrec->jm_parent;
-		refrec->jr_diroff = mvrec->jm_newoff;
-		refrec->jr_mode = rrn->jr_mode;
-		refrec->jr_nlink = rrn->jr_nlink;
-		suj_build_ino(refrec);
-		break;
-	}
-	/*
-	 * Add this mvrec to the queue of pending mvs.
-	 */
-	for (srn = TAILQ_LAST(&sino->si_movs, srechd); srn;
-	    srn = TAILQ_PREV(srn, srechd, sr_next)) {
-		mvrn = (struct jmvrec *)srn->sr_rec;
-		if (mvrn->jm_parent != mvrec->jm_parent ||
-		    mvrn->jm_newoff != mvrec->jm_oldoff)
-			continue;
-		mvrn->jm_newoff = mvrec->jm_newoff;
-		return;
-	}
-	srec = errmalloc(sizeof(*srec));
-	srec->sr_rec = (union jrec *)mvrec;
-	TAILQ_INSERT_TAIL(&sino->si_movs, srec, sr_next);
-}
-
-/*
- * Modify journal records so they refer to the base block number
- * and a start and end frag range.  This is to facilitate the discovery
- * of overlapping fragment allocations.
- */
-static void
-suj_build_blk(struct jblkrec *blkrec)
-{
-	struct suj_rec *srec;
-	struct suj_blk *sblk;
-	struct jblkrec *blkrn;
-	ufs2_daddr_t blk;
-	int frag;
-
-	if (debug)
-		printf("suj_build_blk: op %d blkno %jd frags %d oldfrags %d "
-		    "ino %d lbn %jd\n",
-		    blkrec->jb_op, blkrec->jb_blkno, blkrec->jb_frags,
-		    blkrec->jb_oldfrags, blkrec->jb_ino, blkrec->jb_lbn);
-	blk = blknum(fs, blkrec->jb_blkno);
-	frag = fragnum(fs, blkrec->jb_blkno);
-	sblk = blk_lookup(blk, 1);
-	/*
-	 * Rewrite the record using oldfrags to indicate the offset into
-	 * the block.  Leave jb_frags as the actual allocated count.
-	 */
-	blkrec->jb_blkno -= frag;
-	blkrec->jb_oldfrags = frag;
-	if (blkrec->jb_oldfrags + blkrec->jb_frags > fs->fs_frag)
-		errx(1, "Invalid fragment count %d oldfrags %d",
-		    blkrec->jb_frags, frag);
-	/*
-	 * Detect dups.  If we detect a dup we always discard the oldest
-	 * record as it is superseded by the new record.  This speeds up
-	 * later stages but also eliminates free records which are used
-	 * to indicate that the contents of indirects can be trusted.
-	 */
-	TAILQ_FOREACH(srec, &sblk->sb_recs, sr_next) {
-		blkrn = (struct jblkrec *)srec->sr_rec;
-		if (blkrn->jb_ino != blkrec->jb_ino ||
-		    blkrn->jb_lbn != blkrec->jb_lbn ||
-		    blkrn->jb_blkno != blkrec->jb_blkno ||
-		    blkrn->jb_frags != blkrec->jb_frags ||
-		    blkrn->jb_oldfrags != blkrec->jb_oldfrags)
-			continue;
-		if (debug)
-			printf("Removed dup.\n");
-		/* Discard the free which is a dup with an alloc. */
-		if (blkrec->jb_op == JOP_FREEBLK)
-			return;
-		TAILQ_REMOVE(&sblk->sb_recs, srec, sr_next);
-		free(srec);
-		break;
-	}
-	srec = errmalloc(sizeof(*srec));
-	srec->sr_rec = (union jrec *)blkrec;
-	TAILQ_INSERT_TAIL(&sblk->sb_recs, srec, sr_next);
-}
-
-/*
- * Build up tables of the operations we need to recover.
- */
-static void
-suj_build(void)
-{
-	struct suj_seg *seg;
-	union jrec *rec;
-	int i;
-
-	TAILQ_FOREACH(seg, &allsegs, ss_next) {
-		rec = (union jrec *)seg->ss_blk;
-		rec++;	/* skip the segrec. */
-		if (debug)
-			printf("seg %jd has %d records, oldseq %jd.\n",
-			    seg->ss_rec.jsr_seq, seg->ss_rec.jsr_cnt,
-			    seg->ss_rec.jsr_oldest);
-		for (i = 0; i < seg->ss_rec.jsr_cnt; i++, rec++) {
-			switch (rec->rec_jrefrec.jr_op) {
-			case JOP_ADDREF:
-			case JOP_REMREF:
-				suj_build_ino((struct jrefrec *)rec);
-				break;
-			case JOP_MVREF:
-				suj_move_ino((struct jmvrec *)rec);
-				break;
-			case JOP_NEWBLK:
-			case JOP_FREEBLK:
-				suj_build_blk((struct jblkrec *)rec);
-				break;
-			default:
-				errx(1, "Unknown journal operation %d (%d)",
-				    rec->rec_jrefrec.jr_op, i);
-			}
-		}
-	}
-}
-
-/*
- * Prune the journal segments to those we care about based on the
- * oldest sequence in the newest segment.  Order the segment list
- * based on sequence number.
- */
-static void
-suj_prune(void)
-{
-	struct suj_seg *seg;
-	struct suj_seg *segn;
-	uint64_t newseq;
-	int discard;
-
-	if (debug)
-		printf("Pruning up to %jd\n", oldseq);
-	/* First free the expired segments. */
-	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
-		if (seg->ss_rec.jsr_seq >= oldseq)
-			continue;
-		TAILQ_REMOVE(&allsegs, seg, ss_next);
-		free(seg->ss_blk);
-		free(seg);
-	}
-	/* Next ensure that segments are ordered properly. */
-	seg = TAILQ_FIRST(&allsegs);
-	if (seg == NULL) {
-		if (debug)
-			printf("Empty journal\n");
-		return;
-	}
-	newseq = seg->ss_rec.jsr_seq;
-	for (;;) {
-		seg = TAILQ_LAST(&allsegs, seghd);
-		if (seg->ss_rec.jsr_seq >= newseq)
-			break;
-		TAILQ_REMOVE(&allsegs, seg, ss_next);
-		TAILQ_INSERT_HEAD(&allsegs, seg, ss_next);
-		newseq = seg->ss_rec.jsr_seq;
-		
-	}
-	if (newseq != oldseq)
-		errx(1, "Journal file sequence mismatch %jd != %jd",
-		    newseq, oldseq);
-	/*
-	 * The kernel may asynchronously write segments which can create
-	 * gaps in the sequence space.  Throw away any segments after the
-	 * gap as the kernel guarantees only those that are contiguously
-	 * reachable are marked as completed.
-	 */
-	discard = 0;
-	TAILQ_FOREACH_SAFE(seg, &allsegs, ss_next, segn) {
-		if (!discard && newseq++ == seg->ss_rec.jsr_seq)
-			continue;
-		discard = 1;
-		if (debug)
-			printf("Journal order mismatch %jd != %jd pruning\n",
-			    newseq-1, seg->ss_rec.jsr_seq);
-		TAILQ_REMOVE(&allsegs, seg, ss_next);
-		free(seg->ss_blk);
-		free(seg);
-	}
-	if (debug)
-		printf("Processing journal segments from %jd to %jd\n",
-		    oldseq, newseq-1);
-}
-
-/*
- * Verify the journal inode before attempting to read records.
- */
-static void
-suj_verifyino(union dinode *ip)
-{
-
-	if (DIP(ip, di_nlink) != 1)
-		errx(1, "Invalid link count %d for journal inode %d",
-		    DIP(ip, di_nlink), fs->fs_sujournal);
-
-	if (DIP(ip, di_mode) != IFREG)
-		errx(1, "Invalid mode %d for journal inode %d",
-		    DIP(ip, di_mode), fs->fs_sujournal);
-
-	if (DIP(ip, di_size) < SUJ_MIN || DIP(ip, di_size) > SUJ_MAX)
-		errx(1, "Invalid size %jd for journal inode %d",
-		    DIP(ip, di_size), fs->fs_sujournal);
-
-	if (DIP(ip, di_modrev) != fs->fs_mtime)
-		errx(1, "Journal timestamp does not match fs mount time");
-	/* XXX Add further checks. */
-}
-
-struct jblocks {
-	struct jextent *jb_extent;	/* Extent array. */
-	int		jb_avail;	/* Available extents. */
-	int		jb_used;	/* Last used extent. */
-	int		jb_head;	/* Allocator head. */
-	int		jb_off;		/* Allocator extent offset. */
-};
-struct jextent {
-	ufs2_daddr_t	je_daddr;	/* Disk block address. */
-	int		je_blocks;	/* Disk block count. */
-};
-
-struct jblocks *suj_jblocks;
-
-static struct jblocks *
-jblocks_create(void)
-{
-	struct jblocks *jblocks;
-	int size;
-
-	jblocks = errmalloc(sizeof(*jblocks));
-	jblocks->jb_avail = 10;
-	jblocks->jb_used = 0;
-	jblocks->jb_head = 0;
-	jblocks->jb_off = 0;
-	size = sizeof(struct jextent) * jblocks->jb_avail;
-	jblocks->jb_extent = errmalloc(size);
-	bzero(jblocks->jb_extent, size);
-
-	return (jblocks);
-}
-
-/*
- * Return the next available disk block and the amount of contiguous
- * free space it contains.
- */
-static ufs2_daddr_t
-jblocks_next(struct jblocks *jblocks, int bytes, int *actual)
-{
-	struct jextent *jext;
-	ufs2_daddr_t daddr;
-	int freecnt;
-	int blocks;
-
-	blocks = bytes / DEV_BSIZE;
-	jext = &jblocks->jb_extent[jblocks->jb_head];
-	freecnt = jext->je_blocks - jblocks->jb_off;
-	if (freecnt == 0) {
-		jblocks->jb_off = 0;
-		if (++jblocks->jb_head > jblocks->jb_used)
-			return (0);
-		jext = &jblocks->jb_extent[jblocks->jb_head];
-		freecnt = jext->je_blocks;
-	}
-	if (freecnt > blocks)
-		freecnt = blocks;
-	*actual = freecnt * DEV_BSIZE;
-	daddr = jext->je_daddr + jblocks->jb_off;
-
-	return (daddr);
-}
-
-/*
- * Advance the allocation head by a specified number of bytes, consuming
- * one journal segment.
- */
-static void
-jblocks_advance(struct jblocks *jblocks, int bytes)
-{
-
-	jblocks->jb_off += bytes / DEV_BSIZE;
-}
-
-static void
-jblocks_destroy(struct jblocks *jblocks)
-{
-
-	free(jblocks->jb_extent);
-	free(jblocks);
-}
-
-static void
-jblocks_add(struct jblocks *jblocks, ufs2_daddr_t daddr, int blocks)
-{
-	struct jextent *jext;
-	int size;
-
-	jext = &jblocks->jb_extent[jblocks->jb_used];
-	/* Adding the first block. */
-	if (jext->je_daddr == 0) {
-		jext->je_daddr = daddr;
-		jext->je_blocks = blocks;
-		return;
-	}
-	/* Extending the last extent. */
-	if (jext->je_daddr + jext->je_blocks == daddr) {
-		jext->je_blocks += blocks;
-		return;
-	}
-	/* Adding a new extent. */
-	if (++jblocks->jb_used == jblocks->jb_avail) {
-		jblocks->jb_avail *= 2;
-		size = sizeof(struct jextent) * jblocks->jb_avail;
-		jext = errmalloc(size);
-		bzero(jext, size);
-		bcopy(jblocks->jb_extent, jext,
-		    sizeof(struct jextent) * jblocks->jb_used);
-		free(jblocks->jb_extent);
-		jblocks->jb_extent = jext;
-	}
-	jext = &jblocks->jb_extent[jblocks->jb_used];
-	jext->je_daddr = daddr;
-	jext->je_blocks = blocks;
-
-	return;
-}
-
-/*
- * Add a file block from the journal to the extent map.  We can't read
- * each file block individually because the kernel treats it as a circular
- * buffer and segments may span mutliple contiguous blocks.
- */
-static void
-suj_add_block(ino_t ino, ufs_lbn_t lbn, ufs2_daddr_t blk, int frags)
-{
-
-	jblocks_add(suj_jblocks, fsbtodb(fs, blk), fsbtodb(fs, frags));
-}
-
-static void
-suj_read(void)
-{
-	uint8_t block[1 * 1024 * 1024];
-	struct suj_seg *seg;
-	struct jsegrec *rec;
-	ufs2_daddr_t blk;
-	int recsize;
-	int size;
-
-	/*
-	 * Read records until we exhaust the journal space.  If we find
-	 * an invalid record we start searching for a valid segment header
-	 * at the next block.  This is because we don't have a head/tail
-	 * pointer and must recover the information indirectly.  At the gap
-	 * between the head and tail we won't necessarily have a valid
-	 * segment.
-	 */
-	for (;;) {
-		size = sizeof(block);
-		blk = jblocks_next(suj_jblocks, size, &size);
-		if (blk == 0)
-			return;
-		/*
-		 * Read 1MB at a time and scan for records within this block.
-		 */
-		if (bread(disk, blk, &block, size) == -1)
-			err(1, "Error reading journal block %jd",
-			    (intmax_t)blk);
-		for (rec = (void *)block; size; size -= recsize,
-		    rec = (struct jsegrec *)((uintptr_t)rec + recsize)) {
-			recsize = DEV_BSIZE;
-			if (rec->jsr_time != fs->fs_mtime) {
-				if (debug)
-					printf("Rec time %jd != fs mtime %jd\n",
-					    rec->jsr_time, fs->fs_mtime);
-				jblocks_advance(suj_jblocks, recsize);
-				continue;
-			}
-			if (rec->jsr_cnt == 0) {
-				if (debug)
-					printf("Found illegal count %d\n",
-					    rec->jsr_cnt);
-				jblocks_advance(suj_jblocks, recsize);
-				continue;
-			}
-			recsize = roundup2((rec->jsr_cnt + 1) * JREC_SIZE,
-			    DEV_BSIZE);
-			if (recsize > size) {
-				/*
-				 * We may just have run out of buffer, restart
-				 * the loop to re-read from this spot.
-				 */
-				if (size < fs->fs_bsize && 
-				    recsize <= fs->fs_bsize) {
-					recsize = size;
-					continue;
-				}
-				if (debug)
-					printf("Found invalid segsize %d > %d\n",
-					    recsize, size);
-				recsize = DEV_BSIZE;
-				jblocks_advance(suj_jblocks, recsize);
-				continue;
-			}
-			seg = errmalloc(sizeof(*seg));
-			seg->ss_blk = errmalloc(recsize);
-			seg->ss_rec = *rec;
-			bcopy((void *)rec, seg->ss_blk, recsize);
-			if (rec->jsr_oldest > oldseq)
-				oldseq = rec->jsr_oldest;
-			TAILQ_INSERT_TAIL(&allsegs, seg, ss_next);
-			jrecs += rec->jsr_cnt;
-			jbytes += recsize;
-			jblocks_advance(suj_jblocks, recsize);
-		}
-	}
-}
-
-/*
- * Orchestrate the verification of a filesystem via the softupdates journal.
- */
-void
-suj_check(const char *filesys)
-{
-	union dinode *jip;
-	uint64_t blocks;
-
-	opendisk(filesys);
-	TAILQ_INIT(&allsegs);
-	/*
-	 * Fetch the journal inode and verify it.
-	 */
-	jip = ino_read(fs->fs_sujournal);
-	printf("SU+J Checking %s\n", filesys);
-	suj_verifyino(jip);
-	/*
-	 * Build a list of journal blocks in jblocks before parsing the
-	 * available journal blocks in with suj_read().
-	 */
-	printf("Reading %jd byte journal from inode %d.\n",
-	    DIP(jip, di_size), fs->fs_sujournal);
-	suj_jblocks = jblocks_create();
-	blocks = ino_visit(jip, fs->fs_sujournal, suj_add_block, 0);
-	if (blocks != numfrags(fs, DIP(jip, di_size)))
-		errx(1, "Sparse journal inode %d.\n", fs->fs_sujournal);
-	suj_read();
-	jblocks_destroy(suj_jblocks);
-	suj_jblocks = NULL;
-	if (reply("RECOVER")) {
-		printf("Building recovery table.\n");
-		suj_prune();
-		suj_build();
-		printf("Resolving unreferenced inode list.\n");
-		ino_unlinked();
-		printf("Processing journal entries.\n");
-		cg_apply(cg_check);
-	}
-	if (reply("WRITE CHANGES"))
-		cg_apply(cg_write);
-	printf("%jd journal records in %jd bytes for %.2f%% utilization\n",
-	    jrecs, jbytes, ((float)jrecs / (float)(jbytes / JREC_SIZE)) * 100);
-	printf("Freed %jd inodes (%jd directories) %jd blocks and %jd frags.\n",
-	    freeinos, freedir, freeblocks, freefrags);
-	/* Write back superblock. */
-	closedisk(filesys);
-}

From df9109ee305e47e34c553644b4443f86910640c5 Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Sat, 24 Apr 2010 07:59:45 +0000
Subject: [PATCH 229/532]  - Temporarily lower WARNS until I fix alignment
 warnings on sparc64.

Reported by:	Florian Smeets
---
 sbin/tunefs/Makefile | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sbin/tunefs/Makefile b/sbin/tunefs/Makefile
index d501d104b4b..d5313c44552 100644
--- a/sbin/tunefs/Makefile
+++ b/sbin/tunefs/Makefile
@@ -6,4 +6,6 @@ DPADD=	${LIBUFS}
 LDADD=	-lufs
 MAN=	tunefs.8
 
+WARNS=	3
+
 .include <bsd.prog.mk>

From ed7806879b9fb7bd34a9ec2ae1360760e87d244d Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sat, 24 Apr 2010 12:49:52 +0000
Subject: [PATCH 230/532] Move the constants specifying the size of struct
 kinfo_proc into machine-specific header files. Add KINFO_PROC32_SIZE for
 struct kinfo_proc32 for architectures providing COMPAT_FREEBSD32. Add
 CTASSERT for the size of struct kinfo_proc32.

Submitted by:	pluknet
Reviewed by:	imp, jhb, nwhitehorn
MFC after:	2 weeks
---
 sys/amd64/include/proc.h   |  3 +++
 sys/arm/include/proc.h     |  2 ++
 sys/i386/include/proc.h    |  2 ++
 sys/ia64/include/proc.h    |  3 +++
 sys/kern/kern_proc.c       |  3 +++
 sys/mips/include/proc.h    |  6 ++++++
 sys/powerpc/include/proc.h |  2 ++
 sys/sparc64/include/proc.h |  2 ++
 sys/sun4v/include/proc.h   |  2 ++
 sys/sys/user.h             | 27 ++-------------------------
 10 files changed, 27 insertions(+), 25 deletions(-)

diff --git a/sys/amd64/include/proc.h b/sys/amd64/include/proc.h
index 33d5181d2d5..126412cad45 100644
--- a/sys/amd64/include/proc.h
+++ b/sys/amd64/include/proc.h
@@ -53,6 +53,9 @@ struct mdproc {
 	struct system_segment_descriptor md_ldt_sd;
 };
 
+#define KINFO_PROC_SIZE 1088
+#define KINFO_PROC32_SIZE 768
+
 #ifdef	_KERNEL
 
 /* Get the current kernel thread stack usage. */
diff --git a/sys/arm/include/proc.h b/sys/arm/include/proc.h
index c7b2a4ef5c9..37051bf518b 100644
--- a/sys/arm/include/proc.h
+++ b/sys/arm/include/proc.h
@@ -60,4 +60,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
+#define KINFO_PROC_SIZE 792
+
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/i386/include/proc.h b/sys/i386/include/proc.h
index 1e0b9f9db10..0b9fca2b55f 100644
--- a/sys/i386/include/proc.h
+++ b/sys/i386/include/proc.h
@@ -57,6 +57,8 @@ struct mdproc {
 	struct proc_ldt *md_ldt;	/* (t) per-process ldt */
 };
 
+#define KINFO_PROC_SIZE 768
+
 #ifdef	_KERNEL
 
 /* Get the current kernel thread stack usage. */
diff --git a/sys/ia64/include/proc.h b/sys/ia64/include/proc.h
index 6bf9c78c88b..81e2e4f7bb8 100644
--- a/sys/ia64/include/proc.h
+++ b/sys/ia64/include/proc.h
@@ -38,4 +38,7 @@ struct mdproc {
 	int		__dummy;	/* Avoid having an empty struct. */
 };
 
+#define KINFO_PROC_SIZE 1088
+#define KINFO_PROC32_SIZE 768
+
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index d4943c09d8f..5601ecf67cd 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -151,6 +151,9 @@ int kstack_pages = KSTACK_PAGES;
 SYSCTL_INT(_kern, OID_AUTO, kstack_pages, CTLFLAG_RD, &kstack_pages, 0, "");
 
 CTASSERT(sizeof(struct kinfo_proc) == KINFO_PROC_SIZE);
+#ifdef COMPAT_FREEBSD32
+CTASSERT(sizeof(struct kinfo_proc32) == KINFO_PROC32_SIZE);
+#endif
 
 /*
  * Initialize global process hashing structures.
diff --git a/sys/mips/include/proc.h b/sys/mips/include/proc.h
index 0491b11b02e..1479edab293 100644
--- a/sys/mips/include/proc.h
+++ b/sys/mips/include/proc.h
@@ -68,4 +68,10 @@ struct thread;
 void	mips_cpu_switch(struct thread *, struct thread *, struct mtx *);
 void	mips_cpu_throw(struct thread *, struct thread *);
 
+#ifdef __mips_n64
+#define KINFO_PROC_SIZE 1088
+#else
+#define KINFO_PROC_SIZE 816
+#endif
+
 #endif	/* !_MACHINE_PROC_H_ */
diff --git a/sys/powerpc/include/proc.h b/sys/powerpc/include/proc.h
index c958fb76f0f..cbe6da9ff23 100644
--- a/sys/powerpc/include/proc.h
+++ b/sys/powerpc/include/proc.h
@@ -46,4 +46,6 @@ struct mdthread {
 struct mdproc {
 };
 
+#define KINFO_PROC_SIZE 768
+
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/sparc64/include/proc.h b/sys/sparc64/include/proc.h
index bfd1268e433..84eaed5a00e 100644
--- a/sys/sparc64/include/proc.h
+++ b/sys/sparc64/include/proc.h
@@ -51,4 +51,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
+#define KINFO_PROC_SIZE 1088
+
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/sun4v/include/proc.h b/sys/sun4v/include/proc.h
index bfd1268e433..84eaed5a00e 100644
--- a/sys/sun4v/include/proc.h
+++ b/sys/sun4v/include/proc.h
@@ -51,4 +51,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
+#define KINFO_PROC_SIZE 1088
+
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/sys/user.h b/sys/sys/user.h
index d5e46feb58f..34635e3dc18 100644
--- a/sys/sys/user.h
+++ b/sys/sys/user.h
@@ -87,34 +87,11 @@
 #define	KI_NSPARE_LONG	12
 #define	KI_NSPARE_PTR	7
 
-#ifdef __amd64__
-#define	KINFO_PROC_SIZE	1088
-#endif
-#ifdef __arm__
-#define	KINFO_PROC_SIZE	792
-#endif
-#ifdef __ia64__
-#define	KINFO_PROC_SIZE 1088
-#endif
-#ifdef __i386__
-#define	KINFO_PROC_SIZE	768
-#endif
-#ifdef __mips__
-#ifdef __mips_n64
-#define	KINFO_PROC_SIZE	1088
-#else
-#define	KINFO_PROC_SIZE	816
-#endif
-#endif
-#ifdef __powerpc__
-#define	KINFO_PROC_SIZE	768
-#endif
-#ifdef __sparc64__
-#define	KINFO_PROC_SIZE 1088
-#endif
+#ifndef _KERNEL
 #ifndef KINFO_PROC_SIZE
 #error "Unknown architecture"
 #endif
+#endif /* !_KERNEL */
 
 #define	WMESGLEN	8		/* size of returned wchan message */
 #define	LOCKNAMELEN	8		/* size of returned lock name */

From 24d9be57aebec43468602f26389427474fe2fdb9 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 24 Apr 2010 13:53:12 +0000
Subject: [PATCH 231/532] stat: Allow -f %Sf to display the file flags
 symbolically.

I have changed the patch slightly to show '-' if there are no flags just
like ls -ldo does.

PR:		124349
Submitted by:	Ighighi
MFC after:	1 week
---
 usr.bin/stat/stat.1 |  7 ++++++-
 usr.bin/stat/stat.c | 29 +++++++++++++++++++++++++++--
 2 files changed, 33 insertions(+), 3 deletions(-)

diff --git a/usr.bin/stat/stat.1 b/usr.bin/stat/stat.1
index 6856efae22f..8bbdda48453 100644
--- a/usr.bin/stat/stat.1
+++ b/usr.bin/stat/stat.1
@@ -29,7 +29,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd April 27, 2007
+.Dd April 24, 2010
 .Dt STAT 1
 .Os
 .Sh NAME
@@ -232,6 +232,11 @@ Display date in
 format.
 .It Cm dr
 Display actual device name.
+.It Cm f
+Display the flags of
+.Ar file
+as in
+.Nm ls Fl lTdo .
 .It Cm gu
 Display group or user name.
 .It Cm p
diff --git a/usr.bin/stat/stat.c b/usr.bin/stat/stat.c
index 83d389b8ee8..1ab10ea4b08 100644
--- a/usr.bin/stat/stat.c
+++ b/usr.bin/stat/stat.c
@@ -182,6 +182,9 @@ int	format1(const struct stat *,	/* stat info */
 	    char *, size_t,		/* a place to put the output */
 	    int, int, int, int,		/* the parsed format */
 	    int, int);
+#if HAVE_STRUCT_STAT_ST_FLAGS
+char   *xfflagstostr(unsigned long);
+#endif
 
 char *timefmt;
 int linkfail;
@@ -333,6 +336,25 @@ main(int argc, char *argv[])
 	return (am_readlink ? linkfail : errs);
 }
 
+#if HAVE_STRUCT_STAT_ST_FLAGS
+/*
+ * fflagstostr() wrapper that leaks only once
+ */
+char *
+xfflagstostr(unsigned long fflags)
+{
+	static char *str = NULL;
+
+	if (str != NULL)
+		free(str);
+
+	str = fflagstostr(fflags);
+	if (str == NULL)
+		err(1, "fflagstostr");
+	return (str);
+}
+#endif /* HAVE_STRUCT_STAT_ST_FLAGS */
+
 void
 usage(const char *synopsis)
 {
@@ -725,8 +747,11 @@ format1(const struct stat *st,
 	case SHOW_st_flags:
 		small = (sizeof(st->st_flags) == 4);
 		data = st->st_flags;
-		sdata = NULL;
-		formats = FMTF_DECIMAL | FMTF_OCTAL | FMTF_UNSIGNED | FMTF_HEX;
+		sdata = xfflagstostr(st->st_flags);
+		if (*sdata == '\0')
+			sdata = "-";
+		formats = FMTF_DECIMAL | FMTF_OCTAL | FMTF_UNSIGNED | FMTF_HEX |
+		    FMTF_STRING;
 		if (ofmt == 0)
 			ofmt = FMTF_UNSIGNED;
 		break;

From b901701726a3295ba267a0dec2ab3817621854a7 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Sat, 24 Apr 2010 16:27:43 +0000
Subject: [PATCH 232/532] Fix some sorting and whitespace nits.

---
 sys/conf/files.sparc64 | 8 ++++----
 sys/conf/files.sun4v   | 4 ++--
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/sys/conf/files.sparc64 b/sys/conf/files.sparc64
index a822a0ed817..63ee2c5be7f 100644
--- a/sys/conf/files.sparc64
+++ b/sys/conf/files.sparc64
@@ -22,7 +22,7 @@ ukbdmap.h			optional	ukbd_dflt_keymap	\
 	no-obj no-implicit-rule before-depend				\
 	clean		"ukbdmap.h"
 #
-crypto/blowfish/bf_enc.c	optional	crypto | ipsec 
+crypto/blowfish/bf_enc.c	optional	crypto | ipsec
 crypto/des/des_enc.c		optional	crypto | ipsec | netsmb
 dev/atkbdc/atkbd.c		optional	atkbd atkbdc
 dev/atkbdc/atkbd_atkbdc.c	optional	atkbd atkbdc
@@ -42,11 +42,11 @@ dev/kbd/kbd.c			optional	atkbd | sc | ukbd
 dev/le/if_le_lebuffer.c		optional	le sbus
 dev/le/if_le_ledma.c		optional	le sbus
 dev/le/lebuffer_sbus.c		optional	le sbus
-dev/ofw/ofw_if.m		standard
 dev/ofw/ofw_bus_if.m		standard
 dev/ofw/ofw_bus_subr.c		standard
-dev/ofw/ofw_standard.c		standard
 dev/ofw/ofw_console.c		optional	ofw_console
+dev/ofw/ofw_if.m		standard
+dev/ofw/ofw_standard.c		standard
 dev/ofw/openfirm.c		standard
 dev/ofw/openfirmio.c		standard
 dev/ofw/openpromio.c		standard
@@ -79,7 +79,7 @@ sparc64/pci/ofw_pcib.c		optional	pci
 sparc64/pci/ofw_pcib_subr.c	optional	pci
 sparc64/pci/ofw_pcibus.c	optional	pci
 sparc64/pci/psycho.c		optional	pci
-sparc64/pci/sbbc.c		optional	uart sbbc
+sparc64/pci/sbbc.c		optional	sbbc uart
 sparc64/pci/schizo.c		optional	pci
 sparc64/sbus/dma_sbus.c		optional	sbus
 sparc64/sbus/sbus.c		optional	sbus
diff --git a/sys/conf/files.sun4v b/sys/conf/files.sun4v
index 5964f5659fc..62e9e9be4a5 100644
--- a/sys/conf/files.sun4v
+++ b/sys/conf/files.sun4v
@@ -18,12 +18,12 @@ ukbdmap.h			optional	ukbd_dflt_keymap	\
 	clean		"ukbdmap.h"
 #
 #
-crypto/blowfish/bf_enc.c	optional	crypto | ipsec 
+crypto/blowfish/bf_enc.c	optional	crypto | ipsec
 crypto/des/des_enc.c		optional	crypto | ipsec | netsmb
 dev/ofw/ofw_bus_if.m		standard
-dev/ofw/ofw_if.m		standard
 dev/ofw/ofw_bus_subr.c		standard
 dev/ofw/ofw_console.c		optional	ofw_console
+dev/ofw/ofw_if.m		standard
 dev/ofw/ofw_standard.c		standard
 dev/ofw/openfirm.c		standard
 dev/ofw/openfirmio.c		standard

From 7b85f59183506a536c4a5974bb51c7e956334d62 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sat, 24 Apr 2010 17:32:52 +0000
Subject: [PATCH 233/532] Resurrect pmap_is_referenced() and use it in
 mincore().  Essentially, pmap_ts_referenced() is not always appropriate for
 checking whether or not pages have been referenced because it clears any
 reference bits that it encounters.  For example, in mincore(), clearing the
 reference bits has two negative consequences.  First, it throws off the
 activity count calculations performed by the page daemon.  Specifically, a
 page on which mincore() has called pmap_ts_referenced() looks less active to
 the page daemon than it should.  Consequently, the page could be deactivated
 prematurely by the page daemon.  Arguably, this problem could be fixed by
 having mincore() duplicate the activity count calculation on the page. 
 However, there is a second problem for which that is not a solution.  In
 order to clear a reference on a 4KB page, it may be necessary to demote a
 2/4MB page mapping.  Thus, a mincore() by one process can have the side
 effect of demoting a superpage mapping within another process!

---
 sys/amd64/amd64/pmap.c              | 48 +++++++++++++++++++++++++--
 sys/arm/arm/pmap.c                  | 14 ++++++++
 sys/i386/i386/pmap.c                | 50 +++++++++++++++++++++++++++--
 sys/i386/xen/pmap.c                 | 32 ++++++++++++++++--
 sys/ia64/ia64/pmap.c                | 35 ++++++++++++++++++--
 sys/mips/mips/pmap.c                | 18 +++++++++--
 sys/powerpc/aim/mmu_oea.c           | 11 +++++++
 sys/powerpc/aim/mmu_oea64.c         | 11 +++++++
 sys/powerpc/booke/pmap.c            | 29 +++++++++++++++++
 sys/powerpc/powerpc/mmu_if.m        | 14 ++++++++
 sys/powerpc/powerpc/pmap_dispatch.c |  8 +++++
 sys/sparc64/sparc64/pmap.c          | 21 ++++++++++++
 sys/sun4v/sun4v/pmap.c              | 11 +++++++
 sys/vm/pmap.h                       |  1 +
 sys/vm/vm_mmap.c                    |  4 +--
 15 files changed, 289 insertions(+), 18 deletions(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 4b3f0413191..4d1c9a946ce 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -236,6 +236,7 @@ static void pmap_fill_ptp(pt_entry_t *firstpte, pt_entry_t newpte);
 static void pmap_insert_pt_page(pmap_t pmap, vm_page_t mpte);
 static void pmap_invalidate_cache_range(vm_offset_t sva, vm_offset_t eva);
 static boolean_t pmap_is_modified_pvh(struct md_page *pvh);
+static boolean_t pmap_is_referenced_pvh(struct md_page *pvh);
 static void pmap_kenter_attr(vm_offset_t va, vm_paddr_t pa, int mode);
 static vm_page_t pmap_lookup_pt_page(pmap_t pmap, vm_offset_t va);
 static void pmap_pde_attr(pd_entry_t *pde, int cache_bits);
@@ -4177,6 +4178,49 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t addr)
 	return (rv);
 }
 
+/*
+ *	pmap_is_referenced:
+ *
+ *	Return whether or not the specified physical page was referenced
+ *	in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	if (m->flags & PG_FICTITIOUS)
+		return (FALSE);
+	if (pmap_is_referenced_pvh(&m->md))
+		return (TRUE);
+	return (pmap_is_referenced_pvh(pa_to_pvh(VM_PAGE_TO_PHYS(m))));
+}
+
+/*
+ * Returns TRUE if any of the given mappings were referenced and FALSE
+ * otherwise.  Both page and 2mpage mappings are supported.
+ */
+static boolean_t
+pmap_is_referenced_pvh(struct md_page *pvh)
+{
+	pv_entry_t pv;
+	pt_entry_t *pte;
+	pmap_t pmap;
+	boolean_t rv;
+
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	rv = FALSE;
+	TAILQ_FOREACH(pv, &pvh->pv_list, pv_list) {
+		pmap = PV_PMAP(pv);
+		PMAP_LOCK(pmap);
+		pte = pmap_pte(pmap, pv->pv_va);
+		rv = (*pte & (PG_A | PG_V)) == (PG_A | PG_V);
+		PMAP_UNLOCK(pmap);
+		if (rv)
+			break;
+	}
+	return (rv);
+}
+
 /*
  * Clear the write and modified bits in each of the given page's mappings.
  */
@@ -4893,10 +4937,8 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			 */
 			vm_page_lock_queues();
 			if ((m->flags & PG_REFERENCED) ||
-			    pmap_ts_referenced(m)) {
+			    pmap_is_referenced(m))
 				val |= MINCORE_REFERENCED_OTHER;
-				vm_page_flag_set(m, PG_REFERENCED);
-			}
 			vm_page_unlock_queues();
 		}
 	} 
diff --git a/sys/arm/arm/pmap.c b/sys/arm/arm/pmap.c
index 19b85d88a37..366b43fe3e1 100644
--- a/sys/arm/arm/pmap.c
+++ b/sys/arm/arm/pmap.c
@@ -4492,6 +4492,20 @@ pmap_clear_modify(vm_page_t m)
 }
 
 
+/*
+ *	pmap_is_referenced:
+ *
+ *	Return whether or not the specified physical page was referenced
+ *	in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	return ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) == 0 &&
+	    (m->md.pvh_attrs & PVF_REF) != 0);
+}
+
 /*
  *	pmap_clear_reference:
  *
diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c
index 7c74f00bb2c..f316657d757 100644
--- a/sys/i386/i386/pmap.c
+++ b/sys/i386/i386/pmap.c
@@ -296,6 +296,7 @@ static vm_page_t pmap_enter_quick_locked(pmap_t pmap, vm_offset_t va,
 static void pmap_insert_pt_page(pmap_t pmap, vm_page_t mpte);
 static void pmap_fill_ptp(pt_entry_t *firstpte, pt_entry_t newpte);
 static boolean_t pmap_is_modified_pvh(struct md_page *pvh);
+static boolean_t pmap_is_referenced_pvh(struct md_page *pvh);
 static void pmap_kenter_attr(vm_offset_t va, vm_paddr_t pa, int mode);
 static void pmap_kenter_pde(vm_offset_t va, pd_entry_t newpde);
 static vm_page_t pmap_lookup_pt_page(pmap_t pmap, vm_offset_t va);
@@ -4355,6 +4356,51 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t addr)
 	return (rv);
 }
 
+/*
+ *	pmap_is_referenced:
+ *
+ *	Return whether or not the specified physical page was referenced
+ *	in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	if (m->flags & PG_FICTITIOUS)
+		return (FALSE);
+	if (pmap_is_referenced_pvh(&m->md))
+		return (TRUE);
+	return (pmap_is_referenced_pvh(pa_to_pvh(VM_PAGE_TO_PHYS(m))));
+}
+
+/*
+ * Returns TRUE if any of the given mappings were referenced and FALSE
+ * otherwise.  Both page and 4mpage mappings are supported.
+ */
+static boolean_t
+pmap_is_referenced_pvh(struct md_page *pvh)
+{
+	pv_entry_t pv;
+	pt_entry_t *pte;
+	pmap_t pmap;
+	boolean_t rv;
+
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	rv = FALSE;
+	sched_pin();
+	TAILQ_FOREACH(pv, &pvh->pv_list, pv_list) {
+		pmap = PV_PMAP(pv);
+		PMAP_LOCK(pmap);
+		pte = pmap_pte_quick(pmap, pv->pv_va);
+		rv = (*pte & (PG_A | PG_V)) == (PG_A | PG_V);
+		PMAP_UNLOCK(pmap);
+		if (rv)
+			break;
+	}
+	sched_unpin();
+	return (rv);
+}
+
 /*
  * Clear the write and modified bits in each of the given page's mappings.
  */
@@ -4961,10 +5007,8 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			 */
 			vm_page_lock_queues();
 			if ((m->flags & PG_REFERENCED) ||
-			    pmap_ts_referenced(m)) {
+			    pmap_is_referenced(m))
 				val |= MINCORE_REFERENCED_OTHER;
-				vm_page_flag_set(m, PG_REFERENCED);
-			}
 			vm_page_unlock_queues();
 		}
 	} 
diff --git a/sys/i386/xen/pmap.c b/sys/i386/xen/pmap.c
index ae4d4aa2add..40e36be3dcc 100644
--- a/sys/i386/xen/pmap.c
+++ b/sys/i386/xen/pmap.c
@@ -3718,6 +3718,34 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t addr)
 	return (rv);
 }
 
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+	pv_entry_t pv;
+	pt_entry_t *pte;
+	pmap_t pmap;
+	boolean_t rv;
+
+	rv = FALSE;
+	if (m->flags & PG_FICTITIOUS)
+		return (rv);
+	sched_pin();
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
+		pmap = PV_PMAP(pv);
+		PMAP_LOCK(pmap);
+		pte = pmap_pte_quick(pmap, pv->pv_va);
+		rv = (*pte & (PG_A | PG_V)) == (PG_A | PG_V);
+		PMAP_UNLOCK(pmap);
+		if (rv)
+			break;
+	}
+	if (*PMAP1)
+		PT_SET_MA(PADDR1, 0);
+	sched_unpin();
+	return (rv);
+}
+
 void
 pmap_map_readonly(pmap_t pmap, vm_offset_t va, int len)
 {
@@ -4145,10 +4173,8 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			 */
 			vm_page_lock_queues();
 			if ((m->flags & PG_REFERENCED) ||
-			    pmap_ts_referenced(m)) {
+			    pmap_is_referenced(m))
 				val |= MINCORE_REFERENCED_OTHER;
-				vm_page_flag_set(m, PG_REFERENCED);
-			}
 			vm_page_unlock_queues();
 		}
 	} 
diff --git a/sys/ia64/ia64/pmap.c b/sys/ia64/ia64/pmap.c
index 6fe4cdf987e..a7c47eff842 100644
--- a/sys/ia64/ia64/pmap.c
+++ b/sys/ia64/ia64/pmap.c
@@ -2022,6 +2022,37 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t addr)
 	return (TRUE);
 }
 
+/*
+ *	pmap_is_referenced:
+ *
+ *	Return whether or not the specified physical page was referenced
+ *	in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+	struct ia64_lpte *pte;
+	pmap_t oldpmap;
+	pv_entry_t pv;
+	boolean_t rv;
+
+	rv = FALSE;
+	if (m->flags & PG_FICTITIOUS)
+		return (rv);
+	TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
+		PMAP_LOCK(pv->pv_pmap);
+		oldpmap = pmap_switch(pv->pv_pmap);
+		pte = pmap_find_vhpt(pv->pv_va);
+		pmap_switch(oldpmap);
+		KASSERT(pte != NULL, ("pte"));
+		rv = pmap_accessed(pte) ? TRUE : FALSE;
+		PMAP_UNLOCK(pv->pv_pmap);
+		if (rv)
+			break;
+	}
+	return (rv);
+}
+
 /*
  *	Clear the modify bits on the specified physical page.
  */
@@ -2197,10 +2228,8 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			 * Referenced by someone
 			 */
 			vm_page_lock_queues();
-			if (pmap_ts_referenced(m)) {
+			if (pmap_is_referenced(m))
 				val |= MINCORE_REFERENCED_OTHER;
-				vm_page_flag_set(m, PG_REFERENCED);
-			}
 			vm_page_unlock_queues();
 		}
 	} 
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 9c28f23804f..3634a27e3ee 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -2625,6 +2625,20 @@ pmap_clear_modify(vm_page_t m)
 	}
 }
 
+/*
+ *	pmap_is_referenced:
+ *
+ *	Return whether or not the specified physical page was referenced
+ *	in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	return ((m->flags & PG_FICTITIOUS) == 0 &&
+	    (m->md.pv_flags & PV_TABLE_REF) != 0);
+}
+
 /*
  *	pmap_clear_reference:
  *
@@ -2750,10 +2764,8 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 		 * Referenced by us or someone
 		 */
 		vm_page_lock_queues();
-		if ((m->flags & PG_REFERENCED) || pmap_ts_referenced(m)) {
+		if ((m->flags & PG_REFERENCED) || pmap_is_referenced(m))
 			val |= MINCORE_REFERENCED | MINCORE_REFERENCED_OTHER;
-			vm_page_flag_set(m, PG_REFERENCED);
-		}
 		vm_page_unlock_queues();
 	}
 	return val;
diff --git a/sys/powerpc/aim/mmu_oea.c b/sys/powerpc/aim/mmu_oea.c
index 8357929e4f2..24e7b4e1d40 100644
--- a/sys/powerpc/aim/mmu_oea.c
+++ b/sys/powerpc/aim/mmu_oea.c
@@ -305,6 +305,7 @@ vm_paddr_t moea_extract(mmu_t, pmap_t, vm_offset_t);
 vm_page_t moea_extract_and_hold(mmu_t, pmap_t, vm_offset_t, vm_prot_t);
 void moea_init(mmu_t);
 boolean_t moea_is_modified(mmu_t, vm_page_t);
+boolean_t moea_is_referenced(mmu_t, vm_page_t);
 boolean_t moea_ts_referenced(mmu_t, vm_page_t);
 vm_offset_t moea_map(mmu_t, vm_offset_t *, vm_offset_t, vm_offset_t, int);
 boolean_t moea_page_exists_quick(mmu_t, pmap_t, vm_page_t);
@@ -344,6 +345,7 @@ static mmu_method_t moea_methods[] = {
 	MMUMETHOD(mmu_extract_and_hold,	moea_extract_and_hold),
 	MMUMETHOD(mmu_init,		moea_init),
 	MMUMETHOD(mmu_is_modified,	moea_is_modified),
+	MMUMETHOD(mmu_is_referenced,	moea_is_referenced),
 	MMUMETHOD(mmu_ts_referenced,	moea_ts_referenced),
 	MMUMETHOD(mmu_map,     		moea_map),
 	MMUMETHOD(mmu_page_exists_quick,moea_page_exists_quick),
@@ -1268,6 +1270,15 @@ moea_init(mmu_t mmu)
 	moea_initialized = TRUE;
 }
 
+boolean_t
+moea_is_referenced(mmu_t mmu, vm_page_t m)
+{
+
+	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
+		return (FALSE);
+	return (moea_query_bit(m, PTE_REF));
+}
+
 boolean_t
 moea_is_modified(mmu_t mmu, vm_page_t m)
 {
diff --git a/sys/powerpc/aim/mmu_oea64.c b/sys/powerpc/aim/mmu_oea64.c
index 97fff631277..2571587733a 100644
--- a/sys/powerpc/aim/mmu_oea64.c
+++ b/sys/powerpc/aim/mmu_oea64.c
@@ -379,6 +379,7 @@ vm_paddr_t moea64_extract(mmu_t, pmap_t, vm_offset_t);
 vm_page_t moea64_extract_and_hold(mmu_t, pmap_t, vm_offset_t, vm_prot_t);
 void moea64_init(mmu_t);
 boolean_t moea64_is_modified(mmu_t, vm_page_t);
+boolean_t moea64_is_referenced(mmu_t, vm_page_t);
 boolean_t moea64_ts_referenced(mmu_t, vm_page_t);
 vm_offset_t moea64_map(mmu_t, vm_offset_t *, vm_offset_t, vm_offset_t, int);
 boolean_t moea64_page_exists_quick(mmu_t, pmap_t, vm_page_t);
@@ -416,6 +417,7 @@ static mmu_method_t moea64_bridge_methods[] = {
 	MMUMETHOD(mmu_extract_and_hold,	moea64_extract_and_hold),
 	MMUMETHOD(mmu_init,		moea64_init),
 	MMUMETHOD(mmu_is_modified,	moea64_is_modified),
+	MMUMETHOD(mmu_is_referenced,	moea64_is_referenced),
 	MMUMETHOD(mmu_ts_referenced,	moea64_ts_referenced),
 	MMUMETHOD(mmu_map,     		moea64_map),
 	MMUMETHOD(mmu_page_exists_quick,moea64_page_exists_quick),
@@ -1462,6 +1464,15 @@ moea64_init(mmu_t mmu)
 	moea64_initialized = TRUE;
 }
 
+boolean_t
+moea64_is_referenced(mmu_t mmu, vm_page_t m)
+{
+
+	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
+		return (FALSE);
+	return (moea64_query_bit(m, PTE_REF));
+}
+
 boolean_t
 moea64_is_modified(mmu_t mmu, vm_page_t m)
 {
diff --git a/sys/powerpc/booke/pmap.c b/sys/powerpc/booke/pmap.c
index 13e637ce603..549eaaa94bb 100644
--- a/sys/powerpc/booke/pmap.c
+++ b/sys/powerpc/booke/pmap.c
@@ -288,6 +288,7 @@ static vm_page_t	mmu_booke_extract_and_hold(mmu_t, pmap_t, vm_offset_t,
 static void		mmu_booke_init(mmu_t);
 static boolean_t	mmu_booke_is_modified(mmu_t, vm_page_t);
 static boolean_t	mmu_booke_is_prefaultable(mmu_t, pmap_t, vm_offset_t);
+static boolean_t	mmu_booke_is_referenced(mmu_t, vm_page_t);
 static boolean_t	mmu_booke_ts_referenced(mmu_t, vm_page_t);
 static vm_offset_t	mmu_booke_map(mmu_t, vm_offset_t *, vm_offset_t, vm_offset_t,
     int);
@@ -342,6 +343,7 @@ static mmu_method_t mmu_booke_methods[] = {
 	MMUMETHOD(mmu_init,		mmu_booke_init),
 	MMUMETHOD(mmu_is_modified,	mmu_booke_is_modified),
 	MMUMETHOD(mmu_is_prefaultable,	mmu_booke_is_prefaultable),
+	MMUMETHOD(mmu_is_referenced,	mmu_booke_is_referenced),
 	MMUMETHOD(mmu_ts_referenced,	mmu_booke_ts_referenced),
 	MMUMETHOD(mmu_map,		mmu_booke_map),
 	MMUMETHOD(mmu_mincore,		mmu_booke_mincore),
@@ -2180,6 +2182,33 @@ mmu_booke_is_prefaultable(mmu_t mmu, pmap_t pmap, vm_offset_t addr)
 	return (FALSE);
 }
 
+/*
+ * Return whether or not the specified physical page was referenced
+ * in any physical maps.
+ */
+static boolean_t
+mmu_booke_is_referenced(mmu_t mmu, vm_page_t m)
+{
+	pte_t *pte;
+	pv_entry_t pv;
+	boolean_t rv;
+
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	rv = FALSE;
+	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
+		return (rv);
+	TAILQ_FOREACH(pv, &m->md.pv_list, pv_link) {
+		PMAP_LOCK(pv->pv_pmap);
+		if ((pte = pte_find(mmu, pv->pv_pmap, pv->pv_va)) != NULL &&
+		    PTE_ISVALID(pte))
+			rv = PTE_ISREFERENCED(pte) ? TRUE : FALSE;
+		PMAP_UNLOCK(pv->pv_pmap);
+		if (rv)
+			break;
+	}
+	return (rv);
+}
+
 /*
  * Clear the modify bits on the specified physical page.
  */
diff --git a/sys/powerpc/powerpc/mmu_if.m b/sys/powerpc/powerpc/mmu_if.m
index 5b8ba14d689..a87e5d8c89d 100644
--- a/sys/powerpc/powerpc/mmu_if.m
+++ b/sys/powerpc/powerpc/mmu_if.m
@@ -345,6 +345,20 @@ METHOD boolean_t is_prefaultable {
 } DEFAULT mmu_null_is_prefaultable;
 
 
+/**
+ * @brief Return whether or not the specified physical page was referenced
+ * in any physical maps.
+ *
+ * @params _pg		physical page
+ *
+ * @retval boolean	TRUE if page has been referenced
+ */
+METHOD boolean_t is_referenced {
+	mmu_t		_mmu;
+	vm_page_t	_pg;
+};
+
+
 /**
  * @brief Return a count of referenced bits for a page, clearing those bits.
  * Not all referenced bits need to be cleared, but it is necessary that 0
diff --git a/sys/powerpc/powerpc/pmap_dispatch.c b/sys/powerpc/powerpc/pmap_dispatch.c
index 2b45e17c6f3..c16360f20f1 100644
--- a/sys/powerpc/powerpc/pmap_dispatch.c
+++ b/sys/powerpc/powerpc/pmap_dispatch.c
@@ -194,6 +194,14 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t va)
 	return (MMU_IS_PREFAULTABLE(mmu_obj, pmap, va));
 }
 
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	CTR2(KTR_PMAP, "%s(%p)", __func__, m);
+	return (MMU_IS_REFERENCED(mmu_obj, m));
+}
+
 boolean_t
 pmap_ts_referenced(vm_page_t m)
 {
diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index e46455915c2..d97900c4afd 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -1917,6 +1917,27 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t addr)
 	return (FALSE);
 }
 
+/*
+ * Return whether or not the specified physical page was referenced
+ * in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+	struct tte *tp;
+
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
+		return (FALSE);
+	TAILQ_FOREACH(tp, &m->md.tte_list, tte_link) {
+		if ((tp->tte_data & TD_PV) == 0)
+			continue;
+		if ((tp->tte_data & TD_REF) != 0)
+			return (TRUE);
+	}
+	return (FALSE);
+}
+
 void
 pmap_clear_modify(vm_page_t m)
 {
diff --git a/sys/sun4v/sun4v/pmap.c b/sys/sun4v/sun4v/pmap.c
index d3b8c79467e..2633b8eb76f 100644
--- a/sys/sun4v/sun4v/pmap.c
+++ b/sys/sun4v/sun4v/pmap.c
@@ -1591,6 +1591,17 @@ pmap_is_prefaultable(pmap_t pmap, vm_offset_t va)
 	return (tte_hash_lookup(pmap->pm_hash, va) == 0);
 }
 
+/*
+ * Return whether or not the specified physical page was referenced
+ * in any physical maps.
+ */
+boolean_t
+pmap_is_referenced(vm_page_t m)
+{
+
+	return (tte_get_phys_bit(m, VTD_REF));
+}
+
 /*
  * Extract the physical page address associated with the given kernel virtual
  * address.
diff --git a/sys/vm/pmap.h b/sys/vm/pmap.h
index 51a3bbe4c7c..e4d8e81ec1b 100644
--- a/sys/vm/pmap.h
+++ b/sys/vm/pmap.h
@@ -119,6 +119,7 @@ void		 pmap_growkernel(vm_offset_t);
 void		 pmap_init(void);
 boolean_t	 pmap_is_modified(vm_page_t m);
 boolean_t	 pmap_is_prefaultable(pmap_t pmap, vm_offset_t va);
+boolean_t	 pmap_is_referenced(vm_page_t m);
 boolean_t	 pmap_ts_referenced(vm_page_t m);
 vm_offset_t	 pmap_map(vm_offset_t *, vm_paddr_t, vm_paddr_t, int);
 void		 pmap_object_init_pt(pmap_t pmap, vm_offset_t addr,
diff --git a/sys/vm/vm_mmap.c b/sys/vm/vm_mmap.c
index a47cd6a0f88..88ed3d5655b 100644
--- a/sys/vm/vm_mmap.c
+++ b/sys/vm/vm_mmap.c
@@ -871,10 +871,8 @@ RestartScan:
 						pmap_is_modified(m))
 						mincoreinfo |= MINCORE_MODIFIED_OTHER;
 					if ((m->flags & PG_REFERENCED) ||
-						pmap_ts_referenced(m)) {
-						vm_page_flag_set(m, PG_REFERENCED);
+					    pmap_is_referenced(m))
 						mincoreinfo |= MINCORE_REFERENCED_OTHER;
-					}
 					vm_page_unlock_queues();
 				}
 				VM_OBJECT_UNLOCK(current->object.vm_object);

From 1f4aef21da45dc3100a194f51c1d0e4ddbfb8af8 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Sat, 24 Apr 2010 18:13:12 +0000
Subject: [PATCH 234/532] remove vm obect tracker that slipped in from prior
 work.

MFC after:	2 weeks
---
 sys/dev/drm/drmP.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/drm/drmP.h b/sys/dev/drm/drmP.h
index 3699460a706..af50893163a 100644
--- a/sys/dev/drm/drmP.h
+++ b/sys/dev/drm/drmP.h
@@ -481,7 +481,6 @@ typedef struct drm_sg_mem {
 	vm_offset_t vaddr;
 	vm_paddr_t *busaddr;
 	vm_pindex_t pages;
-	vm_object_t obj;
 } drm_sg_mem_t;
 
 #define DRM_MAP_HANDLE_BITS	(sizeof(void *) == 4 ? 4 : 24)

From 81b47f28bcf0a8c3e91b5439e778833265f80776 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Sat, 24 Apr 2010 21:17:07 +0000
Subject: [PATCH 235/532] apply style(9) changes applied to head_page_lock_2

requested by: kib@
---
 sys/amd64/amd64/pmap.c | 34 +++++++++++++++++-----------------
 1 file changed, 17 insertions(+), 17 deletions(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 4d1c9a946ce..9c7d8b088a9 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -284,7 +284,7 @@ pmap_kmem_choose(vm_offset_t addr)
 	vm_offset_t newaddr = addr;
 
 	newaddr = (addr + (NBPDR - 1)) & ~(NBPDR - 1);
-	return newaddr;
+	return (newaddr);
 }
 
 /********************/
@@ -295,7 +295,7 @@ pmap_kmem_choose(vm_offset_t addr)
 static __inline vm_pindex_t
 pmap_pde_pindex(vm_offset_t va)
 {
-	return va >> PDRSHIFT;
+	return (va >> PDRSHIFT);
 }
 
 
@@ -354,7 +354,7 @@ pmap_pdpe(pmap_t pmap, vm_offset_t va)
 
 	pml4e = pmap_pml4e(pmap, va);
 	if ((*pml4e & PG_V) == 0)
-		return NULL;
+		return (NULL);
 	return (pmap_pml4e_to_pdpe(pml4e, va));
 }
 
@@ -376,7 +376,7 @@ pmap_pde(pmap_t pmap, vm_offset_t va)
 
 	pdpe = pmap_pdpe(pmap, va);
 	if (pdpe == NULL || (*pdpe & PG_V) == 0)
-		 return NULL;
+		return (NULL);
 	return (pmap_pdpe_to_pde(pdpe, va));
 }
 
@@ -398,7 +398,7 @@ pmap_pte(pmap_t pmap, vm_offset_t va)
 
 	pde = pmap_pde(pmap, va);
 	if (pde == NULL || (*pde & PG_V) == 0)
-		return NULL;
+		return (NULL);
 	if ((*pde & PG_PS) != 0)	/* compat with i386 pmap_pte() */
 		return ((pt_entry_t *)pde);
 	return (pmap_pde_to_pte(pde, va));
@@ -1237,7 +1237,7 @@ pmap_kextract(vm_offset_t va)
 			pa = (pa & PG_FRAME) | (va & PAGE_MASK);
 		}
 	}
-	return pa;
+	return (pa);
 }
 
 /***************************************************
@@ -1467,9 +1467,9 @@ pmap_unwire_pte_hold(pmap_t pmap, vm_offset_t va, vm_page_t m, vm_page_t *free)
 
 	--m->wire_count;
 	if (m->wire_count == 0)
-		return _pmap_unwire_pte_hold(pmap, va, m, free);
+		return (_pmap_unwire_pte_hold(pmap, va, m, free));
 	else
-		return 0;
+		return (0);
 }
 
 static int 
@@ -1525,7 +1525,7 @@ _pmap_unwire_pte_hold(pmap_t pmap, vm_offset_t va, vm_page_t m,
 	 */
 	pmap_add_delayed_free_list(m, free, TRUE);
 	
-	return 1;
+	return (1);
 }
 
 /*
@@ -1538,10 +1538,10 @@ pmap_unuse_pt(pmap_t pmap, vm_offset_t va, pd_entry_t ptepde, vm_page_t *free)
 	vm_page_t mpte;
 
 	if (va >= VM_MAXUSER_ADDRESS)
-		return 0;
+		return (0);
 	KASSERT(ptepde != 0, ("pmap_unuse_pt: ptepde != 0"));
 	mpte = PHYS_TO_VM_PAGE(ptepde & PG_FRAME);
-	return pmap_unwire_pte_hold(pmap, va, mpte, free);
+	return (pmap_unwire_pte_hold(pmap, va, mpte, free));
 }
 
 void
@@ -1732,7 +1732,7 @@ _pmap_allocpte(pmap_t pmap, vm_pindex_t ptepindex, int flags)
 
 	pmap->pm_stats.resident_count++;
 
-	return m;
+	return (m);
 }
 
 static vm_page_t
@@ -3510,7 +3510,7 @@ pmap_enter_quick_locked(pmap_t pmap, vm_offset_t va, vm_page_t m,
 		pte_store(pte, pa | PG_V | PG_U);
 	else
 		pte_store(pte, pa | PG_V | PG_U | PG_MANAGED);
-	return mpte;
+	return (mpte);
 }
 
 /*
@@ -3887,12 +3887,12 @@ pmap_page_exists_quick(pmap_t pmap, vm_page_t m)
 	int loops = 0;
 
 	if (m->flags & PG_FICTITIOUS)
-		return FALSE;
+		return (FALSE);
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
 		if (PV_PMAP(pv) == pmap) {
-			return TRUE;
+			return (TRUE);
 		}
 		loops++;
 		if (loops >= 16)
@@ -4908,7 +4908,7 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 	if (pte != 0) {
 		val |= MINCORE_INCORE;
 		if ((pte & PG_MANAGED) == 0)
-			return val;
+			return (val);
 
 		m = PHYS_TO_VM_PAGE(pa);
 
@@ -4942,7 +4942,7 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			vm_page_unlock_queues();
 		}
 	} 
-	return val;
+	return (val);
 }
 
 void

From c5cc832f329433e11b58261161c13d4000ba3a24 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Sat, 24 Apr 2010 21:36:52 +0000
Subject: [PATCH 236/532] - fix style issues on i386 as well

requested by: alc@
---
 sys/i386/i386/pmap.c | 32 ++++++++++++++++----------------
 1 file changed, 16 insertions(+), 16 deletions(-)

diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c
index f316657d757..1ce6454ffbb 100644
--- a/sys/i386/i386/pmap.c
+++ b/sys/i386/i386/pmap.c
@@ -1607,9 +1607,9 @@ pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m, vm_page_t *free)
 
 	--m->wire_count;
 	if (m->wire_count == 0)
-		return _pmap_unwire_pte_hold(pmap, m, free);
+		return (_pmap_unwire_pte_hold(pmap, m, free));
 	else
-		return 0;
+		return (0);
 }
 
 static int 
@@ -1643,7 +1643,7 @@ _pmap_unwire_pte_hold(pmap_t pmap, vm_page_t m, vm_page_t *free)
 	 */
 	pmap_add_delayed_free_list(m, free, TRUE);
 
-	return 1;
+	return (1);
 }
 
 /*
@@ -1657,10 +1657,10 @@ pmap_unuse_pt(pmap_t pmap, vm_offset_t va, vm_page_t *free)
 	vm_page_t mpte;
 
 	if (va >= VM_MAXUSER_ADDRESS)
-		return 0;
+		return (0);
 	ptepde = *pmap_pde(pmap, va);
 	mpte = PHYS_TO_VM_PAGE(ptepde & PG_FRAME);
-	return pmap_unwire_pte_hold(pmap, mpte, free);
+	return (pmap_unwire_pte_hold(pmap, mpte, free));
 }
 
 void
@@ -1811,7 +1811,7 @@ _pmap_allocpte(pmap_t pmap, unsigned ptepindex, int flags)
 	pmap->pm_pdir[ptepindex] =
 		(pd_entry_t) (ptepa | PG_U | PG_RW | PG_V | PG_A | PG_M);
 
-	return m;
+	return (m);
 }
 
 static vm_page_t
@@ -2009,7 +2009,7 @@ kvm_size(SYSCTL_HANDLER_ARGS)
 {
 	unsigned long ksize = VM_MAX_KERNEL_ADDRESS - KERNBASE;
 
-	return sysctl_handle_long(oidp, &ksize, 0, req);
+	return (sysctl_handle_long(oidp, &ksize, 0, req));
 }
 SYSCTL_PROC(_vm, OID_AUTO, kvm_size, CTLTYPE_LONG|CTLFLAG_RD, 
     0, 0, kvm_size, "IU", "Size of KVM");
@@ -2019,7 +2019,7 @@ kvm_free(SYSCTL_HANDLER_ARGS)
 {
 	unsigned long kfree = VM_MAX_KERNEL_ADDRESS - kernel_vm_end;
 
-	return sysctl_handle_long(oidp, &kfree, 0, req);
+	return (sysctl_handle_long(oidp, &kfree, 0, req));
 }
 SYSCTL_PROC(_vm, OID_AUTO, kvm_free, CTLTYPE_LONG|CTLFLAG_RD, 
     0, 0, kvm_free, "IU", "Amount of KVM free");
@@ -2083,7 +2083,7 @@ static __inline struct pv_chunk *
 pv_to_chunk(pv_entry_t pv)
 {
 
-	return (struct pv_chunk *)((uintptr_t)pv & ~(uintptr_t)PAGE_MASK);
+	return ((struct pv_chunk *)((uintptr_t)pv & ~(uintptr_t)PAGE_MASK));
 }
 
 #define PV_PMAP(pv) (pv_to_chunk(pv)->pc_pmap)
@@ -3662,7 +3662,7 @@ pmap_enter_quick_locked(pmap_t pmap, vm_offset_t va, vm_page_t m,
 		pte_store(pte, pa | PG_V | PG_U);
 	else
 		pte_store(pte, pa | PG_V | PG_U | PG_MANAGED);
-	return mpte;
+	return (mpte);
 }
 
 /*
@@ -4059,12 +4059,12 @@ pmap_page_exists_quick(pmap_t pmap, vm_page_t m)
 	int loops = 0;
 
 	if (m->flags & PG_FICTITIOUS)
-		return FALSE;
+		return (FALSE);
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	TAILQ_FOREACH(pv, &m->md.pv_list, pv_list) {
 		if (PV_PMAP(pv) == pmap) {
-			return TRUE;
+			return (TRUE);
 		}
 		loops++;
 		if (loops >= 16)
@@ -4978,7 +4978,7 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 	if (pte != 0) {
 		val |= MINCORE_INCORE;
 		if ((pte & PG_MANAGED) == 0)
-			return val;
+			return (val);
 
 		m = PHYS_TO_VM_PAGE(pa);
 
@@ -5012,7 +5012,7 @@ pmap_mincore(pmap_t pmap, vm_offset_t addr)
 			vm_page_unlock_queues();
 		}
 	} 
-	return val;
+	return (val);
 }
 
 void
@@ -5107,7 +5107,7 @@ pmap_pid_dump(int pid)
 								printf("\n");
 							}
 							sx_sunlock(&allproc_lock);
-							return npte;
+							return (npte);
 						}
 						pte = pmap_pte(pmap, va);
 						if (pte && pmap_pte_v(pte)) {
@@ -5132,7 +5132,7 @@ pmap_pid_dump(int pid)
 		}
 	}
 	sx_sunlock(&allproc_lock);
-	return npte;
+	return (npte);
 }
 #endif
 

From 2d61a2a4777c58c43f04724c2ec5766c3e55a639 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 24 Apr 2010 22:16:01 +0000
Subject: [PATCH 237/532] builtin(1): Mention [ sh builtin.

MFC after:	1 week
---
 share/man/man1/builtin.1 | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/share/man/man1/builtin.1 b/share/man/man1/builtin.1
index 0a93c21ecd7..79b2ceca60b 100644
--- a/share/man/man1/builtin.1
+++ b/share/man/man1/builtin.1
@@ -26,7 +26,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd October 14, 2006
+.Dd April 25, 2010
 .Dt BUILTIN 1
 .Os
 .Sh NAME
@@ -36,6 +36,7 @@
 .Nm \&. ,
 .Nm \&: ,
 .Nm @ ,
+.Nm \&[ ,
 .Nm { ,
 .Nm } ,
 .Nm alias ,
@@ -200,6 +201,7 @@ but are implemented as scripts using a builtin command of the same name.
 .It Ic . Ta \&No Ta \&No Ta Yes
 .It Ic : Ta \&No Ta Yes Ta Yes
 .It Ic @ Ta \&No Ta Yes Ta Yes
+.It Ic \&[ Ta Yes Ta \&No Ta Yes
 .It Ic { Ta \&No Ta \&No Ta Yes
 .It Ic } Ta \&No Ta \&No Ta Yes
 .It Ic alias Ta No** Ta Yes Ta Yes

From bf59faa38570e2cc8034e0c1d9e155c943df2649 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 24 Apr 2010 22:24:26 +0000
Subject: [PATCH 238/532] builtin(1): Add missing escaping for !, . and : in
 the table.

This caused these commands to look differently (not bold) from the other
commands in the table (bold).

MFC after:	1 week
---
 share/man/man1/builtin.1 | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/share/man/man1/builtin.1 b/share/man/man1/builtin.1
index 79b2ceca60b..12ab0a83841 100644
--- a/share/man/man1/builtin.1
+++ b/share/man/man1/builtin.1
@@ -196,10 +196,10 @@ but are implemented as scripts using a builtin command of the same name.
 .It Xo
 .Em "Command	External" Ta Xr csh 1 Ta Xr sh 1
 .Xc
-.It Ic ! Ta \&No Ta \&No Ta Yes
+.It Ic \&! Ta \&No Ta \&No Ta Yes
 .It Ic % Ta \&No Ta Yes Ta \&No
-.It Ic . Ta \&No Ta \&No Ta Yes
-.It Ic : Ta \&No Ta Yes Ta Yes
+.It Ic \&. Ta \&No Ta \&No Ta Yes
+.It Ic \&: Ta \&No Ta Yes Ta Yes
 .It Ic @ Ta \&No Ta Yes Ta Yes
 .It Ic \&[ Ta Yes Ta \&No Ta Yes
 .It Ic { Ta \&No Ta \&No Ta Yes

From 23f929dfe8361d0dbedf909df6fc8e24c1f5caec Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Sat, 24 Apr 2010 22:52:14 +0000
Subject: [PATCH 239/532] An NFSv4 server will reply NFSERR_GRACE for
 non-recovery RPCs during the grace period after startup. This grace period
 must be at least the lease duration, which is typically 1-2 minutes. It seems
 prudent for the experimental NFS client to wait a few seconds before retrying
 such an RPC, so that the server isn't flooded with non-recovery RPCs during
 recovery. This patch adds an argument to nfs_catnap() to implement a 5 second
 delay for this case.

MFC after:	1 week
---
 sys/fs/nfs/nfs_commonkrpc.c      |  2 +-
 sys/fs/nfs/nfs_commonport.c      | 10 +++++++---
 sys/fs/nfs/nfs_var.h             |  2 +-
 sys/fs/nfs/nfsport.h             |  8 ++++----
 sys/fs/nfsclient/nfs_clrpcops.c  | 26 ++++++++++++++------------
 sys/fs/nfsclient/nfs_clstate.c   | 24 +++++++++++++-----------
 sys/fs/nfsclient/nfs_clvfsops.c  |  6 +++---
 sys/fs/nfsclient/nfs_clvnops.c   |  3 ++-
 sys/fs/nfsserver/nfs_nfsdstate.c |  2 +-
 9 files changed, 46 insertions(+), 37 deletions(-)

diff --git a/sys/fs/nfs/nfs_commonkrpc.c b/sys/fs/nfs/nfs_commonkrpc.c
index 8b6ada3865f..384bdb417de 100644
--- a/sys/fs/nfs/nfs_commonkrpc.c
+++ b/sys/fs/nfs/nfs_commonkrpc.c
@@ -650,7 +650,7 @@ tryagain:
 					trylater_delay = NFS_TRYLATERDEL;
 				waituntil = NFSD_MONOSEC + trylater_delay;
 				while (NFSD_MONOSEC < waituntil)
-					(void) nfs_catnap(PZERO, "nfstry");
+					(void) nfs_catnap(PZERO, 0, "nfstry");
 				trylater_delay *= 2;
 				goto tryagain;
 			}
diff --git a/sys/fs/nfs/nfs_commonport.c b/sys/fs/nfs/nfs_commonport.c
index 2849163837d..26765668d24 100644
--- a/sys/fs/nfs/nfs_commonport.c
+++ b/sys/fs/nfs/nfs_commonport.c
@@ -345,17 +345,21 @@ newnfs_timer(void *arg)
 
 
 /*
- * sleep for a short period of time.
+ * Sleep for a short period of time unless errval == NFSERR_GRACE, where
+ * the sleep should be for 5 seconds.
  * Since lbolt doesn't exist in FreeBSD-CURRENT, just use a timeout on
  * an event that never gets a wakeup. Only return EINTR or 0.
  */
 int
-nfs_catnap(int prio, const char *wmesg)
+nfs_catnap(int prio, int errval, const char *wmesg)
 {
 	static int non_event;
 	int ret;
 
-	ret = tsleep(&non_event, prio, wmesg, 1);
+	if (errval == NFSERR_GRACE)
+		ret = tsleep(&non_event, prio, wmesg, 5 * hz);
+	else
+		ret = tsleep(&non_event, prio, wmesg, 1);
 	if (ret != EINTR)
 		ret = 0;
 	return (ret);
diff --git a/sys/fs/nfs/nfs_var.h b/sys/fs/nfs/nfs_var.h
index 5388e13f549..d6ecda21731 100644
--- a/sys/fs/nfs/nfs_var.h
+++ b/sys/fs/nfs/nfs_var.h
@@ -322,7 +322,7 @@ int nfsvno_v4rootexport(struct nfsrv_descript *);
 void newnfs_portinit(void);
 struct ucred *newnfs_getcred(void);
 void newnfs_setroot(struct ucred *);
-int nfs_catnap(int, const char *);
+int nfs_catnap(int, int, const char *);
 struct nfsreferral *nfsv4root_getreferral(vnode_t, vnode_t, u_int32_t);
 int nfsrv_atroot(vnode_t, long *);
 void newnfs_timer(void *);
diff --git a/sys/fs/nfs/nfsport.h b/sys/fs/nfs/nfsport.h
index 3d370040b44..0fa2e14602d 100644
--- a/sys/fs/nfs/nfsport.h
+++ b/sys/fs/nfs/nfsport.h
@@ -143,21 +143,21 @@
 #define	NFSMGET(m)	do { 					\
 		MGET((m), M_TRYWAIT, MT_DATA); 			\
 		while ((m) == NULL ) { 				\
-			(void) nfs_catnap(PZERO, "nfsmget");	\
+			(void) nfs_catnap(PZERO, 0, "nfsmget");	\
 			MGET((m), M_TRYWAIT, MT_DATA); 		\
 		} 						\
 	} while (0)
 #define	NFSMGETHDR(m)	do { 					\
 		MGETHDR((m), M_TRYWAIT, MT_DATA);		\
 		while ((m) == NULL ) { 				\
-			(void) nfs_catnap(PZERO, "nfsmget");	\
+			(void) nfs_catnap(PZERO, 0, "nfsmget");	\
 			MGETHDR((m), M_TRYWAIT, MT_DATA); 	\
 		} 						\
 	} while (0)
 #define	NFSMCLGET(m, w)	do { 					\
 		MGET((m), M_TRYWAIT, MT_DATA); 			\
 		while ((m) == NULL ) { 				\
-			(void) nfs_catnap(PZERO, "nfsmget");	\
+			(void) nfs_catnap(PZERO, 0, "nfsmget");	\
 			MGET((m), M_TRYWAIT, MT_DATA); 		\
 		} 						\
 		MCLGET((m), (w));				\
@@ -165,7 +165,7 @@
 #define	NFSMCLGETHDR(m, w) do { 				\
 		MGETHDR((m), M_TRYWAIT, MT_DATA);		\
 		while ((m) == NULL ) { 				\
-			(void) nfs_catnap(PZERO, "nfsmget");	\
+			(void) nfs_catnap(PZERO, 0, "nfsmget");	\
 			MGETHDR((m), M_TRYWAIT, MT_DATA); 	\
 		} 						\
 	} while (0)
diff --git a/sys/fs/nfsclient/nfs_clrpcops.c b/sys/fs/nfsclient/nfs_clrpcops.c
index 63c15c10fd9..a84a3014abe 100644
--- a/sys/fs/nfsclient/nfs_clrpcops.c
+++ b/sys/fs/nfsclient/nfs_clrpcops.c
@@ -298,7 +298,7 @@ else printf(" fhl=0\n");
 	    nfscl_openrelease(op, error, newone);
 	    if (error == NFSERR_GRACE || error == NFSERR_STALECLIENTID ||
 		error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY) {
-		(void) nfs_catnap(PZERO, "nfs_open");
+		(void) nfs_catnap(PZERO, error, "nfs_open");
 	    } else if ((error == NFSERR_EXPIRED || error == NFSERR_BADSTATEID)
 		&& clidrev != 0) {
 		expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
@@ -460,7 +460,7 @@ nfsrpc_openrpc(struct nfsmount *nmp, vnode_t vp, u_int8_t *nfhp, int fhlen,
 			ret = nfsrpc_openconfirm(vp, newfhp, newfhlen, op,
 			    cred, p);
 			if (ret == NFSERR_DELAY)
-			    (void) nfs_catnap(PZERO, "nfs_open");
+			    (void) nfs_catnap(PZERO, ret, "nfs_open");
 		    } while (ret == NFSERR_DELAY);
 		    error = ret;
 		}
@@ -484,7 +484,7 @@ nfsrpc_openrpc(struct nfsmount *nmp, vnode_t vp, u_int8_t *nfhp, int fhlen,
 			    newfhlen, mode, op, name, namelen, &ndp, 0, 0x0,
 			    cred, p, syscred, 1);
 			if (ret == NFSERR_DELAY)
-			    (void) nfs_catnap(PZERO, "nfs_open2");
+			    (void) nfs_catnap(PZERO, ret, "nfs_open2");
 		    } while (ret == NFSERR_DELAY);
 		    if (ret) {
 			if (ndp != NULL)
@@ -624,6 +624,7 @@ nfsrpc_doclose(struct nfsmount *nmp, struct nfsclopen *op, NFSPROC_T *p)
 					    nd->nd_repstat == NFSERR_DELAY) &&
 					    error == 0)
 						(void) nfs_catnap(PZERO,
+						    (int)nd->nd_repstat,
 						    "nfs_close");
 				} while ((nd->nd_repstat == NFSERR_GRACE ||
 				    nd->nd_repstat == NFSERR_DELAY) &&
@@ -645,7 +646,7 @@ nfsrpc_doclose(struct nfsmount *nmp, struct nfsclopen *op, NFSPROC_T *p)
 	do {
 		error = nfscl_tryclose(op, tcred, nmp, p);
 		if (error == NFSERR_GRACE)
-			(void) nfs_catnap(PZERO, "nfs_close");
+			(void) nfs_catnap(PZERO, error, "nfs_close");
 	} while (error == NFSERR_GRACE);
 	NFSLOCKCLSTATE();
 	nfscl_lockunlock(&op->nfso_own->nfsow_rwlock);
@@ -999,7 +1000,7 @@ nfsrpc_setattr(vnode_t vp, struct vattr *vap, NFSACL_T *aclp,
 		if (error == NFSERR_GRACE || error == NFSERR_STALESTATEID ||
 		    error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY ||
 		    error == NFSERR_OLDSTATEID) {
-			(void) nfs_catnap(PZERO, "nfs_setattr");
+			(void) nfs_catnap(PZERO, error, "nfs_setattr");
 		} else if ((error == NFSERR_EXPIRED ||
 		    error == NFSERR_BADSTATEID) && clidrev != 0) {
 			expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
@@ -1244,7 +1245,7 @@ nfsrpc_read(vnode_t vp, struct uio *uiop, struct ucred *cred,
 		if (error == NFSERR_GRACE || error == NFSERR_STALESTATEID ||
 		    error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY ||
 		    error == NFSERR_OLDSTATEID) {
-			(void) nfs_catnap(PZERO, "nfs_read");
+			(void) nfs_catnap(PZERO, error, "nfs_read");
 		} else if ((error == NFSERR_EXPIRED ||
 		    error == NFSERR_BADSTATEID) && clidrev != 0) {
 			expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
@@ -1409,7 +1410,7 @@ nfscl_dumpstate(nmp, 1, 1, 0, 0);
 		if (error == NFSERR_GRACE || error == NFSERR_STALESTATEID ||
 		    error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY ||
 		    error == NFSERR_OLDSTATEID) {
-			(void) nfs_catnap(PZERO, "nfs_write");
+			(void) nfs_catnap(PZERO, error, "nfs_write");
 		} else if ((error == NFSERR_EXPIRED ||
 		    error == NFSERR_BADSTATEID) && clidrev != 0) {
 			expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
@@ -1736,7 +1737,7 @@ nfsrpc_create(vnode_t dvp, char *name, int namelen, struct vattr *vap,
 		nfscl_ownerrelease(owp, error, newone, unlocked);
 		if (error == NFSERR_GRACE || error == NFSERR_STALECLIENTID ||
 		    error == NFSERR_STALEDONTRECOVER || error == NFSERR_DELAY) {
-			(void) nfs_catnap(PZERO, "nfs_open");
+			(void) nfs_catnap(PZERO, error, "nfs_open");
 		} else if ((error == NFSERR_EXPIRED ||
 		    error == NFSERR_BADSTATEID) && clidrev != 0) {
 			expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
@@ -1969,7 +1970,7 @@ nfsrpc_createv4(vnode_t dvp, char *name, int namelen, struct vattr *vap,
 			ret = nfsrpc_openconfirm(dvp, nfhp->nfh_fh,
 			    nfhp->nfh_len, op, cred, p);
 			if (ret == NFSERR_DELAY)
-			    (void) nfs_catnap(PZERO, "nfs_create");
+			    (void) nfs_catnap(PZERO, ret, "nfs_create");
 		    } while (ret == NFSERR_DELAY);
 		    error = ret;
 		}
@@ -1991,7 +1992,7 @@ nfsrpc_createv4(vnode_t dvp, char *name, int namelen, struct vattr *vap,
 			    (NFSV4OPEN_ACCESSWRITE | NFSV4OPEN_ACCESSREAD), op,
 			    name, namelen, &dp, 0, 0x0, cred, p, 0, 1);
 			if (ret == NFSERR_DELAY)
-			    (void) nfs_catnap(PZERO, "nfs_crt2");
+			    (void) nfs_catnap(PZERO, ret, "nfs_crt2");
 		    } while (ret == NFSERR_DELAY);
 		    if (ret) {
 			if (dp != NULL)
@@ -3533,7 +3534,8 @@ nfsrpc_advlock(vnode_t vp, off_t size, int op, struct flock *fl,
 			    if ((nd->nd_repstat == NFSERR_GRACE ||
 				 nd->nd_repstat == NFSERR_DELAY) &&
 				error == 0)
-				(void) nfs_catnap(PZERO, "nfs_advlock");
+				(void) nfs_catnap(PZERO, (int)nd->nd_repstat,
+				    "nfs_advlock");
 			} while ((nd->nd_repstat == NFSERR_GRACE ||
 			    nd->nd_repstat == NFSERR_DELAY) && error == 0);
 		    }
@@ -3570,7 +3572,7 @@ nfsrpc_advlock(vnode_t vp, off_t size, int op, struct flock *fl,
 	    if (error == NFSERR_GRACE || error == NFSERR_STALESTATEID ||
 		error == NFSERR_STALEDONTRECOVER ||
 		error == NFSERR_STALECLIENTID || error == NFSERR_DELAY) {
-		(void) nfs_catnap(PZERO, "nfs_advlock");
+		(void) nfs_catnap(PZERO, error, "nfs_advlock");
 	    } else if ((error == NFSERR_EXPIRED || error == NFSERR_BADSTATEID)
 		&& clidrev != 0) {
 		expireret = nfscl_hasexpired(nmp->nm_clp, clidrev, p);
diff --git a/sys/fs/nfsclient/nfs_clstate.c b/sys/fs/nfsclient/nfs_clstate.c
index cf8c922ae43..b6fad20d0f5 100644
--- a/sys/fs/nfsclient/nfs_clstate.c
+++ b/sys/fs/nfsclient/nfs_clstate.c
@@ -784,7 +784,7 @@ nfscl_getcl(vnode_t vp, struct ucred *cred, NFSPROC_T *p,
 			if (error == NFSERR_STALECLIENTID ||
 			    error == NFSERR_STALEDONTRECOVER ||
 			    error == NFSERR_CLIDINUSE) {
-				(void) nfs_catnap(PZERO, "nfs_setcl");
+				(void) nfs_catnap(PZERO, error, "nfs_setcl");
 			}
 		} while (((error == NFSERR_STALECLIENTID ||
 		     error == NFSERR_STALEDONTRECOVER) && --trystalecnt > 0) ||
@@ -2046,7 +2046,7 @@ nfscl_recover(struct nfsclclient *clp, struct ucred *cred, NFSPROC_T *p)
 			newnfs_copycred(&op->nfso_cred, tcred);
 			error = nfscl_tryclose(op, tcred, nmp, p);
 			if (error == NFSERR_GRACE)
-				(void) nfs_catnap(PZERO, "nfsexcls");
+				(void) nfs_catnap(PZERO, error, "nfsexcls");
 		} while (error == NFSERR_GRACE);
 		LIST_REMOVE(op, nfso_list);
 		FREE((caddr_t)op, M_NFSCLOPEN);
@@ -2059,7 +2059,7 @@ nfscl_recover(struct nfsclclient *clp, struct ucred *cred, NFSPROC_T *p)
 			newnfs_copycred(&dp->nfsdl_cred, tcred);
 			error = nfscl_trydelegreturn(dp, tcred, nmp, p);
 			if (error == NFSERR_GRACE)
-				(void) nfs_catnap(PZERO, "nfsexdlg");
+				(void) nfs_catnap(PZERO, error, "nfsexdlg");
 		} while (error == NFSERR_GRACE);
 		TAILQ_REMOVE(&extra_deleg, dp, nfsdl_list);
 		FREE((caddr_t)dp, M_NFSCLDELEG);
@@ -3619,7 +3619,7 @@ nfscl_tryopen(struct nfsmount *nmp, vnode_t vp, u_int8_t *fhp, int fhlen,
 		    mode, op, name, namelen, ndpp, reclaim, delegtype, cred, p,
 		    0, 0);
 		if (error == NFSERR_DELAY)
-			(void) nfs_catnap(PZERO, "nfstryop");
+			(void) nfs_catnap(PZERO, error, "nfstryop");
 	} while (error == NFSERR_DELAY);
 	if (error == EAUTH || error == EACCES) {
 		/* Try again using system credentials */
@@ -3629,7 +3629,7 @@ nfscl_tryopen(struct nfsmount *nmp, vnode_t vp, u_int8_t *fhp, int fhlen,
 			newfhlen, mode, op, name, namelen, ndpp, reclaim,
 			delegtype, cred, p, 1, 0);
 		    if (error == NFSERR_DELAY)
-			(void) nfs_catnap(PZERO, "nfstryop");
+			(void) nfs_catnap(PZERO, error, "nfstryop");
 		} while (error == NFSERR_DELAY);
 	}
 	return (error);
@@ -3652,7 +3652,8 @@ nfscl_trylock(struct nfsmount *nmp, vnode_t vp, u_int8_t *fhp,
 		error = nfsrpc_lock(nd, nmp, vp, fhp, fhlen, nlp, newone,
 		    reclaim, off, len, type, cred, p, 0);
 		if (!error && nd->nd_repstat == NFSERR_DELAY)
-			(void) nfs_catnap(PZERO, "nfstrylck");
+			(void) nfs_catnap(PZERO, (int)nd->nd_repstat,
+			    "nfstrylck");
 	} while (!error && nd->nd_repstat == NFSERR_DELAY);
 	if (!error)
 		error = nd->nd_repstat;
@@ -3663,7 +3664,8 @@ nfscl_trylock(struct nfsmount *nmp, vnode_t vp, u_int8_t *fhp,
 			error = nfsrpc_lock(nd, nmp, vp, fhp, fhlen, nlp,
 			    newone, reclaim, off, len, type, cred, p, 1);
 			if (!error && nd->nd_repstat == NFSERR_DELAY)
-				(void) nfs_catnap(PZERO, "nfstrylck");
+				(void) nfs_catnap(PZERO, (int)nd->nd_repstat,
+				    "nfstrylck");
 		} while (!error && nd->nd_repstat == NFSERR_DELAY);
 		if (!error)
 			error = nd->nd_repstat;
@@ -3685,7 +3687,7 @@ nfscl_trydelegreturn(struct nfscldeleg *dp, struct ucred *cred,
 	do {
 		error = nfsrpc_delegreturn(dp, cred, nmp, p, 0);
 		if (error == NFSERR_DELAY)
-			(void) nfs_catnap(PZERO, "nfstrydp");
+			(void) nfs_catnap(PZERO, error, "nfstrydp");
 	} while (error == NFSERR_DELAY);
 	if (error == EAUTH || error == EACCES) {
 		/* Try again using system credentials */
@@ -3693,7 +3695,7 @@ nfscl_trydelegreturn(struct nfscldeleg *dp, struct ucred *cred,
 		do {
 			error = nfsrpc_delegreturn(dp, cred, nmp, p, 1);
 			if (error == NFSERR_DELAY)
-				(void) nfs_catnap(PZERO, "nfstrydp");
+				(void) nfs_catnap(PZERO, error, "nfstrydp");
 		} while (error == NFSERR_DELAY);
 	}
 	return (error);
@@ -3714,7 +3716,7 @@ nfscl_tryclose(struct nfsclopen *op, struct ucred *cred,
 	do {
 		error = nfsrpc_closerpc(nd, nmp, op, cred, p, 0);
 		if (error == NFSERR_DELAY)
-			(void) nfs_catnap(PZERO, "nfstrycl");
+			(void) nfs_catnap(PZERO, error, "nfstrycl");
 	} while (error == NFSERR_DELAY);
 	if (error == EAUTH || error == EACCES) {
 		/* Try again using system credentials */
@@ -3722,7 +3724,7 @@ nfscl_tryclose(struct nfsclopen *op, struct ucred *cred,
 		do {
 			error = nfsrpc_closerpc(nd, nmp, op, cred, p, 1);
 			if (error == NFSERR_DELAY)
-				(void) nfs_catnap(PZERO, "nfstrycl");
+				(void) nfs_catnap(PZERO, error, "nfstrycl");
 		} while (error == NFSERR_DELAY);
 	}
 	return (error);
diff --git a/sys/fs/nfsclient/nfs_clvfsops.c b/sys/fs/nfsclient/nfs_clvfsops.c
index 0bf2bf4b4e5..76154c37180 100644
--- a/sys/fs/nfsclient/nfs_clvfsops.c
+++ b/sys/fs/nfsclient/nfs_clvfsops.c
@@ -652,7 +652,7 @@ nfs_decode_args(struct mount *mp, struct nfsmount *nmp, struct nfs_args *argp,
 			while (newnfs_connect(nmp, &nmp->nm_sockreq,
 			    cred, td, 0)) {
 				printf("newnfs_args: retrying connect\n");
-				(void) nfs_catnap(PSOCK, "newnfscon");
+				(void) nfs_catnap(PSOCK, 0, "newnfscon");
 			}
 		}
 	} else {
@@ -1188,7 +1188,7 @@ mountnfs(struct nfs_args *argp, struct mount *mp, struct sockaddr *nam,
 			error = nfsrpc_getdirpath(nmp, NFSMNT_DIRPATH(nmp),
 			    cred, td);
 			if (error)
-				(void) nfs_catnap(PZERO, "nfsgetdirp");
+				(void) nfs_catnap(PZERO, error, "nfsgetdirp");
 		} while (error && --trycnt > 0);
 		if (error) {
 			error = nfscl_maperr(td, error, (uid_t)0, (gid_t)0);
@@ -1284,7 +1284,7 @@ nfs_unmount(struct mount *mp, int mntflags)
 	do {
 		error = vflush(mp, 1, flags, td);
 		if ((mntflags & MNT_FORCE) && error != 0 && ++trycnt < 30)
-			(void) nfs_catnap(PSOCK, "newndm");
+			(void) nfs_catnap(PSOCK, error, "newndm");
 	} while ((mntflags & MNT_FORCE) && error != 0 && trycnt < 30);
 	if (error)
 		goto out;
diff --git a/sys/fs/nfsclient/nfs_clvnops.c b/sys/fs/nfsclient/nfs_clvnops.c
index 4e844f1024b..d53c8386242 100644
--- a/sys/fs/nfsclient/nfs_clvnops.c
+++ b/sys/fs/nfsclient/nfs_clvnops.c
@@ -2871,7 +2871,8 @@ nfs_advlock(struct vop_advlock_args *ap)
 			if (ret == NFSERR_DENIED && (ap->a_flags & F_WAIT) &&
 			    ap->a_op == F_SETLK) {
 				VOP_UNLOCK(vp, 0);
-				error = nfs_catnap(PZERO | PCATCH, "ncladvl");
+				error = nfs_catnap(PZERO | PCATCH, ret,
+				    "ncladvl");
 				if (error)
 					return (EINTR);
 				vn_lock(vp, LK_EXCLUSIVE | LK_RETRY);
diff --git a/sys/fs/nfsserver/nfs_nfsdstate.c b/sys/fs/nfsserver/nfs_nfsdstate.c
index a4df1ed9c33..e475cb77826 100644
--- a/sys/fs/nfsserver/nfs_nfsdstate.c
+++ b/sys/fs/nfsserver/nfs_nfsdstate.c
@@ -4578,7 +4578,7 @@ nfsd_recalldelegation(vnode_t vp, NFSPROC_T *p)
 			    100000)
 				return;
 			/* Sleep for a short period of time */
-			(void) nfs_catnap(PZERO, "nfsremove");
+			(void) nfs_catnap(PZERO, 0, "nfsremove");
 		}
 	} while (error == NFSERR_DELAY);
 }

From 7d04a6be1268b76a4ddc34a41bcadd3985b62f24 Mon Sep 17 00:00:00 2001
From: Weongyo Jeong <weongyo@FreeBSD.org>
Date: Sat, 24 Apr 2010 23:32:24 +0000
Subject: [PATCH 240/532] ifp->if_ipackets++ when RX packet interrupts are
 occurred.

MFC after:	3 days
---
 sys/dev/bwn/if_bwn.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/dev/bwn/if_bwn.c b/sys/dev/bwn/if_bwn.c
index 983f9972d60..5ca5098115c 100644
--- a/sys/dev/bwn/if_bwn.c
+++ b/sys/dev/bwn/if_bwn.c
@@ -9368,6 +9368,8 @@ bwn_rxeof(struct bwn_mac *mac, struct mbuf *m, const void *_rxhdr)
 	rssi = rxhdr->phy.abg.rssi;	/* XXX incorrect RSSI calculation? */
 	noise = mac->mac_stats.link_noise;
 
+	ifp->if_ipackets++;
+
 	BWN_UNLOCK(sc);
 
 	ni = ieee80211_find_rxnode(ic, wh);

From 509210970393a1a8cd8a65d5340dc4bed069fa68 Mon Sep 17 00:00:00 2001
From: Kirk McKusick <mckusick@FreeBSD.org>
Date: Sun, 25 Apr 2010 00:05:51 +0000
Subject: [PATCH 241/532] The NFS quota-reporting RPC uses 32-bit sized fields.
 We approximate 64-bit quota sizes by scaling down the sizes by the minimum
 amount necessary to fit in a 32-bit field and then upscale the filesystem
 block size to compensate. For example, if the hard block limit is 0x300000008
 then we set the hard block limit to 0xA0000002 and claim that the blocksize
 is 4 * DEV_BSIZE. This will lose the minimal amount of information thus
 delivering nearly correct answers.

---
 libexec/rpc.rquotad/rquotad.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/libexec/rpc.rquotad/rquotad.c b/libexec/rpc.rquotad/rquotad.c
index 3cd129239e7..d9ce06aac82 100644
--- a/libexec/rpc.rquotad/rquotad.c
+++ b/libexec/rpc.rquotad/rquotad.c
@@ -126,6 +126,7 @@ sendquota(struct svc_req *request, SVCXPRT *transp)
 	struct getquota_rslt getq_rslt;
 	struct dqblk dqblk;
 	struct timeval timev;
+	int scale;
 
 	bzero(&getq_args, sizeof(getq_args));
 	if (!svc_getargs(transp, (xdrproc_t)xdr_getquota_args, &getq_args)) {
@@ -142,13 +143,15 @@ sendquota(struct svc_req *request, SVCXPRT *transp)
 		gettimeofday(&timev, NULL);
 		getq_rslt.status = Q_OK;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_active = TRUE;
-		getq_rslt.getquota_rslt_u.gqr_rquota.rq_bsize = DEV_BSIZE;
+		scale = 1 << flsll(dqblk.dqb_bhardlimit >> 32);
+		getq_rslt.getquota_rslt_u.gqr_rquota.rq_bsize =
+		    DEV_BSIZE * scale;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_bhardlimit =
-		    dqblk.dqb_bhardlimit;
+		    dqblk.dqb_bhardlimit / scale;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_bsoftlimit =
-		    dqblk.dqb_bsoftlimit;
+		    dqblk.dqb_bsoftlimit / scale;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_curblocks =
-		    dqblk.dqb_curblocks;
+		    dqblk.dqb_curblocks / scale;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_fhardlimit =
 		    dqblk.dqb_ihardlimit;
 		getq_rslt.getquota_rslt_u.gqr_rquota.rq_fsoftlimit =

From 6f702278e62c0412863e7731c0952314ef874b22 Mon Sep 17 00:00:00 2001
From: Marcel Moolenaar <marcel@FreeBSD.org>
Date: Sun, 25 Apr 2010 00:54:11 +0000
Subject: [PATCH 242/532] Fix undo for schemes that have internal partitions.
 Internal partitions do not constitute user-visible or active partitions and
 as such should not prevent undoing pending operations.

While here, initialize the last usable sector for the placeholder geom
based on the null scheme, created to allow undoing the destruction of
a scheme. This gives consistent output with "gpart show".

Based on a patch from:	"Andrey V. Elsukov" <bu7cher@yandex.ru>
---
 sys/geom/part/g_part.c | 15 ++++++++++++++-
 1 file changed, 14 insertions(+), 1 deletion(-)

diff --git a/sys/geom/part/g_part.c b/sys/geom/part/g_part.c
index 9cddd0dc9fd..7a979da0bdd 100644
--- a/sys/geom/part/g_part.c
+++ b/sys/geom/part/g_part.c
@@ -855,7 +855,9 @@ g_part_ctl_delete(struct gctl_req *req, struct g_part_parms *gpp)
 static int
 g_part_ctl_destroy(struct gctl_req *req, struct g_part_parms *gpp)
 {
+	struct g_consumer *cp;
 	struct g_geom *gp;
+	struct g_provider *pp;
 	struct g_part_entry *entry;
 	struct g_part_table *null, *table;
 	struct sbuf *sb;
@@ -885,6 +887,11 @@ g_part_ctl_destroy(struct gctl_req *req, struct g_part_parms *gpp)
 	null->gpt_gp = gp;
 	null->gpt_scheme = &g_part_null_scheme;
 	LIST_INIT(&null->gpt_entry);
+
+	cp = LIST_FIRST(&gp->consumer);
+	pp = cp->provider;
+	null->gpt_last = pp->mediasize / pp->sectorsize - 1;
+
 	null->gpt_depth = table->gpt_depth;
 	null->gpt_opened = table->gpt_opened;
 	null->gpt_smhead = table->gpt_smhead;
@@ -1141,10 +1148,16 @@ g_part_ctl_undo(struct gctl_req *req, struct g_part_parms *gpp)
 	    table->gpt_created) ? 1 : 0;
 
 	if (reprobe) {
-		if (!LIST_EMPTY(&table->gpt_entry)) {
+		LIST_FOREACH(entry, &table->gpt_entry, gpe_entry) {
+			if (entry->gpe_internal)
+				continue;
 			error = EBUSY;
 			goto fail;
 		}
+		while ((entry = LIST_FIRST(&table->gpt_entry)) != NULL) {
+			LIST_REMOVE(entry, gpe_entry);
+			g_free(entry);
+		}
 		error = g_part_probe(gp, cp, table->gpt_depth);
 		if (error) {
 			g_topology_lock();

From c74f160cb034b52b0d0a1febc804433dee6a3186 Mon Sep 17 00:00:00 2001
From: Marcel Moolenaar <marcel@FreeBSD.org>
Date: Sun, 25 Apr 2010 01:56:39 +0000
Subject: [PATCH 243/532] Re-calculate a geometry when reprobing as well.

PR:		kern/145452
Reported by:	"Andrey V. Elsukov" <bu7cher@yandex.ru>
---
 sys/geom/part/g_part.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/sys/geom/part/g_part.c b/sys/geom/part/g_part.c
index 7a979da0bdd..f47ac08ffe3 100644
--- a/sys/geom/part/g_part.c
+++ b/sys/geom/part/g_part.c
@@ -1166,6 +1166,15 @@ g_part_ctl_undo(struct gctl_req *req, struct g_part_parms *gpp)
 			return (0);
 		}
 		table = gp->softc;
+
+		/*
+		 * Synthesize a disk geometry. Some partitioning schemes
+		 * depend on it and since some file systems need it even
+		 * when the partitition scheme doesn't, we do it here in
+		 * scheme-independent code.
+		 */
+		pp = cp->provider;
+		g_part_geometry(table, cp, pp->mediasize / pp->sectorsize);
 	}
 
 	error = G_PART_READ(table, cp);

From 67c8bc1bfbb250c63acc6022d724c4a01bd8d49a Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 12:35:39 +0000
Subject: [PATCH 244/532] sysctl(3): Update description of various kern.*
 variables. Also add xrefs for confstr(3) (as sysconf(3) but for strings) and
 kvm(3) (which is a more convenient way to access some of the variables).

PR:		116480
MFC after:	1 week
---
 lib/libc/gen/sysctl.3 | 22 ++++++++++------------
 1 file changed, 10 insertions(+), 12 deletions(-)

diff --git a/lib/libc/gen/sysctl.3 b/lib/libc/gen/sysctl.3
index b7ed09139da..9eba6247b71 100644
--- a/lib/libc/gen/sysctl.3
+++ b/lib/libc/gen/sysctl.3
@@ -28,7 +28,7 @@
 .\"	@(#)sysctl.3	8.4 (Berkeley) 5/9/95
 .\" $FreeBSD$
 .\"
-.Dd February 21, 2010
+.Dd April 25, 2010
 .Dt SYSCTL 3
 .Os
 .Sh NAME
@@ -325,7 +325,7 @@ information.
 .It "KERN_BOOTFILE	string	yes"
 .It "KERN_BOOTTIME	struct timeval	no"
 .It "KERN_CLOCKRATE	struct clockinfo	no"
-.It "KERN_FILE	struct file	no"
+.It "KERN_FILE	struct xfile	no"
 .It "KERN_HOSTID	integer	yes"
 .It "KERN_HOSTUUID	string	yes"
 .It "KERN_HOSTNAME	string	yes"
@@ -342,14 +342,14 @@ information.
 .It "KERN_OSREV	integer	no"
 .It "KERN_OSTYPE	string	no"
 .It "KERN_POSIX1	integer	no"
-.It "KERN_PROC	struct proc	no"
+.It "KERN_PROC	node	not applicable"
 .It "KERN_PROF	node	not applicable"
 .It "KERN_QUANTUM	integer	yes"
 .It "KERN_SAVED_IDS	integer	no"
 .It "KERN_SECURELVL	integer	raise only"
 .It "KERN_UPDATEINTERVAL	integer	no"
 .It "KERN_VERSION	string	no"
-.It "KERN_VNODE	struct vnode	no"
+.It "KERN_VNODE	struct xvnode	no"
 .El
 .Bl -tag -width 6n
 .It Li KERN_ARGMAX
@@ -370,10 +370,8 @@ This structure contains the clock, statistics clock and profiling clock
 frequencies, the number of micro-seconds per hz tick and the skew rate.
 .It Li KERN_FILE
 Return the entire file table.
-The returned data consists of a single
-.Va struct filehead
-followed by an array of
-.Va struct file ,
+The returned data consists of an array of
+.Va struct xfile ,
 whose size depends on the current number of such objects in the system.
 .It Li KERN_HOSTID
 Get or set the host ID.
@@ -525,10 +523,8 @@ Note, the vnode table is not necessarily a consistent snapshot of
 the system.
 The returned data consists of an array whose size depends on the
 current number of such objects in the system.
-Each element of the array contains the kernel address of a vnode
-.Va struct vnode *
-followed by the vnode itself
-.Va struct vnode .
+Each element of the array consists of a
+.Va struct xvnode .
 .El
 .Ss CTL_NET
 The string and integer information available for the CTL_NET level
@@ -854,6 +850,8 @@ An attempt is made to set a read-only value.
 A process without appropriate privilege attempts to set a value.
 .El
 .Sh SEE ALSO
+.Xr confstr 3 ,
+.Xr kvm 3 ,
 .Xr sysconf 3 ,
 .Xr sysctl 8
 .Sh HISTORY

From 886ae81fa3357f9f608042aa4a3db0148e56144c Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 12:50:30 +0000
Subject: [PATCH 245/532] kvm(3): Mention that some of the functions use
 sysctl(3) instead of kmem.

Additionally, because of sysctl(3) use (which is generally good), behaviour
for crash dumps differs slightly from behaviour for live kernels and this
will probably never be fixed entirely, so weaken that claim.

MFC after:	1 week
---
 lib/libkvm/kvm.3 | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/lib/libkvm/kvm.3 b/lib/libkvm/kvm.3
index f694fd80073..9dcd772d7e2 100644
--- a/lib/libkvm/kvm.3
+++ b/lib/libkvm/kvm.3
@@ -32,7 +32,7 @@
 .\"     @(#)kvm.3	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd January 29, 2004
+.Dd April 25, 2010
 .Dt KVM 3
 .Os
 .Sh NAME
@@ -46,12 +46,15 @@ The
 library provides a uniform interface for accessing kernel virtual memory
 images, including live systems and crash dumps.
 Access to live systems is via
+.Xr sysctl 3
+for some functions, and
 .Xr mem 4
 and
 .Xr kmem 4
+for other functions,
 while crash dumps can be examined via the core file generated by
 .Xr savecore 8 .
-The interface behaves identically in both cases.
+The interface behaves similarly in both cases.
 Memory can be read and written, kernel symbol addresses can be
 looked up efficiently, and information about user processes can
 be gathered.
@@ -112,5 +115,6 @@ given descriptor.
 .Xr kvm_openfiles 3 ,
 .Xr kvm_read 3 ,
 .Xr kvm_write 3 ,
+.Xr sysctl 3 ,
 .Xr kmem 4 ,
 .Xr mem 4

From 44f78be9241bc78bbf112c1755df9367c3598aaf Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 13:13:23 +0000
Subject: [PATCH 246/532] symlink(7): The ownership of symlinks is used by the
 system, in at least three ways, so do not say it is ignored: * who may
 delete/rename a symlink in a sticky directory * who may do
 lchflags(2)/lchown(2)/lchmod(2) * whose inode quota is charged

MFC after:	1 week
---
 bin/ln/symlink.7 | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/bin/ln/symlink.7 b/bin/ln/symlink.7
index d2e8cbb25e4..93bb27bef15 100644
--- a/bin/ln/symlink.7
+++ b/bin/ln/symlink.7
@@ -29,7 +29,7 @@
 .\"	@(#)symlink.7	8.3 (Berkeley) 3/31/94
 .\" $FreeBSD$
 .\"
-.Dd March 31, 1994
+.Dd April 25, 2010
 .Dt SYMLINK 7
 .Os
 .Sh NAME
@@ -138,8 +138,8 @@ an existing symbolic link can be changed by means of the
 and
 .Xr lutimes 2
 system calls, respectively.
-Of these, only the flags are used by the system;
-the access permissions and ownership are ignored.
+Of these, only the flags and ownership are used by the system;
+the access permissions are ignored.
 .Pp
 The
 .Bx 4.4

From 141c21d9de2657423e318bd44fbb72e4b1c931c8 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 13:29:59 +0000
Subject: [PATCH 247/532] symlink(7): Add lpathconf(2) and *at system calls.

MFC after:	1 week
---
 bin/ln/symlink.7 | 29 +++++++++++++++++++++++++++--
 1 file changed, 27 insertions(+), 2 deletions(-)

diff --git a/bin/ln/symlink.7 b/bin/ln/symlink.7
index 93bb27bef15..8c09f281a51 100644
--- a/bin/ln/symlink.7
+++ b/bin/ln/symlink.7
@@ -103,19 +103,23 @@ the system call
 would return a file descriptor to the file
 .Dq afile .
 .Pp
-There are nine system calls that do not follow links, and which operate
+There are thirteen system calls that do not follow links, and which operate
 on the symbolic link itself.
 They are:
 .Xr lchflags 2 ,
 .Xr lchmod 2 ,
 .Xr lchown 2 ,
+.Xr lpathconf 2 ,
 .Xr lstat 2 ,
 .Xr lutimes 2 ,
 .Xr readlink 2 ,
+.Xr readlinkat 2 ,
 .Xr rename 2 ,
+.Xr renameat 2 ,
 .Xr rmdir 2 ,
+.Xr unlink 2 ,
 and
-.Xr unlink 2 .
+.Xr unlinkat 2 .
 Because
 .Xr remove 3
 is an alias for
@@ -123,9 +127,30 @@ is an alias for
 it also does not follow symbolic links.
 When
 .Xr rmdir 2
+or
+.Xr unlinkat 2
+with the
+.Dv AT_REMOVEDIR
+flag
 is applied to a symbolic link, it fails with the error
 .Er ENOTDIR .
 .Pp
+The
+.Xr linkat 2
+system call does not follow symbolic links
+unless given the
+.Dv AT_SYMLINK_FOLLOW
+flag.
+.Pp
+The following system calls follow symbolic links
+unless given the
+.Dv AT_SYMLINK_NOFOLLOW
+flag:
+.Xr fchmodat 2 ,
+.Xr fchownat 2
+and
+.Xr fstatat 2 .
+.Pp
 The owner and group of an existing symbolic link can be changed by
 means of the
 .Xr lchown 2

From b559746adb8637f516cc54e670f14df45bd175c7 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 13:55:52 +0000
Subject: [PATCH 248/532] unlinkat(2): unlinkat(AT_REMOVEDIR) fails with
 ENOTEMPTY like rmdir() for non-empty directories.

POSIX permits both ENOTEMPTY and EEXIST, but we use the clearer ENOTEMPTY,
following BSD tradition.

MFC after:	1 week
---
 lib/libc/sys/unlink.2 | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/libc/sys/unlink.2 b/lib/libc/sys/unlink.2
index 141f3d320e5..bb27085422b 100644
--- a/lib/libc/sys/unlink.2
+++ b/lib/libc/sys/unlink.2
@@ -28,7 +28,7 @@
 .\"     @(#)unlink.2	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd April 10, 2008
+.Dd April 25, 2010
 .Dt UNLINK 2
 .Os
 .Sh NAME
@@ -167,7 +167,7 @@ argument does not specify an absolute path and the
 argument is neither
 .Dv AT_FDCWD
 nor a valid file descriptor open for searching.
-.It Bq Er EEXIST
+.It Bq Er ENOTEMPTY
 The
 .Fa flag
 parameter has the

From 475d0674a61238df60dd439663d62bb0dd76408e Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Sun, 25 Apr 2010 15:04:57 +0000
Subject: [PATCH 249/532] Undo my lastest fix since that wasn't one at all.

MFC after: 3 days.
---
 sys/netinet/sctp_indata.c | 38 +++++++++++++++++++-------------------
 1 file changed, 19 insertions(+), 19 deletions(-)

diff --git a/sys/netinet/sctp_indata.c b/sys/netinet/sctp_indata.c
index 4b00fc3fc2c..2ed6c1660d7 100644
--- a/sys/netinet/sctp_indata.c
+++ b/sys/netinet/sctp_indata.c
@@ -3514,25 +3514,6 @@ sctp_strike_gap_ack_chunks(struct sctp_tcb *stcb, struct sctp_association *asoc,
 		if (tp1->sent == SCTP_DATAGRAM_RESEND) {
 			struct sctp_nets *alt;
 
-			if ((stcb->asoc.peer_supports_prsctp) &&
-			    (PR_SCTP_RTX_ENABLED(tp1->flags))) {
-				/*
-				 * Has it been retransmitted tv_sec times? -
-				 * we store the retran count there.
-				 */
-				if (tp1->snd_count > tp1->rec.data.timetodrop.tv_sec) {
-					/* Yes, so drop it */
-					if (tp1->data != NULL) {
-						(void)sctp_release_pr_sctp_chunk(stcb, tp1,
-						    (SCTP_RESPONSE_TO_USER_REQ | SCTP_NOTIFY_DATAGRAM_SENT),
-						    SCTP_SO_NOT_LOCKED);
-					}
-					/* Make sure to flag we had a FR */
-					tp1->whoTo->net_ack++;
-					tp1 = TAILQ_NEXT(tp1, sctp_next);
-					continue;
-				}
-			}
 			/* fix counts and things */
 			if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_FLIGHT_LOGGING_ENABLE) {
 				sctp_misc_ints(SCTP_FLIGHT_LOG_DOWN_RSND,
@@ -3555,6 +3536,25 @@ sctp_strike_gap_ack_chunks(struct sctp_tcb *stcb, struct sctp_association *asoc,
 			/* remove from the total flight */
 			sctp_total_flight_decrease(stcb, tp1);
 
+			if ((stcb->asoc.peer_supports_prsctp) &&
+			    (PR_SCTP_RTX_ENABLED(tp1->flags))) {
+				/*
+				 * Has it been retransmitted tv_sec times? -
+				 * we store the retran count there.
+				 */
+				if (tp1->snd_count > tp1->rec.data.timetodrop.tv_sec) {
+					/* Yes, so drop it */
+					if (tp1->data != NULL) {
+						(void)sctp_release_pr_sctp_chunk(stcb, tp1,
+						    (SCTP_RESPONSE_TO_USER_REQ | SCTP_NOTIFY_DATAGRAM_SENT),
+						    SCTP_SO_NOT_LOCKED);
+					}
+					/* Make sure to flag we had a FR */
+					tp1->whoTo->net_ack++;
+					tp1 = TAILQ_NEXT(tp1, sctp_next);
+					continue;
+				}
+			}
 			/* printf("OK, we are now ready to FR this guy\n"); */
 			if (SCTP_BASE_SYSCTL(sctp_logging_level) & SCTP_FR_LOGGING_ENABLE) {
 				sctp_log_fr(tp1->rec.data.TSN_seq, tp1->snd_count,

From 427a928af7c19b884eba6efee02193a7aa373401 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sun, 25 Apr 2010 16:42:47 +0000
Subject: [PATCH 250/532] Provide 32bit compat shims for sysctl net.route
 NET_RT_IFLIST. This allows getifaddrs(3) to work for compat32 binaries.

Submitted by:	jhb (6.x version)
Reviewed by:	emaste
Tested by:	emaste and <pluknet gmail com>
MFC after:	2 weeks
---
 sys/net/rtsock.c | 102 ++++++++++++++++++++++++++++++++++++++++++++++-
 1 file changed, 101 insertions(+), 1 deletion(-)

diff --git a/sys/net/rtsock.c b/sys/net/rtsock.c
index df4f9ae762b..de6548228ea 100644
--- a/sys/net/rtsock.c
+++ b/sys/net/rtsock.c
@@ -29,6 +29,7 @@
  *	@(#)rtsock.c	8.7 (Berkeley) 10/12/95
  * $FreeBSD$
  */
+#include "opt_compat.h"
 #include "opt_sctp.h"
 #include "opt_mpath.h"
 #include "opt_inet.h"
@@ -71,6 +72,49 @@ extern void sctp_addr_change(struct ifaddr *ifa, int cmd);
 #endif /* SCTP */
 #endif
 
+#ifdef COMPAT_FREEBSD32
+#include <sys/mount.h>
+#include <compat/freebsd32/freebsd32.h>
+
+struct if_data32 {
+	uint8_t	ifi_type;
+	uint8_t	ifi_physical;
+	uint8_t	ifi_addrlen;
+	uint8_t	ifi_hdrlen;
+	uint8_t	ifi_link_state;
+	uint8_t	ifi_spare_char1;
+	uint8_t	ifi_spare_char2;
+	uint8_t	ifi_datalen;
+	uint32_t ifi_mtu;
+	uint32_t ifi_metric;
+	uint32_t ifi_baudrate;
+	uint32_t ifi_ipackets;
+	uint32_t ifi_ierrors;
+	uint32_t ifi_opackets;
+	uint32_t ifi_oerrors;
+	uint32_t ifi_collisions;
+	uint32_t ifi_ibytes;
+	uint32_t ifi_obytes;
+	uint32_t ifi_imcasts;
+	uint32_t ifi_omcasts;
+	uint32_t ifi_iqdrops;
+	uint32_t ifi_noproto;
+	uint32_t ifi_hwassist;
+	int32_t	ifi_epoch;
+	struct	timeval32 ifi_lastchange;
+};
+
+struct if_msghdr32 {
+	uint16_t ifm_msglen;
+	uint8_t	ifm_version;
+	uint8_t	ifm_type;
+	int32_t	ifm_addrs;
+	int32_t	ifm_flags;
+	uint16_t ifm_index;
+	struct	if_data32 ifm_data;
+};
+#endif
+
 MALLOC_DEFINE(M_RTABLE, "routetbl", "routing tables");
 
 /* NB: these are not modified */
@@ -1001,6 +1045,12 @@ again:
 		break;
 
 	case RTM_IFINFO:
+#ifdef COMPAT_FREEBSD32
+		if (w != NULL && w->w_req->flags & SCTL_MASK32) {
+			len = sizeof(struct if_msghdr32);
+			break;
+		}
+#endif
 		len = sizeof(struct if_msghdr);
 		break;
 
@@ -1367,6 +1417,38 @@ sysctl_dumpentry(struct radix_node *rn, void *vw)
 	return (error);
 }
 
+#ifdef COMPAT_FREEBSD32
+static void
+copy_ifdata32(struct if_data *src, struct if_data32 *dst)
+{
+
+	bzero(dst, sizeof(*dst));
+	CP(*src, *dst, ifi_type);
+	CP(*src, *dst, ifi_physical);
+	CP(*src, *dst, ifi_addrlen);
+	CP(*src, *dst, ifi_hdrlen);
+	CP(*src, *dst, ifi_link_state);
+	CP(*src, *dst, ifi_datalen);
+	CP(*src, *dst, ifi_mtu);
+	CP(*src, *dst, ifi_metric);
+	CP(*src, *dst, ifi_baudrate);
+	CP(*src, *dst, ifi_ipackets);
+	CP(*src, *dst, ifi_ierrors);
+	CP(*src, *dst, ifi_opackets);
+	CP(*src, *dst, ifi_oerrors);
+	CP(*src, *dst, ifi_collisions);
+	CP(*src, *dst, ifi_ibytes);
+	CP(*src, *dst, ifi_obytes);
+	CP(*src, *dst, ifi_imcasts);
+	CP(*src, *dst, ifi_omcasts);
+	CP(*src, *dst, ifi_iqdrops);
+	CP(*src, *dst, ifi_noproto);
+	CP(*src, *dst, ifi_hwassist);
+	CP(*src, *dst, ifi_epoch);
+	TV_CP(*src, *dst, ifi_lastchange);
+}
+#endif
+
 static int
 sysctl_iflist(int af, struct walkarg *w)
 {
@@ -1387,12 +1469,30 @@ sysctl_iflist(int af, struct walkarg *w)
 		if (w->w_req && w->w_tmem) {
 			struct if_msghdr *ifm;
 
+#ifdef COMPAT_FREEBSD32
+			if (w->w_req->flags & SCTL_MASK32) {
+				struct if_msghdr32 *ifm32;
+
+				ifm32 = (struct if_msghdr32 *)w->w_tmem;
+				ifm32->ifm_index = ifp->if_index;
+				ifm32->ifm_flags = ifp->if_flags |
+				    ifp->if_drv_flags;
+				copy_ifdata32(&ifp->if_data, &ifm32->ifm_data);
+				ifm32->ifm_addrs = info.rti_addrs;
+				error = SYSCTL_OUT(w->w_req, (caddr_t)ifm32,
+				    len);
+				goto sysctl_out;
+			}
+#endif
 			ifm = (struct if_msghdr *)w->w_tmem;
 			ifm->ifm_index = ifp->if_index;
 			ifm->ifm_flags = ifp->if_flags | ifp->if_drv_flags;
 			ifm->ifm_data = ifp->if_data;
 			ifm->ifm_addrs = info.rti_addrs;
-			error = SYSCTL_OUT(w->w_req,(caddr_t)ifm, len);
+			error = SYSCTL_OUT(w->w_req, (caddr_t)ifm, len);
+#ifdef COMPAT_FREEBSD32
+		sysctl_out:
+#endif
 			if (error)
 				goto done;
 		}

From fc0a61a401aecd026b9fee11bc437c9c54817b6b Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sun, 25 Apr 2010 16:43:41 +0000
Subject: [PATCH 251/532] Provide compat32 shims for bpf(4), except zero-copy
 facilities.

bd_compat32 field of struct bpf_d is kept unconditionally to not
impose the requirement of including "opt_compat.h" on all numerous
users of bpfdesc.h.

Submitted by:	jhb (version for 6.x)
Reviewed and tested by:	emaste
MFC after:	2 weeks
---
 sys/net/bpf.c     | 166 +++++++++++++++++++++++++++++++++++++++++++++-
 sys/net/bpfdesc.h |   1 +
 2 files changed, 165 insertions(+), 2 deletions(-)

diff --git a/sys/net/bpf.c b/sys/net/bpf.c
index bac3da40308..1755ec7e8f2 100644
--- a/sys/net/bpf.c
+++ b/sys/net/bpf.c
@@ -38,6 +38,7 @@
 __FBSDID("$FreeBSD$");
 
 #include "opt_bpf.h"
+#include "opt_compat.h"
 #include "opt_netgraph.h"
 
 #include <sys/types.h>
@@ -89,6 +90,43 @@ MALLOC_DEFINE(M_BPF, "BPF", "BPF data");
 
 #define PRINET  26			/* interruptible */
 
+#ifdef COMPAT_FREEBSD32
+#include <sys/mount.h>
+#include <compat/freebsd32/freebsd32.h>
+#define BPF_ALIGNMENT32 sizeof(int32_t)
+#define BPF_WORDALIGN32(x) (((x)+(BPF_ALIGNMENT32-1))&~(BPF_ALIGNMENT32-1))
+
+/*
+ * 32-bit version of structure prepended to each packet.  We use this header
+ * instead of the standard one for 32-bit streams.  We mark the a stream as
+ * 32-bit the first time we see a 32-bit compat ioctl request.
+ */
+struct bpf_hdr32 {
+	struct timeval32 bh_tstamp;	/* time stamp */
+	uint32_t	bh_caplen;	/* length of captured portion */
+	uint32_t	bh_datalen;	/* original length of packet */
+	uint16_t	bh_hdrlen;	/* length of bpf header (this struct
+					   plus alignment padding) */
+};
+
+struct bpf_program32 {
+	u_int bf_len;
+	uint32_t bf_insns;
+};
+
+struct bpf_dltlist32 {
+	u_int	bfl_len;
+	u_int	bfl_list;
+};
+
+#define	BIOCSETF32	_IOW('B', 103, struct bpf_program32)
+#define	BIOCSRTIMEOUT32	_IOW('B',109, struct timeval32)
+#define	BIOCGRTIMEOUT32	_IOR('B',110, struct timeval32)
+#define	BIOCGDLTLIST32	_IOWR('B',121, struct bpf_dltlist32)
+#define	BIOCSETWF32	_IOW('B',123, struct bpf_program32)
+#define	BIOCSETFNR32	_IOW('B',130, struct bpf_program32)
+#endif
+
 /*
  * bpf_iflist is a list of BPF interface structures, each corresponding to a
  * specific DLT.  The same network interface might have several BPF interface
@@ -1005,8 +1043,14 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 		case BIOCFLUSH:
 		case BIOCGDLT:
 		case BIOCGDLTLIST:
+#ifdef COMPAT_FREEBSD32
+		case BIOCGDLTLIST32:
+#endif
 		case BIOCGETIF:
 		case BIOCGRTIMEOUT:
+#ifdef COMPAT_FREEBSD32
+		case BIOCGRTIMEOUT32:
+#endif
 		case BIOCGSTATS:
 		case BIOCVERSION:
 		case BIOCGRSIG:
@@ -1015,6 +1059,9 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 		case FIONREAD:
 		case BIOCLOCK:
 		case BIOCSRTIMEOUT:
+#ifdef COMPAT_FREEBSD32
+		case BIOCSRTIMEOUT32:
+#endif
 		case BIOCIMMEDIATE:
 		case TIOCGPGRP:
 		case BIOCROTZBUF:
@@ -1023,6 +1070,22 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 			return (EPERM);
 		}
 	}
+#ifdef COMPAT_FREEBSD32
+	/*
+	 * If we see a 32-bit compat ioctl, mark the stream as 32-bit so
+	 * that it will get 32-bit packet headers.
+	 */
+	switch (cmd) {
+	case BIOCSETF32:
+	case BIOCSETFNR32:
+	case BIOCSETWF32:
+	case BIOCGDLTLIST32:
+	case BIOCGRTIMEOUT32:
+	case BIOCSRTIMEOUT32:
+		d->bd_compat32 = 1;
+	}
+#endif
+
 	CURVNET_SET(TD_TO_VNET(td));
 	switch (cmd) {
 
@@ -1080,6 +1143,11 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 	case BIOCSETF:
 	case BIOCSETFNR:
 	case BIOCSETWF:
+#ifdef COMPAT_FREEBSD32
+	case BIOCSETF32:
+	case BIOCSETFNR32:
+	case BIOCSETWF32:
+#endif
 		error = bpf_setf(d, (struct bpf_program *)addr, cmd);
 		break;
 
@@ -1123,6 +1191,26 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 	/*
 	 * Get a list of supported data link types.
 	 */
+#ifdef COMPAT_FREEBSD32
+	case BIOCGDLTLIST32:
+		{
+			struct bpf_dltlist32 *list32;
+			struct bpf_dltlist dltlist;
+
+			list32 = (struct bpf_dltlist32 *)addr;
+			dltlist.bfl_len = list32->bfl_len;
+			dltlist.bfl_list = PTRIN(list32->bfl_list);
+			if (d->bd_bif == NULL)
+				error = EINVAL;
+			else {
+				error = bpf_getdltlist(d, &dltlist);
+				if (error == 0)
+					list32->bfl_len = dltlist.bfl_len;
+			}
+			break;
+		}
+#endif
+
 	case BIOCGDLTLIST:
 		if (d->bd_bif == NULL)
 			error = EINVAL;
@@ -1166,8 +1254,23 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 	 * Set read timeout.
 	 */
 	case BIOCSRTIMEOUT:
+#ifdef COMPAT_FREEBSD32
+	case BIOCSRTIMEOUT32:
+#endif
 		{
 			struct timeval *tv = (struct timeval *)addr;
+#ifdef COMPAT_FREEBSD32
+			struct timeval32 *tv32;
+			struct timeval tv64;
+
+			if (cmd == BIOCSRTIMEOUT32) {
+				tv32 = (struct timeval32 *)addr;
+				tv = &tv64;
+				tv->tv_sec = tv32->tv_sec;
+				tv->tv_usec = tv32->tv_usec;
+			} else
+#endif
+				tv = (struct timeval *)addr;
 
 			/*
 			 * Subtract 1 tick from tvtohz() since this isn't
@@ -1182,11 +1285,31 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 	 * Get read timeout.
 	 */
 	case BIOCGRTIMEOUT:
+#ifdef COMPAT_FREEBSD32
+	case BIOCGRTIMEOUT32:
+#endif
 		{
-			struct timeval *tv = (struct timeval *)addr;
+			struct timeval *tv;
+#ifdef COMPAT_FREEBSD32
+			struct timeval32 *tv32;
+			struct timeval tv64;
+
+			if (cmd == BIOCGRTIMEOUT32)
+				tv = &tv64;
+			else
+#endif
+				tv = (struct timeval *)addr;
 
 			tv->tv_sec = d->bd_rtout / hz;
 			tv->tv_usec = (d->bd_rtout % hz) * tick;
+#ifdef COMPAT_FREEBSD32
+			if (cmd == BIOCGRTIMEOUT32) {
+				tv32 = (struct timeval32 *)addr;
+				tv32->tv_sec = tv->tv_sec;
+				tv32->tv_usec = tv->tv_usec;
+			}
+#endif
+
 			break;
 		}
 
@@ -1369,7 +1492,19 @@ bpf_setf(struct bpf_d *d, struct bpf_program *fp, u_long cmd)
 #ifdef BPF_JITTER
 	bpf_jit_filter *ofunc;
 #endif
+#ifdef COMPAT_FREEBSD32
+	struct bpf_program32 *fp32;
+	struct bpf_program fp_swab;
 
+	if (cmd == BIOCSETWF32 || cmd == BIOCSETF32 || cmd == BIOCSETFNR32) {
+		fp32 = (struct bpf_program32 *)fp;
+		fp_swab.bf_len = fp32->bf_len;
+		fp_swab.bf_insns = (struct bpf_insn *)(uintptr_t)fp32->bf_insns;
+		fp = &fp_swab;
+		if (cmd == BIOCSETWF32)
+			cmd = BIOCSETWF;
+	}
+#endif
 	if (cmd == BIOCSETWF) {
 		old = d->bd_wfilter;
 		wfilter = 1;
@@ -1771,6 +1906,9 @@ catchpacket(struct bpf_d *d, u_char *pkt, u_int pktlen, u_int snaplen,
     struct timeval *tv)
 {
 	struct bpf_hdr hdr;
+#ifdef COMPAT_FREEBSD32
+	struct bpf_hdr32 hdr32;
+#endif
 	int totlen, curlen;
 	int hdrlen = d->bd_bif->bif_hdrlen;
 	int do_wakeup = 0;
@@ -1809,7 +1947,12 @@ catchpacket(struct bpf_d *d, u_char *pkt, u_int pktlen, u_int snaplen,
 	 * buffer is considered immutable by the buffer model, try to rotate
 	 * the buffer and wakeup pending processes.
 	 */
-	curlen = BPF_WORDALIGN(d->bd_slen);
+#ifdef COMPAT_FREEBSD32
+	if (d->bd_compat32)
+		curlen = BPF_WORDALIGN32(d->bd_slen);
+	else
+#endif
+		curlen = BPF_WORDALIGN(d->bd_slen);
 	if (curlen + totlen > d->bd_bufsize || !bpf_canwritebuf(d)) {
 		if (d->bd_fbuf == NULL) {
 			/*
@@ -1831,6 +1974,22 @@ catchpacket(struct bpf_d *d, u_char *pkt, u_int pktlen, u_int snaplen,
 		 * reader should be woken up.
 		 */
 		do_wakeup = 1;
+#ifdef COMPAT_FREEBSD32
+	/*
+	 * If this is a 32-bit stream, then stick a 32-bit header at the
+	 * front and copy the data into the buffer.
+	 */
+	if (d->bd_compat32) {
+		bzero(&hdr32, sizeof(hdr32));
+		hdr32.bh_tstamp.tv_sec = tv->tv_sec;
+		hdr32.bh_tstamp.tv_usec = tv->tv_usec;
+		hdr32.bh_datalen = pktlen;
+		hdr32.bh_hdrlen = hdrlen;
+		hdr.bh_caplen = hdr32.bh_caplen = totlen - hdrlen;
+		bpf_append_bytes(d, d->bd_sbuf, curlen, &hdr32, sizeof(hdr32));
+		goto copy;
+	}
+#endif
 
 	/*
 	 * Append the bpf header.  Note we append the actual header size, but
@@ -1846,6 +2005,9 @@ catchpacket(struct bpf_d *d, u_char *pkt, u_int pktlen, u_int snaplen,
 	/*
 	 * Copy the packet data into the store buffer and update its length.
 	 */
+#ifdef COMPAT_FREEBSD32
+ copy:
+#endif
 	(*cpfn)(d, d->bd_sbuf, curlen + hdrlen, pkt, hdr.bh_caplen);
 	d->bd_slen = curlen + totlen;
 
diff --git a/sys/net/bpfdesc.h b/sys/net/bpfdesc.h
index 5784763ed8a..03cb20dd519 100644
--- a/sys/net/bpfdesc.h
+++ b/sys/net/bpfdesc.h
@@ -97,6 +97,7 @@ struct bpf_d {
 	u_int64_t	bd_wfcount;	/* number of packets that matched write filter */
 	u_int64_t	bd_wdcount;	/* number of packets dropped during a write */
 	u_int64_t	bd_zcopy;	/* number of zero copy operations */
+	u_char		bd_compat32;	/* 32-bit stream on LP64 system */
 };
 
 /* Values for bd_state */

From 36f8c7ebef4a5a3d7f459309427531981324b689 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 17:38:53 +0000
Subject: [PATCH 252/532] Make hash, type and ulimit available via execve().

These are specified by POSIX but are not special builtins, and therefore
need to be available via execve() and utilities like time, nohup, xargs.
(Note that hash was moved from the XSI option to the base in the 2008
standard.)

Like most of the POSIX "regular builtin commands", these need to be executed
in a shell environment for full functionality, although they may still be of
some use outside one.

Unlike the POSIX special and regular builtin commands, POSIX does not
require these to be found before a PATH search, although that could be an
oversight.

Like some of the utilities already provided by usr.bin/alias, these may lead
to confusing results when invoked from csh(1).
---
 share/man/man1/builtin.1 | 6 +++---
 usr.bin/alias/Makefile   | 3 +++
 2 files changed, 6 insertions(+), 3 deletions(-)

diff --git a/share/man/man1/builtin.1 b/share/man/man1/builtin.1
index 12ab0a83841..3f2dc1786f0 100644
--- a/share/man/man1/builtin.1
+++ b/share/man/man1/builtin.1
@@ -245,7 +245,7 @@ but are implemented as scripts using a builtin command of the same name.
 .It Ic getopts Ta No** Ta \&No Ta Yes
 .It Ic glob Ta \&No Ta Yes Ta \&No
 .It Ic goto Ta \&No Ta Yes Ta \&No
-.It Ic hash Ta \&No Ta \&No Ta Yes
+.It Ic hash Ta No** Ta \&No Ta Yes
 .It Ic hashstat Ta \&No Ta Yes Ta \&No
 .It Ic history Ta \&No Ta Yes Ta \&No
 .It Ic hup Ta \&No Ta Yes Ta \&No
@@ -290,8 +290,8 @@ but are implemented as scripts using a builtin command of the same name.
 .It Ic times Ta \&No Ta \&No Ta Yes
 .It Ic trap Ta \&No Ta \&No Ta Yes
 .It Ic true Ta Yes Ta \&No Ta Yes
-.It Ic type Ta \&No Ta \&No Ta Yes
-.It Ic ulimit Ta \&No Ta \&No Ta Yes
+.It Ic type Ta No** Ta \&No Ta Yes
+.It Ic ulimit Ta No** Ta \&No Ta Yes
 .It Ic umask Ta No** Ta Yes Ta Yes
 .It Ic unalias Ta No** Ta Yes Ta Yes
 .It Ic uncomplete Ta \&No Ta Yes Ta \&No
diff --git a/usr.bin/alias/Makefile b/usr.bin/alias/Makefile
index 53df717f66c..474499ff802 100644
--- a/usr.bin/alias/Makefile
+++ b/usr.bin/alias/Makefile
@@ -10,8 +10,11 @@ LINKS=	${BINDIR}/alias ${BINDIR}/bg \
 	${BINDIR}/alias ${BINDIR}/fc \
 	${BINDIR}/alias ${BINDIR}/fg \
 	${BINDIR}/alias ${BINDIR}/getopts \
+	${BINDIR}/alias ${BINDIR}/hash \
 	${BINDIR}/alias ${BINDIR}/jobs \
 	${BINDIR}/alias ${BINDIR}/read \
+	${BINDIR}/alias ${BINDIR}/type \
+	${BINDIR}/alias ${BINDIR}/ulimit \
 	${BINDIR}/alias ${BINDIR}/umask \
 	${BINDIR}/alias ${BINDIR}/unalias \
 	${BINDIR}/alias ${BINDIR}/wait

From 6dbd88581dade8b64ff1600605150d727662aa3c Mon Sep 17 00:00:00 2001
From: Michael Tuexen <tuexen@FreeBSD.org>
Date: Sun, 25 Apr 2010 19:00:37 +0000
Subject: [PATCH 253/532] Sending a FWDTSN chunk should not affect the retran
 count.

MFC after: 3 days.
---
 sys/netinet/sctp_output.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/netinet/sctp_output.c b/sys/netinet/sctp_output.c
index c7a806c0d9e..47590c76963 100644
--- a/sys/netinet/sctp_output.c
+++ b/sys/netinet/sctp_output.c
@@ -8973,7 +8973,7 @@ sctp_chunk_retransmission(struct sctp_inpcb *inp,
 		/* (void)SCTP_GETTIME_TIMEVAL(&chk->whoTo->last_sent_time); */
 		*cnt_out += 1;
 		chk->sent = SCTP_DATAGRAM_SENT;
-		sctp_ucount_decr(asoc->sent_queue_retran_cnt);
+		/* sctp_ucount_decr(asoc->sent_queue_retran_cnt); */
 		if (fwd_tsn == 0) {
 			return (0);
 		} else {

From 0d2e1c3e39359263dc3342b3105db2bec383e3cd Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 25 Apr 2010 20:40:45 +0000
Subject: [PATCH 254/532] Clearing a page table entry's accessed bit (PG_A) and
 setting the page's PG_REFERENCED flag in pmap_protect() can't really be
 justified. In contrast to pmap_remove() or pmap_remove_all(), the mapping is
 not being destroyed, so the notion that the page was accessed is not lost.
 Moreover, clearing the page table entry's accessed bit and setting the page's
 PG_REFERENCED flag can throw off the page daemon's activity count
 calculation.  Finally, in my tests, I found that 15% of the atomic memory
 operations being performed by pmap_protect() were only to clear PG_A, and not
 change protection.  This could, by itself, be fixed, but I don't see the
 point given the above argument.

Remove a comment from pmap_protect_pde() that is no longer meaningful
after the above change.
---
 sys/amd64/amd64/pmap.c | 29 ++++++-----------------------
 sys/i386/i386/pmap.c   | 28 ++++++----------------------
 2 files changed, 12 insertions(+), 45 deletions(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 9c7d8b088a9..6cd10b476a8 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -2833,18 +2833,9 @@ retry:
 	if (oldpde & PG_MANAGED) {
 		eva = sva + NBPDR;
 		for (va = sva, m = PHYS_TO_VM_PAGE(oldpde & PG_PS_FRAME);
-		    va < eva; va += PAGE_SIZE, m++) {
-			/*
-			 * In contrast to the analogous operation on a 4KB page
-			 * mapping, the mapping's PG_A flag is not cleared and
-			 * the page's PG_REFERENCED flag is not set.  The
-			 * reason is that pmap_demote_pde() expects that a 2MB
-			 * page mapping with a stored page table page has PG_A
-			 * set.
-			 */
+		    va < eva; va += PAGE_SIZE, m++)
 			if ((oldpde & (PG_M | PG_RW)) == (PG_M | PG_RW))
 				vm_page_dirty(m);
-		}
 	}
 	if ((prot & VM_PROT_WRITE) == 0)
 		newpde &= ~(PG_RW | PG_M);
@@ -2953,23 +2944,15 @@ retry:
 			obits = pbits = *pte;
 			if ((pbits & PG_V) == 0)
 				continue;
-			if (pbits & PG_MANAGED) {
-				m = NULL;
-				if (pbits & PG_A) {
+
+			if ((prot & VM_PROT_WRITE) == 0) {
+				if ((pbits & (PG_MANAGED | PG_M | PG_RW)) ==
+				    (PG_MANAGED | PG_M | PG_RW)) {
 					m = PHYS_TO_VM_PAGE(pbits & PG_FRAME);
-					vm_page_flag_set(m, PG_REFERENCED);
-					pbits &= ~PG_A;
-				}
-				if ((pbits & (PG_M | PG_RW)) == (PG_M | PG_RW)) {
-					if (m == NULL)
-						m = PHYS_TO_VM_PAGE(pbits &
-						    PG_FRAME);
 					vm_page_dirty(m);
 				}
-			}
-
-			if ((prot & VM_PROT_WRITE) == 0)
 				pbits &= ~(PG_RW | PG_M);
+			}
 			if ((prot & VM_PROT_EXECUTE) == 0)
 				pbits |= pg_nx;
 
diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c
index 1ce6454ffbb..d8b9686e738 100644
--- a/sys/i386/i386/pmap.c
+++ b/sys/i386/i386/pmap.c
@@ -2955,18 +2955,9 @@ retry:
 	if (oldpde & PG_MANAGED) {
 		eva = sva + NBPDR;
 		for (va = sva, m = PHYS_TO_VM_PAGE(oldpde & PG_PS_FRAME);
-		    va < eva; va += PAGE_SIZE, m++) {
-			/*
-			 * In contrast to the analogous operation on a 4KB page
-			 * mapping, the mapping's PG_A flag is not cleared and
-			 * the page's PG_REFERENCED flag is not set.  The
-			 * reason is that pmap_demote_pde() expects that a 2/4MB
-			 * page mapping with a stored page table page has PG_A
-			 * set.
-			 */
+		    va < eva; va += PAGE_SIZE, m++)
 			if ((oldpde & (PG_M | PG_RW)) == (PG_M | PG_RW))
 				vm_page_dirty(m);
-		}
 	}
 	if ((prot & VM_PROT_WRITE) == 0)
 		newpde &= ~(PG_RW | PG_M);
@@ -3074,22 +3065,15 @@ retry:
 			obits = pbits = *pte;
 			if ((pbits & PG_V) == 0)
 				continue;
-			if (pbits & PG_MANAGED) {
-				m = NULL;
-				if (pbits & PG_A) {
+
+			if ((prot & VM_PROT_WRITE) == 0) {
+				if ((pbits & (PG_MANAGED | PG_M | PG_RW)) ==
+				    (PG_MANAGED | PG_M | PG_RW)) {
 					m = PHYS_TO_VM_PAGE(pbits & PG_FRAME);
-					vm_page_flag_set(m, PG_REFERENCED);
-					pbits &= ~PG_A;
-				}
-				if ((pbits & (PG_M | PG_RW)) == (PG_M | PG_RW)) {
-					if (m == NULL)
-						m = PHYS_TO_VM_PAGE(pbits & PG_FRAME);
 					vm_page_dirty(m);
 				}
-			}
-
-			if ((prot & VM_PROT_WRITE) == 0)
 				pbits &= ~(PG_RW | PG_M);
+			}
 #ifdef PAE
 			if ((prot & VM_PROT_EXECUTE) == 0)
 				pbits |= pg_nx;

From 593e925a73be1cfa1eb0a9904b7da7493b279d6f Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sun, 25 Apr 2010 20:43:19 +0000
Subject: [PATCH 255/532] sh: Use stalloc for arith variable names.

This is simpler than the custom memory tracker I added earlier, and is also
needed by the dash arith code I plan to import.
---
 bin/sh/arith.y     |  4 ++++
 bin/sh/arith_lex.l | 26 +++-----------------------
 bin/sh/expand.c    |  4 +++-
 3 files changed, 10 insertions(+), 24 deletions(-)

diff --git a/bin/sh/arith.y b/bin/sh/arith.y
index 5db16336b22..28046bab8ed 100644
--- a/bin/sh/arith.y
+++ b/bin/sh/arith.y
@@ -287,7 +287,9 @@ arith_t
 arith(const char *s)
 {
 	arith_t result;
+	struct stackmark smark;
 
+	setstackmark(&smark);
 	arith_buf = arith_startbuf = s;
 
 	INTOFF;
@@ -295,6 +297,8 @@ arith(const char *s)
 	arith_lex_reset();	/* Reprime lex. */
 	INTON;
 
+	popstackmark(&smark);
+
 	return result;
 }
 
diff --git a/bin/sh/arith_lex.l b/bin/sh/arith_lex.l
index f0d9cb34c41..aede40fd93d 100644
--- a/bin/sh/arith_lex.l
+++ b/bin/sh/arith_lex.l
@@ -51,13 +51,6 @@ __FBSDID("$FreeBSD$");
 
 int yylex(void);
 
-struct varname
-{
-	struct varname *next;
-	char name[1];
-};
-static struct varname *varnames;
-
 #undef YY_INPUT
 #define YY_INPUT(buf,result,max) \
 	result = (*buf = *arith_buf++) ? 1 : YY_NULL;
@@ -87,14 +80,11 @@ static struct varname *varnames;
 			 * If variable doesn't exist, we should initialize
 			 * it to zero.
 			 */
-			struct varname *temp;
+			char *temp;
 			if (lookupvar(yytext) == NULL)
 				setvarsafe(yytext, "0", 0);
-			temp = ckmalloc(sizeof(struct varname) +
-			    strlen(yytext));
-			temp->next = varnames;
-			varnames = temp;
-			yylval.s_value = strcpy(temp->name, yytext);
+			temp = stalloc(strlen(yytext) + 1);
+			yylval.s_value = strcpy(temp, yytext);
 
 			return ARITH_VAR;
 		}
@@ -140,15 +130,5 @@ static struct varname *varnames;
 void
 arith_lex_reset(void)
 {
-	struct varname *name, *next;
-
 	YY_NEW_FILE;
-
-	name = varnames;
-	while (name != NULL) {
-		next = name->next;
-		ckfree(name);
-		name = next;
-	}
-	varnames = NULL;
 }
diff --git a/bin/sh/expand.c b/bin/sh/expand.c
index ab885ac1c7b..7d987375ef3 100644
--- a/bin/sh/expand.c
+++ b/bin/sh/expand.c
@@ -360,7 +360,7 @@ removerecordregions(int endoff)
 void
 expari(int flag)
 {
-	char *p, *start;
+	char *p, *q, *start;
 	arith_t result;
 	int begoff;
 	int quotes = flag & (EXP_FULL | EXP_CASE | EXP_REDIR);
@@ -398,7 +398,9 @@ expari(int flag)
 	removerecordregions(begoff);
 	if (quotes)
 		rmescapes(p+2);
+	q = grabstackstr(expdest);
 	result = arith(p+2);
+	ungrabstackstr(q, expdest);
 	fmtstr(p, DIGITS(result), ARITH_FORMAT_STR, result);
 	while (*p++)
 		;

From f6a9d95ba388b7c47a9293217ca1bc68c225ac57 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Sun, 25 Apr 2010 22:01:32 +0000
Subject: [PATCH 256/532] Set USB_DEBUG like the other platforms, I had turned
 it off to test the build before committing r207077.

Spotted by:	marius
---
 sys/amd64/conf/GENERIC | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/amd64/conf/GENERIC b/sys/amd64/conf/GENERIC
index ddfb1ab387b..b5a1930dbba 100644
--- a/sys/amd64/conf/GENERIC
+++ b/sys/amd64/conf/GENERIC
@@ -281,7 +281,7 @@ device		firmware	# firmware assist module
 device		bpf		# Berkeley packet filter
 
 # USB support
-#options 	USB_DEBUG	# enable debug msgs
+options 	USB_DEBUG	# enable debug msgs
 device		uhci		# UHCI PCI->USB interface
 device		ohci		# OHCI PCI->USB interface
 device		ehci		# EHCI PCI->USB interface (USB 2.0)

From c28d264aa0054839491f4965c9c84b5c08852355 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Sun, 25 Apr 2010 23:18:02 +0000
Subject: [PATCH 257/532] Incremental reduction of delta with head_page_lock_2
 branch

- replace modification of pmap resident_count with pmap_resident_count_{inc,dec}
- the pv list is protected by the pmap lock, but in several cases we are relying
  on the vm page queue mutex, move pv_va read under the pmap lock
---
 sys/amd64/amd64/pmap.c | 63 +++++++++++++++++++++++++-----------------
 1 file changed, 38 insertions(+), 25 deletions(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 6cd10b476a8..c4f1e992a0d 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -404,6 +404,21 @@ pmap_pte(pmap_t pmap, vm_offset_t va)
 	return (pmap_pde_to_pte(pde, va));
 }
 
+static __inline void
+pmap_resident_count_inc(pmap_t pmap, int count)
+{
+
+	PMAP_LOCK_ASSERT(pmap, MA_OWNED);
+	pmap->pm_stats.resident_count += count;
+}
+
+static __inline void
+pmap_resident_count_dec(pmap_t pmap, int count)
+{
+
+	PMAP_LOCK_ASSERT(pmap, MA_OWNED);
+	pmap->pm_stats.resident_count -= count;
+}
 
 PMAP_INLINE pt_entry_t *
 vtopte(vm_offset_t va)
@@ -1477,6 +1492,7 @@ _pmap_unwire_pte_hold(pmap_t pmap, vm_offset_t va, vm_page_t m,
     vm_page_t *free)
 {
 
+	PMAP_LOCK_ASSERT(pmap, MA_OWNED);
 	/*
 	 * unmap the page table page
 	 */
@@ -1496,7 +1512,7 @@ _pmap_unwire_pte_hold(pmap_t pmap, vm_offset_t va, vm_page_t m,
 		pd = pmap_pde(pmap, va);
 		*pd = 0;
 	}
-	--pmap->pm_stats.resident_count;
+	pmap_resident_count_dec(pmap, 1);
 	if (m->pindex < NUPDE) {
 		/* We just released a PT, unhold the matching PD */
 		vm_page_t pdpg;
@@ -1613,6 +1629,7 @@ _pmap_allocpte(pmap_t pmap, vm_pindex_t ptepindex, int flags)
 	    (flags & (M_NOWAIT | M_WAITOK)) == M_WAITOK,
 	    ("_pmap_allocpte: flags is neither M_NOWAIT nor M_WAITOK"));
 
+	PMAP_LOCK_ASSERT(pmap, MA_OWNED);
 	/*
 	 * Allocate a page table page.
 	 */
@@ -1730,7 +1747,7 @@ _pmap_allocpte(pmap_t pmap, vm_pindex_t ptepindex, int flags)
 		*pd = VM_PAGE_TO_PHYS(m) | PG_U | PG_RW | PG_V | PG_A | PG_M;
 	}
 
-	pmap->pm_stats.resident_count++;
+	pmap_resident_count_inc(pmap, 1);
 
 	return (m);
 }
@@ -2037,7 +2054,7 @@ pmap_collect(pmap_t locked_pmap, struct vpgqueues *vpq)
 				PMAP_LOCK(pmap);
 			else if (pmap != locked_pmap && !PMAP_TRYLOCK(pmap))
 				continue;
-			pmap->pm_stats.resident_count--;
+			pmap_resident_count_dec(pmap, 1);
 			pde = pmap_pde(pmap, va);
 			KASSERT((*pde & PG_PS) == 0, ("pmap_collect: found"
 			    " a 2mpage in page %p's pv list", m));
@@ -2438,7 +2455,7 @@ pmap_demote_pde(pmap_t pmap, pd_entry_t *pde, vm_offset_t va)
 			return (FALSE);
 		}
 		if (va < VM_MAXUSER_ADDRESS)
-			pmap->pm_stats.resident_count++;
+			pmap_resident_count_inc(pmap, 1);
 	}
 	mptepa = VM_PAGE_TO_PHYS(mpte);
 	firstpte = (pt_entry_t *)PHYS_TO_DMAP(mptepa);
@@ -2530,7 +2547,7 @@ pmap_remove_pde(pmap_t pmap, pd_entry_t *pdq, vm_offset_t sva,
 	 */
 	if (oldpde & PG_G)
 		pmap_invalidate_page(kernel_pmap, sva);
-	pmap->pm_stats.resident_count -= NBPDR / PAGE_SIZE;
+	pmap_resident_count_dec(pmap, NBPDR / PAGE_SIZE);
 	if (oldpde & PG_MANAGED) {
 		pvh = pa_to_pvh(oldpde & PG_PS_FRAME);
 		pmap_pvh_free(pvh, pmap, sva);
@@ -2553,7 +2570,7 @@ pmap_remove_pde(pmap_t pmap, pd_entry_t *pdq, vm_offset_t sva,
 		mpte = pmap_lookup_pt_page(pmap, sva);
 		if (mpte != NULL) {
 			pmap_remove_pt_page(pmap, mpte);
-			pmap->pm_stats.resident_count--;
+			pmap_resident_count_dec(pmap, 1);
 			KASSERT(mpte->wire_count == NPTEPG,
 			    ("pmap_remove_pde: pte page wire count error"));
 			mpte->wire_count = 0;
@@ -2584,7 +2601,7 @@ pmap_remove_pte(pmap_t pmap, pt_entry_t *ptq, vm_offset_t va,
 	 */
 	if (oldpte & PG_G)
 		pmap_invalidate_page(kernel_pmap, va);
-	pmap->pm_stats.resident_count -= 1;
+	pmap_resident_count_dec(pmap, 1);
 	if (oldpte & PG_MANAGED) {
 		m = PHYS_TO_VM_PAGE(oldpte & PG_FRAME);
 		if ((oldpte & (PG_M | PG_RW)) == (PG_M | PG_RW))
@@ -2786,7 +2803,7 @@ pmap_remove_all(vm_page_t m)
 	while ((pv = TAILQ_FIRST(&m->md.pv_list)) != NULL) {
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
-		pmap->pm_stats.resident_count--;
+		pmap_resident_count_dec(pmap, 1);
 		pde = pmap_pde(pmap, pv->pv_va);
 		KASSERT((*pde & PG_PS) == 0, ("pmap_remove_all: found"
 		    " a 2mpage in page %p's pv list", m));
@@ -3192,7 +3209,7 @@ pmap_enter(pmap_t pmap, vm_offset_t va, vm_prot_t access, vm_page_t m,
 			     " va: 0x%lx", va));
 		}
 	} else
-		pmap->pm_stats.resident_count++;
+		pmap_resident_count_inc(pmap, 1);
 
 	/*
 	 * Enter on the PV list if part of our managed memory.
@@ -3326,7 +3343,7 @@ pmap_enter_pde(pmap_t pmap, vm_offset_t va, vm_page_t m, vm_prot_t prot)
 	/*
 	 * Increment counters.
 	 */
-	pmap->pm_stats.resident_count += NBPDR / PAGE_SIZE;
+	pmap_resident_count_inc(pmap, NBPDR / PAGE_SIZE);
 
 	/*
 	 * Map the superpage.
@@ -3480,7 +3497,7 @@ pmap_enter_quick_locked(pmap_t pmap, vm_offset_t va, vm_page_t m,
 	/*
 	 * Increment counters
 	 */
-	pmap->pm_stats.resident_count++;
+	pmap_resident_count_inc(pmap, 1);
 
 	pa = VM_PAGE_TO_PHYS(m) | pmap_cache_bits(m->md.pat_mode, 0);
 	if ((prot & VM_PROT_EXECUTE) == 0)
@@ -3585,8 +3602,7 @@ pmap_object_init_pt(pmap_t pmap, vm_offset_t addr, vm_object_t object,
 			if ((*pde & PG_V) == 0) {
 				pde_store(pde, pa | PG_PS | PG_M | PG_A |
 				    PG_U | PG_RW | PG_V);
-				pmap->pm_stats.resident_count += NBPDR /
-				    PAGE_SIZE;
+				pmap_resident_count_inc(pmap, NBPDR / PAGE_SIZE);
 				pmap_pde_mappings++;
 			} else {
 				/* Continue on if the PDE is already valid. */
@@ -3653,8 +3669,6 @@ out:
 	PMAP_UNLOCK(pmap);
 }
 
-
-
 /*
  *	Copy the range specified by src_addr/len
  *	from the source map to the range dst_addr/len
@@ -3729,8 +3743,7 @@ pmap_copy(pmap_t dst_pmap, pmap_t src_pmap, vm_offset_t dst_addr, vm_size_t len,
 			    pmap_pv_insert_pde(dst_pmap, addr, srcptepaddr &
 			    PG_PS_FRAME))) {
 				*pde = srcptepaddr & ~PG_W;
-				dst_pmap->pm_stats.resident_count +=
-				    NBPDR / PAGE_SIZE;
+				pmap_resident_count_inc(dst_pmap, NBPDR / PAGE_SIZE);
 			} else
 				dstmpde->wire_count--;
 			continue;
@@ -3773,7 +3786,7 @@ pmap_copy(pmap_t dst_pmap, pmap_t src_pmap, vm_offset_t dst_addr, vm_size_t len,
 					 */
 					*dst_pte = ptetemp & ~(PG_W | PG_M |
 					    PG_A);
-					dst_pmap->pm_stats.resident_count++;
+					pmap_resident_count_inc(dst_pmap, 1);
 	 			} else {
 					free = NULL;
 					if (pmap_unwire_pte_hold(dst_pmap,
@@ -4046,7 +4059,7 @@ pmap_remove_pages(pmap_t pmap)
 				pv_entry_count--;
 				pc->pc_map[field] |= bitmask;
 				if ((tpte & PG_PS) != 0) {
-					pmap->pm_stats.resident_count -= NBPDR / PAGE_SIZE;
+					pmap_resident_count_dec(pmap, NBPDR / PAGE_SIZE);
 					pvh = pa_to_pvh(tpte & PG_PS_FRAME);
 					TAILQ_REMOVE(&pvh->pv_list, pv, pv_list);
 					if (TAILQ_EMPTY(&pvh->pv_list)) {
@@ -4057,7 +4070,7 @@ pmap_remove_pages(pmap_t pmap)
 					mpte = pmap_lookup_pt_page(pmap, pv->pv_va);
 					if (mpte != NULL) {
 						pmap_remove_pt_page(pmap, mpte);
-						pmap->pm_stats.resident_count--;
+						pmap_resident_count_dec(pmap, 1);
 						KASSERT(mpte->wire_count == NPTEPG,
 						    ("pmap_remove_pages: pte page wire count error"));
 						mpte->wire_count = 0;
@@ -4065,7 +4078,7 @@ pmap_remove_pages(pmap_t pmap)
 						atomic_subtract_int(&cnt.v_wire_count, 1);
 					}
 				} else {
-					pmap->pm_stats.resident_count--;
+					pmap_resident_count_dec(pmap, 1);
 					TAILQ_REMOVE(&m->md.pv_list, pv, pv_list);
 					if (TAILQ_EMPTY(&m->md.pv_list)) {
 						pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
@@ -4223,9 +4236,9 @@ pmap_remove_write(vm_page_t m)
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
 	TAILQ_FOREACH_SAFE(pv, &pvh->pv_list, pv_list, next_pv) {
-		va = pv->pv_va;
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
+		va = pv->pv_va;
 		pde = pmap_pde(pmap, va);
 		if ((*pde & PG_RW) != 0)
 			(void)pmap_demote_pde(pmap, pde, va);
@@ -4281,9 +4294,9 @@ pmap_ts_referenced(vm_page_t m)
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
 	TAILQ_FOREACH_SAFE(pv, &pvh->pv_list, pv_list, pvn) {
-		va = pv->pv_va;
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
+		va = pv->pv_va;
 		pde = pmap_pde(pmap, va);
 		oldpde = *pde;
 		if ((oldpde & PG_A) != 0) {
@@ -4353,9 +4366,9 @@ pmap_clear_modify(vm_page_t m)
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
 	TAILQ_FOREACH_SAFE(pv, &pvh->pv_list, pv_list, next_pv) {
-		va = pv->pv_va;
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
+		va = pv->pv_va;
 		pde = pmap_pde(pmap, va);
 		oldpde = *pde;
 		if ((oldpde & PG_RW) != 0) {
@@ -4418,9 +4431,9 @@ pmap_clear_reference(vm_page_t m)
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
 	TAILQ_FOREACH_SAFE(pv, &pvh->pv_list, pv_list, next_pv) {
-		va = pv->pv_va;
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
+		va = pv->pv_va;
 		pde = pmap_pde(pmap, va);
 		oldpde = *pde;
 		if ((oldpde & PG_A) != 0) {

From cd4d97b43950ebdd458b7fe88767b57e46e76fea Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Sun, 25 Apr 2010 23:51:05 +0000
Subject: [PATCH 258/532] missed pv access before pmap lock

---
 sys/amd64/amd64/pmap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index c4f1e992a0d..d07292dd2c2 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -2793,9 +2793,9 @@ pmap_remove_all(vm_page_t m)
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
 	while ((pv = TAILQ_FIRST(&pvh->pv_list)) != NULL) {
-		va = pv->pv_va;
 		pmap = PV_PMAP(pv);
 		PMAP_LOCK(pmap);
+		va = pv->pv_va;
 		pde = pmap_pde(pmap, va);
 		(void)pmap_demote_pde(pmap, pde, va);
 		PMAP_UNLOCK(pmap);

From 22b94883101e91c954afcaddecea718a64ac7dda Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Mon, 26 Apr 2010 11:58:15 +0000
Subject: [PATCH 259/532] Mark ATA channel as idle on timeout in non-ATA_CAM
 mode. This should fix possible duplicate request completion.

Submitted by:	mjacob
---
 sys/dev/ata/ata-queue.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/ata/ata-queue.c b/sys/dev/ata/ata-queue.c
index a3b1c4e4435..4aef47190cb 100644
--- a/sys/dev/ata/ata-queue.c
+++ b/sys/dev/ata/ata-queue.c
@@ -513,9 +513,9 @@ ata_timeout(struct ata_request *request)
 	request->flags |= ATA_R_TIMEOUT;
 	if (ch->dma.unload)
 	    ch->dma.unload(request);
-#ifdef ATA_CAM
 	ch->running = NULL;
 	ch->state = ATA_IDLE;
+#ifdef ATA_CAM
 	ata_cam_end_transaction(ch->dev, request);
 #endif
 	mtx_unlock(&ch->state_mtx);

From 699f853bed973d189b2f31161aec01e92c28acfd Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Mon, 26 Apr 2010 12:03:55 +0000
Subject: [PATCH 260/532] MFp4: Move PI_TAG_ABLE check from ada driver to ATA
 XPT.

---
 sys/cam/ata/ata_da.c  |  4 ----
 sys/cam/ata/ata_xpt.c | 29 +++++++++++++++++++----------
 2 files changed, 19 insertions(+), 14 deletions(-)

diff --git a/sys/cam/ata/ata_da.c b/sys/cam/ata/ata_da.c
index 6cb4f2e5509..6eff691de90 100644
--- a/sys/cam/ata/ata_da.c
+++ b/sys/cam/ata/ata_da.c
@@ -686,14 +686,10 @@ adaregister(struct cam_periph *periph, void *arg)
 	else
 		softc->quirks = ADA_Q_NONE;
 
-	/* Check if the SIM does not want queued commands */
 	bzero(&cpi, sizeof(cpi));
 	xpt_setup_ccb(&cpi.ccb_h, periph->path, CAM_PRIORITY_NONE);
 	cpi.ccb_h.func_code = XPT_PATH_INQ;
 	xpt_action((union ccb *)&cpi);
-	if (cpi.ccb_h.status != CAM_REQ_CMP ||
-	    (cpi.hba_inquiry & PI_TAG_ABLE) == 0)
-		softc->flags &= ~ADA_FLAG_CAN_NCQ;
 
 	TASK_INIT(&softc->sysctl_task, 0, adasysctlinit, periph);
 
diff --git a/sys/cam/ata/ata_xpt.c b/sys/cam/ata/ata_xpt.c
index d5e0b32a49b..3bf9c1ef1df 100644
--- a/sys/cam/ata/ata_xpt.c
+++ b/sys/cam/ata/ata_xpt.c
@@ -766,6 +766,7 @@ noerror:
 	}
 	case PROBE_IDENTIFY:
 	{
+		struct ccb_pathinq cpi;
 		int16_t *ptr;
 
 		ident_buf = &softc->ident_data;
@@ -840,16 +841,24 @@ noerror:
 		ata_find_quirk(path->device);
 		if (path->device->mintags != 0 &&
 		    path->bus->sim->max_tagged_dev_openings != 0) {
-			/* Report SIM which tags are allowed. */
-			bzero(&cts, sizeof(cts));
-			xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
-			cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
-			cts.type = CTS_TYPE_CURRENT_SETTINGS;
-			cts.xport_specific.sata.tags = path->device->maxtags;
-			cts.xport_specific.sata.valid = CTS_SATA_VALID_TAGS;
-			xpt_action((union ccb *)&cts);
-			/* Reconfigure queues for tagged queueing. */
-			xpt_start_tags(path);
+			/* Check if the SIM does not want queued commands. */
+			bzero(&cpi, sizeof(cpi));
+			xpt_setup_ccb(&cpi.ccb_h, path, CAM_PRIORITY_NONE);
+			cpi.ccb_h.func_code = XPT_PATH_INQ;
+			xpt_action((union ccb *)&cpi);
+			if (cpi.ccb_h.status == CAM_REQ_CMP &&
+			    (cpi.hba_inquiry & PI_TAG_ABLE)) {
+				/* Report SIM which tags are allowed. */
+				bzero(&cts, sizeof(cts));
+				xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+				cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
+				cts.type = CTS_TYPE_CURRENT_SETTINGS;
+				cts.xport_specific.sata.tags = path->device->maxtags;
+				cts.xport_specific.sata.valid = CTS_SATA_VALID_TAGS;
+				xpt_action((union ccb *)&cts);
+				/* Reconfigure queues for tagged queueing. */
+				xpt_start_tags(path);
+			}
 		}
 		ata_device_transport(path);
 		PROBE_SET_ACTION(softc, PROBE_SETMODE);

From 7d11e744c12499be2ef89e2b0a670bc1847789c8 Mon Sep 17 00:00:00 2001
From: Lawrence Stewart <lstewart@FreeBSD.org>
Date: Mon, 26 Apr 2010 13:48:22 +0000
Subject: [PATCH 261/532] - Rework the underlying ALQ storage to be a circular
 buffer, which amongst other   things allows variable length messages to be
 easily supported.

- Extend KPI with alq_writen() and alq_getn() to support variable length
  messages, which is enabled at ALQ creation time depending on the
  arguments passed to alq_open(). Also add variants of alq_open() and
  alq_post() that accept a flags argument. The KPI is still fully
  backwards compatible and shouldn't require any change in ALQ consumers
  unless they wish to utilise the new features.

- Introduce the ALQ_NOACTIVATE and ALQ_ORDERED flags to allow ALQ consumers
  to have more control over IO scheduling and resource acquisition
  respectively.

- Strengthen invariants checking.

- Document ALQ changes in ALQ(9) man page.

Sponsored by:	FreeBSD Foundation
Reviewed by:	gnn, jeff, rpaulo, rwatson
MFC after:	1 month
---
 share/man/man9/alq.9 | 343 +++++++++++++++++++------
 sys/kern/kern_alq.c  | 579 +++++++++++++++++++++++++++++++++++--------
 sys/sys/alq.h        |  63 +++--
 3 files changed, 782 insertions(+), 203 deletions(-)

diff --git a/share/man/man9/alq.9 b/share/man/man9/alq.9
index ec6ef2688c3..65c2d07d93e 100644
--- a/share/man/man9/alq.9
+++ b/share/man/man9/alq.9
@@ -1,7 +1,13 @@
 .\"
 .\" Copyright (c) 2003 Hiten Pandya <hmp@FreeBSD.org>
+.\" Copyright (c) 2009-2010 The FreeBSD Foundation
 .\" All rights reserved.
 .\"
+.\" Portions of this software were developed at the Centre for Advanced
+.\" Internet Architectures, Swinburne University of Technology, Melbourne,
+.\" Australia by Lawrence Stewart under sponsorship from the FreeBSD
+.\" Foundation.
+.\"
 .\" Redistribution and use in source and binary forms, with or without
 .\" modification, are permitted provided that the following conditions
 .\" are met:
@@ -25,21 +31,34 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd May 16, 2003
+.Dd April 26, 2010
 .Dt ALQ 9
 .Os
 .Sh NAME
 .Nm alq ,
+.Nm alq_open_flags ,
 .Nm alq_open ,
+.Nm alq_writen ,
 .Nm alq_write ,
 .Nm alq_flush ,
 .Nm alq_close ,
+.Nm alq_getn ,
 .Nm alq_get ,
+.Nm alq_post_flags ,
 .Nm alq_post
 .Nd Asynchronous Logging Queues
 .Sh SYNOPSIS
 .In sys/alq.h
 .Ft int
+.Fo alq_open_flags
+.Fa "struct alq **app"
+.Fa "const char *file"
+.Fa "struct ucred *cred"
+.Fa "int cmode"
+.Fa "int size"
+.Fa "int flags"
+.Fc
+.Ft int
 .Fo alq_open
 .Fa "struct alq **app"
 .Fa "const char *file"
@@ -49,19 +68,25 @@
 .Fa "int count"
 .Fc
 .Ft int
-.Fn alq_write "struct alq *alq" "void *data" "int waitok"
+.Fn alq_writen "struct alq *alq" "void *data" "int len" "int flags"
+.Ft int
+.Fn alq_write "struct alq *alq" "void *data" "int flags"
 .Ft void
 .Fn alq_flush "struct alq *alq"
 .Ft void
 .Fn alq_close "struct alq *alq"
 .Ft struct ale *
-.Fn alq_get "struct alq *alq" "int waitok"
+.Fn alq_getn "struct alq *alq" "int len" "int flags"
+.Ft struct ale *
+.Fn alq_get "struct alq *alq" "int flags"
+.Ft void
+.Fn alq_post_flags "struct alq *alq" "struct ale *ale" "int flags"
 .Ft void
 .Fn alq_post "struct alq *alq" "struct ale *ale"
 .Sh DESCRIPTION
 The
 .Nm
-facility provides an asynchronous fixed length recording
+facility provides an asynchronous fixed or variable length recording
 mechanism, known as Asynchronous Logging Queues.
 It can record to any
 .Xr vnode 9 ,
@@ -81,26 +106,37 @@ is defined as
 which has the following members:
 .Bd -literal -offset indent
 struct ale {
-	struct ale	*ae_next;	/* Next Entry */
-	char		*ae_data;	/* Entry buffer */
-	int		ae_flags;	/* Entry flags */
+	intptr_t	ae_bytesused;	/* # bytes written to ALE. */
+	char		*ae_data;	/* Write ptr. */
+	int		ae_pad;		/* Unused, compat. */
 };
 .Ed
 .Pp
-The
-.Va ae_flags
-field is for internal use, clients of the
+An
 .Nm
-interface should not modify this field.
-Behaviour is undefined if this field is modified.
+can be created in either fixed or variable length mode.
+A variable length
+.Nm
+accommodates writes of varying length using
+.Fn alq_writen
+and
+.Fn alq_getn .
+A fixed length
+.Nm
+accommodates a fixed number of writes using
+.Fn alq_write
+and
+.Fn alq_get ,
+each of fixed size (set at queue creation time).
+Fixed length mode is deprecated in favour of variable length mode.
 .Sh FUNCTIONS
 The
-.Fn alq_open
-function creates a new logging queue.
+.Fn alq_open_flags
+function creates a new variable length asynchronous logging queue.
 The
 .Fa file
-argument is the name of the file to open for logging; if the file does not
-yet exist,
+argument is the name of the file to open for logging.
+If the file does not yet exist,
 .Fn alq_open
 will attempt to create it.
 The
@@ -112,33 +148,99 @@ as the requested creation mode, to be used if the file will be created by
 Consumers of this API may wish to pass
 .Dv ALQ_DEFAULT_CMODE ,
 a default creation mode suitable for most applications.
-The argument
-.Fa cred
-specifies the credentials to use when opening and performing I/O on the file.
-The size of each entry in the queue is determined by
-.Fa size .
 The
+.Fa cred
+argument specifies the credentials to use when opening and performing I/O on the file.
+The
+.Fa size
+argument sets the size (in bytes) of the underlying queue.
+The ALQ_ORDERED flag may be passed in via
+.Fa flags
+to indicate that the ordering of writer threads waiting for a busy
+.Nm
+to free up resources should be preserved.
+.Pp
+The deprecated
+.Fn alq_open
+function is implemented as a wrapper around
+.Fn alq_open_flags
+to provide backwards compatibility to consumers that have not been updated to
+utilise the newer
+.Fn alq_open_flags
+function.
+It passes all arguments through to
+.Fn alq_open_flags
+untouched except for
+.Fa size
+and
+.Fa count ,
+and sets
+.Fa flags
+to 0.
+To create a variable length mode
+.Nm ,
+the
+.Fa size
+argument should be set to the size (in bytes) of the underlying queue and the
 .Fa count
-argument determines the number of items to be stored in the
-asynchronous queue over an approximate period of a disk
-write operation.
+argument should be set to 0.
+To create a fixed length mode
+.Nm ,
+the
+.Fa size
+argument should be set to the size (in bytes) of each write and the
+.Fa count
+argument should be set to the number of
+.Fa size
+byte chunks to reserve capacity for.
 .Pp
 The
-.Fn alq_write
+.Fn alq_writen
 function writes
+.Fa len
+bytes from
 .Fa data
-to the designated queue,
+to the designated variable length mode queue
 .Fa alq .
-In the event that
-.Fn alq_write
-could not write the entry immediately, and
+If
+.Fn alq_writen
+could not write the entry immediately and
 .Dv ALQ_WAITOK
-is passed to
-.Fa waitok ,
-then
+is set in
+.Fa flags ,
+the function will be allowed to
+.Xr msleep_spin 9
+with the
+.Dq Li alqwnord
+or
+.Dq Li alqwnres
+wait message.
+A write will automatically schedule the queue
+.Fa alq
+to be flushed to disk.
+This behaviour can be controlled by passing ALQ_NOACTIVATE via
+.Fa flags
+to indicate that the write should not schedule
+.Fa alq
+to be flushed to disk.
+.Pp
+The deprecated
 .Fn alq_write
-will be allowed to
-.Xr tsleep 9 .
+function is implemented as a wrapper around
+.Fn alq_writen
+to provide backwards compatibility to consumers that have not been updated to
+utilise variable length mode queues.
+The function will write
+.Fa size
+bytes of data (where
+.Fa size
+was specified at queue creation time) from the
+.Fa data
+buffer to the
+.Fa alq .
+Note that it is an error to call
+.Fn alq_write
+on a variable length mode queue.
 .Pp
 The
 .Fn alq_flush
@@ -146,61 +248,136 @@ function is used for flushing
 .Fa alq
 to the log medium that was passed to
 .Fn alq_open .
+If
+.Fa alq
+has data to flush and is not already in the process of being flushed, the
+function will block doing IO.
+Otherwise, the function will return immediately.
 .Pp
 The
 .Fn alq_close
-function will close the asynchronous logging queue,
-.Fa alq ,
+function will close the asynchronous logging queue
+.Fa alq
 and flush all pending write requests to the log medium.
 It will free all resources that were previously allocated.
 .Pp
 The
-.Fn alq_get
-function returns the next available asynchronous logging entry
-from the queue,
-.Fa alq .
-This function leaves the queue in a locked state, until a subsequent
+.Fn alq_getn
+function returns an asynchronous log entry from
+.Fa alq ,
+initialised to point at a buffer capable of receiving
+.Fa len
+bytes of data.
+This function leaves
+.Fa alq
+in a locked state, until a subsequent
 .Fn alq_post
+or
+.Fn alq_post_flags
 call is made.
-In the event that
-.Fn alq_get
-could not retrieve an entry immediately, it will
-.Xr tsleep 9
+If
+.Fn alq_getn
+could not obtain
+.Fa len
+bytes of buffer immediately and
+.Dv ALQ_WAITOK
+is set in
+.Fa flags ,
+the function will be allowed to
+.Xr msleep_spin 9
 with the
-.Dq Li alqget
+.Dq Li alqgnord
+or
+.Dq Li alqgnres
 wait message.
+The caller can choose to write less than
+.Fa len
+bytes of data to the returned asynchronous log entry by setting the entry's
+ae_bytesused field to the number of bytes actually written.
+This must be done prior to calling
+.Fn alq_post .
+.Pp
+The deprecated
+.Fn alq_get
+function is implemented as a wrapper around
+.Fn alq_getn
+to provide backwards compatibility to consumers that have not been updated to
+utilise variable length mode queues.
+The asynchronous log entry returned will be initialised to point at a buffer
+capable of receiving
+.Fa size
+bytes of data (where
+.Fa size
+was specified at queue creation time).
+Note that it is an error to call
+.Fn alq_get
+on a variable length mode queue.
+.Pp
+The
+.Fn alq_post_flags
+function schedules the asynchronous log entry
+.Fa ale
+(obtained from
+.Fn alq_getn
+or
+.Fn alq_get )
+for writing to
+.Fa alq .
+The ALQ_NOACTIVATE flag may be passed in via
+.Fa flags
+to indicate that the queue should not be immediately scheduled to be flushed to
+disk.
+This function leaves
+.Fa alq
+in an unlocked state.
 .Pp
 The
 .Fn alq_post
-function schedules the asynchronous logging entry,
-.Fa ale ,
-which is retrieved using the
-.Fn alq_get
-function,
-for writing to the asynchronous logging queue,
-.Fa alq .
-This function leaves the queue,
-.Fa alq ,
-in an unlocked state.
+function is implemented as a wrapper around
+.Fn alq_post_flags
+to provide backwards compatibility to consumers that have not been updated to
+utilise the newer
+.Fn alq_post_flags
+function.
+It simply passes all arguments through to
+.Fn alq_post_flags
+untouched, and sets
+.Fa flags
+to 0.
 .Sh IMPLEMENTATION NOTES
 The
+.Fn alq_writen
+and
 .Fn alq_write
-function is a wrapper around the
+functions both perform a
+.Xr bcopy 3
+from the supplied
+.Fa data
+buffer into the underlying
+.Nm
+buffer.
+Performance critical code paths may wish to consider using
+.Fn alq_getn
+(variable length queues) or
+.Fn alq_get
+(fixed length queues) to avoid the extra memory copy. Note that a queue
+remains locked between calls to
+.Fn alq_getn
+or
 .Fn alq_get
 and
 .Fn alq_post
-functions; by using these functions separately, a call
-to
-.Fn bcopy
-can be avoided for performance critical code paths.
+or
+.Fn alq_post_flags ,
+so this method of writing to a queue is unsuitable for situations where the
+time between calls may be substantial.
 .Sh LOCKING
-Each asynchronous queue is protected by a spin mutex.
+Each asynchronous logging queue is protected by a spin mutex.
 .Pp
 Functions
-.Fn alq_flush ,
-.Fn alq_open
+.Fn alq_flush
 and
-.Fn alq_post
+.Fn alq_open
 may attempt to acquire an internal sleep mutex, and should
 consequently not be used in contexts where sleeping is
 not allowed.
@@ -214,32 +391,36 @@ if it fails to open
 or else it returns 0.
 .Pp
 The
+.Fn alq_writen
+and
 .Fn alq_write
-function returns
+functions return
 .Er EWOULDBLOCK
 if
 .Dv ALQ_NOWAIT
-was provided as a value to
-.Fa waitok
-and either the queue is full, or when the system is shutting down.
+was set in
+.Fa flags
+and either the queue is full or the system is shutting down.
 .Pp
 The
+.Fn alq_getn
+and
 .Fn alq_get
-function returns
-.Dv NULL ,
+functions return
+.Dv NULL
 if
 .Dv ALQ_NOWAIT
-was provided as a value to
-.Fa waitok
-and either the queue is full, or when the system is shutting down.
+was set in
+.Fa flags
+and either the queue is full or the system is shutting down.
 .Pp
 NOTE: invalid arguments to non-void functions will result in
 undefined behaviour.
 .Sh SEE ALSO
-.Xr syslog 3 ,
-.Xr kthread 9 ,
+.Xr kproc 9 ,
 .Xr ktr 9 ,
-.Xr tsleep 9 ,
+.Xr msleep_spin 9 ,
+.Xr syslog 3 ,
 .Xr vnode 9
 .Sh HISTORY
 The
@@ -250,7 +431,11 @@ Asynchronous Logging Queues (ALQ) facility first appeared in
 The
 .Nm
 facility was written by
-.An Jeffrey Roberson Aq jeff@FreeBSD.org .
+.An Jeffrey Roberson Aq jeff@FreeBSD.org
+and extended by
+.An Lawrence Stewart Aq lstewart@freebsd.org .
 .Pp
 This manual page was written by
-.An Hiten Pandya Aq hmp@FreeBSD.org .
+.An Hiten Pandya Aq hmp@FreeBSD.org
+and revised by
+.An Lawrence Stewart Aq lstewart@freebsd.org .
diff --git a/sys/kern/kern_alq.c b/sys/kern/kern_alq.c
index 6722d9a4db1..98e6de885cf 100644
--- a/sys/kern/kern_alq.c
+++ b/sys/kern/kern_alq.c
@@ -55,16 +55,23 @@ __FBSDID("$FreeBSD$");
 
 /* Async. Logging Queue */
 struct alq {
+	char	*aq_entbuf;		/* Buffer for stored entries */
 	int	aq_entmax;		/* Max entries */
 	int	aq_entlen;		/* Entry length */
-	char	*aq_entbuf;		/* Buffer for stored entries */
+	int	aq_freebytes;		/* Bytes available in buffer */
+	int	aq_buflen;		/* Total length of our buffer */
+	int	aq_writehead;		/* Location for next write */
+	int	aq_writetail;		/* Flush starts at this location */
+	int	aq_wrapearly;		/* # bytes left blank at end of buf */
 	int	aq_flags;		/* Queue flags */
+	int	aq_waiters;		/* Num threads waiting for resources
+					 * NB: Used as a wait channel so must
+					 * not be first field in the alq struct
+					 */
+	struct	ale	aq_getpost;	/* ALE for use by get/post */
 	struct mtx	aq_mtx;		/* Queue lock */
 	struct vnode	*aq_vp;		/* Open vnode handle */
 	struct ucred	*aq_cred;	/* Credentials of the opening thread */
-	struct ale	*aq_first;	/* First ent */
-	struct ale	*aq_entfree;	/* First free ent */
-	struct ale	*aq_entvalid;	/* First ent valid for writing */
 	LIST_ENTRY(alq)	aq_act;		/* List of active queues */
 	LIST_ENTRY(alq)	aq_link;	/* List of all queues */
 };
@@ -73,10 +80,14 @@ struct alq {
 #define	AQ_ACTIVE	0x0002		/* on the active list */
 #define	AQ_FLUSHING	0x0004		/* doing IO */
 #define	AQ_SHUTDOWN	0x0008		/* Queue no longer valid */
+#define	AQ_ORDERED	0x0010		/* Queue enforces ordered writes */
+#define	AQ_LEGACY	0x0020		/* Legacy queue (fixed length writes) */
 
 #define	ALQ_LOCK(alq)	mtx_lock_spin(&(alq)->aq_mtx)
 #define	ALQ_UNLOCK(alq)	mtx_unlock_spin(&(alq)->aq_mtx)
 
+#define HAS_PENDING_DATA(alq) ((alq)->aq_freebytes != (alq)->aq_buflen)
+
 static MALLOC_DEFINE(M_ALD, "ALD", "ALD");
 
 /*
@@ -205,7 +216,7 @@ ald_daemon(void)
 		needwakeup = alq_doio(alq);
 		ALQ_UNLOCK(alq);
 		if (needwakeup)
-			wakeup(alq);
+			wakeup_one(alq);
 		ALD_LOCK();
 	}
 
@@ -252,6 +263,20 @@ alq_shutdown(struct alq *alq)
 	/* Stop any new writers. */
 	alq->aq_flags |= AQ_SHUTDOWN;
 
+	/*
+	 * If the ALQ isn't active but has unwritten data (possible if
+	 * the ALQ_NOACTIVATE flag has been used), explicitly activate the
+	 * ALQ here so that the pending data gets flushed by the ald_daemon.
+	 */
+	if (!(alq->aq_flags & AQ_ACTIVE) && HAS_PENDING_DATA(alq)) {
+		alq->aq_flags |= AQ_ACTIVE;
+		ALQ_UNLOCK(alq);
+		ALD_LOCK();
+		ald_activate(alq);
+		ALD_UNLOCK();
+		ALQ_LOCK(alq);
+	}
+
 	/* Drain IO */
 	while (alq->aq_flags & AQ_ACTIVE) {
 		alq->aq_flags |= AQ_WANTED;
@@ -271,7 +296,6 @@ alq_destroy(struct alq *alq)
 	alq_shutdown(alq);
 
 	mtx_destroy(&alq->aq_mtx);
-	free(alq->aq_first, M_ALD);
 	free(alq->aq_entbuf, M_ALD);
 	free(alq, M_ALD);
 }
@@ -287,46 +311,54 @@ alq_doio(struct alq *alq)
 	struct vnode *vp;
 	struct uio auio;
 	struct iovec aiov[2];
-	struct ale *ale;
-	struct ale *alstart;
 	int totlen;
 	int iov;
 	int vfslocked;
+	int wrapearly;
+
+	KASSERT((HAS_PENDING_DATA(alq)), ("%s: queue empty!", __func__));
 
 	vp = alq->aq_vp;
 	td = curthread;
 	totlen = 0;
-	iov = 0;
-
-	alstart = ale = alq->aq_entvalid;
-	alq->aq_entvalid = NULL;
+	iov = 1;
+	wrapearly = alq->aq_wrapearly;
 
 	bzero(&aiov, sizeof(aiov));
 	bzero(&auio, sizeof(auio));
 
-	do {
-		if (aiov[iov].iov_base == NULL)
-			aiov[iov].iov_base = ale->ae_data;
-		aiov[iov].iov_len += alq->aq_entlen;
-		totlen += alq->aq_entlen;
-		/* Check to see if we're wrapping the buffer */
-		if (ale->ae_data + alq->aq_entlen != ale->ae_next->ae_data)
-			iov++;
-		ale->ae_flags &= ~AE_VALID;
-		ale = ale->ae_next;
-	} while (ale->ae_flags & AE_VALID);
+	/* Start the write from the location of our buffer tail pointer. */
+	aiov[0].iov_base = alq->aq_entbuf + alq->aq_writetail;
+
+	if (alq->aq_writetail < alq->aq_writehead) {
+		/* Buffer not wrapped. */
+		totlen = aiov[0].iov_len = alq->aq_writehead - alq->aq_writetail;
+	} else if (alq->aq_writehead == 0) {
+		/* Buffer not wrapped (special case to avoid an empty iov). */
+		totlen = aiov[0].iov_len = alq->aq_buflen - alq->aq_writetail -
+		    wrapearly;
+	} else {
+		/*
+		 * Buffer wrapped, requires 2 aiov entries:
+		 * - first is from writetail to end of buffer
+		 * - second is from start of buffer to writehead
+		 */
+		aiov[0].iov_len = alq->aq_buflen - alq->aq_writetail -
+		    wrapearly;
+		iov++;
+		aiov[1].iov_base = alq->aq_entbuf;
+		aiov[1].iov_len =  alq->aq_writehead;
+		totlen = aiov[0].iov_len + aiov[1].iov_len;
+	}
 
 	alq->aq_flags |= AQ_FLUSHING;
 	ALQ_UNLOCK(alq);
 
-	if (iov == 2 || aiov[iov].iov_base == NULL)
-		iov--;
-
 	auio.uio_iov = &aiov[0];
 	auio.uio_offset = 0;
 	auio.uio_segflg = UIO_SYSSPACE;
 	auio.uio_rw = UIO_WRITE;
-	auio.uio_iovcnt = iov + 1;
+	auio.uio_iovcnt = iov;
 	auio.uio_resid = totlen;
 	auio.uio_td = td;
 
@@ -350,8 +382,28 @@ alq_doio(struct alq *alq)
 	ALQ_LOCK(alq);
 	alq->aq_flags &= ~AQ_FLUSHING;
 
-	if (alq->aq_entfree == NULL)
-		alq->aq_entfree = alstart;
+	/* Adjust writetail as required, taking into account wrapping. */
+	alq->aq_writetail = (alq->aq_writetail + totlen + wrapearly) %
+	    alq->aq_buflen;
+	alq->aq_freebytes += totlen + wrapearly;
+
+	/*
+	 * If we just flushed part of the buffer which wrapped, reset the
+	 * wrapearly indicator.
+	 */
+	if (wrapearly)
+		alq->aq_wrapearly = 0;
+
+	/*
+	 * If we just flushed the buffer completely, reset indexes to 0 to
+	 * minimise buffer wraps.
+	 * This is also required to ensure alq_getn() can't wedge itself.
+	 */
+	if (!HAS_PENDING_DATA(alq))
+		alq->aq_writehead = alq->aq_writetail = 0;
+
+	KASSERT((alq->aq_writetail >= 0 && alq->aq_writetail < alq->aq_buflen),
+	    ("%s: aq_writetail < 0 || aq_writetail >= aq_buflen", __func__));
 
 	if (alq->aq_flags & AQ_WANTED) {
 		alq->aq_flags &= ~AQ_WANTED;
@@ -376,27 +428,27 @@ SYSINIT(ald, SI_SUB_LOCK, SI_ORDER_ANY, ald_startup, NULL);
 /*
  * Create the queue data structure, allocate the buffer, and open the file.
  */
+
 int
-alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
-    int size, int count)
+alq_open_flags(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
+    int size, int flags)
 {
 	struct thread *td;
 	struct nameidata nd;
-	struct ale *ale;
-	struct ale *alp;
 	struct alq *alq;
-	char *bufp;
-	int flags;
+	int oflags;
 	int error;
-	int i, vfslocked;
+	int vfslocked;
+
+	KASSERT((size > 0), ("%s: size <= 0", __func__));
 
 	*alqp = NULL;
 	td = curthread;
 
 	NDINIT(&nd, LOOKUP, NOFOLLOW | MPSAFE, UIO_SYSSPACE, file, td);
-	flags = FWRITE | O_NOFOLLOW | O_CREAT;
+	oflags = FWRITE | O_NOFOLLOW | O_CREAT;
 
-	error = vn_open_cred(&nd, &flags, cmode, 0, cred, NULL);
+	error = vn_open_cred(&nd, &oflags, cmode, 0, cred, NULL);
 	if (error)
 		return (error);
 
@@ -407,31 +459,20 @@ alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
 	VFS_UNLOCK_GIANT(vfslocked);
 
 	alq = malloc(sizeof(*alq), M_ALD, M_WAITOK|M_ZERO);
-	alq->aq_entbuf = malloc(count * size, M_ALD, M_WAITOK|M_ZERO);
-	alq->aq_first = malloc(sizeof(*ale) * count, M_ALD, M_WAITOK|M_ZERO);
 	alq->aq_vp = nd.ni_vp;
 	alq->aq_cred = crhold(cred);
-	alq->aq_entmax = count;
-	alq->aq_entlen = size;
-	alq->aq_entfree = alq->aq_first;
 
 	mtx_init(&alq->aq_mtx, "ALD Queue", NULL, MTX_SPIN|MTX_QUIET);
 
-	bufp = alq->aq_entbuf;
-	ale = alq->aq_first;
-	alp = NULL;
+	alq->aq_buflen = size;
+	alq->aq_entmax = 0;
+	alq->aq_entlen = 0;
 
-	/* Match up entries with buffers */
-	for (i = 0; i < count; i++) {
-		if (alp)
-			alp->ae_next = ale;
-		ale->ae_data = bufp;
-		alp = ale;
-		ale++;
-		bufp += size;
-	}
-
-	alp->ae_next = alq->aq_first;
+	alq->aq_freebytes = alq->aq_buflen;
+	alq->aq_entbuf = malloc(alq->aq_buflen, M_ALD, M_WAITOK|M_ZERO);
+	alq->aq_writehead = alq->aq_writetail = 0;
+	if (flags & ALQ_ORDERED)
+		alq->aq_flags |= AQ_ORDERED;
 
 	if ((error = ald_add(alq)) != 0) {
 		alq_destroy(alq);
@@ -443,77 +484,405 @@ alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
 	return (0);
 }
 
+int
+alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
+    int size, int count)
+{
+	int ret;
+
+	KASSERT((count >= 0), ("%s: count < 0", __func__));
+
+	if (count > 0) {
+		ret = alq_open_flags(alqp, file, cred, cmode, size*count, 0);
+		(*alqp)->aq_flags |= AQ_LEGACY;
+		(*alqp)->aq_entmax = count;
+		(*alqp)->aq_entlen = size;
+	} else
+		ret = alq_open_flags(alqp, file, cred, cmode, size, 0);
+
+	return (ret);
+}
+
+
 /*
  * Copy a new entry into the queue.  If the operation would block either
  * wait or return an error depending on the value of waitok.
  */
 int
-alq_write(struct alq *alq, void *data, int waitok)
+alq_writen(struct alq *alq, void *data, int len, int flags)
 {
-	struct ale *ale;
+	int activate, copy, ret;
+	void *waitchan;
 
-	if ((ale = alq_get(alq, waitok)) == NULL)
-		return (EWOULDBLOCK);
+	KASSERT((len > 0 && len <= alq->aq_buflen),
+	    ("%s: len <= 0 || len > aq_buflen", __func__));
 
-	bcopy(data, ale->ae_data, alq->aq_entlen);
-	alq_post(alq, ale);
-
-	return (0);
-}
-
-struct ale *
-alq_get(struct alq *alq, int waitok)
-{
-	struct ale *ale;
-	struct ale *aln;
-
-	ale = NULL;
+	activate = ret = 0;
+	copy = len;
+	waitchan = NULL;
 
 	ALQ_LOCK(alq);
 
-	/* Loop until we get an entry or we're shutting down */
-	while ((alq->aq_flags & AQ_SHUTDOWN) == 0 && 
-	    (ale = alq->aq_entfree) == NULL &&
-	    (waitok & ALQ_WAITOK)) {
-		alq->aq_flags |= AQ_WANTED;
-		msleep_spin(alq, &alq->aq_mtx, "alqget", 0);
+	/*
+	 * Fail to perform the write and return EWOULDBLOCK if:
+	 * - The message is larger than our underlying buffer.
+	 * - The ALQ is being shutdown.
+	 * - There is insufficient free space in our underlying buffer
+	 *   to accept the message and the user can't wait for space.
+	 * - There is insufficient free space in our underlying buffer
+	 *   to accept the message and the alq is inactive due to prior
+	 *   use of the ALQ_NOACTIVATE flag (which would lead to deadlock).
+	 */
+	if (len > alq->aq_buflen ||
+	    alq->aq_flags & AQ_SHUTDOWN ||
+	    (((flags & ALQ_NOWAIT) || (!(alq->aq_flags & AQ_ACTIVE) &&
+	    HAS_PENDING_DATA(alq))) && alq->aq_freebytes < len)) {
+		ALQ_UNLOCK(alq);
+		return (EWOULDBLOCK);
 	}
 
-	if (ale != NULL) {
-		aln = ale->ae_next;
-		if ((aln->ae_flags & AE_VALID) == 0)
-			alq->aq_entfree = aln;
+	/*
+	 * If we want ordered writes and there is already at least one thread
+	 * waiting for resources to become available, sleep until we're woken.
+	 */
+	if (alq->aq_flags & AQ_ORDERED && alq->aq_waiters > 0) {
+		KASSERT(!(flags & ALQ_NOWAIT),
+		    ("%s: ALQ_NOWAIT set but incorrectly ignored!", __func__));
+		alq->aq_waiters++;
+		msleep_spin(&alq->aq_waiters, &alq->aq_mtx, "alqwnord", 0);
+		alq->aq_waiters--;
+	}
+
+	/*
+	 * (ALQ_WAITOK && aq_freebytes < len) or aq_freebytes >= len, either
+	 * enter while loop and sleep until we have enough free bytes (former)
+	 * or skip (latter). If AQ_ORDERED is set, only 1 thread at a time will
+	 * be in this loop. Otherwise, multiple threads may be sleeping here
+	 * competing for ALQ resources.
+	 */
+	while (alq->aq_freebytes < len && !(alq->aq_flags & AQ_SHUTDOWN)) {
+		KASSERT(!(flags & ALQ_NOWAIT),
+		    ("%s: ALQ_NOWAIT set but incorrectly ignored!", __func__));
+		alq->aq_flags |= AQ_WANTED;
+		alq->aq_waiters++;
+		if (waitchan)
+			wakeup(waitchan);
+		msleep_spin(alq, &alq->aq_mtx, "alqwnres", 0);
+		alq->aq_waiters--;
+
+		/*
+		 * If we're the first thread to wake after an AQ_WANTED wakeup
+		 * but there isn't enough free space for us, we're going to loop
+		 * and sleep again. If there are other threads waiting in this
+		 * loop, schedule a wakeup so that they can see if the space
+		 * they require is available.
+		 */
+		if (alq->aq_waiters > 0 && !(alq->aq_flags & AQ_ORDERED) &&
+		    alq->aq_freebytes < len && !(alq->aq_flags & AQ_WANTED))
+			waitchan = alq;
 		else
-			alq->aq_entfree = NULL;
+			waitchan = NULL;
+	}
+
+	/*
+	 * If there are waiters, we need to signal the waiting threads after we
+	 * complete our work. The alq ptr is used as a wait channel for threads
+	 * requiring resources to be freed up. In the AQ_ORDERED case, threads
+	 * are not allowed to concurrently compete for resources in the above
+	 * while loop, so we use a different wait channel in this case.
+	 */
+	if (alq->aq_waiters > 0) {
+		if (alq->aq_flags & AQ_ORDERED)
+			waitchan = &alq->aq_waiters;
+		else
+			waitchan = alq;
 	} else
-		ALQ_UNLOCK(alq);
+		waitchan = NULL;
 
+	/* Bail if we're shutting down. */
+	if (alq->aq_flags & AQ_SHUTDOWN) {
+		ret = EWOULDBLOCK;
+		goto unlock;
+	}
 
-	return (ale);
-}
+	/*
+	 * If we need to wrap the buffer to accommodate the write,
+	 * we'll need 2 calls to bcopy.
+	 */
+	if ((alq->aq_buflen - alq->aq_writehead) < len)
+		copy = alq->aq_buflen - alq->aq_writehead;
 
-void
-alq_post(struct alq *alq, struct ale *ale)
-{
-	int activate;
+	/* Copy message (or part thereof if wrap required) to the buffer. */
+	bcopy(data, alq->aq_entbuf + alq->aq_writehead, copy);
+	alq->aq_writehead += copy;
 
-	ale->ae_flags |= AE_VALID;
+	if (alq->aq_writehead >= alq->aq_buflen) {
+		KASSERT((alq->aq_writehead == alq->aq_buflen),
+		    ("%s: alq->aq_writehead (%d) > alq->aq_buflen (%d)",
+		    __func__,
+		    alq->aq_writehead,
+		    alq->aq_buflen));
+		alq->aq_writehead = 0;
+	}
 
-	if (alq->aq_entvalid == NULL)
-		alq->aq_entvalid = ale;
+	if (copy != len) {
+		/*
+		 * Wrap the buffer by copying the remainder of our message
+		 * to the start of the buffer and resetting aq_writehead.
+		 */
+		bcopy(((uint8_t *)data)+copy, alq->aq_entbuf, len - copy);
+		alq->aq_writehead = len - copy;
+	}
 
-	if ((alq->aq_flags & AQ_ACTIVE) == 0) {
+	KASSERT((alq->aq_writehead >= 0 && alq->aq_writehead < alq->aq_buflen),
+	    ("%s: aq_writehead < 0 || aq_writehead >= aq_buflen", __func__));
+
+	alq->aq_freebytes -= len;
+
+	if (!(alq->aq_flags & AQ_ACTIVE) && !(flags & ALQ_NOACTIVATE)) {
 		alq->aq_flags |= AQ_ACTIVE;
 		activate = 1;
-	} else
-		activate = 0;
+	}
 
+	KASSERT((HAS_PENDING_DATA(alq)), ("%s: queue empty!", __func__));
+
+unlock:
 	ALQ_UNLOCK(alq);
+
 	if (activate) {
 		ALD_LOCK();
 		ald_activate(alq);
 		ALD_UNLOCK();
 	}
+
+	/* NB: We rely on wakeup_one waking threads in a FIFO manner. */
+	if (waitchan != NULL)
+		wakeup_one(waitchan);
+
+	return (ret);
+}
+
+int
+alq_write(struct alq *alq, void *data, int flags)
+{
+	/* Should only be called in fixed length message (legacy) mode. */
+	KASSERT((alq->aq_flags & AQ_LEGACY),
+	    ("%s: fixed length write on variable length queue", __func__));
+	return (alq_writen(alq, data, alq->aq_entlen, flags));
+}
+
+/*
+ * Retrieve a pointer for the ALQ to write directly into, avoiding bcopy.
+ */
+struct ale *
+alq_getn(struct alq *alq, int len, int flags)
+{
+	int contigbytes;
+	void *waitchan;
+
+	KASSERT((len > 0 && len <= alq->aq_buflen),
+	    ("%s: len <= 0 || len > alq->aq_buflen", __func__));
+
+	waitchan = NULL;
+
+	ALQ_LOCK(alq);
+
+	/*
+	 * Determine the number of free contiguous bytes.
+	 * We ensure elsewhere that if aq_writehead == aq_writetail because
+	 * the buffer is empty, they will both be set to 0 and therefore
+	 * aq_freebytes == aq_buflen and is fully contiguous.
+	 * If they are equal and the buffer is not empty, aq_freebytes will
+	 * be 0 indicating the buffer is full.
+	 */
+	if (alq->aq_writehead <= alq->aq_writetail)
+		contigbytes = alq->aq_freebytes;
+	else {
+		contigbytes = alq->aq_buflen - alq->aq_writehead;
+
+		if (contigbytes < len) {
+			/*
+			 * Insufficient space at end of buffer to handle a
+			 * contiguous write. Wrap early if there's space at
+			 * the beginning. This will leave a hole at the end
+			 * of the buffer which we will have to skip over when
+			 * flushing the buffer to disk.
+			 */
+			if (alq->aq_writetail >= len || flags & ALQ_WAITOK) {
+				/* Keep track of # bytes left blank. */
+				alq->aq_wrapearly = contigbytes;
+				/* Do the wrap and adjust counters. */
+				contigbytes = alq->aq_freebytes =
+				    alq->aq_writetail;
+				alq->aq_writehead = 0;
+			}
+		}
+	}
+
+	/*
+	 * Return a NULL ALE if:
+	 * - The message is larger than our underlying buffer.
+	 * - The ALQ is being shutdown.
+	 * - There is insufficient free space in our underlying buffer
+	 *   to accept the message and the user can't wait for space.
+	 * - There is insufficient free space in our underlying buffer
+	 *   to accept the message and the alq is inactive due to prior
+	 *   use of the ALQ_NOACTIVATE flag (which would lead to deadlock).
+	 */
+	if (len > alq->aq_buflen ||
+	    alq->aq_flags & AQ_SHUTDOWN ||
+	    (((flags & ALQ_NOWAIT) || (!(alq->aq_flags & AQ_ACTIVE) &&
+	    HAS_PENDING_DATA(alq))) && contigbytes < len)) {
+		ALQ_UNLOCK(alq);
+		return (NULL);
+	}
+
+	/*
+	 * If we want ordered writes and there is already at least one thread
+	 * waiting for resources to become available, sleep until we're woken.
+	 */
+	if (alq->aq_flags & AQ_ORDERED && alq->aq_waiters > 0) {
+		KASSERT(!(flags & ALQ_NOWAIT),
+		    ("%s: ALQ_NOWAIT set but incorrectly ignored!", __func__));
+		alq->aq_waiters++;
+		msleep_spin(&alq->aq_waiters, &alq->aq_mtx, "alqgnord", 0);
+		alq->aq_waiters--;
+	}
+
+	/*
+	 * (ALQ_WAITOK && contigbytes < len) or contigbytes >= len, either enter
+	 * while loop and sleep until we have enough contiguous free bytes
+	 * (former) or skip (latter). If AQ_ORDERED is set, only 1 thread at a
+	 * time will be in this loop. Otherwise, multiple threads may be
+	 * sleeping here competing for ALQ resources.
+	 */
+	while (contigbytes < len && !(alq->aq_flags & AQ_SHUTDOWN)) {
+		KASSERT(!(flags & ALQ_NOWAIT),
+		    ("%s: ALQ_NOWAIT set but incorrectly ignored!", __func__));
+		alq->aq_flags |= AQ_WANTED;
+		alq->aq_waiters++;
+		if (waitchan)
+			wakeup(waitchan);
+		msleep_spin(alq, &alq->aq_mtx, "alqgnres", 0);
+		alq->aq_waiters--;
+
+		if (alq->aq_writehead <= alq->aq_writetail)
+			contigbytes = alq->aq_freebytes;
+		else
+			contigbytes = alq->aq_buflen - alq->aq_writehead;
+
+		/*
+		 * If we're the first thread to wake after an AQ_WANTED wakeup
+		 * but there isn't enough free space for us, we're going to loop
+		 * and sleep again. If there are other threads waiting in this
+		 * loop, schedule a wakeup so that they can see if the space
+		 * they require is available.
+		 */
+		if (alq->aq_waiters > 0 && !(alq->aq_flags & AQ_ORDERED) &&
+		    contigbytes < len && !(alq->aq_flags & AQ_WANTED))
+			waitchan = alq;
+		else
+			waitchan = NULL;
+	}
+
+	/*
+	 * If there are waiters, we need to signal the waiting threads after we
+	 * complete our work. The alq ptr is used as a wait channel for threads
+	 * requiring resources to be freed up. In the AQ_ORDERED case, threads
+	 * are not allowed to concurrently compete for resources in the above
+	 * while loop, so we use a different wait channel in this case.
+	 */
+	if (alq->aq_waiters > 0) {
+		if (alq->aq_flags & AQ_ORDERED)
+			waitchan = &alq->aq_waiters;
+		else
+			waitchan = alq;
+	} else
+		waitchan = NULL;
+
+	/* Bail if we're shutting down. */
+	if (alq->aq_flags & AQ_SHUTDOWN) {
+		ALQ_UNLOCK(alq);
+		if (waitchan != NULL)
+			wakeup_one(waitchan);
+		return (NULL);
+	}
+
+	/*
+	 * If we are here, we have a contiguous number of bytes >= len
+	 * available in our buffer starting at aq_writehead.
+	 */
+	alq->aq_getpost.ae_data = alq->aq_entbuf + alq->aq_writehead;
+	alq->aq_getpost.ae_bytesused = len;
+
+	return (&alq->aq_getpost);
+}
+
+struct ale *
+alq_get(struct alq *alq, int flags)
+{
+	/* Should only be called in fixed length message (legacy) mode. */
+	KASSERT((alq->aq_flags & AQ_LEGACY),
+	    ("%s: fixed length get on variable length queue", __func__));
+	return (alq_getn(alq, alq->aq_entlen, flags));
+}
+
+void
+alq_post_flags(struct alq *alq, struct ale *ale, int flags)
+{
+	int activate;
+	void *waitchan;
+
+	activate = 0;
+
+	if (ale->ae_bytesused > 0) {
+		if (!(alq->aq_flags & AQ_ACTIVE) &&
+		    !(flags & ALQ_NOACTIVATE)) {
+			alq->aq_flags |= AQ_ACTIVE;
+			activate = 1;
+		}
+
+		alq->aq_writehead += ale->ae_bytesused;
+		alq->aq_freebytes -= ale->ae_bytesused;
+
+		/* Wrap aq_writehead if we filled to the end of the buffer. */
+		if (alq->aq_writehead == alq->aq_buflen)
+			alq->aq_writehead = 0;
+
+		KASSERT((alq->aq_writehead >= 0 &&
+		    alq->aq_writehead < alq->aq_buflen),
+		    ("%s: aq_writehead < 0 || aq_writehead >= aq_buflen",
+		    __func__));
+
+		KASSERT((HAS_PENDING_DATA(alq)), ("%s: queue empty!", __func__));
+	}
+
+	/*
+	 * If there are waiters, we need to signal the waiting threads after we
+	 * complete our work. The alq ptr is used as a wait channel for threads
+	 * requiring resources to be freed up. In the AQ_ORDERED case, threads
+	 * are not allowed to concurrently compete for resources in the
+	 * alq_getn() while loop, so we use a different wait channel in this case.
+	 */
+	if (alq->aq_waiters > 0) {
+		if (alq->aq_flags & AQ_ORDERED)
+			waitchan = &alq->aq_waiters;
+		else
+			waitchan = alq;
+	} else
+		waitchan = NULL;
+
+	ALQ_UNLOCK(alq);
+
+	if (activate) {
+		ALD_LOCK();
+		ald_activate(alq);
+		ALD_UNLOCK();
+	}
+
+	/* NB: We rely on wakeup_one waking threads in a FIFO manner. */
+	if (waitchan != NULL)
+		wakeup_one(waitchan);
 }
 
 void
@@ -523,16 +892,24 @@ alq_flush(struct alq *alq)
 
 	ALD_LOCK();
 	ALQ_LOCK(alq);
-	if (alq->aq_flags & AQ_ACTIVE) {
-		ald_deactivate(alq);
+
+	/*
+	 * Pull the lever iff there is data to flush and we're
+	 * not already in the middle of a flush operation.
+	 */
+	if (HAS_PENDING_DATA(alq) && !(alq->aq_flags & AQ_FLUSHING)) {
+		if (alq->aq_flags & AQ_ACTIVE)
+			ald_deactivate(alq);
+
 		ALD_UNLOCK();
 		needwakeup = alq_doio(alq);
 	} else
 		ALD_UNLOCK();
+
 	ALQ_UNLOCK(alq);
 
 	if (needwakeup)
-		wakeup(alq);
+		wakeup_one(alq);
 }
 
 /*
diff --git a/sys/sys/alq.h b/sys/sys/alq.h
index bc4961e094c..4a502d29684 100644
--- a/sys/sys/alq.h
+++ b/sys/sys/alq.h
@@ -1,7 +1,13 @@
 /*-
  * Copyright (c) 2002, Jeffrey Roberson <jeff@freebsd.org>
+ * Copyright (c) 2008-2009, Lawrence Stewart <lstewart@freebsd.org>
+ * Copyright (c) 2010, The FreeBSD Foundation
  * All rights reserved.
  *
+ * Portions of this software were developed at the Centre for Advanced
+ * Internet Architectures, Swinburne University of Technology, Melbourne,
+ * Australia by Lawrence Stewart under sponsorship from the FreeBSD Foundation.
+ *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
@@ -41,46 +47,47 @@ extern struct thread *ald_thread;
  * Async. Logging Entry
  */
 struct ale {
-	struct ale	*ae_next;	/* Next Entry */
-	char		*ae_data;	/* Entry buffer */
-	int		ae_flags;	/* Entry flags */
+	intptr_t	ae_bytesused;	/* # bytes written to ALE. */
+	char		*ae_data;	/* Write ptr. */
+	int		ae_pad;		/* Unused, compat. */
 };
 
-#define	AE_VALID	0x0001		/* Entry has valid data */
- 
-
-/* waitok options */
-#define	ALQ_NOWAIT	0x0001
-#define	ALQ_WAITOK	0x0002
+/* Flag options. */
+#define	ALQ_NOWAIT	0x0001	/* ALQ may not sleep. */
+#define	ALQ_WAITOK	0x0002	/* ALQ may sleep. */
+#define	ALQ_NOACTIVATE	0x0004	/* Don't activate ALQ after write. */
+#define	ALQ_ORDERED	0x0010	/* Maintain write ordering between threads. */
 
 /* Suggested mode for file creation. */
 #define	ALQ_DEFAULT_CMODE	0600
 
 /*
- * alq_open:  Creates a new queue
+ * alq_open_flags:  Creates a new queue
  *
  * Arguments:
  *	alq	Storage for a pointer to the newly created queue.
  *	file	The filename to open for logging.
  *	cred	Credential to authorize open and I/O with.
  *	cmode	Creation mode for file, if new.
- *	size	The size of each entry in the queue.
- *	count	The number of items in the buffer, this should be large enough
- *		to store items over the period of a disk write.
+ *	size	The size of the queue in bytes.
+ *	flags	ALQ_ORDERED
  * Returns:
  *	error from open or 0 on success
  */
 struct ucred;
-int alq_open(struct alq **, const char *file, struct ucred *cred, int cmode,
+int alq_open_flags(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
+	    int size, int flags);
+int alq_open(struct alq **alqp, const char *file, struct ucred *cred, int cmode,
 	    int size, int count);
 
 /*
- * alq_write:  Write data into the queue
+ * alq_writen:  Write data into the queue
  *
  * Arguments:
  *	alq	The queue we're writing to
  *	data	The entry to be recorded
- *	waitok	Are we permitted to wait?
+ *	len	The number of bytes to write from *data
+ *	flags	(ALQ_NOWAIT || ALQ_WAITOK), ALQ_NOACTIVATE
  *
  * Returns:
  *	EWOULDBLOCK if:
@@ -88,7 +95,8 @@ int alq_open(struct alq **, const char *file, struct ucred *cred, int cmode,
  *		The system is shutting down.
  *	0 on success.
  */
-int alq_write(struct alq *alq, void *data, int waitok);
+int alq_writen(struct alq *alq, void *data, int len, int flags);
+int alq_write(struct alq *alq, void *data, int flags);
 
 /*
  * alq_flush:  Flush the queue out to disk
@@ -101,27 +109,36 @@ void alq_flush(struct alq *alq);
 void alq_close(struct alq *alq);
 
 /*
- * alq_get:  Return an entry for direct access
+ * alq_getn:  Return an entry for direct access
  *
  * Arguments:
  *	alq	The queue to retrieve an entry from
- *	waitok	Are we permitted to wait?
+ *	len	Max number of bytes required
+ *	flags	(ALQ_NOWAIT || ALQ_WAITOK)
  *
  * Returns:
  *	The next available ale on success.
  *	NULL if:
- *		Waitok is ALQ_NOWAIT and the queue is full.
+ *		flags is ALQ_NOWAIT and the queue is full.
  *		The system is shutting down.
  *
  * This leaves the queue locked until a subsequent alq_post.
  */
-struct ale *alq_get(struct alq *alq, int waitok);
+struct ale *alq_getn(struct alq *alq, int len, int flags);
+struct ale *alq_get(struct alq *alq, int flags);
 
 /*
- * alq_post:  Schedule the ale retrieved by alq_get for writing.
+ * alq_post_flags:  Schedule the ale retrieved by alq_get/alq_getn for writing.
  *	alq	The queue to post the entry to.
  *	ale	An asynch logging entry returned by alq_get.
+ *	flags	ALQ_NOACTIVATE
  */
-void alq_post(struct alq *, struct ale *);
+void alq_post_flags(struct alq *alq, struct ale *ale, int flags);
+
+static __inline void
+alq_post(struct alq *alq, struct ale *ale)
+{
+	alq_post_flags(alq, ale, 0);
+}
 
 #endif	/* _SYS_ALQ_H_ */

From b13cc627c7787c8351eb248a6e63cd8318463dad Mon Sep 17 00:00:00 2001
From: Hajimu UMEMOTO <ume@FreeBSD.org>
Date: Mon, 26 Apr 2010 15:31:58 +0000
Subject: [PATCH 262/532] Better handling of ipv6_default_interface using
 net.inet6.ip6.use_defaultzone=1.  Now, it works IPv6 link-local unicast
 addresses as well as IPv6 link-local multicast addresses.

MFC after:	1 week
---
 etc/defaults/rc.conf     |  4 ++--
 etc/rc.d/routing         | 32 +++++++++++---------------------
 share/man/man5/rc.conf.5 |  4 ++--
 3 files changed, 15 insertions(+), 25 deletions(-)

diff --git a/etc/defaults/rc.conf b/etc/defaults/rc.conf
index ab50920dccb..63c94bbfc4c 100644
--- a/etc/defaults/rc.conf
+++ b/etc/defaults/rc.conf
@@ -470,8 +470,8 @@ route6d_flags=""		# Flags to IPv6 routing daemon.
 #ipv6_prefix_ed0="fec0:0000:0000:0001 fec0:0000:0000:0002"  # Examples for rtr.
 #ipv6_prefix_ep0="fec0:0000:0000:0003 fec0:0000:0000:0004"  # Examples for rtr.
 ipv6_default_interface="NO"	# Default output interface for scoped addrs.
-				# Now this works only for IPv6 link local
-				# multicast addrs.
+				# This works only with
+				# ipv6_gateway_enable="NO".
 rtsol_flags=""			# Flags to IPv6 router solicitation.
 rtsold_enable="NO"		# Set to YES to enable an IPv6 router
 				# solicitation daemon.
diff --git a/etc/rc.d/routing b/etc/rc.d/routing
index befceca0752..30242fc4333 100755
--- a/etc/rc.d/routing
+++ b/etc/rc.d/routing
@@ -210,34 +210,24 @@ static_inet6()
 		;;
 	esac
 
-	# Disallow unicast packets without outgoing scope identifiers,
-	# or route such packets to a "default" interface, if it is specified.
+	# Disallow link-local unicast packets without outgoing scope
+	# identifiers.  However, if you set "ipv6_default_interface",
+	# for the host case, you will allow to omit the identifiers.
+	# Under this configuration, the packets will go to the default
+	# interface.
 	route ${_action} -inet6 fe80:: -prefixlen 10 ::1 -reject
+	route ${_action} -inet6 ff02:: -prefixlen 16 ::1 -reject
 
 	case ${ipv6_default_interface} in
 	'')
-		route ${_action} -inet6 ff02:: -prefixlen 16 ::1 -reject
 		;;
 	*)
-		laddr=`network6_getladdr ${ipv6_default_interface}`
-		route ${_action} -inet6 ff02:: ${laddr} -prefixlen 16 -interface
-
-		# Disable installing the default interface with the
-		# case net.inet6.ip6.forwarding=0 and
-		# the interface with no ND6_IFF_ACCEPT_RTADV
-		# to avoid conflict between the default router list and
-		# the manual configured default route.
+		# Disable installing the default interface when we act
+		# as router to avoid conflict between the default
+		# router list and the manual configured default route.
 		if ! checkyesno ipv6_gateway_enable; then
-			ifconfig ${ipv6_default_interface} nd6 | \
-			while read proto options
-			do
-				case "${proto}:${options}" in
-				nd6:*ACCEPT_RTADV*)
-					ifconfig ${ipv6_default_interface} inet6 defaultif
-					break
-				;;
-				esac
-			done
+			ifconfig ${ipv6_default_interface} inet6 defaultif
+			sysctl net.inet6.ip6.use_defaultzone=1
 		fi
 		;;
 	esac
diff --git a/share/man/man5/rc.conf.5 b/share/man/man5/rc.conf.5
index e93205be539..734e7a375d7 100644
--- a/share/man/man5/rc.conf.5
+++ b/share/man/man5/rc.conf.5
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd April 14, 2010
+.Dd April 27, 2010
 .Dt RC.CONF 5
 .Os
 .Sh NAME
@@ -1370,7 +1370,7 @@ configured on that interface.
 If not set to
 .Dq Li NO ,
 this is the default output interface for scoped addresses.
-Now this works only for IPv6 link local multicast addresses.
+This works only with ipv6_gateway_enable="NO".
 .It Va ip6addrctl_enable
 .Pq Vt bool
 If set to

From 239bcd023794f3c1e4849bb3021dccff3295102b Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 26 Apr 2010 18:44:30 +0000
Subject: [PATCH 263/532] Skip the pseudo-devices found in Fujitsu Siemens
 PRIMEPOWER250.

---
 sys/sparc64/sparc64/nexus.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/sparc64/sparc64/nexus.c b/sys/sparc64/sparc64/nexus.c
index ee01aa86ae0..192251db30d 100644
--- a/sys/sparc64/sparc64/nexus.c
+++ b/sys/sparc64/sparc64/nexus.c
@@ -153,6 +153,7 @@ EARLY_DRIVER_MODULE(nexus, root, nexus_driver, nexus_devclass, 0, 0,
 MODULE_VERSION(nexus, 1);
 
 static const char *const nexus_excl_name[] = {
+	"FJSV,system",
 	"aliases",
 	"associations",
 	"chosen",
@@ -163,6 +164,7 @@ static const char *const nexus_excl_name[] = {
 	"openprom",
 	"options",
 	"packages",
+	"physical-memory",
 	"rsc",
 	"sgcn",
 	"todsg",

From cbafc5aff6932a818d3ad7084052173848a0db2c Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 26 Apr 2010 18:49:06 +0000
Subject: [PATCH 264/532] - Add a missing const. - Map the NS16550 found in
 Fujitsu Siemens PRIMEPOWER250 to PNP0501 as well.

---
 sys/sparc64/isa/isa.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/sparc64/isa/isa.c b/sys/sparc64/isa/isa.c
index 724b4a81f5d..9159cdac942 100644
--- a/sys/sparc64/isa/isa.c
+++ b/sys/sparc64/isa/isa.c
@@ -116,7 +116,7 @@ isa_init(device_t dev)
 }
 
 static const struct {
-	const char	*name;
+	const char	*const name;
 	uint32_t	id;
 } const ofw_isa_pnp_map[] = {
 	{ "SUNW,lomh",	0x0000ae4e }, /* SUN0000 */
@@ -126,6 +126,7 @@ static const struct {
 	{ "flashprom",	0x0100ae4e }, /* SUN0001 */
 	{ "parallel",	0x0104d041 }, /* PNP0401 */
 	{ "serial",	0x0105d041 }, /* PNP0501 */
+	{ "su",		0x0105d041 }, /* PNP0501 */
 	{ "i2c",	0x0200ae4e }, /* SUN0002 */
 	{ "rmc-comm",	0x0300ae4e }, /* SUN0003 */
 	{ "kb_ps2",	0x0303d041 }, /* PNP0303 */

From 2d857b9b194daa83f2a1d6aaffc3812ad4d932f5 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 26 Apr 2010 18:56:06 +0000
Subject: [PATCH 265/532] For the on-board interfaces found in Fujitsu SPARC64
 machines obtain the MAC address via OFW as well.

---
 sys/dev/bge/if_bge.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/bge/if_bge.c b/sys/dev/bge/if_bge.c
index 4333f2044b7..5a51d7988ea 100644
--- a/sys/dev/bge/if_bge.c
+++ b/sys/dev/bge/if_bge.c
@@ -519,7 +519,7 @@ bge_has_eaddr(struct bge_softc *sc)
 	 */
 	if (OF_getprop(ofw_bus_get_node(dev), SPARC64_OFW_SUBVENDOR,
 	    &subvendor, sizeof(subvendor)) == sizeof(subvendor) &&
-	    subvendor == SUN_VENDORID)
+	    (subvendor == FJTSU_VENDORID || subvendor == SUN_VENDORID))
 		return (0);
 	memset(buf, 0, sizeof(buf));
 	if (OF_package_to_path(ofw_bus_get_node(dev), buf, sizeof(buf)) > 0) {

From e2f198273cb2d837f3354cfd8ff2bc84a24c2447 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 26 Apr 2010 19:13:10 +0000
Subject: [PATCH 266/532] Add OF_getscsinitid(), a helper similar to
 OF_getetheraddr() but for obtaining the initiator ID to be used for SPI
 controllers from the Open Firmware device tree.

---
 sys/sparc64/include/ofw_machdep.h |  1 +
 sys/sparc64/sparc64/ofw_machdep.c | 15 ++++++++++++++-
 sys/sun4v/include/ofw_machdep.h   |  1 +
 3 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/sys/sparc64/include/ofw_machdep.h b/sys/sparc64/include/ofw_machdep.h
index 625b13137ad..658d9c7b0d2 100644
--- a/sys/sparc64/include/ofw_machdep.h
+++ b/sys/sparc64/include/ofw_machdep.h
@@ -36,6 +36,7 @@ typedef uint64_t cell_t;
 
 int  OF_decode_addr(phandle_t, int, int *, bus_addr_t *);
 void OF_getetheraddr(device_t, u_char *);
+u_int OF_getscsinitid(device_t);
 void cpu_shutdown(void *);
 int  ofw_entry(void *);
 void ofw_exit(void *);
diff --git a/sys/sparc64/sparc64/ofw_machdep.c b/sys/sparc64/sparc64/ofw_machdep.c
index 07f95e20e51..88ee0726725 100644
--- a/sys/sparc64/sparc64/ofw_machdep.c
+++ b/sys/sparc64/sparc64/ofw_machdep.c
@@ -1,6 +1,6 @@
 /*-
  * Copyright (c) 2001 by Thomas Moestl <tmm@FreeBSD.org>.
- * Copyright (c) 2005 - 2009 by Marius Strobl <marius@FreeBSD.org>.
+ * Copyright (c) 2005 - 2010 by Marius Strobl <marius@FreeBSD.org>.
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -68,6 +68,19 @@ OF_getetheraddr(device_t dev, u_char *addr)
 	bcopy(&idp.id_ether, addr, ETHER_ADDR_LEN);
 }
 
+u_int
+OF_getscsinitid(device_t dev)
+{
+	phandle_t node;
+	uint32_t id;
+
+	for (node = ofw_bus_get_node(dev); node != 0; node = OF_parent(node))
+		if (OF_getprop(node, "scsi-initiator-id", &id,
+		    sizeof(id)) > 0)
+			return (id);
+	return (7);
+}
+
 static __inline uint32_t
 phys_hi_mask_space(const char *bus, uint32_t phys_hi)
 {
diff --git a/sys/sun4v/include/ofw_machdep.h b/sys/sun4v/include/ofw_machdep.h
index 625b13137ad..658d9c7b0d2 100644
--- a/sys/sun4v/include/ofw_machdep.h
+++ b/sys/sun4v/include/ofw_machdep.h
@@ -36,6 +36,7 @@ typedef uint64_t cell_t;
 
 int  OF_decode_addr(phandle_t, int, int *, bus_addr_t *);
 void OF_getetheraddr(device_t, u_char *);
+u_int OF_getscsinitid(device_t);
 void cpu_shutdown(void *);
 int  ofw_entry(void *);
 void ofw_exit(void *);

From 18333f544e0fba84c6d9b33aac5e06e412468d5d Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Mon, 26 Apr 2010 20:05:48 +0000
Subject: [PATCH 267/532] Add a signal handler for SIGINT which removes output
 file when necessary.

While I'm there, move unlink_input() slightly down to after closing the
output file, in uncompression path.

MFC after:	2 weeks
---
 usr.bin/gzip/gzip.c | 25 +++++++++++++++++++++----
 1 file changed, 21 insertions(+), 4 deletions(-)

diff --git a/usr.bin/gzip/gzip.c b/usr.bin/gzip/gzip.c
index ef81442da6a..9503762a587 100644
--- a/usr.bin/gzip/gzip.c
+++ b/usr.bin/gzip/gzip.c
@@ -43,7 +43,6 @@ __RCSID("$FreeBSD$");
  *
  * TODO:
  *	- use mmap where possible
- *	- handle some signals better (remove outfile?)
  *	- make bzip2/compress -v/-t/-l support work as well as possible
  */
 
@@ -194,6 +193,7 @@ static	int	qflag;			/* quiet mode */
 static	int	rflag;			/* recursive mode */
 static	int	tflag;			/* test */
 static	int	vflag;			/* verbose mode */
+static	const char *remove_file = NULL;	/* file to be removed upon SIGINT */
 #else
 #define		qflag	0
 #define		tflag	0
@@ -231,6 +231,7 @@ static	void	usage(void);
 static	void	display_version(void);
 #ifndef SMALL
 static	void	display_license(void);
+static	void	sigint_handler(int);
 #endif
 static	const suffixes_t *check_suffix(char *, int);
 static	ssize_t	read_retry(int, void *, size_t);
@@ -300,11 +301,10 @@ main(int argc, char **argv)
 #endif
 	int ch;
 
-	/* XXX set up signals */
-
 #ifndef SMALL
 	if ((gzip = getenv("GZIP")) != NULL)
 		prepend_gzip(gzip, &argc, &argv);
+	signal(SIGINT, sigint_handler);
 #endif
 
 	/*
@@ -1171,6 +1171,15 @@ unlink_input(const char *file, const struct stat *sb)
 		return;
 	unlink(file);
 }
+
+static void
+sigint_handler(int signo __unused)
+{
+
+	if (remove_file != NULL)
+		unlink(remove_file);
+	exit(2);
+}
 #endif
 
 static const suffixes_t *
@@ -1257,6 +1266,9 @@ file_compress(char *file, char *outfile, size_t outsize)
 			fclose(stdin);
 			return -1;
 		}
+#ifndef SMALL
+		remove_file = outfile;
+#endif
 	} else
 		out = STDOUT_FILENO;
 
@@ -1288,6 +1300,7 @@ file_compress(char *file, char *outfile, size_t outsize)
 	}
 
 	copymodes(out, &isb, outfile);
+	remove_file = NULL;
 #endif
 	if (close(out) == -1)
 		maybe_warn("couldn't close output");
@@ -1424,6 +1437,9 @@ file_uncompress(char *file, char *outfile, size_t outsize)
 			maybe_warn("can't open %s", outfile);
 			goto lose;
 		}
+#ifndef SMALL
+		remove_file = outfile;
+#endif
 	} else
 		zfd = STDOUT_FILENO;
 
@@ -1555,11 +1571,12 @@ file_uncompress(char *file, char *outfile, size_t outsize)
 		unlink(outfile);
 		return -1;
 	}
-	unlink_input(file, &isb);
 #ifndef SMALL
 	copymodes(ofd, &isb, outfile);
+	remove_file = NULL;
 #endif
 	close(ofd);
+	unlink_input(file, &isb);
 	return size;
 
     unexpected_EOF:

From b641222476732f1f99d2362f093b79bbe088d764 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 26 Apr 2010 20:19:49 +0000
Subject: [PATCH 268/532] Don't bother enabling interrupts before we're ready
 to handle them. This prevents the firmware of Fujitsu Siemens PRIMEPOWER250,
 which both causes stray interrupts and erroneously enables interrupts at
 least when calling SUNW,set-trap-table, in the foot.

---
 sys/sparc64/sparc64/cheetah.c    |  9 ---------
 sys/sparc64/sparc64/locore.S     | 17 ++++-------------
 sys/sparc64/sparc64/machdep.c    | 16 +++++++++++++++-
 sys/sparc64/sparc64/mp_locore.S  | 10 +++-------
 sys/sparc64/sparc64/mp_machdep.c | 16 ++++++++++++++++
 5 files changed, 38 insertions(+), 30 deletions(-)

diff --git a/sys/sparc64/sparc64/cheetah.c b/sys/sparc64/sparc64/cheetah.c
index a99a70ca39b..836618737d1 100644
--- a/sys/sparc64/sparc64/cheetah.c
+++ b/sys/sparc64/sparc64/cheetah.c
@@ -58,13 +58,6 @@ void
 cheetah_init(u_int cpu_impl)
 {
 	u_long val;
-	register_t s;
-
-	/*
-	 * Disable interrupts for safety, this shouldn't be actually
-	 * necessary though.
-	 */
-	s = intr_disable();
 
 	/* Ensure the TSB Extension Registers hold 0 as TSB_Base. */
 
@@ -134,8 +127,6 @@ cheetah_init(u_int cpu_impl)
 		val &= ~DCR_DTPE;
 	}
 	wr(asr18, val, 0);
-
-	intr_restore(s);
 }
 
 /*
diff --git a/sys/sparc64/sparc64/locore.S b/sys/sparc64/sparc64/locore.S
index a084a3884d2..41f55a53a8b 100644
--- a/sys/sparc64/sparc64/locore.S
+++ b/sys/sparc64/sparc64/locore.S
@@ -29,6 +29,7 @@ __FBSDID("$FreeBSD$");
 
 #include <machine/asi.h>
 #include <machine/asmacros.h>
+#include <machine/intr_machdep.h>
 #include <machine/pstate.h>
 #include <machine/wstate.h>
 
@@ -47,7 +48,7 @@ ENTRY(btext)
 ENTRY(_start)
 	/*
 	 * Initialize misc. state to known values: interrupts disabled, normal
-	 * globals, windows flushed (cr = 0, cs = nwindows - 1), PIL 0 and
+	 * globals, windows flushed (cr = 0, cs = nwindows - 1), PIL_TICK and
 	 * floating point disabled.
 	 * Note that some firmware versions don't implement a clean window
 	 * trap handler so we unfortunately can't clear the windows by setting
@@ -55,7 +56,7 @@ ENTRY(_start)
 	 */
 	wrpr	%g0, PSTATE_NORMAL, %pstate
 	flushw
-	wrpr	%g0, 0, %pil
+	wrpr	%g0, PIL_TICK, %pil
 	wr	%g0, 0, %fprs
 
 	/*
@@ -65,11 +66,6 @@ ENTRY(_start)
 	SET(pcpu0 + (PCPU_PAGES * PAGE_SIZE) - PC_SIZEOF, %l1, %l0)
 	sub	%l0, SPOFF + CCFSZ, %sp
 
-	/*
-	 * Enable interrupts.
-	 */
-	wrpr	%g0, PSTATE_KERNEL, %pstate
-
 	/*
 	 * Do initial bootstrap to setup pmap and thread0.
 	 */
@@ -97,7 +93,7 @@ ENTRY(cpu_setregs)
 	ldx	[%o0 + PC_CURPCB], %o1
 
 	/*
-	 * Disable interrupts, normal globals.
+	 * Ensure we are on normal globals.
 	 */
 	wrpr	%g0, PSTATE_NORMAL, %pstate
 
@@ -147,11 +143,6 @@ ENTRY(cpu_setregs)
 	wrpr	%o1, 0, %tba
 	stw	%o3, [%o2]
 
-	/*
-	 * Re-enable interrupts.
-	 */
-	wrpr	%g0, PSTATE_KERNEL, %pstate
-
 	retl
 	 nop
 END(cpu_setregs)
diff --git a/sys/sparc64/sparc64/machdep.c b/sys/sparc64/sparc64/machdep.c
index a020fb8038f..c9040601642 100644
--- a/sys/sparc64/sparc64/machdep.c
+++ b/sys/sparc64/sparc64/machdep.c
@@ -341,7 +341,7 @@ sparc64_init(caddr_t mdp, u_long o1, u_long o2, u_long o3, ofw_vec_t *vec)
 	cpu_impl = VER_IMPL(rdpr(ver));
 
 	/*
-	 * Do CPU-specific Initialization.
+	 * Do CPU-specific initialization.
 	 */
 	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 		cheetah_init(cpu_impl);
@@ -477,6 +477,10 @@ sparc64_init(caddr_t mdp, u_long o1, u_long o2, u_long o3, ofw_vec_t *vec)
 	    sizeof(itlb_slots)) == -1)
 		panic("sparc64_init: cannot determine number of iTLB slots");
 
+	/*
+	 * Initialize and enable the caches.  Note that his may include
+	 * applying workarounds.
+	 */
 	cache_init(pc);
 	cache_enable(cpu_impl);
 	uma_set_align(pc->pc_cache.dc_linesize - 1);
@@ -568,8 +572,18 @@ sparc64_init(caddr_t mdp, u_long o1, u_long o2, u_long o3, ofw_vec_t *vec)
 	dpcpu_init(dpcpu0, 0);
 	msgbufinit(msgbufp, MSGBUF_SIZE);
 
+	/*
+	 * Initialize mutexes.
+	 */
 	mutex_init();
+
+	/*
+	 * Finish the interrupt initialization now that mutexes work and
+	 * enable them.
+	 */
 	intr_init2();
+	wrpr(pil, 0, PIL_TICK);
+	wrpr(pstate, 0, PSTATE_KERNEL);
 
 	/*
 	 * Finish pmap initialization now that we're ready for mutexes.
diff --git a/sys/sparc64/sparc64/mp_locore.S b/sys/sparc64/sparc64/mp_locore.S
index 17dc444c7d9..bd4fd13320d 100644
--- a/sys/sparc64/sparc64/mp_locore.S
+++ b/sys/sparc64/sparc64/mp_locore.S
@@ -30,6 +30,7 @@ __FBSDID("$FreeBSD$");
 
 #include <machine/asi.h>
 #include <machine/asmacros.h>
+#include <machine/intr_machdep.h>
 #include <machine/ktr.h>
 #include <machine/pstate.h>
 #include <machine/smp.h>
@@ -44,14 +45,14 @@ __FBSDID("$FreeBSD$");
 	_ALIGN_TEXT
 	/*
 	 * Initialize misc. state to known values: interrupts disabled, normal
-	 * globals, windows flushed (cr = 0, cs = nwindows - 1), PIL 0 and
+	 * globals, windows flushed (cr = 0, cs = nwindows - 1), PIL_TICK and
 	 * floating point disabled.
 	 * Note that some firmware versions don't implement a clean window
 	 * trap handler so we unfortunately can't clear the windows by setting
 	 * %cleanwin to zero here.
 	 */
 1:	wrpr	%g0, PSTATE_NORMAL, %pstate
-	wrpr	%g0, 0, %pil
+	wrpr	%g0, PIL_TICK, %pil
 	wr	%g0, 0, %fprs
 
 	rdpr	%ver, %l7
@@ -262,11 +263,6 @@ ENTRY(mp_startup)
 	add	%l1, %l2, %l1
 	sub	%l1, SPOFF + CCFSZ, %sp
 
-	/*
-	 * Enable interrupts.
-	 */
-	wrpr	%g0, PSTATE_KERNEL, %pstate
-
 #if KTR_COMPILE & KTR_SMP
 	CATR(KTR_SMP,
 	    "mp_startup: bootstrap cpuid=%d mid=%d pcpu=%#lx data=%#lx sp=%#lx"
diff --git a/sys/sparc64/sparc64/mp_machdep.c b/sys/sparc64/sparc64/mp_machdep.c
index 8ea72f3954a..ce66a4b0ee2 100644
--- a/sys/sparc64/sparc64/mp_machdep.c
+++ b/sys/sparc64/sparc64/mp_machdep.c
@@ -409,16 +409,32 @@ cpu_mp_bootstrap(struct pcpu *pc)
 	volatile struct cpu_start_args *csa;
 
 	csa = &cpu_start_args;
+
+	/* Do CPU-specific initialization. */
 	if (pc->pc_impl >= CPU_IMPL_ULTRASPARCIII)
 		cheetah_init(pc->pc_impl);
+	/*
+	 * Enable the caches.  Note that his may include applying workarounds.
+	 */
 	cache_enable(pc->pc_impl);
+
+	/* Lock the kernel TSB in the TLB. */
 	pmap_map_tsb();
+
 	/*
 	 * Flush all non-locked TLB entries possibly left over by the
 	 * firmware.
 	 */
 	tlb_flush_nonlocked();
+
+	/* Initialize global registers. */
 	cpu_setregs(pc);
+
+	/* Enable interrupts. */
+	wrpr(pil, 0, PIL_TICK);
+	wrpr(pstate, 0, PSTATE_KERNEL);
+
+	/* Start the (S)TICK interrupts. */
 	tick_start();
 
 	smp_cpus++;

From 4c7421aadbfda815341786048dd1c57a768023e8 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Mon, 26 Apr 2010 21:08:15 +0000
Subject: [PATCH 269/532] It seems ale(4) controllers do not like to see TCP
 payload in the first descriptor in TSO case. Otherwise controller can
 generate bad frames during TSO. To address it, make sure to pull up ethernet
 + IP + TCP header with options in first buffer. Also ensure the buffer length
 of the first descriptor for TSO covers entire ethernet + IP + TCP with
 options and setup additional Tx descriptor if the first buffer includes TCP
 payload.

Tested by:	Amar Takhar <verm <> darkbeer dot org >
MFC after:	1 week
---
 sys/dev/ale/if_ale.c | 37 +++++++++++++++++++++++++++++++++----
 1 file changed, 33 insertions(+), 4 deletions(-)

diff --git a/sys/dev/ale/if_ale.c b/sys/dev/ale/if_ale.c
index b3e71872d5b..ea6b53b6ff9 100644
--- a/sys/dev/ale/if_ale.c
+++ b/sys/dev/ale/if_ale.c
@@ -1585,7 +1585,7 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 	struct tcphdr *tcp;
 	bus_dma_segment_t txsegs[ALE_MAXTXSEGS];
 	bus_dmamap_t map;
-	uint32_t cflags, ip_off, poff, vtag;
+	uint32_t cflags, hdrlen, ip_off, poff, vtag;
 	int error, i, nsegs, prod, si;
 
 	ALE_LOCK_ASSERT(sc);
@@ -1678,6 +1678,11 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 				return (ENOBUFS);
 			}
 			tcp = (struct tcphdr *)(mtod(m, char *) + poff);
+			m = m_pullup(m, poff + (tcp->th_off << 2));
+			if (m == NULL) {
+				*m_head = NULL;
+				return (ENOBUFS);
+			}
 			/*
 			 * AR81xx requires IP/TCP header size and offset as
 			 * well as TCP pseudo checksum which complicates
@@ -1730,7 +1735,7 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 	}
 
 	/* Check descriptor overrun. */
-	if (sc->ale_cdata.ale_tx_cnt + nsegs >= ALE_TX_RING_CNT - 2) {
+	if (sc->ale_cdata.ale_tx_cnt + nsegs >= ALE_TX_RING_CNT - 3) {
 		bus_dmamap_unload(sc->ale_cdata.ale_tx_tag, map);
 		return (ENOBUFS);
 	}
@@ -1782,8 +1787,32 @@ ale_encap(struct ale_softc *sc, struct mbuf **m_head)
 		cflags |= ALE_TD_INSERT_VLAN_TAG;
 	}
 
-	desc = NULL;
-	for (i = 0; i < nsegs; i++) {
+	i = 0;
+	if ((m->m_pkthdr.csum_flags & CSUM_TSO) != 0) {
+		/*
+		 * Make sure the first fragment contains
+		 * only ethernet and IP/TCP header with options.
+		 */
+		hdrlen =  poff + (tcp->th_off << 2);
+		desc = &sc->ale_cdata.ale_tx_ring[prod];
+		desc->addr = htole64(txsegs[i].ds_addr);
+		desc->len = htole32(ALE_TX_BYTES(hdrlen) | vtag);
+		desc->flags = htole32(cflags);
+		sc->ale_cdata.ale_tx_cnt++;
+		ALE_DESC_INC(prod, ALE_TX_RING_CNT);
+		if (m->m_len - hdrlen > 0) {
+			/* Handle remaining payload of the first fragment. */
+			desc = &sc->ale_cdata.ale_tx_ring[prod];
+			desc->addr = htole64(txsegs[i].ds_addr + hdrlen);
+			desc->len = htole32(ALE_TX_BYTES(m->m_len - hdrlen) |
+			    vtag);
+			desc->flags = htole32(cflags);
+			sc->ale_cdata.ale_tx_cnt++;
+			ALE_DESC_INC(prod, ALE_TX_RING_CNT);
+		}
+		i = 1;
+	}
+	for (; i < nsegs; i++) {
 		desc = &sc->ale_cdata.ale_tx_ring[prod];
 		desc->addr = htole64(txsegs[i].ds_addr);
 		desc->len = htole32(ALE_TX_BYTES(txsegs[i].ds_len) | vtag);

From 90aa4f9e8c26757de0a61b1e03c79e9822bb65c0 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 27 Apr 2010 05:04:32 +0000
Subject: [PATCH 270/532] Move checking the version up from Makefile generation
 to just after we've parsed the config file.  Makefile generation is too late
 if we've introduce changes to the syntax of the metafiles to warn about
 version skew, since we have to try to parse them and we get an parse error
 that's rather baffling to the user rather than a 'your config is too old,
 upgrade' which we should get.

We have to defer doing it until after we've read the user's config
file because we define machinename there.  The version required to
compile the kernel is encoded in Makefile.machinename.  There's no
real reason for this to be the case, but changing it now would
introduce some logistical issues that I'd rather avoid for the moment.
I intend to revisit this if we're still using config in FreeBSD 10.

This also means that we cannot introduce any config metafile changes
that result in a syntax error or other error for the user until 9.0 is
released.  Otherwise, we break the upgrade path, or at least reduce
the usefulness of the error messages we generate.

# This implies that the config file option mapping will need to be redone.

MFC after:	3 days
---
 usr.sbin/config/config.h     |  1 +
 usr.sbin/config/main.c       | 40 ++++++++++++++++++++++++++++++++
 usr.sbin/config/mkmakefile.c | 45 +++++++++++++++++-------------------
 3 files changed, 62 insertions(+), 24 deletions(-)

diff --git a/usr.sbin/config/config.h b/usr.sbin/config/config.h
index ec199862ae7..4abb567e984 100644
--- a/usr.sbin/config/config.h
+++ b/usr.sbin/config/config.h
@@ -179,6 +179,7 @@ void	makehints(void);
 void	headers(void);
 void	cfgfile_add(const char *);
 void	cfgfile_removeall(void);
+FILE	*open_makefile_template(void);
 
 extern STAILQ_HEAD(device_head, device) dtab;
 
diff --git a/usr.sbin/config/main.c b/usr.sbin/config/main.c
index 2b5e055130b..5cffed2c641 100644
--- a/usr.sbin/config/main.c
+++ b/usr.sbin/config/main.c
@@ -90,6 +90,7 @@ static void get_srcdir(void);
 static void usage(void);
 static void cleanheaders(char *);
 static void kernconfdump(const char *);
+static void checkversion(void);
 
 struct hdr_list {
 	char *h_name;
@@ -204,6 +205,7 @@ main(int argc, char **argv)
 		printf("cpu type must be specified\n");
 		exit(1);
 	}
+	checkversion();
 
 	/*
 	 * make symbolic links in compilation directory
@@ -721,3 +723,41 @@ kernconfdump(const char *file)
 	}
 	fclose(fp);
 }
+
+static void 
+badversion(int versreq)
+{
+	fprintf(stderr, "ERROR: version of config(8) does not match kernel!\n");
+	fprintf(stderr, "config version = %d, ", CONFIGVERS);
+	fprintf(stderr, "version required = %d\n\n", versreq);
+	fprintf(stderr, "Make sure that /usr/src/usr.sbin/config is in sync\n");
+	fprintf(stderr, "with your /usr/src/sys and install a new config binary\n");
+	fprintf(stderr, "before trying this again.\n\n");
+	fprintf(stderr, "If running the new config fails check your config\n");
+	fprintf(stderr, "file against the GENERIC or LINT config files for\n");
+	fprintf(stderr, "changes in config syntax, or option/device naming\n");
+	fprintf(stderr, "conventions\n\n");
+	exit(1);
+}
+
+static void
+checkversion(void)
+{
+	FILE *ifp;
+	char line[BUFSIZ];
+	int versreq;
+
+	ifp = open_makefile_template();
+	while (fgets(line, BUFSIZ, ifp) != 0) {
+		if (*line != '%')
+			continue;
+		if (strncmp(line, "%VERSREQ=", 9) != 0)
+			continue;
+		versreq = atoi(line + 9);
+		if (MAJOR_VERS(versreq) == MAJOR_VERS(CONFIGVERS) &&
+		    versreq <= CONFIGVERS)
+			continue;
+		badversion(versreq);
+	}
+	fclose(ifp);
+}
diff --git a/usr.sbin/config/mkmakefile.c b/usr.sbin/config/mkmakefile.c
index 7845089cf55..b5a0e3541ce 100644
--- a/usr.sbin/config/mkmakefile.c
+++ b/usr.sbin/config/mkmakefile.c
@@ -105,17 +105,14 @@ new_fent(void)
 }
 
 /*
- * Build the makefile from the skeleton
+ * Open the correct Makefile and return it, or error out.
  */
-void
-makefile(void)
+FILE *
+open_makefile_template(void)
 {
-	FILE *ifp, *ofp;
+	FILE *ifp;
 	char line[BUFSIZ];
-	struct opt *op, *t;
-	int versreq;
 
-	read_files();
 	snprintf(line, sizeof(line), "../../conf/Makefile.%s", machinename);
 	ifp = fopen(line, "r");
 	if (ifp == 0) {
@@ -124,7 +121,21 @@ makefile(void)
 	}
 	if (ifp == 0)
 		err(1, "%s", line);
+	return (ifp);
+}
 
+/*
+ * Build the makefile from the skeleton
+ */
+void
+makefile(void)
+{
+	FILE *ifp, *ofp;
+	char line[BUFSIZ];
+	struct opt *op, *t;
+
+	read_files();
+	ifp = open_makefile_template();
 	ofp = fopen(path("Makefile.new"), "w");
 	if (ofp == 0)
 		err(1, "%s", path("Makefile.new"));
@@ -156,23 +167,9 @@ makefile(void)
 			do_rules(ofp);
 		else if (eq(line, "%CLEAN\n"))
 			do_clean(ofp);
-		else if (strncmp(line, "%VERSREQ=", sizeof("%VERSREQ=") - 1) == 0) {
-			versreq = atoi(line + sizeof("%VERSREQ=") - 1);
-			if (MAJOR_VERS(versreq) != MAJOR_VERS(CONFIGVERS) ||
-			    versreq > CONFIGVERS) {
-				fprintf(stderr, "ERROR: version of config(8) does not match kernel!\n");
-				fprintf(stderr, "config version = %d, ", CONFIGVERS);
-				fprintf(stderr, "version required = %d\n\n", versreq);
-				fprintf(stderr, "Make sure that /usr/src/usr.sbin/config is in sync\n");
-				fprintf(stderr, "with your /usr/src/sys and install a new config binary\n");
-				fprintf(stderr, "before trying this again.\n\n");
-				fprintf(stderr, "If running the new config fails check your config\n");
-				fprintf(stderr, "file against the GENERIC or LINT config files for\n");
-				fprintf(stderr, "changes in config syntax, or option/device naming\n");
-				fprintf(stderr, "conventions\n\n");
-				exit(1);
-			}
-		} else
+		else if (strncmp(line, "%VERSREQ=", 9) == 0)
+			line[0] = '\0'; /* handled elsewhere */
+		else
 			fprintf(stderr,
 			    "Unknown %% construct in generic makefile: %s",
 			    line);

From 14dd3a29ea7eaa3b02bd2f8ffc894467c025bd4c Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Tue, 27 Apr 2010 05:35:35 +0000
Subject: [PATCH 271/532] MFi386 r207205   Clearing a page table entry's
 accessed bit (PG_A) and setting the   page's PG_REFERENCED flag in
 pmap_protect() can't really be justified,   so don't do it.

---
 sys/i386/xen/pmap.c | 20 +++++++-------------
 1 file changed, 7 insertions(+), 13 deletions(-)

diff --git a/sys/i386/xen/pmap.c b/sys/i386/xen/pmap.c
index 40e36be3dcc..18619c3b0bd 100644
--- a/sys/i386/xen/pmap.c
+++ b/sys/i386/xen/pmap.c
@@ -2600,22 +2600,16 @@ retry:
 			obits = pbits = *pte;
 			if ((pbits & PG_V) == 0)
 				continue;
-			if (pbits & PG_MANAGED) {
-				m = NULL;
-				if (pbits & PG_A) {
-					m = PHYS_TO_VM_PAGE(xpmap_mtop(pbits) & PG_FRAME);
-					vm_page_flag_set(m, PG_REFERENCED);
-					pbits &= ~PG_A;
-				}
-				if ((pbits & PG_M) != 0) {
-					if (m == NULL)
-						m = PHYS_TO_VM_PAGE(xpmap_mtop(pbits) & PG_FRAME);
+
+			if ((prot & VM_PROT_WRITE) == 0) {
+				if ((pbits & (PG_MANAGED | PG_M | PG_RW)) ==
+				    (PG_MANAGED | PG_M | PG_RW)) {
+					m = PHYS_TO_VM_PAGE(xpmap_mtop(pbits) &
+					    PG_FRAME);
 					vm_page_dirty(m);
 				}
-			}
-
-			if ((prot & VM_PROT_WRITE) == 0)
 				pbits &= ~(PG_RW | PG_M);
+			}
 #ifdef PAE
 			if ((prot & VM_PROT_EXECUTE) == 0)
 				pbits |= pg_nx;

From 012ead2f2cfc8c23871e42cccc734ce3c764fb00 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 27 Apr 2010 05:38:26 +0000
Subject: [PATCH 272/532] Redo how we add compat options so as to be compatible
 with old versions of config.  Remove support for the syntax OLD = NEW form
 the options file, and instead have a new file $S/conf/options-compat. This
 file will be parsed as OLD NEW on each line.  Bump version of config.  Since
 nothing in -current ever used this, there's no hazards for current users, so
 I'm not bumping the version in the Makefiles.$MACHINE.  No need, really, for
 this version bump in -current, but this was introduced into -stable before I
 realized the version check was ineffective there, so the verison bump doesn't
 hurt here and keeps the two branches in sync, versionwise, after the MFC.

MFC after:	3 days
---
 usr.sbin/config/configvers.h |   2 +-
 usr.sbin/config/mkoptions.c  | 159 ++++++++++++++++++-----------------
 2 files changed, 82 insertions(+), 79 deletions(-)

diff --git a/usr.sbin/config/configvers.h b/usr.sbin/config/configvers.h
index 73310bd6f41..2d43329b37b 100644
--- a/usr.sbin/config/configvers.h
+++ b/usr.sbin/config/configvers.h
@@ -49,5 +49,5 @@
  *
  * $FreeBSD$
  */
-#define	CONFIGVERS	600008
+#define	CONFIGVERS	600009
 #define	MAJOR_VERS(x)	((x) / 100000)
diff --git a/usr.sbin/config/mkoptions.c b/usr.sbin/config/mkoptions.c
index 1a6ccc8fec0..846e028c4cd 100644
--- a/usr.sbin/config/mkoptions.c
+++ b/usr.sbin/config/mkoptions.c
@@ -166,7 +166,7 @@ do_option(char *name)
 			fprintf(outf, "#define %s %s\n", name, value);
 		} /* else empty file */
 
-		(void) fclose(outf);
+		(void)fclose(outf);
 		return;
 	}
 	basefile = "";
@@ -225,7 +225,7 @@ do_option(char *name)
 		if (cp == (char *)EOF)
 			break;
 	}
-	(void) fclose(inf);
+	(void)fclose(inf);
 	if (!tidy && ((value == NULL && oldvalue == NULL) ||
 	    (value && oldvalue && eq(value, oldvalue)))) {	
 		while (!SLIST_EMPTY(&op_head)) {
@@ -263,7 +263,7 @@ do_option(char *name)
 		free(op->op_value);
 		free(op);
 	}
-	(void) fclose(outf);
+	(void)fclose(outf);
 }
 
 /*
@@ -277,7 +277,7 @@ tooption(char *name)
 	struct opt_list *po;
 
 	/* "cannot happen"?  the otab list should be complete.. */
-	(void) strlcpy(nbuf, "options.h", sizeof(nbuf));
+	(void)strlcpy(nbuf, "options.h", sizeof(nbuf));
 
 	SLIST_FOREACH(po, &otab, o_next) {
 		if (eq(po->o_name, name)) {
@@ -286,78 +286,30 @@ tooption(char *name)
 		}
 	}
 
-	(void) strlcpy(hbuf, path(nbuf), sizeof(hbuf));
+	(void)strlcpy(hbuf, path(nbuf), sizeof(hbuf));
 	return (hbuf);
 }
 
-/*
- * read the options and options.<machine> files
- */
+	
 static void
-read_options(void)
+insert_option(char *this, char *val, int flags)
 {
-	FILE *fp;
-	char fname[MAXPATHLEN];
-	char *wd, *this, *val;
 	struct opt_list *po;
-	int first = 1;
-	char genopt[MAXPATHLEN];
-	int flags = 0;
 
-	SLIST_INIT(&otab);
-	(void) snprintf(fname, sizeof(fname), "../../conf/options");
-openit:
-	fp = fopen(fname, "r");
-	if (fp == 0) {
-		return;
-	}
-next:
-	flags = 0;
-	wd = get_word(fp);
-	if (wd == (char *)EOF) {
-		(void) fclose(fp);
-		if (first == 1) {
-			first++;
-			(void) snprintf(fname, sizeof fname, "../../conf/options.%s", machinename);
-			fp = fopen(fname, "r");
-			if (fp != 0)
-				goto next;
-			(void) snprintf(fname, sizeof fname, "options.%s", machinename);
-			goto openit;
-		}
-		return;
-	}
-	if (wd == 0)
-		goto next;
-	if (wd[0] == '#')
-	{
-		while (((wd = get_word(fp)) != (char *)EOF) && wd)
-		;
-		goto next;
-	}
-	this = ns(wd);
-	val = get_word(fp);
-	if (val == (char *)EOF)
-		return;
-	if (val == 0) {
-		char *s = ns(this);
-		(void) snprintf(genopt, sizeof(genopt), "opt_%s.h", lower(s));
-		val = genopt;
-		free(s);
-	} else if (eq(val, "=")) {
-		val = get_word(fp);
-		if (val == (char *)EOF) {
-			printf("%s: unexpected end of file\n", fname);
-			exit(1);
-		}
-		if (val == 0) {
-			printf("%s: Expected a right hand side at %s\n", fname,
-			    this);
-			exit(1);
-		}
-		flags |= OL_ALIAS;
-	}
-	val = ns(val);
+	po = (struct opt_list *) calloc(1, sizeof *po);
+	if (po == NULL)
+		err(EXIT_FAILURE, "calloc");
+	po->o_name = this;
+	po->o_file = val;
+	po->o_flags = flags;
+	SLIST_INSERT_HEAD(&otab, po, o_next);
+}
+
+
+static void
+check_duplicate(const char *fname, const char *this)
+{
+	struct opt_list *po;
 
 	SLIST_FOREACH(po, &otab, o_next) {
 		if (eq(po->o_name, this)) {
@@ -366,16 +318,67 @@ next:
 			exit(1);
 		}
 	}
-	
-	po = (struct opt_list *) calloc(1, sizeof *po);
-	if (po == NULL)
-		err(EXIT_FAILURE, "calloc");
-	po->o_name = this;
-	po->o_file = val;
-	po->o_flags = flags;
-	SLIST_INSERT_HEAD(&otab, po, o_next);
+}
 
-	goto next;
+static int
+read_option_file(const char *fname, int flags)
+{
+	FILE *fp;
+	char *wd, *this, *val;
+	char genopt[MAXPATHLEN];
+
+	fp = fopen(fname, "r");
+	if (fp == 0)
+		return (0);
+	while ((wd = get_word(fp)) != (char *)EOF) {
+		if (wd == 0)
+			continue;
+		if (wd[0] == '#') {
+			while (((wd = get_word(fp)) != (char *)EOF) && wd)
+				continue;
+			continue;
+		}
+		this = ns(wd);
+		val = get_word(fp);
+		if (val == (char *)EOF)
+			return (1);
+		if (val == 0) {
+			if (flags) {
+				printf("%s: compat file requires two words "
+				    "per line at %s\n", fname, this);
+				exit(1);
+			}
+			char *s = ns(this);
+			(void)snprintf(genopt, sizeof(genopt), "opt_%s.h",
+			    lower(s));
+			val = genopt;
+			free(s);
+		}
+		val = ns(val);
+		check_duplicate(fname, this);
+		insert_option(this, val, flags);
+	}
+	(void)fclose(fp);
+	return (1);
+}
+
+/*
+ * read the options and options.<machine> files
+ */
+static void
+read_options(void)
+{
+	char fname[MAXPATHLEN];
+
+	SLIST_INIT(&otab);
+	read_option_file("../../conf/options", 0);
+	(void)snprintf(fname, sizeof fname, "../../conf/options.%s",
+	    machinename);
+	if (!read_option_file(fname, 0)) {
+		(void)snprintf(fname, sizeof fname, "options.%s", machinename);
+		read_option_file(fname, 0);
+	}
+	read_option_file("../../conf/options-compat", OL_ALIAS);
 }
 
 static char *

From 4bc2b0b76d0257c828db84a98400d7000aa6568a Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 27 Apr 2010 06:14:30 +0000
Subject: [PATCH 273/532] Require the option that's mapped be listed in the
 options file.  This will allow people with old config options to either have
 it just work (if config is new enough), or get a version error (if their
 config is about 7.0 or newer) rather than getting a cryptic error about
 duplicated options in the options file, or getting an error about an unknown
 option, at which point they'd update their config file only to learn they
 need a new config, only to learn they didn't really need to update their
 config file...  All this because our version checking was in the wrong place
 for the past decade...

# hopefully this is the last change, and we'll be able to config with an
# 8.0 GENERIC file on stable/8 after I merge this change and add the
# compat options.

MFC after:	3 days
---
 usr.sbin/config/mkoptions.c | 53 +++++++++++++++++++++++++------------
 1 file changed, 36 insertions(+), 17 deletions(-)

diff --git a/usr.sbin/config/mkoptions.c b/usr.sbin/config/mkoptions.c
index 846e028c4cd..044b669f3a3 100644
--- a/usr.sbin/config/mkoptions.c
+++ b/usr.sbin/config/mkoptions.c
@@ -291,21 +291,6 @@ tooption(char *name)
 }
 
 	
-static void
-insert_option(char *this, char *val, int flags)
-{
-	struct opt_list *po;
-
-	po = (struct opt_list *) calloc(1, sizeof *po);
-	if (po == NULL)
-		err(EXIT_FAILURE, "calloc");
-	po->o_name = this;
-	po->o_file = val;
-	po->o_flags = flags;
-	SLIST_INSERT_HEAD(&otab, po, o_next);
-}
-
-
 static void
 check_duplicate(const char *fname, const char *this)
 {
@@ -320,6 +305,38 @@ check_duplicate(const char *fname, const char *this)
 	}
 }
 
+static void
+insert_option(const char *fname, char *this, char *val)
+{
+	struct opt_list *po;
+
+	check_duplicate(fname, this);
+	po = (struct opt_list *) calloc(1, sizeof *po);
+	if (po == NULL)
+		err(EXIT_FAILURE, "calloc");
+	po->o_name = this;
+	po->o_file = val;
+	po->o_flags = 0;
+	SLIST_INSERT_HEAD(&otab, po, o_next);
+}
+
+static void
+update_option(const char *this, char *val, int flags)
+{
+	struct opt_list *po;
+
+	SLIST_FOREACH(po, &otab, o_next) {
+		if (eq(po->o_name, this)) {
+			free(po->o_file);
+			po->o_file = val;
+			po->o_flags = flags;
+			return;
+		}
+	}
+	printf("Compat option %s not listed in options file.\n", this);
+	exit(1);
+}
+
 static int
 read_option_file(const char *fname, int flags)
 {
@@ -355,8 +372,10 @@ read_option_file(const char *fname, int flags)
 			free(s);
 		}
 		val = ns(val);
-		check_duplicate(fname, this);
-		insert_option(this, val, flags);
+		if (flags == 0)
+			insert_option(fname, this, val);
+		else
+			update_option(this, val, flags);
 	}
 	(void)fclose(fp);
 	return (1);

From 99c750a8141f4125174ba97eeacf3318cab5f667 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 27 Apr 2010 09:47:14 +0000
Subject: [PATCH 274/532] Provide 32bit compat for SIOCGDEFIFACE_IN6.

Based on submission by:	pluknet gmail com
Reviewed by:	emaste
MFC after:	2 weeks
---
 sys/netinet6/in6.c | 25 +++++++++++++++++++++++++
 1 file changed, 25 insertions(+)

diff --git a/sys/netinet6/in6.c b/sys/netinet6/in6.c
index 74c15d58a4a..15ff6b29bd1 100644
--- a/sys/netinet6/in6.c
+++ b/sys/netinet6/in6.c
@@ -63,6 +63,7 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
+#include "opt_compat.h"
 #include "opt_inet.h"
 #include "opt_inet6.h"
 
@@ -176,6 +177,14 @@ in6_mask2len(struct in6_addr *mask, u_char *lim0)
 #define ifa2ia6(ifa)	((struct in6_ifaddr *)(ifa))
 #define ia62ifa(ia6)	(&((ia6)->ia_ifa))
 
+#ifdef COMPAT_FREEBSD32
+struct in6_ndifreq32 {
+        char ifname[IFNAMSIZ];
+        uint32_t ifindex;
+};
+#define	SIOCGDEFIFACE32_IN6     _IOWR('i', 86, struct in6_ndifreq32)
+#endif
+
 int
 in6_control(struct socket *so, u_long cmd, caddr_t data,
     struct ifnet *ifp, struct thread *td)
@@ -226,6 +235,22 @@ in6_control(struct socket *so, u_long cmd, caddr_t data,
 	case SIOCGNBRINFO_IN6:
 	case SIOCGDEFIFACE_IN6:
 		return (nd6_ioctl(cmd, data, ifp));
+
+#ifdef COMPAT_FREEBSD32
+	case SIOCGDEFIFACE32_IN6:
+		{
+			struct in6_ndifreq ndif;
+			struct in6_ndifreq32 *ndif32;
+
+			error = nd6_ioctl(SIOCGDEFIFACE_IN6, (caddr_t)&ndif,
+			    ifp);
+			if (error)
+				return (error);
+			ndif32 = (struct in6_ndifreq32 *)data;
+			ndif32->ifindex = ndif.ifindex;
+			return (0);
+		}
+#endif
 	}
 
 	switch (cmd) {

From 8bac98182aa3f4976b0f5c3067a53f8b29b5a526 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 27 Apr 2010 09:48:43 +0000
Subject: [PATCH 275/532] Style: use #define<TAB> instead of #define<SPACE>.

Noted by:	bde, pluknet gmail com
MFC after:	11 days
---
 sys/amd64/include/proc.h   | 4 ++--
 sys/arm/include/proc.h     | 2 +-
 sys/i386/include/proc.h    | 2 +-
 sys/ia64/include/proc.h    | 4 ++--
 sys/mips/include/proc.h    | 4 ++--
 sys/powerpc/include/proc.h | 2 +-
 sys/sparc64/include/proc.h | 2 +-
 sys/sun4v/include/proc.h   | 2 +-
 8 files changed, 11 insertions(+), 11 deletions(-)

diff --git a/sys/amd64/include/proc.h b/sys/amd64/include/proc.h
index 126412cad45..acea4c0a77f 100644
--- a/sys/amd64/include/proc.h
+++ b/sys/amd64/include/proc.h
@@ -53,8 +53,8 @@ struct mdproc {
 	struct system_segment_descriptor md_ldt_sd;
 };
 
-#define KINFO_PROC_SIZE 1088
-#define KINFO_PROC32_SIZE 768
+#define	KINFO_PROC_SIZE 1088
+#define	KINFO_PROC32_SIZE 768
 
 #ifdef	_KERNEL
 
diff --git a/sys/arm/include/proc.h b/sys/arm/include/proc.h
index 37051bf518b..08857446430 100644
--- a/sys/arm/include/proc.h
+++ b/sys/arm/include/proc.h
@@ -60,6 +60,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
-#define KINFO_PROC_SIZE 792
+#define	KINFO_PROC_SIZE 792
 
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/i386/include/proc.h b/sys/i386/include/proc.h
index 0b9fca2b55f..86be8c81721 100644
--- a/sys/i386/include/proc.h
+++ b/sys/i386/include/proc.h
@@ -57,7 +57,7 @@ struct mdproc {
 	struct proc_ldt *md_ldt;	/* (t) per-process ldt */
 };
 
-#define KINFO_PROC_SIZE 768
+#define	KINFO_PROC_SIZE 768
 
 #ifdef	_KERNEL
 
diff --git a/sys/ia64/include/proc.h b/sys/ia64/include/proc.h
index 81e2e4f7bb8..e9f337cbb48 100644
--- a/sys/ia64/include/proc.h
+++ b/sys/ia64/include/proc.h
@@ -38,7 +38,7 @@ struct mdproc {
 	int		__dummy;	/* Avoid having an empty struct. */
 };
 
-#define KINFO_PROC_SIZE 1088
-#define KINFO_PROC32_SIZE 768
+#define	KINFO_PROC_SIZE 1088
+#define	KINFO_PROC32_SIZE 768
 
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/mips/include/proc.h b/sys/mips/include/proc.h
index 1479edab293..e71c858788d 100644
--- a/sys/mips/include/proc.h
+++ b/sys/mips/include/proc.h
@@ -69,9 +69,9 @@ void	mips_cpu_switch(struct thread *, struct thread *, struct mtx *);
 void	mips_cpu_throw(struct thread *, struct thread *);
 
 #ifdef __mips_n64
-#define KINFO_PROC_SIZE 1088
+#define	KINFO_PROC_SIZE 1088
 #else
-#define KINFO_PROC_SIZE 816
+#define	KINFO_PROC_SIZE 816
 #endif
 
 #endif	/* !_MACHINE_PROC_H_ */
diff --git a/sys/powerpc/include/proc.h b/sys/powerpc/include/proc.h
index cbe6da9ff23..d51116bb013 100644
--- a/sys/powerpc/include/proc.h
+++ b/sys/powerpc/include/proc.h
@@ -46,6 +46,6 @@ struct mdthread {
 struct mdproc {
 };
 
-#define KINFO_PROC_SIZE 768
+#define	KINFO_PROC_SIZE 768
 
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/sparc64/include/proc.h b/sys/sparc64/include/proc.h
index 84eaed5a00e..3c236637fb9 100644
--- a/sys/sparc64/include/proc.h
+++ b/sys/sparc64/include/proc.h
@@ -51,6 +51,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
-#define KINFO_PROC_SIZE 1088
+#define	KINFO_PROC_SIZE 1088
 
 #endif /* !_MACHINE_PROC_H_ */
diff --git a/sys/sun4v/include/proc.h b/sys/sun4v/include/proc.h
index 84eaed5a00e..3c236637fb9 100644
--- a/sys/sun4v/include/proc.h
+++ b/sys/sun4v/include/proc.h
@@ -51,6 +51,6 @@ struct mdproc {
 	void	*md_sigtramp;
 };
 
-#define KINFO_PROC_SIZE 1088
+#define	KINFO_PROC_SIZE 1088
 
 #endif /* !_MACHINE_PROC_H_ */

From fd963b992986bba83e344396c5d7dcff4ea6c1c0 Mon Sep 17 00:00:00 2001
From: Bruce M Simpson <bms@FreeBSD.org>
Date: Tue, 27 Apr 2010 14:14:21 +0000
Subject: [PATCH 276/532] Fix a regression where DVMRP diagnostic traffic, such
 as that used by mrinfo and mtrace, was dropped by the IGMP TTL check. IGMP
 control traffic must always have a TTL of 1.

Submitted by:	Matthew Luckie
MFC after:	3 days
---
 sys/netinet/igmp.c | 17 +++++++++++------
 1 file changed, 11 insertions(+), 6 deletions(-)

diff --git a/sys/netinet/igmp.c b/sys/netinet/igmp.c
index f9f6381f3fc..d4c8e99cdab 100644
--- a/sys/netinet/igmp.c
+++ b/sys/netinet/igmp.c
@@ -1468,12 +1468,6 @@ igmp_input(struct mbuf *m, int off)
 	}
 	ip = mtod(m, struct ip *);
 
-	if (ip->ip_ttl != 1) {
-		IGMPSTAT_INC(igps_rcv_badttl);
-		m_freem(m);
-		return;
-	}
-
 	/*
 	 * Validate checksum.
 	 */
@@ -1488,6 +1482,17 @@ igmp_input(struct mbuf *m, int off)
 	m->m_data -= iphlen;
 	m->m_len += iphlen;
 
+	/*
+	 * IGMP control traffic is link-scope, and must have a TTL of 1.
+	 * DVMRP traffic (e.g. mrinfo, mtrace) is an exception;
+	 * probe packets may come from beyond the LAN.
+	 */
+	if (igmp->igmp_type != IGMP_DVMRP && ip->ip_ttl != 1) {
+		IGMPSTAT_INC(igps_rcv_badttl);
+		m_freem(m);
+		return;
+	}
+
 	switch (igmp->igmp_type) {
 	case IGMP_HOST_MEMBERSHIP_QUERY:
 		if (igmplen == IGMP_MINLEN) {

From 877fc3b64bea982f383636d7bea5fbcda98deb59 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Tue, 27 Apr 2010 15:05:03 +0000
Subject: [PATCH 277/532] Make sure IPv6 source address selection does not
 change interface addresses while walking the IPv6 address list if in the jail
 case something is connecting to ::1.

Reported by:	Pieter de Boer (pieter thedarkside.nl)
Tested by:	Pieter de Boer (pieter thedarkside.nl)
MFC after:	4 days
---
 sys/netinet6/in6_src.c | 25 ++++++++++++++++++++-----
 1 file changed, 20 insertions(+), 5 deletions(-)

diff --git a/sys/netinet6/in6_src.c b/sys/netinet6/in6_src.c
index ea302a52fa4..e6c2cd83726 100644
--- a/sys/netinet6/in6_src.c
+++ b/sys/netinet6/in6_src.c
@@ -182,7 +182,7 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
     struct inpcb *inp, struct route_in6 *ro, struct ucred *cred,
     struct ifnet **ifpp, struct in6_addr *srcp)
 {
-	struct in6_addr dst;
+	struct in6_addr dst, tmp;
 	struct ifnet *ifp = NULL;
 	struct in6_ifaddr *ia = NULL, *ia_best = NULL;
 	struct in6_pktinfo *pi = NULL;
@@ -326,10 +326,9 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 		if (!V_ip6_use_deprecated && IFA6_IS_DEPRECATED(ia))
 			continue;
 
+		/* If jailed only take addresses of the jail into account. */
 		if (cred != NULL &&
-		    prison_local_ip6(cred, &ia->ia_addr.sin6_addr,
-			(inp != NULL &&
-			(inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0)
+		    prison_check_ip6(cred, &ia->ia_addr.sin6_addr) != 0)
 			continue;
 
 		/* Rule 1: Prefer same address */
@@ -476,10 +475,26 @@ in6_selectsrc(struct sockaddr_in6 *dstsock, struct ip6_pktopts *opts,
 		return (EADDRNOTAVAIL);
 	}
 
+	/*
+	 * At this point at least one of the addresses belonged to the jail
+	 * but it could still be, that we want to further restrict it, e.g.
+	 * theoratically IN6_IS_ADDR_LOOPBACK.
+	 * It must not be IN6_IS_ADDR_UNSPECIFIED anymore.
+	 * prison_local_ip6() will fix an IN6_IS_ADDR_LOOPBACK but should
+	 * let all others previously selected pass.
+	 * Use tmp to not change ::1 on lo0 to the primary jail address.
+	 */
+	tmp = ia->ia_addr.sin6_addr;
+	if (cred != NULL && prison_local_ip6(cred, &tmp, (inp != NULL &&
+	    (inp->inp_flags & IN6P_IPV6_V6ONLY) != 0)) != 0) {
+		IN6_IFADDR_RUNLOCK();
+		return (EADDRNOTAVAIL);
+	}
+
 	if (ifpp)
 		*ifpp = ifp;
 
-	bcopy(&ia->ia_addr.sin6_addr, srcp, sizeof(*srcp));
+	bcopy(&tmp, srcp, sizeof(*srcp));
 	IN6_IFADDR_RUNLOCK();
 	return (0);
 }

From 7a657e630d11e68d181d10fce0b44a87804fd037 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Tue, 27 Apr 2010 15:07:08 +0000
Subject: [PATCH 278/532] Enhance the historic behaviour of raw sockets and
 jails in a way that we allow all possible jail IPs as source address rather
 than forcing the "primary". While IPv6 naturally has source address
 selection, for legacy IP we do not go through the pain in case IP_HDRINCL was
 not set. People should bind(2) for that.

This will, for example, allow ping(|6) -S to work correctly for
non-primary addresses.

Reported by:	(ten 211.ru)
Tested by:	(ten 211.ru)
MFC after:	4 days
---
 sys/netinet/raw_ip.c   | 23 ++++++++++++++++++-----
 sys/netinet6/raw_ip6.c |  2 +-
 2 files changed, 19 insertions(+), 6 deletions(-)

diff --git a/sys/netinet/raw_ip.c b/sys/netinet/raw_ip.c
index 1db37746c83..0b77b5b8496 100644
--- a/sys/netinet/raw_ip.c
+++ b/sys/netinet/raw_ip.c
@@ -427,11 +427,24 @@ rip_output(struct mbuf *m, struct socket *so, u_long dst)
 		ip->ip_p = inp->inp_ip_p;
 		ip->ip_len = m->m_pkthdr.len;
 		ip->ip_src = inp->inp_laddr;
-		error = prison_get_ip4(inp->inp_cred, &ip->ip_src);
-		if (error != 0) {
-			INP_RUNLOCK(inp);
-			m_freem(m);
-			return (error);
+		if (jailed(inp->inp_cred)) {
+			/*
+			 * prison_local_ip4() would be good enough but would
+			 * let a source of INADDR_ANY pass, which we do not
+			 * want to see from jails. We do not go through the
+			 * pain of in_pcbladdr() for raw sockets.
+			 */
+			if (ip->ip_src.s_addr == INADDR_ANY)
+				error = prison_get_ip4(inp->inp_cred,
+				    &ip->ip_src);
+			else
+				error = prison_local_ip4(inp->inp_cred,
+				    &ip->ip_src);
+			if (error != 0) {
+				INP_RUNLOCK(inp);
+				m_freem(m);
+				return (error);
+			}
 		}
 		ip->ip_dst.s_addr = dst;
 		ip->ip_ttl = inp->inp_ip_ttl;
diff --git a/sys/netinet6/raw_ip6.c b/sys/netinet6/raw_ip6.c
index 9f1236ab1bd..6052b24e470 100644
--- a/sys/netinet6/raw_ip6.c
+++ b/sys/netinet6/raw_ip6.c
@@ -465,7 +465,7 @@ rip6_output(m, va_alist)
 	    &oifp, &in6a);
 	if (error)
 		goto bad;
-	error = prison_get_ip6(in6p->inp_cred, &in6a);
+	error = prison_check_ip6(in6p->inp_cred, &in6a);
 	if (error != 0)
 		goto bad;
 	ip6->ip6_src = in6a;

From 1b610a749e0cfb4e848d4afbe0d028723650ad88 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Tue, 27 Apr 2010 15:16:54 +0000
Subject: [PATCH 279/532] MFP4: @177254

Add missing CURVNET_RESTORE() calls for multiple code paths, to stop
leaking the currently cached vnet into callers and to the process.

Sponsored by:	The FreeBSD Foundation
Sponsored by:	CK Software GmbH
MFC after:	4 days
---
 sys/net/bpf.c | 11 ++++++++---
 1 file changed, 8 insertions(+), 3 deletions(-)

diff --git a/sys/net/bpf.c b/sys/net/bpf.c
index 1755ec7e8f2..16245c947da 100644
--- a/sys/net/bpf.c
+++ b/sys/net/bpf.c
@@ -1454,6 +1454,7 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 			/* FALLSTHROUGH */
 
 		default:
+			CURVNET_RESTORE();
 			return (EINVAL);
 		}
 
@@ -1461,6 +1462,7 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 		if (d->bd_sbuf != NULL || d->bd_hbuf != NULL ||
 		    d->bd_fbuf != NULL || d->bd_bif != NULL) {
 			BPFD_UNLOCK(d);
+			CURVNET_RESTORE();
 			return (EBUSY);
 		}
 		d->bd_bufmode = *(u_int *)addr;
@@ -1468,13 +1470,16 @@ bpfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags,
 		break;
 
 	case BIOCGETZMAX:
-		return (bpf_ioctl_getzmax(td, d, (size_t *)addr));
+		error = bpf_ioctl_getzmax(td, d, (size_t *)addr);
+		break;
 
 	case BIOCSETZBUF:
-		return (bpf_ioctl_setzbuf(td, d, (struct bpf_zbuf *)addr));
+		error = bpf_ioctl_setzbuf(td, d, (struct bpf_zbuf *)addr);
+		break;
 
 	case BIOCROTZBUF:
-		return (bpf_ioctl_rotzbuf(td, d, (struct bpf_zbuf *)addr));
+		error = bpf_ioctl_rotzbuf(td, d, (struct bpf_zbuf *)addr);
+		break;
 	}
 	CURVNET_RESTORE();
 	return (error);

From 45191d258833672ce2f6438004d4227e7a313e90 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Tue, 27 Apr 2010 15:50:25 +0000
Subject: [PATCH 280/532] Take advantage of OF_getscsinitid().

---
 sys/dev/esp/esp_sbus.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/sys/dev/esp/esp_sbus.c b/sys/dev/esp/esp_sbus.c
index 2033cc65531..3ba0d46ae71 100644
--- a/sys/dev/esp/esp_sbus.c
+++ b/sys/dev/esp/esp_sbus.c
@@ -466,9 +466,7 @@ espattach(struct esp_softc *esc, const struct ncr53c9x_glue *gluep)
 		goto fail_lock;
 	}
 
-	if (OF_getprop(ofw_bus_get_node(esc->sc_dev), "scsi-initiator-id",
-	    &sc->sc_id, sizeof(sc->sc_id)) == -1)
-		sc->sc_id = 7;
+	sc->sc_id = OF_getscsinitid(esc->sc_dev);
 
 #ifdef ESP_SBUS_DEBUG
 	device_printf(esc->sc_dev, "%s: sc_id %d, freq %d\n",

From 2eb4a8dcae2b87613ad730ecb6ad2389fc323249 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Tue, 27 Apr 2010 15:59:38 +0000
Subject: [PATCH 281/532] Update device identify data and serial number when
 device change detected. Reprobe immediately following this should have fresh
 data.

---
 sys/cam/ata/ata_xpt.c | 8 ++++++--
 1 file changed, 6 insertions(+), 2 deletions(-)

diff --git a/sys/cam/ata/ata_xpt.c b/sys/cam/ata/ata_xpt.c
index 3bf9c1ef1df..10e61827ad7 100644
--- a/sys/cam/ata/ata_xpt.c
+++ b/sys/cam/ata/ata_xpt.c
@@ -768,6 +768,7 @@ noerror:
 	{
 		struct ccb_pathinq cpi;
 		int16_t *ptr;
+		int changed = 1;
 
 		ident_buf = &softc->ident_data;
 		for (ptr = (int16_t *)ident_buf;
@@ -809,9 +810,12 @@ noerror:
 			     sizeof(ident_buf->serial))) {
 				/* Device changed. */
 				xpt_async(AC_LOST_DEVICE, path, NULL);
-			} else
+			} else {
 				bcopy(&softc->ident_data, ident_buf, sizeof(struct ata_params));
-		} else {
+				changed = 0;
+			}
+		}
+		if (changed) {
 			bcopy(&softc->ident_data, ident_buf, sizeof(struct ata_params));
 			/* Clean up from previous instance of this device */
 			if (path->device->serial_num != NULL) {

From 1edd4fcd23cf1ca4e2f8bfdd539ee0f60c4d6e65 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Tue, 27 Apr 2010 17:50:43 +0000
Subject: [PATCH 282/532] Language improvements to make the BUGS section easier
 to read.

Reviewed by:	Alexander Best <alexbestms wwu de>
MFC after:	13 days
---
 usr.bin/gzip/gzip.1 | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/usr.bin/gzip/gzip.1 b/usr.bin/gzip/gzip.1
index bb1faddbabb..848a4b391c9 100644
--- a/usr.bin/gzip/gzip.1
+++ b/usr.bin/gzip/gzip.1
@@ -25,7 +25,7 @@
 .\" SUCH DAMAGE.
 .\"
 .\" $FreeBSD$
-.Dd April 7, 2010
+.Dd April 27, 2010
 .Dt GZIP 1
 .Os
 .Sh NAME
@@ -218,8 +218,8 @@ with unpack support written by
 .An Xin LI Aq delphij@FreeBSD.org .
 .Sh BUGS
 According to RFC 1952, the recorded file size is stored in a 32-bit
-integer and therefore it can not represent files that is bigger than
-4GB in size.  This limitation also applies to
+integer, therefore, it can not represent files larger than 4GB.
+This limitation also applies to
 .Fl l
 option of
 .Nm

From 637bfa488b897727534ae4ac65851cb4334cce72 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Tue, 27 Apr 2010 17:57:04 +0000
Subject: [PATCH 283/532] Use _exit(2) system call directly instead of using
 exit(3) in signal handler, as the latter is not guaranteed to be signal safe,
 and we do not really care about flushing the stream during SIGINT.

Suggested by:	Maxim Konovalov <maxim.konovalov gmail com>
MFC after:	13 days
---
 usr.bin/gzip/gzip.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.bin/gzip/gzip.c b/usr.bin/gzip/gzip.c
index 9503762a587..de9fe809aa1 100644
--- a/usr.bin/gzip/gzip.c
+++ b/usr.bin/gzip/gzip.c
@@ -1178,7 +1178,7 @@ sigint_handler(int signo __unused)
 
 	if (remove_file != NULL)
 		unlink(remove_file);
-	exit(2);
+	_exit(2);
 }
 #endif
 

From 699346fc8b4dc86322dae1f9dcacade4560b3e2e Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Tue, 27 Apr 2010 18:05:33 +0000
Subject: [PATCH 284/532] - On sparc64 obtain the initiator ID from the Open
 Firmware device tree   in order to match what the PROM built-in driver uses.
 - Remove some no longer used includes.

---
 sys/dev/sym/sym_hipd.c | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/sys/dev/sym/sym_hipd.c b/sys/dev/sym/sym_hipd.c
index a43c816c65f..5050a5f9a9b 100644
--- a/sys/dev/sym/sym_hipd.c
+++ b/sys/dev/sym/sym_hipd.c
@@ -87,6 +87,12 @@ __FBSDID("$FreeBSD$");
 
 #include <machine/bus.h>
 #include <machine/resource.h>
+
+#ifdef __sparc64__
+#include <dev/ofw/openfirm.h>
+#include <machine/ofw_machdep.h>
+#endif
+
 #include <sys/rman.h>
 
 #include <cam/cam.h>
@@ -98,10 +104,6 @@ __FBSDID("$FreeBSD$");
 #include <cam/scsi/scsi_all.h>
 #include <cam/scsi/scsi_message.h>
 
-#include <vm/vm.h>
-#include <vm/vm_param.h>
-#include <vm/pmap.h>
-
 /* Short and quite clear integer types */
 typedef int8_t    s8;
 typedef int16_t   s16;
@@ -2682,6 +2684,9 @@ static int sym_prepare_setting(hcb_p np, struct sym_nvram *nvram)
 	 */
 	np->myaddr = 255;
 	sym_nvram_setup_host (np, nvram);
+#ifdef __sparc64__
+	np->myaddr = OF_getscsinitid(np->device);
+#endif
 
 	/*
 	 *  Get SCSI addr of host adapter (set by bios?).

From a8217f5c2e29f3e676d274ac99a7982f0dc5c90c Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Tue, 27 Apr 2010 18:06:54 +0000
Subject: [PATCH 285/532] Replace a magic value with the appropriate macro.

---
 sys/dev/mpt/mpt_cam.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/sys/dev/mpt/mpt_cam.c b/sys/dev/mpt/mpt_cam.c
index 5e63ef908bd..56ea3e0f00e 100644
--- a/sys/dev/mpt/mpt_cam.c
+++ b/sys/dev/mpt/mpt_cam.c
@@ -1058,12 +1058,13 @@ mpt_read_config_info_spi(struct mpt_softc *mpt)
 static int
 mpt_set_initial_config_spi(struct mpt_softc *mpt)
 {
-	int i, pp1val = ((1 << mpt->mpt_ini_id) << 16) | mpt->mpt_ini_id;
-	int error;
+	int error, i, pp1val;
 
 	mpt->mpt_disc_enable = 0xff;
 	mpt->mpt_tag_enable = 0;
 
+	pp1val = ((1 << mpt->mpt_ini_id) <<
+	    MPI_SCSIPORTPAGE1_CFG_SHIFT_PORT_RESPONSE_ID) | mpt->mpt_ini_id;
 	if (mpt->mpt_port_page1.Configuration != pp1val) {
 		CONFIG_PAGE_SCSI_PORT_1 tmp;
 

From 8ba5efbd6fa32d1ca0ca6cfcc6cde9201b3c5c73 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Tue, 27 Apr 2010 18:41:16 +0000
Subject: [PATCH 286/532] On sparc64 obtain the initiator ID to be used for SPI
 HBAs from the Open Firmware device tree in order to match what the PROM
 built-in driver uses. This is especially important when netbooting Fujitsu
 Siemens PRIMEPOWER250 as in that case the built-in driver isn't used and the
 port facts PortSCSIID defaults to 0, conflicting with the disk at the same
 address.

---
 sys/dev/mpt/mpt.c     |  2 ++
 sys/dev/mpt/mpt.h     | 10 +++++++++-
 sys/dev/mpt/mpt_pci.c |  5 +++++
 3 files changed, 16 insertions(+), 1 deletion(-)

diff --git a/sys/dev/mpt/mpt.c b/sys/dev/mpt/mpt.c
index 6a74dc3c54d..40c581c6564 100644
--- a/sys/dev/mpt/mpt.c
+++ b/sys/dev/mpt/mpt.c
@@ -2667,6 +2667,8 @@ mpt_configure_ioc(struct mpt_softc *mpt, int tn, int needreset)
 		mpt->is_fc = 0;
 		mpt->is_sas = 0;
 		mpt->is_spi = 1;
+		if (mpt->mpt_ini_id == MPT_INI_ID_NONE)
+			mpt->mpt_ini_id = pfp->PortSCSIID;
 	} else if (pfp->PortType == MPI_PORTFACTS_PORTTYPE_ISCSI) {
 		mpt_prt(mpt, "iSCSI not supported yet\n");
 		return (ENXIO);
diff --git a/sys/dev/mpt/mpt.h b/sys/dev/mpt/mpt.h
index 1fa2d1aaded..ab4cfa6c4ff 100644
--- a/sys/dev/mpt/mpt.h
+++ b/sys/dev/mpt/mpt.h
@@ -130,6 +130,11 @@
 #include <machine/clock.h>
 #endif
 
+#ifdef __sparc64__
+#include <dev/ofw/openfirm.h>
+#include <machine/ofw_machdep.h>
+#endif
+
 #include <sys/rman.h>
 
 #if __FreeBSD_version < 500000  
@@ -172,6 +177,8 @@
 #define	MPT_ROLE_BOTH		3
 #define	MPT_ROLE_DEFAULT	MPT_ROLE_INITIATOR
 
+#define	MPT_INI_ID_NONE		-1
+
 /**************************** Forward Declarations ****************************/
 struct mpt_softc;
 struct mpt_personality;
@@ -637,7 +644,6 @@ struct mpt_softc {
 	 * Port Facts
 	 */
 	MSG_PORT_FACTS_REPLY *	port_facts;
-#define	mpt_ini_id	port_facts[0].PortSCSIID
 #define	mpt_max_tgtcmds	port_facts[0].MaxPostedCmdBuffers
 
 	/*
@@ -650,6 +656,7 @@ struct mpt_softc {
 			CONFIG_PAGE_SCSI_PORT_2		_port_page2;
 			CONFIG_PAGE_SCSI_DEVICE_0	_dev_page0[16];
 			CONFIG_PAGE_SCSI_DEVICE_1	_dev_page1[16];
+			int				_ini_id;
 			uint16_t			_tag_enable;
 			uint16_t			_disc_enable;
 		} spi;
@@ -658,6 +665,7 @@ struct mpt_softc {
 #define	mpt_port_page2		cfg.spi._port_page2
 #define	mpt_dev_page0		cfg.spi._dev_page0
 #define	mpt_dev_page1		cfg.spi._dev_page1
+#define	mpt_ini_id		cfg.spi._ini_id
 #define	mpt_tag_enable		cfg.spi._tag_enable
 #define	mpt_disc_enable		cfg.spi._disc_enable
 		struct mpi_fc_cfg {
diff --git a/sys/dev/mpt/mpt_pci.c b/sys/dev/mpt/mpt_pci.c
index 1e9dc921c09..e3b0ea35994 100644
--- a/sys/dev/mpt/mpt_pci.c
+++ b/sys/dev/mpt/mpt_pci.c
@@ -460,6 +460,11 @@ mpt_pci_attach(device_t dev)
 	mpt->raid_queue_depth = MPT_RAID_QUEUE_DEPTH_DEFAULT;
 	mpt->verbose = MPT_PRT_NONE;
 	mpt->role = MPT_ROLE_NONE;
+	mpt->mpt_ini_id = MPT_INI_ID_NONE;
+#ifdef __sparc64__
+	if (mpt->is_spi)
+		mpt->mpt_ini_id = OF_getscsinitid(dev);
+#endif
 	mpt_set_options(mpt);
 	if (mpt->verbose == MPT_PRT_NONE) {
 		mpt->verbose = MPT_PRT_WARN;

From 3e8b572db47ef252a1061b291c60e11c24ac2223 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Tue, 27 Apr 2010 23:48:50 +0000
Subject: [PATCH 287/532] need to initialize the lock before it is used

MFC after:	3 days
---
 sys/net/flowtable.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/net/flowtable.c b/sys/net/flowtable.c
index 39b6b40a996..0e4beb4a07f 100644
--- a/sys/net/flowtable.c
+++ b/sys/net/flowtable.c
@@ -1626,7 +1626,7 @@ flowtable_init(const void *unused __unused)
 	    EVENTHANDLER_PRI_ANY);
 	flowclean_freq = 20*hz;
 }
-SYSINIT(flowtable_init, SI_SUB_SMP, SI_ORDER_MIDDLE,
+SYSINIT(flowtable_init, SI_SUB_KTHREAD_INIT, SI_ORDER_FIRST,
     flowtable_init, NULL);
 
 

From 030d4cda77676d8dc766b009cda9ef194e28616a Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 28 Apr 2010 04:25:36 +0000
Subject: [PATCH 288/532] Adapt i386 r207205 to mips: Clearing PV_TABLE_REF and
 setting the page's PG_REFERENCED flag in pmap_protect() can't really be
 justified, so don't do it.

---
 sys/mips/mips/pmap.c | 12 +++---------
 1 file changed, 3 insertions(+), 9 deletions(-)

diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 3634a27e3ee..e7c323938dc 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -1690,16 +1690,10 @@ retry:
 		obits = pbits = *pte;
 		pa = mips_tlbpfn_to_paddr(pbits);
 
-		if (page_is_managed(pa)) {
+		if (page_is_managed(pa) && (pbits & PTE_M) != 0) {
 			m = PHYS_TO_VM_PAGE(pa);
-			if (m->md.pv_flags & PV_TABLE_REF) {
-				vm_page_flag_set(m, PG_REFERENCED);
-				m->md.pv_flags &= ~PV_TABLE_REF;
-			}
-			if (pbits & PTE_M) {
-				vm_page_dirty(m);
-				m->md.pv_flags &= ~PV_TABLE_MOD;
-			}
+			vm_page_dirty(m);
+			m->md.pv_flags &= ~PV_TABLE_MOD;
 		}
 		pbits = (pbits & ~PTE_M) | PTE_RO;
 

From 6a2a3d733857a8a4f4b3cea7336e74f8068a6c12 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 28 Apr 2010 04:57:32 +0000
Subject: [PATCH 289/532] Change vm_object_madvise() so that it checks whether
 the page is invalid or unmanaged before acquiring the page queues lock. 
 Neither of these tests require that lock.  Moreover, a better way of testing
 if the page is unmanaged is to test the type of vm object.  This avoids a
 pointless vm_page_lookup().

MFC after:	3 weeks
---
 sys/vm/vm_object.c | 16 ++++++----------
 1 file changed, 6 insertions(+), 10 deletions(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 7e4dbc85350..c25ab775794 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1162,7 +1162,8 @@ shadowlookup:
 			    (tobject->flags & OBJ_ONEMAPPING) == 0) {
 				goto unlock_tobject;
 			}
-		}
+		} else if (tobject->type == OBJT_PHYS)
+			goto unlock_tobject;
 		m = vm_page_lookup(tobject, tpindex);
 		if (m == NULL && advise == MADV_WILLNEED) {
 			/*
@@ -1189,18 +1190,13 @@ shadowlookup:
 				VM_OBJECT_UNLOCK(tobject);
 			tobject = backing_object;
 			goto shadowlookup;
-		}
+		} else if (m->valid != VM_PAGE_BITS_ALL)
+			goto unlock_tobject;
 		/*
-		 * If the page is busy or not in a normal active state,
-		 * we skip it.  If the page is not managed there are no
-		 * page queues to mess with.  Things can break if we mess
-		 * with pages in any of the below states.
+		 * If the page is not in a normal state, skip it.
 		 */
 		vm_page_lock_queues();
-		if (m->hold_count ||
-		    m->wire_count ||
-		    (m->flags & PG_UNMANAGED) ||
-		    m->valid != VM_PAGE_BITS_ALL) {
+		if (m->hold_count != 0 || m->wire_count != 0) {
 			vm_page_unlock_queues();
 			goto unlock_tobject;
 		}

From 77d6d85393d68d8e6752e5708f3e41669b0d7fe2 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 28 Apr 2010 06:34:47 +0000
Subject: [PATCH 290/532] Setting PG_REFERENCED on a page at the end of
 vm_fault() is redundant since the page table entry's accessed bit is either
 preset by the immediately preceding call to pmap_enter() or by hardware (or
 software) upon return from vm_fault() when the faulting access is restarted.

---
 sys/vm/vm_fault.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index f4098567887..058cbb09666 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -924,7 +924,6 @@ vnode_locked:
 		vm_fault_prefault(fs.map->pmap, vaddr, fs.entry);
 	VM_OBJECT_LOCK(fs.object);
 	vm_page_lock_queues();
-	vm_page_flag_set(fs.m, PG_REFERENCED);
 
 	/*
 	 * If the page is not wired down, then put it where the pageout daemon

From 3b32573a9fe54b545b392061f75d9f47fdefa03c Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Wed, 28 Apr 2010 07:26:41 +0000
Subject: [PATCH 291/532]  - Fix builds without SOFTUPDATES defined in the
 kernel config.

---
 sys/ufs/ffs/ffs_softdep.c | 171 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 171 insertions(+)

diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index 4a659f9de7b..b08e7c299d4 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -122,6 +122,21 @@ softdep_uninitialize()
 	return;
 }
 
+void
+softdep_unmount(mp)
+	struct mount *mp;
+{
+
+}
+
+void
+softdep_setup_sbupdate(ump, fs, bp)
+	struct ufsmount *ump;
+	struct fs *fs;
+	struct buf *bp;
+{
+}
+
 void
 softdep_setup_inomapdep(bp, ip, newinum)
 	struct buf *bp;
@@ -267,6 +282,162 @@ softdep_setup_directory_change(bp, dp, ip, newinum, isrmdir)
 	panic("softdep_setup_directory_change called");
 }
 
+void *
+softdep_setup_trunc(vp, length, flags)
+	struct vnode *vp;
+	off_t length;
+	int flags;
+{
+
+	panic("%s called", __FUNCTION__);
+
+	return (NULL);
+}
+
+int
+softdep_complete_trunc(vp, cookie)
+	struct vnode *vp;
+	void *cookie;
+{
+
+	panic("%s called", __FUNCTION__);
+
+	return (0);
+}
+
+void
+softdep_setup_blkfree(mp, bp, blkno, frags, wkhd)
+	struct mount *mp;
+	struct buf *bp;
+	ufs2_daddr_t blkno;
+	int frags;
+	struct workhead *wkhd;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_inofree(mp, bp, ino, wkhd)
+	struct mount *mp;
+	struct buf *bp;
+	ino_t ino;
+	struct workhead *wkhd;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_unlink(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_revert_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_rmdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_revert_rmdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_create(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_revert_create(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_mkdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_revert_mkdir(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+void
+softdep_setup_dotdot_link(dp, ip)
+	struct inode *dp;
+	struct inode *ip;
+{
+
+	panic("%s called", __FUNCTION__);
+}
+
+int
+softdep_prealloc(vp, waitok)
+	struct vnode *vp;
+	int waitok;
+{
+
+	panic("%s called", __FUNCTION__);
+
+	return (0);
+}
+
+int
+softdep_journal_lookup(mp, vpp)
+	struct mount *mp;
+	struct vnode **vpp;
+{
+
+	return (ENOENT);
+}
+
 void
 softdep_change_linkcnt(ip)
 	struct inode *ip;

From 2bd20091e47c9bb7c4d9ae3d4611da6dbf2a406f Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Wed, 28 Apr 2010 07:57:37 +0000
Subject: [PATCH 292/532]  - When canceling jaddrefs they may not yet be in the
 journal if this is via    a revert call.  In this case don't attempt to
 remove something that    has not yet been added.  Otherwise this jaddref must
 hang around    to prevent the bitmap write as normal.

---
 sys/ufs/ffs/ffs_softdep.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index b08e7c299d4..b05339b936c 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -3504,7 +3504,8 @@ cancel_jaddref(jaddref, inodedep, wkhd)
 		WORKLIST_INSERT(wkhd, &jsegdep->jd_list);
 	} else {
 		free_jsegdep(jsegdep);
-		remove_from_journal(&jaddref->ja_list);
+		if (jaddref->ja_state & DEPCOMPLETE)
+			remove_from_journal(&jaddref->ja_list);
 	}
 	/*
 	 * Leave NEWBLOCK jaddrefs on the inodedep so handle_workitem_remove

From f276912e6fa1da1174ba900a7e5d6447d71fc8e7 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Dag-Erling=20Sm=C3=B8rgrav?= <des@FreeBSD.org>
Date: Wed, 28 Apr 2010 08:37:00 +0000
Subject: [PATCH 293/532] Vendor import of OpenSSH 5.5p1

---
 ChangeLog                       | 118 ++++++++++++-
 Makefile.in                     |  29 ++--
 README                          |   4 +-
 auth-options.c                  |   6 +-
 auth2-pubkey.c                  |  31 ++--
 channels.c                      |   6 +-
 clientloop.c                    |   3 +-
 config.h.in                     |  12 +-
 configure                       | 293 +++++++++++++++++++++++---------
 configure.ac                    |  74 ++++----
 contrib/Makefile                |   2 +-
 contrib/caldera/openssh.spec    |   6 +-
 contrib/cygwin/Makefile         |   4 +-
 contrib/cygwin/ssh-host-config  |   2 +-
 contrib/redhat/openssh.spec     |   2 +-
 contrib/ssh-copy-id             |   2 +-
 contrib/suse/openssh.spec       |   2 +-
 defines.h                       |   4 +-
 key.c                           |  15 +-
 key.h                           |   3 +-
 loginrec.c                      |  31 ++++
 logintest.c                     |   2 +-
 moduli.0                        |   2 +-
 openbsd-compat/bsd-arc4random.c |   4 +-
 scp.0                           |   2 +-
 servconf.c                      |  19 ++-
 session.c                       |   8 +-
 sftp-server.0                   |   2 +-
 sftp.0                          |   2 +-
 ssh-add.0                       |   2 +-
 ssh-agent.0                     |   2 +-
 ssh-keygen.0                    |  44 ++---
 ssh-keygen.1                    |  43 +++--
 ssh-keygen.c                    |   5 +-
 ssh-keyscan.0                   |   2 +-
 ssh-keysign.0                   |   2 +-
 ssh-pkcs11-helper.0             |   2 +-
 ssh-pkcs11-helper.c             |   4 +-
 ssh-rand-helper.0               |   2 +-
 ssh.0                           |   9 +-
 ssh.1                           |   9 +-
 ssh_config.0                    |   6 +-
 ssh_config.5                    |  10 +-
 sshd.0                          |   2 +-
 sshd_config.0                   |   2 +-
 version.h                       |   4 +-
 46 files changed, 574 insertions(+), 266 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index d6e4a4a25ae..39e0ba45dd7 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,4 +1,120 @@
-20100307
+20100410
+ - (dtucker) [configure.ac] Put the check for the existence of getaddrinfo
+   back so we disable the IPv6 tests if we don't have it.
+
+20100409
+ - (dtucker) [contrib/cygwin/Makefile] Don't overwrite files with the wrong
+   ones.  Based on a patch from Roumen Petrov.
+ - (dtucker) [configure.ac] Bug #1744: use pkg-config for libedit flags if we
+   have it and the path is not provided to --with-libedit.  Based on a patch
+   from Iain Morgan.
+ - (dtucker) [configure.ac defines.h loginrec.c logintest.c] Bug #1732: enable
+   utmpx support on FreeBSD where possible.  Patch from Ed Schouten, ok djm@
+
+20100326
+ - (djm) [openbsd-compat/bsd-arc4random.c] Fix preprocessor detection
+   for arc4random_buf() and arc4random_uniform(); from Josh Gilkerson
+ - (dtucker) [configure.ac] Bug #1741: Add section for Haiku, patch originally
+   by Ingo Weinhold via Scott McCreary, ok djm@
+ - (djm) OpenBSD CVS Sync
+   - djm@cvs.openbsd.org 2010/03/25 23:38:28
+     [servconf.c]
+     from portable: getcwd(NULL, 0) doesn't work on all platforms, so
+     use a stack buffer; ok dtucker@
+   - djm@cvs.openbsd.org 2010/03/26 00:26:58
+     [ssh.1]
+     mention that -S none disables connection sharing; from Colin Watson
+ - (djm) [session.c] Allow ChrootDirectory to work on SELinux platforms -
+   set up SELinux execution context before chroot() call. From Russell
+   Coker via Colin watson; bz#1726 ok dtucker@
+ - (djm) [channels.c] Check for EPFNOSUPPORT as a socket() errno; bz#1721
+   ok dtucker@
+ - (dtucker) Bug #1725: explicitly link libX11 into gnome-ssh-askpass2 using
+   pkg-config, patch from Colin Watson.  Needed for newer linkers (ie gold).
+ - (djm) [contrib/ssh-copy-id] Don't blow up when the agent has no keys;
+   bz#1723 patch from Adeodato Simóvia Colin Watson; ok dtucker@
+ - (dtucker) OpenBSD CVS Sync
+   - dtucker@cvs.openbsd.org 2010/03/26 01:06:13
+     [ssh_config.5]
+     Reformat default value of PreferredAuthentications entry (current
+     formatting implies ", " is acceptable as a separator, which it's not.
+     ok djm@
+
+20100324
+ - (dtucker) [contrib/cygwin/ssh-host-config] Mount the Windows directory
+   containing the services file explicitely case-insensitive.  This allows to
+   tweak the Windows services file reliably.  Patch from vinschen at redhat.
+
+20100321
+ - (djm) OpenBSD CVS Sync
+   - jmc@cvs.openbsd.org 2010/03/08 09:41:27
+     [ssh-keygen.1]
+     sort the list of constraints (to -O); ok djm
+   - jmc@cvs.openbsd.org 2010/03/10 07:40:35
+     [ssh-keygen.1]
+     typos; from Ross Richardson
+     closes prs 6334 and 6335
+   - djm@cvs.openbsd.org 2010/03/10 23:27:17
+     [auth2-pubkey.c]
+     correct certificate logging and make it more consistent between
+     authorized_keys and TrustedCAKeys; ok markus@
+   - djm@cvs.openbsd.org 2010/03/12 01:06:25
+     [servconf.c]
+     unbreak AuthorizedKeys option with a $HOME-relative path; reported by
+     vinschen AT redhat.com, ok dtucker@
+   - markus@cvs.openbsd.org 2010/03/12 11:37:40
+     [servconf.c]
+     do not prepend AuthorizedKeysFile with getcwd(), unbreaks relative paths
+     free() (not xfree()) the buffer returned by getcwd()
+   - djm@cvs.openbsd.org 2010/03/13 21:10:38
+     [clientloop.c]
+     protocol conformance fix: send language tag when disconnecting normally;
+     spotted by 1.41421 AT gmail.com, ok markus@ deraadt@
+   - djm@cvs.openbsd.org 2010/03/13 21:45:46
+     [ssh-keygen.1]
+     Certificates are named *-cert.pub, not *_cert.pub; committing a diff
+     from stevesk@ ok me
+   - jmc@cvs.openbsd.org 2010/03/13 23:38:13
+     [ssh-keygen.1]
+     fix a formatting error (args need quoted); noted by stevesk
+   - stevesk@cvs.openbsd.org 2010/03/15 19:40:02
+     [key.c key.h ssh-keygen.c]
+     also print certificate type (user or host) for ssh-keygen -L
+     ok djm kettenis
+   - stevesk@cvs.openbsd.org 2010/03/16 15:46:52
+     [auth-options.c]
+     spelling in error message. ok djm kettenis
+   - djm@cvs.openbsd.org 2010/03/16 16:36:49
+     [version.h]
+     crank version to openssh-5.5 since we have a few fixes since 5.4;
+     requested deraadt@ kettenis@
+ - (djm) [README contrib/caldera/openssh.spec contrib/redhat/openssh.spec]
+   [contrib/suse/openssh.spec] Crank version numbers
+
+20100314
+ - (djm) [ssh-pkcs11-helper.c] Move #ifdef to after #defines to fix
+   compilation failure when !HAVE_DLOPEN. Reported by felix-mindrot
+   AT fefe.de
+ - (djm) [Makefile.in] Respecify -lssh after -lopenbsd-compat for
+   ssh-pkcs11-helper to repair static builds (we do the same for
+   ssh-keyscan). Reported by felix-mindrot AT fefe.de
+
+20100312
+ - (tim) [Makefile.in] Now that scard is gone, no need to make $(datadir)
+ - (tim) [Makefile.in] Add missing $(EXEEXT) to install targets.
+   Patch from Corinna Vinschen.
+ - (tim) [contrib/cygwin/Makefile] Fix list of documentation files to install
+   on a Cygwin installation. Patch from Corinna Vinschen.
+
+20100311
+ - (tim) [contrib/suse/openssh.spec] crank version number here too.
+   report by imorgan AT nas.nasa.gov
+
+20100309
+ - (dtucker) [configure.ac] Use a proper AC_CHECK_DECL for BROKEN_GETADDRINFO
+   so setting it in CFLAGS correctly skips IPv6 tests.
+
+20100308
  - (djm) OpenBSD CVS Sync
    - djm@cvs.openbsd.org 2010/03/07 22:16:01
      [ssh-keygen.c]
diff --git a/Makefile.in b/Makefile.in
index 69e3567da2d..476674b0e49 100644
--- a/Makefile.in
+++ b/Makefile.in
@@ -1,4 +1,4 @@
-# $Id: Makefile.in,v 1.306 2010/02/24 07:18:51 djm Exp $
+# $Id: Makefile.in,v 1.309 2010/03/13 21:41:34 djm Exp $
 
 # uncomment if you run a non bourne compatable shell. Ie. csh
 #SHELL = @SH@
@@ -160,7 +160,7 @@ ssh-keysign$(EXEEXT): $(LIBCOMPAT) libssh.a ssh-keysign.o roaming_dummy.o readco
 	$(LD) -o $@ ssh-keysign.o readconf.o roaming_dummy.o $(LDFLAGS) -lssh -lopenbsd-compat $(LIBS)
 
 ssh-pkcs11-helper$(EXEEXT): $(LIBCOMPAT) libssh.a ssh-pkcs11-helper.o ssh-pkcs11.o
-	$(LD) -o $@ ssh-pkcs11-helper.o ssh-pkcs11.o $(LDFLAGS) -lssh -lopenbsd-compat $(LIBS)
+	$(LD) -o $@ ssh-pkcs11-helper.o ssh-pkcs11.o $(LDFLAGS) -lssh -lopenbsd-compat -lssh $(LIBS)
 
 ssh-keyscan$(EXEEXT): $(LIBCOMPAT) libssh.a ssh-keyscan.o roaming_dummy.o
 	$(LD) -o $@ ssh-keyscan.o roaming_dummy.o $(LDFLAGS) -lssh -lopenbsd-compat -lssh $(LIBS)
@@ -249,26 +249,25 @@ install-files:
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(bindir)
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(sbindir)
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(mandir)
-	$(srcdir)/mkinstalldirs $(DESTDIR)$(datadir)
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(mandir)/$(mansubdir)1
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(mandir)/$(mansubdir)5
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(mandir)/$(mansubdir)8
 	$(srcdir)/mkinstalldirs $(DESTDIR)$(libexecdir)
 	(umask 022 ; $(srcdir)/mkinstalldirs $(DESTDIR)$(PRIVSEP_PATH))
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh $(DESTDIR)$(bindir)/ssh
-	$(INSTALL) -m 0755 $(STRIP_OPT) scp $(DESTDIR)$(bindir)/scp
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-add $(DESTDIR)$(bindir)/ssh-add
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-agent $(DESTDIR)$(bindir)/ssh-agent
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-keygen $(DESTDIR)$(bindir)/ssh-keygen
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-keyscan $(DESTDIR)$(bindir)/ssh-keyscan
-	$(INSTALL) -m 0755 $(STRIP_OPT) sshd $(DESTDIR)$(sbindir)/sshd
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh$(EXEEXT) $(DESTDIR)$(bindir)/ssh$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) scp$(EXEEXT) $(DESTDIR)$(bindir)/scp$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-add$(EXEEXT) $(DESTDIR)$(bindir)/ssh-add$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-agent$(EXEEXT) $(DESTDIR)$(bindir)/ssh-agent$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-keygen$(EXEEXT) $(DESTDIR)$(bindir)/ssh-keygen$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-keyscan$(EXEEXT) $(DESTDIR)$(bindir)/ssh-keyscan$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) sshd$(EXEEXT) $(DESTDIR)$(sbindir)/sshd$(EXEEXT)
 	if test ! -z "$(INSTALL_SSH_RAND_HELPER)" ; then \
-		$(INSTALL) -m 0755 $(STRIP_OPT) ssh-rand-helper $(DESTDIR)$(libexecdir)/ssh-rand-helper ; \
+		$(INSTALL) -m 0755 $(STRIP_OPT) ssh-rand-helper$(EXEEXT) $(DESTDIR)$(libexecdir)/ssh-rand-helper$(EXEEXT) ; \
 	fi
-	$(INSTALL) -m 4711 $(STRIP_OPT) ssh-keysign $(DESTDIR)$(SSH_KEYSIGN)
-	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-pkcs11-helper $(DESTDIR)$(SSH_PKCS11_HELPER)
-	$(INSTALL) -m 0755 $(STRIP_OPT) sftp $(DESTDIR)$(bindir)/sftp
-	$(INSTALL) -m 0755 $(STRIP_OPT) sftp-server $(DESTDIR)$(SFTP_SERVER)
+	$(INSTALL) -m 4711 $(STRIP_OPT) ssh-keysign$(EXEEXT) $(DESTDIR)$(SSH_KEYSIGN)$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) ssh-pkcs11-helper$(EXEEXT) $(DESTDIR)$(SSH_PKCS11_HELPER)$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) sftp$(EXEEXT) $(DESTDIR)$(bindir)/sftp$(EXEEXT)
+	$(INSTALL) -m 0755 $(STRIP_OPT) sftp-server$(EXEEXT) $(DESTDIR)$(SFTP_SERVER)$(EXEEXT)
 	$(INSTALL) -m 644 ssh.1.out $(DESTDIR)$(mandir)/$(mansubdir)1/ssh.1
 	$(INSTALL) -m 644 scp.1.out $(DESTDIR)$(mandir)/$(mansubdir)1/scp.1
 	$(INSTALL) -m 644 ssh-add.1.out $(DESTDIR)$(mandir)/$(mansubdir)1/ssh-add.1
diff --git a/README b/README
index 0ecb670b6d8..a29f2007dca 100644
--- a/README
+++ b/README
@@ -1,4 +1,4 @@
-See http://www.openssh.com/txt/release-5.4 for the release notes.
+See http://www.openssh.com/txt/release-5.5 for the release notes.
 
 - A Japanese translation of this document and of the OpenSSH FAQ is
 - available at http://www.unixuser.org/~haruyama/security/openssh/index.html
@@ -62,4 +62,4 @@ References -
 [6] http://www.openbsd.org/cgi-bin/man.cgi?query=style&sektion=9
 [7] http://www.openssh.com/faq.html
 
-$Id: README,v 1.72 2010/03/07 22:41:02 djm Exp $
+$Id: README,v 1.73 2010/03/21 19:11:55 djm Exp $
diff --git a/auth-options.c b/auth-options.c
index 129301765b7..69b314fbd11 100644
--- a/auth-options.c
+++ b/auth-options.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: auth-options.c,v 1.48 2010/03/07 11:57:13 dtucker Exp $ */
+/* $OpenBSD: auth-options.c,v 1.49 2010/03/16 15:46:52 stevesk Exp $ */
 /*
  * Author: Tatu Ylonen <ylo@cs.hut.fi>
  * Copyright (c) 1995 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland
@@ -434,7 +434,7 @@ auth_cert_constraints(Buffer *c_orig, struct passwd *pw)
 				goto out;
 			}
 			if (strlen(command) != clen) {
-				error("force-command constrain contains \\0");
+				error("force-command constraint contains \\0");
 				goto out;
 			}
 			if (cert_forced_command != NULL) {
@@ -454,7 +454,7 @@ auth_cert_constraints(Buffer *c_orig, struct passwd *pw)
 				goto out;
 			}
 			if (strlen(allowed) != clen) {
-				error("source-address constrain contains \\0");
+				error("source-address constraint contains \\0");
 				goto out;
 			}
 			if (cert_source_address_done++) {
diff --git a/auth2-pubkey.c b/auth2-pubkey.c
index 51aa774872b..c4cadf4e78c 100644
--- a/auth2-pubkey.c
+++ b/auth2-pubkey.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: auth2-pubkey.c,v 1.21 2010/03/04 10:36:03 djm Exp $ */
+/* $OpenBSD: auth2-pubkey.c,v 1.22 2010/03/10 23:27:17 djm Exp $ */
 /*
  * Copyright (c) 2000 Markus Friedl.  All rights reserved.
  *
@@ -240,22 +240,26 @@ user_key_allowed2(struct passwd *pw, Key *key, char *file)
 				continue;
 			if (!key_equal(found, key->cert->signature_key))
 				continue;
-			debug("matching CA found: file %s, line %lu",
-			    file, linenum);
 			fp = key_fingerprint(found, SSH_FP_MD5,
 			    SSH_FP_HEX);
-			verbose("Found matching %s CA: %s",
-			    key_type(found), fp);
-			xfree(fp);
+			debug("matching CA found: file %s, line %lu, %s %s",
+			    file, linenum, key_type(found), fp);
 			if (key_cert_check_authority(key, 0, 0, pw->pw_name,
 			    &reason) != 0) {
+				xfree(fp);
 				error("%s", reason);
 				auth_debug_add("%s", reason);
 				continue;
 			}
 			if (auth_cert_constraints(&key->cert->constraints,
-			    pw) != 0)
+			    pw) != 0) {
+				xfree(fp);
 				continue;
+			}
+			verbose("Accepted certificate ID \"%s\" "
+			    "signed by %s CA %s via %s", key->cert->key_id,
+			    key_type(found), fp, file);
+			xfree(fp);
 			found_key = 1;
 			break;
 		} else if (!key_is_cert_authority && key_equal(found, key)) {
@@ -281,15 +285,15 @@ user_key_allowed2(struct passwd *pw, Key *key, char *file)
 static int
 user_cert_trusted_ca(struct passwd *pw, Key *key)
 {
-	char *key_fp, *ca_fp;
+	char *ca_fp;
 	const char *reason;
 	int ret = 0;
 
 	if (!key_is_cert(key) || options.trusted_user_ca_keys == NULL)
 		return 0;
 
-	key_fp = key_fingerprint(key, SSH_FP_MD5, SSH_FP_HEX);
-	ca_fp = key_fingerprint(key, SSH_FP_MD5, SSH_FP_HEX);
+	ca_fp = key_fingerprint(key->cert->signature_key,
+	    SSH_FP_MD5, SSH_FP_HEX);
 
 	if (key_in_file(key->cert->signature_key,
 	    options.trusted_user_ca_keys, 1) != 1) {
@@ -306,13 +310,12 @@ user_cert_trusted_ca(struct passwd *pw, Key *key)
 	if (auth_cert_constraints(&key->cert->constraints, pw) != 0)
 		goto out;
 
-	verbose("%s certificate %s allowed by trusted %s key %s",
-	    key_type(key), key_fp, key_type(key->cert->signature_key), ca_fp);
+	verbose("Accepted certificate ID \"%s\" signed by %s CA %s via %s",
+	    key->cert->key_id, key_type(key->cert->signature_key), ca_fp,
+	    options.trusted_user_ca_keys);
 	ret = 1;
 
  out:
-	if (key_fp != NULL)
-		xfree(key_fp);
 	if (ca_fp != NULL)
 		xfree(ca_fp);
 	return ret;
diff --git a/channels.c b/channels.c
index d8c53a4a8f2..a55d2781757 100644
--- a/channels.c
+++ b/channels.c
@@ -3252,7 +3252,11 @@ x11_create_display_inet(int x11_display_offset, int x11_use_localhost,
 			sock = socket(ai->ai_family, ai->ai_socktype,
 			    ai->ai_protocol);
 			if (sock < 0) {
-				if ((errno != EINVAL) && (errno != EAFNOSUPPORT)) {
+				if ((errno != EINVAL) && (errno != EAFNOSUPPORT)
+#ifdef EPFNOSUPPORT
+				    && (errno != EPFNOSUPPORT)
+#endif 
+				    ) {
 					error("socket: %.100s", strerror(errno));
 					freeaddrinfo(aitop);
 					return -1;
diff --git a/clientloop.c b/clientloop.c
index 6ffef95a2d7..9ab56b44c47 100644
--- a/clientloop.c
+++ b/clientloop.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: clientloop.c,v 1.218 2010/01/28 00:21:18 djm Exp $ */
+/* $OpenBSD: clientloop.c,v 1.219 2010/03/13 21:10:38 djm Exp $ */
 /*
  * Author: Tatu Ylonen <ylo@cs.hut.fi>
  * Copyright (c) 1995 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland
@@ -1484,6 +1484,7 @@ client_loop(int have_pty, int escape_char_arg, int ssh2_chan_id)
 		packet_start(SSH2_MSG_DISCONNECT);
 		packet_put_int(SSH2_DISCONNECT_BY_APPLICATION);
 		packet_put_cstring("disconnected by user");
+		packet_put_cstring(""); /* language tag */
 		packet_send();
 		packet_write_wait();
 	}
diff --git a/config.h.in b/config.h.in
index a61dec6095d..a43ad57115d 100644
--- a/config.h.in
+++ b/config.h.in
@@ -80,9 +80,6 @@
 /* Define if you want to specify the path to your lastlog file */
 #undef CONF_LASTLOG_FILE
 
-/* Define if you want to specify the path to your utmpx file */
-#undef CONF_UTMPX_FILE
-
 /* Define if you want to specify the path to your utmp file */
 #undef CONF_UTMP_FILE
 
@@ -455,6 +452,9 @@
 /* Define to 1 if you have the `getutxline' function. */
 #undef HAVE_GETUTXLINE
 
+/* Define to 1 if you have the `getutxuser' function. */
+#undef HAVE_GETUTXUSER
+
 /* Define to 1 if you have the `get_default_context_with_level' function. */
 #undef HAVE_GET_DEFAULT_CONTEXT_WITH_LEVEL
 
@@ -551,6 +551,9 @@
 /* Define if system has libiaf that supports set_id */
 #undef HAVE_LIBIAF
 
+/* Define to 1 if you have the `network' library (-lnetwork). */
+#undef HAVE_LIBNETWORK
+
 /* Define to 1 if you have the `nsl' library (-lnsl). */
 #undef HAVE_LIBNSL
 
@@ -804,6 +807,9 @@
 /* Define to 1 if you have the `setutent' function. */
 #undef HAVE_SETUTENT
 
+/* Define to 1 if you have the `setutxdb' function. */
+#undef HAVE_SETUTXDB
+
 /* Define to 1 if you have the `setutxent' function. */
 #undef HAVE_SETUTXENT
 
diff --git a/configure b/configure
index 7c19fc204a5..5ebdb15c610 100755
--- a/configure
+++ b/configure
@@ -1,5 +1,5 @@
 #! /bin/sh
-# From configure.ac Revision: 1.444 .
+# From configure.ac Revision: 1.449 .
 # Guess values for system-dependent variables and create Makefiles.
 # Generated by GNU Autoconf 2.61 for OpenSSH Portable.
 #
@@ -693,6 +693,7 @@ LOGIN_PROGRAM_FALLBACK
 PATH_PASSWD_PROG
 LD
 SSHDLIBS
+PKGCONFIG
 LIBEDIT
 INSTALL_SSH_RAND_HELPER
 SSH_PRIVSEP_USER
@@ -7435,6 +7436,85 @@ fi
 *-*-dragonfly*)
 	SSHDLIBS="$SSHDLIBS -lcrypt"
 	;;
+*-*-haiku*)
+    LIBS="$LIBS -lbsd "
+
+{ echo "$as_me:$LINENO: checking for socket in -lnetwork" >&5
+echo $ECHO_N "checking for socket in -lnetwork... $ECHO_C" >&6; }
+if test "${ac_cv_lib_network_socket+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  ac_check_lib_save_LIBS=$LIBS
+LIBS="-lnetwork  $LIBS"
+cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+
+/* Override any GCC internal prototype to avoid an error.
+   Use char because int might match the return type of a GCC
+   builtin and then its argument prototype would still apply.  */
+#ifdef __cplusplus
+extern "C"
+#endif
+char socket ();
+int
+main ()
+{
+return socket ();
+  ;
+  return 0;
+}
+_ACEOF
+rm -f conftest.$ac_objext conftest$ac_exeext
+if { (ac_try="$ac_link"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_link") 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } && {
+	 test -z "$ac_c_werror_flag" ||
+	 test ! -s conftest.err
+       } && test -s conftest$ac_exeext &&
+       $as_test_x conftest$ac_exeext; then
+  ac_cv_lib_network_socket=yes
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+	ac_cv_lib_network_socket=no
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest_ipa8_conftest.oo \
+      conftest$ac_exeext conftest.$ac_ext
+LIBS=$ac_check_lib_save_LIBS
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_lib_network_socket" >&5
+echo "${ECHO_T}$ac_cv_lib_network_socket" >&6; }
+if test $ac_cv_lib_network_socket = yes; then
+  cat >>confdefs.h <<_ACEOF
+#define HAVE_LIBNETWORK 1
+_ACEOF
+
+  LIBS="-lnetwork $LIBS"
+
+fi
+
+    cat >>confdefs.h <<\_ACEOF
+#define HAVE_U_INT64_T 1
+_ACEOF
+
+    MANTYPE=man
+    ;;
 *-*-hpux*)
 	# first we define all of the options common to all HP-UX releases
 	CPPFLAGS="$CPPFLAGS -D_HPUX_SOURCE -D_XOPEN_SOURCE -D_XOPEN_SOURCE_EXTENDED=1"
@@ -12326,7 +12406,61 @@ LIBEDIT_MSG="no"
 # Check whether --with-libedit was given.
 if test "${with_libedit+set}" = set; then
   withval=$with_libedit;  if test "x$withval" != "xno" ; then
-		if test "x$withval" != "xyes"; then
+		if test "x$withval" = "xyes" ; then
+			# Extract the first word of "pkg-config", so it can be a program name with args.
+set dummy pkg-config; ac_word=$2
+{ echo "$as_me:$LINENO: checking for $ac_word" >&5
+echo $ECHO_N "checking for $ac_word... $ECHO_C" >&6; }
+if test "${ac_cv_path_PKGCONFIG+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  case $PKGCONFIG in
+  [\\/]* | ?:[\\/]*)
+  ac_cv_path_PKGCONFIG="$PKGCONFIG" # Let the user override the test with a path.
+  ;;
+  *)
+  as_save_IFS=$IFS; IFS=$PATH_SEPARATOR
+for as_dir in $PATH
+do
+  IFS=$as_save_IFS
+  test -z "$as_dir" && as_dir=.
+  for ac_exec_ext in '' $ac_executable_extensions; do
+  if { test -f "$as_dir/$ac_word$ac_exec_ext" && $as_test_x "$as_dir/$ac_word$ac_exec_ext"; }; then
+    ac_cv_path_PKGCONFIG="$as_dir/$ac_word$ac_exec_ext"
+    echo "$as_me:$LINENO: found $as_dir/$ac_word$ac_exec_ext" >&5
+    break 2
+  fi
+done
+done
+IFS=$as_save_IFS
+
+  test -z "$ac_cv_path_PKGCONFIG" && ac_cv_path_PKGCONFIG="no"
+  ;;
+esac
+fi
+PKGCONFIG=$ac_cv_path_PKGCONFIG
+if test -n "$PKGCONFIG"; then
+  { echo "$as_me:$LINENO: result: $PKGCONFIG" >&5
+echo "${ECHO_T}$PKGCONFIG" >&6; }
+else
+  { echo "$as_me:$LINENO: result: no" >&5
+echo "${ECHO_T}no" >&6; }
+fi
+
+
+			if test "x$PKGCONFIG" != "xno"; then
+				{ echo "$as_me:$LINENO: checking if $PKGCONFIG knows about libedit" >&5
+echo $ECHO_N "checking if $PKGCONFIG knows about libedit... $ECHO_C" >&6; }
+			 	if "$PKGCONFIG" libedit; then
+					{ echo "$as_me:$LINENO: result: yes" >&5
+echo "${ECHO_T}yes" >&6; }
+					use_pkgconfig_for_libedit=yes
+				else
+					{ echo "$as_me:$LINENO: result: no" >&5
+echo "${ECHO_T}no" >&6; }
+				fi
+			fi
+		else
 			CPPFLAGS="$CPPFLAGS -I${withval}/include"
 			if test -n "${need_dash_r}"; then
 				LDFLAGS="-L${withval}/lib -R${withval}/lib ${LDFLAGS}"
@@ -12334,13 +12468,20 @@ if test "${with_libedit+set}" = set; then
 				LDFLAGS="-L${withval}/lib ${LDFLAGS}"
 			fi
 		fi
+		if test "x$use_pkgconfig_for_libedit" == "xyes"; then
+			LIBEDIT=`$PKGCONFIG --libs-only-l libedit`
+			CPPFLAGS="$CPPFLAGS `$PKGCONFIG --cflags libedit`"
+		else
+			LIBEDIT="-ledit -lcurses"
+		fi
+		OTHERLIBS=`echo $LIBEDIT | sed 's/-ledit//'`
 		{ echo "$as_me:$LINENO: checking for el_init in -ledit" >&5
 echo $ECHO_N "checking for el_init in -ledit... $ECHO_C" >&6; }
 if test "${ac_cv_lib_edit_el_init+set}" = set; then
   echo $ECHO_N "(cached) $ECHO_C" >&6
 else
   ac_check_lib_save_LIBS=$LIBS
-LIBS="-ledit  -lcurses
+LIBS="-ledit  $OTHERLIBS
 		 $LIBS"
 cat >conftest.$ac_ext <<_ACEOF
 /* confdefs.h.  */
@@ -12402,7 +12543,6 @@ cat >>confdefs.h <<\_ACEOF
 #define USE_LIBEDIT 1
 _ACEOF
 
-			  LIBEDIT="-ledit -lcurses"
 			  LIBEDIT_MSG="yes"
 
 
@@ -14961,7 +15101,8 @@ done
 
 
 
-for ac_func in endutxent getutxent getutxid getutxline pututxline
+
+for ac_func in endutxent getutxent getutxid getutxline getutxuser pututxline
 do
 as_ac_var=`echo "ac_cv_func_$ac_func" | $as_tr_sh`
 { echo "$as_me:$LINENO: checking for $ac_func" >&5
@@ -15056,7 +15197,8 @@ done
 
 
 
-for ac_func in setutxent utmpxname
+
+for ac_func in setutxdb setutxent utmpxname
 do
 as_ac_var=`echo "ac_cv_func_$ac_func" | $as_tr_sh`
 { echo "$as_me:$LINENO: checking for $ac_func" >&5
@@ -29373,77 +29515,6 @@ _ACEOF
 fi
 
 
-{ echo "$as_me:$LINENO: checking if your system defines UTMPX_FILE" >&5
-echo $ECHO_N "checking if your system defines UTMPX_FILE... $ECHO_C" >&6; }
-cat >conftest.$ac_ext <<_ACEOF
-/* confdefs.h.  */
-_ACEOF
-cat confdefs.h >>conftest.$ac_ext
-cat >>conftest.$ac_ext <<_ACEOF
-/* end confdefs.h.  */
-
-#include <sys/types.h>
-#include <utmp.h>
-#ifdef HAVE_UTMPX_H
-#include <utmpx.h>
-#endif
-#ifdef HAVE_PATHS_H
-#  include <paths.h>
-#endif
-
-int
-main ()
-{
- char *utmpx = UTMPX_FILE;
-  ;
-  return 0;
-}
-_ACEOF
-rm -f conftest.$ac_objext
-if { (ac_try="$ac_compile"
-case "(($ac_try" in
-  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
-  *) ac_try_echo=$ac_try;;
-esac
-eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
-  (eval "$ac_compile") 2>conftest.er1
-  ac_status=$?
-  grep -v '^ *+' conftest.er1 >conftest.err
-  rm -f conftest.er1
-  cat conftest.err >&5
-  echo "$as_me:$LINENO: \$? = $ac_status" >&5
-  (exit $ac_status); } && {
-	 test -z "$ac_c_werror_flag" ||
-	 test ! -s conftest.err
-       } && test -s conftest.$ac_objext; then
-   { echo "$as_me:$LINENO: result: yes" >&5
-echo "${ECHO_T}yes" >&6; }
-else
-  echo "$as_me: failed program was:" >&5
-sed 's/^/| /' conftest.$ac_ext >&5
-
-	 { echo "$as_me:$LINENO: result: no" >&5
-echo "${ECHO_T}no" >&6; }
-	  system_utmpx_path=no
-
-fi
-
-rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
-if test -z "$conf_utmpx_location"; then
-	if test x"$system_utmpx_path" = x"no" ; then
-		cat >>confdefs.h <<\_ACEOF
-#define DISABLE_UTMPX 1
-_ACEOF
-
-	fi
-else
-
-cat >>confdefs.h <<_ACEOF
-#define CONF_UTMPX_FILE "$conf_utmpx_location"
-_ACEOF
-
-fi
-
 { echo "$as_me:$LINENO: checking if your system defines WTMPX_FILE" >&5
 echo $ECHO_N "checking if your system defines WTMPX_FILE... $ECHO_C" >&6; }
 cat >conftest.$ac_ext <<_ACEOF
@@ -29524,14 +29595,69 @@ fi
 
 CFLAGS="$CFLAGS $werror_flags"
 
-if grep "#define BROKEN_GETADDRINFO 1" confdefs.h >/dev/null || \
-    test "x$ac_cv_func_getaddrinfo" != "xyes" ; then
+if test "x$ac_cv_func_getaddrinfo" != "xyes" ; then
 	TEST_SSH_IPV6=no
-
 else
 	TEST_SSH_IPV6=yes
-
 fi
+{ echo "$as_me:$LINENO: checking whether BROKEN_GETADDRINFO is declared" >&5
+echo $ECHO_N "checking whether BROKEN_GETADDRINFO is declared... $ECHO_C" >&6; }
+if test "${ac_cv_have_decl_BROKEN_GETADDRINFO+set}" = set; then
+  echo $ECHO_N "(cached) $ECHO_C" >&6
+else
+  cat >conftest.$ac_ext <<_ACEOF
+/* confdefs.h.  */
+_ACEOF
+cat confdefs.h >>conftest.$ac_ext
+cat >>conftest.$ac_ext <<_ACEOF
+/* end confdefs.h.  */
+$ac_includes_default
+int
+main ()
+{
+#ifndef BROKEN_GETADDRINFO
+  (void) BROKEN_GETADDRINFO;
+#endif
+
+  ;
+  return 0;
+}
+_ACEOF
+rm -f conftest.$ac_objext
+if { (ac_try="$ac_compile"
+case "(($ac_try" in
+  *\"* | *\`* | *\\*) ac_try_echo=\$ac_try;;
+  *) ac_try_echo=$ac_try;;
+esac
+eval "echo \"\$as_me:$LINENO: $ac_try_echo\"") >&5
+  (eval "$ac_compile") 2>conftest.er1
+  ac_status=$?
+  grep -v '^ *+' conftest.er1 >conftest.err
+  rm -f conftest.er1
+  cat conftest.err >&5
+  echo "$as_me:$LINENO: \$? = $ac_status" >&5
+  (exit $ac_status); } && {
+	 test -z "$ac_c_werror_flag" ||
+	 test ! -s conftest.err
+       } && test -s conftest.$ac_objext; then
+  ac_cv_have_decl_BROKEN_GETADDRINFO=yes
+else
+  echo "$as_me: failed program was:" >&5
+sed 's/^/| /' conftest.$ac_ext >&5
+
+	ac_cv_have_decl_BROKEN_GETADDRINFO=no
+fi
+
+rm -f core conftest.err conftest.$ac_objext conftest.$ac_ext
+fi
+{ echo "$as_me:$LINENO: result: $ac_cv_have_decl_BROKEN_GETADDRINFO" >&5
+echo "${ECHO_T}$ac_cv_have_decl_BROKEN_GETADDRINFO" >&6; }
+if test $ac_cv_have_decl_BROKEN_GETADDRINFO = yes; then
+  TEST_SSH_IPV6=no
+fi
+
+TEST_SSH_IPV6=$TEST_SSH_IPV6
+
 
 
 ac_config_files="$ac_config_files Makefile buildpkg.sh opensshd.init openssh.xml openbsd-compat/Makefile openbsd-compat/regress/Makefile ssh_prng_cmds survey.sh"
@@ -30236,6 +30362,7 @@ LOGIN_PROGRAM_FALLBACK!$LOGIN_PROGRAM_FALLBACK$ac_delim
 PATH_PASSWD_PROG!$PATH_PASSWD_PROG$ac_delim
 LD!$LD$ac_delim
 SSHDLIBS!$SSHDLIBS$ac_delim
+PKGCONFIG!$PKGCONFIG$ac_delim
 LIBEDIT!$LIBEDIT$ac_delim
 INSTALL_SSH_RAND_HELPER!$INSTALL_SSH_RAND_HELPER$ac_delim
 SSH_PRIVSEP_USER!$SSH_PRIVSEP_USER$ac_delim
@@ -30255,7 +30382,6 @@ PROG_VMSTAT!$PROG_VMSTAT$ac_delim
 PROG_UPTIME!$PROG_UPTIME$ac_delim
 PROG_IPCS!$PROG_IPCS$ac_delim
 PROG_TAIL!$PROG_TAIL$ac_delim
-INSTALL_SSH_PRNG_CMDS!$INSTALL_SSH_PRNG_CMDS$ac_delim
 _ACEOF
 
   if test `sed -n "s/.*$ac_delim\$/X/p" conf$$subs.sed | grep -c X` = 97; then
@@ -30297,6 +30423,7 @@ _ACEOF
 ac_delim='%!_!# '
 for ac_last_try in false false false false false :; do
   cat >conf$$subs.sed <<_ACEOF
+INSTALL_SSH_PRNG_CMDS!$INSTALL_SSH_PRNG_CMDS$ac_delim
 KRB5CONF!$KRB5CONF$ac_delim
 PRIVSEP_PATH!$PRIVSEP_PATH$ac_delim
 xauth_path!$xauth_path$ac_delim
@@ -30312,7 +30439,7 @@ LIBOBJS!$LIBOBJS$ac_delim
 LTLIBOBJS!$LTLIBOBJS$ac_delim
 _ACEOF
 
-  if test `sed -n "s/.*$ac_delim\$/X/p" conf$$subs.sed | grep -c X` = 13; then
+  if test `sed -n "s/.*$ac_delim\$/X/p" conf$$subs.sed | grep -c X` = 14; then
     break
   elif $ac_last_try; then
     { { echo "$as_me:$LINENO: error: could not make $CONFIG_STATUS" >&5
diff --git a/configure.ac b/configure.ac
index a9960a7baff..0a0e2ea15d0 100644
--- a/configure.ac
+++ b/configure.ac
@@ -1,4 +1,4 @@
-# $Id: configure.ac,v 1.444 2010/03/05 04:04:35 djm Exp $
+# $Id: configure.ac,v 1.449 2010/04/10 12:58:01 dtucker Exp $
 #
 # Copyright (c) 1999-2004 Damien Miller
 #
@@ -15,7 +15,7 @@
 # OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
 
 AC_INIT(OpenSSH, Portable, openssh-unix-dev@mindrot.org)
-AC_REVISION($Revision: 1.444 $)
+AC_REVISION($Revision: 1.449 $)
 AC_CONFIG_SRCDIR([ssh.c])
 
 AC_CONFIG_HEADER(config.h)
@@ -488,6 +488,12 @@ main() { if (NSVersionOfRunTimeLibrary("System") >= (60 << 16))
 *-*-dragonfly*)
 	SSHDLIBS="$SSHDLIBS -lcrypt"
 	;;
+*-*-haiku*) 
+    LIBS="$LIBS -lbsd "
+    AC_CHECK_LIB(network, socket)
+    AC_DEFINE(HAVE_U_INT64_T)
+    MANTYPE=man 
+    ;; 
 *-*-hpux*)
 	# first we define all of the options common to all HP-UX releases
 	CPPFLAGS="$CPPFLAGS -D_HPUX_SOURCE -D_XOPEN_SOURCE -D_XOPEN_SOURCE_EXTENDED=1"
@@ -1248,7 +1254,18 @@ LIBEDIT_MSG="no"
 AC_ARG_WITH(libedit,
 	[  --with-libedit[[=PATH]]   Enable libedit support for sftp],
 	[ if test "x$withval" != "xno" ; then
-		if test "x$withval" != "xyes"; then
+		if test "x$withval" = "xyes" ; then
+			AC_PATH_PROG(PKGCONFIG, pkg-config, no)
+			if test "x$PKGCONFIG" != "xno"; then
+				AC_MSG_CHECKING(if $PKGCONFIG knows about libedit)
+			 	if "$PKGCONFIG" libedit; then
+					AC_MSG_RESULT(yes)
+					use_pkgconfig_for_libedit=yes
+				else
+					AC_MSG_RESULT(no)
+				fi
+			fi
+		else
 			CPPFLAGS="$CPPFLAGS -I${withval}/include"
 			if test -n "${need_dash_r}"; then
 				LDFLAGS="-L${withval}/lib -R${withval}/lib ${LDFLAGS}"
@@ -1256,14 +1273,20 @@ AC_ARG_WITH(libedit,
 				LDFLAGS="-L${withval}/lib ${LDFLAGS}"
 			fi
 		fi
+		if test "x$use_pkgconfig_for_libedit" == "xyes"; then
+			LIBEDIT=`$PKGCONFIG --libs-only-l libedit`
+			CPPFLAGS="$CPPFLAGS `$PKGCONFIG --cflags libedit`"
+		else
+			LIBEDIT="-ledit -lcurses"
+		fi
+		OTHERLIBS=`echo $LIBEDIT | sed 's/-ledit//'`
 		AC_CHECK_LIB(edit, el_init,
 			[ AC_DEFINE(USE_LIBEDIT, 1, [Use libedit for sftp])
-			  LIBEDIT="-ledit -lcurses"
 			  LIBEDIT_MSG="yes"
 			  AC_SUBST(LIBEDIT)
 			],
 			[ AC_MSG_ERROR(libedit not found) ],
-			[ -lcurses ]
+			[ $OTHERLIBS ]
 		)
 		AC_MSG_CHECKING(if libedit version is compatible)
 		AC_COMPILE_IFELSE(
@@ -1534,8 +1557,8 @@ dnl    Checks for utmp functions
 AC_CHECK_FUNCS(endutent getutent getutid getutline pututline setutent)
 AC_CHECK_FUNCS(utmpname)
 dnl    Checks for utmpx functions
-AC_CHECK_FUNCS(endutxent getutxent getutxid getutxline pututxline )
-AC_CHECK_FUNCS(setutxent utmpxname)
+AC_CHECK_FUNCS(endutxent getutxent getutxid getutxline getutxuser pututxline)
+AC_CHECK_FUNCS(setutxdb setutxent utmpxname)
 dnl    Checks for lastlog functions
 AC_CHECK_FUNCS(getlastlogxbyname)
 
@@ -4064,34 +4087,6 @@ if test -n "$conf_wtmp_location"; then
 fi
 
 
-dnl utmpx detection - I don't know any system so perverse as to require
-dnl  utmpx, but not define UTMPX_FILE (ditto wtmpx.) No doubt it's out
-dnl  there, though.
-AC_MSG_CHECKING([if your system defines UTMPX_FILE])
-AC_TRY_COMPILE([
-#include <sys/types.h>
-#include <utmp.h>
-#ifdef HAVE_UTMPX_H
-#include <utmpx.h>
-#endif
-#ifdef HAVE_PATHS_H
-#  include <paths.h>
-#endif
-	],
-	[ char *utmpx = UTMPX_FILE; ],
-	[ AC_MSG_RESULT(yes) ],
-	[ AC_MSG_RESULT(no)
-	  system_utmpx_path=no ]
-)
-if test -z "$conf_utmpx_location"; then
-	if test x"$system_utmpx_path" = x"no" ; then
-		AC_DEFINE(DISABLE_UTMPX)
-	fi
-else
-	AC_DEFINE_UNQUOTED(CONF_UTMPX_FILE, "$conf_utmpx_location",
-		[Define if you want to specify the path to your utmpx file])
-fi
-
 dnl wtmpx detection
 AC_MSG_CHECKING([if your system defines WTMPX_FILE])
 AC_TRY_COMPILE([
@@ -4128,12 +4123,13 @@ dnl Adding -Werror to CFLAGS early prevents configure tests from running.
 dnl Add now.
 CFLAGS="$CFLAGS $werror_flags"
 
-if grep "#define BROKEN_GETADDRINFO 1" confdefs.h >/dev/null || \
-    test "x$ac_cv_func_getaddrinfo" != "xyes" ; then
-	AC_SUBST(TEST_SSH_IPV6, no)
+if test "x$ac_cv_func_getaddrinfo" != "xyes" ; then
+	TEST_SSH_IPV6=no
 else
-	AC_SUBST(TEST_SSH_IPV6, yes)
+	TEST_SSH_IPV6=yes
 fi
+AC_CHECK_DECL(BROKEN_GETADDRINFO,  TEST_SSH_IPV6=no)
+AC_SUBST(TEST_SSH_IPV6, $TEST_SSH_IPV6)
 
 AC_EXEEXT
 AC_CONFIG_FILES([Makefile buildpkg.sh opensshd.init openssh.xml \
diff --git a/contrib/Makefile b/contrib/Makefile
index 2cef46f6c6d..8b34eb221ce 100644
--- a/contrib/Makefile
+++ b/contrib/Makefile
@@ -9,7 +9,7 @@ gnome-ssh-askpass1: gnome-ssh-askpass1.c
 gnome-ssh-askpass2: gnome-ssh-askpass2.c
 	$(CC) `pkg-config --cflags gtk+-2.0` \
 		gnome-ssh-askpass2.c -o gnome-ssh-askpass2 \
-		`pkg-config --libs gtk+-2.0`
+		`pkg-config --libs gtk+-2.0 x11`
 
 clean:
 	rm -f *.o gnome-ssh-askpass1 gnome-ssh-askpass2 gnome-ssh-askpass
diff --git a/contrib/caldera/openssh.spec b/contrib/caldera/openssh.spec
index 7c291a0ebe1..6bea9a40fdf 100644
--- a/contrib/caldera/openssh.spec
+++ b/contrib/caldera/openssh.spec
@@ -17,11 +17,11 @@
 #old cvs stuff.  please update before use.  may be deprecated.
 %define use_stable	1
 %if %{use_stable}
-  %define version 	5.4p1
+  %define version 	5.5p1
   %define cvs		%{nil}
   %define release 	1
 %else
-  %define version 	5.4p1
+  %define version 	5.5p1
   %define cvs		cvs20050315
   %define release 	0r1
 %endif
@@ -360,4 +360,4 @@ fi
 * Mon Jan 01 1998 ...
 Template Version: 1.31
 
-$Id: openssh.spec,v 1.69 2010/03/07 22:41:03 djm Exp $
+$Id: openssh.spec,v 1.70 2010/03/21 19:11:58 djm Exp $
diff --git a/contrib/cygwin/Makefile b/contrib/cygwin/Makefile
index 9f680c16cc2..dc857f2edb6 100644
--- a/contrib/cygwin/Makefile
+++ b/contrib/cygwin/Makefile
@@ -42,11 +42,13 @@ install-sshdoc:
 	$(INSTALL) -m 644 $(srcdir)/OVERVIEW $(DESTDIR)$(sshdocdir)/OVERVIEW
 	$(INSTALL) -m 644 $(srcdir)/PROTOCOL $(DESTDIR)$(sshdocdir)/PROTOCOL
 	$(INSTALL) -m 644 $(srcdir)/PROTOCOL.agent $(DESTDIR)$(sshdocdir)/PROTOCOL.agent
+	$(INSTALL) -m 644 $(srcdir)/PROTOCOL.certkeys $(DESTDIR)$(sshdocdir)/PROTOCOL.certkeys
+	$(INSTALL) -m 644 $(srcdir)/PROTOCOL.mux $(DESTDIR)$(sshdocdir)/PROTOCOL.mux
 	$(INSTALL) -m 644 $(srcdir)/README $(DESTDIR)$(sshdocdir)/README
 	$(INSTALL) -m 644 $(srcdir)/README.dns $(DESTDIR)$(sshdocdir)/README.dns
 	$(INSTALL) -m 644 $(srcdir)/README.platform $(DESTDIR)$(sshdocdir)/README.platform
 	$(INSTALL) -m 644 $(srcdir)/README.privsep $(DESTDIR)$(sshdocdir)/README.privsep
-	$(INSTALL) -m 644 $(srcdir)/README.smartcard $(DESTDIR)$(sshdocdir)/README.smartcard
+	$(INSTALL) -m 644 $(srcdir)/README.tun $(DESTDIR)$(sshdocdir)/README.tun
 	$(INSTALL) -m 644 $(srcdir)/TODO $(DESTDIR)$(sshdocdir)/TODO
 	$(INSTALL) -m 644 $(srcdir)/WARNING.RNG $(DESTDIR)$(sshdocdir)/WARNING.RNG
 
diff --git a/contrib/cygwin/ssh-host-config b/contrib/cygwin/ssh-host-config
index b6f9511e2d0..d968d4619d5 100644
--- a/contrib/cygwin/ssh-host-config
+++ b/contrib/cygwin/ssh-host-config
@@ -90,7 +90,7 @@ update_services_file() {
   fi
   _serv_tmp="${_my_etcdir}/srv.out.$$"
 
-  mount -o text -f "${_win_etcdir}" "${_my_etcdir}"
+  mount -o text,posix=0,noacl -f "${_win_etcdir}" "${_my_etcdir}"
 
   # Depends on the above mount
   _wservices=`cygpath -w "${_services}"`
diff --git a/contrib/redhat/openssh.spec b/contrib/redhat/openssh.spec
index de24f1c5d7e..c13cfe60d32 100644
--- a/contrib/redhat/openssh.spec
+++ b/contrib/redhat/openssh.spec
@@ -1,4 +1,4 @@
-%define ver 5.4p1
+%define ver 5.5p1
 %define rel 1
 
 # OpenSSH privilege separation requires a user & group ID
diff --git a/contrib/ssh-copy-id b/contrib/ssh-copy-id
index df74d25c8c0..65c0a8cd863 100644
--- a/contrib/ssh-copy-id
+++ b/contrib/ssh-copy-id
@@ -19,7 +19,7 @@ if [ "-i" = "$1" ]; then
     shift         # and this should leave $1 as the target name
   fi
 else
-  if [ x$SSH_AUTH_SOCK != x ] ; then
+  if [ x$SSH_AUTH_SOCK != x ] && ssh-add -L >/dev/null 2>&1; then
     GET_ID="$GET_ID ssh-add -L"
   fi
 fi
diff --git a/contrib/suse/openssh.spec b/contrib/suse/openssh.spec
index c09246d1ff0..52ed915dc82 100644
--- a/contrib/suse/openssh.spec
+++ b/contrib/suse/openssh.spec
@@ -13,7 +13,7 @@
 
 Summary:	OpenSSH, a free Secure Shell (SSH) protocol implementation
 Name:		openssh
-Version:	5.3p1
+Version:	5.5p1
 URL:		http://www.openssh.com/
 Release:	1
 Source0:	openssh-%{version}.tar.gz
diff --git a/defines.h b/defines.h
index c9b93bf7139..fe25170e618 100644
--- a/defines.h
+++ b/defines.h
@@ -25,7 +25,7 @@
 #ifndef _DEFINES_H
 #define _DEFINES_H
 
-/* $Id: defines.h,v 1.159 2010/01/13 23:44:34 tim Exp $ */
+/* $Id: defines.h,v 1.160 2010/04/09 08:13:27 dtucker Exp $ */
 
 
 /* Constants */
@@ -674,7 +674,7 @@ struct winsize {
 #else
 /* Simply select your favourite login types. */
 /* Can't do if-else because some systems use several... <sigh> */
-#  if defined(UTMPX_FILE) && !defined(DISABLE_UTMPX)
+#  if !defined(DISABLE_UTMPX)
 #    define USE_UTMPX
 #  endif
 #  if defined(UTMP_FILE) && !defined(DISABLE_UTMP)
diff --git a/key.c b/key.c
index 0d0c912e64f..66592c7ed19 100644
--- a/key.c
+++ b/key.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: key.c,v 1.85 2010/03/04 01:44:57 djm Exp $ */
+/* $OpenBSD: key.c,v 1.86 2010/03/15 19:40:02 stevesk Exp $ */
 /*
  * read_bignum():
  * Copyright (c) 1995 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland
@@ -801,6 +801,19 @@ key_type(const Key *k)
 	return "unknown";
 }
 
+const char *
+key_cert_type(const Key *k)
+{
+	switch (k->cert->type) {
+	case SSH2_CERT_TYPE_USER:
+		return "user";
+	case SSH2_CERT_TYPE_HOST:
+		return "host";
+	default:
+		return "unknown";
+	}
+}
+
 const char *
 key_ssh_name(const Key *k)
 {
diff --git a/key.h b/key.h
index 6a2e049af11..4f17777c088 100644
--- a/key.h
+++ b/key.h
@@ -1,4 +1,4 @@
-/* $OpenBSD: key.h,v 1.28 2010/02/26 20:29:54 djm Exp $ */
+/* $OpenBSD: key.h,v 1.29 2010/03/15 19:40:02 stevesk Exp $ */
 
 /*
  * Copyright (c) 2000, 2001 Markus Friedl.  All rights reserved.
@@ -82,6 +82,7 @@ int		 key_equal(const Key *, const Key *);
 char		*key_fingerprint(Key *, enum fp_type, enum fp_rep);
 u_char		*key_fingerprint_raw(Key *, enum fp_type, u_int *);
 const char	*key_type(const Key *);
+const char	*key_cert_type(const Key *);
 int		 key_write(const Key *, FILE *);
 int		 key_read(Key *, char **);
 u_int		 key_size(const Key *);
diff --git a/loginrec.c b/loginrec.c
index bca95970711..6f655cb16cd 100644
--- a/loginrec.c
+++ b/loginrec.c
@@ -207,6 +207,7 @@ int syslogin_write_entry(struct logininfo *li);
 
 int getlast_entry(struct logininfo *li);
 int lastlog_get_entry(struct logininfo *li);
+int utmpx_get_entry(struct logininfo *li);
 int wtmp_get_entry(struct logininfo *li);
 int wtmpx_get_entry(struct logininfo *li);
 
@@ -508,6 +509,10 @@ getlast_entry(struct logininfo *li)
 #ifdef USE_LASTLOG
 	return(lastlog_get_entry(li));
 #else /* !USE_LASTLOG */
+#if defined(USE_UTMPX) && defined(HAVE_SETUTXDB) && \
+    defined(UTXDB_LASTLOGIN) && defined(HAVE_GETUTXUSER)
+	return (utmpx_get_entry(li));
+#endif
 
 #if defined(DISABLE_LASTLOG)
 	/* On some systems we shouldn't even try to obtain last login
@@ -1608,6 +1613,32 @@ lastlog_get_entry(struct logininfo *li)
 #endif /* HAVE_GETLASTLOGXBYNAME */
 #endif /* USE_LASTLOG */
 
+#if defined(USE_UTMPX) && defined(HAVE_SETUTXDB) && \
+    defined(UTXDB_LASTLOGIN) && defined(HAVE_GETUTXUSER)
+int
+utmpx_get_entry(struct logininfo *li)
+{
+	struct utmpx *utx;
+
+	if (setutxdb(UTXDB_LASTLOGIN, NULL) != 0)
+		return (0);
+	utx = getutxuser(li->username);
+	if (utx == NULL) {
+		endutxent();
+		return (0);
+	}
+
+	line_fullname(li->line, utx->ut_line,
+	    MIN_SIZEOF(li->line, utx->ut_line));
+	strlcpy(li->hostname, utx->ut_host,
+	    MIN_SIZEOF(li->hostname, utx->ut_host));
+	li->tv_sec = utx->ut_tv.tv_sec;
+	li->tv_usec = utx->ut_tv.tv_usec;
+	endutxent();
+	return (1);
+}
+#endif /* USE_UTMPX && HAVE_SETUTXDB && UTXDB_LASTLOGIN && HAVE_GETUTXUSER */
+
 #ifdef USE_BTMP
   /*
    * Logs failed login attempts in _PATH_BTMP if that exists.
diff --git a/logintest.c b/logintest.c
index 7e9fbbfbbdc..4897ae0f9e1 100644
--- a/logintest.c
+++ b/logintest.c
@@ -264,7 +264,7 @@ showOptions(void)
 	printf("\tUSE_UTMP (UTMP_FILE=%s)\n", UTMP_FILE);
 #endif
 #ifdef USE_UTMPX
-	printf("\tUSE_UTMPX (UTMPX_FILE=%s)\n", UTMPX_FILE);
+	printf("\tUSE_UTMPX\n");
 #endif
 #ifdef USE_WTMP
 	printf("\tUSE_WTMP (WTMP_FILE=%s)\n", WTMP_FILE);
diff --git a/moduli.0 b/moduli.0
index 63e6f9eb1a0..e3fea22a9a7 100644
--- a/moduli.0
+++ b/moduli.0
@@ -69,4 +69,4 @@ SEE ALSO
      Diffie-Hellman Group Exchange for the Secure Shell (SSH) Transport Layer
      Protocol, RFC 4419, 2006.
 
-OpenBSD 4.6                      June 26, 2008                               2
+OpenBSD 4.7                      June 26, 2008                               2
diff --git a/openbsd-compat/bsd-arc4random.c b/openbsd-compat/bsd-arc4random.c
index 9d4c8690eb1..d7c5862534f 100644
--- a/openbsd-compat/bsd-arc4random.c
+++ b/openbsd-compat/bsd-arc4random.c
@@ -84,7 +84,7 @@ arc4random_stir(void)
 }
 #endif /* !HAVE_ARC4RANDOM */
 
-#ifndef ARC4RANDOM_BUF
+#ifndef HAVE_ARC4RANDOM_BUF
 void
 arc4random_buf(void *_buf, size_t n)
 {
@@ -102,7 +102,7 @@ arc4random_buf(void *_buf, size_t n)
 }
 #endif /* !HAVE_ARC4RANDOM_BUF */
 
-#ifndef ARC4RANDOM_UNIFORM
+#ifndef HAVE_ARC4RANDOM_UNIFORM
 /*
  * Calculate a uniformly distributed random number less than upper_bound
  * avoiding "modulo bias".
diff --git a/scp.0 b/scp.0
index 74dcf81da32..fc9f75594ab 100644
--- a/scp.0
+++ b/scp.0
@@ -145,4 +145,4 @@ AUTHORS
      Timo Rinne <tri@iki.fi>
      Tatu Ylonen <ylo@cs.hut.fi>
 
-OpenBSD 4.6                    February 8, 2010                              3
+OpenBSD 4.7                    February 8, 2010                              3
diff --git a/servconf.c b/servconf.c
index f9e2f2dfd4c..7d027ddb990 100644
--- a/servconf.c
+++ b/servconf.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: servconf.c,v 1.204 2010/03/04 10:36:03 djm Exp $ */
+/* $OpenBSD: servconf.c,v 1.207 2010/03/25 23:38:28 djm Exp $ */
 /*
  * Copyright (c) 1995 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland
  *                    All rights reserved
@@ -470,15 +470,14 @@ parse_token(const char *cp, const char *filename,
 char *
 derelativise_path(const char *path)
 {
-	char *expanded, *ret, *cwd;
+	char *expanded, *ret, cwd[MAXPATHLEN];
 
 	expanded = tilde_expand_filename(path, getuid());
 	if (*expanded == '/')
 		return expanded;
-	if ((cwd = getcwd(NULL, 0)) == NULL)
+	if (getcwd(cwd, sizeof(cwd)) == NULL)
 		fatal("%s: getcwd: %s", __func__, strerror(errno));
 	xasprintf(&ret, "%s/%s", cwd, expanded);
-	xfree(cwd);
 	xfree(expanded);
 	return ret;
 }
@@ -1223,7 +1222,17 @@ process_server_config_line(ServerOptions *options, char *line,
 		charptr = (opcode == sAuthorizedKeysFile) ?
 		    &options->authorized_keys_file :
 		    &options->authorized_keys_file2;
-		goto parse_filename;
+		arg = strdelim(&cp);
+		if (!arg || *arg == '\0')
+			fatal("%s line %d: missing file name.",
+			    filename, linenum);
+		if (*activep && *charptr == NULL) {
+			*charptr = tilde_expand_filename(arg, getuid());
+			/* increase optional counter */
+			if (intptr != NULL)
+				*intptr = *intptr + 1;
+		}
+		break;
 
 	case sClientAliveInterval:
 		intptr = &options->client_alive_interval;
diff --git a/session.c b/session.c
index 639405fec30..e032de6926f 100644
--- a/session.c
+++ b/session.c
@@ -1551,6 +1551,10 @@ do_setusercontext(struct passwd *pw)
 		}
 #endif /* HAVE_SETPCRED */
 
+#ifdef WITH_SELINUX
+		ssh_selinux_setup_exec_context(pw->pw_name);
+#endif
+
 		if (options.chroot_directory != NULL &&
 		    strcasecmp(options.chroot_directory, "none") != 0) {
                         tmp = tilde_expand_filename(options.chroot_directory,
@@ -1575,10 +1579,6 @@ do_setusercontext(struct passwd *pw)
 
 	if (getuid() != pw->pw_uid || geteuid() != pw->pw_uid)
 		fatal("Failed to set uids to %u.", (u_int) pw->pw_uid);
-
-#ifdef WITH_SELINUX
-	ssh_selinux_setup_exec_context(pw->pw_name);
-#endif
 }
 
 static void
diff --git a/sftp-server.0 b/sftp-server.0
index 6628dcfca17..0c2654c8d2a 100644
--- a/sftp-server.0
+++ b/sftp-server.0
@@ -60,4 +60,4 @@ HISTORY
 AUTHORS
      Markus Friedl <markus@openbsd.org>
 
-OpenBSD 4.6                     January 9, 2010                              1
+OpenBSD 4.7                     January 9, 2010                              1
diff --git a/sftp.0 b/sftp.0
index cb1bb033ea4..488548a6ed7 100644
--- a/sftp.0
+++ b/sftp.0
@@ -316,4 +316,4 @@ SEE ALSO
      T. Ylonen and S. Lehtinen, SSH File Transfer Protocol, draft-ietf-secsh-
      filexfer-00.txt, January 2001, work in progress material.
 
-OpenBSD 4.6                    February 8, 2010                              5
+OpenBSD 4.7                    February 8, 2010                              5
diff --git a/ssh-add.0 b/ssh-add.0
index 8fbd1e1009f..0d49c5ea34e 100644
--- a/ssh-add.0
+++ b/ssh-add.0
@@ -106,4 +106,4 @@ AUTHORS
      ated OpenSSH.  Markus Friedl contributed the support for SSH protocol
      versions 1.5 and 2.0.
 
-OpenBSD 4.6                      March 5, 2010                               2
+OpenBSD 4.7                      March 5, 2010                               2
diff --git a/ssh-agent.0 b/ssh-agent.0
index c178dcd0177..536eac756ac 100644
--- a/ssh-agent.0
+++ b/ssh-agent.0
@@ -115,4 +115,4 @@ AUTHORS
      ated OpenSSH.  Markus Friedl contributed the support for SSH protocol
      versions 1.5 and 2.0.
 
-OpenBSD 4.6                    January 17, 2010                              2
+OpenBSD 4.7                    January 17, 2010                              2
diff --git a/ssh-keygen.0 b/ssh-keygen.0
index c9877300e7f..aed4a14ad91 100644
--- a/ssh-keygen.0
+++ b/ssh-keygen.0
@@ -165,8 +165,14 @@ DESCRIPTION
              section for details.  The constraints that are valid for user
              certificates are:
 
-             no-x11-forwarding
-                     Disable X11 forwarding (permitted by default).
+             clear   Clear all enabled permissions.  This is useful for clear-
+                     ing the default set of permissions so permissions may be
+                     added individually.
+
+             force-command=command
+                     Forces the execution of command instead of any shell or
+                     command specified by the user when the certificate is
+                     used for authentication.
 
              no-agent-forwarding
                      Disable ssh-agent(1) forwarding (permitted by default).
@@ -180,12 +186,8 @@ DESCRIPTION
                      Disable execution of ~/.ssh/rc by sshd(8) (permitted by
                      default).
 
-             clear   Clear all enabled permissions.  This is useful for clear-
-                     ing the default set of permissions so permissions may be
-                     added individually.
-
-             permit-x11-forwarding
-                     Allows X11 forwarding.
+             no-x11-forwarding
+                     Disable X11 forwarding (permitted by default).
 
              permit-agent-forwarding
                      Allows ssh-agent(1) forwarding.
@@ -199,16 +201,14 @@ DESCRIPTION
              permit-user-rc
                      Allows execution of ~/.ssh/rc by sshd(8).
 
-             force-command=command
-                     Forces the execution of command instead of any shell or
-                     command specified by the user when the certificate is
-                     used for authentication.
+             permit-x11-forwarding
+                     Allows X11 forwarding.
 
              source-address=address_list
                      Restrict the source addresses from which the certificate
-                     is considered valid from.  The address_list is a comma-
-                     separated list of one or more address/netmask pairs in
-                     CIDR format.
+                     is considered valid.  The address_list is a comma-sepa-
+                     rated list of one or more address/netmask pairs in CIDR
+                     format.
 
              At present, no constraints are valid for host keys.
 
@@ -257,9 +257,9 @@ DESCRIPTION
              in YYYYMMDD format, a time in YYYYMMDDHHMMSS format or a relative
              time (to the current time) consisting of a minus sign followed by
              a relative time in the format described in the TIME FORMATS sec-
-             tion of ssh_config(5).  The end time may be specified as a YYYYM-
-             MDD date, a YYYYMMDDHHMMSS time or a relative time starting with
-             a plus character.
+             tion of sshd_config(5).  The end time may be specified as a
+             YYYYMMDD date, a YYYYMMDDHHMMSS time or a relative time starting
+             with a plus character.
 
              For example: ``+52w1d'' (valid from now to 52 weeks and one day
              from now), ``-4w:+4w'' (valid from four weeks ago to four weeks
@@ -329,12 +329,12 @@ CERTIFICATES
 
            $ ssh-keygen -s /path/to/ca_key -I key_id /path/to/user_key.pub
 
-     The resultant certificate will be placed in /path/to/user_key_cert.pub.
+     The resultant certificate will be placed in /path/to/user_key-cert.pub.
      A host certificate requires the -h option:
 
            $ ssh-keygen -s /path/to/ca_key -I key_id -h /path/to/host_key.pub
 
-     The host certificate will be output to /path/to/host_key_cert.pub.  In
+     The host certificate will be output to /path/to/host_key-cert.pub.  In
      both cases, key_id is a "key identifier" that is logged by the server
      when the certificate is used for authentication.
 
@@ -344,7 +344,7 @@ CERTIFICATES
      pals:
 
            $ ssh-keygen -s ca_key -I key_id -n user1,user2 user_key.pub
-           $ ssh-keygen -s ca_key -I key_id -h -n host.domain $0
+           $ ssh-keygen -s ca_key -I key_id -h -n host.domain user_key.pub
 
      Additional limitations on the validity and use of user certificates may
      be specified through certificate constraints.  A constrained certificate
@@ -431,4 +431,4 @@ AUTHORS
      created OpenSSH.  Markus Friedl contributed the support for SSH protocol
      versions 1.5 and 2.0.
 
-OpenBSD 4.6                      March 8, 2010                               7
+OpenBSD 4.7                     March 13, 2010                               7
diff --git a/ssh-keygen.1 b/ssh-keygen.1
index 6557f9336c0..3e03a9bd0cd 100644
--- a/ssh-keygen.1
+++ b/ssh-keygen.1
@@ -1,4 +1,4 @@
-.\"	$OpenBSD: ssh-keygen.1,v 1.88 2010/03/08 00:28:55 djm Exp $
+.\"	$OpenBSD: ssh-keygen.1,v 1.92 2010/03/13 23:38:13 jmc Exp $
 .\"
 .\"  -*- nroff -*-
 .\"
@@ -37,7 +37,7 @@
 .\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 .\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .\"
-.Dd $Mdocdate: March 8 2010 $
+.Dd $Mdocdate: March 13 2010 $
 .Dt SSH-KEYGEN 1
 .Os
 .Sh NAME
@@ -307,8 +307,15 @@ Please see the
 section for details.
 The constraints that are valid for user certificates are:
 .Bl -tag -width Ds
-.It Ic no-x11-forwarding
-Disable X11 forwarding (permitted by default).
+.It Ic clear
+Clear all enabled permissions.
+This is useful for clearing the default set of permissions so permissions may
+be added individually.
+.It Ic force-command Ns = Ns Ar command
+Forces the execution of
+.Ar command
+instead of any shell or command specified by the user when
+the certificate is used for authentication.
 .It Ic no-agent-forwarding
 Disable
 .Xr ssh-agent 1
@@ -323,12 +330,8 @@ Disable execution of
 by
 .Xr sshd 8
 (permitted by default).
-.It Ic clear
-Clear all enabled permissions.
-This is useful for clearing the default set of permissions so permissions may
-be added individually.
-.It Ic permit-x11-forwarding
-Allows X11 forwarding.
+.It Ic no-x11-forwarding
+Disable X11 forwarding (permitted by default).
 .It Ic permit-agent-forwarding
 Allows
 .Xr ssh-agent 1
@@ -342,14 +345,10 @@ Allows execution of
 .Pa ~/.ssh/rc
 by
 .Xr sshd 8 .
-.It Ic force-command=command
-Forces the execution of
-.Ar command
-instead of any shell or command specified by the user when
-the certificate is used for authentication.
-.It Ic source-address=address_list
-Restrict the source addresses from which the certificate is considered valid
-from.
+.It Ic permit-x11-forwarding
+Allows X11 forwarding.
+.It Ic source-address Ns = Ns Ar address_list
+Restrict the source addresses from which the certificate is considered valid.
 The
 .Ar address_list
 is a comma-separated list of one or more address/netmask pairs in CIDR
@@ -414,7 +413,7 @@ in YYYYMMDDHHMMSS format or a relative time (to the current time) consisting
 of a minus sign followed by a relative time in the format described in the
 .Sx TIME FORMATS
 section of
-.Xr ssh_config 5 .
+.Xr sshd_config 5 .
 The end time may be specified as a YYYYMMDD date, a YYYYMMDDHHMMSS time or
 a relative time starting with a plus character.
 .Pp
@@ -519,7 +518,7 @@ To generate a user certificate:
 .Dl $ ssh-keygen -s /path/to/ca_key -I key_id /path/to/user_key.pub
 .Pp
 The resultant certificate will be placed in
-.Pa /path/to/user_key_cert.pub .
+.Pa /path/to/user_key-cert.pub .
 A host certificate requires the
 .Fl h
 option:
@@ -527,7 +526,7 @@ option:
 .Dl $ ssh-keygen -s /path/to/ca_key -I key_id -h /path/to/host_key.pub
 .Pp
 The host certificate will be output to
-.Pa /path/to/host_key_cert.pub .
+.Pa /path/to/host_key-cert.pub .
 In both cases,
 .Ar key_id
 is a "key identifier" that is logged by the server when the certificate
@@ -539,7 +538,7 @@ By default, generated certificates are valid for all users or hosts.
 To generate a certificate for a specified set of principals:
 .Pp
 .Dl $ ssh-keygen -s ca_key -I key_id -n user1,user2 user_key.pub
-.Dl $ ssh-keygen -s ca_key -I key_id -h -n host.domain user_key.pub
+.Dl "$ ssh-keygen -s ca_key -I key_id -h -n host.domain user_key.pub"
 .Pp
 Additional limitations on the validity and use of user certificates may
 be specified through certificate constraints.
diff --git a/ssh-keygen.c b/ssh-keygen.c
index dd662c90760..37e516ff2f1 100644
--- a/ssh-keygen.c
+++ b/ssh-keygen.c
@@ -1,4 +1,4 @@
-/* $OpenBSD: ssh-keygen.c,v 1.184 2010/03/07 22:16:01 djm Exp $ */
+/* $OpenBSD: ssh-keygen.c,v 1.185 2010/03/15 19:40:02 stevesk Exp $ */
 /*
  * Author: Tatu Ylonen <ylo@cs.hut.fi>
  * Copyright (c) 1994 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland
@@ -1393,7 +1393,8 @@ do_show_cert(struct passwd *pw)
 	    SSH_FP_MD5, SSH_FP_HEX);
 
 	printf("%s:\n", identity_file);
-	printf("        %s certificate %s\n", key_type(key), key_fp);
+	printf("        %s %s certificate %s\n", key_type(key),
+	    key_cert_type(key), key_fp);
 	printf("        Signed by %s CA %s\n",
 	    key_type(key->cert->signature_key), ca_fp);
 	printf("        Key ID \"%s\"\n", key->cert->key_id);
diff --git a/ssh-keyscan.0 b/ssh-keyscan.0
index 1100a69b64b..8a0ef60e485 100644
--- a/ssh-keyscan.0
+++ b/ssh-keyscan.0
@@ -104,4 +104,4 @@ BUGS
      This is because it opens a connection to the ssh port, reads the public
      key, and drops the connection as soon as it gets the key.
 
-OpenBSD 4.6                     January 9, 2010                              2
+OpenBSD 4.7                     January 9, 2010                              2
diff --git a/ssh-keysign.0 b/ssh-keysign.0
index ab15e88bc7e..20a8eaa8794 100644
--- a/ssh-keysign.0
+++ b/ssh-keysign.0
@@ -39,4 +39,4 @@ HISTORY
 AUTHORS
      Markus Friedl <markus@openbsd.org>
 
-OpenBSD 4.6                      May 31, 2007                                1
+OpenBSD 4.7                      May 31, 2007                                1
diff --git a/ssh-pkcs11-helper.0 b/ssh-pkcs11-helper.0
index 2760cad94d4..9eb2bc96a33 100644
--- a/ssh-pkcs11-helper.0
+++ b/ssh-pkcs11-helper.0
@@ -22,4 +22,4 @@ HISTORY
 AUTHORS
      Markus Friedl <markus@openbsd.org>
 
-OpenBSD 4.6                    February 10, 2010                             1
+OpenBSD 4.7                    February 10, 2010                             1
diff --git a/ssh-pkcs11-helper.c b/ssh-pkcs11-helper.c
index d3bfb98384e..8e3f57acef9 100644
--- a/ssh-pkcs11-helper.c
+++ b/ssh-pkcs11-helper.c
@@ -17,8 +17,6 @@
 
 #include "includes.h"
 
-#ifdef ENABLE_PKCS11
-
 #include <sys/types.h>
 #ifdef HAVE_SYS_TIME_H
 # include <sys/time.h>
@@ -39,6 +37,8 @@
 #include "authfd.h"
 #include "ssh-pkcs11.h"
 
+#ifdef ENABLE_PKCS11
+
 /* borrows code from sftp-server and ssh-agent */
 
 struct pkcs11_keyinfo {
diff --git a/ssh-rand-helper.0 b/ssh-rand-helper.0
index 72bfcdf6404..51b6f557158 100644
--- a/ssh-rand-helper.0
+++ b/ssh-rand-helper.0
@@ -48,4 +48,4 @@ AUTHORS
 SEE ALSO
      ssh(1), ssh-add(1), ssh-keygen(1), sshd(8)
 
-OpenBSD 4.6                     April 14, 2002                               1
+OpenBSD 4.7                     April 14, 2002                               1
diff --git a/ssh.0 b/ssh.0
index deab90e7383..9c793b8438d 100644
--- a/ssh.0
+++ b/ssh.0
@@ -308,9 +308,10 @@ DESCRIPTION
              allocated on the server and reported to the client at run time.
 
      -S ctl_path
-             Specifies the location of a control socket for connection shar-
-             ing.  Refer to the description of ControlPath and ControlMaster
-             in ssh_config(5) for details.
+             Specifies the location of a control socket for connection sharing
+             or the string ``none'' to disable connection sharing.  Refer to
+             the description of ControlPath and ControlMaster in ssh_config(5)
+             for details.
 
      -s      May be used to request invocation of a subsystem on the remote
              system.  Subsystems are a feature of the SSH2 protocol which fa-
@@ -876,4 +877,4 @@ AUTHORS
      created OpenSSH.  Markus Friedl contributed the support for SSH protocol
      versions 1.5 and 2.0.
 
-OpenBSD 4.6                      March 5, 2010                              14
+OpenBSD 4.7                     March 26, 2010                              14
diff --git a/ssh.1 b/ssh.1
index 3f815b8e716..c03771a3eba 100644
--- a/ssh.1
+++ b/ssh.1
@@ -34,8 +34,8 @@
 .\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 .\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .\"
-.\" $OpenBSD: ssh.1,v 1.302 2010/03/05 10:28:21 djm Exp $
-.Dd $Mdocdate: March 5 2010 $
+.\" $OpenBSD: ssh.1,v 1.303 2010/03/26 00:26:58 djm Exp $
+.Dd $Mdocdate: March 26 2010 $
 .Dt SSH 1
 .Os
 .Sh NAME
@@ -558,7 +558,10 @@ argument is
 the listen port will be dynamically allocated on the server and reported
 to the client at run time.
 .It Fl S Ar ctl_path
-Specifies the location of a control socket for connection sharing.
+Specifies the location of a control socket for connection sharing
+or the string
+.Dq none
+to disable connection sharing.
 Refer to the description of
 .Cm ControlPath
 and
diff --git a/ssh_config.0 b/ssh_config.0
index 48969a22a48..1a2c64ce152 100644
--- a/ssh_config.0
+++ b/ssh_config.0
@@ -425,8 +425,8 @@ DESCRIPTION
              Specifies the order in which the client should try protocol 2 au-
              thentication methods.  This allows a client to prefer one method
              (e.g. keyboard-interactive) over another method (e.g. password)
-             The default for this option is: ``gssapi-with-mic,hostbased,
-             publickey, keyboard-interactive, password''.
+             The default for this option is: ``gssapi-with-
+             mic,hostbased,publickey,keyboard-interactive,password''.
 
      Protocol
              Specifies the protocol versions ssh(1) should support in order of
@@ -673,4 +673,4 @@ AUTHORS
      ated OpenSSH.  Markus Friedl contributed the support for SSH protocol
      versions 1.5 and 2.0.
 
-OpenBSD 4.6                      March 5, 2010                              11
+OpenBSD 4.7                     March 26, 2010                              11
diff --git a/ssh_config.5 b/ssh_config.5
index 8cf02597d06..41085356040 100644
--- a/ssh_config.5
+++ b/ssh_config.5
@@ -34,8 +34,8 @@
 .\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
 .\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
 .\"
-.\" $OpenBSD: ssh_config.5,v 1.129 2010/03/05 10:28:21 djm Exp $
-.Dd $Mdocdate: March 5 2010 $
+.\" $OpenBSD: ssh_config.5,v 1.130 2010/03/26 01:06:13 dtucker Exp $
+.Dd $Mdocdate: March 26 2010 $
 .Dt SSH_CONFIG 5
 .Os
 .Sh NAME
@@ -734,11 +734,7 @@ This allows a client to prefer one method (e.g.\&
 over another method (e.g.\&
 .Cm password )
 The default for this option is:
-.Do gssapi-with-mic ,
-hostbased,
-publickey,
-keyboard-interactive,
-password
+.Do gssapi-with-mic,hostbased,publickey,keyboard-interactive,password
 .Dc .
 .It Cm Protocol
 Specifies the protocol versions
diff --git a/sshd.0 b/sshd.0
index 34eaafc9243..30bf6dedeea 100644
--- a/sshd.0
+++ b/sshd.0
@@ -614,4 +614,4 @@ CAVEATS
      System security is not improved unless rshd, rlogind, and rexecd are dis-
      abled (thus completely disabling rlogin and rsh into the machine).
 
-OpenBSD 4.6                      March 5, 2010                              10
+OpenBSD 4.7                      March 5, 2010                              10
diff --git a/sshd_config.0 b/sshd_config.0
index a9162f18d2f..94935c07f5f 100644
--- a/sshd_config.0
+++ b/sshd_config.0
@@ -656,4 +656,4 @@ AUTHORS
      versions 1.5 and 2.0.  Niels Provos and Markus Friedl contributed support
      for privilege separation.
 
-OpenBSD 4.6                      March 4, 2010                              10
+OpenBSD 4.7                      March 4, 2010                              10
diff --git a/version.h b/version.h
index c604c753c08..5d2b0b741f0 100644
--- a/version.h
+++ b/version.h
@@ -1,6 +1,6 @@
-/* $OpenBSD: version.h,v 1.57 2010/03/07 22:01:32 djm Exp $ */
+/* $OpenBSD: version.h,v 1.58 2010/03/16 16:36:49 djm Exp $ */
 
-#define SSH_VERSION	"OpenSSH_5.4"
+#define SSH_VERSION	"OpenSSH_5.5"
 
 #define SSH_PORTABLE	"p1"
 #define SSH_RELEASE	SSH_VERSION SSH_PORTABLE

From e7699af0af56af669401bec39f1a19ee314aa584 Mon Sep 17 00:00:00 2001
From: Robert Noland <rnoland@FreeBSD.org>
Date: Wed, 28 Apr 2010 10:33:41 +0000
Subject: [PATCH 294/532] fx->lock is used as an index, correct test case.

MFC after:	1 week
---
 sys/dev/drm/via_video.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/drm/via_video.c b/sys/dev/drm/via_video.c
index f903cf55f10..3bd96051057 100644
--- a/sys/dev/drm/via_video.c
+++ b/sys/dev/drm/via_video.c
@@ -78,7 +78,7 @@ int via_decoder_futex(struct drm_device *dev, void *data, struct drm_file *file_
 
 	DRM_DEBUG("\n");
 
-	if (fx->lock > VIA_NR_XVMC_LOCKS)
+	if (fx->lock >= VIA_NR_XVMC_LOCKS)
 		return -EFAULT;
 
 	lock = (volatile int *)XVMCLOCKPTR(sAPriv, fx->lock);

From f30f9d40066dfc9bddfbaf89872381dd8aa13389 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 10:57:27 +0000
Subject: [PATCH 295/532] Fix typos.

Obtained from:	DragonFlyBSD
---
 sbin/ifconfig/ifconfig.8    | 4 ++--
 sbin/ifconfig/ifieee80211.c | 2 +-
 2 files changed, 3 insertions(+), 3 deletions(-)

diff --git a/sbin/ifconfig/ifconfig.8 b/sbin/ifconfig/ifconfig.8
index eafd5d82ada..e26cbda11f4 100644
--- a/sbin/ifconfig/ifconfig.8
+++ b/sbin/ifconfig/ifconfig.8
@@ -1163,9 +1163,9 @@ In particular the information elements included in management frames
 for old devices are different.
 When compatibility support is enabled both standard and compatible data
 will be provided.
-Stations that associate using the compatiblity mechanisms are flagged
+Stations that associate using the compatibility mechanisms are flagged
 in ``list sta''.
-To disable compatiblity support use
+To disable compatibility support use
 .Fl htcompat .
 .It Cm htprotmode Ar technique
 For interfaces operating in 802.11n, use the specified
diff --git a/sbin/ifconfig/ifieee80211.c b/sbin/ifconfig/ifieee80211.c
index 6cb4d2c7ac6..6ec45ff46f4 100644
--- a/sbin/ifconfig/ifieee80211.c
+++ b/sbin/ifconfig/ifieee80211.c
@@ -5027,7 +5027,7 @@ DECL_CMD_FUNC(set80211clone_wlanaddr, arg, d)
 
 	ea = ether_aton(arg);
 	if (ea == NULL)
-		errx(1, "%s: cannot parse addres", arg);
+		errx(1, "%s: cannot parse address", arg);
 	memcpy(params.icp_macaddr, ea->octet, IEEE80211_ADDR_LEN);
 	params.icp_flags |= IEEE80211_CLONE_MACADDR;
 }

From 2fb1e09dee2a798fa8186cca1c4d5884ed9f9077 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 10:58:03 +0000
Subject: [PATCH 296/532] Bump man page date.

---
 sbin/ifconfig/ifconfig.8 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sbin/ifconfig/ifconfig.8 b/sbin/ifconfig/ifconfig.8
index e26cbda11f4..65c240fe170 100644
--- a/sbin/ifconfig/ifconfig.8
+++ b/sbin/ifconfig/ifconfig.8
@@ -28,7 +28,7 @@
 .\"     From: @(#)ifconfig.8	8.3 (Berkeley) 1/5/94
 .\" $FreeBSD$
 .\"
-.Dd February 20, 2010
+.Dd April 28, 2010
 .Dt IFCONFIG 8
 .Os
 .Sh NAME

From 48e1bda05fdb902f240d8eda19055ac92db2bb95 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 10:58:50 +0000
Subject: [PATCH 297/532] Fix typo. No functional change.

Obtained from:	DragonFlyBSD
---
 sys/net80211/ieee80211_node.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/net80211/ieee80211_node.c b/sys/net80211/ieee80211_node.c
index b17f42fd51a..992ea0f314c 100644
--- a/sys/net80211/ieee80211_node.c
+++ b/sys/net80211/ieee80211_node.c
@@ -432,7 +432,7 @@ ieee80211_reset_bss(struct ieee80211vap *vap)
 	ieee80211_reset_erp(ic);
 
 	ni = ieee80211_alloc_node(&ic->ic_sta, vap, vap->iv_myaddr);
-	KASSERT(ni != NULL, ("unable to setup inital BSS node"));
+	KASSERT(ni != NULL, ("unable to setup initial BSS node"));
 	obss = vap->iv_bss;
 	vap->iv_bss = ieee80211_ref_node(ni);
 	if (obss != NULL) {

From 380fe2df605569b6a4d0a4485c9fc7ab490c6a3e Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 13:25:53 +0000
Subject: [PATCH 298/532] When in the RUN -> AUTH -> RUN FSM transition
 happens, we'll call the ratectl_node_init() functions and since ni_rtctls was
 already malloc'ed() we will panic. Fix this by using the already malloc'ed
 pointer.

Found by:	bschmidt
Reviewed by:	bschmidt
---
 sys/net80211/ieee80211_amrr.c     | 20 ++++++++++----------
 sys/net80211/ieee80211_rssadapt.c | 18 +++++++++++-------
 2 files changed, 21 insertions(+), 17 deletions(-)

diff --git a/sys/net80211/ieee80211_amrr.c b/sys/net80211/ieee80211_amrr.c
index 61110586999..7902545c280 100644
--- a/sys/net80211/ieee80211_amrr.c
+++ b/sys/net80211/ieee80211_amrr.c
@@ -136,16 +136,16 @@ amrr_node_init(struct ieee80211_node *ni)
 	struct ieee80211_amrr *amrr = vap->iv_rs;
 	struct ieee80211_amrr_node *amn;
 
-	KASSERT(ni->ni_rctls == NULL, ("%s: ni_rctls already initialized",
-	    __func__));
-
-	ni->ni_rctls = amn = malloc(sizeof(struct ieee80211_amrr_node),
-	    M_80211_RATECTL, M_NOWAIT|M_ZERO);
-	if (amn == NULL) {
-		if_printf(vap->iv_ifp, "couldn't alloc per-node ratectl "
-		    "structure\n");
-		return;
-	}
+	if (ni->ni_rctls == NULL) {
+		ni->ni_rctls = amn = malloc(sizeof(struct ieee80211_amrr_node),
+		    M_80211_RATECTL, M_NOWAIT|M_ZERO);
+		if (amn == NULL) {
+			if_printf(vap->iv_ifp, "couldn't alloc per-node ratectl "
+			    "structure\n");
+			return;
+		}
+	} else
+		amn = ni->ni_rctls;
 	amn->amn_amrr = amrr;
 	amn->amn_success = 0;
 	amn->amn_recovery = 0;
diff --git a/sys/net80211/ieee80211_rssadapt.c b/sys/net80211/ieee80211_rssadapt.c
index ad329e0470d..f82c01ee930 100644
--- a/sys/net80211/ieee80211_rssadapt.c
+++ b/sys/net80211/ieee80211_rssadapt.c
@@ -169,13 +169,17 @@ rssadapt_node_init(struct ieee80211_node *ni)
 	struct ieee80211_rssadapt *rsa = vap->iv_rs;
 	const struct ieee80211_rateset *rs = &ni->ni_rates;
 
-	ni->ni_rctls = ra = malloc(sizeof(struct ieee80211_rssadapt_node),
-	    M_80211_RATECTL, M_NOWAIT|M_ZERO);
-	if (ra == NULL) {
-		if_printf(vap->iv_ifp, "couldn't alloc per-node ratectl "
-		    "structure\n");
-		return;
-	}
+	if (ni->ni_rctls == NULL) {
+		ni->ni_rctls = ra = 
+		    malloc(sizeof(struct ieee80211_rssadapt_node),
+		        M_80211_RATECTL, M_NOWAIT|M_ZERO);
+		if (ra == NULL) {
+			if_printf(vap->iv_ifp, "couldn't alloc per-node ratectl "
+			    "structure\n");
+			return;
+		}
+	} else
+		ra = ni->ni_rctls;
 	ra->ra_rs = rsa;
 	ra->ra_rates = *rs;
 	rssadapt_updatestats(ra);

From 54862540ba828d9c51c101c4915c0c11f03c25bf Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 13:29:48 +0000
Subject: [PATCH 299/532] Couple of fixes from DragonFly (typos).

Obtained from:	DragonFlyBSD
---
 share/man/man9/ieee80211.9      | 6 +++---
 share/man/man9/ieee80211_node.9 | 4 ++--
 2 files changed, 5 insertions(+), 5 deletions(-)

diff --git a/share/man/man9/ieee80211.9 b/share/man/man9/ieee80211.9
index ae92d5e5965..8fd6e0da655 100644
--- a/share/man/man9/ieee80211.9
+++ b/share/man/man9/ieee80211.9
@@ -25,11 +25,11 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd March 29, 2010
-.Dt NET80211 9
+.Dd April 28, 2010
+.Dt IEEE80211 9
 .Os
 .Sh NAME
-.Nm net80211
+.Nm IEEE80211 
 .Nd 802.11 network layer
 .Sh SYNOPSIS
 .In net80211/ieee80211_var.h
diff --git a/share/man/man9/ieee80211_node.9 b/share/man/man9/ieee80211_node.9
index be0b506a02f..9389411a2ff 100644
--- a/share/man/man9/ieee80211_node.9
+++ b/share/man/man9/ieee80211_node.9
@@ -26,7 +26,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd March 29, 2010
+.Dd April 28, 2010
 .Dt IEEE80211_NODE 9
 .Os
 .Sh NAME
@@ -152,7 +152,7 @@ displays the contents of a single node while
 .Fn ieee80211_dump_nodes
 displays the contents of the specified node table.
 Nodes may also be displayed using
-.Xr ddb 9
+.Xr ddb 4
 with the
 .Dq show node
 directive and the station node table can be displayed with

From dfe4a64b123deb8dbe208ee83d8d4fa73f29541b Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 13:42:13 +0000
Subject: [PATCH 300/532] Add ic_rxstream and ic_txstream. These variables hold
 the number of RX streams and the number of TX streams respectively.

Obtained from:	//depot/user/rpaulo/80211n/...
Sponsored by:	iXsystems, inc.
---
 sys/net80211/ieee80211_var.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/net80211/ieee80211_var.h b/sys/net80211/ieee80211_var.h
index ff3694f56a5..eea8dbe2855 100644
--- a/sys/net80211/ieee80211_var.h
+++ b/sys/net80211/ieee80211_var.h
@@ -213,6 +213,8 @@ struct ieee80211com {
 	enum ieee80211_protmode	ic_htprotmode;	/* HT protection mode */
 	int			ic_lastnonerp;	/* last time non-ERP sta noted*/
 	int			ic_lastnonht;	/* last time non-HT sta noted */
+	uint8_t			ic_rxstream;    /* # RX streams */
+	uint8_t			ic_txstream;    /* # TX streams */
 
 	/* optional state for Atheros SuperG protocol extensions */
 	struct ieee80211_superg	*ic_superg;
@@ -315,7 +317,6 @@ struct ieee80211com {
 	void			(*ic_ampdu_rx_stop)(struct ieee80211_node *,
 				    struct ieee80211_rx_ampdu *);
 	uint64_t		ic_spare[7];
-	uint32_t		ic_spare2;
 };
 
 struct ieee80211_aclator;

From ee6b5ace1f9b0df8fbf3ad4cd28c522032994834 Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Wed, 28 Apr 2010 14:59:05 +0000
Subject: [PATCH 301/532] Add a private field to struct ieee80211_rx_ampdu do
 hold driver specific data.

Sponsored by:	iXsystems, inc
---
 sys/net80211/ieee80211_ht.h | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/net80211/ieee80211_ht.h b/sys/net80211/ieee80211_ht.h
index 7b0eab7ace1..21e7ff631d8 100644
--- a/sys/net80211/ieee80211_ht.h
+++ b/sys/net80211/ieee80211_ht.h
@@ -142,6 +142,7 @@ struct ieee80211_rx_ampdu {
 	int		rxa_age;	/* age of oldest frame in window */
 	int		rxa_nframes;	/* frames since ADDBA */
 	struct mbuf *rxa_m[IEEE80211_AGGR_BAWMAX];
+	void		*rxa_private;
 	uint64_t	rxa_pad[4];
 };
 

From d8b878873e7aa8df1972cc6a642804b17eb61087 Mon Sep 17 00:00:00 2001
From: Attilio Rao <attilio@FreeBSD.org>
Date: Wed, 28 Apr 2010 15:38:01 +0000
Subject: [PATCH 302/532] - Extract the IODEV_PIO interface from ia64 and make
 it MI.   In the end, it does help fixing /dev/io usage from multithreaded  
 processes. - On i386 and amd64 the old behaviour is kept but multithreaded  
 processes must use the new interface in order to work well. - Support for the
 other architectures is greatly improved, where   necessary, by the necessity
 to define very small things now.

Manpage update will happen shortly.

Sponsored by:	Sandvine Incorporated
PR:		threads/116181
Reviewed by:	emaste, marcel
MFC after:	3 weeks
---
 sys/amd64/amd64/io.c          |  40 ++--------
 sys/amd64/include/iodev.h     |  21 ++++-
 sys/dev/io/iodev.c            | 146 +++++++++++++++++++++++++++++++---
 sys/dev/io/iodev.h            |  44 ++++++++++
 sys/i386/i386/io.c            |  40 ++--------
 sys/i386/include/iodev.h      |  21 ++++-
 sys/ia64/ia64/iodev_machdep.c | 114 +++-----------------------
 sys/ia64/include/iodev.h      |  33 ++++----
 8 files changed, 256 insertions(+), 203 deletions(-)
 create mode 100644 sys/dev/io/iodev.h

diff --git a/sys/amd64/amd64/io.c b/sys/amd64/amd64/io.c
index 09d6e896582..c2d0d515555 100644
--- a/sys/amd64/amd64/io.c
+++ b/sys/amd64/amd64/io.c
@@ -28,60 +28,32 @@
 __FBSDID("$FreeBSD$");
 
 #include <sys/param.h>
-#include <sys/conf.h>
-#include <sys/fcntl.h>
-#include <sys/lock.h>
-#include <sys/malloc.h>
-#include <sys/mutex.h>
-#include <sys/priv.h>
 #include <sys/proc.h>
-#include <sys/signalvar.h>
-#include <sys/systm.h>
 
-#include <machine/db_machdep.h>
 #include <machine/frame.h>
-#include <machine/psl.h>
-#include <machine/specialreg.h>
-
-#include <vm/vm.h>
-#include <vm/pmap.h>
-
 #include <machine/iodev.h>
+#include <machine/psl.h>
 
-/* ARGSUSED */
 int
-ioopen(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td)
+iodev_open(struct thread *td)
 {
-	int error;
-
-	error = priv_check(td, PRIV_IO);
-	if (error != 0)
-		return (error);
-	error = securelevel_gt(td->td_ucred, 0);
-	if (error != 0)
-		return (error);
 
 	td->td_frame->tf_rflags |= PSL_IOPL;
-
 	return (0);
 }
 
-/* ARGSUSED */
 int
-ioclose(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td)
+iodev_close(struct thread *td)
 {
+
 	td->td_frame->tf_rflags &= ~PSL_IOPL;
-
 	return (0);
 }
 
 /* ARGSUSED */
 int
-ioioctl(struct cdev *dev __unused, u_long cmd __unused, caddr_t data __unused,
-    int fflag __unused, struct thread *td __unused)
+iodev_ioctl(u_long cmd __unused, caddr_t data __unused)
 {
 
-	return (ENXIO);
+	return (ENOIOCTL);
 }
diff --git a/sys/amd64/include/iodev.h b/sys/amd64/include/iodev.h
index 1a0a17a9561..9f53cacff00 100644
--- a/sys/amd64/include/iodev.h
+++ b/sys/amd64/include/iodev.h
@@ -25,7 +25,22 @@
  *
  * $FreeBSD$
  */
+#ifndef _MACHINE_IODEV_H_
+#define	_MACHINE_IODEV_H_
 
-d_open_t	ioopen;
-d_close_t	ioclose;
-d_ioctl_t	ioioctl;
+#ifdef _KERNEL
+#include <machine/cpufunc.h>
+
+#define	iodev_read_1	inb
+#define	iodev_read_2	inw
+#define	iodev_read_4	inl
+#define	iodev_write_1	outb
+#define	iodev_write_2	outw
+#define	iodev_write_4	outl
+
+int	 iodev_open(struct thread *td);
+int	 iodev_close(struct thread *td);
+int	 iodev_ioctl(u_long cmd, caddr_t data);
+
+#endif /* _KERNEL */
+#endif /* _MACHINE_IODEV_H_ */
diff --git a/sys/dev/io/iodev.c b/sys/dev/io/iodev.c
index b142a39c610..eae69f484a4 100644
--- a/sys/dev/io/iodev.c
+++ b/sys/dev/io/iodev.c
@@ -30,22 +30,27 @@ __FBSDID("$FreeBSD$");
 
 #include <sys/param.h>
 #include <sys/conf.h>
-#include <sys/fcntl.h>
 #include <sys/kernel.h>
-#include <sys/lock.h>
-#include <sys/malloc.h>
+#include <sys/ioccom.h>
 #include <sys/module.h>
-#include <sys/mutex.h>
+#include <sys/priv.h>
 #include <sys/proc.h>
-#include <sys/signalvar.h>
 #include <sys/systm.h>
-#include <sys/uio.h>
-
-#include <vm/vm.h>
-#include <vm/pmap.h>
 
 #include <machine/iodev.h>
 
+#include <dev/io/iodev.h>
+
+static int	 ioopen(struct cdev *dev, int flags, int fmt,
+		    struct thread *td);
+static int	 ioclose(struct cdev *dev, int flags, int fmt,
+		    struct thread *td);
+static int	 ioioctl(struct cdev *dev, u_long cmd, caddr_t data,
+		    int fflag, struct thread *td);
+
+static int	 iopio_read(struct iodev_pio_req *req);
+static int	 iopio_write(struct iodev_pio_req *req);
+
 static struct cdev *iodev;
 
 static struct cdevsw io_cdevsw = {
@@ -56,6 +61,129 @@ static struct cdevsw io_cdevsw = {
 	.d_name =	"io",
 };
 
+/* ARGSUSED */
+static int
+ioopen(struct cdev *dev __unused, int flags __unused, int fmt __unused,
+    struct thread *td)
+{
+	int error;
+
+	error = priv_check(td, PRIV_IO);
+	if (error != 0)
+		return (error);
+	error = securelevel_gt(td->td_ucred, 0);
+	if (error != 0)
+		return (error);
+	error = iodev_open(td);
+
+        return (error);
+}
+
+/* ARGSUSED */
+static int
+ioclose(struct cdev *dev __unused, int flags __unused, int fmt __unused,
+    struct thread *td)
+{
+
+	return (iodev_close(td));
+}
+
+/* ARGSUSED */
+static int
+ioioctl(struct cdev *dev __unused, u_long cmd, caddr_t data,
+    int fflag __unused, struct thread *td __unused)
+{
+	struct iodev_pio_req *pio_req;
+	int error;
+
+	switch (cmd) {
+	case IODEV_PIO:
+		pio_req = (struct iodev_pio_req *)data;
+		switch (pio_req->access) {
+		case IODEV_PIO_READ:
+			error = iopio_read(pio_req);
+			break;
+		case IODEV_PIO_WRITE:
+			error = iopio_write(pio_req);
+			break;
+		default:
+			error = EINVAL;
+			break;
+		}
+		break;
+	default:
+		error = iodev_ioctl(cmd, data);
+	}
+
+	return (error);
+}
+
+static int
+iopio_read(struct iodev_pio_req *req)
+{
+
+	switch (req->width) {
+	case 1:
+		req->val = iodev_read_1(req->port);
+		break;
+	case 2:
+		if (req->port & 1) {
+			req->val = iodev_read_1(req->port);
+			req->val |= iodev_read_1(req->port + 1) << 8;
+		} else
+			req->val = iodev_read_2(req->port);
+		break;
+	case 4:
+		if (req->port & 1) {
+			req->val = iodev_read_1(req->port);
+			req->val |= iodev_read_2(req->port + 1) << 8;
+			req->val |= iodev_read_1(req->port + 3) << 24;
+		} else if (req->port & 2) {
+			req->val = iodev_read_2(req->port);
+			req->val |= iodev_read_2(req->port + 2) << 16;
+		} else
+			req->val = iodev_read_4(req->port);
+		break;
+	default:
+		return (EINVAL);
+	}
+
+	return (0);
+}
+
+static int
+iopio_write(struct iodev_pio_req *req)
+{
+
+	switch (req->width) {
+	case 1:
+		iodev_write_1(req->port, req->val);
+		break;
+	case 2:
+		if (req->port & 1) {
+			iodev_write_1(req->port, req->val);
+			iodev_write_1(req->port + 1, req->val >> 8);
+		} else
+			iodev_write_2(req->port, req->val);
+		break;
+	case 4:
+		if (req->port & 1) {
+			iodev_write_1(req->port, req->val);
+			iodev_write_2(req->port + 1, req->val >> 8);
+			iodev_write_1(req->port + 3, req->val >> 24);
+		} else if (req->port & 2) {
+			iodev_write_2(req->port, req->val);
+			iodev_write_2(req->port + 2, req->val >> 16);
+		} else
+			iodev_write_4(req->port, req->val);
+		break;
+	default:
+		return (EINVAL);
+	}
+
+	return (0);
+}
+
 /* ARGSUSED */
 static int
 io_modevent(module_t mod __unused, int type, void *data __unused)
diff --git a/sys/dev/io/iodev.h b/sys/dev/io/iodev.h
new file mode 100644
index 00000000000..d040fcccf48
--- /dev/null
+++ b/sys/dev/io/iodev.h
@@ -0,0 +1,44 @@
+/*-
+ * Copyright (c) 2010 Marcel Moolenaar
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef _DEV_IODEV_H_
+#define	_DEV_IODEV_H_
+
+#define	IODEV_PIO_READ		0
+#define	IODEV_PIO_WRITE		1
+
+struct iodev_pio_req {
+	u_int access;
+	u_int port;
+	u_int width;
+	u_int val;
+};
+
+#define	IODEV_PIO	_IOWR('I', 0, struct iodev_pio_req)
+
+#endif /* _DEV_IODEV_H_ */
diff --git a/sys/i386/i386/io.c b/sys/i386/i386/io.c
index c392af5f38d..152f6b1dc5e 100644
--- a/sys/i386/i386/io.c
+++ b/sys/i386/i386/io.c
@@ -28,60 +28,32 @@
 __FBSDID("$FreeBSD$");
 
 #include <sys/param.h>
-#include <sys/conf.h>
-#include <sys/fcntl.h>
-#include <sys/lock.h>
-#include <sys/malloc.h>
-#include <sys/mutex.h>
-#include <sys/priv.h>
 #include <sys/proc.h>
-#include <sys/signalvar.h>
-#include <sys/systm.h>
 
-#include <machine/db_machdep.h>
 #include <machine/frame.h>
-#include <machine/psl.h>
-#include <machine/specialreg.h>
-
-#include <vm/vm.h>
-#include <vm/pmap.h>
-
 #include <machine/iodev.h>
+#include <machine/psl.h>
 
-/* ARGSUSED */
 int
-ioopen(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td)
+iodev_open(struct thread *td)
 {
-	int error;
-
-	error = priv_check(td, PRIV_IO);
-	if (error != 0)
-		return (error);
-	error = securelevel_gt(td->td_ucred, 0);
-	if (error != 0)
-		return (error);
 
 	td->td_frame->tf_eflags |= PSL_IOPL;
-
 	return (0);
 }
 
-/* ARGSUSED */
 int
-ioclose(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td)
+iodev_close(struct thread *td)
 {
+
 	td->td_frame->tf_eflags &= ~PSL_IOPL;
-
 	return (0);
 }
 
 /* ARGSUSED */
 int
-ioioctl(struct cdev *dev __unused, u_long cmd __unused, caddr_t data __unused,
-    int fflag __unused, struct thread *td __unused)
+iodev_ioctl(u_long cmd __unused, caddr_t data __unused)
 {
 
-	return (ENXIO);
+	return (ENOIOCTL);
 }
diff --git a/sys/i386/include/iodev.h b/sys/i386/include/iodev.h
index 1a0a17a9561..9f53cacff00 100644
--- a/sys/i386/include/iodev.h
+++ b/sys/i386/include/iodev.h
@@ -25,7 +25,22 @@
  *
  * $FreeBSD$
  */
+#ifndef _MACHINE_IODEV_H_
+#define	_MACHINE_IODEV_H_
 
-d_open_t	ioopen;
-d_close_t	ioclose;
-d_ioctl_t	ioioctl;
+#ifdef _KERNEL
+#include <machine/cpufunc.h>
+
+#define	iodev_read_1	inb
+#define	iodev_read_2	inw
+#define	iodev_read_4	inl
+#define	iodev_write_1	outb
+#define	iodev_write_2	outw
+#define	iodev_write_4	outl
+
+int	 iodev_open(struct thread *td);
+int	 iodev_close(struct thread *td);
+int	 iodev_ioctl(u_long cmd, caddr_t data);
+
+#endif /* _KERNEL */
+#endif /* _MACHINE_IODEV_H_ */
diff --git a/sys/ia64/ia64/iodev_machdep.c b/sys/ia64/ia64/iodev_machdep.c
index d255aae127b..9d9057dc76e 100644
--- a/sys/ia64/ia64/iodev_machdep.c
+++ b/sys/ia64/ia64/iodev_machdep.c
@@ -40,31 +40,13 @@ __FBSDID("$FreeBSD$");
 #include <machine/efi.h>
 #include <machine/iodev.h>
 
-static int iodev_pio_read(struct iodev_pio_req *req);
-static int iodev_pio_write(struct iodev_pio_req *req);
-
 static int iodev_efivar_getvar(struct iodev_efivar_req *req);
 static int iodev_efivar_nextname(struct iodev_efivar_req *req);
 static int iodev_efivar_setvar(struct iodev_efivar_req *req);
 
 /* ARGSUSED */
 int
-ioopen(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td)
-{
-	int error;
-
-	error = priv_check(td, PRIV_IO);
-	if (error == 0)
-		error = securelevel_gt(td->td_ucred, 0);
-
-	return (error);
-}
-
-/* ARGSUSED */
-int
-ioclose(struct cdev *dev __unused, int flags __unused, int fmt __unused,
-    struct thread *td __unused)
+iodev_open(struct thread *td __unused)
 {
 
 	return (0);
@@ -72,29 +54,19 @@ ioclose(struct cdev *dev __unused, int flags __unused, int fmt __unused,
 
 /* ARGSUSED */
 int
-ioioctl(struct cdev *dev __unused, u_long cmd, caddr_t data,
-    int fflag __unused, struct thread *td __unused)
+iodev_close(struct thread *td __unused)
+{
+
+	return (0);
+}
+
+int
+iodev_ioctl(u_long cmd, caddr_t data)
 {
 	struct iodev_efivar_req *efivar_req;
-	struct iodev_pio_req *pio_req;
 	int error;
 
-	error = ENOIOCTL;
 	switch (cmd) {
-	case IODEV_PIO:
-		pio_req = (struct iodev_pio_req *)data;
-		switch (pio_req->access) {
-		case IODEV_PIO_READ:
-			error = iodev_pio_read(pio_req);
-			break;
-		case IODEV_PIO_WRITE:
-			error = iodev_pio_write(pio_req);
-			break;
-		default:
-			error = EINVAL;
-			break;
-		}
-		break;
 	case IODEV_EFIVAR:
 		efivar_req = (struct iodev_efivar_req *)data;
 		efivar_req->result = 0;		/* So it's well-defined */
@@ -113,77 +85,13 @@ ioioctl(struct cdev *dev __unused, u_long cmd, caddr_t data,
 			break;
 		}
 		break;
+	default:
+		error = ENOIOCTL;
 	}
 
 	return (error);
 }
 
-static int
-iodev_pio_read(struct iodev_pio_req *req)
-{
-
-	switch (req->width) {
-	case 1:
-		req->val = bus_space_read_io_1(req->port);
-		break;
-	case 2:
-		if (req->port & 1) {
-			req->val = bus_space_read_io_1(req->port);
-			req->val |= bus_space_read_io_1(req->port + 1) << 8;
-		} else
-			req->val = bus_space_read_io_2(req->port);
-		break;
-	case 4:
-		if (req->port & 1) {
-			req->val = bus_space_read_io_1(req->port);
-			req->val |= bus_space_read_io_2(req->port + 1) << 8;
-			req->val |= bus_space_read_io_1(req->port + 3) << 24;
-		} else if (req->port & 2) {
-			req->val = bus_space_read_io_2(req->port);
-			req->val |= bus_space_read_io_2(req->port + 2) << 16;
-		} else
-			req->val = bus_space_read_io_4(req->port);
-		break;
-	default:
-		return (EINVAL);
-	}
-
-	return (0);
-}
-
-static int
-iodev_pio_write(struct iodev_pio_req *req)
-{
-
-	switch (req->width) {
-	case 1:
-		bus_space_write_io_1(req->port, req->val);
-		break;
-	case 2:
-		if (req->port & 1) {
-			bus_space_write_io_1(req->port, req->val);
-			bus_space_write_io_1(req->port + 1, req->val >> 8);
-		} else
-			bus_space_write_io_2(req->port, req->val);
-		break;
-	case 4:
-		if (req->port & 1) {
-			bus_space_write_io_1(req->port, req->val);
-			bus_space_write_io_2(req->port + 1, req->val >> 8);
-			bus_space_write_io_1(req->port + 3, req->val >> 24);
-		} else if (req->port & 2) {
-			bus_space_write_io_2(req->port, req->val);
-			bus_space_write_io_2(req->port + 2, req->val >> 16);
-		} else
-			bus_space_write_io_4(req->port, req->val);
-		break;
-	default:
-		return (EINVAL);
-	}
-
-	return (0);
-}
-
 static int
 iodev_efivar_getvar(struct iodev_efivar_req *req)
 {
diff --git a/sys/ia64/include/iodev.h b/sys/ia64/include/iodev.h
index 6d2ae19d36d..cf349d9b068 100644
--- a/sys/ia64/include/iodev.h
+++ b/sys/ia64/include/iodev.h
@@ -31,22 +31,16 @@
 
 #include <sys/uuid.h>
 
-struct iodev_pio_req {
-	u_int access;
-#define	IODEV_PIO_READ		0
-#define	IODEV_PIO_WRITE		1
-	u_int port;
-	u_int width;
-	u_int val;
-};
+#ifdef _KERNEL
+#include <machine/bus.h>
+#endif
 
-#define	IODEV_PIO	_IOWR('I', 0, struct iodev_pio_req)
-
-struct iodev_efivar_req {
-	u_int	access;
 #define	IODEV_EFIVAR_GETVAR	0
 #define	IODEV_EFIVAR_NEXTNAME	1
 #define	IODEV_EFIVAR_SETVAR	2
+
+struct iodev_efivar_req {
+	u_int	access;
 	u_int	result;			/* errno value */
 	size_t	namesize;
 	u_short	*name;			/* UCS-2 */
@@ -59,11 +53,16 @@ struct iodev_efivar_req {
 #define	IODEV_EFIVAR	_IOWR('I', 1, struct iodev_efivar_req)
 
 #ifdef _KERNEL
+#define	iodev_read_1	bus_space_read_io_1
+#define	iodev_read_2	bus_space_read_io_2
+#define	iodev_read_4	bus_space_read_io_4
+#define	iodev_write_1	bus_space_write_io_1
+#define	iodev_write_2	bus_space_write_io_2
+#define	iodev_write_4	bus_space_write_io_4
 
-d_open_t	ioopen;
-d_close_t	ioclose;
-d_ioctl_t	ioioctl;
-
-#endif
+int	 iodev_open(struct thread *td);
+int	 iodev_close(struct thread *td);
+int	 iodev_ioctl(u_long, caddr_t data);
 
+#endif /* _KERNEL */
 #endif /* _MACHINE_IODEV_H_ */

From 8c975e70afa1d9d6224d2204cc0ea0ca309022e5 Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Wed, 28 Apr 2010 17:26:05 +0000
Subject: [PATCH 303/532] Remove dead code.  Calculated greatest common divisor
 was not used at all. On top of that, LLVM+Clang mis-compiles this code
 because of its register allocator bug.

Analyzed by:	Andrew Reilly (areilly at bigpond dot net dot au)
Reviewed by:	ariff, rdivacky
MFC after:	3 days
---
 sys/dev/sound/pcm/buffer.c | 11 -----------
 1 file changed, 11 deletions(-)

diff --git a/sys/dev/sound/pcm/buffer.c b/sys/dev/sound/pcm/buffer.c
index 6e8259202dc..a9053d751d8 100644
--- a/sys/dev/sound/pcm/buffer.c
+++ b/sys/dev/sound/pcm/buffer.c
@@ -568,7 +568,6 @@ sndbuf_updateprevtotal(struct snd_dbuf *b)
 unsigned int
 snd_xbytes(unsigned int v, unsigned int from, unsigned int to)
 {
-	unsigned int w, x, y;
 
 	if (from == to)
 		return v;
@@ -576,16 +575,6 @@ snd_xbytes(unsigned int v, unsigned int from, unsigned int to)
 	if (from == 0 || to == 0 || v == 0)
 		return 0;
 
-	x = from;
-	y = to;
-	while (y != 0) {
-		w = x % y;
-		x = y;
-		y = w;
-	}
-	from /= x;
-	to /= x;
-
 	return (unsigned int)(((u_int64_t)v * to) / from);
 }
 

From 1655af0a720bd444ef75cc351cd381c33b98888a Mon Sep 17 00:00:00 2001
From: Jack F Vogel <jfv@FreeBSD.org>
Date: Wed, 28 Apr 2010 17:37:30 +0000
Subject: [PATCH 304/532] Change default WOL back to MAGIC only, having
 multicast enabled causes problems in man environments.

---
 sys/dev/e1000/if_em.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
index 1db90121ddb..fcebc02930a 100644
--- a/sys/dev/e1000/if_em.c
+++ b/sys/dev/e1000/if_em.c
@@ -2707,10 +2707,10 @@ em_setup_interface(device_t dev, struct adapter *adapter)
 	ifp->if_capabilities |= IFCAP_POLLING;
 #endif
 
-	/* Enable All WOL methods by default */
+	/* Enable only WOL MAGIC by default */
 	if (adapter->wol) {
 		ifp->if_capabilities |= IFCAP_WOL;
-		ifp->if_capenable |= IFCAP_WOL;
+		ifp->if_capenable |= IFCAP_WOL_MAGIC;
 	}
 		
 	/*

From b19b0de471acdb479b26017e573ce8babb43f7d0 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Wed, 28 Apr 2010 18:29:48 +0000
Subject: [PATCH 305/532] Backport fix for 'zfs_znode_dmu_init: existing znode
 for dbuf' panic from OpenSolaris.

PR:		kern/144402
Reported by:	Alex Bakhtin <alex.bakhtin@gmail.com>
Tested by:	Alex Bakhtin <alex.bakhtin@gmail.com>
Obtained from:	OpenSolaris, Bug ID 6895088
MFC after:	3 days
---
 .../opensolaris/uts/common/fs/zfs/zfs_znode.c | 43 +++++++++++++------
 1 file changed, 31 insertions(+), 12 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c
index 900087b8b32..947f9dd39f4 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c
@@ -704,6 +704,8 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
 			    DMU_OT_ZNODE, sizeof (znode_phys_t) + bonuslen, tx);
 		}
 	}
+
+	ZFS_OBJ_HOLD_ENTER(zfsvfs, obj);
 	VERIFY(0 == dmu_bonus_hold(zfsvfs->z_os, obj, NULL, &db));
 	dmu_buf_will_dirty(db, tx);
 
@@ -765,9 +767,7 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
 
 	pzp->zp_mode = MAKEIMODE(vap->va_type, vap->va_mode);
 	if (!(flag & IS_ROOT_NODE)) {
-		ZFS_OBJ_HOLD_ENTER(zfsvfs, obj);
 		*zpp = zfs_znode_alloc(zfsvfs, db, 0);
-		ZFS_OBJ_HOLD_EXIT(zfsvfs, obj);
 	} else {
 		/*
 		 * If we are creating the root node, the "parent" we
@@ -776,6 +776,7 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr,
 		*zpp = dzp;
 	}
 	zfs_perm_init(*zpp, dzp, flag, vap, tx, cr, setaclp, fuidp);
+	ZFS_OBJ_HOLD_EXIT(zfsvfs, obj);
 	if (!(flag & IS_ROOT_NODE)) {
 		vnode_t *vp;
 
@@ -939,19 +940,31 @@ again:
 
 	/*
 	 * Not found create new znode/vnode
+	 * but only if file exists.
+	 *
+	 * There is a small window where zfs_vget() could
+	 * find this object while a file create is still in
+	 * progress.  Since a gen number can never be zero
+	 * we will check that to determine if its an allocated
+	 * file.
 	 */
-	zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size);
-
-	vp = ZTOV(zp);
-	vp->v_vflag |= VV_FORCEINSMQ;
-	err = insmntque(vp, zfsvfs->z_vfs);
-	vp->v_vflag &= ~VV_FORCEINSMQ;
-	KASSERT(err == 0, ("insmntque() failed: error %d", err));
-	VOP_UNLOCK(vp, 0);
 
+	if (((znode_phys_t *)db->db_data)->zp_gen != 0) {
+		zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size);
+		*zpp = zp;
+		vp = ZTOV(zp);
+		vp->v_vflag |= VV_FORCEINSMQ;
+		err = insmntque(vp, zfsvfs->z_vfs);
+		vp->v_vflag &= ~VV_FORCEINSMQ;
+		KASSERT(err == 0, ("insmntque() failed: error %d", err));
+		VOP_UNLOCK(vp, 0);
+		err = 0;
+	} else {
+		dmu_buf_rele(db, NULL);
+		err = ENOENT;
+	}
 	ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num);
-	*zpp = zp;
-	return (0);
+	return (err);
 }
 
 int
@@ -1440,6 +1453,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
 	uint64_t	norm = 0;
 	nvpair_t	*elem;
 	int		error;
+	int		i;
 	znode_t		*rootzp = NULL;
 	vnode_t		vnode;
 	vattr_t		vattr;
@@ -1537,6 +1551,9 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
 	list_create(&zfsvfs.z_all_znodes, sizeof (znode_t),
 	    offsetof(znode_t, z_link_node));
 
+	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
+		mutex_init(&zfsvfs.z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL);
+
 	ASSERT(!POINTER_IS_VALID(rootzp->z_zfsvfs));
 	rootzp->z_zfsvfs = &zfsvfs;
 	zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, 0, NULL, NULL);
@@ -1547,6 +1564,8 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx)
 
 	dmu_buf_rele(rootzp->z_dbuf, NULL);
 	rootzp->z_dbuf = NULL;
+	for (i = 0; i != ZFS_OBJ_MTX_SZ; i++)
+		mutex_destroy(&zfsvfs.z_hold_mtx[i]);
 	mutex_destroy(&zfsvfs.z_znodes_lock);
 	rootzp->z_vnode = NULL;
 	kmem_cache_free(znode_cache, rootzp);

From aa1385cd7b366e2224febb0d3fce71f872ea8838 Mon Sep 17 00:00:00 2001
From: John Baldwin <jhb@FreeBSD.org>
Date: Wed, 28 Apr 2010 18:49:45 +0000
Subject: [PATCH 306/532] Initialize the callout structure earlier in attach
 before calling any routines that can fail since ciss_free() always tries to
 stop and drain the callout.

---
 sys/dev/ciss/ciss.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/ciss/ciss.c b/sys/dev/ciss/ciss.c
index 7293bb1682d..2a4fb27ae55 100644
--- a/sys/dev/ciss/ciss.c
+++ b/sys/dev/ciss/ciss.c
@@ -417,6 +417,7 @@ ciss_attach(device_t dev)
     sc = device_get_softc(dev);
     sc->ciss_dev = dev;
     mtx_init(&sc->ciss_mtx, "cissmtx", NULL, MTX_DEF);
+    callout_init_mtx(&sc->ciss_periodic, &sc->ciss_mtx, 0);
 
     /*
      * Do PCI-specific init.
@@ -429,7 +430,6 @@ ciss_attach(device_t dev)
      */
     ciss_initq_free(sc);
     ciss_initq_notify(sc);
-    callout_init_mtx(&sc->ciss_periodic, &sc->ciss_mtx, 0);
 
     /*
      * Initalize device sysctls.

From beef45ff880d30654101f4e20aed97762d2c70f1 Mon Sep 17 00:00:00 2001
From: Jack F Vogel <jfv@FreeBSD.org>
Date: Wed, 28 Apr 2010 19:22:52 +0000
Subject: [PATCH 307/532] Address the LOD that some are seeing, put the RX lock
 back in rxeof (I could see little point in taking it out), and now release it
 before the stack entry.

Also, make it so the 82574 does not configure for multiqueue
when its not used in the stack.
---
 sys/dev/e1000/if_em.c | 23 +++++++++++++----------
 1 file changed, 13 insertions(+), 10 deletions(-)

diff --git a/sys/dev/e1000/if_em.c b/sys/dev/e1000/if_em.c
index fcebc02930a..2fb4f174464 100644
--- a/sys/dev/e1000/if_em.c
+++ b/sys/dev/e1000/if_em.c
@@ -347,8 +347,13 @@ static int em_debug_sbp = FALSE;
 TUNABLE_INT("hw.em.sbp", &em_debug_sbp);
 
 /* Local controls for MSI/MSIX */
+#ifdef EM_MULTIQUEUE
 static int em_enable_msix = TRUE;
 static int em_msix_queues = 2; /* for 82574, can be 1 or 2 */
+#else
+static int em_enable_msix = FALSE;
+static int em_msix_queues = 0; /* disable */
+#endif
 TUNABLE_INT("hw.em.enable_msix", &em_enable_msix);
 TUNABLE_INT("hw.em.msix_queues", &em_msix_queues);
 
@@ -1371,9 +1376,7 @@ em_poll(struct ifnet *ifp, enum poll_cmd cmd, int count)
 	}
 	EM_CORE_UNLOCK(adapter);
 
-	EM_RX_LOCK(rxr);
 	rx_done = em_rxeof(rxr, count);
-	EM_RX_UNLOCK(rxr);
 
 	EM_TX_LOCK(txr);
 	em_txeof(txr);
@@ -1449,9 +1452,7 @@ em_handle_que(void *context, int pending)
 
 
 	if (ifp->if_drv_flags & IFF_DRV_RUNNING) {
-		EM_RX_LOCK(rxr);
 		more_rx = em_rxeof(rxr, adapter->rx_process_limit);
-		EM_RX_UNLOCK(rxr);
 
 		EM_TX_LOCK(txr);
 		em_txeof(txr);
@@ -1511,10 +1512,8 @@ em_msix_rx(void *arg)
 	struct adapter	*adapter = rxr->adapter;
 	bool		more;
 
-	EM_RX_LOCK(rxr);
 	++rxr->rx_irq;
 	more = em_rxeof(rxr, adapter->rx_process_limit);
-	EM_RX_UNLOCK(rxr);
 	if (more)
 		taskqueue_enqueue(rxr->tq, &rxr->rx_task);
 	else
@@ -1553,9 +1552,7 @@ em_handle_rx(void *context, int pending)
 	struct adapter	*adapter = rxr->adapter;
         bool            more;
 
-	EM_RX_LOCK(rxr);
 	more = em_rxeof(rxr, adapter->rx_process_limit);
-	EM_RX_UNLOCK(rxr);
 	if (more)
 		taskqueue_enqueue(rxr->tq, &rxr->rx_task);
 	else
@@ -4100,7 +4097,7 @@ em_rxeof(struct rx_ring *rxr, int count)
 	bool			eop;
 	struct e1000_rx_desc	*cur;
 
-	EM_RX_LOCK_ASSERT(rxr);
+	EM_RX_LOCK(rxr);
 
 	for (i = rxr->next_to_check, processed = 0; count != 0;) {
 
@@ -4194,8 +4191,13 @@ skip:
 			i = 0;
 
 		/* Send to the stack */
-		if (sendmp != NULL)
+		if (sendmp != NULL) {
+			rxr->next_to_check = i;
+			EM_RX_UNLOCK(rxr);
 			(*ifp->if_input)(ifp, sendmp);
+			EM_RX_LOCK(rxr);
+			i = rxr->next_to_check;
+		}
 
 		/* Only refresh mbufs every 8 descriptors */
 		if (processed == 8) {
@@ -4211,6 +4213,7 @@ skip:
 	}
 
 	rxr->next_to_check = i;
+	EM_RX_UNLOCK(rxr);
 
 #ifdef DEVICE_POLLING
 	return (rxdone);

From 5b663f0c698a9ecf7e562f7f7f730d056e499b5f Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Wed, 28 Apr 2010 21:50:57 +0000
Subject: [PATCH 308/532] Import ACPICA 20100428.

---
 changes.txt                           |    66 +
 common/adisasm.c                      |     6 +-
 compiler/Makefile                     |    38 +-
 compiler/aslcompile.c                 |     6 +-
 compiler/aslcompiler.h                |   126 +-
 compiler/asldefine.h                  |     2 +-
 compiler/aslfiles.c                   |    30 +
 compiler/aslglobal.h                  |     3 +
 compiler/asllisting.c                 |   269 +-
 compiler/aslmain.c                    |    31 +-
 compiler/aslpredef.c                  |    82 +-
 compiler/aslresource.c                |   294 +-
 compiler/aslrestype1.c                |   562 +-
 compiler/aslrestype1i.c               |   668 ++
 compiler/aslrestype2.c                |  2396 +-----
 compiler/aslrestype2d.c               |   814 ++
 compiler/aslrestype2e.c               |   646 ++
 compiler/aslrestype2q.c               |   793 ++
 compiler/aslrestype2w.c               |   774 ++
 compiler/asltypes.h                   |    22 +-
 debugger/dbcmds.c                     |     2 +-
 debugger/dbdisply.c                   |    14 +-
 events/evgpe.c                        |     2 +-
 events/evgpeblk.c                     |   774 +-
 events/evgpeinit.c                    |   763 ++
 events/evgpeutil.c                    |   452 ++
 executer/exconfig.c                   |    21 +-
 executer/exoparg1.c                   |     2 +-
 executer/exsystem.c                   |     8 +-
 include/acevents.h                    |    74 +-
 include/acglobal.h                    |     7 +-
 include/acinterp.h                    |     2 +-
 include/aclocal.h                     |    16 +-
 include/acpiosxf.h                    |     4 +-
 include/acpixf.h                      |    30 +-
 include/actypes.h                     |     2 +-
 include/amlresrc.h                    |     6 +
 os_specific/service_layers/osunixxf.c |     4 +-
 os_specific/service_layers/oswinxf.c  |     4 +-
 osunixxf.c                            |     4 +-
 tables/tbfind.c                       |     2 +-
 tables/tbinstal.c                     |    52 +-
 tables/tbutils.c                      |    15 +-
 tables/tbxface.c                      |    20 +-
 tests/misc/badcode.asl                |   295 +
 tests/misc/grammar.asl                | 10266 ++++++++++++++++++++++++
 tools/acpiexec/Makefile               |     4 +-
 tools/acpiexec/aeexec.c               |     1 -
 tools/acpiexec/aetables.c             |   141 +-
 tools/acpisrc/astable.c               |     1 -
 utilities/utcopy.c                    |    20 +-
 utilities/utglobal.c                  |     4 +-
 utilities/uttrack.c                   |   114 +-
 53 files changed, 16803 insertions(+), 3951 deletions(-)
 create mode 100644 compiler/aslrestype1i.c
 create mode 100644 compiler/aslrestype2d.c
 create mode 100644 compiler/aslrestype2e.c
 create mode 100644 compiler/aslrestype2q.c
 create mode 100644 compiler/aslrestype2w.c
 create mode 100644 events/evgpeinit.c
 create mode 100644 events/evgpeutil.c
 create mode 100644 tests/misc/badcode.asl
 create mode 100644 tests/misc/grammar.asl

diff --git a/changes.txt b/changes.txt
index d2c8c0e41ed..fed0395e61a 100644
--- a/changes.txt
+++ b/changes.txt
@@ -1,3 +1,69 @@
+----------------------------------------
+28 April 2010. Summary of changes for version 20100428:
+
+1) ACPI CA Core Subsystem:
+
+Implemented GPE support for dynamically loaded ACPI tables. For all GPEs, 
+including FADT-based and GPE Block Devices, execute any _PRW methods in the 
+new table, and process any _Lxx/_Exx GPE methods in the new table. Any 
+runtime GPE that is referenced by an _Lxx/_Exx method in the new table is 
+immediately enabled. Handles the FADT-defined GPEs as well as GPE Block 
+Devices. Provides compatibility with other ACPI implementations. Two new 
+files added, evgpeinit.c and evgpeutil.c. ACPICA BZ 833. Lin Ming, Bob Moore.
+
+Fixed a regression introduced in version 20100331 within the table manager 
+where initial table loading could fail. This was introduced in the fix for 
+AcpiReallocateRootTable. Also, renamed some of fields in the table manager 
+data structures to clarify their meaning and use.
+
+Fixed a possible allocation overrun during internal object copy in 
+AcpiUtCopySimpleObject. The original code did not correctly handle the case 
+where the object to be copied was a namespace node. Lin Ming. ACPICA BZ 847.
+
+Updated the allocation dump routine, AcpiUtDumpAllocation and fixed a 
+possible access beyond end-of-allocation. Also, now fully validate descriptor 
+(size and type) before output. Lin Ming, Bob Moore. ACPICA BZ 847
+
+Example Code and Data Size: These are the sizes for the OS-independent 
+acpica.lib produced by the Microsoft Visual C++ 6.0 32-bit compiler. The 
+debug version of the code includes the debug output trace mechanism and has a 
+much larger code and data size.
+
+  Previous Release:
+    Non-Debug Version:  87.9K Code, 18.6K Data, 106.5K Total
+    Debug Version:     163.5K Code, 51.3K Data, 214.8K Total
+  Current Release:
+    Non-Debug Version:  88.4K Code, 18.8K Data, 107.2K Total
+    Debug Version:     164.2K Code, 51.5K Data, 215.7K Total
+
+2) iASL Compiler/Disassembler and Tools:
+
+iASL: Implemented Min/Max/Len/Gran validation for address resource 
+descriptors. This change implements validation for the address fields that 
+are common to all address-type resource descriptors. These checks are 
+implemented: Checks for valid Min/Max, length within the Min/Max window, 
+valid granularity, Min/Max a multiple of granularity, and _MIF/_MAF as per 
+table 6-40 in the ACPI 4.0a specification. Also split the large aslrestype1.c 
+and aslrestype2.c files into five new files. ACPICA BZ 840.
+
+iASL: Added support for the _Wxx predefined names. This support was missing 
+and these names were not recognized by the compiler as valid predefined 
+names. ACPICA BZ 851.
+
+iASL: Added an error for all predefined names that are defined to return no 
+value and thus must be implemented as Control Methods. These include all of 
+the _Lxx, _Exx, _Wxx, and _Qxx names, as well as some other miscellaneous 
+names such as _DIS, _INI, _IRC, _OFF, _ON, and _PSx. ACPICA BZ 850, 856.
+
+iASL: Implemented the -ts option to emit hex AML data in ASL format, as an 
+ASL Buffer. Allows ACPI tables to be easily included within ASL files, to be 
+dynamically loaded via the Load() operator. Also cleaned up output for the -
+ta and -tc options. ACPICA BZ 853.
+
+Tests: Added a new file with examples of extended iASL error checking. 
+Demonstrates the advanced error checking ability of the iASL compiler. 
+Available at tests/misc/badcode.asl.
+
 ----------------------------------------
 31 March 2010. Summary of changes for version 20100331:
 
diff --git a/common/adisasm.c b/common/adisasm.c
index 093cfe5536e..eced93d536b 100644
--- a/common/adisasm.c
+++ b/common/adisasm.c
@@ -282,8 +282,8 @@ AdInitialize (
 
     /* Setup the Table Manager (cheat - there is no RSDT) */
 
-    AcpiGbl_RootTableList.Size = 1;
-    AcpiGbl_RootTableList.Count = 0;
+    AcpiGbl_RootTableList.MaxTableCount = 1;
+    AcpiGbl_RootTableList.CurrentTableCount = 0;
     AcpiGbl_RootTableList.Tables = LocalTables;
 
     return (Status);
@@ -1156,7 +1156,7 @@ AdParseTable (
 
     /* If LoadTable is FALSE, we are parsing the last loaded table */
 
-    TableIndex = AcpiGbl_RootTableList.Count - 1;
+    TableIndex = AcpiGbl_RootTableList.CurrentTableCount - 1;
 
     /* Pass 2 */
 
diff --git a/compiler/Makefile b/compiler/Makefile
index b4c03fd0627..2bdd0d31d26 100644
--- a/compiler/Makefile
+++ b/compiler/Makefile
@@ -1,11 +1,37 @@
  
-
 PROG=	iasl
-SRCS=	aslcompilerparse.c aslcompilerlex.c aslanalyze.c aslcodegen.c \
-	aslcompile.c aslerror.c aslfiles.c asllength.c \
-	asllisting.c aslload.c asllookup.c aslmain.c aslmap.c aslopcodes.c \
-	asloperands.c aslpredef.c aslresource.c aslrestype1.c aslrestype2.c aslstartup.c \
-	asltree.c aslutils.c asltransform.c aslfold.c aslstubs.c aslopt.c \
+SRCS=	\
+	aslcompilerparse.c \
+	aslcompilerlex.c \
+	aslanalyze.c \
+	aslcodegen.c \
+	aslcompile.c \
+	aslerror.c \
+	aslfiles.c \
+	aslfold.c \
+	asllength.c \
+	asllisting.c \
+	aslload.c \
+	asllookup.c \
+	aslmain.c \
+	aslmap.c \
+	aslopcodes.c \
+	asloperands.c \
+	aslopt.c \
+	aslpredef.c \
+	aslresource.c \
+	aslrestype1.c \
+	aslrestype1i.c \
+	aslrestype2.c \
+	aslrestype2d.c \
+	aslrestype2e.c \
+	aslrestype2q.c \
+	aslrestype2w.c \
+	aslstartup.c \
+	aslstubs.c \
+	asltransform.c \
+	asltree.c \
+	aslutils.c \
 	../common/getopt.c \
 	../utilities/utalloc.c \
 	../utilities/utcache.c \
diff --git a/compiler/aslcompile.c b/compiler/aslcompile.c
index 90e90991b7e..d6cb8adc738 100644
--- a/compiler/aslcompile.c
+++ b/compiler/aslcompile.c
@@ -177,7 +177,8 @@ AslCompilerSignon (
         {
             Prefix = "; ";
         }
-        else if (Gbl_HexOutputFlag == HEX_OUTPUT_C)
+        else if ((Gbl_HexOutputFlag == HEX_OUTPUT_C) ||
+                 (Gbl_HexOutputFlag == HEX_OUTPUT_ASL))
         {
             FlPrintFile (ASL_FILE_HEX_OUTPUT, "/*\n");
             Prefix = " * ";
@@ -265,7 +266,8 @@ AslCompilerFileHeader (
         {
             Prefix = "; ";
         }
-        else if (Gbl_HexOutputFlag == HEX_OUTPUT_C)
+        else if ((Gbl_HexOutputFlag == HEX_OUTPUT_C) ||
+                 (Gbl_HexOutputFlag == HEX_OUTPUT_ASL))
         {
             Prefix = " * ";
         }
diff --git a/compiler/aslcompiler.h b/compiler/aslcompiler.h
index 50837982989..804713b00d9 100644
--- a/compiler/aslcompiler.h
+++ b/compiler/aslcompiler.h
@@ -592,6 +592,10 @@ FlFileError (
     UINT32                  FileId,
     UINT8                   ErrorId);
 
+UINT32
+FlGetFileSize (
+    UINT32                  FileId);
+
 ACPI_STATUS
 FlReadFile (
     UINT32                  FileId,
@@ -750,6 +754,34 @@ UtDoConstant (
 /*
  * aslresource - Resource template generation utilities
  */
+void
+RsSmallAddressCheck (
+    UINT8                   Type,
+    UINT32                  Minimum,
+    UINT32                  Maximum,
+    UINT32                  Length,
+    UINT32                  Alignment,
+    ACPI_PARSE_OBJECT       *MinOp,
+    ACPI_PARSE_OBJECT       *MaxOp,
+    ACPI_PARSE_OBJECT       *LengthOp,
+    ACPI_PARSE_OBJECT       *AlignOp);
+
+void
+RsLargeAddressCheck (
+    UINT64                  Minimum,
+    UINT64                  Maximum,
+    UINT64                  Length,
+    UINT64                  Granularity,
+    UINT8                   Flags,
+    ACPI_PARSE_OBJECT       *MinOp,
+    ACPI_PARSE_OBJECT       *MaxOp,
+    ACPI_PARSE_OBJECT       *LengthOp,
+    ACPI_PARSE_OBJECT       *GranOp);
+
+UINT16
+RsGetStringDataLength (
+    ACPI_PARSE_OBJECT       *InitializerOp);
+
 ASL_RESOURCE_NODE *
 RsAllocateResourceNode (
     UINT32                  Size);
@@ -805,43 +837,18 @@ RsDoResourceTemplate (
 
 
 /*
- * aslrestype1 - generate Small descriptors
+ * aslrestype1 - Miscellaneous Small descriptors
  */
 ASL_RESOURCE_NODE *
 RsDoEndTagDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
-ASL_RESOURCE_NODE *
-RsDoDmaDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
 ASL_RESOURCE_NODE *
 RsDoEndDependentDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
-ASL_RESOURCE_NODE *
-RsDoFixedIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
-ASL_RESOURCE_NODE *
-RsDoIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
-ASL_RESOURCE_NODE *
-RsDoIrqDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
-ASL_RESOURCE_NODE *
-RsDoIrqNoFlagsDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
 ASL_RESOURCE_NODE *
 RsDoMemory24Descriptor (
     ACPI_PARSE_OBJECT       *Op,
@@ -874,7 +881,36 @@ RsDoVendorSmallDescriptor (
 
 
 /*
- * aslrestype2 - generate Large descriptors
+ * aslrestype1i - I/O-related Small descriptors
+ */
+ASL_RESOURCE_NODE *
+RsDoDmaDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+ASL_RESOURCE_NODE *
+RsDoFixedIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+ASL_RESOURCE_NODE *
+RsDoIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+ASL_RESOURCE_NODE *
+RsDoIrqDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+ASL_RESOURCE_NODE *
+RsDoIrqNoFlagsDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+
+/*
+ * aslrestype2 - Large resource descriptors
  */
 ASL_RESOURCE_NODE *
 RsDoInterruptDescriptor (
@@ -882,6 +918,20 @@ RsDoInterruptDescriptor (
     UINT32                  CurrentByteOffset);
 
 ASL_RESOURCE_NODE *
+RsDoVendorLargeDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+ASL_RESOURCE_NODE *
+RsDoGeneralRegisterDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset);
+
+
+/*
+ * aslrestype2d - DWord address descriptors
+ */
+ASL_RESOURCE_NODE *
 RsDoDwordIoDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
@@ -896,6 +946,10 @@ RsDoDwordSpaceDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
+
+/*
+ * aslrestype2e - Extended address descriptors
+ */
 ASL_RESOURCE_NODE *
 RsDoExtendedIoDescriptor (
     ACPI_PARSE_OBJECT       *Op,
@@ -911,6 +965,10 @@ RsDoExtendedSpaceDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
+
+/*
+ * aslrestype2q - QWord address descriptors
+ */
 ASL_RESOURCE_NODE *
 RsDoQwordIoDescriptor (
     ACPI_PARSE_OBJECT       *Op,
@@ -926,6 +984,10 @@ RsDoQwordSpaceDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
+
+/*
+ * aslrestype2w - Word address descriptors
+ */
 ASL_RESOURCE_NODE *
 RsDoWordIoDescriptor (
     ACPI_PARSE_OBJECT       *Op,
@@ -941,15 +1003,5 @@ RsDoWordBusNumberDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset);
 
-ASL_RESOURCE_NODE *
-RsDoVendorLargeDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
-ASL_RESOURCE_NODE *
-RsDoGeneralRegisterDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset);
-
 #endif /*  __ASLCOMPILER_H */
 
diff --git a/compiler/asldefine.h b/compiler/asldefine.h
index 75345a5d7f4..cff4ecf410e 100644
--- a/compiler/asldefine.h
+++ b/compiler/asldefine.h
@@ -128,7 +128,7 @@
 #define CompilerId                  "ASL Optimizing Compiler"
 #define DisassemblerId              "AML Disassembler"
 #define CompilerCopyright           "Copyright (c) 2000 - 2010 Intel Corporation"
-#define CompilerCompliance          "Supports ACPI Specification Revision 4.0"
+#define CompilerCompliance          "Supports ACPI Specification Revision 4.0a"
 #define CompilerName                "iasl"
 #define CompilerCreatorId           "INTL"
 
diff --git a/compiler/aslfiles.c b/compiler/aslfiles.c
index ebf59a408e8..8f498a1b9df 100644
--- a/compiler/aslfiles.c
+++ b/compiler/aslfiles.c
@@ -234,6 +234,36 @@ FlOpenFile (
 }
 
 
+/*******************************************************************************
+ *
+ * FUNCTION:    FlGetFileSize
+ *
+ * PARAMETERS:  FileId              - Index into file info array
+ *
+ * RETURN:      File Size
+ *
+ * DESCRIPTION: Get current file size. Uses seek-to-EOF. File must be open.
+ *
+ ******************************************************************************/
+
+UINT32
+FlGetFileSize (
+    UINT32                  FileId)
+{
+    FILE                    *fp;
+    UINT32                  FileSize;
+
+
+    fp = Gbl_Files[FileId].Handle;
+
+    fseek (fp, 0, SEEK_END);
+    FileSize = (UINT32) ftell (fp);
+    fseek (fp, 0, SEEK_SET);
+
+    return (FileSize);
+}
+
+
 /*******************************************************************************
  *
  * FUNCTION:    FlReadFile
diff --git a/compiler/aslglobal.h b/compiler/aslglobal.h
index 01ab931e0d0..85bb3e2d9e6 100644
--- a/compiler/aslglobal.h
+++ b/compiler/aslglobal.h
@@ -188,6 +188,7 @@ ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_CompileTimesFlag, FALSE
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_FoldConstants, TRUE);
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_VerboseErrors, TRUE);
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_NoErrors, FALSE);
+ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_NoResourceChecking, FALSE);
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_DisasmFlag, FALSE);
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_GetAllTables, FALSE);
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_IntegerOptimizationFlag, TRUE);
@@ -200,6 +201,8 @@ ASL_EXTERN UINT8                    ASL_INIT_GLOBAL (Gbl_WarningLevel, ASL_WARNI
 #define HEX_OUTPUT_NONE         0
 #define HEX_OUTPUT_C            1
 #define HEX_OUTPUT_ASM          2
+#define HEX_OUTPUT_ASL          3
+
 ASL_EXTERN BOOLEAN                  ASL_INIT_GLOBAL (Gbl_HexOutputFlag, HEX_OUTPUT_NONE);
 
 
diff --git a/compiler/asllisting.c b/compiler/asllisting.c
index c4253c2e115..93f83ad8427 100644
--- a/compiler/asllisting.c
+++ b/compiler/asllisting.c
@@ -198,6 +198,10 @@ static void
 LsDoHexOutputAsm (
     void);
 
+static void
+LsDoHexOutputAsl (
+    void);
+
 ACPI_STATUS
 LsTreeWriteWalk (
     ACPI_PARSE_OBJECT       *Op,
@@ -1337,6 +1341,11 @@ LsDoHexOutput (
         LsDoHexOutputAsm ();
         break;
 
+    case HEX_OUTPUT_ASL:
+
+        LsDoHexOutputAsl ();
+        break;
+
     default:
         /* No other output types supported */
         break;
@@ -1362,60 +1371,160 @@ static void
 LsDoHexOutputC (
     void)
 {
-    UINT32                  j;
-    UINT8                   FileByte[HEX_TABLE_LINE_SIZE];
-    UINT8                   Buffer[4];
+    UINT8                   FileData[HEX_TABLE_LINE_SIZE];
+    UINT32                  LineLength;
     UINT32                  Offset = 0;
+    UINT32                  AmlFileSize;
+    UINT32                  i;
 
 
-    FlPrintFile (ASL_FILE_HEX_OUTPUT, " * C source code output\n *\n */\n");
+    /* Get AML size, seek back to start */
+
+    AmlFileSize = FlGetFileSize (ASL_FILE_AML_OUTPUT);
+
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, " * C source code output\n");
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, " * AML code block contains 0x%X bytes\n *\n */\n",
+        AmlFileSize);
     FlPrintFile (ASL_FILE_HEX_OUTPUT, "unsigned char AmlCode[] =\n{\n");
 
-    /* Start at the beginning of the AML file */
-
-    FlSeekFile (ASL_FILE_AML_OUTPUT, 0);
-
-    /* Process all AML bytes in the AML file */
-
-    j = 0;
-    while (FlReadFile (ASL_FILE_AML_OUTPUT, &FileByte[j], 1) == AE_OK)
+    while (Offset < AmlFileSize)
     {
-        if (j == 0)
+        /* Read enough bytes needed for one output line */
+
+        LineLength = fread (FileData, 1, HEX_TABLE_LINE_SIZE,
+                        Gbl_Files[ASL_FILE_AML_OUTPUT].Handle);
+        if (!LineLength)
         {
-            FlPrintFile (ASL_FILE_HEX_OUTPUT, "    ");
+            break;
         }
 
-        /* Convert each AML byte to hex */
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "    ");
 
-        UtConvertByteToHex (FileByte[j], Buffer);
-        FlWriteFile (ASL_FILE_HEX_OUTPUT, Buffer, 4);
-        FlPrintFile (ASL_FILE_HEX_OUTPUT, ",");
-
-        /* An occasional linefeed improves readability */
-
-        Offset++;
-        j++;
-
-        if (j >= HEX_TABLE_LINE_SIZE)
+        for (i = 0; i < LineLength; i++)
         {
-            /* End of line, emit the ascii dump of the entire line */
-
-            FlPrintFile (ASL_FILE_HEX_OUTPUT,
-                "  /* %8.8X", Offset - HEX_TABLE_LINE_SIZE);
-
-            /* Write the ASCII character associated with each of the bytes */
-
-            LsDumpAsciiInComment (ASL_FILE_HEX_OUTPUT,
-                HEX_TABLE_LINE_SIZE, FileByte);
-            FlPrintFile (ASL_FILE_HEX_OUTPUT, " */\n");
-
-            /* Start new line */
-
-            j = 0;
+            /*
+             * Print each hex byte.
+             * Add a comma until the very last byte of the AML file
+             * (Some C compilers complain about a trailing comma)
+             */
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "0x%2.2X", FileData[i]);
+            if ((Offset + i + 1) < AmlFileSize)
+            {
+                FlPrintFile (ASL_FILE_HEX_OUTPUT, ",");
+            }
+            else
+            {
+                FlPrintFile (ASL_FILE_HEX_OUTPUT, " ");
+            }
         }
+
+        /* Add fill spaces if needed for last line */
+
+        if (LineLength < HEX_TABLE_LINE_SIZE)
+        {
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "%*s",
+                5 * (HEX_TABLE_LINE_SIZE - LineLength), " ");
+        }
+
+        /* Emit the offset and ascii dump for the entire line */
+
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "  /* %8.8X", Offset);
+        LsDumpAsciiInComment (ASL_FILE_HEX_OUTPUT, LineLength, FileData);
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "%*s*/\n",
+            HEX_TABLE_LINE_SIZE - LineLength + 1, " ");
+
+        Offset += LineLength;
     }
 
-    FlPrintFile (ASL_FILE_HEX_OUTPUT, "\n};\n");
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, "};\n");
+    FlCloseFile (ASL_FILE_HEX_OUTPUT);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    LsDoHexOutputAsl
+ *
+ * PARAMETERS:  None
+ *
+ * RETURN:      None.
+ *
+ * DESCRIPTION: Create the hex output file.  This is the same data as the AML
+ *              output file, but formatted into hex/ascii bytes suitable for
+ *              inclusion into a C source file.
+ *
+ ******************************************************************************/
+
+static void
+LsDoHexOutputAsl (
+    void)
+{
+    UINT8                   FileData[HEX_TABLE_LINE_SIZE];
+    UINT32                  LineLength;
+    UINT32                  Offset = 0;
+    UINT32                  AmlFileSize;
+    UINT32                  i;
+
+
+    /* Get AML size, seek back to start */
+
+    AmlFileSize = FlGetFileSize (ASL_FILE_AML_OUTPUT);
+
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, " * ASL source code output\n");
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, " * AML code block contains 0x%X bytes\n *\n */\n",
+        AmlFileSize);
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, "    Name (BUF1, Buffer()\n    {\n");
+
+    while (Offset < AmlFileSize)
+    {
+        /* Read enough bytes needed for one output line */
+
+        LineLength = fread (FileData, 1, HEX_TABLE_LINE_SIZE,
+                        Gbl_Files[ASL_FILE_AML_OUTPUT].Handle);
+        if (!LineLength)
+        {
+            break;
+        }
+
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "        ");
+
+        for (i = 0; i < LineLength; i++)
+        {
+            /*
+             * Print each hex byte.
+             * Add a comma until the very last byte of the AML file
+             * (Some C compilers complain about a trailing comma)
+             */
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "0x%2.2X", FileData[i]);
+            if ((Offset + i + 1) < AmlFileSize)
+            {
+                FlPrintFile (ASL_FILE_HEX_OUTPUT, ",");
+            }
+            else
+            {
+                FlPrintFile (ASL_FILE_HEX_OUTPUT, " ");
+            }
+        }
+
+        /* Add fill spaces if needed for last line */
+
+        if (LineLength < HEX_TABLE_LINE_SIZE)
+        {
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "%*s",
+                5 * (HEX_TABLE_LINE_SIZE - LineLength), " ");
+        }
+
+        /* Emit the offset and ascii dump for the entire line */
+
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "  /* %8.8X", Offset);
+        LsDumpAsciiInComment (ASL_FILE_HEX_OUTPUT, LineLength, FileData);
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "%*s*/\n",
+            HEX_TABLE_LINE_SIZE - LineLength + 1, " ");
+
+        Offset += LineLength;
+    }
+
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, "    })\n");
     FlCloseFile (ASL_FILE_HEX_OUTPUT);
 }
 
@@ -1438,58 +1547,64 @@ static void
 LsDoHexOutputAsm (
     void)
 {
-    UINT32                  j;
-    UINT8                   FileByte[HEX_TABLE_LINE_SIZE];
-    UINT8                   Buffer[4];
+    UINT8                   FileData[HEX_TABLE_LINE_SIZE];
+    UINT32                  LineLength;
     UINT32                  Offset = 0;
-    BOOLEAN                 DoComma = FALSE;
+    UINT32                  AmlFileSize;
+    UINT32                  i;
 
 
-    FlPrintFile (ASL_FILE_HEX_OUTPUT, "; Assembly code source output\n;\n");
+    /* Get AML size, seek back to start */
 
-    /* Start at the beginning of the AML file */
+    AmlFileSize = FlGetFileSize (ASL_FILE_AML_OUTPUT);
 
-    FlSeekFile (ASL_FILE_AML_OUTPUT, 0);
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, "; Assembly code source output\n");
+    FlPrintFile (ASL_FILE_HEX_OUTPUT, "; AML code block contains 0x%X bytes\n;\n",
+        AmlFileSize);
 
-    /* Process all AML bytes in the AML file */
-
-    j = 0;
-    while (FlReadFile (ASL_FILE_AML_OUTPUT, &FileByte[j], 1) == AE_OK)
+    while (Offset < AmlFileSize)
     {
-        if (j == 0)
+        /* Read enough bytes needed for one output line */
+
+        LineLength = fread (FileData, 1, HEX_TABLE_LINE_SIZE,
+                        Gbl_Files[ASL_FILE_AML_OUTPUT].Handle);
+        if (!LineLength)
         {
-            FlPrintFile (ASL_FILE_HEX_OUTPUT, "  db  ");
-        }
-        else if (DoComma)
-        {
-            FlPrintFile (ASL_FILE_HEX_OUTPUT, ",");
-            DoComma = FALSE;
+            break;
         }
 
-        /* Convert each AML byte to hex */
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "  db  ");
 
-        UtConvertByteToAsmHex (FileByte[j], Buffer);
-        FlWriteFile (ASL_FILE_HEX_OUTPUT, Buffer, 4);
-
-        /* An occasional linefeed improves readability */
-
-        Offset++;
-        j++;
-        if (j >= HEX_TABLE_LINE_SIZE)
+        for (i = 0; i < LineLength; i++)
         {
-            FlPrintFile (ASL_FILE_HEX_OUTPUT,
-                "  ;%8.8X", Offset - HEX_TABLE_LINE_SIZE);
-
-            /* Write the ASCII character associated with each of the bytes */
-
-            LsDumpAscii (ASL_FILE_HEX_OUTPUT, HEX_TABLE_LINE_SIZE, FileByte);
-            FlPrintFile (ASL_FILE_HEX_OUTPUT, "\n");
-            j = 0;
+            /*
+             * Print each hex byte.
+             * Add a comma until the last byte of the line
+             */
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "0%2.2Xh", FileData[i]);
+            if ((i + 1) < LineLength)
+            {
+                FlPrintFile (ASL_FILE_HEX_OUTPUT, ",");
+            }
         }
-        else
+
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, " ");
+
+        /* Add fill spaces if needed for last line */
+
+        if (LineLength < HEX_TABLE_LINE_SIZE)
         {
-            DoComma = TRUE;
+            FlPrintFile (ASL_FILE_HEX_OUTPUT, "%*s",
+                5 * (HEX_TABLE_LINE_SIZE - LineLength), " ");
         }
+
+        /* Emit the offset and ascii dump for the entire line */
+
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "  ; %8.8X", Offset);
+        LsDumpAsciiInComment (ASL_FILE_HEX_OUTPUT, LineLength, FileData);
+        FlPrintFile (ASL_FILE_HEX_OUTPUT, "\n");
+
+        Offset += LineLength;
     }
 
     FlPrintFile (ASL_FILE_HEX_OUTPUT, "\n");
diff --git a/compiler/aslmain.c b/compiler/aslmain.c
index 59b5880e76d..42127c09618 100644
--- a/compiler/aslmain.c
+++ b/compiler/aslmain.c
@@ -167,7 +167,7 @@ AslDoResponseFile (
 
 
 #define ASL_TOKEN_SEPARATORS    " \t\n"
-#define ASL_SUPPORTED_OPTIONS   "@:2b:cd^e:fgh^i^I:l^o:p:r:s:t:v:w:x:"
+#define ASL_SUPPORTED_OPTIONS   "@:2b:c:d^e:fgh^i^I:l^no:p:r:s:t:v:w:x:"
 
 
 /*******************************************************************************
@@ -203,13 +203,14 @@ Options (
     printf ("\nAML Output Files:\n");
     printf ("  -s<a|c>        Create AML in assembler or C source file (*.asm or *.c)\n");
     printf ("  -i<a|c>        Create assembler or C include file (*.inc or *.h)\n");
-    printf ("  -t<a|c>        Create AML in assembler or C hex table (*.hex)\n");
+    printf ("  -t<a|c|s>      Create AML in assembler, C, or ASL hex table (*.hex)\n");
 
     printf ("\nAML Code Generation:\n");
     printf ("  -oa            Disable all optimizations (compatibility mode)\n");
     printf ("  -of            Disable constant folding\n");
     printf ("  -oi            Disable integer optimization to Zero/One/Ones\n");
     printf ("  -on            Disable named reference string optimization\n");
+    printf ("  -cr            Disable Resource Descriptor error checking\n");
     printf ("  -r<Revision>   Override table header Revision (1-255)\n");
 
     printf ("\nListings:\n");
@@ -264,7 +265,7 @@ HelpMessage (
     printf ("  -b<p|t|b>      Create compiler debug/trace file (*.txt)\n");
     printf ("                   Types: Parse/Tree/Both\n");
     printf ("  -f             Ignore errors, force creation of AML output file(s)\n");
-    printf ("  -c             Parse only, no output generation\n");
+    printf ("  -n             Parse only, no output generation\n");
     printf ("  -ot            Display compile times\n");
     printf ("  -x<level>      Set debug level for trace output\n");
 }
@@ -507,10 +508,16 @@ AslDoOptions (
 
 
     case 'c':
+        switch (AcpiGbl_Optarg[0])
+        {
+        case 'r':
+            Gbl_NoResourceChecking = TRUE;
+            break;
 
-        /* Parse only */
-
-        Gbl_ParseOnlyFlag = TRUE;
+        default:
+            printf ("Unknown option: -c%s\n", AcpiGbl_Optarg);
+            return (-1);
+        }
         break;
 
 
@@ -688,6 +695,14 @@ AslDoOptions (
         break;
 
 
+    case 'n':
+
+        /* Parse only */
+
+        Gbl_ParseOnlyFlag = TRUE;
+        break;
+
+
     case 'p':
 
         /* Override default AML output filename */
@@ -741,6 +756,10 @@ AslDoOptions (
             Gbl_HexOutputFlag = HEX_OUTPUT_C;
             break;
 
+        case 's':
+            Gbl_HexOutputFlag = HEX_OUTPUT_ASL;
+            break;
+
         default:
             printf ("Unknown option: -t%s\n", AcpiGbl_Optarg);
             return (-1);
diff --git a/compiler/aslpredef.c b/compiler/aslpredef.c
index 6346bafd10e..ae1b7738148 100644
--- a/compiler/aslpredef.c
+++ b/compiler/aslpredef.c
@@ -243,11 +243,11 @@ ApCheckForPredefinedMethod (
         break;
 
 
-    case ACPI_EVENT_RESERVED_NAME:      /* _Lxx, _Exx, and _Qxx methods */
+    case ACPI_EVENT_RESERVED_NAME:      /* _Lxx/_Exx/_Wxx/_Qxx methods */
 
         Gbl_ReservedMethods++;
 
-        /* NumArguments must be zero for all _Lxx, _Exx, and _Qxx methods */
+        /* NumArguments must be zero for all _Lxx/_Exx/_Wxx/_Qxx methods */
 
         if (MethodInfo->NumArguments != 0)
         {
@@ -346,12 +346,12 @@ ApCheckPredefinedReturnValue (
     case ACPI_NOT_RESERVED_NAME:        /* No underscore or _Txx or _xxx name not matched */
     case ACPI_PREDEFINED_NAME:          /* Resource Name or reserved scope name */
     case ACPI_COMPILER_RESERVED_NAME:   /* A _Txx that was not emitted by compiler */
-    case ACPI_EVENT_RESERVED_NAME:      /* _Lxx, _Exx, and _Qxx methods */
+    case ACPI_EVENT_RESERVED_NAME:      /* _Lxx/_Exx/_Wxx/_Qxx methods */
 
         /* Just return, nothing to do */
         return;
 
-    default: /* a real predefined ACPI name */
+    default: /* A standard predefined ACPI name */
 
         /* Exit if no return value expected */
 
@@ -425,29 +425,59 @@ ApCheckForPredefinedObject (
      * or a predefined scope name
      */
     Index = ApCheckForPredefinedName (Op, Name);
-    if (Index > ACPI_VALID_RESERVED_NAME_MAX)
-    {
-        return;
-    }
 
-    /*
-     * We found a matching predefind name.
-     * Check if this predefined name requires input arguments
-     */
-    if (PredefinedNames[Index].Info.ParamCount > 0)
+    switch (Index)
     {
+    case ACPI_NOT_RESERVED_NAME:        /* No underscore or _Txx or _xxx name not matched */
+    case ACPI_PREDEFINED_NAME:          /* Resource Name or reserved scope name */
+    case ACPI_COMPILER_RESERVED_NAME:   /* A _Txx that was not emitted by compiler */
+
+        /* Nothing to do */
+        return;
+
+    case ACPI_EVENT_RESERVED_NAME:      /* _Lxx/_Exx/_Wxx/_Qxx methods */
+
         /*
-         * This predefined name must always be defined as a control
-         * method because it is required to have input arguments.
+         * These names must be control methods, by definition in ACPI spec.
+         * Also because they are defined to return no value. None of them
+         * require any arguments.
          */
         AslError (ASL_ERROR, ASL_MSG_RESERVED_METHOD, Op,
-            "with arguments");
+            "with zero arguments");
+        return;
+
+    default: /* A standard predefined ACPI name */
+
+        /*
+         * If this predefined name requires input arguments, then
+         * it must be implemented as a control method
+         */
+        if (PredefinedNames[Index].Info.ParamCount > 0)
+        {
+            AslError (ASL_ERROR, ASL_MSG_RESERVED_METHOD, Op,
+                "with arguments");
+            return;
+        }
+
+        /*
+         * If no return value is expected from this predefined name, then
+         * it follows that it must be implemented as a control method
+         * (with zero args, because the args > 0 case was handled above)
+         * Examples are: _DIS, _INI, _IRC, _OFF, _ON, _PSx
+         */
+        if (!PredefinedNames[Index].Info.ExpectedBtypes)
+        {
+            AslError (ASL_ERROR, ASL_MSG_RESERVED_METHOD, Op,
+                "with zero arguments");
+            return;
+        }
+
+        /* Typecheck the actual object, it is the next argument */
+
+        ApCheckObjectType (Op->Asl.Child->Asl.Next,
+            PredefinedNames[Index].Info.ExpectedBtypes);
+        return;
     }
-
-    /* Typecheck the actual object, it is the next argument */
-
-    ApCheckObjectType (Op->Asl.Child->Asl.Next,
-        PredefinedNames[Index].Info.ExpectedBtypes);
 }
 
 
@@ -514,7 +544,7 @@ ApCheckForPredefinedName (
         }
     }
 
-    /* Check for _Lxx, _Exx, _Qxx, _T_x. Warning if unknown predefined name */
+    /* Check for _Lxx/_Exx/_Wxx/_Qxx/_T_x. Warning if unknown predefined name */
 
     return (ApCheckForSpecialName (Op, Name));
 }
@@ -530,7 +560,7 @@ ApCheckForPredefinedName (
  * RETURN:      None
  *
  * DESCRIPTION: Check for the "special" predefined names -
- *              _Lxx, _Exx, _Qxx, and _T_x
+ *              _Lxx, _Exx, _Qxx, _Wxx, and _T_x
  *
  ******************************************************************************/
 
@@ -541,14 +571,16 @@ ApCheckForSpecialName (
 {
 
     /*
-     * Check for the "special" predefined names. We know the first char is an
-     * underscore already.
+     * Check for the "special" predefined names. We already know that the
+     * first character is an underscore.
      *   GPE:  _Lxx
      *   GPE:  _Exx
+     *   GPE:  _Wxx
      *   EC:   _Qxx
      */
     if ((Name[1] == 'L') ||
         (Name[1] == 'E') ||
+        (Name[1] == 'W') ||
         (Name[1] == 'Q'))
     {
         /* The next two characters must be hex digits */
diff --git a/compiler/aslresource.c b/compiler/aslresource.c
index 421fde3a367..67aabcadae1 100644
--- a/compiler/aslresource.c
+++ b/compiler/aslresource.c
@@ -1,7 +1,7 @@
 
 /******************************************************************************
  *
- * Module Name: aslresource - Resource templates and descriptors
+ * Module Name: aslresource - Resource template/descriptor utilities
  *
  *****************************************************************************/
 
@@ -124,6 +124,298 @@
         ACPI_MODULE_NAME    ("aslresource")
 
 
+/*******************************************************************************
+ *
+ * FUNCTION:    RsSmallAddressCheck
+ *
+ * PARAMETERS:  Minimum             - Address Min value
+ *              Maximum             - Address Max value
+ *              Length              - Address range value
+ *              Alignment           - Address alignment value
+ *              MinOp               - Original Op for Address Min
+ *              MaxOp               - Original Op for Address Max
+ *              LengthOp            - Original Op for address range
+ *              AlignOp             - Original Op for address alignment. If
+ *                                    NULL, means "zero value for alignment is
+ *                                    OK, and means 64K alignment" (for
+ *                                    Memory24 descriptor)
+ *
+ * RETURN:      None. Adds error messages to error log if necessary
+ *
+ * DESCRIPTION: Perform common value checks for "small" address descriptors.
+ *              Currently:
+ *                  Io, Memory24, Memory32
+ *
+ ******************************************************************************/
+
+void
+RsSmallAddressCheck (
+    UINT8                   Type,
+    UINT32                  Minimum,
+    UINT32                  Maximum,
+    UINT32                  Length,
+    UINT32                  Alignment,
+    ACPI_PARSE_OBJECT       *MinOp,
+    ACPI_PARSE_OBJECT       *MaxOp,
+    ACPI_PARSE_OBJECT       *LengthOp,
+    ACPI_PARSE_OBJECT       *AlignOp)
+{
+
+    if (Gbl_NoResourceChecking)
+    {
+        return;
+    }
+
+    /* Special case for Memory24, values are compressed */
+
+    if (Type == ACPI_RESOURCE_NAME_MEMORY24)
+    {
+        if (!Alignment) /* Alignment==0 means 64K - no invalid alignment */
+        {
+            Alignment = ACPI_UINT16_MAX + 1;
+        }
+
+        Minimum <<= 8;
+        Maximum <<= 8;
+        Length *= 256;
+    }
+
+    /* IO descriptor has different definition of min/max, don't check */
+
+    if (Type != ACPI_RESOURCE_NAME_IO)
+    {
+        /* Basic checks on Min/Max/Length */
+
+        if (Minimum > Maximum)
+        {
+            AslError (ASL_ERROR, ASL_MSG_INVALID_MIN_MAX, MinOp, NULL);
+        }
+        else if (Length > (Maximum - Minimum + 1))
+        {
+            AslError (ASL_ERROR, ASL_MSG_INVALID_LENGTH, LengthOp, NULL);
+        }
+    }
+
+    /* Alignment of zero is not in ACPI spec, but is used to mean byte acc */
+
+    if (!Alignment)
+    {
+        Alignment = 1;
+    }
+
+    /* Addresses must be an exact multiple of the alignment value */
+
+    if (Minimum % Alignment)
+    {
+        AslError (ASL_ERROR, ASL_MSG_ALIGNMENT, MinOp, NULL);
+    }
+    if (Maximum % Alignment)
+    {
+        AslError (ASL_ERROR, ASL_MSG_ALIGNMENT, MaxOp, NULL);
+    }
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsLargeAddressCheck
+ *
+ * PARAMETERS:  Minimum             - Address Min value
+ *              Maximum             - Address Max value
+ *              Length              - Address range value
+ *              Granularity         - Address granularity value
+ *              Flags               - General flags for address descriptors:
+ *                                    _MIF, _MAF, _DEC
+ *              MinOp               - Original Op for Address Min
+ *              MaxOp               - Original Op for Address Max
+ *              LengthOp            - Original Op for address range
+ *              GranOp              - Original Op for address granularity
+ *
+ * RETURN:      None. Adds error messages to error log if necessary
+ *
+ * DESCRIPTION: Perform common value checks for "large" address descriptors.
+ *              Currently:
+ *                  WordIo,     WordBusNumber,  WordSpace
+ *                  DWordIo,    DWordMemory,    DWordSpace
+ *                  QWordIo,    QWordMemory,    QWordSpace
+ *                  ExtendedIo, ExtendedMemory, ExtendedSpace
+ *
+ * _MIF flag set means that the minimum address is fixed and is not relocatable
+ * _MAF flag set means that the maximum address is fixed and is not relocatable
+ * Length of zero means that the record size is variable
+ *
+ * This function implements the LEN/MIF/MAF/MIN/MAX/GRA rules within Table 6-40
+ * of the ACPI 4.0a specification. Added 04/2010.
+ *
+ ******************************************************************************/
+
+void
+RsLargeAddressCheck (
+    UINT64                  Minimum,
+    UINT64                  Maximum,
+    UINT64                  Length,
+    UINT64                  Granularity,
+    UINT8                   Flags,
+    ACPI_PARSE_OBJECT       *MinOp,
+    ACPI_PARSE_OBJECT       *MaxOp,
+    ACPI_PARSE_OBJECT       *LengthOp,
+    ACPI_PARSE_OBJECT       *GranOp)
+{
+
+    if (Gbl_NoResourceChecking)
+    {
+        return;
+    }
+
+    /* Basic checks on Min/Max/Length */
+
+    if (Minimum > Maximum)
+    {
+        AslError (ASL_ERROR, ASL_MSG_INVALID_MIN_MAX, MinOp, NULL);
+        return;
+    }
+    else if (Length > (Maximum - Minimum + 1))
+    {
+        AslError (ASL_ERROR, ASL_MSG_INVALID_LENGTH, LengthOp, NULL);
+        return;
+    }
+
+    /* If specified (non-zero), ensure granularity is a power-of-two minus one */
+
+    if (Granularity)
+    {
+        if ((Granularity + 1) &
+             Granularity)
+        {
+            AslError (ASL_ERROR, ASL_MSG_INVALID_GRANULARITY, GranOp, NULL);
+            return;
+        }
+    }
+
+    /*
+     * Check the various combinations of Length, MinFixed, and MaxFixed
+     */
+    if (Length)
+    {
+        /* Fixed non-zero length */
+
+        switch (Flags & (ACPI_RESOURCE_FLAG_MIF | ACPI_RESOURCE_FLAG_MAF))
+        {
+        case 0:
+            /*
+             * Fixed length, variable locations (both _MIN and _MAX).
+             * Length must be a multiple of granularity
+             */
+            if (Granularity & Length)
+            {
+                AslError (ASL_ERROR, ASL_MSG_ALIGNMENT, LengthOp, NULL);
+            }
+            break;
+
+        case (ACPI_RESOURCE_FLAG_MIF | ACPI_RESOURCE_FLAG_MAF):
+
+            /* Fixed length, fixed location. Granularity must be zero */
+
+            if (Granularity != 0)
+            {
+                AslError (ASL_ERROR, ASL_MSG_INVALID_GRAN_FIXED, GranOp, NULL);
+            }
+
+            /* Length must be exactly the size of the min/max window */
+
+            if (Length != (Maximum - Minimum + 1))
+            {
+                AslError (ASL_ERROR, ASL_MSG_INVALID_LENGTH_FIXED, LengthOp, NULL);
+            }
+            break;
+
+        /* All other combinations are invalid */
+
+        case ACPI_RESOURCE_FLAG_MIF:
+        case ACPI_RESOURCE_FLAG_MAF:
+        default:
+            AslError (ASL_ERROR, ASL_MSG_INVALID_ADDR_FLAGS, LengthOp, NULL);
+        }
+    }
+    else
+    {
+        /* Variable length (length==0) */
+
+        switch (Flags & (ACPI_RESOURCE_FLAG_MIF | ACPI_RESOURCE_FLAG_MAF))
+        {
+        case 0:
+            /*
+             * Both _MIN and _MAX are variable.
+             * No additional requirements, just exit
+             */
+            break;
+
+        case ACPI_RESOURCE_FLAG_MIF:
+
+            /* _MIN is fixed. _MIN must be multiple of _GRA */
+
+            /*
+             * The granularity is defined by the ACPI specification to be a
+             * power-of-two minus one, therefore the granularity is a
+             * bitmask which can be used to easily validate the addresses.
+             */
+            if (Granularity & Minimum)
+            {
+                AslError (ASL_ERROR, ASL_MSG_ALIGNMENT, MinOp, NULL);
+            }
+            break;
+
+        case ACPI_RESOURCE_FLAG_MAF:
+
+            /* _MAX is fixed. (_MAX + 1) must be multiple of _GRA */
+
+            if (Granularity & (Maximum + 1))
+            {
+                AslError (ASL_ERROR, ASL_MSG_ALIGNMENT, MaxOp, "-1");
+            }
+            break;
+
+        /* Both MIF/MAF set is invalid if length is zero */
+
+        case (ACPI_RESOURCE_FLAG_MIF | ACPI_RESOURCE_FLAG_MAF):
+        default:
+            AslError (ASL_ERROR, ASL_MSG_INVALID_ADDR_FLAGS, LengthOp, NULL);
+        }
+    }
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsGetStringDataLength
+ *
+ * PARAMETERS:  InitializerOp     - Start of a subtree of init nodes
+ *
+ * RETURN:      Valid string length if a string node is found (otherwise 0)
+ *
+ * DESCRIPTION: In a list of peer nodes, find the first one that contains a
+ *              string and return the length of the string.
+ *
+ ******************************************************************************/
+
+UINT16
+RsGetStringDataLength (
+    ACPI_PARSE_OBJECT       *InitializerOp)
+{
+
+    while (InitializerOp)
+    {
+        if (InitializerOp->Asl.ParseOpcode == PARSEOP_STRING_LITERAL)
+        {
+            return ((UINT16) (strlen (InitializerOp->Asl.Value.String) + 1));
+        }
+        InitializerOp = ASL_GET_PEER_NODE (InitializerOp);
+    }
+
+    return 0;
+}
+
+
 /*******************************************************************************
  *
  * FUNCTION:    RsAllocateResourceNode
diff --git a/compiler/aslrestype1.c b/compiler/aslrestype1.c
index 1dbff0310fd..381fb53cc25 100644
--- a/compiler/aslrestype1.c
+++ b/compiler/aslrestype1.c
@@ -1,7 +1,7 @@
 
 /******************************************************************************
  *
- * Module Name: aslrestype1 - Short (type1) resource templates and descriptors
+ * Module Name: aslrestype1 - Miscellaneous small resource descriptors
  *
  *****************************************************************************/
 
@@ -121,6 +121,18 @@
 #define _COMPONENT          ACPI_COMPILER
         ACPI_MODULE_NAME    ("aslrestype1")
 
+/*
+ * This module contains miscellaneous small resource descriptors:
+ *
+ * EndTag
+ * EndDependentFn
+ * Memory24
+ * Memory32
+ * Memory32Fixed
+ * StartDependentFn
+ * StartDependentFnNoPri
+ * VendorShort
+ */
 
 /*******************************************************************************
  *
@@ -156,127 +168,6 @@ RsDoEndTagDescriptor (
 }
 
 
-/*******************************************************************************
- *
- * FUNCTION:    RsDoDmaDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a short "DMA" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoDmaDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT32                  i;
-    UINT8                   DmaChannelMask = 0;
-    UINT8                   DmaChannels = 0;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_DMA));
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Dma.DescriptorType  = ACPI_RESOURCE_NAME_DMA |
-                                        ASL_RDESC_DMA_SIZE;
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* DMA type */
-
-            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DMATYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 5);
-            break;
-
-        case 1: /* Bus Master */
-
-            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_BUSMASTER,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 2);
-            break;
-
-        case 2: /* Xfer Type (transfer width) */
-
-            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 0, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_XFERTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 0);
-            break;
-
-        case 3: /* Name */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            /* All DMA channel bytes are handled here, after flags and name */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                /* Up to 8 channels can be specified in the list */
-
-                DmaChannels++;
-                if (DmaChannels > 8)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_DMA_LIST,
-                        InitializerOp, NULL);
-                    return (Rnode);
-                }
-
-                /* Only DMA channels 0-7 are allowed (mask is 8 bits) */
-
-                if (InitializerOp->Asl.Value.Integer > 7)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_DMA_CHANNEL,
-                        InitializerOp, NULL);
-                }
-
-                /* Build the mask */
-
-                DmaChannelMask |=
-                    (1 << ((UINT8) InitializerOp->Asl.Value.Integer));
-            }
-
-            if (i == 4) /* case 4: First DMA byte */
-            {
-                /* Check now for duplicates in list */
-
-                RsCheckListForDuplicates (InitializerOp);
-
-                /* Create a named field at the start of the list */
-
-                RsCreateByteField (InitializerOp, ACPI_RESTAG_DMA,
-                    CurrentByteOffset +
-                    ASL_RESDESC_OFFSET (Dma.DmaChannelMask));
-            }
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    /* Now we can set the channel mask */
-
-    Descriptor->Dma.DmaChannelMask = DmaChannelMask;
-    return (Rnode);
-}
-
-
 /*******************************************************************************
  *
  * FUNCTION:    RsDoEndDependentDescriptor
@@ -309,396 +200,6 @@ RsDoEndDependentDescriptor (
 }
 
 
-/*******************************************************************************
- *
- * FUNCTION:    RsDoFixedIoDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a short "FixedIO" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoFixedIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_FIXED_IO));
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Io.DescriptorType  = ACPI_RESOURCE_NAME_FIXED_IO |
-                                      ASL_RDESC_FIXED_IO_SIZE;
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Base Address */
-
-            Descriptor->FixedIo.Address =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_BASEADDRESS,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (FixedIo.Address));
-            break;
-
-        case 1: /* Length */
-
-            Descriptor->FixedIo.AddressLength =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (FixedIo.AddressLength));
-            break;
-
-        case 2: /* Name */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoIoDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a short "IO" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IO));
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Io.DescriptorType  = ACPI_RESOURCE_NAME_IO |
-                                      ASL_RDESC_IO_SIZE;
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Decode size */
-
-            RsSetFlagBits (&Descriptor->Io.Flags, InitializerOp, 0, 1);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Flags), 0);
-            break;
-
-        case 1:  /* Min Address */
-
-            Descriptor->Io.Minimum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Minimum));
-            break;
-
-        case 2: /* Max Address */
-
-            Descriptor->Io.Maximum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Maximum));
-            break;
-
-        case 3: /* Alignment */
-
-            Descriptor->Io.Alignment =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_ALIGNMENT,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Alignment));
-            break;
-
-        case 4: /* Length */
-
-            Descriptor->Io.AddressLength =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.AddressLength));
-            break;
-
-        case 5: /* Name */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoIrqDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a short "IRQ" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoIrqDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT32                  Interrupts = 0;
-    UINT16                  IrqMask = 0;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IRQ));
-
-    /* Length = 3 (with flag byte) */
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Irq.DescriptorType  = ACPI_RESOURCE_NAME_IRQ |
-                                      (ASL_RDESC_IRQ_SIZE + 0x01);
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Interrupt Type (or Mode - edge/level) */
-
-            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 0, 1);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 0);
-            break;
-
-        case 1: /* Interrupt Level (or Polarity - Active high/low) */
-
-            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTLEVEL,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 3);
-            break;
-
-        case 2: /* Share Type - Default: exclusive (0) */
-
-            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 4, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTSHARE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 4);
-            break;
-
-        case 3: /* Name */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            /* All IRQ bytes are handled here, after the flags and name */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                /* Up to 16 interrupts can be specified in the list */
-
-                Interrupts++;
-                if (Interrupts > 16)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_LIST,
-                        InitializerOp, NULL);
-                    return (Rnode);
-                }
-
-                /* Only interrupts 0-15 are allowed (mask is 16 bits) */
-
-                if (InitializerOp->Asl.Value.Integer > 15)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_NUMBER,
-                        InitializerOp, NULL);
-                }
-                else
-                {
-                    IrqMask |= (1 << (UINT8) InitializerOp->Asl.Value.Integer);
-                }
-            }
-
-            /* Case 4: First IRQ value in list */
-
-            if (i == 4)
-            {
-                /* Check now for duplicates in list */
-
-                RsCheckListForDuplicates (InitializerOp);
-
-                /* Create a named field at the start of the list */
-
-                RsCreateByteField (InitializerOp, ACPI_RESTAG_INTERRUPT,
-                    CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.IrqMask));
-            }
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    /* Now we can set the channel mask */
-
-    Descriptor->Irq.IrqMask = IrqMask;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoIrqNoFlagsDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a short "IRQNoFlags" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoIrqNoFlagsDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT16                  IrqMask = 0;
-    UINT32                  Interrupts = 0;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IRQ_NOFLAGS));
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Irq.DescriptorType  = ACPI_RESOURCE_NAME_IRQ |
-                                      ASL_RDESC_IRQ_SIZE;
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Name */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            /* IRQ bytes are handled here, after the flags and name */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                /* Up to 16 interrupts can be specified in the list */
-
-                Interrupts++;
-                if (Interrupts > 16)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_LIST,
-                        InitializerOp, NULL);
-                    return (Rnode);
-                }
-
-                /* Only interrupts 0-15 are allowed (mask is 16 bits) */
-
-                if (InitializerOp->Asl.Value.Integer > 15)
-                {
-                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_NUMBER,
-                        InitializerOp, NULL);
-                }
-                else
-                {
-                    IrqMask |= (1 << ((UINT8) InitializerOp->Asl.Value.Integer));
-                }
-            }
-
-            /* Case 1: First IRQ value in list */
-
-            if (i == 1)
-            {
-                /* Check now for duplicates in list */
-
-                RsCheckListForDuplicates (InitializerOp);
-
-                /* Create a named field at the start of the list */
-
-                RsCreateByteField (InitializerOp, ACPI_RESTAG_INTERRUPT,
-                    CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.IrqMask));
-            }
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    /* Now we can set the interrupt mask */
-
-    Descriptor->Irq.IrqMask = IrqMask;
-    return (Rnode);
-}
-
-
 /*******************************************************************************
  *
  * FUNCTION:    RsDoMemory24Descriptor
@@ -720,6 +221,10 @@ RsDoMemory24Descriptor (
 {
     AML_RESOURCE            *Descriptor;
     ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *AlignOp = NULL;
     ASL_RESOURCE_NODE       *Rnode;
     UINT32                  i;
 
@@ -749,6 +254,7 @@ RsDoMemory24Descriptor (
             Descriptor->Memory24.Minimum = (UINT16) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory24.Minimum));
+            MinOp = InitializerOp;
             break;
 
         case 2: /* Max Address */
@@ -756,6 +262,7 @@ RsDoMemory24Descriptor (
             Descriptor->Memory24.Maximum = (UINT16) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory24.Maximum));
+            MaxOp = InitializerOp;
             break;
 
         case 3: /* Alignment */
@@ -763,6 +270,7 @@ RsDoMemory24Descriptor (
             Descriptor->Memory24.Alignment = (UINT16) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_ALIGNMENT,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory24.Alignment));
+            AlignOp = InitializerOp;
             break;
 
         case 4: /* Length */
@@ -770,6 +278,7 @@ RsDoMemory24Descriptor (
             Descriptor->Memory24.AddressLength = (UINT16) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory24.AddressLength));
+            LengthOp = InitializerOp;
             break;
 
         case 5: /* Name */
@@ -786,6 +295,15 @@ RsDoMemory24Descriptor (
         InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
     }
 
+    /* Validate the Min/Max/Len/Align values (Alignment==0 means 64K) */
+
+    RsSmallAddressCheck (ACPI_RESOURCE_NAME_MEMORY24,
+        Descriptor->Memory24.Minimum,
+        Descriptor->Memory24.Maximum,
+        Descriptor->Memory24.AddressLength,
+        Descriptor->Memory24.Alignment,
+        MinOp, MaxOp, LengthOp, NULL);
+
     return (Rnode);
 }
 
@@ -811,6 +329,10 @@ RsDoMemory32Descriptor (
 {
     AML_RESOURCE            *Descriptor;
     ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *AlignOp = NULL;
     ASL_RESOURCE_NODE       *Rnode;
     UINT32                  i;
 
@@ -840,6 +362,7 @@ RsDoMemory32Descriptor (
             Descriptor->Memory32.Minimum = (UINT32) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory32.Minimum));
+            MinOp = InitializerOp;
             break;
 
         case 2: /* Max Address */
@@ -847,6 +370,7 @@ RsDoMemory32Descriptor (
             Descriptor->Memory32.Maximum = (UINT32) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory32.Maximum));
+            MaxOp = InitializerOp;
             break;
 
         case 3: /* Alignment */
@@ -854,6 +378,7 @@ RsDoMemory32Descriptor (
             Descriptor->Memory32.Alignment = (UINT32) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_ALIGNMENT,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory32.Alignment));
+            AlignOp = InitializerOp;
             break;
 
         case 4: /* Length */
@@ -861,6 +386,7 @@ RsDoMemory32Descriptor (
             Descriptor->Memory32.AddressLength = (UINT32) InitializerOp->Asl.Value.Integer;
             RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
                 CurrentByteOffset + ASL_RESDESC_OFFSET (Memory32.AddressLength));
+            LengthOp = InitializerOp;
             break;
 
         case 5: /* Name */
@@ -877,6 +403,15 @@ RsDoMemory32Descriptor (
         InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
     }
 
+    /* Validate the Min/Max/Len/Align values */
+
+    RsSmallAddressCheck (ACPI_RESOURCE_NAME_MEMORY32,
+        Descriptor->Memory32.Minimum,
+        Descriptor->Memory32.Maximum,
+        Descriptor->Memory32.AddressLength,
+        Descriptor->Memory32.Alignment,
+        MinOp, MaxOp, LengthOp, AlignOp);
+
     return (Rnode);
 }
 
@@ -1027,6 +562,7 @@ RsDoStartDependentDescriptor (
             break;
 
         default:
+
             NextRnode = RsDoOneResourceDescriptor  (InitializerOp,
                         CurrentByteOffset, &State);
 
@@ -1036,7 +572,6 @@ RsDoStartDependentDescriptor (
              * must keep track of the offset of not only each descriptor, but each
              * element (field) within each descriptor as well.
              */
-
             CurrentByteOffset += RsLinkDescriptorChain (&PreviousRnode,
                                     NextRnode);
             break;
@@ -1182,4 +717,3 @@ RsDoVendorSmallDescriptor (
     return (Rnode);
 }
 
-
diff --git a/compiler/aslrestype1i.c b/compiler/aslrestype1i.c
new file mode 100644
index 00000000000..239c500cb6f
--- /dev/null
+++ b/compiler/aslrestype1i.c
@@ -0,0 +1,668 @@
+
+/******************************************************************************
+ *
+ * Module Name: aslrestype1i - Small I/O-related resource descriptors
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "aslcompiler.h"
+#include "aslcompiler.y.h"
+
+#define _COMPONENT          ACPI_COMPILER
+        ACPI_MODULE_NAME    ("aslrestype1i")
+
+/*
+ * This module contains the I/O-related small resource descriptors:
+ *
+ * DMA
+ * FixedIO
+ * IO
+ * IRQ
+ * IRQNoFlags
+ */
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoDmaDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a short "DMA" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoDmaDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT32                  i;
+    UINT8                   DmaChannelMask = 0;
+    UINT8                   DmaChannels = 0;
+
+
+    InitializerOp = Op->Asl.Child;
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_DMA));
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Dma.DescriptorType  = ACPI_RESOURCE_NAME_DMA |
+                                        ASL_RDESC_DMA_SIZE;
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* DMA type */
+
+            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DMATYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 5);
+            break;
+
+        case 1: /* Bus Master */
+
+            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_BUSMASTER,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 2);
+            break;
+
+        case 2: /* Xfer Type (transfer width) */
+
+            RsSetFlagBits (&Descriptor->Dma.Flags, InitializerOp, 0, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_XFERTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Dma.Flags), 0);
+            break;
+
+        case 3: /* Name */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            /* All DMA channel bytes are handled here, after flags and name */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                /* Up to 8 channels can be specified in the list */
+
+                DmaChannels++;
+                if (DmaChannels > 8)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_DMA_LIST,
+                        InitializerOp, NULL);
+                    return (Rnode);
+                }
+
+                /* Only DMA channels 0-7 are allowed (mask is 8 bits) */
+
+                if (InitializerOp->Asl.Value.Integer > 7)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_DMA_CHANNEL,
+                        InitializerOp, NULL);
+                }
+
+                /* Build the mask */
+
+                DmaChannelMask |=
+                    (1 << ((UINT8) InitializerOp->Asl.Value.Integer));
+            }
+
+            if (i == 4) /* case 4: First DMA byte */
+            {
+                /* Check now for duplicates in list */
+
+                RsCheckListForDuplicates (InitializerOp);
+
+                /* Create a named field at the start of the list */
+
+                RsCreateByteField (InitializerOp, ACPI_RESTAG_DMA,
+                    CurrentByteOffset +
+                    ASL_RESDESC_OFFSET (Dma.DmaChannelMask));
+            }
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Now we can set the channel mask */
+
+    Descriptor->Dma.DmaChannelMask = DmaChannelMask;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoFixedIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a short "FixedIO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoFixedIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *AddressOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_FIXED_IO));
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Io.DescriptorType  = ACPI_RESOURCE_NAME_FIXED_IO |
+                                      ASL_RDESC_FIXED_IO_SIZE;
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Base Address */
+
+            Descriptor->FixedIo.Address =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_BASEADDRESS,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (FixedIo.Address));
+            AddressOp = InitializerOp;
+            break;
+
+        case 1: /* Length */
+
+            Descriptor->FixedIo.AddressLength =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (FixedIo.AddressLength));
+            break;
+
+        case 2: /* Name */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Error checks */
+
+    if (Descriptor->FixedIo.Address > 0x03FF)
+    {
+        AslError (ASL_WARNING, ASL_MSG_ISA_ADDRESS, AddressOp, NULL);
+    }
+
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a short "IO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *AlignOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IO));
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Io.DescriptorType  = ACPI_RESOURCE_NAME_IO |
+                                      ASL_RDESC_IO_SIZE;
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Decode size */
+
+            RsSetFlagBits (&Descriptor->Io.Flags, InitializerOp, 0, 1);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Flags), 0);
+            break;
+
+        case 1:  /* Min Address */
+
+            Descriptor->Io.Minimum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 2: /* Max Address */
+
+            Descriptor->Io.Maximum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 3: /* Alignment */
+
+            Descriptor->Io.Alignment =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_ALIGNMENT,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.Alignment));
+            AlignOp = InitializerOp;
+            break;
+
+        case 4: /* Length */
+
+            Descriptor->Io.AddressLength =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Io.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 5: /* Name */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Align values */
+
+    RsSmallAddressCheck (ACPI_RESOURCE_NAME_IO,
+        Descriptor->Io.Minimum,
+        Descriptor->Io.Maximum,
+        Descriptor->Io.AddressLength,
+        Descriptor->Io.Alignment,
+        MinOp, MaxOp, LengthOp, AlignOp);
+
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoIrqDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a short "IRQ" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoIrqDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT32                  Interrupts = 0;
+    UINT16                  IrqMask = 0;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IRQ));
+
+    /* Length = 3 (with flag byte) */
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Irq.DescriptorType  = ACPI_RESOURCE_NAME_IRQ |
+                                      (ASL_RDESC_IRQ_SIZE + 0x01);
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Interrupt Type (or Mode - edge/level) */
+
+            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 0, 1);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 0);
+            break;
+
+        case 1: /* Interrupt Level (or Polarity - Active high/low) */
+
+            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTLEVEL,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 3);
+            break;
+
+        case 2: /* Share Type - Default: exclusive (0) */
+
+            RsSetFlagBits (&Descriptor->Irq.Flags, InitializerOp, 4, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_INTERRUPTSHARE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.Flags), 4);
+            break;
+
+        case 3: /* Name */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            /* All IRQ bytes are handled here, after the flags and name */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                /* Up to 16 interrupts can be specified in the list */
+
+                Interrupts++;
+                if (Interrupts > 16)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_LIST,
+                        InitializerOp, NULL);
+                    return (Rnode);
+                }
+
+                /* Only interrupts 0-15 are allowed (mask is 16 bits) */
+
+                if (InitializerOp->Asl.Value.Integer > 15)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_NUMBER,
+                        InitializerOp, NULL);
+                }
+                else
+                {
+                    IrqMask |= (1 << (UINT8) InitializerOp->Asl.Value.Integer);
+                }
+            }
+
+            /* Case 4: First IRQ value in list */
+
+            if (i == 4)
+            {
+                /* Check now for duplicates in list */
+
+                RsCheckListForDuplicates (InitializerOp);
+
+                /* Create a named field at the start of the list */
+
+                RsCreateByteField (InitializerOp, ACPI_RESTAG_INTERRUPT,
+                    CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.IrqMask));
+            }
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Now we can set the channel mask */
+
+    Descriptor->Irq.IrqMask = IrqMask;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoIrqNoFlagsDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a short "IRQNoFlags" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoIrqNoFlagsDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT16                  IrqMask = 0;
+    UINT32                  Interrupts = 0;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_IRQ_NOFLAGS));
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Irq.DescriptorType  = ACPI_RESOURCE_NAME_IRQ |
+                                      ASL_RDESC_IRQ_SIZE;
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Name */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            /* IRQ bytes are handled here, after the flags and name */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                /* Up to 16 interrupts can be specified in the list */
+
+                Interrupts++;
+                if (Interrupts > 16)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_LIST,
+                        InitializerOp, NULL);
+                    return (Rnode);
+                }
+
+                /* Only interrupts 0-15 are allowed (mask is 16 bits) */
+
+                if (InitializerOp->Asl.Value.Integer > 15)
+                {
+                    AslError (ASL_ERROR, ASL_MSG_INTERRUPT_NUMBER,
+                        InitializerOp, NULL);
+                }
+                else
+                {
+                    IrqMask |= (1 << ((UINT8) InitializerOp->Asl.Value.Integer));
+                }
+            }
+
+            /* Case 1: First IRQ value in list */
+
+            if (i == 1)
+            {
+                /* Check now for duplicates in list */
+
+                RsCheckListForDuplicates (InitializerOp);
+
+                /* Create a named field at the start of the list */
+
+                RsCreateByteField (InitializerOp, ACPI_RESTAG_INTERRUPT,
+                    CurrentByteOffset + ASL_RESDESC_OFFSET (Irq.IrqMask));
+            }
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Now we can set the interrupt mask */
+
+    Descriptor->Irq.IrqMask = IrqMask;
+    return (Rnode);
+}
diff --git a/compiler/aslrestype2.c b/compiler/aslrestype2.c
index 7ca8eeb70c8..a9aeda73ad9 100644
--- a/compiler/aslrestype2.c
+++ b/compiler/aslrestype2.c
@@ -1,7 +1,7 @@
 
 /******************************************************************************
  *
- * Module Name: aslrestype2 - Long (type2) resource templates and descriptors
+ * Module Name: aslrestype2 - Miscellaneous Large resource descriptors
  *
  *****************************************************************************/
 
@@ -117,51 +117,22 @@
 
 #include "aslcompiler.h"
 #include "aslcompiler.y.h"
+#include "amlcode.h"
 
 #define _COMPONENT          ACPI_COMPILER
         ACPI_MODULE_NAME    ("aslrestype2")
 
-/* Local prototypes */
-
-static UINT16
-RsGetStringDataLength (
-    ACPI_PARSE_OBJECT       *InitializerOp);
-
+/*
+ * This module contains miscellaneous large resource descriptors:
+ *
+ * Register
+ * Interrupt
+ * VendorLong
+ */
 
 /*******************************************************************************
  *
- * FUNCTION:    RsGetStringDataLength
- *
- * PARAMETERS:  InitializerOp     - Start of a subtree of init nodes
- *
- * RETURN:      Valid string length if a string node is found (otherwise 0)
- *
- * DESCRIPTION: In a list of peer nodes, find the first one that contains a
- *              string and return the length of the string.
- *
- ******************************************************************************/
-
-static UINT16
-RsGetStringDataLength (
-    ACPI_PARSE_OBJECT       *InitializerOp)
-{
-
-    while (InitializerOp)
-    {
-        if (InitializerOp->Asl.ParseOpcode == PARSEOP_STRING_LITERAL)
-        {
-            return ((UINT16) (strlen (InitializerOp->Asl.Value.String) + 1));
-        }
-        InitializerOp = ASL_GET_PEER_NODE (InitializerOp);
-    }
-
-    return 0;
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoDwordIoDescriptor
+ * FUNCTION:    RsDoGeneralRegisterDescriptor
  *
  * PARAMETERS:  Op                  - Parent resource descriptor parse node
  *              CurrentByteOffset   - Offset into the resource template AML
@@ -169,43 +140,27 @@ RsGetStringDataLength (
  *
  * RETURN:      Completed resource node
  *
- * DESCRIPTION: Construct a long "DwordIO" descriptor
+ * DESCRIPTION: Construct a long "Register" descriptor
  *
  ******************************************************************************/
 
 ASL_RESOURCE_NODE *
-RsDoDwordIoDescriptor (
+RsDoGeneralRegisterDescriptor (
     ACPI_PARSE_OBJECT       *Op,
     UINT32                  CurrentByteOffset)
 {
     AML_RESOURCE            *Descriptor;
     ACPI_PARSE_OBJECT       *InitializerOp;
     ASL_RESOURCE_NODE       *Rnode;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT8                   *OptionalFields;
     UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
 
 
     InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
+    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_GENERIC_REGISTER));
 
     Descriptor = Rnode->Buffer;
-    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
-    Descriptor->Address32.ResourceType   = ACPI_ADDRESS_TYPE_IO_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
-    Descriptor->Address32.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS32) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
+    Descriptor->GenericReg.DescriptorType = ACPI_RESOURCE_NAME_GENERIC_REGISTER;
+    Descriptor->GenericReg.ResourceLength = 12;
 
     /* Process all child initialization nodes */
 
@@ -213,1637 +168,48 @@ RsDoDwordIoDescriptor (
     {
         switch (i)
         {
-        case 0: /* Resource Usage */
+        case 0: /* Address space */
 
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
-            break;
-
-        case 2: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
-            break;
-
-        case 3: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
-            break;
-
-        case 4: /* Range Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 0, 3);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 0);
-            break;
-
-        case 5: /* Address Granularity */
-
-            Descriptor->Address32.Granularity =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
-            break;
-
-        case 6: /* Address Min */
-
-            Descriptor->Address32.Minimum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
-            break;
-
-        case 7: /* Address Max */
-
-            Descriptor->Address32.Maximum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
-            break;
-
-        case 8: /* Translation Offset */
-
-            Descriptor->Address32.TranslationOffset =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
-            break;
-
-        case 9: /* Address Length */
-
-            Descriptor->Address32.AddressLength =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
-            break;
-
-        case 10: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                /* Found a valid ResourceSourceIndex */
-
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address32.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 11: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    /* Found a valid ResourceSource */
-
-                    Descriptor->Address32.ResourceLength = (UINT16)
-                        (Descriptor->Address32.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 12: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        case 13: /* Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 4, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 4);
-            break;
-
-        case 14: /* Translation Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoDwordMemoryDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "DwordMemory" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoDwordMemoryDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
-    Descriptor->Address32.ResourceType   = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
-    Descriptor->Address32.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS32) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
-            break;
-
-        case 2: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
-            break;
-
-        case 3: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
-            break;
-
-        case 4: /* Memory Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 1);
-            break;
-
-        case 5: /* Read/Write Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 0, 1);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 0);
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->Address32.Granularity =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->Address32.Minimum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->Address32.Maximum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->Address32.TranslationOffset =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->Address32.AddressLength =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
-            break;
-
-        case 11: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address32.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 12: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address32.ResourceLength = (UINT16)
-                        (Descriptor->Address32.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 13: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-
-        case 14: /* Address Range */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 3);
-            break;
-
-        case 15: /* Type */
-
-            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoDwordSpaceDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "DwordSpace" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoDwordSpaceDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
-    Descriptor->Address32.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS32) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Type */
-
-            Descriptor->Address32.ResourceType =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 1: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 2: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
-            break;
-
-        case 3: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
-            break;
-
-        case 4: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
-            break;
-
-        case 5: /* Type-Specific flags */
-
-            Descriptor->Address32.SpecificFlags =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->Address32.Granularity =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->Address32.Minimum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->Address32.Maximum =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->Address32.TranslationOffset =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->Address32.AddressLength =
-                (UINT32) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
-            break;
-
-        case 11: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address32.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 12: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address32.ResourceLength = (UINT16)
-                        (Descriptor->Address32.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 13: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST,
-                InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoExtendedIoDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "ExtendedIO" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoExtendedIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT16                  StringLength = 0;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
-    Descriptor->ExtAddress64.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
-    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
-
-    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
-        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* MinType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
-            break;
-
-        case 2: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
-            break;
-
-        case 3: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
-            break;
-
-        case 4: /* Range Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 0, 3);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 0);
-            break;
-
-        case 5: /* Address Granularity */
-
-            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
+            Descriptor->GenericReg.AddressSpaceId = (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_ADDRESSSPACE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.AddressSpaceId));
            break;
 
-        case 6: /* Address Min */
+        case 1: /* Register Bit Width */
 
-            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
+            Descriptor->GenericReg.BitWidth = (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_REGISTERBITWIDTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.BitWidth));
             break;
 
-        case 7: /* Address Max */
+        case 2: /* Register Bit Offset */
 
-            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
+            Descriptor->GenericReg.BitOffset = (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_REGISTERBITOFFSET,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.BitOffset));
             break;
 
-        case 8: /* Translation Offset */
+        case 3: /* Register Address */
 
-            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
+            Descriptor->GenericReg.Address = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_ADDRESS,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.Address));
             break;
 
-        case 9: /* Address Length */
+        case 4: /* Access Size (ACPI 3.0) */
 
-            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
-            break;
+            Descriptor->GenericReg.AccessSize = (UINT8) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_ACCESSSIZE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.AccessSize));
 
-        case 10: /* Type-Specific Attributes */
-
-            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
-            break;
-
-        case 11: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        case 12: /* Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 4, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 4);
-            break;
-
-        case 13: /* Translation Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoExtendedMemoryDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "ExtendedMemory" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoExtendedMemoryDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT16                  StringLength = 0;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
-    Descriptor->ExtAddress64.ResourceType    = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
-    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
-
-    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
-        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
-            break;
-
-        case 2: /* MinType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
-            break;
-
-        case 3: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
-            break;
-
-        case 4: /* Memory Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 1);
-            break;
-
-        case 5: /* Read/Write Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 0, 1);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 0);
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
-            break;
-
-        case 11: /* Type-Specific Attributes */
-
-            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
-            break;
-
-        case 12: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-
-        case 13: /* Address Range */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 3);
-            break;
-
-        case 14: /* Type */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoExtendedSpaceDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "ExtendedSpace" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoExtendedSpaceDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT16                  StringLength = 0;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
-    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
-
-    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
-        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Type */
-
-            Descriptor->ExtAddress64.ResourceType =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 1: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 2: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
-            break;
-
-        case 3: /* MinType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
-            break;
-
-        case 4: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
-            break;
-
-        case 5: /* Type-Specific flags */
-
-            Descriptor->ExtAddress64.SpecificFlags =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
-            break;
-
-        case 11: /* Type-Specific Attributes */
-
-            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
-            break;
-
-        case 12: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoQwordIoDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "QwordIO" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoQwordIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address64.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS64;
-    Descriptor->Address64.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
-    Descriptor->Address64.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
-            break;
-
-        case 2: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
-            break;
-
-        case 3: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
-            break;
-
-        case 4: /* Range Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 0, 3);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 0);
-            break;
-
-        case 5: /* Address Granularity */
-
-            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
-           break;
-
-        case 6: /* Address Min */
-
-            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
-            break;
-
-        case 7: /* Address Max */
-
-            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
-            break;
-
-        case 8: /* Translation Offset */
-
-            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
-            break;
-
-        case 9: /* Address Length */
-
-            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
-            break;
-
-        case 10: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            if (Descriptor->GenericReg.AccessSize > AML_FIELD_ACCESS_QWORD)
             {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address64.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 11: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address64.ResourceLength = (UINT16)
-                        (Descriptor->Address64.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                AslError (ASL_ERROR, ASL_MSG_INVALID_ACCESS_SIZE,
                     InitializerOp, NULL);
             }
-#endif
             break;
 
-        case 12: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        case 13: /* Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 4, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 4);
-            break;
-
-        case 14: /* Translation Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoQwordMemoryDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "QwordMemory" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoQwordMemoryDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address64.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS64;
-    Descriptor->Address64.ResourceType    = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
-    Descriptor->Address64.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
-            break;
-
-        case 2: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
-            break;
-
-        case 3: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
-            break;
-
-        case 4: /* Memory Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 1);
-            break;
-
-        case 5: /* Read/Write Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 0, 1);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 0);
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
-            break;
-
-        case 11: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address64.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 12: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address64.ResourceLength = (UINT16)
-                        (Descriptor->Address64.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 13: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-
-        case 14: /* Address Range */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 3);
-            break;
-
-        case 15: /* Type */
-
-            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoQwordSpaceDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "QwordSpace" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoQwordSpaceDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address64.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS64;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
-    Descriptor->Address64.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS64) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Type */
-
-            Descriptor->Address64.ResourceType =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 1: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 2: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
-            break;
-
-        case 3: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
-            break;
-
-        case 4: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
-            break;
-
-        case 5: /* Type-Specific flags */
-
-            Descriptor->Address64.SpecificFlags =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
-            break;
-
-        case 11: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address64.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 12: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address64.ResourceLength = (UINT16)
-                        (Descriptor->Address64.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 13: /* ResourceTag */
+        case 5: /* ResourceTag (ACPI 3.0b) */
 
             UtAttachNamepathToOwner (Op, InitializerOp);
             break;
@@ -1856,600 +222,6 @@ RsDoQwordSpaceDescriptor (
 
         InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
     }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoWordIoDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "WordIO" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoWordIoDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
-    Descriptor->Address16.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
-    Descriptor->Address16.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS16) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
-            break;
-
-        case 2: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
-            break;
-
-        case 3: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
-            break;
-
-        case 4: /* Range Type */
-
-            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 0, 3);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 0);
-            break;
-
-        case 5: /* Address Granularity */
-
-            Descriptor->Address16.Granularity = (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
-            break;
-
-        case 6: /* Address Min */
-
-            Descriptor->Address16.Minimum = (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
-            break;
-
-        case 7: /* Address Max */
-
-            Descriptor->Address16.Maximum = (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
-            break;
-
-        case 8: /* Translation Offset */
-
-            Descriptor->Address16.TranslationOffset = (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
-            break;
-
-        case 9: /* Address Length */
-
-            Descriptor->Address16.AddressLength = (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                 CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
-            break;
-
-        case 10: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address16.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 11: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address16.ResourceLength = (UINT16)
-                        (Descriptor->Address16.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 12: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        case 13: /* Type */
-
-            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 4, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 4);
-            break;
-
-        case 14: /* Translation Type */
-
-            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 5, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 5);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoWordBusNumberDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "WordBusNumber" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoWordBusNumberDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
-    Descriptor->Address16.ResourceType    = ACPI_ADDRESS_TYPE_BUS_NUMBER_RANGE;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
-    Descriptor->Address16.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS16) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 1: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
-            break;
-
-        case 2: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
-            break;
-
-        case 3: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
-            break;
-
-        case 4: /* Address Granularity */
-
-            Descriptor->Address16.Granularity =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
-            break;
-
-        case 5: /* Min Address */
-
-            Descriptor->Address16.Minimum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
-            break;
-
-        case 6: /* Max Address */
-
-            Descriptor->Address16.Maximum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
-            break;
-
-        case 7: /* Translation Offset */
-
-            Descriptor->Address16.TranslationOffset =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
-            break;
-
-        case 8: /* Address Length */
-
-            Descriptor->Address16.AddressLength =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                 CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
-            break;
-
-        case 9: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address16.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 10: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address16.ResourceLength = (UINT16)
-                        (Descriptor->Address16.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 11: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
-                            OptionIndex + StringLength;
-    return (Rnode);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoWordSpaceDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "WordSpace" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoWordSpaceDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT8                   *OptionalFields;
-    UINT16                  StringLength = 0;
-    UINT32                  OptionIndex = 0;
-    UINT32                  i;
-    BOOLEAN                 ResSourceIndex = FALSE;
-
-
-    InitializerOp = Op->Asl.Child;
-    StringLength = RsGetStringDataLength (InitializerOp);
-
-    Rnode = RsAllocateResourceNode (
-                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
-
-    /*
-     * Initial descriptor length -- may be enlarged if there are
-     * optional fields present
-     */
-    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
-    Descriptor->Address16.ResourceLength = (UINT16)
-        (sizeof (AML_RESOURCE_ADDRESS16) -
-         sizeof (AML_RESOURCE_LARGE_HEADER));
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Resource Type */
-
-            Descriptor->Address16.ResourceType =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 1: /* Resource Usage */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
-            break;
-
-        case 2: /* DecodeType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
-            break;
-
-        case 3: /* MinType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
-            break;
-
-        case 4: /* MaxType */
-
-            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
-            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
-            break;
-
-        case 5: /* Type-Specific flags */
-
-            Descriptor->Address16.SpecificFlags =
-                (UINT8) InitializerOp->Asl.Value.Integer;
-            break;
-
-        case 6: /* Address Granularity */
-
-            Descriptor->Address16.Granularity =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
-            break;
-
-        case 7: /* Min Address */
-
-            Descriptor->Address16.Minimum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
-            break;
-
-        case 8: /* Max Address */
-
-            Descriptor->Address16.Maximum =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
-            break;
-
-        case 9: /* Translation Offset */
-
-            Descriptor->Address16.TranslationOffset =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
-            break;
-
-        case 10: /* Address Length */
-
-            Descriptor->Address16.AddressLength =
-                (UINT16) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
-            break;
-
-        case 11: /* ResSourceIndex [Optional Field - BYTE] */
-
-            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
-            {
-                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
-                OptionIndex++;
-                Descriptor->Address16.ResourceLength++;
-                ResSourceIndex = TRUE;
-            }
-            break;
-
-        case 12: /* ResSource [Optional Field - STRING] */
-
-            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
-                (InitializerOp->Asl.Value.String))
-            {
-                if (StringLength)
-                {
-                    Descriptor->Address16.ResourceLength = (UINT16)
-                        (Descriptor->Address16.ResourceLength + StringLength);
-
-                    strcpy ((char *)
-                        &OptionalFields[OptionIndex],
-                        InitializerOp->Asl.Value.String);
-
-                    /* ResourceSourceIndex must also be valid */
-
-                    if (!ResSourceIndex)
-                    {
-                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
-                            InitializerOp, NULL);
-                    }
-                }
-            }
-
-#if 0
-            /*
-             * Not a valid ResourceSource, ResourceSourceIndex must also
-             * be invalid
-             */
-            else if (ResSourceIndex)
-            {
-                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
-                    InitializerOp, NULL);
-            }
-#endif
-            break;
-
-        case 13: /* ResourceTag */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-
-    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
-                            OptionIndex + StringLength;
     return (Rnode);
 }
 
@@ -2760,95 +532,3 @@ RsDoVendorLargeDescriptor (
 
     return (Rnode);
 }
-
-
-/*******************************************************************************
- *
- * FUNCTION:    RsDoGeneralRegisterDescriptor
- *
- * PARAMETERS:  Op                  - Parent resource descriptor parse node
- *              CurrentByteOffset   - Offset into the resource template AML
- *                                    buffer (to track references to the desc)
- *
- * RETURN:      Completed resource node
- *
- * DESCRIPTION: Construct a long "Register" descriptor
- *
- ******************************************************************************/
-
-ASL_RESOURCE_NODE *
-RsDoGeneralRegisterDescriptor (
-    ACPI_PARSE_OBJECT       *Op,
-    UINT32                  CurrentByteOffset)
-{
-    AML_RESOURCE            *Descriptor;
-    ACPI_PARSE_OBJECT       *InitializerOp;
-    ASL_RESOURCE_NODE       *Rnode;
-    UINT32                  i;
-
-
-    InitializerOp = Op->Asl.Child;
-    Rnode = RsAllocateResourceNode (sizeof (AML_RESOURCE_GENERIC_REGISTER));
-
-    Descriptor = Rnode->Buffer;
-    Descriptor->GenericReg.DescriptorType = ACPI_RESOURCE_NAME_GENERIC_REGISTER;
-    Descriptor->GenericReg.ResourceLength = 12;
-
-    /* Process all child initialization nodes */
-
-    for (i = 0; InitializerOp; i++)
-    {
-        switch (i)
-        {
-        case 0: /* Address space */
-
-            Descriptor->GenericReg.AddressSpaceId = (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_ADDRESSSPACE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.AddressSpaceId));
-           break;
-
-        case 1: /* Register Bit Width */
-
-            Descriptor->GenericReg.BitWidth = (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_REGISTERBITWIDTH,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.BitWidth));
-            break;
-
-        case 2: /* Register Bit Offset */
-
-            Descriptor->GenericReg.BitOffset = (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_REGISTERBITOFFSET,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.BitOffset));
-            break;
-
-        case 3: /* Register Address */
-
-            Descriptor->GenericReg.Address = InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_ADDRESS,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.Address));
-            break;
-
-        case 4: /* Access Size (ACPI 3.0) */
-
-            Descriptor->GenericReg.AccessSize = (UINT8) InitializerOp->Asl.Value.Integer;
-            RsCreateByteField (InitializerOp, ACPI_RESTAG_ACCESSSIZE,
-                CurrentByteOffset + ASL_RESDESC_OFFSET (GenericReg.AccessSize));
-            break;
-
-        case 5: /* ResourceTag (ACPI 3.0b) */
-
-            UtAttachNamepathToOwner (Op, InitializerOp);
-            break;
-
-        default:
-
-            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
-            break;
-        }
-
-        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
-    }
-    return (Rnode);
-}
-
-
diff --git a/compiler/aslrestype2d.c b/compiler/aslrestype2d.c
new file mode 100644
index 00000000000..7725bf9a8da
--- /dev/null
+++ b/compiler/aslrestype2d.c
@@ -0,0 +1,814 @@
+
+/******************************************************************************
+ *
+ * Module Name: aslrestype2d - Large DWord address resource descriptors
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "aslcompiler.h"
+#include "aslcompiler.y.h"
+
+#define _COMPONENT          ACPI_COMPILER
+        ACPI_MODULE_NAME    ("aslrestype2d")
+
+/*
+ * This module contains the Dword (32-bit) address space descriptors:
+ *
+ * DwordIO
+ * DwordMemory
+ * DwordSpace
+ */
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoDwordIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "DwordIO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoDwordIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT8                   *OptionalFields;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
+    Descriptor->Address32.ResourceType   = ACPI_ADDRESS_TYPE_IO_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
+    Descriptor->Address32.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS32) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
+            break;
+
+        case 2: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
+            break;
+
+        case 3: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
+            break;
+
+        case 4: /* Range Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 0, 3);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 0);
+            break;
+
+        case 5: /* Address Granularity */
+
+            Descriptor->Address32.Granularity =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 6: /* Address Min */
+
+            Descriptor->Address32.Minimum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 7: /* Address Max */
+
+            Descriptor->Address32.Maximum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 8: /* Translation Offset */
+
+            Descriptor->Address32.TranslationOffset =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
+            break;
+
+        case 9: /* Address Length */
+
+            Descriptor->Address32.AddressLength =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 10: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                /* Found a valid ResourceSourceIndex */
+
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address32.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 11: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    /* Found a valid ResourceSource */
+
+                    Descriptor->Address32.ResourceLength = (UINT16)
+                        (Descriptor->Address32.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 12: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        case 13: /* Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 4, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 4);
+            break;
+
+        case 14: /* Translation Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address32.Minimum,
+        Descriptor->Address32.Maximum,
+        Descriptor->Address32.AddressLength,
+        Descriptor->Address32.Granularity,
+        Descriptor->Address32.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoDwordMemoryDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "DwordMemory" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoDwordMemoryDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
+    Descriptor->Address32.ResourceType   = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
+    Descriptor->Address32.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS32) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
+            break;
+
+        case 2: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
+            break;
+
+        case 3: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
+            break;
+
+        case 4: /* Memory Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 1);
+            break;
+
+        case 5: /* Read/Write Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 0, 1);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 0);
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->Address32.Granularity =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->Address32.Minimum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->Address32.Maximum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->Address32.TranslationOffset =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->Address32.AddressLength =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address32.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 12: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address32.ResourceLength = (UINT16)
+                        (Descriptor->Address32.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 13: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+
+        case 14: /* Address Range */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 3);
+            break;
+
+        case 15: /* Type */
+
+            RsSetFlagBits (&Descriptor->Address32.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address32.Minimum,
+        Descriptor->Address32.Maximum,
+        Descriptor->Address32.AddressLength,
+        Descriptor->Address32.Granularity,
+        Descriptor->Address32.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoDwordSpaceDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "DwordSpace" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoDwordSpaceDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS32) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address32.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS32;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS32);
+    Descriptor->Address32.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS32) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Type */
+
+            Descriptor->Address32.ResourceType =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 1: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 2: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 1);
+            break;
+
+        case 3: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 2);
+            break;
+
+        case 4: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address32.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Flags), 3);
+            break;
+
+        case 5: /* Type-Specific flags */
+
+            Descriptor->Address32.SpecificFlags =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->Address32.Granularity =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->Address32.Minimum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->Address32.Maximum =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->Address32.TranslationOffset =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->Address32.AddressLength =
+                (UINT32) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address32.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address32.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 12: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address32.ResourceLength = (UINT16)
+                        (Descriptor->Address32.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 13: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST,
+                InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address32.Minimum,
+        Descriptor->Address32.Maximum,
+        Descriptor->Address32.AddressLength,
+        Descriptor->Address32.Granularity,
+        Descriptor->Address32.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS32) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
diff --git a/compiler/aslrestype2e.c b/compiler/aslrestype2e.c
new file mode 100644
index 00000000000..46f56cec136
--- /dev/null
+++ b/compiler/aslrestype2e.c
@@ -0,0 +1,646 @@
+
+/******************************************************************************
+ *
+ * Module Name: aslrestype2e - Large Extended address resource descriptors
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "aslcompiler.h"
+#include "aslcompiler.y.h"
+
+#define _COMPONENT          ACPI_COMPILER
+        ACPI_MODULE_NAME    ("aslrestype2e")
+
+/*
+ * This module contains the Extended (64-bit) address space descriptors:
+ *
+ * ExtendedIO
+ * ExtendedMemory
+ * ExtendedSpace
+ */
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoExtendedIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "ExtendedIO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoExtendedIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT16                  StringLength = 0;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
+    Descriptor->ExtAddress64.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
+    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
+
+    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
+        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* MinType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
+            break;
+
+        case 2: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
+            break;
+
+        case 3: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
+            break;
+
+        case 4: /* Range Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 0, 3);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 0);
+            break;
+
+        case 5: /* Address Granularity */
+
+            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
+            GranOp = InitializerOp;
+           break;
+
+        case 6: /* Address Min */
+
+            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 7: /* Address Max */
+
+            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 8: /* Translation Offset */
+
+            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
+            break;
+
+        case 9: /* Address Length */
+
+            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 10: /* Type-Specific Attributes */
+
+            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
+            break;
+
+        case 11: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        case 12: /* Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 4, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 4);
+            break;
+
+        case 13: /* Translation Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->ExtAddress64.Minimum,
+        Descriptor->ExtAddress64.Maximum,
+        Descriptor->ExtAddress64.AddressLength,
+        Descriptor->ExtAddress64.Granularity,
+        Descriptor->ExtAddress64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoExtendedMemoryDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "ExtendedMemory" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoExtendedMemoryDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT16                  StringLength = 0;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
+    Descriptor->ExtAddress64.ResourceType    = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
+    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
+
+    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
+        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
+            break;
+
+        case 2: /* MinType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
+            break;
+
+        case 3: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
+            break;
+
+        case 4: /* Memory Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 1);
+            break;
+
+        case 5: /* Read/Write Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 0, 1);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 0);
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* Type-Specific Attributes */
+
+            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
+            break;
+
+        case 12: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+
+        case 13: /* Address Range */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 3);
+            break;
+
+        case 14: /* Type */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->ExtAddress64.Minimum,
+        Descriptor->ExtAddress64.Maximum,
+        Descriptor->ExtAddress64.AddressLength,
+        Descriptor->ExtAddress64.Granularity,
+        Descriptor->ExtAddress64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoExtendedSpaceDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "ExtendedSpace" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoExtendedSpaceDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT16                  StringLength = 0;
+    UINT32                  i;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->ExtAddress64.DescriptorType  = ACPI_RESOURCE_NAME_EXTENDED_ADDRESS64;
+    Descriptor->ExtAddress64.RevisionID      = AML_RESOURCE_EXTENDED_ADDRESS_REVISION;
+
+    Descriptor->ExtAddress64.ResourceLength  = (UINT16)
+        (sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Type */
+
+            Descriptor->ExtAddress64.ResourceType =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 1: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 2: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 1);
+            break;
+
+        case 3: /* MinType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 2);
+            break;
+
+        case 4: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->ExtAddress64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Flags), 3);
+            break;
+
+        case 5: /* Type-Specific flags */
+
+            Descriptor->ExtAddress64.SpecificFlags =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->ExtAddress64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->ExtAddress64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->ExtAddress64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->ExtAddress64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->ExtAddress64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* Type-Specific Attributes */
+
+            Descriptor->ExtAddress64.TypeSpecific = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TYPESPECIFICATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (ExtAddress64.TypeSpecific));
+            break;
+
+        case 12: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->ExtAddress64.Minimum,
+        Descriptor->ExtAddress64.Maximum,
+        Descriptor->ExtAddress64.AddressLength,
+        Descriptor->ExtAddress64.Granularity,
+        Descriptor->ExtAddress64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_EXTENDED_ADDRESS64) + StringLength;
+    return (Rnode);
+}
diff --git a/compiler/aslrestype2q.c b/compiler/aslrestype2q.c
new file mode 100644
index 00000000000..591ac25c42b
--- /dev/null
+++ b/compiler/aslrestype2q.c
@@ -0,0 +1,793 @@
+
+/******************************************************************************
+ *
+ * Module Name: aslrestype2q - Large QWord address resource descriptors
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "aslcompiler.h"
+#include "aslcompiler.y.h"
+
+#define _COMPONENT          ACPI_COMPILER
+        ACPI_MODULE_NAME    ("aslrestype2q")
+
+/*
+ * This module contains the QWord (64-bit) address space descriptors:
+ *
+ * QWordIO
+ * QWordMemory
+ * QWordSpace
+ */
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoQwordIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "QwordIO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoQwordIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address64.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS64;
+    Descriptor->Address64.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
+    Descriptor->Address64.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
+            break;
+
+        case 2: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
+            break;
+
+        case 3: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
+            break;
+
+        case 4: /* Range Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 0, 3);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 0);
+            break;
+
+        case 5: /* Address Granularity */
+
+            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 6: /* Address Min */
+
+            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 7: /* Address Max */
+
+            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 8: /* Translation Offset */
+
+            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
+            break;
+
+        case 9: /* Address Length */
+
+            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 10: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address64.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 11: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address64.ResourceLength = (UINT16)
+                        (Descriptor->Address64.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 12: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        case 13: /* Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 4, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 4);
+            break;
+
+        case 14: /* Translation Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address64.Minimum,
+        Descriptor->Address64.Maximum,
+        Descriptor->Address64.AddressLength,
+        Descriptor->Address64.Granularity,
+        Descriptor->Address64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoQwordMemoryDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "QwordMemory" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoQwordMemoryDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address64.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS64;
+    Descriptor->Address64.ResourceType    = ACPI_ADDRESS_TYPE_MEMORY_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
+    Descriptor->Address64.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
+            break;
+
+        case 2: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
+            break;
+
+        case 3: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
+            break;
+
+        case 4: /* Memory Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 1);
+            break;
+
+        case 5: /* Read/Write Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 0, 1);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_READWRITETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 0);
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address64.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 12: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address64.ResourceLength = (UINT16)
+                        (Descriptor->Address64.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 13: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+
+        case 14: /* Address Range */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MEMATTRIBUTES,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 3);
+            break;
+
+        case 15: /* Type */
+
+            RsSetFlagBits (&Descriptor->Address64.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address64.Minimum,
+        Descriptor->Address64.Maximum,
+        Descriptor->Address64.AddressLength,
+        Descriptor->Address64.Granularity,
+        Descriptor->Address64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoQwordSpaceDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "QwordSpace" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoQwordSpaceDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS64) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address64.DescriptorType = ACPI_RESOURCE_NAME_ADDRESS64;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS64);
+    Descriptor->Address64.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS64) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Type */
+
+            Descriptor->Address64.ResourceType =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 1: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 2: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 1);
+            break;
+
+        case 3: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 2);
+            break;
+
+        case 4: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address64.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Flags), 3);
+            break;
+
+        case 5: /* Type-Specific flags */
+
+            Descriptor->Address64.SpecificFlags =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->Address64.Granularity = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->Address64.Minimum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->Address64.Maximum = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->Address64.TranslationOffset = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->Address64.AddressLength = InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address64.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address64.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 12: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address64.ResourceLength = (UINT16)
+                        (Descriptor->Address64.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 13: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address64.Minimum,
+        Descriptor->Address64.Maximum,
+        Descriptor->Address64.AddressLength,
+        Descriptor->Address64.Granularity,
+        Descriptor->Address64.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS64) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
diff --git a/compiler/aslrestype2w.c b/compiler/aslrestype2w.c
new file mode 100644
index 00000000000..45f1858fccf
--- /dev/null
+++ b/compiler/aslrestype2w.c
@@ -0,0 +1,774 @@
+
+/******************************************************************************
+ *
+ * Module Name: aslrestype2w - Large Word address resource descriptors
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "aslcompiler.h"
+#include "aslcompiler.y.h"
+
+#define _COMPONENT          ACPI_COMPILER
+        ACPI_MODULE_NAME    ("aslrestype2w")
+
+/*
+ * This module contains the Word (16-bit) address space descriptors:
+ *
+ * WordIO
+ * WordMemory
+ * WordSpace
+ */
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoWordIoDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "WordIO" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoWordIoDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
+    Descriptor->Address16.ResourceType    = ACPI_ADDRESS_TYPE_IO_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
+    Descriptor->Address16.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS16) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
+            break;
+
+        case 2: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
+            break;
+
+        case 3: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
+            break;
+
+        case 4: /* Range Type */
+
+            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 0, 3);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_RANGETYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 0);
+            break;
+
+        case 5: /* Address Granularity */
+
+            Descriptor->Address16.Granularity = (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 6: /* Address Min */
+
+            Descriptor->Address16.Minimum = (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 7: /* Address Max */
+
+            Descriptor->Address16.Maximum = (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 8: /* Translation Offset */
+
+            Descriptor->Address16.TranslationOffset = (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
+            break;
+
+        case 9: /* Address Length */
+
+            Descriptor->Address16.AddressLength = (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                 CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 10: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address16.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 11: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address16.ResourceLength = (UINT16)
+                        (Descriptor->Address16.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 12: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        case 13: /* Type */
+
+            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 4, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 4);
+            break;
+
+        case 14: /* Translation Type */
+
+            RsSetFlagBits (&Descriptor->Address16.SpecificFlags, InitializerOp, 5, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_TRANSTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.SpecificFlags), 5);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address16.Minimum,
+        Descriptor->Address16.Maximum,
+        Descriptor->Address16.AddressLength,
+        Descriptor->Address16.Granularity,
+        Descriptor->Address16.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoWordBusNumberDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "WordBusNumber" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoWordBusNumberDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
+    Descriptor->Address16.ResourceType    = ACPI_ADDRESS_TYPE_BUS_NUMBER_RANGE;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
+    Descriptor->Address16.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS16) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 1: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
+            break;
+
+        case 2: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
+            break;
+
+        case 3: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
+            break;
+
+        case 4: /* Address Granularity */
+
+            Descriptor->Address16.Granularity =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 5: /* Min Address */
+
+            Descriptor->Address16.Minimum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 6: /* Max Address */
+
+            Descriptor->Address16.Maximum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 7: /* Translation Offset */
+
+            Descriptor->Address16.TranslationOffset =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
+            break;
+
+        case 8: /* Address Length */
+
+            Descriptor->Address16.AddressLength =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                 CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 9: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address16.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 10: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address16.ResourceLength = (UINT16)
+                        (Descriptor->Address16.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 11: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address16.Minimum,
+        Descriptor->Address16.Maximum,
+        Descriptor->Address16.AddressLength,
+        Descriptor->Address16.Granularity,
+        Descriptor->Address16.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    RsDoWordSpaceDescriptor
+ *
+ * PARAMETERS:  Op                  - Parent resource descriptor parse node
+ *              CurrentByteOffset   - Offset into the resource template AML
+ *                                    buffer (to track references to the desc)
+ *
+ * RETURN:      Completed resource node
+ *
+ * DESCRIPTION: Construct a long "WordSpace" descriptor
+ *
+ ******************************************************************************/
+
+ASL_RESOURCE_NODE *
+RsDoWordSpaceDescriptor (
+    ACPI_PARSE_OBJECT       *Op,
+    UINT32                  CurrentByteOffset)
+{
+    AML_RESOURCE            *Descriptor;
+    ACPI_PARSE_OBJECT       *InitializerOp;
+    ACPI_PARSE_OBJECT       *MinOp = NULL;
+    ACPI_PARSE_OBJECT       *MaxOp = NULL;
+    ACPI_PARSE_OBJECT       *LengthOp = NULL;
+    ACPI_PARSE_OBJECT       *GranOp = NULL;
+    ASL_RESOURCE_NODE       *Rnode;
+    UINT8                   *OptionalFields;
+    UINT16                  StringLength = 0;
+    UINT32                  OptionIndex = 0;
+    UINT32                  i;
+    BOOLEAN                 ResSourceIndex = FALSE;
+
+
+    InitializerOp = Op->Asl.Child;
+    StringLength = RsGetStringDataLength (InitializerOp);
+
+    Rnode = RsAllocateResourceNode (
+                sizeof (AML_RESOURCE_ADDRESS16) + 1 + StringLength);
+
+    Descriptor = Rnode->Buffer;
+    Descriptor->Address16.DescriptorType  = ACPI_RESOURCE_NAME_ADDRESS16;
+
+    /*
+     * Initial descriptor length -- may be enlarged if there are
+     * optional fields present
+     */
+    OptionalFields = ((UINT8 *) Descriptor) + sizeof (AML_RESOURCE_ADDRESS16);
+    Descriptor->Address16.ResourceLength = (UINT16)
+        (sizeof (AML_RESOURCE_ADDRESS16) -
+         sizeof (AML_RESOURCE_LARGE_HEADER));
+
+    /* Process all child initialization nodes */
+
+    for (i = 0; InitializerOp; i++)
+    {
+        switch (i)
+        {
+        case 0: /* Resource Type */
+
+            Descriptor->Address16.ResourceType =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 1: /* Resource Usage */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 0, 1);
+            break;
+
+        case 2: /* DecodeType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 1, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_DECODE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 1);
+            break;
+
+        case 3: /* MinType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 2, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MINTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 2);
+            break;
+
+        case 4: /* MaxType */
+
+            RsSetFlagBits (&Descriptor->Address16.Flags, InitializerOp, 3, 0);
+            RsCreateBitField (InitializerOp, ACPI_RESTAG_MAXTYPE,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Flags), 3);
+            break;
+
+        case 5: /* Type-Specific flags */
+
+            Descriptor->Address16.SpecificFlags =
+                (UINT8) InitializerOp->Asl.Value.Integer;
+            break;
+
+        case 6: /* Address Granularity */
+
+            Descriptor->Address16.Granularity =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_GRANULARITY,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Granularity));
+            GranOp = InitializerOp;
+            break;
+
+        case 7: /* Min Address */
+
+            Descriptor->Address16.Minimum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MINADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Minimum));
+            MinOp = InitializerOp;
+            break;
+
+        case 8: /* Max Address */
+
+            Descriptor->Address16.Maximum =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_MAXADDR,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.Maximum));
+            MaxOp = InitializerOp;
+            break;
+
+        case 9: /* Translation Offset */
+
+            Descriptor->Address16.TranslationOffset =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_TRANSLATION,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.TranslationOffset));
+            break;
+
+        case 10: /* Address Length */
+
+            Descriptor->Address16.AddressLength =
+                (UINT16) InitializerOp->Asl.Value.Integer;
+            RsCreateByteField (InitializerOp, ACPI_RESTAG_LENGTH,
+                CurrentByteOffset + ASL_RESDESC_OFFSET (Address16.AddressLength));
+            LengthOp = InitializerOp;
+            break;
+
+        case 11: /* ResSourceIndex [Optional Field - BYTE] */
+
+            if (InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG)
+            {
+                OptionalFields[0] = (UINT8) InitializerOp->Asl.Value.Integer;
+                OptionIndex++;
+                Descriptor->Address16.ResourceLength++;
+                ResSourceIndex = TRUE;
+            }
+            break;
+
+        case 12: /* ResSource [Optional Field - STRING] */
+
+            if ((InitializerOp->Asl.ParseOpcode != PARSEOP_DEFAULT_ARG) &&
+                (InitializerOp->Asl.Value.String))
+            {
+                if (StringLength)
+                {
+                    Descriptor->Address16.ResourceLength = (UINT16)
+                        (Descriptor->Address16.ResourceLength + StringLength);
+
+                    strcpy ((char *)
+                        &OptionalFields[OptionIndex],
+                        InitializerOp->Asl.Value.String);
+
+                    /* ResourceSourceIndex must also be valid */
+
+                    if (!ResSourceIndex)
+                    {
+                        AslError (ASL_ERROR, ASL_MSG_RESOURCE_INDEX,
+                            InitializerOp, NULL);
+                    }
+                }
+            }
+
+#if 0
+            /*
+             * Not a valid ResourceSource, ResourceSourceIndex must also
+             * be invalid
+             */
+            else if (ResSourceIndex)
+            {
+                AslError (ASL_ERROR, ASL_MSG_RESOURCE_SOURCE,
+                    InitializerOp, NULL);
+            }
+#endif
+            break;
+
+        case 13: /* ResourceTag */
+
+            UtAttachNamepathToOwner (Op, InitializerOp);
+            break;
+
+        default:
+
+            AslError (ASL_ERROR, ASL_MSG_RESOURCE_LIST, InitializerOp, NULL);
+            break;
+        }
+
+        InitializerOp = RsCompleteNodeAndGetNext (InitializerOp);
+    }
+
+    /* Validate the Min/Max/Len/Gran values */
+
+    RsLargeAddressCheck (
+        Descriptor->Address16.Minimum,
+        Descriptor->Address16.Maximum,
+        Descriptor->Address16.AddressLength,
+        Descriptor->Address16.Granularity,
+        Descriptor->Address16.Flags,
+        MinOp, MaxOp, LengthOp, GranOp);
+
+    Rnode->BufferLength = sizeof (AML_RESOURCE_ADDRESS16) +
+                            OptionIndex + StringLength;
+    return (Rnode);
+}
diff --git a/compiler/asltypes.h b/compiler/asltypes.h
index 30b8fdfe7d0..815763499d5 100644
--- a/compiler/asltypes.h
+++ b/compiler/asltypes.h
@@ -420,7 +420,16 @@ typedef enum
     ASL_MSG_SERIALIZED,
     ASL_MSG_COMPILER_RESERVED,
     ASL_MSG_NAMED_OBJECT_IN_WHILE,
-    ASL_MSG_LOCAL_OUTSIDE_METHOD
+    ASL_MSG_LOCAL_OUTSIDE_METHOD,
+    ASL_MSG_ALIGNMENT,
+    ASL_MSG_ISA_ADDRESS,
+    ASL_MSG_INVALID_MIN_MAX,
+    ASL_MSG_INVALID_LENGTH,
+    ASL_MSG_INVALID_LENGTH_FIXED,
+    ASL_MSG_INVALID_GRANULARITY,
+    ASL_MSG_INVALID_GRAN_FIXED,
+    ASL_MSG_INVALID_ACCESS_SIZE,
+    ASL_MSG_INVALID_ADDR_FLAGS
 
 } ASL_MESSAGE_IDS;
 
@@ -540,7 +549,16 @@ char                        *AslMessages [] = {
 /*    ASL_MSG_SERIALIZED */                 "Control Method marked Serialized",
 /*    ASL_MSG_COMPILER_RESERVED */          "Use of compiler reserved name",
 /*    ASL_MSG_NAMED_OBJECT_IN_WHILE */      "Creating a named object in a While loop",
-/*    ASL_MSG_LOCAL_OUTSIDE_METHOD */       "Local or Arg used outside a control method"
+/*    ASL_MSG_LOCAL_OUTSIDE_METHOD */       "Local or Arg used outside a control method",
+/*    ASL_MSG_ALIGNMENT */                  "Must be a multiple of alignment/granularity value",
+/*    ASL_MSG_ISA_ADDRESS */                "Maximum 10-bit ISA address (0x3FF)",
+/*    ASL_MSG_INVALID_MIN_MAX */            "Address Min is greater than Address Max",
+/*    ASL_MSG_INVALID_LENGTH */             "Length is larger than Min/Max window",
+/*    ASL_MSG_INVALID_LENGTH_FIXED */       "Length is not equal to fixed Min/Max window",
+/*    ASL_MSG_INVALID_GRANULARITY */        "Granularity must be zero or a power of two minus one",
+/*    ASL_MSG_INVALID_GRAN_FIXED */         "Granularity must be zero for fixed Min/Max",
+/*    ASL_MSG_INVALID_ACCESS_SIZE */        "Invalid AccessSize (Maximum is 4 - QWord access)",
+/*    ASL_MSG_INVALID_ADDR_FLAGS */         "Invalid combination of Length and Min/Max fixed flags"
 
 };
 
diff --git a/debugger/dbcmds.c b/debugger/dbcmds.c
index 50a8b2e037e..f8c6496a7df 100644
--- a/debugger/dbcmds.c
+++ b/debugger/dbcmds.c
@@ -675,7 +675,7 @@ AcpiDbDisplayTableInfo (
 
     /* Walk the entire root table list */
 
-    for (i = 0; i < AcpiGbl_RootTableList.Count; i++)
+    for (i = 0; i < AcpiGbl_RootTableList.CurrentTableCount; i++)
     {
         TableDesc = &AcpiGbl_RootTableList.Tables[i];
         AcpiOsPrintf ("%d ", i);
diff --git a/debugger/dbdisply.c b/debugger/dbdisply.c
index ec4715bddf6..a610d802c8d 100644
--- a/debugger/dbdisply.c
+++ b/debugger/dbdisply.c
@@ -816,6 +816,7 @@ AcpiDbDisplayGpes (
     ACPI_GPE_XRUPT_INFO     *GpeXruptInfo;
     ACPI_GPE_EVENT_INFO     *GpeEventInfo;
     ACPI_GPE_REGISTER_INFO  *GpeRegisterInfo;
+    char                    *GpeType;
     UINT32                  GpeIndex;
     UINT32                  Block = 0;
     UINT32                  i;
@@ -844,8 +845,17 @@ AcpiDbDisplayGpes (
                 AcpiOsPrintf ("Could not convert name to pathname\n");
             }
 
-            AcpiOsPrintf ("\nBlock %d - Info %p  DeviceNode %p [%s]\n",
-                Block, GpeBlock, GpeBlock->Node, Buffer);
+            if (GpeBlock->Node == AcpiGbl_FadtGpeDevice)
+            {
+                GpeType = "FADT-defined GPE block";
+            }
+            else
+            {
+                GpeType = "GPE Block Device";
+            }
+
+            AcpiOsPrintf ("\nBlock %d - Info %p  DeviceNode %p [%s] - %s\n",
+                Block, GpeBlock, GpeBlock->Node, Buffer, GpeType);
 
             AcpiOsPrintf ("    Registers:    %u (%u GPEs)\n",
                 GpeBlock->RegisterCount, GpeBlock->GpeCount);
diff --git a/events/evgpe.c b/events/evgpe.c
index d3569b18840..604f23d6c20 100644
--- a/events/evgpe.c
+++ b/events/evgpe.c
@@ -370,7 +370,7 @@ AcpiEvGetGpeEventInfo (
     ACPI_FUNCTION_ENTRY ();
 
 
-    /* A NULL GpeBlock means use the FADT-defined GPE block(s) */
+    /* A NULL GpeDevice means use the FADT-defined GPE block(s) */
 
     if (!GpeDevice)
     {
diff --git a/events/evgpeblk.c b/events/evgpeblk.c
index b109673ba8e..721411703e8 100644
--- a/events/evgpeblk.c
+++ b/events/evgpeblk.c
@@ -123,28 +123,6 @@
 
 /* Local prototypes */
 
-static ACPI_STATUS
-AcpiEvMatchGpeMethod (
-    ACPI_HANDLE             ObjHandle,
-    UINT32                  Level,
-    void                    *ObjDesc,
-    void                    **ReturnValue);
-
-static ACPI_STATUS
-AcpiEvMatchPrwAndGpe (
-    ACPI_HANDLE             ObjHandle,
-    UINT32                  Level,
-    void                    *Info,
-    void                    **ReturnValue);
-
-static ACPI_GPE_XRUPT_INFO *
-AcpiEvGetGpeXruptBlock (
-    UINT32                  InterruptNumber);
-
-static ACPI_STATUS
-AcpiEvDeleteGpeXrupt (
-    ACPI_GPE_XRUPT_INFO     *GpeXrupt);
-
 static ACPI_STATUS
 AcpiEvInstallGpeBlock (
     ACPI_GPE_BLOCK_INFO     *GpeBlock,
@@ -155,581 +133,6 @@ AcpiEvCreateGpeInfoBlocks (
     ACPI_GPE_BLOCK_INFO     *GpeBlock);
 
 
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvValidGpeEvent
- *
- * PARAMETERS:  GpeEventInfo                - Info for this GPE
- *
- * RETURN:      TRUE if the GpeEvent is valid
- *
- * DESCRIPTION: Validate a GPE event. DO NOT CALL FROM INTERRUPT LEVEL.
- *              Should be called only when the GPE lists are semaphore locked
- *              and not subject to change.
- *
- ******************************************************************************/
-
-BOOLEAN
-AcpiEvValidGpeEvent (
-    ACPI_GPE_EVENT_INFO     *GpeEventInfo)
-{
-    ACPI_GPE_XRUPT_INFO     *GpeXruptBlock;
-    ACPI_GPE_BLOCK_INFO     *GpeBlock;
-
-
-    ACPI_FUNCTION_ENTRY ();
-
-
-    /* No need for spin lock since we are not changing any list elements */
-
-    /* Walk the GPE interrupt levels */
-
-    GpeXruptBlock = AcpiGbl_GpeXruptListHead;
-    while (GpeXruptBlock)
-    {
-        GpeBlock = GpeXruptBlock->GpeBlockListHead;
-
-        /* Walk the GPE blocks on this interrupt level */
-
-        while (GpeBlock)
-        {
-            if ((&GpeBlock->EventInfo[0] <= GpeEventInfo) &&
-                (&GpeBlock->EventInfo[GpeBlock->GpeCount] > GpeEventInfo))
-            {
-                return (TRUE);
-            }
-
-            GpeBlock = GpeBlock->Next;
-        }
-
-        GpeXruptBlock = GpeXruptBlock->Next;
-    }
-
-    return (FALSE);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvWalkGpeList
- *
- * PARAMETERS:  GpeWalkCallback     - Routine called for each GPE block
- *              Context             - Value passed to callback
- *
- * RETURN:      Status
- *
- * DESCRIPTION: Walk the GPE lists.
- *
- ******************************************************************************/
-
-ACPI_STATUS
-AcpiEvWalkGpeList (
-    ACPI_GPE_CALLBACK       GpeWalkCallback,
-    void                    *Context)
-{
-    ACPI_GPE_BLOCK_INFO     *GpeBlock;
-    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo;
-    ACPI_STATUS             Status = AE_OK;
-    ACPI_CPU_FLAGS          Flags;
-
-
-    ACPI_FUNCTION_TRACE (EvWalkGpeList);
-
-
-    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
-
-    /* Walk the interrupt level descriptor list */
-
-    GpeXruptInfo = AcpiGbl_GpeXruptListHead;
-    while (GpeXruptInfo)
-    {
-        /* Walk all Gpe Blocks attached to this interrupt level */
-
-        GpeBlock = GpeXruptInfo->GpeBlockListHead;
-        while (GpeBlock)
-        {
-            /* One callback per GPE block */
-
-            Status = GpeWalkCallback (GpeXruptInfo, GpeBlock, Context);
-            if (ACPI_FAILURE (Status))
-            {
-                if (Status == AE_CTRL_END) /* Callback abort */
-                {
-                    Status = AE_OK;
-                }
-                goto UnlockAndExit;
-            }
-
-            GpeBlock = GpeBlock->Next;
-        }
-
-        GpeXruptInfo = GpeXruptInfo->Next;
-    }
-
-UnlockAndExit:
-    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
-    return_ACPI_STATUS (Status);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvDeleteGpeHandlers
- *
- * PARAMETERS:  GpeXruptInfo        - GPE Interrupt info
- *              GpeBlock            - Gpe Block info
- *
- * RETURN:      Status
- *
- * DESCRIPTION: Delete all Handler objects found in the GPE data structs.
- *              Used only prior to termination.
- *
- ******************************************************************************/
-
-ACPI_STATUS
-AcpiEvDeleteGpeHandlers (
-    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo,
-    ACPI_GPE_BLOCK_INFO     *GpeBlock,
-    void                    *Context)
-{
-    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
-    UINT32                  i;
-    UINT32                  j;
-
-
-    ACPI_FUNCTION_TRACE (EvDeleteGpeHandlers);
-
-
-    /* Examine each GPE Register within the block */
-
-    for (i = 0; i < GpeBlock->RegisterCount; i++)
-    {
-        /* Now look at the individual GPEs in this byte register */
-
-        for (j = 0; j < ACPI_GPE_REGISTER_WIDTH; j++)
-        {
-            GpeEventInfo = &GpeBlock->EventInfo[((ACPI_SIZE) i *
-                ACPI_GPE_REGISTER_WIDTH) + j];
-
-            if ((GpeEventInfo->Flags & ACPI_GPE_DISPATCH_MASK) ==
-                    ACPI_GPE_DISPATCH_HANDLER)
-            {
-                ACPI_FREE (GpeEventInfo->Dispatch.Handler);
-                GpeEventInfo->Dispatch.Handler = NULL;
-                GpeEventInfo->Flags &= ~ACPI_GPE_DISPATCH_MASK;
-            }
-        }
-    }
-
-    return_ACPI_STATUS (AE_OK);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvMatchGpeMethod
- *
- * PARAMETERS:  Callback from WalkNamespace
- *
- * RETURN:      Status
- *
- * DESCRIPTION: Called from AcpiWalkNamespace. Expects each object to be a
- *              control method under the _GPE portion of the namespace.
- *              Extract the name and GPE type from the object, saving this
- *              information for quick lookup during GPE dispatch
- *
- *              The name of each GPE control method is of the form:
- *              "_Lxx" or "_Exx", where:
- *                  L      - means that the GPE is level triggered
- *                  E      - means that the GPE is edge triggered
- *                  xx     - is the GPE number [in HEX]
- *
- ******************************************************************************/
-
-static ACPI_STATUS
-AcpiEvMatchGpeMethod (
-    ACPI_HANDLE             ObjHandle,
-    UINT32                  Level,
-    void                    *ObjDesc,
-    void                    **ReturnValue)
-{
-    ACPI_NAMESPACE_NODE     *MethodNode = ACPI_CAST_PTR (ACPI_NAMESPACE_NODE, ObjHandle);
-    ACPI_GPE_BLOCK_INFO     *GpeBlock = (void *) ObjDesc;
-    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
-    UINT32                  GpeNumber;
-    char                    Name[ACPI_NAME_SIZE + 1];
-    UINT8                   Type;
-
-
-    ACPI_FUNCTION_TRACE (EvMatchGpeMethod);
-
-
-    /*
-     * Match and decode the _Lxx and _Exx GPE method names
-     *
-     * 1) Extract the method name and null terminate it
-     */
-    ACPI_MOVE_32_TO_32 (Name, &MethodNode->Name.Integer);
-    Name[ACPI_NAME_SIZE] = 0;
-
-    /* 2) Name must begin with an underscore */
-
-    if (Name[0] != '_')
-    {
-        return_ACPI_STATUS (AE_OK); /* Ignore this method */
-    }
-
-    /*
-     * 3) Edge/Level determination is based on the 2nd character
-     *    of the method name
-     *
-     * NOTE: Default GPE type is RUNTIME only. Later, if a _PRW object is
-     * found that points to this GPE, the ACPI_GPE_CAN_WAKE flag is set.
-     */
-    switch (Name[1])
-    {
-    case 'L':
-        Type = ACPI_GPE_LEVEL_TRIGGERED;
-        break;
-
-    case 'E':
-        Type = ACPI_GPE_EDGE_TRIGGERED;
-        break;
-
-    default:
-        /* Unknown method type, just ignore it */
-
-        ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
-            "Ignoring unknown GPE method type: %s "
-            "(name not of form _Lxx or _Exx)", Name));
-        return_ACPI_STATUS (AE_OK);
-    }
-
-    /* 4) The last two characters of the name are the hex GPE Number */
-
-    GpeNumber = ACPI_STRTOUL (&Name[2], NULL, 16);
-    if (GpeNumber == ACPI_UINT32_MAX)
-    {
-        /* Conversion failed; invalid method, just ignore it */
-
-        ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
-            "Could not extract GPE number from name: %s "
-            "(name is not of form _Lxx or _Exx)", Name));
-        return_ACPI_STATUS (AE_OK);
-    }
-
-    /* Ensure that we have a valid GPE number for this GPE block */
-
-    GpeEventInfo = AcpiEvLowGetGpeInfo (GpeNumber, GpeBlock);
-    if (!GpeEventInfo)
-    {
-        /*
-         * This GpeNumber is not valid for this GPE block, just ignore it.
-         * However, it may be valid for a different GPE block, since GPE0
-         * and GPE1 methods both appear under \_GPE.
-         */
-        return_ACPI_STATUS (AE_OK);
-    }
-
-    /*
-     * Add the GPE information from above to the GpeEventInfo block for
-     * use during dispatch of this GPE.
-     */
-    GpeEventInfo->Flags = (UINT8) (Type | ACPI_GPE_DISPATCH_METHOD);
-    GpeEventInfo->Dispatch.MethodNode = MethodNode;
-
-    ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
-        "Registered GPE method %s as GPE number 0x%.2X\n",
-        Name, GpeNumber));
-    return_ACPI_STATUS (AE_OK);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvMatchPrwAndGpe
- *
- * PARAMETERS:  Callback from WalkNamespace
- *
- * RETURN:      Status. NOTE: We ignore errors so that the _PRW walk is
- *              not aborted on a single _PRW failure.
- *
- * DESCRIPTION: Called from AcpiWalkNamespace. Expects each object to be a
- *              Device. Run the _PRW method. If present, extract the GPE
- *              number and mark the GPE as a CAN_WAKE GPE.
- *
- ******************************************************************************/
-
-static ACPI_STATUS
-AcpiEvMatchPrwAndGpe (
-    ACPI_HANDLE             ObjHandle,
-    UINT32                  Level,
-    void                    *Info,
-    void                    **ReturnValue)
-{
-    ACPI_GPE_WALK_INFO      *GpeInfo = (void *) Info;
-    ACPI_NAMESPACE_NODE     *GpeDevice;
-    ACPI_GPE_BLOCK_INFO     *GpeBlock;
-    ACPI_NAMESPACE_NODE     *TargetGpeDevice;
-    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
-    ACPI_OPERAND_OBJECT     *PkgDesc;
-    ACPI_OPERAND_OBJECT     *ObjDesc;
-    UINT32                  GpeNumber;
-    ACPI_STATUS             Status;
-
-
-    ACPI_FUNCTION_TRACE (EvMatchPrwAndGpe);
-
-
-    /* Check for a _PRW method under this device */
-
-    Status = AcpiUtEvaluateObject (ObjHandle, METHOD_NAME__PRW,
-                ACPI_BTYPE_PACKAGE, &PkgDesc);
-    if (ACPI_FAILURE (Status))
-    {
-        /* Ignore all errors from _PRW, we don't want to abort the walk */
-
-        return_ACPI_STATUS (AE_OK);
-    }
-
-    /* The returned _PRW package must have at least two elements */
-
-    if (PkgDesc->Package.Count < 2)
-    {
-        goto Cleanup;
-    }
-
-    /* Extract pointers from the input context */
-
-    GpeDevice = GpeInfo->GpeDevice;
-    GpeBlock = GpeInfo->GpeBlock;
-
-    /*
-     * The _PRW object must return a package, we are only interested in the
-     * first element
-     */
-    ObjDesc = PkgDesc->Package.Elements[0];
-
-    if (ObjDesc->Common.Type == ACPI_TYPE_INTEGER)
-    {
-        /* Use FADT-defined GPE device (from definition of _PRW) */
-
-        TargetGpeDevice = AcpiGbl_FadtGpeDevice;
-
-        /* Integer is the GPE number in the FADT described GPE blocks */
-
-        GpeNumber = (UINT32) ObjDesc->Integer.Value;
-    }
-    else if (ObjDesc->Common.Type == ACPI_TYPE_PACKAGE)
-    {
-        /* Package contains a GPE reference and GPE number within a GPE block */
-
-        if ((ObjDesc->Package.Count < 2) ||
-            ((ObjDesc->Package.Elements[0])->Common.Type !=
-                ACPI_TYPE_LOCAL_REFERENCE) ||
-            ((ObjDesc->Package.Elements[1])->Common.Type !=
-                ACPI_TYPE_INTEGER))
-        {
-            goto Cleanup;
-        }
-
-        /* Get GPE block reference and decode */
-
-        TargetGpeDevice = ObjDesc->Package.Elements[0]->Reference.Node;
-        GpeNumber = (UINT32) ObjDesc->Package.Elements[1]->Integer.Value;
-    }
-    else
-    {
-        /* Unknown type, just ignore it */
-
-        goto Cleanup;
-    }
-
-    /*
-     * Is this GPE within this block?
-     *
-     * TRUE if and only if these conditions are true:
-     *     1) The GPE devices match.
-     *     2) The GPE index(number) is within the range of the Gpe Block
-     *          associated with the GPE device.
-     */
-    if (GpeDevice != TargetGpeDevice)
-    {
-        goto Cleanup;
-    }
-
-    GpeEventInfo = AcpiEvLowGetGpeInfo (GpeNumber, GpeBlock);
-    if (GpeEventInfo)
-    {
-        /* This GPE can wake the system */
-
-        GpeEventInfo->Flags |= ACPI_GPE_CAN_WAKE;
-    }
-
-Cleanup:
-    AcpiUtRemoveReference (PkgDesc);
-    return_ACPI_STATUS (AE_OK);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvGetGpeXruptBlock
- *
- * PARAMETERS:  InterruptNumber      - Interrupt for a GPE block
- *
- * RETURN:      A GPE interrupt block
- *
- * DESCRIPTION: Get or Create a GPE interrupt block. There is one interrupt
- *              block per unique interrupt level used for GPEs. Should be
- *              called only when the GPE lists are semaphore locked and not
- *              subject to change.
- *
- ******************************************************************************/
-
-static ACPI_GPE_XRUPT_INFO *
-AcpiEvGetGpeXruptBlock (
-    UINT32                  InterruptNumber)
-{
-    ACPI_GPE_XRUPT_INFO     *NextGpeXrupt;
-    ACPI_GPE_XRUPT_INFO     *GpeXrupt;
-    ACPI_STATUS             Status;
-    ACPI_CPU_FLAGS          Flags;
-
-
-    ACPI_FUNCTION_TRACE (EvGetGpeXruptBlock);
-
-
-    /* No need for lock since we are not changing any list elements here */
-
-    NextGpeXrupt = AcpiGbl_GpeXruptListHead;
-    while (NextGpeXrupt)
-    {
-        if (NextGpeXrupt->InterruptNumber == InterruptNumber)
-        {
-            return_PTR (NextGpeXrupt);
-        }
-
-        NextGpeXrupt = NextGpeXrupt->Next;
-    }
-
-    /* Not found, must allocate a new xrupt descriptor */
-
-    GpeXrupt = ACPI_ALLOCATE_ZEROED (sizeof (ACPI_GPE_XRUPT_INFO));
-    if (!GpeXrupt)
-    {
-        return_PTR (NULL);
-    }
-
-    GpeXrupt->InterruptNumber = InterruptNumber;
-
-    /* Install new interrupt descriptor with spin lock */
-
-    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
-    if (AcpiGbl_GpeXruptListHead)
-    {
-        NextGpeXrupt = AcpiGbl_GpeXruptListHead;
-        while (NextGpeXrupt->Next)
-        {
-            NextGpeXrupt = NextGpeXrupt->Next;
-        }
-
-        NextGpeXrupt->Next = GpeXrupt;
-        GpeXrupt->Previous = NextGpeXrupt;
-    }
-    else
-    {
-        AcpiGbl_GpeXruptListHead = GpeXrupt;
-    }
-    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
-
-    /* Install new interrupt handler if not SCI_INT */
-
-    if (InterruptNumber != AcpiGbl_FADT.SciInterrupt)
-    {
-        Status = AcpiOsInstallInterruptHandler (InterruptNumber,
-                    AcpiEvGpeXruptHandler, GpeXrupt);
-        if (ACPI_FAILURE (Status))
-        {
-            ACPI_ERROR ((AE_INFO,
-                "Could not install GPE interrupt handler at level 0x%X",
-                InterruptNumber));
-            return_PTR (NULL);
-        }
-    }
-
-    return_PTR (GpeXrupt);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvDeleteGpeXrupt
- *
- * PARAMETERS:  GpeXrupt        - A GPE interrupt info block
- *
- * RETURN:      Status
- *
- * DESCRIPTION: Remove and free a GpeXrupt block. Remove an associated
- *              interrupt handler if not the SCI interrupt.
- *
- ******************************************************************************/
-
-static ACPI_STATUS
-AcpiEvDeleteGpeXrupt (
-    ACPI_GPE_XRUPT_INFO     *GpeXrupt)
-{
-    ACPI_STATUS             Status;
-    ACPI_CPU_FLAGS          Flags;
-
-
-    ACPI_FUNCTION_TRACE (EvDeleteGpeXrupt);
-
-
-    /* We never want to remove the SCI interrupt handler */
-
-    if (GpeXrupt->InterruptNumber == AcpiGbl_FADT.SciInterrupt)
-    {
-        GpeXrupt->GpeBlockListHead = NULL;
-        return_ACPI_STATUS (AE_OK);
-    }
-
-    /* Disable this interrupt */
-
-    Status = AcpiOsRemoveInterruptHandler (
-                GpeXrupt->InterruptNumber, AcpiEvGpeXruptHandler);
-    if (ACPI_FAILURE (Status))
-    {
-        return_ACPI_STATUS (Status);
-    }
-
-    /* Unlink the interrupt block with lock */
-
-    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
-    if (GpeXrupt->Previous)
-    {
-        GpeXrupt->Previous->Next = GpeXrupt->Next;
-    }
-    else
-    {
-        /* No previous, update list head */
-
-        AcpiGbl_GpeXruptListHead = GpeXrupt->Next;
-    }
-
-    if (GpeXrupt->Next)
-    {
-        GpeXrupt->Next->Previous = GpeXrupt->Previous;
-    }
-    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
-
-    /* Free the block */
-
-    ACPI_FREE (GpeXrupt);
-    return_ACPI_STATUS (AE_OK);
-}
-
-
 /*******************************************************************************
  *
  * FUNCTION:    AcpiEvInstallGpeBlock
@@ -1041,6 +444,7 @@ AcpiEvCreateGpeBlock (
 {
     ACPI_STATUS             Status;
     ACPI_GPE_BLOCK_INFO     *GpeBlock;
+    ACPI_GPE_WALK_INFO      WalkInfo;
 
 
     ACPI_FUNCTION_TRACE (EvCreateGpeBlock);
@@ -1089,11 +493,16 @@ AcpiEvCreateGpeBlock (
         return_ACPI_STATUS (Status);
     }
 
-    /* Find all GPE methods (_Lxx, _Exx) for this block */
+    /* Find all GPE methods (_Lxx or_Exx) for this block */
+
+    WalkInfo.GpeBlock = GpeBlock;
+    WalkInfo.GpeDevice = GpeDevice;
+    WalkInfo.EnableThisGpe = FALSE;
+    WalkInfo.ExecuteByOwnerId = FALSE;
 
     Status = AcpiNsWalkNamespace (ACPI_TYPE_METHOD, GpeDevice,
                 ACPI_UINT32_MAX, ACPI_NS_WALK_NO_UNLOCK,
-                AcpiEvMatchGpeMethod, NULL, GpeBlock, NULL);
+                AcpiEvMatchGpeMethod, NULL, &WalkInfo, NULL);
 
     /* Return the new block */
 
@@ -1139,7 +548,7 @@ AcpiEvInitializeGpeBlock (
 {
     ACPI_STATUS             Status;
     ACPI_GPE_EVENT_INFO     *GpeEventInfo;
-    ACPI_GPE_WALK_INFO      GpeInfo;
+    ACPI_GPE_WALK_INFO      WalkInfo;
     UINT32                  WakeGpeCount;
     UINT32                  GpeEnabledCount;
     UINT32                  GpeIndex;
@@ -1170,12 +579,13 @@ AcpiEvInitializeGpeBlock (
          * definition a wake GPE and will not be enabled while the machine
          * is running.
          */
-        GpeInfo.GpeBlock = GpeBlock;
-        GpeInfo.GpeDevice = GpeDevice;
+        WalkInfo.GpeBlock = GpeBlock;
+        WalkInfo.GpeDevice = GpeDevice;
+        WalkInfo.ExecuteByOwnerId = FALSE;
 
         Status = AcpiNsWalkNamespace (ACPI_TYPE_DEVICE, ACPI_ROOT_OBJECT,
                     ACPI_UINT32_MAX, ACPI_NS_WALK_UNLOCK,
-                    AcpiEvMatchPrwAndGpe, NULL, &GpeInfo, NULL);
+                    AcpiEvMatchPrwAndGpe, NULL, &WalkInfo, NULL);
         if (ACPI_FAILURE (Status))
         {
             ACPI_EXCEPTION ((AE_INFO, Status, "While executing _PRW methods"));
@@ -1238,165 +648,13 @@ AcpiEvInitializeGpeBlock (
         }
     }
 
-    ACPI_DEBUG_PRINT ((ACPI_DB_INIT,
-        "Found %u Wake, Enabled %u Runtime GPEs in this block\n",
-        WakeGpeCount, GpeEnabledCount));
-
-    return_ACPI_STATUS (AE_OK);
-}
-
-
-/*******************************************************************************
- *
- * FUNCTION:    AcpiEvGpeInitialize
- *
- * PARAMETERS:  None
- *
- * RETURN:      Status
- *
- * DESCRIPTION: Initialize the GPE data structures
- *
- ******************************************************************************/
-
-ACPI_STATUS
-AcpiEvGpeInitialize (
-    void)
-{
-    UINT32                  RegisterCount0 = 0;
-    UINT32                  RegisterCount1 = 0;
-    UINT32                  GpeNumberMax = 0;
-    ACPI_STATUS             Status;
-
-
-    ACPI_FUNCTION_TRACE (EvGpeInitialize);
-
-
-    Status = AcpiUtAcquireMutex (ACPI_MTX_NAMESPACE);
-    if (ACPI_FAILURE (Status))
+    if (GpeEnabledCount || WakeGpeCount)
     {
-        return_ACPI_STATUS (Status);
-    }
-
-    /*
-     * Initialize the GPE Block(s) defined in the FADT
-     *
-     * Why the GPE register block lengths are divided by 2:  From the ACPI
-     * Spec, section "General-Purpose Event Registers", we have:
-     *
-     * "Each register block contains two registers of equal length
-     *  GPEx_STS and GPEx_EN (where x is 0 or 1). The length of the
-     *  GPE0_STS and GPE0_EN registers is equal to half the GPE0_LEN
-     *  The length of the GPE1_STS and GPE1_EN registers is equal to
-     *  half the GPE1_LEN. If a generic register block is not supported
-     *  then its respective block pointer and block length values in the
-     *  FADT table contain zeros. The GPE0_LEN and GPE1_LEN do not need
-     *  to be the same size."
-     */
-
-    /*
-     * Determine the maximum GPE number for this machine.
-     *
-     * Note: both GPE0 and GPE1 are optional, and either can exist without
-     * the other.
-     *
-     * If EITHER the register length OR the block address are zero, then that
-     * particular block is not supported.
-     */
-    if (AcpiGbl_FADT.Gpe0BlockLength &&
-        AcpiGbl_FADT.XGpe0Block.Address)
-    {
-        /* GPE block 0 exists (has both length and address > 0) */
-
-        RegisterCount0 = (UINT16) (AcpiGbl_FADT.Gpe0BlockLength / 2);
-
-        GpeNumberMax = (RegisterCount0 * ACPI_GPE_REGISTER_WIDTH) - 1;
-
-        /* Install GPE Block 0 */
-
-        Status = AcpiEvCreateGpeBlock (AcpiGbl_FadtGpeDevice,
-                    &AcpiGbl_FADT.XGpe0Block, RegisterCount0, 0,
-                    AcpiGbl_FADT.SciInterrupt, &AcpiGbl_GpeFadtBlocks[0]);
-
-        if (ACPI_FAILURE (Status))
-        {
-            ACPI_EXCEPTION ((AE_INFO, Status,
-                "Could not create GPE Block 0"));
-        }
-    }
-
-    if (AcpiGbl_FADT.Gpe1BlockLength &&
-        AcpiGbl_FADT.XGpe1Block.Address)
-    {
-        /* GPE block 1 exists (has both length and address > 0) */
-
-        RegisterCount1 = (UINT16) (AcpiGbl_FADT.Gpe1BlockLength / 2);
-
-        /* Check for GPE0/GPE1 overlap (if both banks exist) */
-
-        if ((RegisterCount0) &&
-            (GpeNumberMax >= AcpiGbl_FADT.Gpe1Base))
-        {
-            ACPI_ERROR ((AE_INFO,
-                "GPE0 block (GPE 0 to %u) overlaps the GPE1 block "
-                "(GPE %u to %u) - Ignoring GPE1",
-                GpeNumberMax, AcpiGbl_FADT.Gpe1Base,
-                AcpiGbl_FADT.Gpe1Base +
-                ((RegisterCount1 * ACPI_GPE_REGISTER_WIDTH) - 1)));
-
-            /* Ignore GPE1 block by setting the register count to zero */
-
-            RegisterCount1 = 0;
-        }
-        else
-        {
-            /* Install GPE Block 1 */
-
-            Status = AcpiEvCreateGpeBlock (AcpiGbl_FadtGpeDevice,
-                        &AcpiGbl_FADT.XGpe1Block, RegisterCount1,
-                        AcpiGbl_FADT.Gpe1Base,
-                        AcpiGbl_FADT.SciInterrupt, &AcpiGbl_GpeFadtBlocks[1]);
-
-            if (ACPI_FAILURE (Status))
-            {
-                ACPI_EXCEPTION ((AE_INFO, Status,
-                    "Could not create GPE Block 1"));
-            }
-
-            /*
-             * GPE0 and GPE1 do not have to be contiguous in the GPE number
-             * space. However, GPE0 always starts at GPE number zero.
-             */
-            GpeNumberMax = AcpiGbl_FADT.Gpe1Base +
-                            ((RegisterCount1 * ACPI_GPE_REGISTER_WIDTH) - 1);
-        }
-    }
-
-    /* Exit if there are no GPE registers */
-
-    if ((RegisterCount0 + RegisterCount1) == 0)
-    {
-        /* GPEs are not required by ACPI, this is OK */
-
         ACPI_DEBUG_PRINT ((ACPI_DB_INIT,
-            "There are no GPE blocks defined in the FADT\n"));
-        Status = AE_OK;
-        goto Cleanup;
+            "Enabled %u Runtime GPEs, added %u Wake GPEs in this block\n",
+            GpeEnabledCount, WakeGpeCount));
     }
 
-    /* Check for Max GPE number out-of-range */
-
-    if (GpeNumberMax > ACPI_GPE_MAX)
-    {
-        ACPI_ERROR ((AE_INFO,
-            "Maximum GPE number from FADT is too large: 0x%X",
-            GpeNumberMax));
-        Status = AE_BAD_VALUE;
-        goto Cleanup;
-    }
-
-Cleanup:
-    (void) AcpiUtReleaseMutex (ACPI_MTX_NAMESPACE);
     return_ACPI_STATUS (AE_OK);
 }
 
-
diff --git a/events/evgpeinit.c b/events/evgpeinit.c
new file mode 100644
index 00000000000..1720250cc85
--- /dev/null
+++ b/events/evgpeinit.c
@@ -0,0 +1,763 @@
+/******************************************************************************
+ *
+ * Module Name: evgpeinit - System GPE initialization and update
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+#include "acpi.h"
+#include "accommon.h"
+#include "acevents.h"
+#include "acnamesp.h"
+#include "acinterp.h"
+
+#define _COMPONENT          ACPI_EVENTS
+        ACPI_MODULE_NAME    ("evgpeinit")
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvGpeInitialize
+ *
+ * PARAMETERS:  None
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Initialize the GPE data structures and the FADT GPE 0/1 blocks
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvGpeInitialize (
+    void)
+{
+    UINT32                  RegisterCount0 = 0;
+    UINT32                  RegisterCount1 = 0;
+    UINT32                  GpeNumberMax = 0;
+    ACPI_STATUS             Status;
+
+
+    ACPI_FUNCTION_TRACE (EvGpeInitialize);
+
+
+    Status = AcpiUtAcquireMutex (ACPI_MTX_NAMESPACE);
+    if (ACPI_FAILURE (Status))
+    {
+        return_ACPI_STATUS (Status);
+    }
+
+    /*
+     * Initialize the GPE Block(s) defined in the FADT
+     *
+     * Why the GPE register block lengths are divided by 2:  From the ACPI
+     * Spec, section "General-Purpose Event Registers", we have:
+     *
+     * "Each register block contains two registers of equal length
+     *  GPEx_STS and GPEx_EN (where x is 0 or 1). The length of the
+     *  GPE0_STS and GPE0_EN registers is equal to half the GPE0_LEN
+     *  The length of the GPE1_STS and GPE1_EN registers is equal to
+     *  half the GPE1_LEN. If a generic register block is not supported
+     *  then its respective block pointer and block length values in the
+     *  FADT table contain zeros. The GPE0_LEN and GPE1_LEN do not need
+     *  to be the same size."
+     */
+
+    /*
+     * Determine the maximum GPE number for this machine.
+     *
+     * Note: both GPE0 and GPE1 are optional, and either can exist without
+     * the other.
+     *
+     * If EITHER the register length OR the block address are zero, then that
+     * particular block is not supported.
+     */
+    if (AcpiGbl_FADT.Gpe0BlockLength &&
+        AcpiGbl_FADT.XGpe0Block.Address)
+    {
+        /* GPE block 0 exists (has both length and address > 0) */
+
+        RegisterCount0 = (UINT16) (AcpiGbl_FADT.Gpe0BlockLength / 2);
+
+        GpeNumberMax = (RegisterCount0 * ACPI_GPE_REGISTER_WIDTH) - 1;
+
+        /* Install GPE Block 0 */
+
+        Status = AcpiEvCreateGpeBlock (AcpiGbl_FadtGpeDevice,
+                    &AcpiGbl_FADT.XGpe0Block, RegisterCount0, 0,
+                    AcpiGbl_FADT.SciInterrupt, &AcpiGbl_GpeFadtBlocks[0]);
+
+        if (ACPI_FAILURE (Status))
+        {
+            ACPI_EXCEPTION ((AE_INFO, Status,
+                "Could not create GPE Block 0"));
+        }
+    }
+
+    if (AcpiGbl_FADT.Gpe1BlockLength &&
+        AcpiGbl_FADT.XGpe1Block.Address)
+    {
+        /* GPE block 1 exists (has both length and address > 0) */
+
+        RegisterCount1 = (UINT16) (AcpiGbl_FADT.Gpe1BlockLength / 2);
+
+        /* Check for GPE0/GPE1 overlap (if both banks exist) */
+
+        if ((RegisterCount0) &&
+            (GpeNumberMax >= AcpiGbl_FADT.Gpe1Base))
+        {
+            ACPI_ERROR ((AE_INFO,
+                "GPE0 block (GPE 0 to %u) overlaps the GPE1 block "
+                "(GPE %u to %u) - Ignoring GPE1",
+                GpeNumberMax, AcpiGbl_FADT.Gpe1Base,
+                AcpiGbl_FADT.Gpe1Base +
+                ((RegisterCount1 * ACPI_GPE_REGISTER_WIDTH) - 1)));
+
+            /* Ignore GPE1 block by setting the register count to zero */
+
+            RegisterCount1 = 0;
+        }
+        else
+        {
+            /* Install GPE Block 1 */
+
+            Status = AcpiEvCreateGpeBlock (AcpiGbl_FadtGpeDevice,
+                        &AcpiGbl_FADT.XGpe1Block, RegisterCount1,
+                        AcpiGbl_FADT.Gpe1Base,
+                        AcpiGbl_FADT.SciInterrupt, &AcpiGbl_GpeFadtBlocks[1]);
+
+            if (ACPI_FAILURE (Status))
+            {
+                ACPI_EXCEPTION ((AE_INFO, Status,
+                    "Could not create GPE Block 1"));
+            }
+
+            /*
+             * GPE0 and GPE1 do not have to be contiguous in the GPE number
+             * space. However, GPE0 always starts at GPE number zero.
+             */
+            GpeNumberMax = AcpiGbl_FADT.Gpe1Base +
+                            ((RegisterCount1 * ACPI_GPE_REGISTER_WIDTH) - 1);
+        }
+    }
+
+    /* Exit if there are no GPE registers */
+
+    if ((RegisterCount0 + RegisterCount1) == 0)
+    {
+        /* GPEs are not required by ACPI, this is OK */
+
+        ACPI_DEBUG_PRINT ((ACPI_DB_INIT,
+            "There are no GPE blocks defined in the FADT\n"));
+        Status = AE_OK;
+        goto Cleanup;
+    }
+
+    /* Check for Max GPE number out-of-range */
+
+    if (GpeNumberMax > ACPI_GPE_MAX)
+    {
+        ACPI_ERROR ((AE_INFO,
+            "Maximum GPE number from FADT is too large: 0x%X",
+            GpeNumberMax));
+        Status = AE_BAD_VALUE;
+        goto Cleanup;
+    }
+
+Cleanup:
+    (void) AcpiUtReleaseMutex (ACPI_MTX_NAMESPACE);
+    return_ACPI_STATUS (AE_OK);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvUpdateGpes
+ *
+ * PARAMETERS:  TableOwnerId        - ID of the newly-loaded ACPI table
+ *
+ * RETURN:      None
+ *
+ * DESCRIPTION: Check for new GPE methods (_Lxx/_Exx) made available as a
+ *              result of a Load() or LoadTable() operation. If new GPE
+ *              methods have been installed, register the new methods and
+ *              enable and runtime GPEs that are associated with them. Also,
+ *              run any newly loaded _PRW methods in order to discover any
+ *              new CAN_WAKE GPEs.
+ *
+ ******************************************************************************/
+
+void
+AcpiEvUpdateGpes (
+    ACPI_OWNER_ID           TableOwnerId)
+{
+    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo;
+    ACPI_GPE_BLOCK_INFO     *GpeBlock;
+    ACPI_GPE_WALK_INFO      WalkInfo;
+    ACPI_STATUS             Status = AE_OK;
+    UINT32                  NewWakeGpeCount = 0;
+
+
+    /* We will examine only _PRW/_Lxx/_Exx methods owned by this table */
+
+    WalkInfo.OwnerId = TableOwnerId;
+    WalkInfo.ExecuteByOwnerId = TRUE;
+    WalkInfo.Count = 0;
+
+    if (AcpiGbl_LeaveWakeGpesDisabled)
+    {
+        /*
+         * 1) Run any newly-loaded _PRW methods to find any GPEs that
+         * can now be marked as CAN_WAKE GPEs. Note: We must run the
+         * _PRW methods before we process the _Lxx/_Exx methods because
+         * we will enable all runtime GPEs associated with the new
+         * _Lxx/_Exx methods at the time we process those methods.
+         *
+         * Unlock interpreter so that we can run the _PRW methods.
+         */
+        WalkInfo.GpeBlock = NULL;
+        WalkInfo.GpeDevice = NULL;
+
+        AcpiExExitInterpreter ();
+
+        Status = AcpiNsWalkNamespace (ACPI_TYPE_DEVICE, ACPI_ROOT_OBJECT,
+                    ACPI_UINT32_MAX, ACPI_NS_WALK_NO_UNLOCK,
+                    AcpiEvMatchPrwAndGpe, NULL, &WalkInfo, NULL);
+        if (ACPI_FAILURE (Status))
+        {
+            ACPI_EXCEPTION ((AE_INFO, Status,
+                "While executing _PRW methods"));
+        }
+
+        AcpiExEnterInterpreter ();
+        NewWakeGpeCount = WalkInfo.Count;
+    }
+
+    /*
+     * 2) Find any _Lxx/_Exx GPE methods that have just been loaded.
+     *
+     * Any GPEs that correspond to new _Lxx/_Exx methods and are not
+     * marked as CAN_WAKE are immediately enabled.
+     *
+     * Examine the namespace underneath each GpeDevice within the
+     * GpeBlock lists.
+     */
+    Status = AcpiUtAcquireMutex (ACPI_MTX_EVENTS);
+    if (ACPI_FAILURE (Status))
+    {
+        return;
+    }
+
+    WalkInfo.Count = 0;
+    WalkInfo.EnableThisGpe = TRUE;
+
+    /* Walk the interrupt level descriptor list */
+
+    GpeXruptInfo = AcpiGbl_GpeXruptListHead;
+    while (GpeXruptInfo)
+    {
+        /* Walk all Gpe Blocks attached to this interrupt level */
+
+        GpeBlock = GpeXruptInfo->GpeBlockListHead;
+        while (GpeBlock)
+        {
+            WalkInfo.GpeBlock = GpeBlock;
+            WalkInfo.GpeDevice = GpeBlock->Node;
+
+            Status = AcpiNsWalkNamespace (ACPI_TYPE_METHOD,
+                        WalkInfo.GpeDevice, ACPI_UINT32_MAX,
+                        ACPI_NS_WALK_NO_UNLOCK, AcpiEvMatchGpeMethod,
+                        NULL, &WalkInfo, NULL);
+            if (ACPI_FAILURE (Status))
+            {
+                ACPI_EXCEPTION ((AE_INFO, Status,
+                    "While decoding _Lxx/_Exx methods"));
+            }
+
+            GpeBlock = GpeBlock->Next;
+        }
+
+        GpeXruptInfo = GpeXruptInfo->Next;
+    }
+
+    if (WalkInfo.Count || NewWakeGpeCount)
+    {
+        ACPI_INFO ((AE_INFO,
+            "Enabled %u new runtime GPEs, added %u new wakeup GPEs",
+            WalkInfo.Count, NewWakeGpeCount));
+    }
+
+    (void) AcpiUtReleaseMutex (ACPI_MTX_EVENTS);
+    return;
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvMatchGpeMethod
+ *
+ * PARAMETERS:  Callback from WalkNamespace
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Called from AcpiWalkNamespace. Expects each object to be a
+ *              control method under the _GPE portion of the namespace.
+ *              Extract the name and GPE type from the object, saving this
+ *              information for quick lookup during GPE dispatch. Allows a
+ *              per-OwnerId evaluation if ExecuteByOwnerId is TRUE in the
+ *              WalkInfo parameter block.
+ *
+ *              The name of each GPE control method is of the form:
+ *              "_Lxx" or "_Exx", where:
+ *                  L      - means that the GPE is level triggered
+ *                  E      - means that the GPE is edge triggered
+ *                  xx     - is the GPE number [in HEX]
+ *
+ * If WalkInfo->ExecuteByOwnerId is TRUE, we only execute examine GPE methods
+ *    with that owner.
+ * If WalkInfo->EnableThisGpe is TRUE, the GPE that is referred to by a GPE
+ *    method is immediately enabled (Used for Load/LoadTable operators)
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvMatchGpeMethod (
+    ACPI_HANDLE             ObjHandle,
+    UINT32                  Level,
+    void                    *Context,
+    void                    **ReturnValue)
+{
+    ACPI_NAMESPACE_NODE     *MethodNode = ACPI_CAST_PTR (ACPI_NAMESPACE_NODE, ObjHandle);
+    ACPI_GPE_WALK_INFO      *WalkInfo = ACPI_CAST_PTR (ACPI_GPE_WALK_INFO, Context);
+    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
+    ACPI_NAMESPACE_NODE     *GpeDevice;
+    ACPI_STATUS             Status;
+    UINT32                  GpeNumber;
+    char                    Name[ACPI_NAME_SIZE + 1];
+    UINT8                   Type;
+
+
+    ACPI_FUNCTION_TRACE (EvMatchGpeMethod);
+
+
+    /* Check if requested OwnerId matches this OwnerId */
+
+    if ((WalkInfo->ExecuteByOwnerId) &&
+        (MethodNode->OwnerId != WalkInfo->OwnerId))
+    {
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /*
+     * Match and decode the _Lxx and _Exx GPE method names
+     *
+     * 1) Extract the method name and null terminate it
+     */
+    ACPI_MOVE_32_TO_32 (Name, &MethodNode->Name.Integer);
+    Name[ACPI_NAME_SIZE] = 0;
+
+    /* 2) Name must begin with an underscore */
+
+    if (Name[0] != '_')
+    {
+        return_ACPI_STATUS (AE_OK); /* Ignore this method */
+    }
+
+    /*
+     * 3) Edge/Level determination is based on the 2nd character
+     *    of the method name
+     *
+     * NOTE: Default GPE type is RUNTIME only. Later, if a _PRW object is
+     * found that points to this GPE, the ACPI_GPE_CAN_WAKE flag is set.
+     */
+    switch (Name[1])
+    {
+    case 'L':
+        Type = ACPI_GPE_LEVEL_TRIGGERED;
+        break;
+
+    case 'E':
+        Type = ACPI_GPE_EDGE_TRIGGERED;
+        break;
+
+    default:
+        /* Unknown method type, just ignore it */
+
+        ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
+            "Ignoring unknown GPE method type: %s "
+            "(name not of form _Lxx or _Exx)", Name));
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* 4) The last two characters of the name are the hex GPE Number */
+
+    GpeNumber = ACPI_STRTOUL (&Name[2], NULL, 16);
+    if (GpeNumber == ACPI_UINT32_MAX)
+    {
+        /* Conversion failed; invalid method, just ignore it */
+
+        ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
+            "Could not extract GPE number from name: %s "
+            "(name is not of form _Lxx or _Exx)", Name));
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* Ensure that we have a valid GPE number for this GPE block */
+
+    GpeEventInfo = AcpiEvLowGetGpeInfo (GpeNumber, WalkInfo->GpeBlock);
+    if (!GpeEventInfo)
+    {
+        /*
+         * This GpeNumber is not valid for this GPE block, just ignore it.
+         * However, it may be valid for a different GPE block, since GPE0
+         * and GPE1 methods both appear under \_GPE.
+         */
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    if ((GpeEventInfo->Flags & ACPI_GPE_DISPATCH_MASK) ==
+            ACPI_GPE_DISPATCH_HANDLER)
+    {
+        /* If there is already a handler, ignore this GPE method */
+
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    if ((GpeEventInfo->Flags & ACPI_GPE_DISPATCH_MASK) ==
+            ACPI_GPE_DISPATCH_METHOD)
+    {
+        /*
+         * If there is already a method, ignore this method. But check
+         * for a type mismatch (if both the _Lxx AND _Exx exist)
+         */
+        if (Type != (GpeEventInfo->Flags & ACPI_GPE_XRUPT_TYPE_MASK))
+        {
+            ACPI_ERROR ((AE_INFO,
+                "For GPE 0x%.2X, found both _L%2.2X and _E%2.2X methods",
+                GpeNumber, GpeNumber, GpeNumber));
+        }
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /*
+     * Add the GPE information from above to the GpeEventInfo block for
+     * use during dispatch of this GPE.
+     */
+    GpeEventInfo->Flags |= (UINT8) (Type | ACPI_GPE_DISPATCH_METHOD);
+    GpeEventInfo->Dispatch.MethodNode = MethodNode;
+
+    /*
+     * Enable this GPE if requested. This only happens when during the
+     * execution of a Load or LoadTable operator. We have found a new
+     * GPE method and want to immediately enable the GPE if it is a
+     * runtime GPE.
+     */
+    if (WalkInfo->EnableThisGpe)
+    {
+        /* Ignore GPEs that can wake the system */
+
+        if (!(GpeEventInfo->Flags & ACPI_GPE_CAN_WAKE) ||
+            !AcpiGbl_LeaveWakeGpesDisabled)
+        {
+            WalkInfo->Count++;
+            GpeDevice = WalkInfo->GpeDevice;
+
+            if (GpeDevice == AcpiGbl_FadtGpeDevice)
+            {
+                GpeDevice = NULL;
+            }
+
+            Status = AcpiEnableGpe (GpeDevice, GpeNumber,
+                        ACPI_GPE_TYPE_RUNTIME);
+            if (ACPI_FAILURE (Status))
+            {
+                ACPI_EXCEPTION ((AE_INFO, Status,
+                    "Could not enable GPE 0x%02X", GpeNumber));
+            }
+        }
+    }
+
+    ACPI_DEBUG_PRINT ((ACPI_DB_LOAD,
+        "Registered GPE method %s as GPE number 0x%.2X\n",
+        Name, GpeNumber));
+    return_ACPI_STATUS (AE_OK);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvMatchPrwAndGpe
+ *
+ * PARAMETERS:  Callback from WalkNamespace
+ *
+ * RETURN:      Status. NOTE: We ignore errors so that the _PRW walk is
+ *              not aborted on a single _PRW failure.
+ *
+ * DESCRIPTION: Called from AcpiWalkNamespace. Expects each object to be a
+ *              Device. Run the _PRW method. If present, extract the GPE
+ *              number and mark the GPE as a CAN_WAKE GPE. Allows a
+ *              per-OwnerId execution if ExecuteByOwnerId is TRUE in the
+ *              WalkInfo parameter block.
+ *
+ * If WalkInfo->ExecuteByOwnerId is TRUE, we only execute _PRWs with that
+ *    owner.
+ * If WalkInfo->GpeDevice is NULL, we execute every _PRW found. Otherwise,
+ *    we only execute _PRWs that refer to the input GpeDevice.
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvMatchPrwAndGpe (
+    ACPI_HANDLE             ObjHandle,
+    UINT32                  Level,
+    void                    *Context,
+    void                    **ReturnValue)
+{
+    ACPI_GPE_WALK_INFO      *WalkInfo = ACPI_CAST_PTR (ACPI_GPE_WALK_INFO, Context);
+    ACPI_NAMESPACE_NODE     *GpeDevice;
+    ACPI_GPE_BLOCK_INFO     *GpeBlock;
+    ACPI_NAMESPACE_NODE     *TargetGpeDevice;
+    ACPI_NAMESPACE_NODE     *PrwNode;
+    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
+    ACPI_OPERAND_OBJECT     *PkgDesc;
+    ACPI_OPERAND_OBJECT     *ObjDesc;
+    UINT32                  GpeNumber;
+    ACPI_STATUS             Status;
+
+
+    ACPI_FUNCTION_TRACE (EvMatchPrwAndGpe);
+
+
+    /* Check for a _PRW method under this device */
+
+    Status = AcpiNsGetNode (ObjHandle, METHOD_NAME__PRW,
+                ACPI_NS_NO_UPSEARCH, &PrwNode);
+    if (ACPI_FAILURE (Status))
+    {
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* Check if requested OwnerId matches this OwnerId */
+
+    if ((WalkInfo->ExecuteByOwnerId) &&
+        (PrwNode->OwnerId != WalkInfo->OwnerId))
+    {
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* Execute the _PRW */
+
+    Status = AcpiUtEvaluateObject (PrwNode, NULL,
+                ACPI_BTYPE_PACKAGE, &PkgDesc);
+    if (ACPI_FAILURE (Status))
+    {
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* The returned _PRW package must have at least two elements */
+
+    if (PkgDesc->Package.Count < 2)
+    {
+        goto Cleanup;
+    }
+
+    /* Extract pointers from the input context */
+
+    GpeDevice = WalkInfo->GpeDevice;
+    GpeBlock = WalkInfo->GpeBlock;
+
+    /*
+     * The _PRW object must return a package, we are only interested
+     * in the first element
+     */
+    ObjDesc = PkgDesc->Package.Elements[0];
+
+    if (ObjDesc->Common.Type == ACPI_TYPE_INTEGER)
+    {
+        /* Use FADT-defined GPE device (from definition of _PRW) */
+
+        TargetGpeDevice = NULL;
+        if (GpeDevice)
+        {
+            TargetGpeDevice = AcpiGbl_FadtGpeDevice;
+        }
+
+        /* Integer is the GPE number in the FADT described GPE blocks */
+
+        GpeNumber = (UINT32) ObjDesc->Integer.Value;
+    }
+    else if (ObjDesc->Common.Type == ACPI_TYPE_PACKAGE)
+    {
+        /* Package contains a GPE reference and GPE number within a GPE block */
+
+        if ((ObjDesc->Package.Count < 2) ||
+            ((ObjDesc->Package.Elements[0])->Common.Type !=
+                ACPI_TYPE_LOCAL_REFERENCE) ||
+            ((ObjDesc->Package.Elements[1])->Common.Type !=
+                ACPI_TYPE_INTEGER))
+        {
+            goto Cleanup;
+        }
+
+        /* Get GPE block reference and decode */
+
+        TargetGpeDevice = ObjDesc->Package.Elements[0]->Reference.Node;
+        GpeNumber = (UINT32) ObjDesc->Package.Elements[1]->Integer.Value;
+    }
+    else
+    {
+        /* Unknown type, just ignore it */
+
+        goto Cleanup;
+    }
+
+    /* Get the GpeEventInfo for this GPE */
+
+    if (GpeDevice)
+    {
+        /*
+         * Is this GPE within this block?
+         *
+         * TRUE if and only if these conditions are true:
+         *     1) The GPE devices match.
+         *     2) The GPE index(number) is within the range of the Gpe Block
+         *          associated with the GPE device.
+         */
+        if (GpeDevice != TargetGpeDevice)
+        {
+            goto Cleanup;
+        }
+
+        GpeEventInfo = AcpiEvLowGetGpeInfo (GpeNumber, GpeBlock);
+    }
+    else
+    {
+        /* GpeDevice is NULL, just match the TargetDevice and GpeNumber */
+
+        GpeEventInfo = AcpiEvGetGpeEventInfo (TargetGpeDevice, GpeNumber);
+    }
+
+    if (GpeEventInfo)
+    {
+        if (!(GpeEventInfo->Flags & ACPI_GPE_CAN_WAKE))
+        {
+            /* This GPE can wake the system */
+
+            GpeEventInfo->Flags |= ACPI_GPE_CAN_WAKE;
+            WalkInfo->Count++;
+        }
+    }
+
+Cleanup:
+    AcpiUtRemoveReference (PkgDesc);
+    return_ACPI_STATUS (AE_OK);
+}
+
diff --git a/events/evgpeutil.c b/events/evgpeutil.c
new file mode 100644
index 00000000000..2ac0ad44809
--- /dev/null
+++ b/events/evgpeutil.c
@@ -0,0 +1,452 @@
+/******************************************************************************
+ *
+ * Module Name: evgpeutil - GPE utilities
+ *
+ *****************************************************************************/
+
+/******************************************************************************
+ *
+ * 1. Copyright Notice
+ *
+ * Some or all of this work - Copyright (c) 1999 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * 2. License
+ *
+ * 2.1. This is your license from Intel Corp. under its intellectual property
+ * rights.  You may have additional license terms from the party that provided
+ * you this software, covering your right to use that party's intellectual
+ * property rights.
+ *
+ * 2.2. Intel grants, free of charge, to any person ("Licensee") obtaining a
+ * copy of the source code appearing in this file ("Covered Code") an
+ * irrevocable, perpetual, worldwide license under Intel's copyrights in the
+ * base code distributed originally by Intel ("Original Intel Code") to copy,
+ * make derivatives, distribute, use and display any portion of the Covered
+ * Code in any form, with the right to sublicense such rights; and
+ *
+ * 2.3. Intel grants Licensee a non-exclusive and non-transferable patent
+ * license (with the right to sublicense), under only those claims of Intel
+ * patents that are infringed by the Original Intel Code, to make, use, sell,
+ * offer to sell, and import the Covered Code and derivative works thereof
+ * solely to the minimum extent necessary to exercise the above copyright
+ * license, and in no event shall the patent license extend to any additions
+ * to or modifications of the Original Intel Code.  No other license or right
+ * is granted directly or by implication, estoppel or otherwise;
+ *
+ * The above copyright and patent license is granted only if the following
+ * conditions are met:
+ *
+ * 3. Conditions
+ *
+ * 3.1. Redistribution of Source with Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification with rights to further distribute source must include
+ * the above Copyright Notice, the above License, this list of Conditions,
+ * and the following Disclaimer and Export Compliance provision.  In addition,
+ * Licensee must cause all Covered Code to which Licensee contributes to
+ * contain a file documenting the changes Licensee made to create that Covered
+ * Code and the date of any change.  Licensee must include in that file the
+ * documentation of any changes made by any predecessor Licensee.  Licensee
+ * must include a prominent statement that the modification is derived,
+ * directly or indirectly, from Original Intel Code.
+ *
+ * 3.2. Redistribution of Source with no Rights to Further Distribute Source.
+ * Redistribution of source code of any substantial portion of the Covered
+ * Code or modification without rights to further distribute source must
+ * include the following Disclaimer and Export Compliance provision in the
+ * documentation and/or other materials provided with distribution.  In
+ * addition, Licensee may not authorize further sublicense of source of any
+ * portion of the Covered Code, and must include terms to the effect that the
+ * license from Licensee to its licensee is limited to the intellectual
+ * property embodied in the software Licensee provides to its licensee, and
+ * not to intellectual property embodied in modifications its licensee may
+ * make.
+ *
+ * 3.3. Redistribution of Executable. Redistribution in executable form of any
+ * substantial portion of the Covered Code or modification must reproduce the
+ * above Copyright Notice, and the following Disclaimer and Export Compliance
+ * provision in the documentation and/or other materials provided with the
+ * distribution.
+ *
+ * 3.4. Intel retains all right, title, and interest in and to the Original
+ * Intel Code.
+ *
+ * 3.5. Neither the name Intel nor any other trademark owned or controlled by
+ * Intel shall be used in advertising or otherwise to promote the sale, use or
+ * other dealings in products derived from or relating to the Covered Code
+ * without prior written authorization from Intel.
+ *
+ * 4. Disclaimer and Export Compliance
+ *
+ * 4.1. INTEL MAKES NO WARRANTY OF ANY KIND REGARDING ANY SOFTWARE PROVIDED
+ * HERE.  ANY SOFTWARE ORIGINATING FROM INTEL OR DERIVED FROM INTEL SOFTWARE
+ * IS PROVIDED "AS IS," AND INTEL WILL NOT PROVIDE ANY SUPPORT,  ASSISTANCE,
+ * INSTALLATION, TRAINING OR OTHER SERVICES.  INTEL WILL NOT PROVIDE ANY
+ * UPDATES, ENHANCEMENTS OR EXTENSIONS.  INTEL SPECIFICALLY DISCLAIMS ANY
+ * IMPLIED WARRANTIES OF MERCHANTABILITY, NONINFRINGEMENT AND FITNESS FOR A
+ * PARTICULAR PURPOSE.
+ *
+ * 4.2. IN NO EVENT SHALL INTEL HAVE ANY LIABILITY TO LICENSEE, ITS LICENSEES
+ * OR ANY OTHER THIRD PARTY, FOR ANY LOST PROFITS, LOST DATA, LOSS OF USE OR
+ * COSTS OF PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES, OR FOR ANY INDIRECT,
+ * SPECIAL OR CONSEQUENTIAL DAMAGES ARISING OUT OF THIS AGREEMENT, UNDER ANY
+ * CAUSE OF ACTION OR THEORY OF LIABILITY, AND IRRESPECTIVE OF WHETHER INTEL
+ * HAS ADVANCE NOTICE OF THE POSSIBILITY OF SUCH DAMAGES.  THESE LIMITATIONS
+ * SHALL APPLY NOTWITHSTANDING THE FAILURE OF THE ESSENTIAL PURPOSE OF ANY
+ * LIMITED REMEDY.
+ *
+ * 4.3. Licensee shall not export, either directly or indirectly, any of this
+ * software or system incorporating such software without first obtaining any
+ * required license or other approval from the U. S. Department of Commerce or
+ * any other agency or department of the United States Government.  In the
+ * event Licensee exports any such software from the United States or
+ * re-exports any such software from a foreign destination, Licensee shall
+ * ensure that the distribution and export/re-export of the software is in
+ * compliance with all laws, regulations, orders, or other restrictions of the
+ * U.S. Export Administration Regulations. Licensee agrees that neither it nor
+ * any of its subsidiaries will export/re-export any technical data, process,
+ * software, or service, directly or indirectly, to any country for which the
+ * United States government or any agency thereof requires an export license,
+ * other governmental approval, or letter of assurance, without first obtaining
+ * such license, approval or letter.
+ *
+ *****************************************************************************/
+
+
+
+#include "acpi.h"
+#include "accommon.h"
+#include "acevents.h"
+
+#define _COMPONENT          ACPI_EVENTS
+        ACPI_MODULE_NAME    ("evgpeutil")
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvWalkGpeList
+ *
+ * PARAMETERS:  GpeWalkCallback     - Routine called for each GPE block
+ *              Context             - Value passed to callback
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Walk the GPE lists.
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvWalkGpeList (
+    ACPI_GPE_CALLBACK       GpeWalkCallback,
+    void                    *Context)
+{
+    ACPI_GPE_BLOCK_INFO     *GpeBlock;
+    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo;
+    ACPI_STATUS             Status = AE_OK;
+    ACPI_CPU_FLAGS          Flags;
+
+
+    ACPI_FUNCTION_TRACE (EvWalkGpeList);
+
+
+    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
+
+    /* Walk the interrupt level descriptor list */
+
+    GpeXruptInfo = AcpiGbl_GpeXruptListHead;
+    while (GpeXruptInfo)
+    {
+        /* Walk all Gpe Blocks attached to this interrupt level */
+
+        GpeBlock = GpeXruptInfo->GpeBlockListHead;
+        while (GpeBlock)
+        {
+            /* One callback per GPE block */
+
+            Status = GpeWalkCallback (GpeXruptInfo, GpeBlock, Context);
+            if (ACPI_FAILURE (Status))
+            {
+                if (Status == AE_CTRL_END) /* Callback abort */
+                {
+                    Status = AE_OK;
+                }
+                goto UnlockAndExit;
+            }
+
+            GpeBlock = GpeBlock->Next;
+        }
+
+        GpeXruptInfo = GpeXruptInfo->Next;
+    }
+
+UnlockAndExit:
+    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
+    return_ACPI_STATUS (Status);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvValidGpeEvent
+ *
+ * PARAMETERS:  GpeEventInfo                - Info for this GPE
+ *
+ * RETURN:      TRUE if the GpeEvent is valid
+ *
+ * DESCRIPTION: Validate a GPE event. DO NOT CALL FROM INTERRUPT LEVEL.
+ *              Should be called only when the GPE lists are semaphore locked
+ *              and not subject to change.
+ *
+ ******************************************************************************/
+
+BOOLEAN
+AcpiEvValidGpeEvent (
+    ACPI_GPE_EVENT_INFO     *GpeEventInfo)
+{
+    ACPI_GPE_XRUPT_INFO     *GpeXruptBlock;
+    ACPI_GPE_BLOCK_INFO     *GpeBlock;
+
+
+    ACPI_FUNCTION_ENTRY ();
+
+
+    /* No need for spin lock since we are not changing any list elements */
+
+    /* Walk the GPE interrupt levels */
+
+    GpeXruptBlock = AcpiGbl_GpeXruptListHead;
+    while (GpeXruptBlock)
+    {
+        GpeBlock = GpeXruptBlock->GpeBlockListHead;
+
+        /* Walk the GPE blocks on this interrupt level */
+
+        while (GpeBlock)
+        {
+            if ((&GpeBlock->EventInfo[0] <= GpeEventInfo) &&
+                (&GpeBlock->EventInfo[GpeBlock->GpeCount] > GpeEventInfo))
+            {
+                return (TRUE);
+            }
+
+            GpeBlock = GpeBlock->Next;
+        }
+
+        GpeXruptBlock = GpeXruptBlock->Next;
+    }
+
+    return (FALSE);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvGetGpeXruptBlock
+ *
+ * PARAMETERS:  InterruptNumber      - Interrupt for a GPE block
+ *
+ * RETURN:      A GPE interrupt block
+ *
+ * DESCRIPTION: Get or Create a GPE interrupt block. There is one interrupt
+ *              block per unique interrupt level used for GPEs. Should be
+ *              called only when the GPE lists are semaphore locked and not
+ *              subject to change.
+ *
+ ******************************************************************************/
+
+ACPI_GPE_XRUPT_INFO *
+AcpiEvGetGpeXruptBlock (
+    UINT32                  InterruptNumber)
+{
+    ACPI_GPE_XRUPT_INFO     *NextGpeXrupt;
+    ACPI_GPE_XRUPT_INFO     *GpeXrupt;
+    ACPI_STATUS             Status;
+    ACPI_CPU_FLAGS          Flags;
+
+
+    ACPI_FUNCTION_TRACE (EvGetGpeXruptBlock);
+
+
+    /* No need for lock since we are not changing any list elements here */
+
+    NextGpeXrupt = AcpiGbl_GpeXruptListHead;
+    while (NextGpeXrupt)
+    {
+        if (NextGpeXrupt->InterruptNumber == InterruptNumber)
+        {
+            return_PTR (NextGpeXrupt);
+        }
+
+        NextGpeXrupt = NextGpeXrupt->Next;
+    }
+
+    /* Not found, must allocate a new xrupt descriptor */
+
+    GpeXrupt = ACPI_ALLOCATE_ZEROED (sizeof (ACPI_GPE_XRUPT_INFO));
+    if (!GpeXrupt)
+    {
+        return_PTR (NULL);
+    }
+
+    GpeXrupt->InterruptNumber = InterruptNumber;
+
+    /* Install new interrupt descriptor with spin lock */
+
+    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
+    if (AcpiGbl_GpeXruptListHead)
+    {
+        NextGpeXrupt = AcpiGbl_GpeXruptListHead;
+        while (NextGpeXrupt->Next)
+        {
+            NextGpeXrupt = NextGpeXrupt->Next;
+        }
+
+        NextGpeXrupt->Next = GpeXrupt;
+        GpeXrupt->Previous = NextGpeXrupt;
+    }
+    else
+    {
+        AcpiGbl_GpeXruptListHead = GpeXrupt;
+    }
+    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
+
+    /* Install new interrupt handler if not SCI_INT */
+
+    if (InterruptNumber != AcpiGbl_FADT.SciInterrupt)
+    {
+        Status = AcpiOsInstallInterruptHandler (InterruptNumber,
+                    AcpiEvGpeXruptHandler, GpeXrupt);
+        if (ACPI_FAILURE (Status))
+        {
+            ACPI_ERROR ((AE_INFO,
+                "Could not install GPE interrupt handler at level 0x%X",
+                InterruptNumber));
+            return_PTR (NULL);
+        }
+    }
+
+    return_PTR (GpeXrupt);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvDeleteGpeXrupt
+ *
+ * PARAMETERS:  GpeXrupt        - A GPE interrupt info block
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Remove and free a GpeXrupt block. Remove an associated
+ *              interrupt handler if not the SCI interrupt.
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvDeleteGpeXrupt (
+    ACPI_GPE_XRUPT_INFO     *GpeXrupt)
+{
+    ACPI_STATUS             Status;
+    ACPI_CPU_FLAGS          Flags;
+
+
+    ACPI_FUNCTION_TRACE (EvDeleteGpeXrupt);
+
+
+    /* We never want to remove the SCI interrupt handler */
+
+    if (GpeXrupt->InterruptNumber == AcpiGbl_FADT.SciInterrupt)
+    {
+        GpeXrupt->GpeBlockListHead = NULL;
+        return_ACPI_STATUS (AE_OK);
+    }
+
+    /* Disable this interrupt */
+
+    Status = AcpiOsRemoveInterruptHandler (
+                GpeXrupt->InterruptNumber, AcpiEvGpeXruptHandler);
+    if (ACPI_FAILURE (Status))
+    {
+        return_ACPI_STATUS (Status);
+    }
+
+    /* Unlink the interrupt block with lock */
+
+    Flags = AcpiOsAcquireLock (AcpiGbl_GpeLock);
+    if (GpeXrupt->Previous)
+    {
+        GpeXrupt->Previous->Next = GpeXrupt->Next;
+    }
+    else
+    {
+        /* No previous, update list head */
+
+        AcpiGbl_GpeXruptListHead = GpeXrupt->Next;
+    }
+
+    if (GpeXrupt->Next)
+    {
+        GpeXrupt->Next->Previous = GpeXrupt->Previous;
+    }
+    AcpiOsReleaseLock (AcpiGbl_GpeLock, Flags);
+
+    /* Free the block */
+
+    ACPI_FREE (GpeXrupt);
+    return_ACPI_STATUS (AE_OK);
+}
+
+
+/*******************************************************************************
+ *
+ * FUNCTION:    AcpiEvDeleteGpeHandlers
+ *
+ * PARAMETERS:  GpeXruptInfo        - GPE Interrupt info
+ *              GpeBlock            - Gpe Block info
+ *
+ * RETURN:      Status
+ *
+ * DESCRIPTION: Delete all Handler objects found in the GPE data structs.
+ *              Used only prior to termination.
+ *
+ ******************************************************************************/
+
+ACPI_STATUS
+AcpiEvDeleteGpeHandlers (
+    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo,
+    ACPI_GPE_BLOCK_INFO     *GpeBlock,
+    void                    *Context)
+{
+    ACPI_GPE_EVENT_INFO     *GpeEventInfo;
+    UINT32                  i;
+    UINT32                  j;
+
+
+    ACPI_FUNCTION_TRACE (EvDeleteGpeHandlers);
+
+
+    /* Examine each GPE Register within the block */
+
+    for (i = 0; i < GpeBlock->RegisterCount; i++)
+    {
+        /* Now look at the individual GPEs in this byte register */
+
+        for (j = 0; j < ACPI_GPE_REGISTER_WIDTH; j++)
+        {
+            GpeEventInfo = &GpeBlock->EventInfo[((ACPI_SIZE) i *
+                ACPI_GPE_REGISTER_WIDTH) + j];
+
+            if ((GpeEventInfo->Flags & ACPI_GPE_DISPATCH_MASK) ==
+                    ACPI_GPE_DISPATCH_HANDLER)
+            {
+                ACPI_FREE (GpeEventInfo->Dispatch.Handler);
+                GpeEventInfo->Dispatch.Handler = NULL;
+                GpeEventInfo->Flags &= ~ACPI_GPE_DISPATCH_MASK;
+            }
+        }
+    }
+
+    return_ACPI_STATUS (AE_OK);
+}
+
diff --git a/executer/exconfig.c b/executer/exconfig.c
index 034afb65c34..ab2be8a83de 100644
--- a/executer/exconfig.c
+++ b/executer/exconfig.c
@@ -164,8 +164,9 @@ AcpiExAddTable (
     ACPI_NAMESPACE_NODE     *ParentNode,
     ACPI_OPERAND_OBJECT     **DdbHandle)
 {
-    ACPI_STATUS             Status;
     ACPI_OPERAND_OBJECT     *ObjDesc;
+    ACPI_STATUS             Status;
+    ACPI_OWNER_ID           OwnerId;
 
 
     ACPI_FUNCTION_TRACE (ExAddTable);
@@ -205,7 +206,15 @@ AcpiExAddTable (
     AcpiNsExecModuleCodeList ();
     AcpiExEnterInterpreter ();
 
-    return_ACPI_STATUS (Status);
+    /* Update GPEs for any new _PRW or _Lxx/_Exx methods. Ignore errors */
+
+    Status = AcpiTbGetOwnerId (TableIndex, &OwnerId);
+    if (ACPI_SUCCESS (Status))
+    {
+        AcpiEvUpdateGpes (OwnerId);
+    }
+
+    return_ACPI_STATUS (AE_OK);
 }
 
 
@@ -347,9 +356,8 @@ AcpiExLoadTableOp (
     Status = AcpiGetTableByIndex (TableIndex, &Table);
     if (ACPI_SUCCESS (Status))
     {
-        ACPI_INFO ((AE_INFO,
-            "Dynamic OEM Table Load - [%.4s] OemId [%.6s] OemTableId [%.8s]",
-            Table->Signature, Table->OemId, Table->OemTableId));
+        ACPI_INFO ((AE_INFO, "Dynamic OEM Table Load:"));
+        AcpiTbPrintTableHeader (0, Table);
     }
 
     /* Invoke table handler if present */
@@ -644,6 +652,9 @@ AcpiExLoadOp (
         return_ACPI_STATUS (Status);
     }
 
+    ACPI_INFO ((AE_INFO, "Dynamic OEM Table Load:"));
+    AcpiTbPrintTableHeader (0, TableDesc.Pointer);
+
     /* Remove the reference by added by AcpiExStore above */
 
     AcpiUtRemoveReference (DdbHandle);
diff --git a/executer/exoparg1.c b/executer/exoparg1.c
index 1bff09b09cc..7a47bb102df 100644
--- a/executer/exoparg1.c
+++ b/executer/exoparg1.c
@@ -268,7 +268,7 @@ AcpiExOpcode_1A_0T_0R (
 
     case AML_SLEEP_OP:      /*  Sleep (MsecTime) */
 
-        Status = AcpiExSystemDoSuspend (Operand[0]->Integer.Value);
+        Status = AcpiExSystemDoSleep (Operand[0]->Integer.Value);
         break;
 
 
diff --git a/executer/exsystem.c b/executer/exsystem.c
index f43ef215527..8f4ccfe3714 100644
--- a/executer/exsystem.c
+++ b/executer/exsystem.c
@@ -280,19 +280,19 @@ AcpiExSystemDoStall (
 
 /*******************************************************************************
  *
- * FUNCTION:    AcpiExSystemDoSuspend
+ * FUNCTION:    AcpiExSystemDoSleep
  *
- * PARAMETERS:  HowLong         - The amount of time to suspend,
+ * PARAMETERS:  HowLong         - The amount of time to sleep,
  *                                in milliseconds
  *
  * RETURN:      None
  *
- * DESCRIPTION: Suspend running thread for specified amount of time.
+ * DESCRIPTION: Sleep the running thread for specified amount of time.
  *
  ******************************************************************************/
 
 ACPI_STATUS
-AcpiExSystemDoSuspend (
+AcpiExSystemDoSleep (
     UINT64                  HowLong)
 {
     ACPI_FUNCTION_ENTRY ();
diff --git a/include/acevents.h b/include/acevents.h
index 4d945eff8aa..40515579e5e 100644
--- a/include/acevents.h
+++ b/include/acevents.h
@@ -167,8 +167,12 @@ AcpiEvQueueNotifyRequest (
 
 
 /*
- * evgpe - GPE handling and dispatch
+ * evgpe - Low-level GPE support
  */
+UINT32
+AcpiEvGpeDetect (
+    ACPI_GPE_XRUPT_INFO     *GpeXruptList);
+
 ACPI_STATUS
 AcpiEvUpdateGpeEnableMasks (
     ACPI_GPE_EVENT_INFO     *GpeEventInfo);
@@ -193,23 +197,8 @@ AcpiEvLowGetGpeInfo (
 
 
 /*
- * evgpeblk
+ * evgpeblk - Upper-level GPE block support
  */
-BOOLEAN
-AcpiEvValidGpeEvent (
-    ACPI_GPE_EVENT_INFO     *GpeEventInfo);
-
-ACPI_STATUS
-AcpiEvWalkGpeList (
-    ACPI_GPE_CALLBACK       GpeWalkCallback,
-    void                    *Context);
-
-ACPI_STATUS
-AcpiEvDeleteGpeHandlers (
-    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo,
-    ACPI_GPE_BLOCK_INFO     *GpeBlock,
-    void                    *Context);
-
 ACPI_STATUS
 AcpiEvCreateGpeBlock (
     ACPI_NAMESPACE_NODE     *GpeDevice,
@@ -233,14 +222,57 @@ AcpiEvGpeDispatch (
     ACPI_GPE_EVENT_INFO     *GpeEventInfo,
     UINT32                  GpeNumber);
 
-UINT32
-AcpiEvGpeDetect (
-    ACPI_GPE_XRUPT_INFO     *GpeXruptList);
-
+/*
+ * evgpeinit - GPE initialization and update
+ */
 ACPI_STATUS
 AcpiEvGpeInitialize (
     void);
 
+void
+AcpiEvUpdateGpes (
+    ACPI_OWNER_ID           TableOwnerId);
+
+ACPI_STATUS
+AcpiEvMatchGpeMethod (
+    ACPI_HANDLE             ObjHandle,
+    UINT32                  Level,
+    void                    *Context,
+    void                    **ReturnValue);
+
+ACPI_STATUS
+AcpiEvMatchPrwAndGpe (
+    ACPI_HANDLE             ObjHandle,
+    UINT32                  Level,
+    void                    *Context,
+    void                    **ReturnValue);
+
+/*
+ * evgpeutil - GPE utilities
+ */
+ACPI_STATUS
+AcpiEvWalkGpeList (
+    ACPI_GPE_CALLBACK       GpeWalkCallback,
+    void                    *Context);
+
+BOOLEAN
+AcpiEvValidGpeEvent (
+    ACPI_GPE_EVENT_INFO     *GpeEventInfo);
+
+ACPI_GPE_XRUPT_INFO *
+AcpiEvGetGpeXruptBlock (
+    UINT32                  InterruptNumber);
+
+ACPI_STATUS
+AcpiEvDeleteGpeXrupt (
+    ACPI_GPE_XRUPT_INFO     *GpeXrupt);
+
+ACPI_STATUS
+AcpiEvDeleteGpeHandlers (
+    ACPI_GPE_XRUPT_INFO     *GpeXruptInfo,
+    ACPI_GPE_BLOCK_INFO     *GpeBlock,
+    void                    *Context);
+
 
 /*
  * evregion - Address Space handling
diff --git a/include/acglobal.h b/include/acglobal.h
index 4f0769e6f67..e0f8871fcd0 100644
--- a/include/acglobal.h
+++ b/include/acglobal.h
@@ -216,11 +216,10 @@ ACPI_NAME                   AcpiGbl_TraceMethodName;
  ****************************************************************************/
 
 /*
- * AcpiGbl_RootTableList is the master list of ACPI tables found in the
- * RSDT/XSDT.
- *
+ * AcpiGbl_RootTableList is the master list of ACPI tables that were
+ * found in the RSDT/XSDT.
  */
-ACPI_EXTERN ACPI_INTERNAL_RSDT          AcpiGbl_RootTableList;
+ACPI_EXTERN ACPI_TABLE_LIST             AcpiGbl_RootTableList;
 ACPI_EXTERN ACPI_TABLE_FACS            *AcpiGbl_FACS;
 
 /* These addresses are calculated from the FADT Event Block addresses */
diff --git a/include/acinterp.h b/include/acinterp.h
index 67142c8911b..e3fa940a102 100644
--- a/include/acinterp.h
+++ b/include/acinterp.h
@@ -434,7 +434,7 @@ AcpiExSystemDoNotifyOp (
     ACPI_OPERAND_OBJECT     *ObjDesc);
 
 ACPI_STATUS
-AcpiExSystemDoSuspend(
+AcpiExSystemDoSleep(
     UINT64                  Time);
 
 ACPI_STATUS
diff --git a/include/aclocal.h b/include/aclocal.h
index 8818a6ea9b5..1deced01136 100644
--- a/include/aclocal.h
+++ b/include/aclocal.h
@@ -307,16 +307,16 @@ typedef struct acpi_namespace_node
 #define ANOBJ_IS_REFERENCED             0x80    /* iASL only: Object was referenced */
 
 
-/* One internal RSDT for table management */
+/* Internal ACPI table management - master table list */
 
-typedef struct acpi_internal_rsdt
+typedef struct acpi_table_list
 {
-    ACPI_TABLE_DESC                 *Tables;
-    UINT32                          Count;
-    UINT32                          Size;
+    ACPI_TABLE_DESC                 *Tables;            /* Table descriptor array */
+    UINT32                          CurrentTableCount;  /* Tables currently in the array */
+    UINT32                          MaxTableCount;      /* Max tables array will hold */
     UINT8                           Flags;
 
-} ACPI_INTERNAL_RSDT;
+} ACPI_TABLE_LIST;
 
 /* Flags for above */
 
@@ -612,6 +612,10 @@ typedef struct acpi_gpe_walk_info
 {
     ACPI_NAMESPACE_NODE             *GpeDevice;
     ACPI_GPE_BLOCK_INFO             *GpeBlock;
+    UINT16                          Count;
+    ACPI_OWNER_ID                   OwnerId;
+    BOOLEAN                         EnableThisGpe;
+    BOOLEAN                         ExecuteByOwnerId;
 
 } ACPI_GPE_WALK_INFO;
 
diff --git a/include/acpiosxf.h b/include/acpiosxf.h
index 0c22b63bb69..3a9da6c512e 100644
--- a/include/acpiosxf.h
+++ b/include/acpiosxf.h
@@ -410,8 +410,8 @@ AcpiOsWritePciConfiguration (
  */
 void
 AcpiOsDerivePciId(
-    ACPI_HANDLE             Rhandle,
-    ACPI_HANDLE             Chandle,
+    ACPI_HANDLE             Device,
+    ACPI_HANDLE             Region,
     ACPI_PCI_ID             **PciId);
 
 
diff --git a/include/acpixf.h b/include/acpixf.h
index 300b2c73f5f..0e68fabfeb0 100644
--- a/include/acpixf.h
+++ b/include/acpixf.h
@@ -120,7 +120,7 @@
 
 /* Current ACPICA subsystem version in YYYYMMDD format */
 
-#define ACPI_CA_VERSION                 0x20100331
+#define ACPI_CA_VERSION                 0x20100428
 
 #include "actypes.h"
 #include "actbl.h"
@@ -283,7 +283,7 @@ AcpiGetDevices (
 
 ACPI_STATUS
 AcpiGetName (
-    ACPI_HANDLE             Handle,
+    ACPI_HANDLE             Object,
     UINT32                  NameType,
     ACPI_BUFFER             *RetPathPtr);
 
@@ -295,18 +295,18 @@ AcpiGetHandle (
 
 ACPI_STATUS
 AcpiAttachData (
-    ACPI_HANDLE             ObjHandle,
+    ACPI_HANDLE             Object,
     ACPI_OBJECT_HANDLER     Handler,
     void                    *Data);
 
 ACPI_STATUS
 AcpiDetachData (
-    ACPI_HANDLE             ObjHandle,
+    ACPI_HANDLE             Object,
     ACPI_OBJECT_HANDLER     Handler);
 
 ACPI_STATUS
 AcpiGetData (
-    ACPI_HANDLE             ObjHandle,
+    ACPI_HANDLE             Object,
     ACPI_OBJECT_HANDLER     Handler,
     void                    **Data);
 
@@ -338,7 +338,7 @@ AcpiEvaluateObjectTyped (
 
 ACPI_STATUS
 AcpiGetObjectInfo (
-    ACPI_HANDLE             Handle,
+    ACPI_HANDLE             Object,
     ACPI_DEVICE_INFO        **ReturnBuffer);
 
 ACPI_STATUS
@@ -527,36 +527,36 @@ ACPI_STATUS (*ACPI_WALK_RESOURCE_CALLBACK) (
 
 ACPI_STATUS
 AcpiGetVendorResource (
-    ACPI_HANDLE             DeviceHandle,
+    ACPI_HANDLE             Device,
     char                    *Name,
     ACPI_VENDOR_UUID        *Uuid,
     ACPI_BUFFER             *RetBuffer);
 
 ACPI_STATUS
-AcpiGetCurrentResources(
-    ACPI_HANDLE             DeviceHandle,
+AcpiGetCurrentResources (
+    ACPI_HANDLE             Device,
     ACPI_BUFFER             *RetBuffer);
 
 ACPI_STATUS
-AcpiGetPossibleResources(
-    ACPI_HANDLE             DeviceHandle,
+AcpiGetPossibleResources (
+    ACPI_HANDLE             Device,
     ACPI_BUFFER             *RetBuffer);
 
 ACPI_STATUS
 AcpiWalkResources (
-    ACPI_HANDLE                 DeviceHandle,
+    ACPI_HANDLE                 Device,
     char                        *Name,
     ACPI_WALK_RESOURCE_CALLBACK UserFunction,
     void                        *Context);
 
 ACPI_STATUS
 AcpiSetCurrentResources (
-    ACPI_HANDLE             DeviceHandle,
+    ACPI_HANDLE             Device,
     ACPI_BUFFER             *InBuffer);
 
 ACPI_STATUS
-AcpiGetIrqRoutingTable  (
-    ACPI_HANDLE             BusDeviceHandle,
+AcpiGetIrqRoutingTable (
+    ACPI_HANDLE             Device,
     ACPI_BUFFER             *RetBuffer);
 
 ACPI_STATUS
diff --git a/include/actypes.h b/include/actypes.h
index 3c0626ab1e3..601e87e908d 100644
--- a/include/actypes.h
+++ b/include/actypes.h
@@ -1099,7 +1099,7 @@ ACPI_STATUS (*ACPI_ADR_SPACE_SETUP) (
 
 typedef
 ACPI_STATUS (*ACPI_WALK_CALLBACK) (
-    ACPI_HANDLE                     ObjHandle,
+    ACPI_HANDLE                     Object,
     UINT32                          NestingLevel,
     void                            *Context,
     void                            **ReturnValue);
diff --git a/include/amlresrc.h b/include/amlresrc.h
index 9084e44540a..4cd2617e22b 100644
--- a/include/amlresrc.h
+++ b/include/amlresrc.h
@@ -300,6 +300,12 @@ typedef struct aml_resource_large_header
 } AML_RESOURCE_LARGE_HEADER;
 
 
+/* General Flags for address space resource descriptors */
+
+#define ACPI_RESOURCE_FLAG_DEC      2
+#define ACPI_RESOURCE_FLAG_MIF      4
+#define ACPI_RESOURCE_FLAG_MAF      8
+
 typedef struct aml_resource_memory24
 {
     AML_RESOURCE_LARGE_HEADER_COMMON
diff --git a/os_specific/service_layers/osunixxf.c b/os_specific/service_layers/osunixxf.c
index e62ed738be1..b457745122a 100644
--- a/os_specific/service_layers/osunixxf.c
+++ b/os_specific/service_layers/osunixxf.c
@@ -986,8 +986,8 @@ AcpiOsWritePciConfiguration (
 /* TEMPORARY STUB FUNCTION */
 void
 AcpiOsDerivePciId(
-    ACPI_HANDLE             rhandle,
-    ACPI_HANDLE             chandle,
+    ACPI_HANDLE             Device,
+    ACPI_HANDLE             Region,
     ACPI_PCI_ID             **PciId)
 {
 
diff --git a/os_specific/service_layers/oswinxf.c b/os_specific/service_layers/oswinxf.c
index 0f117772eaa..4c50416267a 100644
--- a/os_specific/service_layers/oswinxf.c
+++ b/os_specific/service_layers/oswinxf.c
@@ -1234,8 +1234,8 @@ AcpiOsWritePciConfiguration (
 /* TEMPORARY STUB FUNCTION */
 void
 AcpiOsDerivePciId(
-    ACPI_HANDLE             rhandle,
-    ACPI_HANDLE             chandle,
+    ACPI_HANDLE             Device,
+    ACPI_HANDLE             Region,
     ACPI_PCI_ID             **PciId)
 {
 
diff --git a/osunixxf.c b/osunixxf.c
index e62ed738be1..b457745122a 100644
--- a/osunixxf.c
+++ b/osunixxf.c
@@ -986,8 +986,8 @@ AcpiOsWritePciConfiguration (
 /* TEMPORARY STUB FUNCTION */
 void
 AcpiOsDerivePciId(
-    ACPI_HANDLE             rhandle,
-    ACPI_HANDLE             chandle,
+    ACPI_HANDLE             Device,
+    ACPI_HANDLE             Region,
     ACPI_PCI_ID             **PciId)
 {
 
diff --git a/tables/tbfind.c b/tables/tbfind.c
index e1c110f4523..4fd988ef9e8 100644
--- a/tables/tbfind.c
+++ b/tables/tbfind.c
@@ -164,7 +164,7 @@ AcpiTbFindTable (
 
     /* Search for the table */
 
-    for (i = 0; i < AcpiGbl_RootTableList.Count; ++i)
+    for (i = 0; i < AcpiGbl_RootTableList.CurrentTableCount; ++i)
     {
         if (ACPI_MEMCMP (&(AcpiGbl_RootTableList.Tables[i].Signature),
                             Header.Signature, ACPI_NAME_SIZE))
diff --git a/tables/tbinstal.c b/tables/tbinstal.c
index 78cf164a99e..2179b6bf7b7 100644
--- a/tables/tbinstal.c
+++ b/tables/tbinstal.c
@@ -227,7 +227,7 @@ AcpiTbAddTable (
 
     /* Check if table is already registered */
 
-    for (i = 0; i < AcpiGbl_RootTableList.Count; ++i)
+    for (i = 0; i < AcpiGbl_RootTableList.CurrentTableCount; ++i)
     {
         if (!AcpiGbl_RootTableList.Tables[i].Pointer)
         {
@@ -370,7 +370,7 @@ AcpiTbResizeRootTableList (
     /* Increase the Table Array size */
 
     Tables = ACPI_ALLOCATE_ZEROED (
-        ((ACPI_SIZE) AcpiGbl_RootTableList.Size +
+        ((ACPI_SIZE) AcpiGbl_RootTableList.MaxTableCount +
             ACPI_ROOT_TABLE_SIZE_INCREMENT) *
         sizeof (ACPI_TABLE_DESC));
     if (!Tables)
@@ -384,7 +384,7 @@ AcpiTbResizeRootTableList (
     if (AcpiGbl_RootTableList.Tables)
     {
         ACPI_MEMCPY (Tables, AcpiGbl_RootTableList.Tables,
-            (ACPI_SIZE) AcpiGbl_RootTableList.Size * sizeof (ACPI_TABLE_DESC));
+            (ACPI_SIZE) AcpiGbl_RootTableList.MaxTableCount * sizeof (ACPI_TABLE_DESC));
 
         if (AcpiGbl_RootTableList.Flags & ACPI_ROOT_ORIGIN_ALLOCATED)
         {
@@ -393,7 +393,7 @@ AcpiTbResizeRootTableList (
     }
 
     AcpiGbl_RootTableList.Tables = Tables;
-    AcpiGbl_RootTableList.Size += ACPI_ROOT_TABLE_SIZE_INCREMENT;
+    AcpiGbl_RootTableList.MaxTableCount += ACPI_ROOT_TABLE_SIZE_INCREMENT;
     AcpiGbl_RootTableList.Flags |= (UINT8) ACPI_ROOT_ORIGIN_ALLOCATED;
 
     return_ACPI_STATUS (AE_OK);
@@ -423,12 +423,14 @@ AcpiTbStoreTable (
     UINT8                   Flags,
     UINT32                  *TableIndex)
 {
-    ACPI_STATUS             Status = AE_OK;
+    ACPI_STATUS             Status;
+    ACPI_TABLE_DESC         *NewTable;
 
 
     /* Ensure that there is room for the table in the Root Table List */
 
-    if (AcpiGbl_RootTableList.Count >= AcpiGbl_RootTableList.Size)
+    if (AcpiGbl_RootTableList.CurrentTableCount >=
+        AcpiGbl_RootTableList.MaxTableCount)
     {
         Status = AcpiTbResizeRootTableList();
         if (ACPI_FAILURE (Status))
@@ -437,21 +439,21 @@ AcpiTbStoreTable (
         }
     }
 
+    NewTable = &AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.CurrentTableCount];
+
     /* Initialize added table */
 
-    AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Address = Address;
-    AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Pointer = Table;
-    AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Length = Length;
-    AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].OwnerId = 0;
-    AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Flags = Flags;
+    NewTable->Address = Address;
+    NewTable->Pointer = Table;
+    NewTable->Length = Length;
+    NewTable->OwnerId = 0;
+    NewTable->Flags = Flags;
 
-    ACPI_MOVE_32_TO_32 (
-        &(AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Signature),
-        Table->Signature);
+    ACPI_MOVE_32_TO_32 (&NewTable->Signature, Table->Signature);
 
-    *TableIndex = AcpiGbl_RootTableList.Count;
-    AcpiGbl_RootTableList.Count++;
-    return (Status);
+    *TableIndex = AcpiGbl_RootTableList.CurrentTableCount;
+    AcpiGbl_RootTableList.CurrentTableCount++;
+    return (AE_OK);
 }
 
 
@@ -523,7 +525,7 @@ AcpiTbTerminate (
 
     /* Delete the individual tables */
 
-    for (i = 0; i < AcpiGbl_RootTableList.Count; i++)
+    for (i = 0; i < AcpiGbl_RootTableList.CurrentTableCount; i++)
     {
         AcpiTbDeleteTable (&AcpiGbl_RootTableList.Tables[i]);
     }
@@ -539,7 +541,7 @@ AcpiTbTerminate (
 
     AcpiGbl_RootTableList.Tables = NULL;
     AcpiGbl_RootTableList.Flags = 0;
-    AcpiGbl_RootTableList.Count = 0;
+    AcpiGbl_RootTableList.CurrentTableCount = 0;
 
     ACPI_DEBUG_PRINT ((ACPI_DB_INFO, "ACPI Tables freed\n"));
     (void) AcpiUtReleaseMutex (ACPI_MTX_TABLES);
@@ -575,7 +577,7 @@ AcpiTbDeleteNamespaceByOwner (
         return_ACPI_STATUS (Status);
     }
 
-    if (TableIndex >= AcpiGbl_RootTableList.Count)
+    if (TableIndex >= AcpiGbl_RootTableList.CurrentTableCount)
     {
         /* The table index does not exist */
 
@@ -634,7 +636,7 @@ AcpiTbAllocateOwnerId (
 
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    if (TableIndex < AcpiGbl_RootTableList.Count)
+    if (TableIndex < AcpiGbl_RootTableList.CurrentTableCount)
     {
         Status = AcpiUtAllocateOwnerId
                     (&(AcpiGbl_RootTableList.Tables[TableIndex].OwnerId));
@@ -668,7 +670,7 @@ AcpiTbReleaseOwnerId (
 
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    if (TableIndex < AcpiGbl_RootTableList.Count)
+    if (TableIndex < AcpiGbl_RootTableList.CurrentTableCount)
     {
         AcpiUtReleaseOwnerId (
             &(AcpiGbl_RootTableList.Tables[TableIndex].OwnerId));
@@ -705,7 +707,7 @@ AcpiTbGetOwnerId (
 
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    if (TableIndex < AcpiGbl_RootTableList.Count)
+    if (TableIndex < AcpiGbl_RootTableList.CurrentTableCount)
     {
         *OwnerId = AcpiGbl_RootTableList.Tables[TableIndex].OwnerId;
         Status = AE_OK;
@@ -734,7 +736,7 @@ AcpiTbIsTableLoaded (
 
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    if (TableIndex < AcpiGbl_RootTableList.Count)
+    if (TableIndex < AcpiGbl_RootTableList.CurrentTableCount)
     {
         IsLoaded = (BOOLEAN)
             (AcpiGbl_RootTableList.Tables[TableIndex].Flags &
@@ -766,7 +768,7 @@ AcpiTbSetTableLoadedFlag (
 {
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    if (TableIndex < AcpiGbl_RootTableList.Count)
+    if (TableIndex < AcpiGbl_RootTableList.CurrentTableCount)
     {
         if (IsLoaded)
         {
diff --git a/tables/tbutils.c b/tables/tbutils.c
index eb8bc43597d..100a91d9e51 100644
--- a/tables/tbutils.c
+++ b/tables/tbutils.c
@@ -184,7 +184,7 @@ AcpiTbTablesLoaded (
     void)
 {
 
-    if (AcpiGbl_RootTableList.Count >= 3)
+    if (AcpiGbl_RootTableList.CurrentTableCount >= 3)
     {
         return (TRUE);
     }
@@ -765,14 +765,15 @@ AcpiTbParseRootTable (
      * come from the FADT
      */
     TableEntry = ACPI_CAST_PTR (UINT8, Table) + sizeof (ACPI_TABLE_HEADER);
-    AcpiGbl_RootTableList.Count = 2;
+    AcpiGbl_RootTableList.CurrentTableCount = 2;
 
     /*
      * Initialize the root table array from the RSDT/XSDT
      */
     for (i = 0; i < TableCount; i++)
     {
-        if (AcpiGbl_RootTableList.Count >= AcpiGbl_RootTableList.Size)
+        if (AcpiGbl_RootTableList.CurrentTableCount >=
+            AcpiGbl_RootTableList.MaxTableCount)
         {
             /* There is no more room in the root table array, attempt resize */
 
@@ -781,18 +782,18 @@ AcpiTbParseRootTable (
             {
                 ACPI_WARNING ((AE_INFO, "Truncating %u table entries!",
                     (unsigned) (TableCount -
-                    (AcpiGbl_RootTableList.Count - 2))));
+                    (AcpiGbl_RootTableList.CurrentTableCount - 2))));
                 break;
             }
         }
 
         /* Get the table physical address (32-bit for RSDT, 64-bit for XSDT) */
 
-        AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.Count].Address =
+        AcpiGbl_RootTableList.Tables[AcpiGbl_RootTableList.CurrentTableCount].Address =
             AcpiTbGetRootTableEntry (TableEntry, TableEntrySize);
 
         TableEntry += TableEntrySize;
-        AcpiGbl_RootTableList.Count++;
+        AcpiGbl_RootTableList.CurrentTableCount++;
     }
 
     /*
@@ -805,7 +806,7 @@ AcpiTbParseRootTable (
      * Complete the initialization of the root table array by examining
      * the header of each table
      */
-    for (i = 2; i < AcpiGbl_RootTableList.Count; i++)
+    for (i = 2; i < AcpiGbl_RootTableList.CurrentTableCount; i++)
     {
         AcpiTbInstallTable (AcpiGbl_RootTableList.Tables[i].Address,
             NULL, i);
diff --git a/tables/tbxface.c b/tables/tbxface.c
index b3e09a0271f..ba1a026e868 100644
--- a/tables/tbxface.c
+++ b/tables/tbxface.c
@@ -150,7 +150,7 @@ AcpiAllocateRootTable (
     UINT32                  InitialTableCount)
 {
 
-    AcpiGbl_RootTableList.Size = InitialTableCount;
+    AcpiGbl_RootTableList.MaxTableCount = InitialTableCount;
     AcpiGbl_RootTableList.Flags = ACPI_ROOT_ALLOW_RESIZE;
 
     return (AcpiTbResizeRootTableList ());
@@ -216,7 +216,7 @@ AcpiInitializeTables (
             (ACPI_SIZE) InitialTableCount * sizeof (ACPI_TABLE_DESC));
 
         AcpiGbl_RootTableList.Tables = InitialTableArray;
-        AcpiGbl_RootTableList.Size = InitialTableCount;
+        AcpiGbl_RootTableList.MaxTableCount = InitialTableCount;
         AcpiGbl_RootTableList.Flags = ACPI_ROOT_ORIGIN_UNKNOWN;
         if (AllowResize)
         {
@@ -285,7 +285,7 @@ AcpiReallocateRootTable (
      * increment to create the new table size.
      */
     CurrentSize = (ACPI_SIZE)
-        AcpiGbl_RootTableList.Count * sizeof (ACPI_TABLE_DESC);
+        AcpiGbl_RootTableList.CurrentTableCount * sizeof (ACPI_TABLE_DESC);
 
     NewSize = CurrentSize +
         (ACPI_ROOT_TABLE_SIZE_INCREMENT * sizeof (ACPI_TABLE_DESC));
@@ -306,8 +306,8 @@ AcpiReallocateRootTable (
      * size of the original table list.
      */
     AcpiGbl_RootTableList.Tables = Tables;
-    AcpiGbl_RootTableList.Size =
-        AcpiGbl_RootTableList.Count + ACPI_ROOT_TABLE_SIZE_INCREMENT;
+    AcpiGbl_RootTableList.MaxTableCount =
+        AcpiGbl_RootTableList.CurrentTableCount + ACPI_ROOT_TABLE_SIZE_INCREMENT;
     AcpiGbl_RootTableList.Flags =
         ACPI_ROOT_ORIGIN_ALLOCATED | ACPI_ROOT_ALLOW_RESIZE;
 
@@ -354,7 +354,7 @@ AcpiGetTableHeader (
 
     /* Walk the root table list */
 
-    for (i = 0, j = 0; i < AcpiGbl_RootTableList.Count; i++)
+    for (i = 0, j = 0; i < AcpiGbl_RootTableList.CurrentTableCount; i++)
     {
         if (!ACPI_COMPARE_NAME (&(AcpiGbl_RootTableList.Tables[i].Signature),
                     Signature))
@@ -439,7 +439,7 @@ AcpiGetTable (
 
     /* Walk the root table list */
 
-    for (i = 0, j = 0; i < AcpiGbl_RootTableList.Count; i++)
+    for (i = 0, j = 0; i < AcpiGbl_RootTableList.CurrentTableCount; i++)
     {
         if (!ACPI_COMPARE_NAME (&(AcpiGbl_RootTableList.Tables[i].Signature),
                 Signature))
@@ -502,7 +502,7 @@ AcpiGetTableByIndex (
 
     /* Validate index */
 
-    if (TableIndex >= AcpiGbl_RootTableList.Count)
+    if (TableIndex >= AcpiGbl_RootTableList.CurrentTableCount)
     {
         (void) AcpiUtReleaseMutex (ACPI_MTX_TABLES);
         return_ACPI_STATUS (AE_BAD_PARAMETER);
@@ -559,7 +559,7 @@ AcpiTbLoadNamespace (
      * Load the namespace. The DSDT is required, but any SSDT and
      * PSDT tables are optional. Verify the DSDT.
      */
-    if (!AcpiGbl_RootTableList.Count ||
+    if (!AcpiGbl_RootTableList.CurrentTableCount ||
         !ACPI_COMPARE_NAME (
             &(AcpiGbl_RootTableList.Tables[ACPI_TABLE_INDEX_DSDT].Signature),
             ACPI_SIG_DSDT) ||
@@ -613,7 +613,7 @@ AcpiTbLoadNamespace (
     /* Load any SSDT or PSDT tables. Note: Loop leaves tables locked */
 
     (void) AcpiUtAcquireMutex (ACPI_MTX_TABLES);
-    for (i = 0; i < AcpiGbl_RootTableList.Count; ++i)
+    for (i = 0; i < AcpiGbl_RootTableList.CurrentTableCount; ++i)
     {
         if ((!ACPI_COMPARE_NAME (&(AcpiGbl_RootTableList.Tables[i].Signature),
                     ACPI_SIG_SSDT) &&
diff --git a/tests/misc/badcode.asl b/tests/misc/badcode.asl
new file mode 100644
index 00000000000..36ca0d582f2
--- /dev/null
+++ b/tests/misc/badcode.asl
@@ -0,0 +1,295 @@
+/*
+ * badcode.asl
+ *
+ * This file contains examples of the extended error checking and
+ * typechecking capabilities of the iASL compiler. Other ASL compilers
+ * may ignore these errors completely. Note - this is not an exhaustive
+ * list of errors detected by iASL, it shows many of the errors that
+ * are not detected by other ASL compilers.
+ *
+ * To compile, use:
+ * iasl badcode.asl
+ */
+DefinitionBlock ("badcode.aml", "DSDT", 1, "Intel", "Example", 0x00000001)
+{
+    Name (INT1, 0)
+    Name (BUF1, Buffer() {0,1,2,3})
+    Event (EVT1)
+
+    // Invalid SyncLevel in Mutex declaration
+
+    Mutex (MTX1, 32)
+
+    // Integer beyond the table integer size (32 bits)
+
+    Name (BIG, 0x1234567887654321)
+
+    // CPackage length does not match initializer list length
+
+    Name (PKG1, Package(5) {0,1})
+
+    // Inadvertent use of single backslash in a string
+
+    Name (PATH, Buffer() {"\_SB_.PCI2._CRS"})
+
+    // Invalid hex escape sequence
+
+    Name (ESC1, "abcdefg\x00hijklmn")
+    
+    // Field access beyond region bounds
+
+    OperationRegion (OPR1, SystemMemory, 0x2000, 6)
+    Field (OPR1, DWordAcc, NoLock, Preserve)
+    {
+        Offset (4),
+        FLD1, 8
+    }
+
+    // Some address spaces support only ByteAcc or BufferAcc
+
+    OperationRegion (OPR2, EmbeddedControl, 0x4000, 8)
+    Field (OPR2, DWordAcc, NoLock, Preserve)
+    {
+        FLD2, 8
+    }
+    OperationRegion (OPR3, SMBus, 0x8000, 16)
+    Field (OPR3, WordAcc, NoLock, Preserve)
+    {
+        FLD3, 8
+    }
+
+    // Invalid SyncLevel in method declaration
+
+    Method (MTH1, 0, NotSerialized, 32)
+    {
+        // Invalid arguments and uninitialized locals
+
+        Store (Arg3, Local0)
+        Store (Local1, Local2)
+
+        // Parameter typechecking (MTX1 is invalid type)
+
+        Subtract (MTX1, 4, Local3)
+
+        // Various invalid parameters
+
+        CreateField (BUF1, 0, Subtract (4, 4), FLD1)
+
+        // Unchecked mutex and event timeouts
+
+        Acquire (MTX1, 100)
+        Wait (EVT1, 1)
+
+        // Result from operation is not used - statement has no effect
+
+        Add (INT1, 8)
+
+        // Unreachable code
+
+        Return (0)
+        Store (5, INT1)
+    }
+
+    Method (MTH2)
+    {
+        // Switch with no Case statements
+
+        Switch (ToInteger (INT1))
+        {
+            Default
+            {
+            }
+        }
+
+        if (LEqual (INT1, 0))
+        {
+            Return (INT1)
+        }
+
+        // Fallthrough exit path does not return a value
+    }
+
+    Method (MTH3)
+    {
+        // Method MTH2 above does not always return a value
+
+        Store (MTH2 (), Local0)
+    }
+
+    // Invalid _HID value
+
+    Name (_HID, "*PNP0C0A")
+
+    // Predefined Name typechecking
+
+    Name (_PRW, 4)
+    Name (_FDI, Buffer () {0})
+
+    // Predefined Name argument count validation
+    // and return value validation
+
+    Method (_OSC, 5)
+    {
+    }
+
+    // Predefined Names that must be implemented as control methods
+
+    Name (_L01, 1)
+    Name (_E02, 2)
+    Name (_Q03, 3)
+    Name (_ON,  0)
+    Name (_INI, 1)
+    Name (_PTP, 2)
+
+    /*
+     * Resource Descriptor error checking
+     */
+    Name (RSC1, ResourceTemplate ()
+    {
+        // Illegal nested StartDependent macros
+
+		StartDependentFn (0, 0)
+		{
+		    StartDependentFn (0, 0)
+		    {
+            }
+        }
+
+        // Missing EndDependentFn macro 
+    })
+
+    Name (RSC2, ResourceTemplate ()
+    {
+        // AddressMin is larger than AddressMax
+        IO (Decode16,
+            0x07D0,             // Range Minimum
+            0x03E8,             // Range Maximum
+            0x01,               // Alignment
+            0x20,               // Length
+            )
+
+        // Length larger than Min/Max window size
+        Memory32 (ReadOnly,
+            0x00001000,         // Range Minimum
+            0x00002000,         // Range Maximum
+            0x00000004,         // Alignment
+            0x00002000,         // Length
+            )
+
+        // Min and Max not multiples of alignment value
+        Memory32 (ReadOnly,
+            0x00001001,         // Range Minimum
+            0x00002002,         // Range Maximum
+            0x00000004,         // Alignment
+            0x00000200,         // Length
+            )
+
+        // 10-bit ISA I/O address has a max of 0x3FF
+        FixedIO (
+            0xFFFF,             // Address
+            0x20,               // Length
+            )
+
+        // Invalid AccessSize parameter
+        Register (SystemIO, 
+            0x08,               // Bit Width
+            0x00,               // Bit Offset
+            0x0000000000000100, // Address
+            0x05                // Access Size
+            )
+
+        // Invalid ResourceType (0xB0)
+        QWordSpace (0xB0, ResourceConsumer, PosDecode, MinFixed, MaxFixed, 0xA5,
+            0x0000,             // Granularity
+            0xA000,             // Range Minimum
+            0xBFFF,             // Range Maximum
+            0x0000,             // Translation Offset
+            0x2000,             // Length
+            ,, )
+
+        // AddressMin is larger than AddressMax
+        WordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+            0x0000,             // Granularity
+            0x0200,             // Range Minimum
+            0x0100,             // Range Maximum
+            0x0000,             // Translation Offset
+            0x0100,             // Length
+            ,, , TypeStatic)
+
+        // Length larger than Min/Max window size
+        DWordSpace (0xC3, ResourceConsumer, PosDecode, MinFixed, MaxFixed, 0xA5,
+            0x00000000,         // Granularity
+            0x000C8000,         // Range Minimum
+            0x000C9000,         // Range Maximum
+            0x00000000,         // Translation Offset
+            0x00001002,         // Length
+            ,, )
+
+        // Granularity must be (power-of-two -1)
+        DWordMemory (ResourceProducer, PosDecode, MinFixed, MaxNotFixed, NonCacheable, ReadWrite,
+            0x00000010,
+            0x40000000,
+            0xFED9FFFF,
+            0x00000000,
+            0xBECA0000)
+
+        // Address Min (with zero length) not on granularity boundary
+        QWordIO (ResourceProducer, MinFixed, MaxNotFixed, PosDecode, EntireRange,
+            0x0000000000000003, // Granularity
+            0x0000000000000B02, // Range Minimum
+            0x0000000000000C00, // Range Maximum
+            0x0000000000000000, // Translation Offset
+            0x0000000000000000, // Length
+            ,, , TypeStatic)
+
+        // Address Max (with zero length) not on (granularity boundary -1)
+        QWordMemory (ResourceProducer, PosDecode, MinNotFixed, MaxFixed, Cacheable, ReadWrite,
+            0x0000000000000001, // Granularity
+            0x0000000000100000, // Range Minimum
+            0x00000000002FFFFE, // Range Maximum
+            0x0000000000000000, // Translation Offset
+            0x0000000000000000, // Length
+            ,, , AddressRangeMemory, TypeStatic)
+
+        // Invalid combination: zero length, both Min and Max are fixed
+        DWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+            0x00000000,         // Granularity
+            0x000C8000,         // Range Minimum
+            0x000C8FFF,         // Range Maximum
+            0x00000000,         // Translation Offset
+            0x00000000,         // Length
+            ,, )
+
+        // Invalid combination: non-zero length, Min Fixed, Max not fixed
+        DWordIO (ResourceProducer, MinFixed, MaxNotFixed, PosDecode, EntireRange,
+            0x00000001,         // Granularity
+            0x000C8000,         // Range Minimum
+            0x000C8FFF,         // Range Maximum
+            0x00000000,         // Translation Offset
+            0x00000100,         // Length
+            ,, )
+
+        // Invalid combination: non-zero length, Min not Fixed, Max fixed
+        DWordIO (ResourceProducer, MinNotFixed, MaxFixed, PosDecode, EntireRange,
+            0x00000001,         // Granularity
+            0x000C8000,         // Range Minimum
+            0x000C8FFF,         // Range Maximum
+            0x00000000,         // Translation Offset
+            0x00000200,         // Length
+            ,, )
+
+        // Granularity must be zero if non-zero length, min/max fixed
+        DWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+            0x0000000F,         // Granularity
+            0x000C8000,         // Range Minimum
+            0x000C8FFF,         // Range Maximum
+            0x00000000,         // Translation Offset
+            0x00001000,         // Length
+            ,, )
+
+        // Missing StartDependentFn macro
+
+        EndDependentFn ()
+    })
+}
+
diff --git a/tests/misc/grammar.asl b/tests/misc/grammar.asl
new file mode 100644
index 00000000000..8a13fd1a20c
--- /dev/null
+++ b/tests/misc/grammar.asl
@@ -0,0 +1,10266 @@
+/*
+ * Some or all of this work - Copyright (c) 2006 - 2010, Intel Corp.
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without modification,
+ * are permitted provided that the following conditions are met:
+ *
+ * Redistributions of source code must retain the above copyright notice,
+ * this list of conditions and the following disclaimer.
+ * Redistributions in binary form must reproduce the above copyright notice,
+ * this list of conditions and the following disclaimer in the documentation
+ * and/or other materials provided with the distribution.
+ * Neither the name of Intel Corporation nor the names of its contributors
+ * may be used to endorse or promote products derived from this software
+ * without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
+ * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
+ * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES;
+ * LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+ * NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE,
+ * EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+//
+//
+// Grammar.asl - Minimally excercises most ASL constructs
+//
+// NOTE -- use: iasl -f -of grammar.asl to compile
+//
+//         This 1) Ignores errors (checks compiler error handling)
+//              2) Disables constant folding
+//
+//
+
+/*******************************************************************************
+Compilation should look like this:
+
+C:\acpica\tests\misc>iasl -f -of grammar.asl
+
+Intel ACPI Component Architecture
+ASL Optimizing Compiler version 20090422 [Apr 22 2009]
+Copyright (C) 2000 - 2009 Intel Corporation
+Supports ACPI Specification Revision 3.0a
+
+grammar.asl   187:     Name (_NPK, Package (8)
+Warning  1098 -                 ^ Unknown reserved name (_NPK)
+
+grammar.asl   510:     NAME (ESC1, "abcdefg\x00hijklmn")
+Warning  1042 -                                ^ Invalid Hex/Octal Escape - Non-ASCII or NULL
+
+grammar.asl   511:     NAME (ESC2, "abcdefg\000hijklmn")
+Warning  1042 -                                ^ Invalid Hex/Octal Escape - Non-ASCII or NULL
+
+grammar.asl   601:     Method (RCIV, 1)
+Warning  1087 -                   ^ Not all control paths return a value (RCIV)
+
+grammar.asl   608:         RCIV (Subtract (Arg0, 1))
+Remark   5073 -               ^ Recursive method call (RCIV)
+
+grammar.asl   937:     Method (_ERR, 2)
+Warning  1077 -                   ^ Reserved method has too few arguments (_ERR requires 3)
+
+grammar.asl  1377:         Store (0x1234567887654321, QWD2)
+Warning  1032 -                                    ^ 64-bit integer in 32-bit table, truncating
+
+grammar.asl  1379:         if (LNotEqual (Local0, 0x1234567887654321))
+Warning  1032 -         64-bit integer in 32-bit table, truncating ^
+
+grammar.asl  1459:         SizeOf (BUFO)
+Warning  1105 -                       ^ Result is not used, operator has no effect
+
+grammar.asl  1485:         Acquire (MTX2, 1)
+Warning  1104 -                           ^ Possible operator timeout is ignored
+
+grammar.asl  1633:         Add (Local0, Local1)
+Warning  1105 -                      ^ Result is not used, operator has no effect
+
+grammar.asl  1804:     Method (COND)
+Warning  1087 -                   ^ Not all control paths return a value (COND)
+
+grammar.asl  6010:             Name (_HID, "*PNP0A06")
+Error    4001 -                                     ^ String must be entirely alphanumeric (*PNP0A06)
+
+grammar.asl  6461:             Name (_CRS, Buffer(26)  {"\_SB_.PCI2._CRS..........."})
+Warning  1038 -        Invalid or unknown escape sequence ^
+
+grammar.asl  6800:                 And (Local0, 1, Local0) //  Local0 &= 1
+Error    4050 -                              ^ Method local variable is not initialized (Local0)
+
+grammar.asl  6886:             Name (_HID, "*PNP0C0A")     //  Control Method Battey ID
+Error    4001 -                                     ^ String must be entirely alphanumeric (*PNP0C0A)
+
+ASL Input:  grammar.asl - 10254 lines, 322162 bytes, 4810 keywords
+AML Output: grammar.aml - 43392 bytes, 669 named objects, 4141 executable opcodes
+
+Compilation complete. 3 Errors, 12 Warnings, 1 Remarks, 1101 Optimizations
+
+***************************************************************************************************/
+
+DefinitionBlock (
+    "grammar.aml",      //Output filename
+    "DSDT",             //Signature
+    0x01,               //DSDT Revision ---> 32-bit table
+    "Intel",            //OEMID
+    "GRMTEST",          //TABLE ID
+    0x20090511          //OEM Revision
+    )
+{
+
+    External (\ABCD, UnknownObj)
+
+
+    /* Device with _STA and _INI */
+
+    Device (A1)
+    {
+        Method (_STA)
+        {
+            Return (0x0F)
+        }
+
+        Method (_INI)
+        {
+            Return
+        }
+    }
+
+    /* Device with no _STA, has _INI */
+
+    Device (A2)
+    {
+        Method (_INI)
+        {
+            Return
+        }
+    }
+
+    /* Device with _STA, no _INI */
+
+    Device (A3)
+    {
+        Method (_STA)
+        {
+            Return (0x0F)
+        }
+    }
+
+    /* Device with _STA and _INI, but not present */
+
+    Device (A4)
+    {
+        Method (_STA)
+        {
+            Return (Zero)
+        }
+
+        Method (_INI)
+        {
+            Return
+        }
+    }
+
+
+    /* Resource descriptors */
+
+    Device (IRES)
+    {
+        Name (PRT0, ResourceTemplate ()
+        {
+            IRQ (Edge, ActiveHigh, Exclusive) {3,4,5,6,7,9,10,11,14,15}
+
+            StartDependentFn (1,1)
+            {
+                IRQNoFlags () {0,1,2}
+            }
+            EndDependentFn ()
+        })
+
+        Method (_CRS, 0, NotSerialized)
+        {
+            Store ("_CRS:", Debug)
+            Store (PRT0, Debug)
+            Return (PRT0)
+        }
+
+        Method (_SRS, 1, Serialized)
+        {
+            Store ("_SRS:", Debug)
+            Store (Arg0, Debug)
+            Return (Zero)
+        }
+    }
+
+    Name (_NPK, Package ()
+    {
+        0x1111,
+        0x2222,
+        0x3333,
+        0x4444
+    })
+
+
+    Device (RES)
+    {
+        Name (_PRT, Package (0x04)
+        {
+            Package (0x04)
+            {
+                0x0002FFFF,
+                Zero,
+                Zero,
+                Zero
+            },
+
+            Package (0x04)
+            {
+                0x0002FFFF,
+                One,
+                Zero,
+                Zero
+            },
+
+            Package (0x04)
+            {
+                0x000AFFFF,
+                Zero,
+                Zero,
+                Zero
+            },
+
+            Package (0x04)
+            {
+                0x000BFFFF,
+                Zero,
+                Zero,
+                Zero
+            }
+        })
+
+        Method (_CRS, 0, NotSerialized)
+        {
+            Name (PRT0, ResourceTemplate ()
+            {
+                WordBusNumber (ResourceConsumer, MinFixed, MaxFixed, SubDecode,
+                    0x0000, // Address Space Granularity
+                    0xFFF2, // Address Range Minimum
+                    0xFFF3, // Address Range Maximum
+                    0x0032, // Address Translation Offset
+                    0x0002,,,)
+                WordBusNumber (ResourceProducer, MinFixed, MaxFixed, PosDecode,
+                    0x0000, // Address Space Granularity
+                    0x0000, // Address Range Minimum
+                    0x00FF, // Address Range Maximum
+                    0x0000, // Address Translation Offset
+                    0x0100,,,)
+                WordSpace (0xC3, ResourceConsumer, PosDecode, MinFixed, MaxFixed, 0xA5,
+                    0x0000, // Address Space Granularity
+                    0xA000, // Address Range Minimum
+                    0xBFFF, // Address Range Maximum
+                    0x0000, // Address Translation Offset
+                    0x2000,,,)
+                IO (Decode16, 0x0CF8, 0x0CFF, 0x01, 0x08)
+                WordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x0000, // Address Space Granularity
+                    0x0000, // Address Range Minimum
+                    0x0CF7, // Address Range Maximum
+                    0x0000, // Address Translation Offset
+                    0x0CF8,,,
+                    , TypeStatic)
+                WordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x0000, // Address Space Granularity
+                    0x0D00, // Address Range Minimum
+                    0xFFFF, // Address Range Maximum
+                    0x0000, // Address Translation Offset
+                    0xF300,,,
+                    , TypeStatic)
+                DWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x00000000, // Address Space Granularity
+                    0x00000000, // Address Range Minimum
+                    0x00000CF7, // Address Range Maximum
+                    0x00000000, // Address Translation Offset
+                    0x00000CF8,,,
+                    , TypeStatic)
+                DWordMemory (ResourceProducer, PosDecode, MinFixed, MaxFixed, Cacheable, ReadWrite,
+                    0x00000000, // Address Space Granularity
+                    0x000C8000, // Address Range Minimum
+                    0x000EFFFF, // Address Range Maximum
+                    0x00000000, // Address Translation Offset
+                    0x00028000,,,
+                    , AddressRangeMemory, TypeStatic)
+                DWordSpace (0xC3, ResourceConsumer, PosDecode, MinFixed, MaxFixed, 0xA5,
+                    0x00000000, // Address Space Granularity
+                    0x000C8000, // Address Range Minimum
+                    0x000EFFFF, // Address Range Maximum
+                    0x00000000, // Address Translation Offset
+                    0x00028000,,,)
+                QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000000000, // Address Range Minimum
+                    0x0000000000000CF7, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x0000000000000CF8, 0x44, "This is a ResouceSource string",
+                    , TypeStatic)
+                QWordIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000000000, // Address Range Minimum
+                    0x0000000000000CF7, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x0000000000000CF8,,,
+                    , TypeStatic)
+                QWordMemory (ResourceProducer, PosDecode, MinFixed, MaxFixed, Cacheable, ReadWrite,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000100000, // Address Range Minimum
+                    0x00000000FFDFFFFF, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x00000000FFD00000,,,
+                    , AddressRangeMemory, TypeStatic)
+                QWordSpace (0xC3, ResourceConsumer, PosDecode, MinFixed, MaxFixed, 0xA5,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000000000, // Address Range Minimum
+                    0x0000000000000CF7, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x0000000000000CF8,,,)
+                ExtendedIO (ResourceProducer, MinFixed, MaxFixed, PosDecode, EntireRange,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000000000, // Address Range Minimum
+                    0x0000000000000CF7, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x0000000000000CF8, // Address Length
+                    0x0000000000000000, // Type Specific Attributes
+                    , TypeStatic)
+                ExtendedMemory (ResourceProducer, PosDecode, MinFixed, MaxFixed, Cacheable, ReadWrite,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000100000, // Address Range Minimum
+                    0x00000000FFDFFFFF, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x00000000FFD00000, // Address Length
+                    0x0000000000000000, // Type Specific Attributes
+                    , AddressRangeMemory, TypeStatic)
+                ExtendedSpace (0xC3, ResourceProducer, PosDecode, MinFixed, MaxFixed, 0xA3,
+                    0x0000000000000000, // Address Space Granularity
+                    0x0000000000100000, // Address Range Minimum
+                    0x00000000FFDFFFFF, // Address Range Maximum
+                    0x0000000000000000, // Address Translation Offset
+                    0x00000000FFD00000, // Address Length
+                    0x0000000000000000) // Type Specific Attributes
+                IO (Decode16, 0x0010, 0x0020, 0x01, 0x10)
+                IO (Decode16, 0x0090, 0x00A0, 0x01, 0x10)
+                FixedIO (0x0061, 0x01)
+                IRQNoFlags () {2}
+                DMA (Compatibility, BusMaster, Transfer8_16) {4}
+                DMA (Compatibility, BusMaster, Transfer8) {2,5,7}
+                Memory32Fixed (ReadWrite, 0x00100000, 0x00000000)
+                Memory32Fixed (ReadOnly, 0xFFFE0000, 0x00020000)
+                Memory32 (ReadOnly, 0x00020000, 0xFFFE0000, 0x00000004, 0x00000200)
+                Memory24 (ReadOnly, 0x1111, 0x2222, 0x0004, 0x0200)
+                Interrupt (ResourceConsumer, Level, ActiveLow, Exclusive, 0xE, "\\_SB_.TEST")
+                {
+                    0x00000E01,
+                }
+                Interrupt (ResourceConsumer, Edge, ActiveHigh, Exclusive, 0x6, "xxxx")
+                {
+                    0x00000601,
+                    0x00000003,
+                    0x00000002,
+                    0x00000001,
+                }
+                Interrupt (ResourceProducer, Edge, ActiveHigh, Exclusive)
+                {
+                    0xFFFF0000,
+                    0x00000003,
+                    0x00000002,
+                    0x00000001,
+                    0x00000005,
+                    0x00000007,
+                    0x00000009,
+                }
+                VendorShort () {0x01, 0x02, 0x03}
+                VendorLong ()
+                {
+                    0x01, 0x02, 0x03, 0x04, 0x05, 0x06, 0x07, 0x08,
+                    0x09
+                }
+                Register (SystemIO, 0x08, 0x00, 0x00000000000000B2, , R000)
+                Register (SystemMemory, 0x08, 0x00, 0x00000000000000B2)
+                StartDependentFnNoPri ()
+                {
+                    IRQNoFlags () {0,1,2}
+                    IRQ (Level, ActiveLow, Shared) {3,4,5,6,7,9,10,11,14,15}
+                }
+                EndDependentFn ()
+            })
+            CreateWordField (PRT0, 0x08, BMIN)
+            CreateByteField (PRT0, R000._ASZ, RSIZ)
+            Store (0x03, BMIN)
+            Return (PRT0)
+        }
+
+        Method (_PRS, 0, NotSerialized)
+        {
+            Name (BUF0, ResourceTemplate ()
+            {
+                StartDependentFn (0x01, 0x02)
+                {
+                    IO (Decode16, 0x03D8, 0x03F8, 0x01, 0x08)
+                    IRQNoFlags () {4}
+                }
+                StartDependentFn (0x02, 0x01)
+                {
+                    IO (Decode16, 0x03D8, 0x03E8, 0x01, 0x08)
+                    IRQNoFlags () {4}
+                }
+                StartDependentFn (0x00, 0x02)
+                {
+                    IO (Decode16, 0x02E8, 0x02F8, 0x01, 0x08)
+                    IRQNoFlags () {3}
+                }
+                StartDependentFn (0x00, 0x02)
+                {
+                    IO (Decode16, 0x02D8, 0x02E8, 0x01, 0x08)
+                    IRQNoFlags () {3}
+                }
+                StartDependentFn (0x02, 0x00)
+                {
+                    IO (Decode16, 0x0100, 0x03F8, 0x08, 0x08)
+                    IRQNoFlags () {1,3,4,5,6,7,8,10,11,12,13,14,15}
+                }
+                EndDependentFn ()
+            })
+            Return (BUF0)
+        }
+
+        Method (_SRS, 1, Serialized)
+        {
+            Return (Zero)
+        }
+    }
+
+
+    Name(\_S0,Package(0x04){
+        0x00,
+        0x00,
+        0x00,
+        0x00
+    })
+    Name(\_S3,Package(0x04){
+        0x05,
+        0x05,
+        0x00,
+        0x00
+    })
+    Name(\_S4,Package(0x04){
+        0x06,
+        0x06,
+        0x00,
+        0x00
+    })
+    Name(\_S5,Package(0x04){
+        0x07,
+        0x07,
+        0x00,
+        0x00
+    })
+
+/* Examine this table header (DSDT) */
+
+/*
+    DataTableRegion (HDR, "DSDT", "", "")
+    Field (HDR, AnyAcc, NoLock, Preserve)
+    {
+        SIG,  32,
+        LENG, 32,
+        REV,  8,
+        SUM,  8,
+        OID,  48,
+        OTID, 64,
+        OREV, 32,
+        CID,  32,
+        CREV, 32
+    }
+
+    Method (SIZE)
+    {
+        If (LLess (REV, 2))
+        {
+            Store ("32-bit table", Debug)
+        }
+        else
+        {
+            Store ("64-bit table", Debug)
+        }
+        Return (0)
+    }
+
+*/
+    Name (SIZE, 0)
+
+    /* Custom operation region */
+
+    OperationRegion(MYOP,0x80,0xFD60,0x6)
+    Field(MYOP,ByteAcc,NoLock,Preserve)
+    {
+        MFLD,8
+    }
+
+    Method (TCOP)
+    {
+        Name (_STR, Unicode ("test"))
+        Store (4, MFLD)
+        Store (MFLD, Local0)
+    }
+
+    Name (ERRS, 0x0)
+
+    /* Warning should be issued for premature string termination */
+
+    NAME (ESC1, "abcdefg\x00hijklmn")
+    NAME (ESC2, "abcdefg\000hijklmn")
+    Name (ESC3, "abc\a\bdef\f\n\r\t\v\x03ffff\432")
+
+
+    Name(CRSA,ResourceTemplate()
+    {
+        WORDBusNumber(ResourceProducer,MinFixed,MaxFixed,PosDecode,0x0000,0x0019,0x001D,0x0000,0x0005)
+        WORDIO(ResourceProducer,MinFixed,MaxFixed,PosDecode,NonISAOnlyRanges,0x0000,0xC000,0xCFFF,0x0000,0x1000)
+        DWORDMemory(ResourceProducer,PosDecode,MinFixed,MaxFixed,NonCacheable,ReadWrite,0x00000000,0xD8000000,0xDBFFFFFF,0x00000000,0x04000000)
+
+    })
+    Name(CRSB,ResourceTemplate()
+    {
+        DWORDMemory(ResourceProducer,PosDecode,MinFixed,MaxFixed,NonCacheable,ReadWrite,0x00000000,0xD8000000,0xDBFFFFFF,0x00000000,0x04000000)
+
+    })
+
+    Name(CRSC,ResourceTemplate()
+    {
+        VendorShort () {0x1, 0x2, 0x3}
+    })
+    Name(CRSD,ResourceTemplate()
+    {
+        VendorLong (VNDL) {0x1, 0x2, 0x3, 0x4, 0x5, 0x6, 0x7, 0x8, 0x9}
+    })
+
+    Name(CRSE,ResourceTemplate()
+    {
+        IRQNoFlags(){3,4,10,11}
+        IRQNoFlags(xxxt){3,4,10,11}
+    })
+    Name(CRSR, Buffer (Add (SizeOf(CRSA),SizeOf(CRSB))){})
+    Method(_CRS,0,NotSerialized)
+    {
+        Return(CRSR)
+    }
+
+
+    //
+    // Unnamed scope
+    //
+    Scope (\)
+    {
+        Name(Bxxx,0xFFFFFFFF)
+    }
+
+    Name (LANS, 0x0)
+
+    PowerResource(LANP,1,0)
+    {
+        Method(_STA){
+            If(LEqual(And(LANS,0x30),0x30)){
+                Return(One)
+            } Else {
+                Return(Zero)
+            }
+        }
+        Method(_ON){
+            If(LNot(_STA())){
+                Store (0x30, LANS)
+            }
+        }
+        Method(_OFF){
+            If(_STA()){
+                Store (0, LANS)
+            }
+        }
+    }
+
+
+    /* Can a method define another method? */
+
+    /**********************************
+    Method (TASK, 2, SERIALIZED)
+    {
+        Sleep (100)
+
+        Method (TAS2)
+        {
+            Sleep (100)
+        }
+
+        TAS2()
+        Return
+
+    }
+    ************************************/
+
+    /* A recursive method */
+
+    Method (RCIV, 1)
+    {
+        Store (Arg0, Debug)
+        If (Lequal (Arg0, 0))
+        {
+            Return ()
+        }
+        RCIV (Subtract (Arg0, 1))
+    }
+
+    Method (RTOP)
+    {
+        RCIV (100)
+    }
+
+
+    Scope(\_PR)
+    {
+        Processor(CPU0,0x0,0xFFFFFFFF,0x0) {}
+    }
+
+    Name(B1TP,0xFFFFFFFF)
+
+    Name(B2TP,0xFFFFFFFF)
+    Name(ADPS,0xFFFFFFFF)
+    Name(B1PS,0xFFFFFFFF)
+    Name(B1RS,0xFFFFFFFF)
+    Name(B1CS,0xFFFFFFFF)
+    Name(B2PS,0xFFFFFFFF)
+    Name(B2RS,0xFFFFFFFF)
+    Name(B2CS,0xFFFFFFFF)
+    Name(B1DC,3000)
+    Name(B2DC,2600)
+    Name(B1LF,3000)
+    Name(B2LF,2600)
+    Name(BPIF,0)
+    Name(PBLL,0)
+
+    Name(RBIF,Package()
+    {
+        0x1,
+        2200,
+        2200,
+        0x1,
+        10800,
+        0,
+        0,
+        1,
+        1,
+        "CA54200-5003/5",
+        "1",
+        "LION",
+        "Fujitsu"
+    })
+
+    Method(SMWE, 4)
+    {
+       return(ONES)
+    }
+
+    Method(SMRE, 4)
+    {
+       return(ONES)
+    }
+
+/*
+    Method(RDBT,0,Serialized){
+        If(LNot(SMWE(0x09,0x15,1,1))){
+                    Store(0x18,Local2)
+            }
+    }
+*/
+    Scope(_SB)
+    {
+
+        Name (SBUF, Buffer (128) {})
+
+        CreateBitField (SBUF, 3, BITY)
+        CreateByteField (SBUF, 1, BYTY)
+        CreateWordField (SBUF, 2, WRDZ)
+        CreateDwordField (SBUF, 4, DWDZ)
+        CreateQwordField (SBUF, 8, QWDZ)
+        CreateField (SBUF, 128, 12, FLDZ)
+        CreateField (SBUF, 148, 96, FLDY)
+        CreateField (SBUF, 148, 96, \_SB_.FLDW)
+
+        Method (_INI)
+        {
+            CreateField (\_SB_.SBUF, 148, 96, FLDV)
+        }
+
+
+        Device(PCI0)
+        {
+            Name(_HID,EISAID("PNP0A03"))
+            Name(_ADR,0x0)
+
+            Method(_CRS)
+            {
+                Name(PRT0, ResourceTemplate() {
+                    WORDBusNumber(                          // Bus number resource(0)
+                            ResourceConsumer,               // bit 0 of general flags is 1
+                            MinFixed,                       // Range is notfixed
+                            MaxFixed,                       // Range is not fixed
+                            SubDecode,                      // SubDecode
+                            0x0000,                           // Granularity
+                            0xfff1,                           // Min
+                            0xfff2,                           // Max
+                            0x0032,                           // Translation
+                            0x0002,,,                         // Range Length
+                            BUS0
+                    ) } )// PRT0
+
+                CreateWordField(PRT0, BUS0._MIN, BMIN)          //Minimum bus number suported under this bridge.
+
+                Store(3, BMIN)
+                Return(PRT0)
+
+            } // _CRS
+
+            Method(_SRS)
+            {
+                Return ()
+            }
+
+            Device(EIO)
+            {
+                OperationRegion(FJIO,SystemIO,0xFD60,0x6)
+                    Field(FJIO,ByteAcc,NoLock,Preserve)
+                    {
+                        GIDX,8,
+
+                        GDTA,8,
+
+                        PIDX,8,
+
+                        PDTA,8,
+
+                        SIDX,8,
+
+                        SDTA,8
+                    }
+                    IndexField(GIDX,GDTA,ByteAcc,NoLock,Preserve)
+                    {
+                        Offset(0x2),
+                         ,5,
+                        VGAS,2,
+                        Offset(0x4),
+                         ,4,
+                        DCKE,1,
+                        Offset(0x5),
+                         ,6,
+                        ACPW,1,
+
+                        Offset(0xA),
+                        B1P,1,
+
+                        B2P,1,
+
+                        B1C,1,
+
+                        B2C,1,
+
+                        B1ER,1,
+
+                        B2ER,1,
+
+                        Offset(0xB),
+                        B1CP,8,
+
+                        B2CP,8,
+
+                        BCP,8,
+
+                        B1VH,8,
+
+                        B1VL,8,
+
+                        B2VH,8,
+
+                        B2VL,8,
+
+                        B1TM,8,
+
+                        B2TM,8,
+
+                        B1CH,8,
+
+                        B1CL,8,
+
+                        B2CH,8,
+
+                        B2CL,8
+                    }
+                }
+            }
+        }
+
+        Method(RDBT,3,Serialized){
+            Store(0x1FFF,Local1)
+            If( Arg0 ){
+                Store(0x2FFF,Local1)
+            }
+            Store(0x18,Local2)
+            If( Arg1 ){
+                Store(0x10,Local2)
+            }
+            If(LNot(SMRE(0x09,0x15,1,RefOf(Local0)))){
+                If(LNot(SMWE(0x08,0x14,1,Local1))){
+                    If(LNot(SMRE(0x09,0x17,Local2,RefOf(Local3)))){
+                        Store(Local1,Arg2)
+                    }
+                }
+                Or(Local0,0xFFF,Local0)
+                SMWE(0x08,0x14,1,Local0)
+            }
+        }
+        Method(MKWD,2)
+        {
+            If(And(Arg1,0x80)) {
+                Or(0xFFFF0000,Arg0,Local0)
+                Or(Local0,ShiftLeft(Arg1,8),Local0)
+                Subtract(Zero,Local0,Local0)
+            } else {
+                Store(Arg0,Local0)
+                Or(Local0,ShiftLeft(Arg1,8),Local0)
+            }
+            Return(Local0)
+        }
+
+        Device(CMB1)
+        {
+            Name(_HID,EISAID("PNP0C0A"))
+            Name(_UID,0x1)
+            Alias(\_SB.PCI0.EIO.B1P,\_SB_.PCI0.XXXX)
+            Alias(\_SB.PCI0.EIO.B1P,B1P)
+            Alias(\_SB.PCI0.EIO.B1C,B1C)
+            Alias(\_SB.PCI0.EIO.B1CH,B1CH)
+            Alias(\_SB.PCI0.EIO.B1CL,B1CL)
+            Alias(\_SB.PCI0.EIO.B1VH,B1VH)
+            Alias(\_SB.PCI0.EIO.B1VL,B1VL)
+            Alias(\_SB.PCI0.EIO.B1CP,B1CP)
+
+            Method(_INI)
+                        {
+                Store(B1P, B1PS)
+                Store(B1CP,B1RS)
+                Store(B1C, B1CS)
+            }
+
+            Method(_BIF){
+                RDBT(Zero,Zero,RefOf(B1DC))
+                RDBT(Zero,One,RefOf(B1LF))
+                Store(B1DC,Index(RBIF,1))
+                Store(B1LF,Index(RBIF,2))
+                Store("CA54200-5003/5",Index(RBIF,9))
+                Store("1",Index(RBIF,10))
+                Return(RBIF)
+            }
+
+            Method(_BST) {
+
+                _INI()
+
+                Store(Zero,Local0)
+
+                if (LAnd(B1P,LNot(B1C))){
+                    Or(Local0,1,Local0)
+                }
+
+                if (LAnd(B1P,B1C)) {
+                    Or(Local0,2,Local0)
+                }
+
+                if (LLessEqual(B1CP,1)) {
+                    Or(Local0,4,Local0)
+                }
+
+                Store(MKWD(B1CL,B1CH),Local1)
+
+                Store(Divide(Add(Multiply(B1CP,B1LF),99),100),Local2)
+
+                Store(MKWD(B1VL,B1VH),Local3)
+
+                Name(STAT,Package(4){})
+                Store(Local0,Index(STAT,0))
+                Store(Local1,Index(STAT,1))
+                Store(Local2,Index(STAT,2))
+                Store(Local3,Index(STAT,3))
+
+                If(LNot(BPIF)){
+//                    \_SB.PCI0.EIO.EC0.IECT()
+//                    \_SB.PCI0.EIO.EC0.SECT()
+                    Store(One,BPIF)
+                }
+                return(STAT)
+            }
+
+        }
+
+    Device (DEV1)
+    {
+    }
+
+    Scope(\_TZ)
+    {
+        ThermalZone(TZ1)
+        {
+            Name(_PSL,Package()
+            {
+                \_PR.CPU0
+            })
+        }
+    }
+
+    Method (TZ2, 0, SERIALIZED)
+    {
+        Name(_PSL,Package()
+        {
+            \_PR.CPU0
+        })
+
+        Return (_PSL)
+    }
+
+    ThermalZone (THM1)
+    {
+    }
+
+    Method (NOTI)
+    {
+        Notify (\DEV1, 0)
+        Notify (\THM1, 0)
+        Notify (\_PR.CPU0, 0)
+    }
+
+    Method (_ERR, 2)
+    {
+        Increment (ERRS)
+        Store ("Run-time exception:", Debug)
+        Store (Arg0, Debug)
+        Store (Arg1, Debug)
+
+        Return (0)          // Map error to AE_OK
+    }
+
+    Method (DIV0)
+    {
+        Store (1, Local0)
+        Store (0, Local1)
+        Divide (Local0, Local1, Local3)
+
+        Store ("DIV0 - noabort", Debug)
+    }
+
+    Method (ERR_, 1)
+    {
+        if (LEqual (Arg0, 0))
+        {
+            Store ("+*+*+*+* MTHD_ERROR: Results not equal!", Debug)
+        }
+        if (LEqual (Arg0, 1))
+        {
+            Store ("+*+*+*+* MTHD_ERROR: Numeric result is incorrect!", Debug)
+        }
+        if (LEqual (Arg0, 2))
+        {
+            Store ("+*+*+*+* MTHD_ERROR: Operand was clobbered!", Debug)
+        }
+
+        Notify (DEV1, Arg0)
+        Increment (ERRS)
+    }
+
+    Method (R226, 2)
+    {
+    }
+    Method (R225, 2)
+    {
+        R226 (Arg0, Arg1)
+    }
+    Method (R224, 2)
+    {
+        R225 (Arg1, Arg0)
+    }
+    Method (R223, 2)
+    {
+        R224 (Arg0, Arg1)
+    }
+    Method (R222, 2)
+    {
+        R223 (Arg1, Arg0)
+    }
+    Method (R111)
+    {
+        Store (0x01010101, Local0)
+        R222 (0xABAB, Local0)
+        Store (Local0, Local1)
+    }
+
+    Method (MAIN)
+    {
+
+//      SIZE()
+        Store (NUM1(), Local0)
+        \CMB1._BST()
+        RDBT(1,2,3)
+        OBJ1(1)
+        OBJ2(2)
+        CHEK()
+        RETZ()
+        BITZ()
+        LOGS()
+        REFS()
+        COND()
+        TZ2()
+
+        //
+        // iPCO tests added
+        //
+        Store (\IFEL.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\NOSV.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\IDXF.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.NSTL.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\RTBF.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.RTLV.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.RETP.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\WHLR.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\ANDO.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\BRKP.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\ADSU.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\INDC.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\LOPS.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\FDSO.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\MLDV.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\NBIT.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\SHFT.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\XORD.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\CRBF.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\IDX4.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\EVNT.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\SZLV.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.BYTF.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\DWDF.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\DVAX.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\IDX6.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\IDX5.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.IDX0.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.IDX3.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\IDX7.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\MTCH.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\WHLB.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.IDX2.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\SIZO.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        Store (\_SB_.SMIS.TEST(), Local0)
+        if (LGreater (Local0, 0))
+        {
+            ERR_ (1)
+            Return(Local0)
+        }
+
+        if (LGreater (ERRS, 0))
+        {
+            Store ("****** There were errors during the execution of the test ******", Debug)
+        }
+
+        //
+        // Last Test
+        //
+
+        Return(0) // Success
+    }
+
+
+    Method (OBJ1, 1, SERIALIZED)
+    {
+
+        Store (3, Local0)
+        Name(BUFR, Buffer (Local0) {})
+        Name(BUF1, Buffer (4) {1,2,3,4})
+        Name(BUF2, Buffer (4) {})
+
+        Store (BUF1, BUF2)
+        Mutex (MTX1, 4)
+
+        Alias (MTX1, MTX2)
+    }
+
+
+    Mutex (MTXT, 0)
+    Mutex (MTXX, 0)
+
+    /*
+     * Field Creation
+     */
+
+    Method (FLDS)
+    {
+        Store ("++++++++ Creating BufferFields", Debug)
+        Name (BUF2, Buffer (128) {})
+
+        CreateBitField (BUF2, 3, BIT2)
+        CreateByteField (BUF2, 1, BYT2)
+        CreateWordField (BUF2, 2, WRD2)
+        CreateDwordField (BUF2, 4, DWD2)
+        CreateQwordField (BUF2, 8, QWD2)
+        CreateField (BUF2, 128, 12, FLD2)
+        CreateField (BUF2, 148, 96, FLD3)
+
+        Store (0x1, BIT2)
+        Store (BIT2, Local0)
+        if (LNotEqual (Local0, 0x1))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store (DerefOf (Index (BUF2, 0)), Local0)
+            if (LNotEqual (Local0, 0x08))
+            {
+                ERR_ (1)
+            }
+            else
+            {
+                Store ("++++++++ Bit BufferField I/O PASS", Debug)
+            }
+        }
+
+        Store (0x1A, BYT2)
+        Store (BYT2, Local0)
+        if (LNotEqual (Local0, 0x1A))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store ("++++++++ Byte BufferField I/O PASS", Debug)
+        }
+
+        Store (0x1234, WRD2)
+        Store (WRD2, Local0)
+        if (LNotEqual (Local0, 0x1234))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store ("++++++++ Word BufferField I/O PASS", Debug)
+        }
+
+        Store (0x123, FLD2)
+        Store (FLD2, Local0)
+        if (LNotEqual (Local0, 0x123))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store ("++++++++ 12-bit BufferField I/O PASS", Debug)
+        }
+
+        Store (0x12345678, DWD2)
+        Store (DWD2, Local0)
+        if (LNotEqual (Local0, 0x12345678))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store ("++++++++ Dword BufferField I/O PASS", Debug)
+        }
+
+        Store (0x1234567887654321, QWD2)
+        Store (QWD2, Local0)
+        if (LNotEqual (Local0, 0x1234567887654321))
+        {
+            ERR_ (1)
+        }
+        else
+        {
+            Store ("++++++++ Qword BufferField I/O PASS", Debug)
+        }
+    }
+
+
+    /* Field execution */
+
+    Method (FLDX)
+    {
+        Field (\_SB_.MEM.SMEM, AnyAcc, NoLock, Preserve)
+        {   //  Field:  SMEM overlay using 32-bit field elements
+            SMD0,   32, //  32-bits
+            SMD1,   32,     //  32-bits
+            SMD2,   32,     //  32-bits
+            SMD3,   32  //  32-bits
+        }   //  Field:  SMEM overlay using 32-bit field elements
+        Field (\_SB_.MEM.SMEM, AnyAcc, NoLock, Preserve)
+        {   //  Field:  SMEM overlay using greater than 32-bit field elements
+            SME0,   69, //  larger than an integer (32 or 64)
+            SME1,   97  //  larger than an integer
+        }   //  Field:  SMEM overlay using greater than 32-bit field elements
+    }
+
+
+    Method (MTX_, )
+    {
+        /* Test "Force release" of mutex on method exit */
+
+        Acquire (MTXT, 0xFFFF)
+        Acquire (MTXX, 0xFFFF)
+
+        Store ("++++++++ Acquiring Mutex MTX2", Debug)
+        Acquire (_GL_, 0xFFFF)
+
+
+        Store ("++++++++ Releasing Mutex MTX2", Debug)
+        Release (_GL_)
+    }
+
+
+    Method (OBJ2, 1)
+    {
+        Store ("++++++++ Creating Buffer BUFO", Debug)
+        Name (BUFO, Buffer (32) {})
+
+        Store ("++++++++ Creating OpRegion OPR2", Debug)
+        OperationRegion (OPR2, SystemMemory, Arg0, 256)
+
+        Store ("++++++++ Creating Field(s) in OpRegion OPR2", Debug)
+        Field (OPR2, ByteAcc, NoLock, Preserve)
+        {
+            IDX2, 8,
+            DAT2, 8,
+            BNK2, 4
+        }
+
+        Store ("++++++++ Creating BankField BNK2 in OpRegion OPR2", Debug)
+        //
+        // mcw 3/20/00 - changed FET0, 4, FET1, 3 to FET0, 1, FET1, 1
+        //
+        BankField (OPR2, BNK2, 0, ByteAcc, NoLock, Preserve)
+        {
+            FET0, 4,
+            FET1, 3
+        }
+
+        Store ("++++++++ Creating IndexField", Debug)
+        IndexField (IDX2, DAT2, ByteAcc, NoLock, Preserve)
+        {
+            FET2, 4,
+            FET3, 3
+        }
+
+        Store ("++++++++ SizeOf (BUFO)", Debug)
+        SizeOf (BUFO)
+
+        Store ("++++++++ Store (SizeOf (BUFO), Local0)", Debug)
+        Store (SizeOf (BUFO), Local0)
+
+        Store ("++++++++ Concatenate (\"abd\", \"def\", Local0)", Debug)
+        Concatenate ("abd", "def", Local0)
+        Store (Local0, Debug)
+
+        Store ("++++++++ Concatenate (\"abd\", 0x7B, Local0)", Debug)
+        Concatenate ("abd", 0x7B, Local0)
+        Store (Local0, Debug)
+
+        Store ("++++++++ Creating Event EVT2", Debug)
+        Event (EVT2)
+
+        Store ("++++++++ Creating Mutex MTX2", Debug)
+        Mutex (MTX2, 0)
+
+        Store ("++++++++ Creating Alias MTXA to MTX2", Debug)
+        Alias (MTX2, MTXA)
+
+        Store ("++++++++ Acquiring Mutex MTX2", Debug)
+        Acquire (MTX2, 0xFFFF)
+
+        Store ("++++++++ Acquiring Mutex MTX2 (2nd acquire)", Debug)
+        Acquire (MTX2, 1)
+
+        Store ("++++++++ Releasing Mutex MTX2", Debug)
+        Release (MTX2)
+
+        // Type 1 opcodes
+
+        Store ("++++++++ Signalling Event EVT2", Debug)
+        Signal (EVT2)
+
+        Store ("++++++++ Resetting Event EVT2", Debug)
+        Reset (EVT2)
+
+        Store ("++++++++ Signalling Event EVT2", Debug)
+        Signal (EVT2)
+
+        Store ("++++++++ Waiting Event EVT2", Debug)
+        Wait (EVT2, 0xFFFF)
+
+        Store ("++++++++ Sleep", Debug)
+        Sleep (100)
+
+        Store ("++++++++ Stall", Debug)
+        Stall (254)
+
+        Store ("++++++++ NoOperation", Debug)
+        Noop
+
+        // Type 2 Opcodes
+
+        Store ("++++++++ Return from Method OBJ2", Debug)
+        return (4)
+    }
+
+
+    Method (NUM1, 0)
+    {
+        /* ADD */
+
+        Store ("++++++++ Add (0x12345678, 0x11111111, Local0)", Debug)
+        Add (0x12345678, 0x11111111, Local0)
+
+        Store ("++++++++ Store (Add (0x12345678, 0x11111111), Local1)", Debug)
+        Store (Add (0x12345678, 0x11111111), Local1)
+
+        Store ("++++++++ Checking result from ADD", Debug)
+        if (LNotEqual (Local0, Local1))
+        {
+            ERR_ (0)
+        }
+
+
+        /* SUBTRACT */
+
+        Store ("++++++++ Subtract (0x87654321, 0x11111111, Local4)", Debug)
+        Subtract (0x87654321, 0x11111111, Local4)
+
+        Store ("++++++++ Store (Subtract (0x87654321, 0x11111111), Local5)", Debug)
+        Store (Subtract (0x87654321, 0x11111111), Local5)
+
+        Store ("++++++++ Checking result from SUBTRACT", Debug)
+        if (LNotEqual (Local4, Local5))
+        {
+            ERR_ (0)
+        }
+
+
+        /* MULTIPLY */
+
+        Store ("++++++++ Multiply (33, 10, Local6)", Debug)
+        Multiply (33, 10, Local6)
+
+        Store ("++++++++ Store (Multiply (33, 10), Local7)", Debug)
+        Store (Multiply (33, 10), Local7)
+
+
+        Store ("++++++++ Checking result from MULTIPLY", Debug)
+        if (LNotEqual (Local6, Local7))
+        {
+            ERR_ (0)
+        }
+
+
+        /* DIVIDE */
+
+        Store ("++++++++ Divide (100, 9, Local1, Local2)", Debug)
+        Divide (100, 9, Local1, Local2)
+
+        Store ("++++++++ Store (Divide (100, 9), Local3)", Debug)
+        Store (Divide (100, 9), Local3)
+
+        Store ("++++++++ Checking (quotient) result from DIVIDE", Debug)
+        if (LNotEqual (Local2, Local3))
+        {
+            ERR_ (0)
+        }
+
+
+        /* INCREMENT */
+
+        Store ("++++++++ Increment (Local0)", Debug)
+        Store (1, Local0)
+        Store (2, Local1)
+        Increment (Local0)
+
+        Store ("++++++++ Checking result from INCREMENT", Debug)
+        if (LNotEqual (Local0, Local1))
+        {
+            ERR_ (0)
+        }
+
+
+        /* DECREMENT */
+
+        Store ("++++++++ Decrement (Local0)", Debug)
+        Store (2, Local0)
+        Store (1, Local1)
+        Decrement (Local0)
+
+        Store ("++++++++ Checking result from DECREMENT", Debug)
+        if (LNotEqual (Local0, Local1))
+        {
+            ERR_ (0)
+        }
+
+
+        /* TOBCD */
+        /* FROMBCD */
+
+        Store ("++++++++ ToBCD (0x1234, Local5)", Debug)
+        ToBCD (0x1234, Local5)
+
+        Store ("++++++++ FromBCD (Local5, Local6)", Debug)
+        FromBCD (Local5, Local6)
+
+        Store ("++++++++ Return (Local6)", Debug)
+        Return (Local6)
+    }
+
+
+    Method (CHEK)
+    {
+
+        Store (3, Local0)
+        Store (3, Debug)
+        Store (Local0, Debug)
+        Store (7, Local1)
+
+        Add (Local0, Local1)
+        if (LNotEqual (Local0, 3))
+        {
+            ERR_ (2)
+        }
+        if (LNotEqual (Local1, 7))
+        {
+            ERR_ (2)
+        }
+
+
+        Add (Local0, Local1, Local2)
+        if (LNotEqual (Local0, 3))
+        {
+            ERR_ (2)
+        }
+        if (LNotEqual (Local1, 7))
+        {
+            ERR_ (2)
+        }
+    }
+
+
+    Method (RET1)
+    {
+        Store (3, Local0)
+        Return (Local0)
+    }
+
+    Method (RET2)
+    {
+        Return (RET1())
+    }
+
+    Method (RETZ)
+    {
+        RET2 ()
+    }
+
+
+    Method (BITZ)
+    {
+        Store ("++++++++ FindSetLeftBit (0x00100100, Local0)", Debug)
+        FindSetLeftBit (0x00100100, Local0)
+        if (LNotEqual (Local0, 21))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ FindSetRightBit (0x00100100, Local1)", Debug)
+        FindSetRightBit (0x00100100, Local1)
+        if (LNotEqual (Local1, 9))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ And (0xF0F0F0F0, 0x11111111, Local2)", Debug)
+        And (0xF0F0F0F0, 0x11111111, Local2)
+        if (LNotEqual (Local2, 0x10101010))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ NAnd (0xF0F0F0F0, 0x11111111, Local3)", Debug)
+        NAnd (0xF0F0F0F0, 0x11111111, Local3)
+        if (LNotEqual (Local3, 0xEFEFEFEF))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ Or (0x11111111, 0x22222222, Local4)", Debug)
+        Or (0x11111111, 0x22222222, Local4)
+        if (LNotEqual (Local4, 0x33333333))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ NOr (0x11111111, 0x22222222, Local5)", Debug)
+        NOr (0x11111111, 0x22222222, Local5)
+        if (LNotEqual (Local5, 0xCCCCCCCC))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ XOr (0x11113333, 0x22222222, Local6)", Debug)
+        XOr (0x11113333, 0x22222222, Local6)
+        if (LNotEqual (Local6, 0x33331111))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ ShiftLeft (0x11112222, 2, Local7)", Debug)
+        ShiftLeft (0x11112222, 2, Local7)
+        if (LNotEqual (Local7, 0x44448888))
+        {
+            ERR_ (1)
+        }
+
+        Store ("++++++++ ShiftRight (Local7, 2, Local7)", Debug)
+        ShiftRight (Local7, 2, Local7)
+        if (LNotEqual (Local7, 0x11112222))
+        {
+            ERR_ (1)
+        }
+
+
+        Store ("++++++++ Not (Local0, Local1)", Debug)
+        Store (0x22224444, Local0)
+        Not (Local0, Local1)
+        if (LNotEqual (Local0, 0x22224444))
+        {
+            ERR_ (2)
+        }
+
+        if (LNotEqual (Local1, 0xDDDDBBBB))
+        {
+            ERR_ (1)
+        }
+
+        Return (Local7)
+    }
+
+
+    Method (LOGS)
+    {
+
+        Store ("++++++++ Store (LAnd (0xFFFFFFFF, 0x11111111), Local0)", Debug)
+        Store (LAnd (0xFFFFFFFF, 0x11111111), Local0)
+
+        Store ("++++++++ Store (LEqual (0xFFFFFFFF, 0x11111111), Local)", Debug)
+        Store (LEqual (0xFFFFFFFF, 0x11111111), Local1)
+
+        Store ("++++++++ Store (LGreater (0xFFFFFFFF, 0x11111111), Local2)", Debug)
+        Store (LGreater (0xFFFFFFFF, 0x11111111), Local2)
+
+        Store ("++++++++ Store (LGreaterEqual (0xFFFFFFFF, 0x11111111), Local3)", Debug)
+        Store (LGreaterEqual (0xFFFFFFFF, 0x11111111), Local3)
+
+        Store ("++++++++ Store (LLess (0xFFFFFFFF, 0x11111111), Local4)", Debug)
+        Store (LLess (0xFFFFFFFF, 0x11111111), Local4)
+
+        Store ("++++++++ Store (LLessEqual (0xFFFFFFFF, 0x11111111), Local5)", Debug)
+        Store (LLessEqual (0xFFFFFFFF, 0x11111111), Local5)
+
+        Store ("++++++++ Store (LNot (0x31313131), Local6)", Debug)
+        Store (0x00001111, Local6)
+        Store (LNot (Local6), Local7)
+        if (LNotEqual (Local6, 0x00001111))
+        {
+            ERR_ (2)
+        }
+
+        if (LNotEqual (Local7, 0x0))
+        {
+            ERR_ (1)
+        }
+
+
+        Store ("++++++++ Store (LNotEqual (0xFFFFFFFF, 0x11111111), Local7)", Debug)
+        Store (LNotEqual (0xFFFFFFFF, 0x11111111), Local7)
+
+        Store ("++++++++ Lor (0x0, 0x1)", Debug)
+        if (Lor (0x0, 0x1))
+        {
+            Store ("+_+_+_+_+ Lor (0x0, 0x1) returned TRUE", Debug)
+        }
+
+        return (Local7)
+    }
+
+
+    Method (COND)
+    {
+        Store ("++++++++ Store (0x4, Local0)", Debug)
+        Store (0x4, Local0)
+
+        Store ("++++++++ While (Local0)", Debug)
+        While (Local0)
+        {
+            Store ("++++++++ Decrement (Local0)", Debug)
+            Decrement (Local0)
+        }
+
+
+        Store ("++++++++ Store (0x3, Local6)", Debug)
+        Store (0x3, Local6)
+
+        Store ("++++++++ While (Subtract (Local6, 1))", Debug)
+        While (Subtract (Local6, 1))
+        {
+            Store ("++++++++ Decrement (Local6)", Debug)
+            Decrement (Local6)
+        }
+
+
+        Store ("++++++++ [LVL1] If (LGreater (0x2, 0x1))", Debug)
+        If (LGreater (0x2, 0x1))
+        {
+            Store ("++++++++ [LVL2] If (LEqual (0x11111111, 0x22222222))", Debug)
+            If (LEqual (0x11111111, 0x22222222))
+            {
+                Store ("++++++++ ERROR: If (LEqual (0x11111111, 0x22222222)) returned TRUE", Debug)
+            }
+
+            else
+            {
+                Store ("++++++++ [LVL3] If (LNot (0x0))", Debug)
+                If (LNot (0x0))
+                {
+                    Store ("++++++++ [LVL4] If (LAnd (0xEEEEEEEE, 0x2))", Debug)
+                    If (LAnd (0xEEEEEEEE, 0x2))
+                    {
+                        Store ("++++++++ [LVL5] If (LLess (0x44444444, 0x3))", Debug)
+                        If (LLess (0x44444444, 0x3))
+                        {
+                            Store ("++++++++ ERROR: If (LLess (0x44444444, 0x3)) returned TRUE", Debug)
+                        }
+
+                        else
+                        {
+                            Store ("++++++++ Exiting from nested IF/ELSE statements", Debug)
+                        }
+                    }
+                }
+            }
+        }
+
+
+        Store ("++++++++ [LVL1] If (LGreater (0x2, 0x1))", Debug)
+        If (LGreater (0x2, 0x1))
+        {
+            Store ("++++++++ [LVL2] If (LEqual (0x11111111, 0x22222222))", Debug)
+            If (LEqual (0x11111111, 0x22222222))
+            {
+                Store ("++++++++ ERROR: If (LEqual (0x11111111, 0x22222222)) returned TRUE", Debug)
+            }
+
+            else
+            {
+                Store ("++++++++ [LVL3] If (LNot (0x0))", Debug)
+                If (LNot (0x0))
+                {
+                    Store ("++++++++ [LVL4] If (LAnd (0xEEEEEEEE, 0x2))", Debug)
+                    If (LAnd (0xEEEEEEEE, 0x2))
+                    {
+                        Store ("++++++++ [LVL5] If (LLess (0x44444444, 0x3))", Debug)
+                        If (LLess (0x44444444, 0x3))
+                        {
+                            Store ("++++++++ ERROR: If (LLess (0x44444444, 0x3)) returned TRUE", Debug)
+                        }
+
+                        else
+                        {
+                            Store ("++++++++ Returning from nested IF/ELSE statements", Debug)
+                            Return (Local6)
+                        }
+                    }
+                }
+            }
+        }
+
+    }
+
+
+    Method (REFS)
+    {
+        Name (BBUF, Buffer() {0xB0, 0xB1, 0xB2, 0xB3, 0xB4, 0xB5, 0xB6, 0xB7})
+
+        Name (NEST, Package ()
+        {
+            Package ()
+            {
+                0x01, 0x02, 0x03, 0x04, 0x05, 0x06
+            },
+            Package ()
+            {
+                0x11, 0x12, 0x12, 0x14, 0x15, 0x16
+            }
+        })
+
+    /* Parser thinks this is a method invocation!! */
+
+        Store (RefOf (MAIN), Local5)
+
+        // For this to work, ABCD must NOT exist.
+
+        Store (CondRefOf (ABCD, Local0), Local1)
+        if (LNotEqual (Local1, 0))
+        {
+            ERR_ (2)
+        }
+
+        Store (CondRefOf (BBUF, Local0), Local1)
+        if (LNotEqual (Local1, Ones))
+        {
+            ERR_ (2)
+        }
+
+        Store (DeRefOf (Index (BBUF, 3)), Local6)
+        if (LNotEqual (Local6, 0xB3))
+        {
+            ERR_ (2)
+        }
+
+        Store (DeRefOf (Index (DeRefOf (Index (NEST, 1)), 3)), Local0)
+        if (LNotEqual (Local0, 0x14))
+        {
+            ERR_ (2)
+        }
+
+
+        Store (0x11223344, Local0)
+        Store (RefOf (Local0), Local1)
+
+        Store (DerefOf (Local1), Local2)
+        If (LNotEqual (Local2, 0x11223344))
+        {
+            ERR_ (2)
+        }
+
+
+    /* Parser thinks this is a method invocation!! */
+
+    //  RefOf (MAIN)
+
+
+    //  RefOf (R___)
+    //  RefOf (BBUF)
+
+    //  Store (RefOf (Local0), Local1)
+
+    //  CondRefOf (BBUF, Local2)
+    //  CondRefOf (R___, Local3)
+
+    //  Store (DerefOf (Local1), Local4)
+
+    //  Return (Local4)
+    }
+
+
+    Method (INDX, 0)
+    {
+        Name(STAT,Package(4){})
+        Store(0x44443333,Index(STAT,0))
+    }
+
+//=================================================================
+//=================================================================
+//===================== iPCO TESTS ================================
+//=================================================================
+//=================================================================
+//
+//
+// test IfElseOp.asl
+//
+//  test for IfOp and ElseOp, including validation of object stack cleanup
+//
+    Device (IFEL)
+    {
+        Name (DWRD, 1)
+        Name (RSLT, 0)
+
+        //  IFNR control method executes IfOp branch with NO nested Return
+        //  and no Else branch
+        Method (IFNR)
+        {
+            Store (DWRD, RSLT)
+            If (LEqual (DWRD, 1))
+            {
+                Store (0, RSLT)
+            }
+        }   //  IFNR
+
+        //  NINR control method does not execute If branch and has no Else branch
+        Method (NINR)
+        {
+            Store (0, RSLT)
+            If (LNotEqual (DWRD, 1))
+            {
+                Store (DWRD, RSLT)
+            }
+        }   //  NINR
+
+        //  IENR control method executes IfOp branch with NO nested Return
+        Method (IENR)
+        {
+            If (LEqual (DWRD, 1))
+            {
+                Store (0, RSLT)
+            }
+            Else
+            {
+                Store (DWRD, RSLT)
+            }
+        }   //  IENR
+
+        //  ELNR control method executes ElseOp branch with NO nested Return
+        Method (ELNR)
+        {
+            If (LNotEqual (DWRD, 1))
+            {
+                Store (DWRD, RSLT)
+            }
+            Else
+            {
+                Store (0, RSLT)
+            }
+        }   //  ELNR
+
+        //  IFRT control method executes IfOp branch with nested Return with
+        //  no Else branch
+        Method (IFRT)
+
+        {
+            If (LEqual (DWRD, 1))
+            {
+                Return (0)
+            }
+            Return (DWRD)
+        }   //  IFRT
+
+        //  IERT control method executes IfOp branch with nested Return with
+        //  Else branch
+        Method (IERT)
+        {
+            If (LEqual (DWRD, 1))
+            {
+                Return (0)
+            }
+            Else
+            {
+                Return (DWRD)
+            }
+        }   //  IERT
+
+        //  ELRT control method executes ElseOp branch with nested Return
+        Method (ELRT)
+        {
+            If (LNotEqual (DWRD, 1))
+            {
+                Return (DWRD)
+            }
+            Else
+            {
+                Return (0)
+            }
+        }   //  ELRT
+
+        Method (TEST)
+        {
+            Store ("++++++++ IfElseOp Test", Debug)
+
+            //  IfOp with NO return value
+            IFNR()
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  IfOp with NO return value
+            NINR()
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  IfOp with NO return value
+            IENR()
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  ElseOp with NO return value
+            ELNR()
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  IfOp with return value
+            Store (IFRT, RSLT)
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  IfOp with return value
+            Store (IERT, RSLT)
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            //  ElseOp with return value
+            Store (ELRT, RSLT)
+            If (LNotEqual (RSLT, 0))
+            {
+                Return (RSLT)
+            }
+
+            Return (0)
+        }   //  TEST
+    }   //  IFEL
+
+//
+// test NoSave.asl
+//
+//
+//  Internal test cases to validate IfOp (Operator (,,)) where Operator
+//  target is ZeroOp to throw away the results.
+//  Includes internal test cases for logical operators with no destination
+//  operands.
+//
+    Device (NOSV)
+    {
+        Method (TEST)
+        {
+            Store ("++++++++ NoSave Test", Debug)
+
+            Name (WRD, 0x1234)
+
+            //
+            //  Begin test of nested operators without saving results
+            //
+
+            //  Test If (And ()) with no save of And result
+            If (And (3, 1, ))
+            {
+                Store (1, WRD)  //  pass -- just do something
+            }
+            else
+            {
+                Return (1)      //  fail
+            }
+
+            //  Test If (And ()) with no save of And result
+            If (And (4, 1, ))
+            {
+                Return (2)      //  fail
+            }
+            else
+            {
+                Store (2, WRD)  //  pass -- just do something
+            }
+
+
+            //  Test If (NAnd ()) with no save of NAnd result
+            If (NAnd (3, 1, ))
+            {
+                Store (3, WRD)  //  pass -- just do something
+            }
+            else
+            {
+                Return (3)      //  fail
+            }
+
+            //  Test If (NAnd ()) with no save of NAnd result
+            If (NAnd (0xFFFFFFFF, 0xFFFFFFFF, ))
+            {
+                Return (4)      // fail
+            }
+            else
+            {
+                Store (4, WRD)  //  pass -- just do something
+            }
+
+
+            //  Test If (NOr ()) with no save of NOr result
+            If (NOr (0, 1, ))
+            {
+                Store (5, WRD)  //  pass -- just do something
+            }
+            else
+            {
+                Return (5)      //  fail
+            }
+
+            //  Test If (NOr ()) with no save of NOr result
+            If (NOr (0xFFFFFFFE, 1, ))
+            {
+                Return (6)      // fail
+            }
+            else
+            {
+                Store (6, WRD)  //  pass -- just do something
+            }
+
+
+            //  Test If (Not ()) with no save of Not result
+            If (Not (1, ))
+            {
+                Store (7, WRD)  //  pass -- just do something
+            }
+            else
+            {
+                Return (7)      //  fail
+            }
+
+            //  Test If (Not ()) with no save of Not result
+            If (Not (0xFFFFFFFF, ))
+            {
+                Return (8)      // fail
+            }
+            else
+            {
+                Store (8, WRD)  //  pass -- just do something
+            }
+
+
+            //  Test If (Or ()) with no save of Or result
+            If (Or (3, 1, ))
+            {
+                Store (9, WRD)  //  pass -- just do something
+            }
+            else
+            {
+                Return (9)      //  fail
+            }
+
+            //  Test If (Or ()) with no save of Or result
+            If (Or (0, 0, ))
+            {
+                Return (10)     //  fail
+            }
+            else
+            {
+                Store (10, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (XOr ()) with no save of XOr result
+            If (XOr (3, 1, ))
+            {
+                Store (11, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (11)     // fail
+            }
+
+            //  Test If (XOr ()) with no save of XOr result
+            If (XOr (3, 3, ))
+            {
+                Return (12)     // fail
+            }
+            else
+            {
+                Store (12, WRD) //  pass -- just do something
+            }
+
+
+            //
+            //  Begin test of logical operators with no destination operands
+            //
+
+            //  Test If (LAnd ()) with no save of LAnd result
+            If (LAnd (3, 3))
+            {
+                Store (21, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (21)     // fail
+            }
+
+            //  Test If (LAnd ()) with no save of LAnd result
+            If (LAnd (3, 0))
+            {
+                Return (22)     // fail
+            }
+            else
+            {
+                Store (22, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LAnd ()) with no save of LAnd result
+            If (LAnd (0, 3))
+            {
+                Return (23)     //  fail
+            }
+            else
+            {
+                Store (23, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LAnd ()) with no save of LAnd result
+            If (LAnd (0, 0))
+            {
+                Return (24)     //  fail
+            }
+            else
+            {
+                Store (24, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LEqual ()) with no save of LEqual result
+            If (LEqual (3, 3))
+            {
+                Store (31, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (31)     //  fail
+            }
+
+            //  Test If (LEqual ()) with no save of LEqual result
+            If (LEqual (1, 3))
+            {
+                Return (32)     //  fail
+            }
+            else
+            {
+                Store (32, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LGreater ()) with no save of LGreater result
+            If (LGreater (3, 1))
+            {
+                Store (41, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (41)     //  fail
+            }
+
+            //  Test If (LGreater ()) with no save of LGreater result
+            If (LGreater (4, 4))
+            {
+                Return (42)     //  fail
+            }
+            else
+            {
+                Store (42, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LGreater ()) with no save of LGreater result
+            If (LGreater (1, 4))
+            {
+                Return (43)     //  fail
+            }
+            else
+            {
+                Store (43, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LGreaterEqual ()) with no save of LGreaterEqual result
+            If (LGreaterEqual (3, 1))
+            {
+                Store (44, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (44)     //  fail
+            }
+
+            //  Test If (LGreaterEqual ()) with no save of LGreaterEqual result
+            If (LGreaterEqual (3, 3))
+            {
+                Store (45, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (45)     //  fail
+            }
+
+            //  Test If (LGreaterEqual ()) with no save of LGreaterEqual result
+            If (LGreaterEqual (3, 4))
+            {
+                Return (46)     //  fail
+            }
+            else
+            {
+                Store (46, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LLess ()) with no save of LLess result
+            If (LLess (1, 3))
+            {
+                Store (51, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (51)     //  fail
+            }
+
+            //  Test If (LLess ()) with no save of LLess result
+            If (LLess (2, 2))
+            {
+                Return (52)     //  fail
+            }
+            else
+            {
+                Store (52, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LLess ()) with no save of LLess result
+            If (LLess (4, 2))
+            {
+                Return (53)     //  fail
+            }
+            else
+            {
+                Store (53, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LLessEqual ()) with no save of LLessEqual result
+            If (LLessEqual (1, 3))
+            {
+                Store (54, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (54)     //  fail
+            }
+
+            //  Test If (LLessEqual ()) with no save of LLessEqual result
+            If (LLessEqual (2, 2))
+            {
+                Store (55, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (55)     //  fail
+            }
+
+            //  Test If (LLessEqual ()) with no save of LLessEqual result
+            If (LLessEqual (4, 2))
+            {
+                Return (56)     //  fail
+            }
+            else
+            {
+                Store (56, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LNot ()) with no save of LNot result
+            If (LNot (0))
+            {
+                Store (61, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (61)     //  fail
+            }
+
+            //  Test If (LNot ()) with no save of LNot result
+            If (LNot (1))
+            {
+                Return (62)     //  fail
+            }
+            else
+            {
+                Store (62, WRD) //  pass -- just do something
+            }
+
+
+            //  Test If (LNotEqual ()) with no save of LNotEqual result
+            If (LNotEqual (3, 3))
+            {
+                Return (63)     //  fail
+            }
+            else
+            {
+                Store (63, WRD) //  pass -- just do something
+            }
+
+            //  Test If (LNotEqual ()) with no save of LNotEqual result
+            If (LNotEqual (1, 3))
+            {
+                Store (64, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (64)     //  fail
+            }
+
+
+            //  Test If (LOr ()) with no save of LOr result
+            If (LOr (3, 1))
+            {
+                Store (71, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (71)     //  fail
+            }
+
+            //  Test If (LOr ()) with no save of LOr result
+            If (LOr (0, 1))
+            {
+                Store (72, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (72)     //  fail
+            }
+
+            //  Test If (LOr ()) with no save of LOr result
+            If (LOr (3, 0))
+            {
+                Store (73, WRD) //  pass -- just do something
+            }
+            else
+            {
+                Return (73)     //  fail
+            }
+
+            //  Test If (LOr ()) with no save of LOr result
+            If (LOr (0, 0))
+            {
+                Return (74)     //  fail
+            }
+            else
+            {
+                Store (74, WRD) //  pass -- just do something
+            }
+
+            Return (0)
+        }   //  TEST
+    }   //  NOSV
+
+
+//
+// test IndxFld.asl
+//
+//  IndexFld test
+//      This is just a subset of the many RegionOp/Index Field test cases.
+//      Tests index field element AccessAs macro.
+//
+    Device (IDXF)
+    {   //  Test device name
+
+        OperationRegion (SIO, SystemIO, 0x100, 2)
+        Field (SIO, ByteAcc, NoLock, Preserve)
+        {
+            INDX,   8,
+            DATA,   8
+        }
+        IndexField (INDX, DATA, AnyAcc, NoLock, WriteAsOnes)
+        {
+            AccessAs (ByteAcc, 0),
+            IFE0,   8,
+            IFE1,   8,
+            IFE2,   8,
+            IFE3,   8,
+            IFE4,   8,
+            IFE5,   8,
+            IFE6,   8,
+            IFE7,   8,
+            IFE8,   8,
+            IFE9,   8,
+        }
+
+        Method (TEST)
+        {
+            Store ("++++++++ IndxFld Test", Debug)
+
+            Store (IFE0, Local0)
+            Store (IFE1, Local1)
+            Store (IFE2, Local2)
+
+            Return (0)
+        }   //  TEST
+    }   //  IDXF
+
+//
+// test NestdLor.asl
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+
+        Name (ZER0, 0)
+        Name (ZER1, 0)
+        Name (ZER2, 0)
+        Name (ONE0, 1)
+
+        Device (NSTL)
+        {
+            Method (TEST)
+            {
+                Store ("++++++++ NestdLor Test", Debug)
+
+                If (Lor (ZER0, Lor (ZER1, Lor (ZER2, ONE0))))
+                {   //  Indicate Pass
+                    Store (0x00, Local0)
+                }
+
+                Else
+                {   //  Indicate Fail
+                    Store (0x01, Local0)
+                }
+
+                Return (Local0)
+            }   //  End Method TEST
+        }   //  Device NSTL
+    }   //  _SB system bus
+
+//
+// test RetBuf.asl
+//
+//  Test ReturnOp(Buffer)
+//      This is required to support Control Method Batteries on
+//          Dell Latitude Laptops (e.g., CP1-A)
+//
+    Device (RTBF)
+    {
+        Method (SUBR, 1)
+        {
+            Return (Arg0)
+        }
+
+        Method (RBUF)
+        {   //  RBUF: Return Buffer from local variable
+            Name (ABUF, Buffer() {"ARBITRARY_BUFFER"})
+
+            //  store local buffer ABUF into Local0
+            Store (ABUF, Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Buffer
+            If (LNotEqual (Local1, 3))  //  Buffer type is 3
+            {
+                Return (1)      //  failure
+            }
+
+            //  store value returned by control method SUBR into Local0
+            Store (SUBR (ABUF), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Buffer
+            If (LNotEqual (Local1, 3))  //  Buffer type is 3
+            {
+                Return (2)      //  failure
+            }
+
+            //  allocate buffer using Local1 as buffer size (run-time evaluation)
+            Store (5, Local1)
+            Name (BUFR, Buffer(Local1) {})
+
+            //  store value returned by control method SUBR into Local0
+            Store (SUBR (BUFR), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Buffer
+            If (LNotEqual (Local1, 3))  //  Buffer type is 3
+            {
+                Return (3)      //  failure
+            }
+
+            //  store BUFR Buffer into Local0
+            Store (BUFR, Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Buffer
+            If (LNotEqual (Local1, 3))  //  Buffer type is 3
+            {
+                Return (4)      //  failure
+            }
+
+
+            //  return Local0 Buffer
+            Return (Local0)
+        }   //  RBUF
+
+        Method (TEST)
+        {
+            Store ("++++++++ RetBuf Test", Debug)
+
+            //  store RBUF Buffer return value into Local0
+            Store (RBUF, Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Buffer
+            If (LNotEqual (Local1, 3))  //  Buffer type is 3
+            {
+                Return (10)     //  failure
+            }
+            Else
+            {
+                Return (0)      //  success
+            }
+        }   //  TEST
+    }   //  RTBF
+
+//
+// test RetLVal.asl
+//
+//  Test ReturnOp(Lvalue)
+//      This is required to support _PSR on IBM ThinkPad 560D and
+//      _DCK on Toshiba Tecra 8000.
+//
+
+    Device (GPE2)
+    {
+        Method (_L03)
+        {
+            Store ("Method GPE2._L03 invoked", Debug)
+            Return ()
+        }
+
+        Method (_E05)
+        {
+            Store ("Method GPE2._E05 invoked", Debug)
+            Return ()
+        }
+    }
+
+    Device (PRW2)
+    {
+        Name (_PRW, Package(2) {Package(2){\GPE2, 0x05}, 3})
+    }
+
+
+    Scope (\_GPE)
+    {
+        Name (ACST, 0xFF)
+
+        Method (_L08)
+        {
+            Store ("Method _GPE._L08 invoked", Debug)
+            Return ()
+        }
+
+        Method (_E09)
+        {
+            Store ("Method _GPE._E09 invoked", Debug)
+            Return ()
+        }
+
+        Method (_E11)
+        {
+            Store ("Method _GPE._E11 invoked", Debug)
+            Notify (\PRW1, 2)
+        }
+
+        Method (_L22)
+        {
+            Store ("Method _GPE._L22 invoked", Debug)
+            Return ()
+        }
+
+        Method (_L33)
+        {
+            Store ("Method _GPE._L33 invoked", Debug)
+            Return ()
+        }
+
+        Method (_E64)
+        {
+            Store ("Method _GPE._E64 invoked", Debug)
+            Return ()
+        }
+
+    }   //  _GPE
+
+    Device (PRW1)
+    {
+        Name (_PRW, Package(2) {0x11, 3})
+    }
+
+    Device (PWRB)
+    {
+        Name (_HID, EISAID("PNP0C0C"))
+        Name (_PRW, Package(2) {0x33, 3})
+    }
+
+
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+
+        Device (ACAD)
+        {   //  ACAD:   AC adapter device
+            Name (_HID, "ACPI0003") //  AC adapter device
+
+            Name (_PCL, Package () {\_SB})
+
+            OperationRegion (AREG, SystemIO, 0x0372, 2)
+            Field (AREG, ByteAcc, NoLock, Preserve)
+            {
+                AIDX,   8,
+                ADAT,   8
+            }
+            IndexField (AIDX, ADAT, ByteAcc, NoLock, Preserve)
+            {
+                     ,  1,  //  skips
+                ACIN,   1,
+                     ,  2,  //  skips
+                CHAG,   1,
+                     ,  3,  //  skips
+                     ,  7,  //  skips
+                ABAT,   1,
+            }   //  IndexField
+
+            Method (_PSR)
+            {
+                Store (\_GPE.ACST, Local0)
+                Store (ACIN, Local1)
+                If (LNotEqual (\_GPE.ACST, Local1))
+                {
+                    Store (Local1, \_GPE.ACST)
+                    // This Notify is commented because it causes a
+                    //  method error when running on a system without the
+                    //  specific device.
+                    // Notify (\_SB_.ACAD, 0)
+                }
+                Return (Local0)
+            }   //  _PSR
+
+            Method (_STA)
+            {
+                Return (0x0F)
+            }
+
+            Method (_INI)
+            {
+                Store (ACIN, \_GPE.ACST)
+            }
+        }   //  ACAD:   AC adapter device
+
+        //  test implicit return from control method
+        Method (DIS_, 1)
+        {
+            Store (Arg0, Local0)
+        }
+
+        Device (RTLV)
+        {
+            //  test implicit return inside nested if with explicit return of Lvalue
+            Method (_DCK, 1)
+            //  Arg0:   1 == dock, 0 == undock
+            {
+                If (Arg0)
+                {   //  dock
+                    Store (0x87, Local0)
+
+                    If (Local0)
+                    {
+                        DIS_ (0x23)
+                        Return (1)
+                    }
+
+                    Return (0)
+                }   //  dock
+                Else
+                {   //  undock
+                    Store (Arg0, Local0)
+
+                    If (Local0)
+                    {
+                        DIS_ (0x23)
+                        Return (1)
+                    }
+
+                    Return (0)
+                }   //  undock
+            }   //  _DCK control method
+
+            Method (TEST)
+            {
+                Store ("++++++++ RetLVal Test", Debug)
+
+                //  store _PSR return value into Local0
+                Store (\_SB_.ACAD._PSR, Local0)
+
+                //  save Local0 object type value into Local1
+                Store (ObjectType (Local0), Local1)
+
+                //  validate Local0 is a Number
+                If (LNotEqual (Local1, 1))  //  Number/Integer type is 1
+                {
+                    Return (1)      //  failure
+                }
+
+                //  test implicit return inside nested if with explicit return of Lvalue
+                Store (_DCK (1), Local2)
+
+                //  save Local2 object type value into Local3
+                Store (ObjectType (Local2), Local3)
+
+                //  validate Local2 is a Number
+                If (LNotEqual (Local3, 1))  //  Number/Integer type is 1
+                {
+                    Return (2)      //  failure
+                }
+
+                If (LNotEqual (Local2, 1))
+                {
+                    Return (3)      //  failure
+                }
+
+                Return (0)  //  success
+            }   //  TEST
+        }   //  RTLV
+    }   //  _SB system bus
+
+//
+// test RetPkg.asl
+//
+//  Test ReturnOp(Package)
+//      This is required to support _PRT on Dell Optiplex Workstations (e.g. GX1)
+//
+
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+        Device(LNKA)
+        {
+            Name (_HID, EISAID("PNP0C0F"))  //  PCI interrupt link
+            Name (_UID, 1)
+        }
+        Device(LNKB)
+        {
+            Name (_HID, EISAID("PNP0C0F"))  //  PCI interrupt link
+            Name (_UID, 2)
+        }
+        Device(LNKC)
+        {
+            Name (_HID, EISAID("PNP0C0F"))  //  PCI interrupt link
+            Name (_UID, 3)
+        }
+        Device(LNKD)
+        {
+            Name (_HID, EISAID("PNP0C0F"))  //  PCI interrupt link
+            Name (_UID, 4)
+        }
+
+        Device (PCI1)
+        {   //  PCI1:   Root PCI Bus
+            Name (_HID, "PNP0A03")  //  Need _HID for root device (String format)
+            Name (_ADR,0x00000000)
+            Name (_CRS,0)
+
+            Name (_PRT, Package ()
+            {
+                Package () {0x0004ffff, 0, LNKA, 0},            //  Slot 1, INTA
+                Package () {0x0004ffff, 1, LNKB, 0},            //  Slot 1, INTB
+                Package () {0x0004ffff, 2, LNKC, 0},            //  Slot 1, INTC
+                Package () {0x0004ffff, 3, LNKD, 0},            //  Slot 1, INTD
+                Package () {0x0005ffff, 0, \_SB_.LNKB, 0},  //  Slot 2, INTA
+                Package () {0x0005ffff, 1, \_SB_.LNKC, 0},  //  Slot 2, INTB
+                Package () {0x0005ffff, 2, \_SB_.LNKD, 0},  //  Slot 2, INTC
+                Package () {0x0006ffff, 3, \_SB_.LNKA, 0},  //  Slot 2, INTD
+                Package () {0x0006ffff, 0, LNKC, 0},            //  Slot 3, INTA
+                Package () {0x0006ffff, 1, LNKD, 0},            //  Slot 3, INTB
+                Package () {0x0006ffff, 2, LNKA, 0},            //  Slot 3, INTC
+                Package () {0x0006ffff, 3, LNKB, 0},            //  Slot 3, INTD
+            })
+
+            Device (PX40)
+            {   // Map f0 space, Start PX40
+                Name (_ADR,0x00070000)  //  Address+function.
+            }
+        }   //  PCI0:   Root PCI Bus
+
+        Device (RETP)
+        {
+            Method (RPKG)
+            {   //  RPKG: Return Package from local variable
+
+                //  store _PRT package into Local0
+                Store (\_SB_.PCI1._PRT, Local0)
+
+                //  return Local0 Package
+                Return (Local0)
+            }   //  RPKG
+
+            Method (TEST)
+            {
+                Store ("++++++++ RetPkg Test", Debug)
+
+                //  store RPKG package return value into Local0
+                Store (RPKG, Local0)
+
+                //  save Local0 object type value into Local1
+                Store (ObjectType (Local0), Local1)
+
+                //  validate Local0 is a Package
+                If (LNotEqual (Local1, 4))  //  Package type is 4
+                    {   Return (1)  }   //  failure
+                Else
+                    {   Return (0)  }   //  success
+            }   //  TEST
+        }   //  RETP
+    } // _SB_
+
+//
+// test WhileRet.asl
+//
+//  WhileRet.asl tests a ReturnOp nested in a IfOp nested in a WhileOp.
+//
+    Device (WHLR)
+    {
+        Name (LCNT, 0)
+        Method (WIR)
+        {   //  WIR:    control method that returns inside of IfOp inside of WhileOp
+            While (LLess (LCNT, 4))
+            {
+                    If (LEqual (LCNT, 2))
+                    {
+                        Return (0)
+                    }
+
+                Increment (LCNT)
+            }
+
+            Return (LCNT)
+        }   //  WIR:    control method that returns inside of IfOp inside of WhileOp
+
+        Method (TEST)
+        {
+            Store ("++++++++ WhileRet Test", Debug)
+
+            Store (WIR, Local0)
+
+            Return (Local0)
+        }   //  TEST
+    }   //  WHLR
+
+//
+// test AndOrOp.asl
+//
+//This code tests the bitwise AndOp and OrOp Operator terms
+//
+//Syntax of Andop term
+//And - Bitwise And
+//AndTerm   := And(
+//  Source1,    //TermArg=>Integer
+//  Source2,    //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source1 and Source2 are evaluated as integer data types,
+// a bit-wise AND is performed, and the result is optionally
+//stored into Result.
+//
+//
+//Syntax of OrOp
+//Or - Bit-wise Or
+//OrTerm    := Or(
+//  Source1,    //TermArg=>Integer
+//  Source2 //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source1 and Source2 are evaluated as integer data types,
+// a bit-wide OR is performed, and the result is optionally
+//stored in Result
+//
+    Device (ANDO)
+    {
+        OperationRegion (TMEM, SystemMemory, 0xC4, 0x02)
+        Field (TMEM, ByteAcc, NoLock, Preserve)
+        {
+                ,   3,
+            TOUD,   13
+        }
+
+        //Create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }// Field(RAM)
+
+
+        //And with Byte Data
+        Name (BYT1, 0xff)
+        Name (BYT2, 0xff)
+        Name (BRSL, 0x00)
+
+        //And with Word Data
+        Name (WRD1, 0xffff)
+        Name (WRD2, 0xffff)
+        Name (WRSL, 0x0000)
+
+        //And with DWord Data
+        Name (DWD1, 0xffffffff)
+        Name (DWD2, 0xffffffff)
+        Name (DRSL, 0x00000000)
+
+        Method (ANDP)
+        {
+            //Check with 1 And 1 on byte data
+            And(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0xff))
+            {Return(1)}
+
+            //Check with 1 And 1 on Word data
+            And(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0xffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 And 1 Dword
+            And(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0xffffffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 And 0 on byte data
+            Store(0x00,BYT1)
+            Store(0x00,BYT2)
+            Store(0x00,BRSL)
+            And(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0x00))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 And 0 on Word data
+            Store (0x0000,WRD1)
+            Store (0x0000,WRD2)
+            Store (0x0000,WRSL)
+            And(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0x0000))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 And 0 Dword
+            Store (0x00000000,DWD1)
+            Store (0x00000000,DWD2)
+            Store (0x00000000,DRSL)
+            And(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+            {
+                Return (1)      //  failure
+            }
+
+
+            //Check with 1 And 0 on byte data
+            Store(0x55,BYT1)
+            Store(0xAA,BYT2)
+            Store(0x00,BRSL)
+            And(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0x00))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 And 0 on Word data
+            Store (0x5555,WRD1)
+            Store (0xAAAA,WRD2)
+            Store (0x0000,WRSL)
+            And(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0x0000))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 And 0 on Dword
+            Store (0x55555555,DWD1)
+            Store (0xAAAAAAAA,DWD2)
+            Store (0x00000000,DRSL)
+            And(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+            {
+                Return (1)      //  failure
+            }
+
+            Store (0x1FFF, TOUD)
+            Store (TOUD, Local0)
+            if(LNotEqual(Local0,0x1FFF))
+            {
+                Return (1)      //  failure
+            }
+
+            //TBD- Do We need to check for system memory data also for each test case ??
+
+            Return(0)
+
+        }//ANDP
+
+        Method (OROP)
+        {
+
+            //Check with 1 Ored with 1 on byte data
+            Store(0xff,BYT1)
+            Store(0xff,BYT2)
+            Store(0x00,BRSL)
+            Or(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0xff))
+            {
+                Return (1)      //  failure
+            }
+
+
+            //Check with 1 Ored with 1 on Word data
+            Store(0xffff,WRD1)
+            Store(0xffff,WRD2)
+            Store(0x0000,WRSL)
+            Or(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0xffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 Ored with 1 on Dword data
+            Store(0xffffffff,DWD1)
+            Store(0xffffffff,DWD2)
+            Store(0x00000000,DRSL)
+            Or(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0xffffffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 Ored with 0 on byte data
+            Store(0x00,BYT1)
+            Store(0x00,BYT2)
+            Store(0x00,BRSL)
+            Or(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0x00))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 Ored with 0 on Word data
+            Store (0x0000,WRD1)
+            Store (0x0000,WRD2)
+            Store (0x0000,WRSL)
+            Or(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0x0000))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 0 Ored with  0 Dword data
+            Store (0x00000000,DWD1)
+            Store (0x00000000,DWD2)
+            Store (0x00000000,DRSL)
+            Or(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+            {
+                Return (1)      //  failure
+            }
+
+
+            //Check with 1 Ored with 0 on byte data
+            Store(0x55,BYT1)
+            Store(0xAA,BYT2)
+            Store(0x00,BRSL)
+            Or(BYT1, BYT2, BRSL)
+            if(LNotEqual(BRSL,0xff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 Ored with 0 on Word data
+            Store (0x5555,WRD1)
+            Store (0xAAAA,WRD2)
+            Store (0x0000,WRSL)
+            Or(WRD1, WRD2, WRSL)
+            if(LNotEqual(WRSL,0xffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //Check with 1 Ored with 0 on Dword data
+            Store (0x55555555,DWD1)
+            Store (0xAAAAAAAA,DWD2)
+            Store (0x00000000,DRSL)
+            Or(DWD1, DWD2, DRSL)
+            if(LNotEqual(DRSL,0xffffffff))
+            {
+                Return (1)      //  failure
+            }
+
+            //TBD - Do We need to check for system memory data also for each test case ??
+
+            Return(0)
+
+        }//OROP
+
+        Method(TEST)
+        {
+            Store ("++++++++ AndOrOp Test", Debug)
+
+            Name(RSLT,1)
+            //Call Andop method
+            Store(ANDP,RSLT)
+            if(LEqual(RSLT,1))
+            {
+                Return (RSLT)
+            }
+
+            //Call OrOp Method
+            Store(OROP,RSLT)
+            if(LEqual(RSLT,1))
+            {
+                Return(RSLT)
+            }
+
+            //
+            // Return original conditions to allow iterative execution
+            //
+            Store(0xff,BYT1)
+            Store(0xff,BYT2)
+            Store(0x00,BRSL)
+            Store (0xffff,WRD1)
+            Store (0xffff,WRD2)
+            Store (0x0000,WRSL)
+            Store (0xffffffff,DWD1)
+            Store (0xffffffff,DWD2)
+            Store (0x00000000,DRSL)
+
+            Return(0)
+        }   //TEST
+    }   //ANDO
+
+//
+// test BreakPnt.asl
+//
+// This code tests the BreakPoint opcode term. The syntax of BreakPoint Term is
+// BreakPointTerm    := BreakPoint
+// Used for debugging, the Breakpoint opcode stops the execution and enters the AML debugger.
+// In the non-debug version of the interpreter, BreakPoint is equivalent to Noop.
+//
+    Device (BRKP)
+    {
+        Name(CNT0,0)
+
+        Method (BK1)
+        {
+            BreakPoint
+            Return(0)
+        }
+
+        Method (TEST)
+        {
+            Store ("++++++++ BreakPnt Test", Debug)
+
+            Store(0,CNT0)
+
+            //Check BreakPoint statement
+            While(LLess(CNT0,10))
+            {
+                Increment(CNT0)
+            }
+
+            //Check the BreakPoint statement
+            If(LEqual(CNT0,10))
+            {
+    //            BreakPoint
+                Return(0)
+            }
+
+            //failed
+            Return(1)
+        }
+    }
+
+//
+// test AddSubOp.asl
+//
+    Device (ADSU)
+    {
+        //  create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }   //  Field(RAM)
+
+        Method (TEST)
+        {
+            Store ("++++++++ AddSubOp Test", Debug)
+
+            Name (DWRD, 0x12345678)
+            Name (WRD, 0x1234)
+            Name (BYT, 0x12)
+
+            //  Test AddOp with DWORD data
+            Store (0x12345678, DWRD)
+            Add (DWRD, 7, DWRD)
+            If (LNotEqual (DWRD, 0x1234567F))
+                {   Return (DWRD)   }
+
+            //  Test AddOp with WORD data
+            Add (WRD, 5, WRD)
+            If (LNotEqual (WRD, 0x1239))
+                {   Return (WRD)    }
+
+            //  Test AddOp with BYTE data
+            Add (BYT, 3, BYT)
+            If (LNotEqual (BYT, 0x15))
+                {   Return (BYT)    }
+
+            //  Test SubtractOp with DWORD data
+            Subtract (DWRD, 7, DWRD)
+            If (LNotEqual (DWRD, 0x12345678))
+                {   Return (DWRD)   }
+
+            //  Test SubtractOp with WORD data
+            Subtract (WRD, 3, WRD)
+            If (LNotEqual (WRD, 0x1236))
+                {   Return (WRD)    }
+
+            //  Test SubtractOp with BYTE data
+            Subtract (BYT, 3, BYT)
+            If (LNotEqual (BYT, 0x12))
+                {   Return (BYT)    }
+
+
+            //  test AddOp with DWORD SystemMemory OpRegion
+            Store (0x01234567, SMDW)
+            Add (SMDW, 8, SMDW)
+            If (LNotEqual (SMDW, 0x0123456F))
+                {   Return (SMDW)   }
+
+            //  test SubtractOp with DWORD SystemMemory OpRegion
+            Subtract (SMDW, 7, SMDW)
+            If (LNotEqual (SMDW, 0x01234568))
+                {   Return (SMDW)   }
+
+
+            //  test AddOp with WORD SystemMemory OpRegion
+            Store (0x0123, SMWD)
+            Add (SMWD, 6, SMWD)
+            If (LNotEqual (SMWD, 0x0129))
+                {   Return (SMWD)   }
+
+            //  test SubtractOp with WORD SystemMemory OpRegion
+            Subtract (SMWD, 5, SMWD)
+            If (LNotEqual (SMWD, 0x0124))
+                {   Return (SMWD)   }
+
+
+            //  test AddOp with BYTE SystemMemory OpRegion
+            Store (0x01, SMBY)
+            Add (SMBY, 4, SMBY)
+            If (LNotEqual (SMBY, 0x05))
+                {   Return (SMBY)   }
+
+            //  test SubtractOp with BYTE SystemMemory OpRegion
+            Subtract (SMBY, 3, SMBY)
+            If (LNotEqual (SMBY, 0x02))
+                {   Return (SMBY)   }
+
+            Return (0)
+        }   //  TEST
+    }   //  ADSU
+
+//
+// test IncDecOp.asl
+//
+    Device (INDC)
+    {
+        //  create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }   //  Field(RAM)
+
+        Method (TEST)
+        {
+            Store ("++++++++ IncDecOp Test", Debug)
+
+            Name (DWRD, 0x12345678)
+            Name (WRD, 0x1234)
+            Name (BYT, 0x12)
+
+            //  Test IncrementOp with DWORD data
+            Store (0x12345678, DWRD)
+            Increment (DWRD)
+            If (LNotEqual (DWRD, 0x12345679))
+                {   Return (DWRD)   }
+
+            //  Test IncrementOp with WORD data
+            Increment (WRD)
+            If (LNotEqual (WRD, 0x1235))
+                {   Return (WRD)    }
+
+            //  Test IncrementOp with BYTE data
+            Increment (BYT)
+            If (LNotEqual (BYT, 0x13))
+                {   Return (BYT)    }
+
+            //  Test DecrementOp with DWORD data
+            Decrement (DWRD)
+            If (LNotEqual (DWRD, 0x12345678))
+                {   Return (DWRD)   }
+
+            //  Test DecrementOp with WORD data
+            Decrement (WRD)
+            If (LNotEqual (WRD, 0x1234))
+                {   Return (WRD)    }
+
+            //  Test DecrementOp with BYTE data
+            Decrement (BYT)
+            If (LNotEqual (BYT, 0x12))
+                {   Return (BYT)    }
+
+
+            //  test IncrementOp with DWORD SystemMemory OpRegion
+            Store (0x01234567, SMDW)
+            Increment (SMDW)
+            If (LNotEqual (SMDW, 0x01234568))
+                {   Return (SMDW)   }
+
+            //  test DecrementOp with DWORD SystemMemory OpRegion
+            Decrement (SMDW)
+            If (LNotEqual (SMDW, 0x01234567))
+                {   Return (SMDW)   }
+
+
+            //  test IncrementOp with WORD SystemMemory OpRegion
+            Store (0x0123, SMWD)
+            Increment (SMWD)
+            If (LNotEqual (SMWD, 0x0124))
+                {   Return (SMWD)   }
+
+            //  test DecrementOp with WORD SystemMemory OpRegion
+            Decrement (SMWD)
+            If (LNotEqual (SMWD, 0x0123))
+                {   Return (SMWD)   }
+
+
+            //  test IncrementOp with BYTE SystemMemory OpRegion
+            Store (0x01, SMBY)
+            Increment (SMBY)
+            If (LNotEqual (SMBY, 0x02))
+                {   Return (SMBY)   }
+
+            //  test DecrementOp with BYTE SystemMemory OpRegion
+            Decrement (SMBY)
+            If (LNotEqual (SMBY, 0x01))
+                {   Return (SMBY)   }
+
+            Return (0)
+        }   //  TEST
+    }   //  INDC
+
+//
+// test LOps.asl
+//
+//This source tests all the logical operators. Logical operators in ASL are as follows.
+//LAnd, LEqual, LGreater, LLess, LNot, LNotEqual, LOr.
+// Success will return 0 and failure will return a non zero number. Check the source code for
+// non zero number to find where the test failed
+
+    Device (LOPS)
+    {
+        //Create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }// Field(RAM)
+
+        //And with Byte Data
+        Name (BYT1, 0xff)
+        Name (BYT2, 0xff)
+        Name (BRSL, 0x00)
+
+        //And with Word Data
+        Name (WRD1, 0xffff)
+        Name (WRD2, 0xffff)
+        Name (WRSL, 0x0000)
+
+        //And with DWord Data
+        Name (DWD1, 0xffffffff)
+        Name (DWD2, 0xffffffff)
+        Name (DRSL, 0x00000000)
+
+        Name(RSLT,1)
+
+        Method (ANDL,2) // Test Logical And
+        {
+            //test with the arguments passed
+            if(LEqual(Arg0,Arg1))
+            { Store(LAnd(Arg0,Arg1),RSLT)
+                if(LNotEqual(Ones,RSLT))
+                {Return(11)}
+            }
+
+            //test with he locals
+            Store(Arg0,Local0)
+            Store(Arg1,Local1)
+
+            if(LEqual(Local0,Local1))
+            {
+                Store(LAnd(Local0,Local1),RSLT)
+                if(LNotEqual(Ones,RSLT))
+                    {Return(12)}
+            }
+
+            //test with BYTE data
+            if(LEqual(BYT1,BYT2))
+            { Store(LAnd(BYT1,BYT2),BRSL)
+                if(LNotEqual(Ones,BRSL))
+                {Return(13)}
+            }
+
+            //test with WORD data
+            if(LEqual(WRD1,WRD2))
+            { Store(LAnd(WRD1,WRD2),WRSL)
+                if(LNotEqual(Ones,WRSL))
+                {Return(14)}
+            }
+
+            //test with DWORD data
+            if(LEqual(DWD1,DWD2))
+            { Store(LAnd(DWD1,DWD2),DRSL)
+                if(LNotEqual(Ones,DRSL))
+                {Return(15)}
+            }
+
+            //Test for system memory data for each test case.
+
+                Store(0xff,BYT1)
+                Store(0xff,SMBY)
+                Store(0x00,BRSL)
+
+            //test with BYTE system memory data
+            if(LEqual(BYT1,SMBY))
+            { Store(LAnd(BYT1,SMBY),BRSL)
+                if(LNotEqual(Ones,BRSL))
+                {Return(16)}
+            }
+
+            Store (0xffff,WRD1)
+            Store(0xffff,SMWD)
+            Store(0x0000,WRSL)
+            //test with WORD system memory data
+            if(LEqual(WRD1,SMWD))
+            { Store(LAnd(WRD1,SMWD),WRSL)
+                if(LNotEqual(Ones,WRSL))
+                {Return(17)}
+            }
+
+            Store(0x000000,DRSL)
+            Store (0xffffff,DWD1)
+            Store(0xffffff,SMDW)
+
+            //test with DWORD system memory data
+            if(LEqual(DWD1,SMDW))
+            { Store(LAnd(DWD1,SMDW),DRSL)
+                if(LNotEqual(Ones,DRSL))
+                {Return(18)}
+            }
+
+            Return(0)
+
+        }//ANDL
+
+        //Test the LOr Operator
+
+        Method (ORL_,2)
+        {//ORL_
+
+            //test with the arguments passed
+            if(LEqual(Arg0,Arg1))
+            {
+                Store(LOr(Arg0,Arg1),RSLT)
+                if(LNotEqual(Ones,RSLT))
+                {
+                    Return(21)
+                }
+            }
+
+            //test with he locals
+            Store(Arg0,Local0)
+            Store(Arg1,Local1)
+
+            if(LEqual(Local0,Local1))
+            {
+                Store(LOr(Local0,Local1),RSLT)
+                if(LNotEqual(Ones,RSLT))
+                    {Return(22)}
+            }
+
+            //Check with 1 LOred with 0 on byte data
+            Store(0xff,BYT1)
+            Store(0x00,BYT2)
+            Store(0x00,BRSL)
+
+            if(LNotEqual(BYT1, BYT2))
+            {
+                Store(LOr(BYT1, BYT2), BRSL)
+                if(LNotEqual(Ones,BRSL))
+                {Return(23)}
+            }
+
+            //Check with 1 LOred with 0 on WORD data
+            Store(0xffff,WRD1)
+            Store(0x0000,WRD2)
+            Store(0x0000,WRSL)
+
+            if(LNotEqual(WRD1, WRD2))
+            {
+                Store(LOr(WRD1, WRD2), WRSL)
+                if(LNotEqual(Ones,WRSL))
+                {Return(24)}
+            }
+
+            //Check with 1 LOred with 0 on DWORD data
+            Store(0xffffffff,DWD1)
+            Store(0x00000000,DWD2)
+            Store(0x00000000,DRSL)
+
+            if(LNotEqual(DWD1, DWD2))
+            {
+                Store(LOr(DWD1, DWD2), DRSL)
+                if(LNotEqual(Ones,DRSL))
+                {Return(25)}
+            }
+
+            Store(0x00,BYT1)
+            Store(0xff,SMBY)
+            Store(0x00,BRSL)
+
+            //test with BYTE system memory data
+            if(LEqual(BYT1,SMBY))
+            { Store(LOr(BYT1,SMBY),BRSL)
+                if(LNotEqual(Ones,BRSL))
+                {Return(26)}
+            }
+
+            Store (0x0000,WRD1)
+            Store(0xffff,SMWD)
+            Store(0x0000,WRSL)
+
+            //test with WORD system memory data
+            if(LEqual(WRD1,SMWD))
+            { Store(LOr(WRD1,SMWD),WRSL)
+                if(LNotEqual(Ones,WRSL))
+                {Return(27)}
+            }
+
+
+            Store(0x00000000,DWD1)
+            Store(0xffffffff,SMDW)
+            Store(0x00000000,DRSL)
+
+            //test with DWORD system memory data
+            if(LEqual(DWD1,SMDW))
+            { Store(LAnd(DWD1,SMDW),DRSL)
+                if(LNotEqual(Ones,DRSL))
+                {Return(28)}
+            }
+            Return(0)
+
+        }//ORL_
+
+        //This method tests LGreater and LNot operator
+        Method(LSGR,2)
+        {//LSGR
+
+            //Test on arguements passed
+
+            //in test data, Arg1 > Arg0
+            if(LEqual(Ones,LNot(LGreater(Arg1,Arg0))))
+            {Return(31)}
+
+            //test LLessEqual
+            if(LEqual(Ones,LNot(LGreaterEqual(Arg1,Arg0))))
+            {Return(32)}
+
+            if(LEqual(Ones,LLess(Arg1,Arg0)))
+            {Return(33)}
+
+            //test LLessEqual
+            if(LEqual(Ones,LLessEqual(Arg1,Arg0)))
+            {Return(34)}
+
+            Store(Arg0,Local0)
+            Store(Arg1,Local1)
+
+            //test with the locals
+            if(LNot(LGreater(Local1,Local0)))
+                {Return(35)}
+
+            //test on Byte data
+            Store(0x12,BYT1)
+            Store(0x21,BYT2)
+
+            if(LNot(LGreater(BYT2,BYT1)))
+                {Return(36)}
+
+            if(LNot(LLess(BYT1,BYT2)))
+                {Return(37)}
+
+            //test LGreaterEqual with byte data
+            if(LNot(LGreaterEqual(BYT2,BYT1)))
+                {Return(38)}
+
+            //test LLessEqual byte data
+            if(LNot(LLessEqual(BYT1,BYT2)))
+                {Return(39)}
+
+
+            //test on Word data
+            Store(0x1212,WRD1)
+            Store(0x2121,WRD2)
+
+            if(LNot(LGreater(WRD2,WRD1)))
+                {Return(310)}
+
+            if(LNot(LLess(WRD1,WRD2)))
+                {Return(311)}
+
+            //Test LGreaterEqual with Word Data
+            if(LNot(LGreaterEqual(WRD2,WRD1)))
+                {Return(312)}
+
+
+            //Test LLessEqual with Word Data
+            if(LNot(LLessEqual(WRD1,WRD2)))
+                {Return(313)}
+
+            //test on DWord data
+            Store(0x12121212,DWD1)
+            Store(0x21212121,DWD2)
+
+            if(LNot(LGreater(DWD2,DWD1)))
+                {Return(314)}
+
+            if(LNot(LLess(DWD1,DWD2)))
+                {Return(315)}
+
+
+            //Test LGreaterEqual with Dword
+            if(LNot(LGreaterEqual(DWD2,DWD1)))
+                {Return(316)}
+
+            //Test LLessEqual DWord
+            if(LNot(LLessEqual(DWD1,DWD2)))
+                {Return(317)}
+
+            Return(0)
+        }//LSGR
+
+        //The test method
+        Method(TEST)
+        {
+            Store ("++++++++ LOps Test", Debug)
+
+            Store(0,RSLT)
+            //Call LAndOp method
+            Store(ANDL(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+             {Return(RSLT)}
+
+            //Call LOrOp Method
+            Store(ORL_(5,5),RSLT)
+            if(LNotEqual(RSLT,0))
+            {Return(RSLT)}
+
+            //Call LSGR Method
+            Store(LSGR(5,7),RSLT)
+            if(LNotEqual(RSLT,0))
+            {Return(RSLT)}
+
+            Return(0)
+        }//TEST
+    }//LOPS
+
+//
+// test FdSetOps.asl
+//
+//  FindSetLeftBit - Find Set Left Bit
+//  FindSetLeftBitTerm  := FindSetLeftBit
+//  (   Source, //TermArg=>Integer
+//      Result  //Nothing | SuperName
+//  ) => Integer
+//  Source is evaluated as integer data type, and the one-based bit location of
+//  the first MSb (most significant set bit) is optionally stored into Result.
+//  The result of 0 means no bit was set, 1 means the left-most bit set is the
+//  first bit, 2 means the left-most bit set is the second bit, and so on.
+//  FindSetRightBit - Find Set Right Bit
+
+//  FindSetRightBitTerm := FindSetRightBit
+//  (   Source, //TermArg=>Integer
+//      Result  //Nothing | SuperName
+//  ) => Integer
+//  Source is evaluated as integer data type, and the one-based bit location of
+//  the most LSb (least significant set bit) is optionally stored in Result.
+//  The result of 0 means no bit was set, 32 means the first bit set is the
+//  32nd bit, 31 means the first bit set is the 31st bit, and so on.
+
+//  If the Control method is success Zero is returned. Otherwise a non-zero
+//  number is returned.
+//
+    Device (FDSO)
+    {   //  FDSO
+
+        //  Create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,      //  8-bit BYTE
+        }   //  Field(RAM)
+
+        //  Byte Data
+        Name (BYT1, 1)
+        Name (BRSL, 0x00)
+
+        //  Word Data
+        Name (WRD1, 0x100)
+        Name (WRSL, 0x0000)
+
+        //  DWord Data
+        Name (DWD1, 0x10000)
+        Name (DRSL, 0x00000000)
+        Name (RSLT, 1)
+        Name (CNTR, 1)
+
+        Method (SHFT,2)
+        //  Arg0 is the actual data and Arg1 is the bit position
+        {   //  SHFT
+            Store (Arg0, Local0)
+            Store (Arg1, Local1)
+
+            FindSetLeftBit (Arg0, BRSL)
+            If (LNotEqual (BRSL, Arg1))
+                {   Return (0x11)   }
+            If (LNotEqual (Arg0, Local0))
+                {   Return (0x12)   }
+
+            FindSetLeftBit (Local0, BRSL)
+            If (LNotEqual (BRSL, Local1))
+                {   Return (0x13)   }
+            If (LNotEqual (Arg0, Local0))
+                {   Return (0x14)   }
+
+            //  test the byte value for SetLeftBit
+            Store (7, BYT1)
+            FindSetLeftBit (BYT1, BRSL)
+            If (LNotEqual (BRSL, 3))
+                {   Return (0x15)   }
+            If (LNotEqual (BYT1, 7))
+                {   Return (0x16)   }
+
+            Store (1, BYT1)
+            Store (1, CNTR)
+            While (LLessEqual (CNTR, 8))
+            {   //  FindSetLeftBit check loop for byte data
+                FindSetLeftBit (BYT1, BRSL)
+                If (LNotEqual (BRSL, CNTR))
+                    {   Return (0x17)   }
+
+                //  Shift the bits to check the same
+                ShiftLeft (BYT1, 1, BYT1)
+                Increment (CNTR)
+            }   //  FindSetLeftBit check loop for byte data
+
+
+            //  Check BYTE value for SetRightBit
+            Store (7, BYT1)
+            FindSetRightBit (BYT1, BRSL)
+            If (LNotEqual (BRSL, 1))
+                {   Return (0x21)   }
+            If (LNotEqual (BYT1, 7))
+                {   Return (0x22)   }
+
+            Store (1, CNTR)
+            Store (0xFF, BYT1)
+            While (LLessEqual (CNTR, 8))
+            {   //  FindSetRightBit check loop for byte data
+                FindSetRightBit (BYT1, BRSL)
+                If (LNotEqual (BRSL, CNTR))
+                    {   Return (0x23)   }
+
+                ShiftLeft (BYT1, 1, BYT1)
+                Increment (CNTR)
+            }   //  FindSetRightBit check loop for byte data
+
+
+            //  Test Word value for SetLeftBit
+            Store (9, CNTR)
+            Store (0x100, WRD1)
+            While (LLessEqual (CNTR, 16))
+            {
+                //  FindSetLeftBit check loop for Word data
+                FindSetLeftBit (WRD1, WRSL)
+                If (LNotEqual (WRSL, CNTR))
+                    {   Return (0x31)   }
+
+                //  Shift the bits to check the same
+                ShiftLeft (WRD1, 1, WRD1)
+                Increment (CNTR)
+            }   //  FindSetLeftBit check loop for Word data
+
+            //  Check Word value for SetRightBit
+            Store (9, CNTR)
+            Store (0xFF00, WRD1)
+            While (LLessEqual (CNTR, 16))
+            {
+                //  FindSetRightBit check loop for Word data
+                FindSetRightBit (WRD1, WRSL)
+                If (LNotEqual (WRSL, CNTR))
+                    {   Return (0x32)   }
+
+                ShiftLeft (WRD1, 1, WRD1)
+                Increment (CNTR)
+            }   //  FindSetRightBit check loop for Word data
+
+            //  Test the DWord value for SetLeftBit
+            Store (17, CNTR)
+            Store (0x10000, DWD1)
+            While (LLessEqual (CNTR, 32))
+            {
+                //  FindSetLeftBit check loop for Dword
+                FindSetLeftBit (DWD1, DRSL)
+                If (LNotEqual (DRSL, CNTR))
+                    {   Return (0x41)   }
+
+                //  Shift the bits to check the same
+                ShiftLeft (DWD1, 1, DWD1)
+                Increment (CNTR)
+            }   //  FindSetLeftBit check loop for Dword
+
+            //  Check DWord value for SetRightBit
+            Store (17, CNTR)
+            Store (0xFFFF0000, DWD1)
+            While (LLessEqual (CNTR, 32))
+            {   //  FindSetRightBit Check loop for DWORD
+                FindSetRightBit (DWD1, DRSL)
+                If (LNotEqual (DRSL, CNTR))
+                    {   Return (0x42)   }
+
+                ShiftLeft (DWD1, 1, DWD1)
+                Increment (CNTR)
+            }   //  FindSetRightBit Check loop for DWORD
+
+            Return (0)
+        }   //  SHFT
+
+        //  Test method called from amlexec
+        Method (TEST)
+        {   //  TEST
+
+            Store ("++++++++ FdSetOps Test", Debug)
+
+            Store (SHFT (0x80, 8), RSLT)
+            If (LNotEqual (RSLT, 0))
+                {   Return (RSLT)   }
+
+            Return (0)  //  pass
+        }   //  TEST
+    }   //  Device FDSO
+
+//
+// test MulDivOp.asl
+//
+    Device (MLDV)
+    {
+        //  create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }   //  Field(RAM)
+
+        Method (TEST)
+        {
+            Store ("++++++++ MulDivOp Test", Debug)
+
+            Name (RMDR, 0)
+            Name (DWRD, 0x12345678)
+            Name (WRD, 0x1234)
+            Name (BYT, 0x12)
+
+            //  Test MultiplyOp with DWORD data
+            Store (0x12345678, DWRD)
+            Multiply (DWRD, 3, DWRD)
+            If (LNotEqual (DWRD, 0x369D0368))
+                {   Return (DWRD)   }
+
+            //  Test MultiplyOp with WORD data
+            Multiply (WRD, 4, WRD)
+            If (LNotEqual (WRD, 0x48D0))
+                {   Return (WRD)    }
+
+            //  Test MultiplyOp with BYTE data
+            Multiply (BYT, 5, BYT)
+            If (LNotEqual (BYT, 0x5A))
+                {   Return (BYT)    }
+
+            //  Test DivideOp with DWORD data
+            Divide (DWRD, 3, DWRD, RMDR)
+            If (LNotEqual (DWRD, 0x12345678))
+                {   Return (DWRD)   }
+            If (LNotEqual (RMDR, 0))
+                {   Return (RMDR)   }
+
+            //  Test DivideOp with WORD data
+            Divide (WRD, 4, WRD, RMDR)
+            If (LNotEqual (WRD, 0x1234))
+                {   Return (WRD)    }
+            If (LNotEqual (RMDR, 0))
+                {   Return (RMDR)   }
+
+            //  Test DivideOp with BYTE data
+            Divide (BYT, 5, BYT, RMDR)
+            If (LNotEqual (BYT, 0x12))
+                {   Return (BYT)    }
+            If (LNotEqual (RMDR, 0))
+                {   Return (RMDR)   }
+
+
+            //  test MultiplyOp with DWORD SystemMemory OpRegion
+            Store (0x01234567, SMDW)
+            Multiply (SMDW, 2, SMDW)
+            If (LNotEqual (SMDW, 0x02468ACE))
+                {   Return (SMDW)   }
+
+            //  test DivideOp with DWORD SystemMemory OpRegion
+            Divide (SMDW, 3, SMDW, RMDR)
+            If (LNotEqual (SMDW, 0x00C22E44))
+                {   Return (SMDW)   }
+            If (LNotEqual (RMDR, 2))
+                {   Return (RMDR)   }
+
+
+            //  test MultiplyOp with WORD SystemMemory OpRegion
+            Store (0x0123, SMWD)
+            Multiply (SMWD, 3, SMWD)
+            If (LNotEqual (SMWD, 0x369))
+                {   Return (SMWD)   }
+
+            //  test DivideOp with WORD SystemMemory OpRegion
+            Divide (SMWD, 2, SMWD, RMDR)
+            If (LNotEqual (SMWD, 0x01B4))
+                {   Return (SMWD)   }
+            If (LNotEqual (RMDR, 1))
+                {   Return (RMDR)   }
+
+
+            //  test MultiplyOp with BYTE SystemMemory OpRegion
+            Store (0x01, SMBY)
+            Multiply (SMBY, 7, SMBY)
+            If (LNotEqual (SMBY, 0x07))
+                {   Return (SMBY)   }
+
+            //  test DivideOp with BYTE SystemMemory OpRegion
+            Divide (SMBY, 4, SMBY, RMDR)
+            If (LNotEqual (SMBY, 0x01))
+                {   Return (SMBY)   }
+            If (LNotEqual (RMDR, 3))
+                {   Return (RMDR)   }
+
+            Return (0)
+        }   //  TEST
+    }   //  MLDV
+
+//
+// test NBitOps.asl
+//
+//NAnd - Bit-wise NAnd
+//NAndTerm  := NAnd(
+//  Source1,    //TermArg=>Integer
+//  Source2 //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source1 and Source2 are evaluated as integer data types, a bit-wise NAND is performed, and the result is optionally
+//stored in Result.
+
+//NOr - Bitwise NOr
+//NOrTerm   := NOr(
+//  Source1,    //TermArg=>Integer
+//  Source2 //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source1 and Source2 are evaluated as integer data types, a bit-wise NOR is performed, and the result is optionally
+//stored in Result.
+// Not - Not
+//NotTerm   := Not(
+//  Source, //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source1 is evaluated as an integer data type, a bit-wise NOT is performed, and the result is optionally stored in
+//Result.
+
+//If the Control method is success Zero is returned else a non-zero number is returned
+
+    Device (NBIT)
+    {//NBIT
+
+        //Create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }// Field(RAM)
+
+
+        //And with Byte Data
+        Name (BYT1, 0xff)
+        Name (BYT2, 0xff)
+        Name (BRSL, 0x00)
+
+        //And with Word Data
+        Name (WRD1, 0xffff)
+        Name (WRD2, 0xffff)
+        Name (WRSL, 0x0000)
+
+        //And with DWord Data
+        Name (DWD1, 0xffffffff)
+        Name (DWD2, 0xffffffff)
+        Name (DRSL, 0x00000000)
+        Name(RSLT,1)
+
+
+        Name(ARSL,0x00)
+        Name(LRSL,0x00)
+
+        Method(NNDB,2)
+        {//NNDB
+
+            Store(0xffffffff,SMDW)
+            Store(0xffff,SMWD)
+            Store(0xff,SMBY)
+
+
+            NAnd(Arg0,Arg1,ARSL)
+            if(LNotEqual(ARSL,0xfffffffd))
+             {Return(11)}
+
+             Store(Arg0,local0)
+             Store(Arg1,Local1)
+
+             NAnd(Local0,Local1,LRSL)
+                if(LNotEqual(LRSL,0xfffffffd))
+             {Return(12)}
+
+
+            //Byte data
+            NAnd(BYT1,BYT2,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(13)}
+
+            //Word Data
+             NAnd(WRD1,WRD2,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(14)}
+
+             //DWord Data
+             NAnd(DWD1,DWD2,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(15)}
+
+             //Byte data
+            NAnd(SMBY,0xff,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(16)}
+
+            //Word Data
+             NAnd(SMWD,0xffff,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(17)}
+
+             //DWord Data
+             NAnd(SMDW,0xffffffff,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(18)}
+
+            Return(0)
+
+        }//NNDB
+
+        Method(NNOR,2)
+        {//NNOR
+
+            NOr(Arg0,Arg1,ARSL)
+            if(LNotEqual(ARSL,0xfffffffd))
+             {Return(21)}
+
+            Store(Arg0,local0)
+            Store(Arg1,Local1)
+
+            NOr(Local0,Local1,LRSL)
+            if(LNotEqual(LRSL,0xfffffffd))
+             {Return(22)}
+
+
+            //Byte data
+            NOr(BYT1,BYT2,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(23)}
+
+            //Word Data
+            NOr(WRD1,WRD2,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(24)}
+
+            //DWord Data
+            NOr(DWD1,DWD2,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(25)}
+
+             //System Memory Byte data
+            NOr(SMBY,0xff,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(26)}
+
+            //System Memory Word Data
+            NOr(SMWD,0xffff,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(27)}
+
+            //System Memory DWord Data
+            NOr(SMDW,0xffffffff,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(28)}
+
+            Return(0)
+
+        }//NNOR
+
+        Method(NNOT,2)
+        {//NNOT
+
+            Or(Arg0,Arg1,ARSL)
+            Not(ARSL,ARSL)
+            if(LNotEqual(ARSL,0xfffffffd))
+             {Return(31)}
+
+            Store(Arg0,local0)
+            Store(Arg1,Local1)
+
+            Or(Local0,Local1,LRSL)
+            Not(LRSL,LRSL)
+            if(LNotEqual(LRSL,0xfffffffd))
+             {Return(32)}
+
+
+            //Byte data
+            Or(BYT1,BYT2,BRSL)
+            Not(BRSL,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(33)}
+
+            //Word Data
+            Or(WRD1,WRD2,WRSL)
+            Not(WRSL,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(34)}
+
+            //DWord Data
+            Or(DWD1,DWD2,DRSL)
+            Not(DRSL,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(35)}
+
+             //System Memory Byte data
+            Or(SMBY,0xff,BRSL)
+            Not(BRSL,BRSL)
+            if(LNotEqual(BRSL,0xffffff00))
+             {Return(36)}
+
+            //System Memory Word Data
+            Or(SMWD,0xffff,WRSL)
+            Not(WRSL,WRSL)
+            if(LNotEqual(WRSL,0xffff0000))
+             {Return(37)}
+
+            //System Memory DWord Data
+            Or(SMDW,0xffffffff,DRSL)
+            Not(DRSL,DRSL)
+            if(LNotEqual(DRSL,0x00000000))
+             {Return(38)}
+
+            Return(0)
+        }//NNOT
+
+
+        Method(TEST)
+        {
+
+            Store ("++++++++ NBitOps Test", Debug)
+
+            Store(NNDB(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+                {Return(RSLT)}
+
+            Store(NNOR(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+                {Return(RSLT)}
+
+            Store(NNOT(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+                {Return(RSLT)}
+
+
+           Return(0)
+        }
+
+    }//Device NBIT
+
+//
+// test ShftOp.asl
+//
+//ShiftRightTerm    := ShiftRight(
+//  Source, //TermArg=>Integer
+//  ShiftCount  //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source and ShiftCount are evaluated as integer data types. Source is shifted right with the most significant bit
+//zeroed ShiftCount times.  The result is optionally stored into Result.
+
+//ShiftLeft(
+//  Source, //TermArg=>Integer
+//  ShiftCount  //TermArg=>Integer
+//  Result  //Nothing | SuperName
+//) => Integer
+//Source and ShiftCount are evaluated as integer data types. Source is shifted left with the least significant
+//bit zeroed ShiftCount times. The result is optionally stored into Result.
+
+//If the Control method is success Zero is returned else a non-zero number is returned
+    Device (SHFT)
+    {//SHFT
+
+        //Create System Memory Operation Region and field overlays
+        OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {
+            SMDW,   32, //  32-bit DWORD
+            SMWD,   16, //  16-bit WORD
+            SMBY,   8,  //  8-bit BYTE
+        }// Field(RAM)
+
+
+        Name(SHFC,0x00)
+
+        //And with Byte Data
+        Name (BYT1, 0xff)
+        Name (BRSL, 0x00)
+
+        //And with Word Data
+        Name (WRD1, 0xffff)
+        Name (WRSL, 0x0000)
+
+        //And with DWord Data
+        Name (DWD1, 0xffffffff)
+        Name (DRSL, 0x00000000)
+
+        Name(RSLT,1)
+
+        Name(ARSL,0x00)
+        Name(LRSL,0x00)
+
+        Method(SLFT,2)
+        {//SLFT
+
+            Store(0xffffffff,SMDW)
+            Store(0xffff,SMWD)
+            Store(0xff,SMBY)
+
+
+            //Arg0-> 2 & Arg1->2
+            ShiftLeft(Arg0,Arg1,ARSL)
+            if(LNotEqual(ARSL,8))
+            {Return(11)}
+
+             Store(Arg0,local0)
+             Store(Arg1,Local1)
+
+             //Local0->8 and Local1->2
+             ShiftLeft(Local0,Local1,LRSL)
+                if(LNotEqual(LRSL,8))
+             {Return(12)}
+
+            Store(2,SHFC)
+            //Byte data
+            ShiftLeft(BYT1,SHFC,BRSL)
+            if(LNotEqual(BRSL,0x3FC))
+             {Return(13)}
+
+            Store(4,SHFC)
+            //Word Data
+             ShiftLeft(WRD1,SHFC,WRSL)
+            if(LNotEqual(WRSL,0xFFFF0))
+             {Return(14)}
+
+            Store(8,SHFC)
+            //DWord Data
+            ShiftLeft(DWD1,SHFC,DRSL)
+            if(LNotEqual(DRSL,0xFFFFFF00))
+             {Return(15)}
+
+
+             //System Memory Byte data
+            Store(4,SHFC)
+            ShiftLeft(SMBY,SHFC,BRSL)
+            if(LNotEqual(BRSL,0xFF0))
+            {Return(16)}
+
+            //Word Data
+            Store(4,SHFC)
+            ShiftLeft(SMWD,SHFC,WRSL)
+            if(LNotEqual(WRSL,0xffff0))
+             {Return(17)}
+
+            //DWord Data
+            Store(8,SHFC)
+            ShiftLeft(SMDW,SHFC,DRSL)
+            if(LNotEqual(DRSL,0xFFFFFF00))
+                {Return(18)}
+
+            Return(0)
+
+        }//SLFT
+
+        Method(SRGT,2)
+        {//SRGT
+            //And with Byte Data
+            Store (0xff,BYT1)
+            Store (0x00,BRSL)
+
+            //And with Word Data
+            Store (0xffff,WRD1)
+            Store (0x0000,WRSL)
+
+            //And with DWord Data
+            Store(0xffffffff,DWD1)
+            Store (0x00000000,DRSL)
+
+            //Reinitialize the result objects
+            Store(0x00,ARSL)
+            Store(0x00,LRSL)
+
+            Store(0xffffffff,SMDW)
+            Store(0xffff,SMWD)
+            Store(0xff,SMBY)
+
+            //Arg0-> 2 & Arg1->2
+            ShiftRight(Arg0,Arg1,ARSL)
+            if(LNotEqual(ARSL,0))
+            {Return(21)}
+
+             Store(Arg0,local0)
+             Store(Arg1,Local1)
+
+             //Local0->8 and Local1->2
+             ShiftRight(Local0,Local1,LRSL)
+                if(LNotEqual(LRSL,0))
+             {Return(22)}
+
+            Store(2,SHFC)
+            //Byte data
+            ShiftRight(BYT1,SHFC,BRSL)
+            if(LNotEqual(BRSL,0x3F))
+             {Return(23)}
+
+            Store(4,SHFC)
+            //Word Data
+             ShiftRight(WRD1,SHFC,WRSL)
+            if(LNotEqual(WRSL,0xFFF))
+             {Return(24)}
+
+            Store(8,SHFC)
+            //DWord Data
+            ShiftRight(DWD1,SHFC,DRSL)
+            if(LNotEqual(DRSL,0xFFFFFF))
+             {Return(25)}
+
+            //System Memory Byte data
+            Store(4,SHFC)
+            ShiftRight(SMBY,SHFC,BRSL)
+            if(LNotEqual(BRSL,0xF))
+            {Return(26)}
+
+            //Word Data
+            Store(4,SHFC)
+            ShiftRight(SMWD,SHFC,WRSL)
+            if(LNotEqual(WRSL,0xFFF))
+             {Return(27)}
+
+            //DWord Data
+            Store(8,SHFC)
+            ShiftRight(SMDW,SHFC,DRSL)
+            if(LNotEqual(DRSL,0xFFFFFF))
+                {Return(28)}
+
+            Return(0)
+        }//SRGT
+
+        //Test method called from amlexec
+        Method(TEST)
+        {
+            Store ("++++++++ ShftOp Test", Debug)
+
+            Store(SLFT(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+                {Return(RSLT)}
+            Store(SRGT(2,2),RSLT)
+            if(LNotEqual(RSLT,0))
+                {Return(RSLT)}
+           Return(0)
+        }
+
+    }//Device SHFT
+
+//
+// test Xor.asl and slightly modified
+//
+//This code tests the XOR opcode term
+//Syntax of XOR term
+//          XOr(
+//                  Source1  //TermArg=>BufferTerm
+//                  Source2  //TermArg=>Integer
+//                  Result //NameString
+//              )
+//"Source1" and "Source2" are evaluated as integers, a bit-wise XOR is performed, and the result is optionally stored in
+// Result
+    Device (XORD)
+    {
+        //This Method tests XOr operator for all the data types i.e. BYTE, WORD and DWORD
+        Method (TEST)
+        {
+            Store ("++++++++ Xor Test", Debug)
+
+            //Overlay in system memory
+            OperationRegion (RAM, SystemMemory, 0x800000, 256)
+            Field (RAM, ByteAcc, NoLock, Preserve)
+            {
+                RES1,   1,  //Offset
+                BYT1,   8,  //First BYTE
+                BYT2,   8,  //Second BYTE
+                RBYT,   8,  //Result Byte
+                RES2,   1,  //Offset
+                WRD1,   16, //First WORD field
+                WRD2,   16, //Second WORD field
+                RWRD,   16, //RSLT WORD field
+                RES3,   1,  //Offset
+                DWD1,   32, //First Dword
+                DWD2,   32, //Second Dword
+                RDWD,   32, //Result Dword
+                RES4,   1,  //Offset
+            }
+
+            // Store bits in the single bit fields for checking
+            //  at the end
+            Store(1, RES1)
+            Store(1, RES2)
+            Store(1, RES3)
+            Store(1, RES4)
+
+            // Check the stored single bits
+            if(LNotEqual(RES1, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES2, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES3, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES4, 1))
+            {
+                Return(1)
+            }
+
+            //************************************************
+            // (BYT1) Bit1 ->0 and (BYT2)Bit2 -> 0 condition
+            Store(0x00,BYT1)
+            Store(0x00,BYT2)
+            XOr(BYT1,BYT2,Local0)
+            Store (Local0, RBYT)
+            if(LNotEqual(RBYT,0))
+            {   Return(1)}
+
+            // (BYT1) Bit1 ->1 and (BYT2)Bit2 -> 1 condition
+            Store(0xff,BYT1)
+            Store(0xff,BYT2)
+            XOr(BYT1,BYT2,Local0)
+            Store (Local0, RBYT)
+            if(LNotEqual(RBYT,0))
+            {   Return(1)}
+
+            // (BYT1) Bit1 ->1 and (BYT)Bit2 -> 0 condition
+            Store(0x55,BYT1)
+            Store(0xAA,BYT2)
+            XOr(BYT1,BYT2,Local0)
+            Store (Local0, RBYT)
+            if(LNotEqual(RBYT,0xFF))
+            {   Return(1)}
+
+            //(BYT1) Bit1 ->0 and (BYT2)Bit2 -> 1 condition
+            Store(0xAA,BYT1)
+            Store(0x55,BYT2)
+            XOr(BYT1,BYT2,Local0)
+            Store (Local0, RBYT)
+            if(LNotEqual(RBYT,0xFF))
+            {   Return(1)}
+
+            Store(0x12,BYT1)
+            Store(0xED,BYT2)
+
+            XOr(BYT1,BYT2,Local0)
+            Store (Local0, RBYT)
+            if(LNotEqual(RBYT,0xFF))
+            {
+                Return(1)
+            }
+
+            // Store known values for checking later
+            Store(0x12, BYT1)
+            if(LNotEqual(BYT1, 0x12))
+            {
+                Return(1)
+            }
+
+            Store(0xFE, BYT2)
+            if(LNotEqual(BYT2, 0xFE))
+            {
+                Return(1)
+            }
+
+            Store(0xAB, RBYT)
+            if(LNotEqual(RBYT, 0xAB))
+            {
+                Return(1)
+            }
+
+            //***********************************************
+            // (WRD1) Bit1 ->0 and (WRD2)Bit2 -> 0 condition
+            Store(0x0000,WRD1)
+            Store(0x0000,WRD2)
+            XOr(WRD1,WRD2,RWRD)
+            if(LNotEqual(RWRD,0))
+            {   Return(1)}
+
+            // (WRD1) Bit1 ->1 and (WRD2)Bit2 -> 1 condition
+            Store(0xffff,WRD1)
+            Store(0xffff,WRD2)
+            XOr(WRD1,WRD2,RWRD)
+            if(LNotEqual(RWRD,0))
+            {   Return(1)}
+
+            // (WRD1) Bit1 ->1 and (WRD2)Bit2 -> 0 condition
+            Store(0x5555,WRD1)
+            Store(0xAAAA,WRD2)
+            XOr(WRD1,WRD2,RWRD)
+            if(LNotEqual(RWRD,0xFFFF))
+            {   Return(1)}
+
+            //(WRD1) Bit1 ->0 and (WRD2)Bit2 -> 1 condition
+            Store(0xAAAA,WRD1)
+            Store(0x5555,WRD2)
+            XOr(WRD1,WRD2,RWRD)
+            if(LNotEqual(RWRD,0xFFFF))
+            {   Return(1)}
+
+            Store(0x1234,WRD1)
+            Store(0xEDCB,WRD2)
+            XOr(WRD1,WRD2,RWRD)
+            if(LNotEqual(RWRD,0xFFFF))
+            {   Return(1)}
+
+            // Store known values for checking later
+            Store(0x1234, WRD1)
+            if(LNotEqual(WRD1, 0x1234))
+            {
+                Return(1)
+            }
+
+            Store(0xFEDC, WRD2)
+            if(LNotEqual(WRD2, 0xFEDC))
+            {
+                Return(1)
+            }
+
+            Store(0x87AB, RWRD)
+            if(LNotEqual(RWRD, 0x87AB))
+            {
+                Return(1)
+            }
+
+
+            //**************************************************
+            // (DWD1) Bit1 ->0 and (DWD2)Bit2 -> 0 condition
+            Store(0x00000000,DWD1)
+            Store(0x00000000,DWD2)
+            XOr(DWD1,DWD2,RDWD)
+            if(LNotEqual(RDWD,0))
+            {   Return(1)}
+
+            // (DWD1) Bit1 ->1 and (DWD2)Bit2 -> 1 condition
+            Store(0xffffffff,DWD1)
+            Store(0xffffffff,DWD2)
+            XOr(DWD1,DWD2,RDWD)
+            if(LNotEqual(RDWD,0))
+            {   Return(1)}
+
+            // (DWD1) Bit1 ->1 and (DWD2)Bit2 -> 0 condition
+            Store(0x55555555,DWD1)
+            Store(0xAAAAAAAA,DWD2)
+            XOr(DWD1,DWD2,RDWD)
+            if(LNotEqual(RDWD,0xFFFFFFFF))
+            {   Return(1)}
+
+            //(DWD1) Bit1 ->0 and (DWD2)Bit2 -> 1 condition
+            Store(0xAAAAAAAA,DWD1)
+            Store(0x55555555,DWD2)
+            XOr(DWD1,DWD2,RDWD)
+            if(LNotEqual(RDWD,0xFFFFFFFF))
+            {   Return(1)}
+
+            // (DWD1) Bit1 ->1 and (DWD2)Bit2 -> 0 condition
+            Store(0x12345678,DWD1)
+            Store(0xEDCBA987,DWD2)
+            XOr(DWD1,DWD2,RDWD)
+            if(LNotEqual(RDWD,0xFFFFFFFF))
+            {   Return(1)}
+
+            Store(0x12345678,DWD1)
+            if(LNotEqual(DWD1,0x12345678))
+            {
+                Return(1)
+            }
+
+            Store(0xFEDCBA98,DWD2)
+            if(LNotEqual(DWD2,0xFEDCBA98))
+            {
+                Return(1)
+            }
+
+            Store(0x91827364,RDWD)
+            if(LNotEqual(RDWD,0x91827364))
+            {
+                Return(1)
+            }
+
+            //****************************************************
+            // Check the stored single bits
+            if(LNotEqual(RES1, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES2, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES3, 1))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES4, 1))
+            {
+                Return(1)
+            }
+
+            // Change all of the single bit fields to zero
+            Store(0, RES1)
+            Store(0, RES2)
+            Store(0, RES3)
+            Store(0, RES4)
+
+            // Now, check all of the fields
+
+            // Byte
+            if(LNotEqual(BYT1, 0x12))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(BYT2, 0xFE))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RBYT, 0xAB))
+            {
+                Return(1)
+            }
+
+            // Word
+            if(LNotEqual(WRD1, 0x1234))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(WRD2, 0xFEDC))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RWRD, 0x87AB))
+            {
+                Return(1)
+            }
+
+            // Dword
+            if(LNotEqual(DWD1, 0x12345678))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(DWD2, 0xFEDCBA98))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RDWD, 0x91827364))
+            {
+                Return(1)
+            }
+
+            // Bits
+            if(LNotEqual(RES1, 0))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES2, 0))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES3, 0))
+            {
+                Return(1)
+            }
+
+            if(LNotEqual(RES4, 0))
+            {
+                Return(1)
+            }
+
+
+            Return(0)
+        }   //  TEST
+    }   //  XORD
+
+//
+// test CrBytFld.asl
+//
+//  CrBytFld test
+//      Test for CreateByteField.
+//      Tests creating byte field overlay of buffer stored in Local0.
+//      Tests need to be added for Arg0 and Name buffers.
+//
+    Device (CRBF)
+    {   //  Test device name
+        Method (TEST)
+        {
+            Store ("++++++++ CrBytFld Test", Debug)
+
+            //  Local0 is unitialized buffer with 4 elements
+            Store (Buffer (4) {}, Local0)
+
+            //  create Byte Field named BF0 based on Local0 element 0
+            CreateByteField (Local0, 0, BF0)
+
+            //  validate CreateByteField did not alter Local0
+            Store (ObjectType (Local0), Local1) //  Local1 = Local0 object type
+            If (LNotEqual (Local1, 3))  //  Buffer object type value is 3
+                {   Return (2)  }
+
+            //  store something into BF0
+            Store (1, BF0)
+
+            //  validate Store did not alter Local0 object type
+            Store (ObjectType (Local0), Local1) //  Local1 = Local0 object type
+            If (LNotEqual (Local1, 3))  //  Buffer object type value is 3
+                {   Return (3)  }
+
+            //  verify that the Store into BF0 was successful
+            If (LNotEqual (BF0, 1))
+                {   Return (4)  }
+
+
+            //  create Byte Field named BF1 based on Local0 element 1
+            CreateByteField (Local0, 1, BF1)
+
+            //  validate CreateByteField did not alter Local0
+            Store (ObjectType (Local0), Local1) //  Local1 = Local0 object type
+            If (LNotEqual (Local1, 3))  //  Buffer object type value is 3
+                {   Return (10) }
+
+            //  store something into BF1
+            Store (5, BF1)
+
+            //  validate Store did not alter Local0 object type
+            Store (ObjectType (Local0), Local1) //  Local1 = Local0 object type
+            If (LNotEqual (Local1, 3))  //  Buffer object type value is 3
+                {   Return (11) }
+
+            //  verify that the Store into BF1 was successful
+            If (LNotEqual (BF1, 5))
+                {   Return (12) }
+
+            //  verify that the Store into BF1 did not alter BF0
+            If (LNotEqual (BF0, 1))
+                {   Return (13) }
+
+
+            //  store something into BF0
+            Store (0xFFFF, BF0)
+
+            //  verify that the Store into BF0 was successful
+            If (LNotEqual (BF0, 0xFF))
+                {   Return (20) }
+
+            //  verify that the Store into BF0 did not alter BF1
+            If (LNotEqual (BF1, 5))
+                {   Return (21) }
+
+
+            Return (0)
+        }   //  TEST
+    }   //  CRBF
+
+//
+// test IndexOp4.asl
+//
+//  IndexOp4 test
+//      This is just a subset of the many RegionOp/Index Field test cases.
+//      Tests access of index fields smaller than 8 bits.
+//
+    Device (IDX4)
+    {   //  Test device name
+
+        //  MADM:   Misaligned Dynamic RAM SystemMemory OperationRegion
+        //          Tests OperationRegion memory access using misaligned BYTE,
+        //          WORD, and DWORD field element accesses. Validation is performed
+        //          using both misaligned field entries and aligned field entries.
+        //
+        //          MADM returns 0 if all test cases pass or non-zero identifying
+        //          the failing test case for debug purposes. This non-zero numbers
+        //          are not guaranteed to be in perfect sequence (i.e., test case
+        //          index), but are guaranteed to be unique so the failing test
+        //          case can be uniquely identified.
+        //
+        Method (MADM, 1)    //  Misaligned Dynamic RAM SystemMemory OperationRegion
+        //  Arg0    --  SystemMemory OperationRegion base address
+        {   //  MADM:   Misaligned Dynamic RAM SystemMemory OperationRegion
+            OperationRegion (RAM, SystemMemory, Arg0, 0x100)
+            Field (RAM, DwordAcc, NoLock, Preserve)
+            {   //  aligned field definition (for verification)
+                DWD0,   32, //  aligned DWORD field
+                DWD1,   32      //  aligned DWORD field
+            }
+            Field (RAM, ByteAcc, NoLock, Preserve)
+            {   //  bit access field definition
+                BIT0,   1,      //  single bit field entry
+                BIT1,   1,      //  single bit field entry
+                BIT2,   1,      //  single bit field entry
+                BIT3,   1,      //  single bit field entry
+                BIT4,   1,      //  single bit field entry
+                BIT5,   1,      //  single bit field entry
+                BIT6,   1,      //  single bit field entry
+                BIT7,   1,      //  single bit field entry
+                BIT8,   1,      //  single bit field entry
+                BIT9,   1,      //  single bit field entry
+                BITA,   1,      //  single bit field entry
+                BITB,   1,      //  single bit field entry
+                BITC,   1,      //  single bit field entry
+                BITD,   1,      //  single bit field entry
+                BITE,   1,      //  single bit field entry
+                BITF,   1,      //  single bit field entry
+                BI10,   1,      //  single bit field entry
+                BI11,   1,      //  single bit field entry
+                BI12,   1,      //  single bit field entry
+                BI13,   1,      //  single bit field entry
+                BI14,   1,      //  single bit field entry
+                BI15,   1,      //  single bit field entry
+                BI16,   1,      //  single bit field entry
+                BI17,   1,      //  single bit field entry
+                BI18,   1,      //  single bit field entry
+                BI19,   1,      //  single bit field entry
+                BI1A,   1,      //  single bit field entry
+                BI1B,   1,      //  single bit field entry
+                BI1C,   1,      //  single bit field entry
+                BI1D,   1,      //  single bit field entry
+                BI1E,   1,      //  single bit field entry
+                BI1F,   1       //  single bit field entry
+            }   //  bit access field definition
+
+            Field (RAM, ByteAcc, NoLock, Preserve)
+            {   //  two-bit access field definition
+                B2_0,   2,      //  single bit field entry
+                B2_1,   2,      //  single bit field entry
+                B2_2,   2,      //  single bit field entry
+                B2_3,   2,      //  single bit field entry
+                B2_4,   2,      //  single bit field entry
+                B2_5,   2,      //  single bit field entry
+                B2_6,   2,      //  single bit field entry
+                B2_7,   2,      //  single bit field entry
+                B2_8,   2,      //  single bit field entry
+                B2_9,   2,      //  single bit field entry
+                B2_A,   2,      //  single bit field entry
+                B2_B,   2,      //  single bit field entry
+                B2_C,   2,      //  single bit field entry
+                B2_D,   2,      //  single bit field entry
+                B2_E,   2,      //  single bit field entry
+                B2_F,   2       //  single bit field entry
+            }   //  bit access field definition
+
+            //  initialize memory contents using aligned field entries
+            Store (0x5AA55AA5, DWD0)
+            Store (0x5AA55AA5, DWD1)
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BIT0)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT0, 0))
+                    {   Return (1)  }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55AA4))
+                    {   Return (2)  }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BIT1)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT1, 1))
+                    {   Return (3)  }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55AA6))
+                    {   Return (4)  }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BIT2)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT2, 0))
+                    {   Return (5)  }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55AA2))
+                    {   Return (6)  }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BIT3)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT3, 1))
+                    {   Return (7)  }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55AAA))
+                    {   Return (8)  }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BIT4)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT4, 1))
+                    {   Return (9)  }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55ABA))
+                    {   Return (10) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BIT5)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT5, 0))
+                    {   Return (11) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55A9A))
+                    {   Return (12) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BIT6)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT6, 1))
+                    {   Return (13) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55ADA))
+                    {   Return (14) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BIT7)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT7, 0))
+                    {   Return (15) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55A5A))
+                    {   Return (16) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BIT8)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT8, 1))
+                    {   Return (17) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55B5A))
+                    {   Return (18) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BIT9)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BIT9, 0))
+                    {   Return (19) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5595A))
+                    {   Return (20) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BITA)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITA, 1))
+                    {   Return (21) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA55D5A))
+                    {   Return (22) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BITB)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITB, 0))
+                    {   Return (23) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5555A))
+                    {   Return (24) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BITC)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITC, 0))
+                    {   Return (25) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5455A))
+                    {   Return (26) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BITD)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITD, 1))
+                    {   Return (27) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5655A))
+                    {   Return (28) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BITE)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITE, 0))
+                    {   Return (29) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5255A))
+                    {   Return (30) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BITF)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BITF, 1))
+                    {   Return (31) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA5A55A))
+                    {   Return (32) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI10)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI10, 0))
+                    {   Return (33) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA4A55A))
+                    {   Return (34) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI11)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI11, 1))
+                    {   Return (35) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA6A55A))
+                    {   Return (36) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI12)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI12, 0))
+                    {   Return (37) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AA2A55A))
+                    {   Return (38) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI13)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI13, 1))
+                    {   Return (39) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5AAAA55A))
+                    {   Return (40) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI14)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI14, 1))
+                    {   Return (41) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5ABAA55A))
+                    {   Return (42) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI15)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI15, 0))
+                    {   Return (43) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5A9AA55A))
+                    {   Return (44) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI16)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI16, 1))
+                    {   Return (45) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5ADAA55A))
+                    {   Return (46) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI17)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI17, 0))
+                    {   Return (47) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5A5AA55A))
+                    {   Return (48) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI18)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI18, 1))
+                    {   Return (49) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5B5AA55A))
+                    {   Return (50) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI19)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI19, 0))
+                    {   Return (51) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x595AA55A))
+                    {   Return (52) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI1A)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1A, 1))
+                    {   Return (53) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x5D5AA55A))
+                    {   Return (54) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI1B)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1B, 0))
+                    {   Return (55) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x555AA55A))
+                    {   Return (56) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI1C)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1C, 0))
+                    {   Return (57) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x455AA55A))
+                    {   Return (58) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI1D)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1D, 1))
+                    {   Return (59) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x655AA55A))
+                    {   Return (60) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, BI1E)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1E, 0))
+                    {   Return (61) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x255AA55A))
+                    {   Return (62) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, BI1F)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (BI1F, 1))
+                    {   Return (63) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA55A))
+                    {   Return (64) }
+
+
+            //  set memory contents to known values using misaligned field entries
+            Store (3, B2_0)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_0, 3))
+                    {   Return (65) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA55B))
+                    {   Return (66) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, B2_1)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_1, 1))
+                    {   Return (67) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA557))
+                    {   Return (68) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, B2_2)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_2, 0))
+                    {   Return (69) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA547))
+                    {   Return (70) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (3, B2_3)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_3, 3))
+                    {   Return (71) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA5C7))
+                    {   Return (72) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (3, B2_4)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_4, 3))
+                    {   Return (73) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA7C7))
+                    {   Return (74) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, B2_5)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_5, 0))
+                    {   Return (75) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55AA3C7))
+                    {   Return (76) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, B2_6)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_6, 1))
+                    {   Return (77) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55A93C7))
+                    {   Return (78) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, B2_7)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_7, 1))
+                    {   Return (79) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55A53C7))
+                    {   Return (80) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, B2_8)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_8, 0))
+                    {   Return (81) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55853C7))
+                    {   Return (82) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, B2_9)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_9, 1))
+                    {   Return (83) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA55453C7))
+                    {   Return (84) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (2, B2_A)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_A, 2))
+                    {   Return (85) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA56453C7))
+                    {   Return (86) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (2, B2_B)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_B, 2))
+                    {   Return (87) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA5A453C7))
+                    {   Return (88) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (3, B2_C)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_C, 3))
+                    {   Return (89) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xA7A453C7))
+                    {   Return (90) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (3, B2_D)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_D, 3))
+                    {   Return (91) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0xAFA453C7))
+                    {   Return (92) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (1, B2_E)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_E, 1))
+                    {   Return (93) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x9FA453C7))
+                    {   Return (94) }
+
+            //  set memory contents to known values using misaligned field entries
+            Store (0, B2_F)
+                //  verify memory contents using misaligned field entries
+                If (LNotEqual (B2_F, 0))
+                    {   Return (95) }
+                //  verify memory contents using aligned field entries
+                If (LNotEqual (DWD0, 0x1FA453C7))
+                    {   Return (96) }
+
+
+            Return (0)  //  pass
+        }   //  MADM:   Misaligned Dynamic RAM SystemMemory OperationRegion
+
+        Method (TEST)
+        {
+            Store ("++++++++ IndexOp4 Test", Debug)
+
+            //  MADM (Misaligned Dynamic RAM SystemMemory OperationRegion) arguments:
+            //      Arg0    --  SystemMemory OperationRegion base address
+            Store (MADM (0x800000), Local0)
+            If (LNotEqual (Local0, 0))      //  MADM returns zero if successful
+                {   Return (Local0) }       //  failure:    return MADM error code
+
+            Return (Local0)
+        }   //  TEST
+    }   //  IDX4
+
+//
+// test Event.asl
+//
+//  EventOp, ResetOp, SignalOp, and WaitOp test cases.
+//
+    Device (EVNT)
+    {
+        Event (EVNT)    //  event synchronization object
+
+        Method (TEVN, 1)
+        //  Arg0:   time to Wait for event in milliseconds
+        {   //  TEVN control method to test ResetOp, SignalOp, and WaitOp
+            //  reset EVNT to initialization (zero) state
+            Reset (EVNT)
+
+            //  prime EVNT with two outstanding signals
+            Signal (EVNT)
+            Signal (EVNT)
+
+
+            //  acquire existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x21)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LNotEqual (Local0, 0))  //  Number is type 1
+                {   Return (0x22)   }       //  timeout occurred without acquiring signal
+
+            Store ("Acquire 1st existing signal PASS", Debug)
+
+
+            //  acquire existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x31)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LNotEqual (Local0, 0))  //  Number is type 1
+                {   Return (0x32)   }       //  timeout occurred without acquiring signal
+
+            Store ("Acquire 2nd existing signal PASS", Debug)
+
+
+            //  ensure WaitOp timeout test cases do not hang
+            if (LEqual (Arg0, 0xFFFF))
+                {   Store (0xFFFE, Arg0)    }
+
+            //  acquire non-existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x41)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LEqual (Local0, 0))     //  Number is type 1
+                {   Return (0x42)   }       //  non-existant signal was acquired
+
+            Store ("Acquire signal timeout PASS", Debug)
+
+
+            //  prime EVNT with two outstanding signals
+            Signal (EVNT)
+            Signal (EVNT)
+
+            //  reset EVNT to initialization (zero) state
+            Reset (EVNT)
+
+            //  acquire non-existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x51)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LEqual (Local0, 0))     //  Number is type 1
+                {   Return (0x52)   }       //  non-existant signal was acquired
+
+            Store ("Reset signal PASS", Debug)
+
+
+            //  acquire non-existing signal using Lvalue timeout
+            Store (Wait (EVNT, Zero), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x61)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LEqual (Local0, 0))     //  Number is type 1
+                {   Return (0x62)   }       //  non-existant signal was acquired
+
+            Store ("Zero Lvalue PASS", Debug)
+
+
+            //  acquire non-existing signal using Lvalue timeout
+            Store (Wait (EVNT, One), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x71)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LEqual (Local0, 0))     //  Number is type 1
+                {   Return (0x72)   }       //  non-existant signal was acquired
+
+            Store ("One Lvalue PASS", Debug)
+
+            //  Lvalue Event test cases
+    // ILLEGAL SOURCE OPERAND        Store (EVNT, Local2)
+
+            //  validate Local2 is an Event
+            Store (ObjectType (EVNT), Local1)
+            If (LNotEqual (Local1, 7))  //  Event is type 7
+                {   Return (0x81)   }       //  Local1 indicates Local0 is not a Number
+
+            //  reset EVNT to initialization (zero) state
+            Reset (EVNT)
+
+            //  prime EVNT with two outstanding signals
+            Signal (EVNT)
+
+            //  acquire existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x82)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LNotEqual (Local0, 0))  //  Number is type 1
+                {   Return (0x83)   }       //  timeout occurred without acquiring signal
+
+            Store ("Acquire Lvalue existing signal PASS", Debug)
+
+
+            //  acquire non-existing signal
+            Store (Wait (EVNT, Arg0), Local0)
+
+            //  validate Local0 is a Number
+            Store (ObjectType (Local0), Local1)
+            If (LNotEqual (Local1, 1))  //  Number is type 1
+                {   Return (0x84)   }       //  Local1 indicates Local0 is not a Number
+
+            If (LEqual (Local0, 0))     //  Number is type 1
+                {   Return (0x85)   }       //  non-existant signal was acquired
+
+            Store ("Acquire Lvalue signal timeout PASS", Debug)
+
+
+            Return (0)  //  success
+        }   //  TEVN control method to test ResetOp, SignalOp, and WaitOp
+
+        Method (TEST)
+        {
+            Store ("++++++++ Event Test", Debug)
+
+            Store (TEVN (100), Local0)
+
+            Return (Local0)
+        }   //  TEST
+    }   //  EVNT
+
+//
+// test SizeOfLv.asl
+//
+//  Test for SizeOf (Lvalue)
+//
+//  This next section will contain the packages that the SizeOfOp will be
+//  exercised on.  The first one, PKG0, is a regular package of 3 elements.
+//  The 2nd one, PKG1, is a nested package with 3 packages inside it, each
+//  with 3 elements.  It is expected that SizeOf operator will return the
+//  same value for these two packages since they both have 3 elements.  The
+//  final package, PKG2, has 4 elements and the SizeOf operator is expected
+//  to return different results for this package.
+
+    Name (PKG0,
+        Package (3)
+        {0x0123, 0x4567, 0x89AB}
+    )   //  PKG0
+
+    Name (PKG1,
+        Package (3)
+        {
+            Package (3) {0x0123, 0x4567, 0x89AB},
+            Package (3) {0xCDEF, 0xFEDC, 0xBA98},
+            Package (3) {0x7654, 0x3210, 0x1234}
+        }
+    )   //  PKG1
+
+    Name (PKG2,
+        Package (4)
+        {0x0123, 0x4567, 0x89AB, 0x8888}
+    )   //  PKG2
+
+    Name (PKG3,
+        Package (5)
+        {0x0123, 0x4567, 0x89AB, 0x8888, 0x7777}
+    )   //  PKG3
+
+//  End Packages    **********************************************************
+
+//  The following section will declare the data strings that will be used to
+//  exercise the SizeOf operator.  STR0 and STR1 are expected to be equal,
+//  STR2 is expected to have a different SizeOf value than STR0 and STR1.
+
+    Name (STR0, "ACPI permits very flexible methods of expressing a system")
+
+    Name (STR1, "MIKE permits very flexible methods of expressing a system")
+
+    Name (STR2, "Needless to say, Mike and ACPI are frequently at odds")
+
+//  This string is being made in case we want to do a SizeOf comparison
+//  between strings and packages or buffers
+    Name (STR3, "12345")
+
+//  End Strings     **********************************************************
+
+//  The following section will declare the buffers that will be used to exercise
+//  the SizeOf operator.
+
+    Name (BUF0, Buffer (10) {})
+    Name (BUF1, Buffer (10) {})
+    Name (BUF2, Buffer (8)  {})
+    Name (BUF3, Buffer (5)  {})
+
+//  End Buffers     **********************************************************
+    Device (SZLV)
+    {
+
+        Method (CMPR, 2)
+        {
+            //  CMPR is passed two arguments.  If unequal, return 1 to indicate
+            //  that, otherwise return 0 to indicate SizeOf each is equal.
+
+            Store (0x01, Local0)
+
+            if (LEqual (SizeOf(Arg0), SizeOf(Arg1)))
+            {
+                Store (0x00, Local0)
+            }
+
+            return (Local0)
+        }   //  CMPR
+
+
+        Method (TEST)
+        {
+
+            Store ("++++++++ SizeOfLv Test", Debug)
+
+            //  TBD:    SizeOf ("string")
+            //          SizeOf (Buffer)
+            //          SizeOf (Package)
+            //          SizeOf (String)
+            //          SizeOf (STR0)   --  where Name (STR0,...) -- lot's of cases
+            //              buffer, string, package,
+            //          SizeOf (METH) -- where METH control method returns
+            //              buffer, string, package,
+
+            //  TBD:    SLOC [SizeOf (Local0)] -- dup SARG
+
+            //  Compare the elements that we expect to be the same.  Exit out with an error
+            //  code on the first failure.
+            if (LNotEqual (0x00, CMPR (STR0, STR1)))
+            {
+                Return (0x01)
+            }
+
+            if (LNotEqual (0x00, CMPR (STR3, BUF3)))
+            {
+                Return (0x02)
+            }
+
+            if (LNotEqual (0x00, CMPR (STR3, PKG3)))
+            {
+                Return (0x03)
+            }
+
+            //  In the following section, this test will cover the SizeOf
+            //  operator for Local values.
+            //  In this case, both Local0 and Local1 should have the same Size
+            Store (STR0, Local0)
+            Store (STR1, Local1)
+
+            if (LNotEqual (SizeOf (Local0), SizeOf (Local1)))
+            {
+                Return (0x04)
+            }
+
+            //  Now create a case where Local0 and Local1 are different
+            Store (STR2, Local1)
+
+            if (LEqual (SizeOf (Local0), SizeOf (Local1)))
+            {
+                Return (0x05)
+            }
+
+            //  Finally, check for the return of SizeOf for a known Buffer.  Just
+            //  in case we magically pass above cases due to all Buffers being Zero
+            //  bytes in size, or Infinity, etc.
+            if (LNotEqual (0x05, SizeOf (BUF3)))
+            {
+                Return (0x06)
+            }
+
+            Return (0)
+        }   //  TEST
+    }   //  SZLV
+
+
+//
+// test BytField.asl
+//
+//  BytField test
+//      This is just a subset of the many RegionOp/Index Field test cases.
+//      Tests access of TBD.
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+        Device (BYTF)
+        {   //  Test device name
+            Method (TEST)
+            {
+                Store ("++++++++ BytField Test", Debug)
+
+                Return (\_TZ.C19B.RSLT)
+            }   //  TEST
+        }   //  BYTF
+
+        Device (C005)
+        {   //  Device C005
+            Device (C013)
+            {   //  Device C013
+            }   //  Device C013
+        }   //  Device C005
+
+        Method (C115)
+        {   //  C115 control method
+            Acquire (\_GL, 0xFFFF)
+            Store (\_SB.C005.C013.C058.C07E, Local0)
+            Release (\_GL)
+            And (Local0, 16, Local0)
+            Store (ShiftRight (Local0, 4, ), Local1)
+            If (LEqual (Local1, 0))
+                {   Return (1)  }
+            Else
+                {   Return (0)  }
+        }   //  C115 control method
+    }   //  _SB system bus
+
+    OperationRegion (C018, SystemIO, 0x5028, 4)
+    Field (C018, AnyAcc, NoLock, Preserve)
+    {   //  Field overlaying C018
+        C019,   32
+    }   //  Field overlaying C018
+
+    OperationRegion (C01A, SystemIO, 0x5030, 4)
+    Field (C01A, ByteAcc, NoLock, Preserve)
+    {   //  Field overlaying C01A
+        C01B,   8,
+        C01C,   8,
+        C01D,   8,
+        C01E,   8
+    }   //  Field overlaying C01A
+
+    Mutex (\C01F, 0)
+    Name (\C020, 0)
+    Name (\C021, 0)
+
+    Method (\C022, 0)
+    {   //  \C022 control method
+        Acquire (\C01F, 0xFFFF)
+        If (LEqual (\C021, 0))
+        {
+            Store (C019, Local0)
+            And (Local0, 0xFFFEFFFE, Local0)
+            Store (Local0, C019)
+            Increment (\C021)
+        }
+        Release (\C01F)
+    }   //  \C022 control method
+
+    Scope (\_SB.C005.C013)
+    {   //  Scope \_SB.C005.C013
+        Device (C058)
+        {   //  Device C058
+            Name (_HID, "*PNP0A06")
+
+            OperationRegion (C059, SystemIO, 0xE0, 2)
+            Field (C059, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C059
+                C05A,   8,
+                C05B,   8
+            }   //  Field overlaying C059
+
+            OperationRegion (C05C, SystemIO, 0xE2, 2)
+            Field (C05C, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C05C
+                C05D,   8,
+                C05E,   8
+            }   //  Field overlaying C05C
+            IndexField (C05D, C05E, ByteAcc, NoLock, Preserve)
+            {   //  IndexField overlaying C05D/C05E
+                ,       0x410,  //  skip
+                C05F,   8,
+                C060,   8,
+                C061,   8,
+                C062,   8,
+                C063,   8,
+                C064,   8,
+                C065,   8,
+                C066,   8,
+                C067,   8,
+                C068,   8,
+                C069,   8,
+                C06A,   8,
+                C06B,   8,
+                C06C,   8,
+                C06D,   8,
+                C06E,   8,
+                ,       0x70,       //  skip
+                C06F,   8,
+                C070,   8,
+                C071,   8,
+                C072,   8,
+                C073,   8,
+                C074,   8,
+                C075,   8,
+                C076,   8,
+                C077,   8,
+                C078,   8,
+                C079,   8,
+                C07A,   8,
+                C07B,   8,
+                C07C,   8,
+                C07D,   8,
+                C07E,   8
+            }   //  IndexField overlaying C05D/C05E
+
+            OperationRegion (C07F, SystemIO, 0xE4, 2)
+            Field (C07F, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C07F
+                C080,   8,
+                C081,   8
+            }   //  Field overlaying C07F
+
+            OperationRegion (C082, SystemIO, 0xE0, 1)
+            Field (C082, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C082
+                C083,   8
+            }   //  Field overlaying C082
+
+            OperationRegion (C084, SystemIO, 0xFF, 1)
+            Field (C084, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C084
+                C085,   8
+            }   //  Field overlaying C084
+
+            OperationRegion (C086, SystemIO, 0xFD, 1)
+            Field (C086, ByteAcc, NoLock, Preserve)
+            {   //  Field overlaying C086
+                C087,   8
+            }   //  Field overlaying C086
+
+            Mutex (C088, 0)
+            Mutex (C089, 0)
+            Mutex (C08A, 0)
+            Mutex (C08B, 0)
+            Mutex (C08C, 0)
+            Mutex (C08D, 0)
+
+            Name (C08E, 0xFFFFFFFD)
+            Name (C08F, 0)
+
+            Method (C0AA, 4)
+            {   //  C0AA control method
+                Store (Buffer (4) {}, Local7)
+                CreateByteField (Local7, 0, C0AB)
+                CreateByteField (Local7, 1, C0AC)
+                CreateByteField (Local7, 2, C0AD)
+                CreateByteField (Local7, 3, C0AE)
+                Acquire (^C08B, 0xFFFF)
+                Acquire (\_GL, 0xFFFF)
+                \C022 ()
+                Store (1, \_SB.C005.C013.C058.C06B)
+                While (LNot (LEqual (0, \_SB.C005.C013.C058.C06B)))
+                    {   Stall (100) }
+                Store (Arg3, \_SB.C005.C013.C058.C06E)
+                Store (Arg2, \_SB.C005.C013.C058.C06D)
+                Store (Arg1, \_SB.C005.C013.C058.C06C)
+                Store (Arg0, \_SB.C005.C013.C058.C06B)
+                While (LNot (LEqual (0, \_SB.C005.C013.C058.C06B)))
+                    {   Stall (100) }
+                Store (\_SB.C005.C013.C058.C06E, C0AB)
+                Store (\_SB.C005.C013.C058.C06D, C0AC)
+                Store (\_SB.C005.C013.C058.C06C, C0AD)
+                Store (\_SB.C005.C013.C058.C06B, C0AE)
+                If (LNot (LEqual (Arg0, 23)))
+                {
+                    Store (2, \_SB.C005.C013.C058.C06B)
+                    Stall (100)
+                }
+                Release (\_GL)
+                Release (^C08B)
+                Return (Local7)
+            }   //  C0AA control method
+        }   //  Device C058
+    }   //  Scope \_SB.C005.C013
+
+    Scope (\_TZ)
+    {   //  \_TZ thermal zone scope
+        Name (C18B, Package (2)
+        {
+            Package (2)
+            {
+                Package (5) {0x05AC, 0x0CD2, 0x0D68, 0x0DE0, 0x0E4E},
+                Package (5) {0x0D04, 0x0D9A, 0x0DFE, 0x0E80, 0x0FA2}
+            },
+            Package (2)
+            {
+                Package (5) {0x05AC, 0x0CD2, 0x0D68, 0x0DE0, 0x0E4E},
+                Package (5) {0x0D04, 0x0D9A, 0x0DFE, 0x0E80, 0x0FA2}
+            }
+        })  //  C18B
+
+        Name (C18C, Package (2)
+        {
+            Package (2)
+            {
+                Package (3) {0x64, 0x4B, 0x32},
+                Package (3) {0x64, 0x4B, 0x32}
+            }
+        })  //  C81C
+
+        Name (C18D, 0)
+        Name (C18E, 0)
+        Name (C18F, 0)
+        Name (C190, 0)
+        Name (C191, 3)
+        Name (C192, 0)
+        Name (C193, 1)
+        Name (C194, 2)
+        Mutex (C195, 0)
+        Name (C196, 1)
+        Name (C197, 0x0B9C)
+        Name (C198, 0x0B9C)
+        Name (C199, 0xFFFFFFFD)
+        Name (C19A, 0)
+
+        Device (C19B)
+        {   //  Device C19B
+            Name (RSLT, 0)  //  default to zero
+
+            Method (XINI)
+            {   //  _INI control method (Uses Global Lock -- can't run under AcpiExec)
+                Store (\_SB.C115, C19A)
+                \_TZ.C19C._SCP (0)
+                Subtract (0x0EB2, 0x0AAC, Local1)   //  Local1 = AACh - EB2h
+                Divide (Local1, 10, Local0, Local2) //  Local0 = Local1 / 10
+                                                                //  Local2 = Local1 % 10
+                \_SB.C005.C013.C058.C0AA (14, Local2, 0, 0)
+                Store
+                    (DerefOf (Index (DerefOf (Index (\_TZ.C18C, C19A, )), 0, )), C18D)
+                Store
+                    (DerefOf (Index (DerefOf (Index (\_TZ.C18C, C19A, )), 1, )), C18E)
+                Store
+                    (DerefOf (Index (DerefOf (Index (\_TZ.C18C, C19A, )), 2, )), C18F)
+
+                Store (1, RSLT) //  set RSLT to 1 if _INI control method completes
+            }   //  _INI control method
+
+            //  PowerResource (C19D) {...}
+        }   //  Device C19B
+
+        ThermalZone (C19C)
+        {
+            Method (_SCP, 1)
+            {   //  _SCP control method
+                Store (Arg0, Local0)
+                If (LEqual (Local0, 0))
+                {
+                    Store (0, \_TZ.C192)
+                    Store (1, \_TZ.C193)
+                    Store (2, \_TZ.C194)
+                    Store (3, \_TZ.C191)
+                }
+                Else
+                {
+                    Store (0, \_TZ.C191)
+                    Store (1, \_TZ.C192)
+                    Store (2, \_TZ.C193)
+                    Store (3, \_TZ.C194)
+                }
+            }   //  _SCP control method
+        }   //  ThermalZone C19C
+    }   //  \_TZ thermal zone scope
+
+
+//
+// test DwrdFld.asl
+//
+    Name (BUFR, buffer(10) {0,0,0,0,0,0,0,0,0,0} )
+
+    Device (DWDF)
+    {
+        Method (TEST)
+        {
+            Store ("++++++++ DwrdFld Test", Debug)
+
+            CreateByteField (BUFR, 0, BYTE)
+            Store (0xAA, BYTE)
+
+            CreateWordField (BUFR, 1, WORD)
+            Store (0xBBCC, WORD)
+
+            CreateDWordField (BUFR, 3, DWRD)
+            Store (0xDDEEFF00, DWRD)
+
+            CreateByteField (BUFR, 7, BYT2)
+            Store (0x11, BYT2)
+
+            CreateWordField (BUFR, 8, WRD2)
+            Store (0x2233, WRD2)
+
+            Return (0)
+
+        }   //  End Method TEST
+    }   //  Device DWDF
+
+    //
+    // test DivAddx.asl
+    //
+    Name (B1LO, 0xAA)
+    Name (B1HI, 0xBB)
+
+    Method (MKW_, 2)
+    {   //  This control method will take two bytes and make them into a WORD
+
+        Multiply (B1HI, 256, Local0)    //  Make high byte.....high
+        Or (Local0, B1LO, Local0)       //  OR in the low byte
+        Return (Local0)                 //  Return the WORD
+
+    }   //  MKW_
+
+    Device (DVAX)
+    {
+        Method (TEST)
+        {
+
+            Store ("++++++++ DivAddx Test", Debug)
+
+            Store (25, B1LO)
+            Store (0, B1HI)
+
+            //  We'll multiply 25 * 3 to get 75, add 99 to it then divide
+            //  by 100.  We expect to get 74 for the remainder and 1 for
+            //  the quotient.
+            Divide(
+                Add (Multiply (3, MKW_ (B1LO, B1HI)), 0x63),
+                            //  Dividend,
+                100,        //  Divisor
+                Local4,     //  Remainder
+                Local2)     //  Quotient
+
+            If (LAnd (LEqual (74, Local4), LEqual (1, Local2)))
+            {   //  Indicate Pass
+                Store (0x00, Local0)
+            }
+
+            Else
+            {   //  Indicate Fail
+                Store (0x01, Local0)
+            }
+
+            Return (Local0)
+        }   //  End Method TEST
+    }   //  Device DVAX
+
+//
+// test IndexFld.asl (IndexOp6.asl)
+//
+//  IndexFld test
+//      This is just a subset of the many RegionOp/Index Field test cases.
+//      Tests index field element AccessAs macro.
+//      Also tests name resolution of index field elements with same names
+//      but different namespace scopes.
+//
+    Device (IDX6)
+    {   //  Test device name
+
+        OperationRegion (SIO, SystemIO, 0x100, 2)
+        Field (SIO, ByteAcc, NoLock, Preserve)
+        {
+            INDX,   8,
+            DATA,   8
+        }
+        IndexField (INDX, DATA, AnyAcc, NoLock, WriteAsOnes)
+        {
+            AccessAs (ByteAcc, 0),
+            IFE0,   8,
+            IFE1,   8,
+            IFE2,   8,
+            IFE3,   8,
+            IFE4,   8,
+            IFE5,   8,
+            IFE6,   8,
+            IFE7,   8,
+            IFE8,   8,
+            IFE9,   8,
+        }
+
+        Device (TST_)
+        {   //  TST_:   provides a different namespace scope for IFE0 and IFE1
+            OperationRegion (SIO2, SystemIO, 0x100, 2)
+            Field (SIO2, ByteAcc, NoLock, Preserve)
+            {
+                IND2,   8,
+                DAT2,   8
+            }
+            IndexField (IND2, DAT2, AnyAcc, NoLock, WriteAsOnes)
+            {
+                IFE0,   8,  //  duplicate IndexField name with different scope
+                IFE1,   8
+            }
+        }   //  TST_:   provides a different namespace scope for IFE0 and IFE1
+
+        Method (TEST)
+        {
+            Store ("++++++++ IndexOp6 Test", Debug)
+
+            Store (IFE0, Local0)
+            Store (IFE1, Local1)
+            Store (IFE2, Local2)
+
+            //  validate name resolution of IndexFields with different scopes
+            Store (\IDX6.IFE0, Local3)
+            Store (\IDX6.IFE1, Local4)
+            //  verioading of namespace can resolve following names
+            Store (\IDX6.TST_.IFE0, Local5)
+            Store (\IDX6.TST_.IFE1, Local6)
+
+            Return (0)
+        }   //  TEST
+    }   //  IDX6
+
+//
+// test IndexOp5.asl
+//
+//  IndexOp5 test
+//      This is just a subset of the many RegionOp/Index Field test cases.
+//      Tests copying string into buffer then performing IndexOp on result.
+//
+    Device (IDX5)
+    {   //  Test device name
+
+        Name (OSFL, 0)  //  0 == Windows 98, 1 == Windows NT
+
+        //  MCTH is a control method to compare two strings. It returns
+        //  zero if the strings mismatch, or 1 if the strings match.
+        //  This exercises the test case of copying a string into a buffer
+        //  and performing an IndexOp on the resulting buffer.
+        Method (MCTH, 2)    //  Control Method to compare two strings
+        {   //  MCTH:   Control Method to compare two strings
+            //  Arg0:       first string to compare
+            //  Arg1:       second string to compare
+            //  Return: zero if strings mismatch, 1 if strings match
+
+            //  check if first string's length is less than second string's length
+            If (LLess (SizeOf (Arg0), SizeOf (Arg1)))
+                {   Return (0)  }
+
+            //  increment length to include NULL termination character
+            Add (SizeOf (Arg0), 1, Local0)  //  Local0 = strlen(Arg0) + 1
+
+            //  create two buffers of size Local0 [strlen(Arg0)+1]
+            Name (BUF0, Buffer (Local0) {})
+            Name (BUF1, Buffer (Local0) {})
+
+            //  copy strings into buffers
+            Store (Arg0, BUF0)
+            Store (Arg1, BUF1)
+
+            //  validate BUF0 and BUF1 are still buffers
+            Store (ObjectType (BUF0), Local1)
+            If (LNotEqual (Local1, 3))  //  Buffer is type 3
+                {   Return (20) }
+            Store (ObjectType (BUF1), Local1)
+            If (LNotEqual (Local1, 3))  //  Buffer is type 3
+                {   Return (21) }
+
+            // Decrement because the Index base below is zero based
+            //  while Local0 length is one based.
+            Decrement (Local0)
+
+            While (Local0)
+            {   //  loop through all BUF0 buffer elements
+                Decrement (Local0)
+
+                //  check if BUF0[n] == BUF1[n]
+                If (LEqual (DerefOf (Index (BUF0, Local0, )),
+                        DerefOf (Index (BUF1, Local0, ))))
+                    {   }   //  this is how the code was really implemented
+                Else
+                    {   Return (Zero)   }
+            }   //  loop through all BUF0 buffer elements
+
+            Return (One)    //  strings / buffers match
+        }   //  MCTH:   Control Method to compare two strings
+
+
+        Method (TEST)
+        {
+            Store ("++++++++ IndexOp5 Test", Debug)
+
+            If (MCTH (\_OS, "Microsoft Windows NT"))
+                {   Store (1, OSFL) }
+
+            If (LNotEqual (OSFL, 1))
+                {   Return (11) }
+
+            Return (0)
+        }   //  TEST
+    }   //  IDX5
+
+//
+// test IndexOp.asl
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+
+        Method (C097)
+            {   Return (1)  }
+
+        Device (PCI2)
+        {   //  Root PCI Bus
+            Name (_HID, EISAID("PNP0A03"))
+            Name (_ADR, 0x00000000)
+            Name (_CRS, Buffer(26)  {"\_SB_.PCI2._CRS..........."})
+            Method (_STA)   {Return (0x0F)}
+
+            Device (ISA)
+            {   //  ISA bridge
+                Name (_ADR, 0x00030000)     //  ISA bus ID
+
+                Device (EC0)
+                {   //  Embedded Controller
+                    Name (_GPE, 0)              //  EC use GPE0
+                    Name (_ADR, 0x0030000)  //  PCI address
+
+                    Method (_STA,0)         //  EC Status
+                        {   Return(0xF) }       //  EC is functioning
+
+                    Name (_CRS, ResourceTemplate()
+                        {
+                            IO (Decode16, 0x62, 0x62, 1, 1)
+                            IO (Decode16, 0x66, 0x66, 1, 1)
+                        }
+                    )
+
+                //  create EC's region and field
+                    OperationRegion (RAM, SystemMemory, 0x400000, 0x100)
+                    Field (RAM, AnyAcc, NoLock, Preserve)
+                    {
+                        //  AC information
+                        ADP,    1,      //  AC Adapter 1:On-line, 0:Off-line
+                        AFLT,   1,      //  AC Adapter Fault  1:Fault  0:Normal
+                        BAT0,   1,      //  BAT0  1:present, 0:not present
+                        ,       1,      //  reserved
+                        ,       28, //  filler to force DWORD alignment
+
+                        //  CMBatt information
+                        BPU0,   32, //  Power Unit
+                        BDC0,   32, //  Designed Capacity
+                        BFC0,   32, //  Last Full Charge Capacity
+                        BTC0,   32, //  Battery Technology
+                        BDV0,   32, //  Design Voltage
+                        BST0,   32, //  Battery State
+                        BPR0,   32, //  Battery Present Rate
+                                        //  (Designed Capacity)x(%)/{(h)x100}
+                        BRC0,   32, //  Battery Remaining Capacity
+                                        //  (Designed Capacity)(%)^100
+                        BPV0,   32, //  Battery Present Voltage
+                        BTP0,   32, //  Trip Point
+                        BCW0,   32, //  Design capacity of Warning
+                        BCL0,   32, //  Design capacity of Low
+                        BCG0,   32, //  capacity granularity 1
+                        BG20,   32, //  capacity granularity 2
+                        BMO0,   32, //  Battery model number field
+                        BIF0,   32, //  OEM Information(00h)
+                        BSN0,   32, //  Battery Serial Number
+                        BTY0,   32, //  Battery Type (e.g., "Li-Ion")
+                        BTY1,   32      //  Battery Type (e.g., "Li-Ion")
+                    }   //  Field
+                }   //  EC0: Embedded Controller
+            }   //  ISA bridge
+        }   //  PCI2 Root PCI Bus
+
+        Device (IDX0)
+        {   //  Test device name
+            Name (_HID, EISAID ("PNP0C0A"))     //  Control Method Battey ID
+            Name (_PCL, Package() {\_SB})
+            Method (_STA)
+            {
+                //  _STA bits 0-3 indicate existence of battery slot
+                //  _STA bit 4 indicates battery (not) present
+                If (\_SB.PCI2.ISA.EC0.BAT0)
+                    {   Return (0x1F)   }   //  Battery present
+                else
+                    {   Return (0x0F)   }   //  Battery not present
+            }   //  _STA
+
+            Method (_BIF)
+            {
+                Name (BUFR, Package(13) {})
+                Store (\_SB.PCI2.ISA.EC0.BPU0, Index (BUFR,0))  //  Power Unit
+                Store (\_SB.PCI2.ISA.EC0.BDC0, Index (BUFR,1))  //  Designed Capacity
+                Store (\_SB.PCI2.ISA.EC0.BFC0, Index (BUFR,2))  //  Last Full Charge Capa.
+                Store (\_SB.PCI2.ISA.EC0.BTC0, Index (BUFR,3))  //  Battery Technology
+                Store (\_SB.PCI2.ISA.EC0.BDV0, Index (BUFR,4))  //  Designed Voltage
+                Store (\_SB.PCI2.ISA.EC0.BCW0, Index (BUFR,5))  //  Designed warning level
+                Store (\_SB.PCI2.ISA.EC0.BCL0, Index (BUFR,6))  //  Designed Low level
+                Store (\_SB.PCI2.ISA.EC0.BCG0, Index (BUFR,7))  //  Capacity granularity 1
+                Store (\_SB.PCI2.ISA.EC0.BG20, Index (BUFR,8))  //  Capacity granularity 2
+
+                Store ("", Index (BUFR,9))                              //  Model Number
+
+                Store ("", Index (BUFR,10))                         //  Serial Number
+
+                Store ("LiOn", Index (BUFR,11))                     //  Battery Type
+
+                Store ("Chicony", Index (BUFR,12))                  //  OEM Information
+
+                Return (BUFR)
+            }   //  _BIF
+
+            Method (_BST)
+            {
+                Name (BUFR, Package(4) {1, 0x100, 0x76543210, 0x180})
+                Return (BUFR)
+            }   //  _BST
+
+            Method (_BTP,1)
+            {
+                Store (arg0, \_SB.PCI2.ISA.EC0.BTP0)    //  Set Battery Trip point
+            }
+
+            Method (TEST)
+            {
+
+                Store ("++++++++ IndexOp Test", Debug)
+
+                //  test storing into uninitialized package elements
+                Name (PBUF, Package(4) {})  //  leave unitialized
+                Store (0x01234567, Index (PBUF,0))
+                Store (0x89ABCDEF, Index (PBUF,1))
+                Store (0xFEDCBA98, Index (PBUF,2))
+                Store (0x76543210, Index (PBUF,3))
+
+                //  verify values stored into uninitialized package elements
+                If (LNotEqual (DerefOf (Index (PBUF,0)), 0x01234567))
+                    {   Return (0x10)   }
+
+                If (LNotEqual (DerefOf (Index (PBUF,1)), 0x89ABCDEF))
+                    {   Return (0x11)   }
+
+                If (LNotEqual (DerefOf (Index (PBUF,2)), 0xFEDCBA98))
+                    {   Return (0x12)   }
+
+                If (LNotEqual (DerefOf (Index (PBUF,3)), 0x76543210))
+                    {   Return (0x13)   }
+
+
+                //  store _BIF package return value into Local0
+                Store (_BIF, Local0)
+
+                //  save Local0 object type value into Local1
+                Store (ObjectType (Local0), Local1)
+
+                //  validate Local0 is a Package
+                If (LNotEqual (Local1, 4))  //  Package type is 4
+                    {   Return (0x21)   }   //  failure
+
+
+                //  test storing into buffer field elements
+                Name (BUFR, Buffer(16)
+                    {   //  initial values
+                        00, 00, 00, 00, 00, 00, 00, 00,
+                        00, 00, 00, 00, 00, 00, 00, 00,
+                    }
+                )   //  BUFR
+                //  test storing into buffer field elements
+                Store (0x01234567, Index (BUFR,0))  //  should only store 0x67
+                Store (0x89ABCDEF, Index (BUFR,4))  //  should only store 0xEF
+                Store (0xFEDCBA98, Index (BUFR,8))  //  should only store 0x98
+                Store (0x76543210, Index (BUFR,12)) //  should only store 0x10
+
+                //  verify storing into buffer field elements
+                If (LNotEqual (DerefOf (Index (BUFR,0)), 0x67))
+                    {   Return (0x30)   }
+
+                If (LNotEqual (DerefOf (Index (BUFR,1)), 0))
+                    {   Return (0x31)   }
+
+                If (LNotEqual (DerefOf (Index (BUFR,4)), 0xEF))
+                    {   Return (0x34)   }
+
+                If (LNotEqual (DerefOf (Index (BUFR,8)), 0x98))
+                    {   Return (0x38)   }
+
+                If (LNotEqual (DerefOf (Index (BUFR,12)), 0x10))
+                    {   Return (0x3C)   }
+
+
+                Return (0)  //  pass
+            }   //  TEST
+        }   //  IDX0
+    }   //  _SB system bus
+
+//
+// test BitIndex.asl
+//
+//  BitIndex test
+//  This is a test case for accessing fields defined as single bits in
+//  memory.  This is done by creating two index fields that overlay the
+//  same DWORD in memory.  One field accesses the DWORD as a DWORD, the
+//  other accesses individual bits of the same DWORD field in memory.
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+        OperationRegion (RAM, SystemMemory, 0x800000, 0x100)
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {   //  Any access
+            TREE,   3,
+            WRD0,   16,
+            WRD1,   16,
+            WRD2,   16,
+            WRD3,   16,
+            WRD4,   16,
+            DWRD,   32, //  DWORD field
+        }
+        Field (RAM, AnyAcc, NoLock, Preserve)
+        {   //  Any access
+            THRE,   3,
+            WD00,   16,
+            WD01,   16,
+            WD02,   16,
+            WD03,   16,
+            WD04,   16,
+            BYT0,   8,  //  Start off with a BYTE
+            BIT0,   1,  //  single-bit field
+            BIT1,   1,  //  single-bit field
+            BIT2,   1,  //  single-bit field
+            BIT3,   1,  //  single-bit field
+            BIT4,   1,  //  single-bit field
+            BIT5,   1,  //  single-bit field
+            BIT6,   1,  //  single-bit field
+            BIT7,   1,  //  single-bit field
+            BIT8,   1,  //  single-bit field
+            BIT9,   1,  //  single-bit field
+            BITA,   1,  //  single-bit field
+            BITB,   1,  //  single-bit field
+            BITC,   1,  //  single-bit field
+            BITD,   1,  //  single-bit field
+            BITE,   1,  //  single-bit field
+            BITF,   1,  //  single-bit field
+            BYTZ,   8,  //  End with a BYTE for a total of 32 bits
+        }
+
+        Device (BITI)
+        {   //  Test device name
+
+            Method (MBIT)   //  Test single bit memory accesses
+            {
+
+                If (LNotEqual (DWRD, 0x00))
+                {
+                    Store (0xFF00, Local0)
+                }
+                Else
+                {
+                    //  Prime Local0 with 0...assume passing condition
+                    Store (0, Local0)
+
+                    //  set memory contents to known values using DWORD field
+                    Store (0x5A5A5A5A, DWRD)
+
+                    //  Given the value programmed into DWRD, only the odd bits
+                    //  of the lower nibble should be set. BIT1, BIT3 should be set.
+                    //  BIT0 and BIT2 should be clear
+
+                    If (BIT0)
+                    {
+                        Or (Local0, 0x01, Local0)
+                    }
+
+                    If (LNot (BIT1))
+                    {
+                        Or (Local0, 0x02, Local0)
+                    }
+
+                    If (BIT2)
+                    {
+                        Or (Local0, 0x04, Local0)
+                    }
+
+                    If (LNot (BIT3))
+                    {
+                        Or (Local0, 0x08, Local0)
+                    }
+
+                    //  Now check the upper nibble.  Only the "even" bits should
+                    //  be set.  BIT4, BIT6.  BIT5 and BIT7 should be clear.
+                    If (LNot (BIT4))
+                    {
+                        Or (Local0, 0x10, Local0)
+                    }
+
+                    If (BIT5)
+                    {
+                        Or (Local0, 0x20, Local0)
+                    }
+
+                    If (LNot (BIT6))
+                    {
+                        Or (Local0, 0x40, Local0)
+                    }
+
+                    If (BIT7)
+                    {
+                        Or (Local0, 0x80, Local0)
+                    }
+                }   //  End Else DWRD zeroed out
+
+                Return (Local0)
+            }   //  MBIT:   Test single bit memory accesses
+
+            Method (TEST)
+            {
+
+                Store ("++++++++ BitIndex Test", Debug)
+
+                //  Zero out DWRD
+                Store (0x00000000, DWRD)
+
+                //  MBIT returns zero if successful
+                //  This may be causing problems -- Return (MBIT)
+                Store (MBIT, Local0)
+
+                Return (Local0)
+            }   //  TEST
+        }   //  BITI
+    }   //  _SB system bus
+
+//
+// test IndexOp3.asl
+//
+//  Additional IndexOp test cases to support ACPICMB (control method battery
+//  test) on Compaq laptops. Test cases include storing a package into
+//  an IndexOp target and validating that changing source and destination
+//  package contents are independent of each other.
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+
+        Name (C174, 13)
+        Name (C175, 8)
+
+        Device (C158)
+        {   //  C158:   AC Adapter device
+            Name (_HID, "ACPI0003") //  AC Adapter device
+            Name (_PCL, Package (1) {\_SB})
+
+            Method (_PSR)
+            {
+                Acquire (\_GL, 0xFFFF)
+                Release (\_GL)
+                And (Local0, 1, Local0) //  Local0 &= 1
+                Return (Local0)
+            }   //  _PSR
+        }   //  C158:   AC Adapter device
+
+        Name (C176, Package (4) {"Primary", "MultiBay", "DockRight", "DockLeft"})
+
+        Name (C177, Package (4) {0x99F5, 0x99F5, 0x995F, 0x995F})
+
+        Name (C178, Package (4)
+        {
+            Package (4) {0, 0, 0x966B, 0x4190},
+            Package (4) {0, 0, 0x966B, 0x4190},
+            Package (4) {0, 0, 0x966B, 0x4190},
+            Package (4) {0, 0, 0x966B, 0x4190}
+        })  //  C178
+
+        Name (C179, Package (4) {0, 0, 0x966B, 0x4190})
+
+        Name (C17A, Package (4)
+        {
+            Package (3) {0, 0, 0},
+            Package (3) {0, 0, 0},
+            Package (3) {0, 0, 0},
+            Package (3) {0, 0, 0}
+        })  //  C17A
+
+        Method (C17B, 1)
+        {   //  C17B:   _BIF implementation
+            Name (C17C, Package (13)
+            {   //  C17C:   _BIF control method return package
+                0,                  //  Power Unit (0 ==> mWh and mW)
+                0x99F5,         //  Design Capacity
+                0x99F5,         //  Last Full Charge Capacity
+                1,                  //  Battery Technology (1 ==> rechargeable)
+                0x3840,         //  Design Voltage
+                0x1280,         //  Design Capacity of Warning
+                0x0AC7,         //  Design Capacity of Low
+                1,                  //  Battery Capacity Granularity 1 (Low -- Warning)
+                1,                  //  Battery Capacity Granularity 2 (Warning -- Full)
+                "2891",         //  Model Number (ASCIIZ)
+                "(-Unknown-)",  //  Serial Number (ASCIIZ)
+                "LIon",         //  Battery Type (ASCIIZ)
+                0                   //  OEM Information (ASCIIZ)
+            })  //  C17C:   _BIF control method return package
+
+            And (Arg0, 7, Local0)                       //  Local0 = Arg0 & 7
+
+            ShiftRight (Local0, 1, Local4)          //  Local4 = Local0 >> 1
+
+            Store (C179, Index (C178, Local4, ))    //  C178->Local4 = C179
+
+            //  verify source and destination packages can be altered independent
+            //  of each other (i.e., changing one's contents does NOT change other's
+            //  contents)
+            Store (0x1234, Index (C179, 2, ))               //  C179[2] = 0x1234
+            Store (DerefOf (Index (C179, 2, )), Local2) //  Local2 = C179[2]
+            if (LNotEqual (Local2, 0x1234))
+                {   Return (0x1234) }
+                                                                        //  Local2 = C178[0,2]
+            Store (DerefOf (Index (DerefOf (Index (C178, 0, )), 2, )), Local2)
+            if (LNotEqual (Local2, 0x966B))
+                {   Return (0x1234) }
+
+            // Restore data to allow iterative execution
+            Store (0x966B, Index (C179, 2, ))               //  C179[2] = 0x966B
+
+                                                                        //  C178[0,3] = 0x5678
+            Store (0x5678, Index (DerefOf (Index (C178, 0, )), 3, ))
+                                                                        //  Local2 = C178[0,3]
+            Store (DerefOf (Index (DerefOf (Index (C178, 0, )), 3, )), Local2)
+            if (LNotEqual (Local2, 0x5678))
+                {   Return (0x5678) }
+
+            Store (DerefOf (Index (C179, 3, )), Local2) //  Local2 = C179[3]
+            if (LNotEqual (Local2, 0x4190))
+                {   Return (0x5678) }
+
+            // Restore data to allow iterative execution
+            Store (0x4190, Index (DerefOf (Index (C178, 0, )), 3, ))    //  C179[2] = 0x4190
+
+            Return (C17C)
+        }   //  C17B:   _BIF implementation
+
+        Device (C154)
+        {   //  C154:   Battery 0
+            Name (_HID, "*PNP0C0A")     //  Control Method Battey ID
+            Name (_UID, 0)                  //  first instance
+
+            Method (_BIF)
+            {   //  _BIF
+                Return (C17B (48))
+            }   //  _BIF
+        }   //  C154:   Battery 0
+
+        Device (IDX3)
+        {
+            Method (LCLB)
+            {   //  LCLB control method: test Index(Local#) where Local# is buffer
+                //  Local0 is index counter
+                //  Local1 is buffer
+                //  Local2 receives BUFR[Local0] via Deref(Index(Local1...))
+                //  Local3 is Local1 or Local2 object type
+                //  Local4 is return error code
+
+                Name (BUFR, Buffer ()   {0, 1, 2, 3, 4, 5, 6, 7, 8, 9})
+
+                //  save PKG into Local1
+                Store (BUFR, Local1)
+
+                //  save Local2 object type value into Local3
+                Store (ObjectType (Local1), Local3)
+
+                //  validate Local1 is a Buffer
+                If (LNotEqual (Local3, 3))      //  Buffer type is 3
+                    {   Return (0x9F)   }
+
+
+                Store (0, Local0)
+                While (LLess (Local0, 5))
+                {   //  While (Local0 < 5)
+                    //  Local2 = Local1[Local0]
+                    Store (DerefOf (Index (Local1, Local0, )), Local2)
+
+                    //  save Local2 object type value into Local3
+                    Store (ObjectType (Local2), Local3)
+
+                    //  validate Local2 is a Number
+                    If (LNotEqual (Local3, 1))      //  Number type is 1
+                        {   Return (0x9E)   }
+
+                    //  validate Local1[Local0] value == Local0
+                    If (LNotEqual (Local0, Local2))
+                    {   //  Local0 != Local2 == PKG[Local0]
+                        //  Local4 = 0x90 + loop index (Local0)
+                        Add (0x90, Local0, Local4)
+
+                        //  return 0x90 + loop index
+                        Return (Local4)
+                    }
+
+                    Increment (Local0)
+                }   //  While (Local0 < 5)
+
+                Store ("DerefOf(Index(LocalBuffer,,)) PASS", Debug)
+
+                Return (0)  //  Pass
+            }   //  LCLB control method: test Index(Local#) where Local# is buffer
+
+            Method (LCLP)
+            {   //  LCLP control method: test Index(Local#) where Local# is package
+                //  Local0 is index counter
+                //  Local1 is package
+                //  Local2 receives PKG[Local0] via Deref(Index(Local1...))
+                //  Local3 is Local1 or Local2 object type
+                //  Local4 is return error code
+
+                Name (PKG, Package ()   {0, 1, 2, 3, 4, 5, 6, 7, 8, 9})
+
+                //  save PKG into Local1
+                Store (PKG, Local1)
+
+                //  save Local2 object type value into Local3
+                Store (ObjectType (Local1), Local3)
+
+                //  validate Local1 is a Package
+                If (LNotEqual (Local3, 4))      //  Package type is 4
+                    {   Return (0x8F)   }
+
+
+                Store (0, Local0)
+                While (LLess (Local0, 5))
+                {   //  While (Local0 < 5)
+                    //  Local2 = Local1[Local0]
+                    Store (DerefOf (Index (Local1, Local0, )), Local2)
+
+                    //  save Local2 object type value into Local3
+                    Store (ObjectType (Local2), Local3)
+
+                    //  validate Local2 is a Number
+                    If (LNotEqual (Local3, 1))      //  Number type is 1
+                        {   Return (0x8E)   }
+
+                    //  validate Local1[Local0] value == Local0
+                    If (LNotEqual (Local0, Local2))
+                    {   //  Local0 != Local2 == PKG[Local0]
+                        //  Local4 = 0x80 + loop index (Local0)
+                        Add (0x80, Local0, Local4)
+
+                        //  return 0x80 + loop index
+                        Return (Local4)
+                    }
+
+                    Increment (Local0)
+                }   //  While (Local0 < 5)
+
+                Store ("DerefOf(Index(LocalPackage,,)) PASS", Debug)
+
+                Return (0)  //  Pass
+            }   //  LCLP control method: test Index(Local#) where Local# is package
+
+            Method (TEST)
+            {
+
+                Store ("++++++++ IndexOp3 Test", Debug)
+
+                //  store _BIF package return value into Local0
+                Store (\_SB.C154._BIF, Local0)
+
+                //  save Local0 object type value into Local1
+                Store (ObjectType (Local0), Local1)
+
+                //  validate Local0 is a Package
+                If (LNotEqual (Local1, 4))      //  Package type is 4
+                {   //  failure: did not return a Package (type 4)
+                    //  if Local0 is a Number, it contains an error code
+                    If (LEqual (Local1, 1))     //  Number type is 1
+                        {   Return (Local0) }   //  return Local0 error code
+                    Else                                //  Local0 is not a Number
+                        {   Return (1)  }           //  return default error code
+                }   //  failure: did not return a Package (type 4)
+
+                //  save LCLB control method return value into Local2
+                Store (LCLB, Local2)
+                If (LNotEqual (Local2, 0))
+                    {   Return (Local2) }   //  return failure code
+
+                //  save LCLP control method return value into Local2
+                Store (LCLP, Local2)
+                If (LNotEqual (Local2, 0))
+                    {   Return (Local2) }   //  return failure code
+
+                Return (0)  //  Pass
+            }   //  TEST
+        }   //  IDX3:   Test device name
+    }   //  _SB system bus
+
+//
+// MTL developed test to exercise Indexes into buffers
+//
+    Device(IDX7)
+    {
+
+        Name (PKG4, Package() {
+                0x2,
+                "A short string",
+                Buffer() {0xA, 0xB, 0xC, 0xD},
+                0x1234,
+                Package() {IDX7, 0x3}
+                })
+
+        //
+        // Generic Test method
+        //
+        // This test returns 0xE (14) - ObjectType = Buffer Field
+        Method(TST1)
+        {
+            Name (DEST, Buffer ()                           //  62 characters plus NULL
+                {"Destination buffer that is longer than the short source buffer"})
+
+            //  verify object type returned by Index(Buffer,Element,)
+            Store (Index (DEST, 2, ), Local1)
+            Store (ObjectType (Local1), Local2)
+            If (LEqual(Local2, 14))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x1)
+            }
+
+        }
+
+        Method(TST2)
+        {
+            Name (BUF0, Buffer() {0x1, 0x2, 0x3, 0x4, 0x5})
+            Store(0x55, Index(BUF0, 2))
+            Store(DerefOf(Index(BUF0, 2)), Local0)
+            If (LEqual(Local0, 0x55))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x2)
+            }
+
+
+        }
+
+        Method(TST3)
+        {
+            Name (BUF1, Buffer() {0x1, 0x2, 0x3, 0x4, 0x5})
+            Store(Index(BUF1, 1), Local0)
+            Store(DerefOf(Local0), Local1)
+            If (LEqual(Local1, 0x2))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x3)
+            }
+
+        }
+
+        Method(TST4)
+        {
+            // Index (PKG4, 0) is a Number
+            Store (Index (PKG4, 0), Local0)
+            Store (ObjectType(Local0), Local1)
+            If (LEqual(Local1, 0x1))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x4)
+            }
+
+        }
+
+        Method(TST5)
+        {
+            // Index (PKG4, 1) is a String
+            Store (Index (PKG4, 1), Local0)
+            Store (ObjectType(Local0), Local1)
+            If (LEqual(Local1, 0x2))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x5)
+            }
+
+        }
+
+        Method(TST6)
+        {
+            // Index (PKG4, 2) is a Buffer
+            Store (Index (PKG4, 2), Local0)
+            Store (ObjectType(Local0), Local1)
+            If (LEqual(Local1, 0x3))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x6)
+            }
+
+        }
+
+        Method(TST7)
+        {
+            // Index (PKG4, 3) is a Number
+            Store (Index (PKG4, 3), Local0)
+            Store (ObjectType(Local0), Local1)
+            If (LEqual(Local1, 0x1))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x7)
+            }
+
+        }
+
+        Method(TST8)
+        {
+            // Index (PKG4, 4) is a Package
+            Store (Index (PKG4, 4), Local0)
+            Store (ObjectType(Local0), Local1)
+            If (LEqual(Local1, 0x4))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x8)
+            }
+
+        }
+
+        Method(TST9)
+        {
+            // DerefOf (Index (PKG4, 0)) is a Number
+            Store (DerefOf (Index (PKG4, 0)), Local0)
+            If (LEqual(Local0, 0x2))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0x9)
+            }
+
+        }
+
+        Method(TSTA)
+        {
+            // DerefOf (Index (PKG4, 1)) is a String
+            Store (DerefOf (Index (PKG4, 1)), Local0)
+            Store (SizeOf(Local0), Local1)
+            If (LEqual(Local1, 0xE))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0xA)
+            }
+
+        }
+
+        Method(TSTB)
+        {
+            // DerefOf (Index (PKG4, 2)) is a Buffer
+            Store (DerefOf (Index (PKG4, 2)), Local0)
+            Store (SizeOf(Local0), Local1)
+            If (LEqual(Local1, 0x4))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0xB)
+            }
+
+        }
+
+        Method(TSTC)
+        {
+            // DerefOf (Index (PKG4, 3)) is a Number
+            Store (DerefOf (Index (PKG4, 3)), Local0)
+            If (LEqual(Local0, 0x1234))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0xC)
+            }
+
+        }
+
+        Method(TSTD)
+        {
+            // DerefOf (Index (PKG4, 4)) is a Package
+            Store (DerefOf (Index (PKG4, 4)), Local0)
+            Store (SizeOf(Local0), Local1)
+            If (LEqual(Local1, 0x2))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0xD)
+            }
+
+        }
+
+        Method(TSTE)
+        {
+            // DerefOf (Index (PKG4, 2)) is a Buffer
+            Store (DerefOf (Index (PKG4, 2)), Local0)
+            // DerefOf (Index (Local0, 1)) is a Number
+            Store (DerefOf (Index (Local0, 1)), Local1)
+            If (LEqual(Local1, 0xB))
+            {
+                Return(0)
+            }
+            Else
+            {
+                Return(0xE)
+            }
+
+        }
+
+        Method (TSTF)
+        {
+            Name (SRCB, Buffer (12) {}) //  12 characters
+            Store ("Short Buffer", SRCB)
+
+            Name (DEST, Buffer ()                       //  62 characters plus NULL
+                {"Destination buffer that is longer than the short source buffer"})
+
+            //  overwrite DEST contents, starting at buffer position 2
+            Store (SRCB, Index (DEST, 2))
+
+            //
+            //  The DEST buffer element should be replaced with the last element of
+            //      the SRCB element (i.e. 's'->'r')
+            Store (DerefOf (Index (DEST, 2)), Local0)
+
+            If (LNotEqual (Local0, 0x72))       //  'r'
+            {
+                //  DEST element does not match the value from SRCB
+                Return(Or(Local0, 0x1000))
+            }
+
+            Return(0)
+        }
+
+        Method (TSTG)
+        {
+
+            Name (SRCB, Buffer (12) {}) //  12 characters
+            Store ("Short Buffer", SRCB)
+
+            Name (DEST, Buffer ()                       //  62 characters plus NULL
+                {"Destination buffer that is longer than the short source buffer"})
+
+            //  overwrite DEST contents, starting at buffer position 2
+            Store (SRCB, Index (DEST, 2))
+
+            //
+            // The next element of DEST should be unchanged
+            //
+            Store (DerefOf (Index (DEST, 3)), Local0)
+
+            If (LNotEqual (Local0, 0x74))       //  't'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2000))
+            }
+
+            //
+            // The next element of DEST should be unchanged
+            //
+            Store (DerefOf (Index (DEST, 4)), Local0)
+
+            If (LNotEqual (Local0, 0x69))       //  'i'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2100))
+            }
+
+            //
+            // The next element of DEST should be unchanged
+            //
+            Store (DerefOf (Index (DEST, 5)), Local0)
+
+            If (LNotEqual (Local0, 0x6E))       //  'n'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2200))
+            }
+
+            //
+            // The next element of DEST should be unchanged
+            //
+            Store (DerefOf (Index (DEST, 6)), Local0)
+
+            If (LNotEqual (Local0, 0x61))       //  'a'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2300))
+            }
+
+            //
+            // The next element of DEST should be unchanged
+            //
+            Store (DerefOf (Index (DEST, 7)), Local0)
+
+            If (LNotEqual (Local0, 0x74))       //  't'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2400))
+            }
+
+            //
+            //  Verify DEST elements beyond end of SRCB buffer copy
+            //  have not been changed
+            Store (DerefOf (Index (DEST, 14)), Local0)
+
+            If (LNotEqual (Local0, 0x66))       // 'f'
+            {
+                //  DEST has been changed
+                Return(Or(Local0, 0x2400))
+            }
+
+            Return(0)
+        }
+
+        //
+        // This test shows that MS ACPI.SYS stores only the lower 8-bits of a 32-bit
+        //  number into the index'ed buffer
+        //
+        Method (TSTH)
+        {
+            // Create a Destination Buffer
+            Name (DBUF, Buffer () {"abcdefghijklmnopqrstuvwxyz"})
+
+            // Store a number > UINT8 into an index of the buffer
+            Store (0x12345678, Index(DBUF, 2))
+
+            // Check the results
+            Store (DerefOf (Index (DBUF, 2)), Local0)
+            If (LNotEqual (Local0, 0x78))   // 0x78
+            {
+                Return(Or(Local0, 0x3000))
+            }
+
+            Store (DerefOf (Index (DBUF, 3)), Local0)
+            If (LNotEqual (Local0, 0x64))   // 'd'
+            {
+                Return(Or(Local0, 0x3100))
+            }
+
+            Store (DerefOf (Index (DBUF, 4)), Local0)
+            If (LNotEqual (Local0, 0x65))   // 'e'
+            {
+                Return(Or(Local0, 0x3200))
+            }
+
+            Store (DerefOf (Index (DBUF, 5)), Local0)
+            If (LNotEqual (Local0, 0x66))   // 'f'
+            {
+                Return(Or(Local0, 0x3300))
+            }
+
+            Return(0)
+        }
+
+        Method (TSTI)
+        {
+            // Create a Destination Buffer
+            Name (DBUF, Buffer () {"abcdefghijklmnopqrstuvwxyz"})
+
+            // Store a String into an index of the buffer
+            Store ("ABCDEFGH", Index(DBUF, 2))
+
+            // Check the results
+            Store (DerefOf (Index (DBUF, 2)), Local0)
+            If (LNotEqual (Local0, 0x48))   // 'H'
+            {
+                Return(Or(Local0, 0x4000))
+            }
+
+            Store (DerefOf (Index (DBUF, 3)), Local0)
+            If (LNotEqual (Local0, 0x64))   // 'd'
+            {
+                Return(Or(Local0, 0x4100))
+            }
+
+            Store (DerefOf (Index (DBUF, 4)), Local0)
+            If (LNotEqual (Local0, 0x65))   // 'e'
+            {
+                Return(Or(Local0, 0x4200))
+            }
+
+            Store (DerefOf (Index (DBUF, 5)), Local0)
+            If (LNotEqual (Local0, 0x66))   // 'f'
+            {
+                Return(Or(Local0, 0x4300))
+            }
+
+            Return(0)
+        }
+
+        Method(TSTJ)
+        {
+            // Create a Destination Buffer
+            Name (DBUF, Buffer () {"abcdefghijklmnopqrstuvwxyz"})
+
+            // Store a number > UINT8 into an index of the buffer
+            Store (0x1234, Index(DBUF, 2))
+
+            // Check the results
+            Store (DerefOf (Index (DBUF, 2)), Local0)
+            If (LNotEqual (Local0, 0x34))   // 0x34
+            {
+                Return(Or(Local0, 0x3000))
+            }
+
+            Store (DerefOf (Index (DBUF, 3)), Local0)
+            If (LNotEqual (Local0, 0x64))   // 'd'
+            {
+                Return(Or(Local0, 0x3100))
+            }
+
+            Store (DerefOf (Index (DBUF, 4)), Local0)
+            If (LNotEqual (Local0, 0x65))   // 'e'
+            {
+                Return(Or(Local0, 0x3200))
+            }
+
+            Store (DerefOf (Index (DBUF, 5)), Local0)
+            If (LNotEqual (Local0, 0x66))   // 'f'
+            {
+                Return(Or(Local0, 0x3300))
+            }
+
+            Return(0)
+        }
+
+        Method(TSTK)
+        {
+            // Create a Destination Buffer
+            Name (DBUF, Buffer () {"abcdefghijklmnopqrstuvwxyz"})
+
+            // Store a number > UINT8 into an index of the buffer
+            Store (0x123456, Index(DBUF, 2))
+
+            // Check the results
+            Store (DerefOf (Index (DBUF, 2)), Local0)
+            If (LNotEqual (Local0, 0x56))   // 0x56
+            {
+                Return(Or(Local0, 0x3000))
+            }
+
+            Store (DerefOf (Index (DBUF, 3)), Local0)
+            If (LNotEqual (Local0, 0x64))   // 'd'
+            {
+                Return(Or(Local0, 0x3100))
+            }
+
+            Store (DerefOf (Index (DBUF, 4)), Local0)
+            If (LNotEqual (Local0, 0x65))   // 'e'
+            {
+                Return(Or(Local0, 0x3200))
+            }
+
+            Store (DerefOf (Index (DBUF, 5)), Local0)
+            If (LNotEqual (Local0, 0x66))   // 'f'
+            {
+                Return(Or(Local0, 0x3300))
+            }
+
+            Return(0)
+        }
+
+        Method(TSTL)
+        {
+            // Create a Destination Buffer
+            Name (DBUF, Buffer () {"abcdefghijklmnopqrstuvwxyz"})
+
+            // Store a number > UINT8 into an index of the buffer
+            Store (0x12, Index(DBUF, 2))
+
+            // Check the results
+            Store (DerefOf (Index (DBUF, 2)), Local0)
+            If (LNotEqual (Local0, 0x12))   // 0x12
+            {
+                Return(Or(Local0, 0x3000))
+            }
+
+            Store (DerefOf (Index (DBUF, 3)), Local0)
+            If (LNotEqual (Local0, 0x64))   // 'd'
+            {
+                Return(Or(Local0, 0x3100))
+            }
+
+            Store (DerefOf (Index (DBUF, 4)), Local0)
+            If (LNotEqual (Local0, 0x65))   // 'e'
+            {
+                Return(Or(Local0, 0x3200))
+            }
+
+            Store (DerefOf (Index (DBUF, 5)), Local0)
+            If (LNotEqual (Local0, 0x66))   // 'f'
+            {
+                Return(Or(Local0, 0x3300))
+            }
+
+            Return(0)
+        }
+
+        Method(TEST)
+        {
+            Store ("++++++++ IndexOp7 Test", Debug)
+
+            Store(TST1(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST2(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST3(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST4(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST5(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST6(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST7(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST8(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TST9(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTA(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTB(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTC(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTD(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTE(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+    /* No longer ACPI compliant */
+    /*
+            Store(TSTF(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+    */
+
+            Store(TSTG(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTH(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+    /* No longer ACPI compliant */
+    /*
+            Store(TSTI(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+    */
+            Store(TSTJ(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTK(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Store(TSTL(), Local0)
+            if (LGreater (Local0, 0))
+            {
+                Return(Local0)
+            }
+
+            Return(Local0)
+
+        }
+
+    } // Device(IDX7)
+
+//
+// test MatchOp.asl
+//
+//  MatchOp test cases that utilize nested DerefOf(Index(...)) to validate
+//  MatchOp, DerefOfOp, and IndexOp of nested packages.
+//
+    Device (MTCH)
+    {
+
+        Method (TEST)
+        {
+            Store ("++++++++ MatchOp Test", Debug)
+
+            Name (TIM0, Package ()
+                {
+                    Package ()  {0x78, 0xB4, 0xF0, 0x0384},
+                    Package ()  {0x23, 0x21, 0x10, 0},
+                    Package ()  {0x0B, 9, 4, 0},
+                    Package ()  {0x70, 0x49, 0x36, 0x27, 0x19},
+                    Package ()  {0, 1, 2, 1, 2},
+                    Package ()  {0, 0, 0, 1, 1},
+                    Package ()  {4, 3, 2, 0},
+                    Package ()  {2, 1, 0, 0}
+                })  //  TIM0
+
+            Name (TMD0, Buffer (20) {0xFF, 0xFF, 0xFF, 0xFF })
+            CreateDWordField (TMD0, 0, PIO0)    //  0xFFFFFFFF
+            CreateDWordField (TMD0, 4, DMA0)
+            CreateDWordField (TMD0, 8, PIO1)
+            CreateDWordField (TMD0, 12, DMA1)
+            CreateDWordField (TMD0, 16, CHNF)
+
+
+            //  validate PIO0 value
+            Store (PIO0, Local3)
+
+            //  save Local3 object type value into Local2
+            Store (ObjectType (Local3), Local2)
+
+            //  validate Local3 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+                {   Return (2)  }   //  failure
+
+            //  validate Local3 Number value
+            If (LNotEqual (Local3, 0xFFFFFFFF)) //  Number value 0xFFFFFFFF
+                {   Return (3)  }   //  failure
+
+            Store ("DWordField PASS", Debug)
+
+
+            Store (0, Local5)
+            Store (Match (DerefOf (Index (TIM0, 1, )), MLE, Local5, MTR, 0, 0), Local6)
+
+            //  save Local6 object type value into Local2
+            Store (ObjectType (Local6), Local2)
+
+            //  validate Local6 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+                {   Return (4)  }   //  failure
+
+            Store ("Match(DerefOf(Index(TIM0,1)),... PASS", Debug)
+
+
+            //  validate following produces a nested package to validate
+            //  that MatchOp did not corrupt SearchPackage (TIM0)
+            Store (DerefOf (Index (TIM0, 1, )), Local4)
+
+            //  save Local4 object type value into Local2
+            Store (ObjectType (Local4), Local2)
+
+            //  validate Local4 is a Package
+            If (LNotEqual (Local2, 4))  //  Package type is 4
+                {   Return (5)  }   //  failure
+
+            Store ("DerefOf(Index(TIM0,1)),... PASS", Debug)
+
+
+            And (Match (DerefOf (Index (TIM0, 0, )), MGE, PIO0, MTR, 0, 0), 3, Local0)
+
+            //  save Local0 object type value into Local2
+            Store (ObjectType (Local0), Local2)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+                {   Return (6)  }   //  failure
+
+            //  validate Local0 Number value
+            If (LNotEqual (Local0, 3))  //  Number value 3
+                {   Return (7)  }   //  failure
+
+            Store ("And(Match(DerefOf(Index(TIM0,0)),... PASS", Debug)
+
+
+            //  again, validate following produces a nested package
+            Store (DerefOf (Index (TIM0, 1, )), Local4)
+
+            //  save Local4 object type value into Local2
+            Store (ObjectType (Local4), Local2)
+
+            //  validate Local4 is a Package
+            If (LNotEqual (Local2, 4))  //  Package type is 4
+                {   Return (8)  }   //  failure
+
+            Store ("DerefOf(Index(TIM0,1)),... PASS again", Debug)
+
+
+            //  again, validate following produces a nested package
+            Store (DerefOf (Index (TIM0, 1, )), Local4)
+
+            //  save Local4 object type value into Local2
+            Store (ObjectType (Local4), Local2)
+
+            //  validate Local4 is a Package
+            If (LNotEqual (Local2, 4))  //  Package type is 4
+                {   Return (9)  }   //  failure
+
+            Store ("DerefOf(Index(TIM0,1)),... PASS again", Debug)
+
+
+            //  test nested DerefOf(Index) operators
+            Store (DerefOf (Index (DerefOf (Index (TIM0, 1, )), Local0, )), Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+                {   Return (10) }   //  failure
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+                {   Return (11) }   //  failure
+
+            Store ("DerefOf(Index(DerefOf(Index(TIM0,1)),... PASS", Debug)
+
+
+            //  again, validate following produces a nested package
+            Store (DerefOf (Index (TIM0, 1, )), Local4)
+
+            //  save Local4 object type value into Local2
+            Store (ObjectType (Local4), Local2)
+
+            //  validate Local4 is a Package
+            If (LNotEqual (Local2, 4))  //  Package type is 4
+                {   Return (12) }   //  failure
+
+            Store ("DerefOf(Index(TIM0,1)),... PASS again", Debug)
+
+
+            //  retest nested DerefOf(Index) operators
+            Store (DerefOf (Index (DerefOf (Index (TIM0, 1, )), Local0, )), Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+                {   Return (13) }   //  failure
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+                {   Return (14) }   //  failure
+
+            Store ("DerefOf(Index(DerefOf(Index(TIM0,1)),... PASS again", Debug)
+
+
+            //  again, validate following produces a nested package
+            Store (DerefOf (Index (TIM0, 1, )), Local4)
+
+            //  save Local4 object type value into Local2
+            Store (ObjectType (Local4), Local2)
+
+            //  validate Local4 is a Package
+            If (LNotEqual (Local2, 4))  //  Package type is 4
+                {   Return (15) }   //  failure
+
+            Store ("DerefOf(Index(TIM0,1)),... PASS again", Debug)
+
+
+            Return (0)  //  pass
+        }   //  TEST
+    }   // MTCH
+
+//
+// test WhileBrk.asl
+//
+//  This code tests the Break term and While term
+//
+//  Syntax of Break term
+//      BreakTerm := Break
+//  The break operation causes the current package execution to complete.
+//
+//  Syntax of While Term
+//      WhileTerm   := While(
+//          Predicate   //TermArg=>Integer
+//      ) {TermList}
+//  Predicate is evaluated as an integer.
+//  If the integer is non-zero, the list of terms in TermList is executed.
+//  The operation repeats until the Predicate evaluates to zero.
+//
+// MTL NOTE: This test has been modified to reflect ACPI 2.0 break
+// NOTE: This test, when run under the MS ACPI.SYS grinds the system to
+//  a halt.
+//
+    Device (WHLB)
+    {
+        Name (CNT0, 0)
+        Name (CNT1, 0)
+
+        Method (TEST)
+        {
+            //  Check Break statement nested in If nested in While nested in
+            //  While only exits inner-most While loop
+            Store (0, CNT0)
+
+            While (LLess (CNT0, 4))
+            {
+                Store (0, CNT1)
+                While (LLess (CNT1, 10))
+                {
+                    if (LEqual (CNT1, 1))
+                    {
+                        Break       //  exit encompassing loop
+                    }
+
+                    Increment (CNT1)
+                }
+
+                If (LNotEqual (CNT1, 1))
+                {
+                    //  failure
+                    Return (7)
+                }
+
+                Increment (CNT0)
+            }
+
+            //  Verify Break only exited inner-most While loop
+
+            If (LNotEqual (CNT0, 4))
+            {
+                //  failure
+                Return (8)
+            }
+
+            Store ("While/While/If/Break PASS", Debug)
+
+            Store ("++++++++ WhileBrk Test", Debug)
+
+            //  Check Break statement nested in While
+            Store (0, CNT0)
+
+            While (LLess (CNT0, 10))
+            {
+                Break       //  exit encompassing package
+                Increment (CNT0)
+            }
+
+            If (LNotEqual (CNT0, 0))    //  instruction after Break executed
+            {
+                Return (4)
+            }
+
+
+            Store (0, CNT0)
+
+            //  Test While Term
+            While (LLess (CNT0, 10))
+            {
+                Increment (CNT0)
+            }
+
+            //  Check if the while loop was executed until the condition is satisfied.
+            If (LNotEqual (CNT0, 10))
+            {
+                Return (1)
+            }
+
+
+            //  While loop in a reverse order
+            While (LGreater (CNT0, 0))
+            {
+                Decrement (CNT0)
+            }
+
+            //  Check if the while loop was executed until the condition is satisfied.
+            If (LNotEqual (CNT0, 0))
+            {
+                Return (2)
+            }
+
+
+            Store ("While/Break PASS", Debug)
+
+
+            //  Check Break statement nested in If nested in While
+            Store (0, CNT0)
+
+            While (LLess (CNT0, 10))
+            {
+                if (LEqual (CNT0, 5))
+                {
+                    Break       //  exit encompassing Package (If)
+
+                    //  if we execute the next instruction,
+                    //  Break did not exit the loop
+                    Store (20, CNT0)    //  exit While loop with value larger
+                                            //  than above
+                }
+
+                Increment (CNT0)    //  check if Break exited both If and While
+            }   //  While
+
+            If (LGreater (CNT0, 19))
+            {   //  instruction after Break inside IfOp executed
+                Return (5)
+            }
+
+            //
+            // Break will exit out of the while loop, therefore
+            //  the CNT0 counter should still Increment until 5
+            //
+            If (LNotEqual (CNT0, 5))
+            {   //  instruction after Break inside WhileOp executed
+                Return (6)
+            }
+            Store ("While/If/Break PASS", Debug)
+
+
+            //  All the conditions passed
+            Return (0)
+        }   //  TEST
+    }   //  WHLB
+
+
+//
+// test IndexOp2.asl
+//
+//  Additional IndexOp test cases to support ACPICMB (control method battery
+//  test) on Toshiba Portege 7020CT. Test cases include appropriate bit
+//  shifting of Field elements and reading Field elements greater than 64 bits.
+//
+// MTL NOTE: This test has been modified slightly from the original test
+//  to take into account ACPI specification limitations.
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+
+        Device (MEM)
+        {   //  MEM
+            Name (_HID, 0x010CD041)
+            Name (_STA, 0x0F)
+
+            OperationRegion (SMEM, SystemMemory, 0x800000, 0x100)
+            Field (SMEM, AnyAcc, NoLock, Preserve)
+            {   //  Field:  SMEM overlay using 32-bit field elements
+                SMD0,   32, //  32-bits
+                SMD1,   32,     //  32-bits
+                SMD2,   32,     //  32-bits
+                SMD3,   32  //  32-bits
+            }   //  Field:  SMEM overlay using 32-bit field elements
+            Field (SMEM, AnyAcc, NoLock, Preserve)
+            {   //  Field:  SMEM overlay using greater than 32-bit field elements
+                SME0,   69, //  larger than an integer (32 or 64)
+                SME1,   97  //  larger than an integer
+            }   //  Field:  SMEM overlay using greater than 32-bit field elements
+
+            OperationRegion (SRAM, SystemMemory, 0x100B0000, 0xF000)
+            Field (SRAM, AnyAcc, NoLock, Preserve)
+            {   //  Field:  SRAM overlay
+                    ,   0x34000,    //  skip
+                IEAX,   0x20,
+                IEBX,   0x20,
+                IECX,   0x20,
+                IEDX,   0x20,
+                IESI,   0x20,
+                IEDI,   0x20,
+                IEBP,   0x20,
+                    ,   0x20,
+                OEAX,   0x20,
+                OEBX,   0x20,
+                OECX,   0x20,
+                OEDX,   0x20,
+                OESI,   0x20,
+                OEDI,   0x20,
+                OEBP,   0x20,
+                    ,   0x618,  //  skip
+                ACST,   1,
+                BES1,   1,
+                BES2,   1,
+                    ,   5,          //  skip
+                BMN1,   0x68,
+                BSN1,   0x58,
+                BTP1,   0x48,
+                BPU1,   0x20,
+                BDC1,   0x20,
+                BLF1,   0x20,
+                BTC1,   0x20,
+                BDV1,   0x20,
+                BST1,   0x20,
+                BPR1,   0x20,
+                BRC1,   0x20,
+                BPV1,   0x20,
+                    ,   0x20,
+                BCW1,   0x20,
+                BCL1,   0x20,
+                BG11,   0x20,
+                BG21,   0x20,
+                BOI1,   0x20,
+                    ,   0x530,  //  skip
+                BMN2,   0x68,
+                BSN2,   0x58,
+                BTP2,   0x48,
+                BPU2,   0x20,
+                BDC2,   0x20,
+                BLF2,   0x20,
+                BTC2,   0x20,
+                BDV2,   0x20,
+                BST2,   0x20,
+                BPR2,   0x20,
+                BRC2,   0x20,
+                BPV2,   0x20,
+                    ,   0x20,
+                BCW2,   0x20,
+                BCL2,   0x20,
+                BG12,   0x20,
+                BG22,   0x20,
+                BOI2,   0x20,
+                    ,   0x518,  //  skip
+                AC01,   0x10,
+                AC11,   0x10,
+                PSV1,   0x10,
+                CRT1,   0x10,
+                TMP1,   0x10,
+                AST1,   0x10,
+                AC21,   0x10,
+                AC31,   0x10,
+                AC02,   0x10,
+                AC12,   0x10,
+                PSV2,   0x10,
+                CRT2,   0x10,
+                TMP2,   0x10,
+                AST2,   0x10,
+                AC22,   0x10,
+                AC32,   0x10,
+                AC03,   0x10,
+                AC13,   0x10,
+                PSV3,   0x10,
+                CRT3,   0x10,
+                TMP3,   0x10,
+                AST3,   0x10,
+                AC23,   0x10,
+                AC33,   0x10,
+                    ,   0x80,       //  skip
+                TMPF,   0x10,
+                    ,   0x570,  //  skip
+                FANH,   1,
+                FANL,   7,
+                TF11,   1,
+                TF21,   1,
+                TF31,   1,
+                    ,   1,
+                TF10,   1,
+                TF20,   1,
+                TF30,   1,
+                    ,   1,
+                TP11,   1,
+                TP21,   1,
+                TP31,   1,
+                    ,   0x6D,   //  109
+                GP50,   1,
+                GP51,   1,
+                GP52,   1,
+                GP53,   1,
+                    ,   4,
+                GP60,   1,
+                GP61,   1,
+                GP62,   1,
+                GP63,   1,
+                GP64,   1,
+                GP65,   1,
+                GP66,   1,
+                    ,   1,
+                GP70,   1,
+                GP71,   1,
+                GP72,   1,
+                GP73,   1,
+                GP74,   1,
+                GP75,   1,
+                GP76,   1,
+                    ,   1,
+                WED0,   1,
+                WED1,   1,
+                WED2,   1,
+                WED3,   1,
+                WED4,   1,
+                    ,   3,
+                SBL0,   1,
+                SBL1,   1,
+                SBL2,   1,
+                SBL3,   1,
+                    ,   4,
+                LIDS,   1,
+                VALF,   1,
+                    ,   2,
+                DCKI,   1,
+                DCKF,   1,
+                BT1F,   1,
+                BT2F,   1,
+                    ,   0x7D0,  //  skip
+                HKCD,   8,
+                    ,   8,
+                DLID,   0x20,
+                DSRN,   0x20,
+                    ,   0x20,
+                BDID,   0x20,
+                DSPW,   1,
+                VGAF,   1,
+                VWE0,   1,
+                VWE1,   1,
+                PPSC,   1,
+                SPSC,   1,
+                EWLD,   1,
+                EWPS,   1,
+                    ,   0x1768, //  skip
+                PRES,   0x8000
+            }   //  Field:  SRAM overlay
+        }   //  MEM
+
+        Device (BAT1)
+        {   //  BAT1
+            Name (_HID, EISAID ("PNP0C0A"))     //  Control Method Battey ID
+            Name (_UID, 1)
+            Name (_PCL, Package (1) {\_SB})
+
+            Method (_STA)
+            {   //  _STA
+                If (\_SB.MEM.BES1)
+                    {   Return (0x1F)   }   //  battery present
+                Else
+                    {   Return (0x0F)   }   //  battery not present
+            }   //  _STA
+
+            Method (_BIF)
+            {   //  _BIF
+                Name (BUFR, Package (13)    {})
+
+                Store (\_SB.MEM.BPU1, Index (BUFR, 0))
+                Store (\_SB.MEM.BDC1, Index (BUFR, 1))
+                Store (\_SB.MEM.BLF1, Index (BUFR, 2))
+                Store (\_SB.MEM.BTC1, Index (BUFR, 3))
+                Store (\_SB.MEM.BDV1, Index (BUFR, 4))
+                Store (\_SB.MEM.BCW1, Index (BUFR, 5))
+                Store (\_SB.MEM.BCL1, Index (BUFR, 6))
+                Store (\_SB.MEM.BG11, Index (BUFR, 7))
+                Store (\_SB.MEM.BG21, Index (BUFR, 8))
+                Store (\_SB.MEM.BMN1, Index (BUFR, 9))
+                Store (\_SB.MEM.BSN1, Index (BUFR, 10))
+                Store (\_SB.MEM.BTP1, Index (BUFR, 11))
+                Store (\_SB.MEM.BOI1, Index (BUFR, 12))
+
+                Return (BUFR)
+            }   //  _BIF
+        }   //  BAT1
+
+        Device (IDX2)
+        {
+            Method (B2IB)
+            {   //  B2IB:   store from Buffer into Index'ed Buffer
+
+                Name (SRCB, Buffer ()   {"Short Buffer"})   //  12 characters plus NULL
+
+                Name (DEST, Buffer ()                           //  62 characters plus NULL
+                    {"Destination buffer that is longer than the short source buffer"})
+
+
+                //  verify object type returned by Index(Buffer,Element,)
+
+                Store (Index (DEST, 2, ), Local1)
+                Store (ObjectType (Local1), Local2)
+
+                If (LNotEqual (Local2, 14))     //  Buffer Field is type 14
+                {
+                    //  Local2 indicates Local1 is not a Buffer Field
+
+                    Return (0x61)
+                }
+
+                //  verify object type and value returned by DerefOf(Index(Buffer,Element,))
+                //  should return Number containing element value
+
+                Store (DerefOf (Local1), Local3)
+                Store (ObjectType (Local3), Local4)
+
+                If (LNotEqual (Local4, 1))          //  Number is type 1
+                {
+                    //  Local2 indicates Local1 is not a Number
+                    Return (0x62)
+                }
+                Else
+                {
+                    If (LNotEqual (Local3, 0x73))       //  expect 's' element from DEST
+                    {
+                        Return (0x63)
+                    }
+                }
+
+                Store ("DerefOf(Index(Buffer,,)) PASS", Debug)
+
+
+                //
+                // The following sections have been rewritten because storing into
+                // an Indexed buffer only changes one byte - the FIRST byte of the
+                // buffer is written to the source index.  This is the ONLY byte
+                // written -- as per ACPI 2.0
+                //
+                // Overwrite DEST contents, at buffer position 2 [only]
+
+                Store (SRCB, Index (DEST, 2, ))
+
+                //
+                // Check that the next byte is not changed
+                //
+                Store (DerefOf (Index (DEST, 3, )), Local0)
+                If (LNotEqual (Local0, 0x74))       //  't'
+                {
+                    //  DEST element is not matching original value
+                    If (LEqual (Local0, 0x68))
+                    {
+                        //  DEST element was altered to 'h'
+                        Return (0x68)
+                    }
+                    Else
+                    {
+                        // DEST element is an unknown value
+                        Return (0x69)
+                    }
+                }
+
+                //
+                // Check that the elements beyond the SRCB buffer copy
+                //  have not been altered.
+                //
+                Store (DerefOf (Index (DEST, 14)), Local0)
+
+                //
+                // This should be an 'f'.
+                //
+                If (LNotEqual (Local0, 0x66))
+                {
+                    //  DEST element was zero'd by buffer copy
+                    If (LEqual (Local0, 0))
+                    {
+                        //  DEST element is zero
+                        Return (0x6A)
+                    }
+                    Else
+                    {
+                        //  DEST element is unknown value
+                        Return (0x6B)
+                    }
+                }
+
+                Store ("Store(SRCB,Index(Buffer,,)) PASS", Debug)
+
+                //
+                //  verify altering SRCB does NOT alter DEST
+                //
+                Store (0x6A, Index (SRCB, 1))   //  SRCB = "Sjort Buffer"
+
+                Store (DerefOf (Index (SRCB, 1)), Local0)
+
+                If (LNotEqual (Local0, 0x6A))       //  'j'
+                {
+                    //  SRCB element is unaltered
+                    Return (0x71)
+                }
+
+                Store (DerefOf (Index (DEST, 3)), Local0) // DEST = "Destination buffer that...
+
+                If (LNotEqual (Local0, 0x74))       //  't'
+                {
+                    //  DEST element is altered
+                    If (LEqual (Local0, 0x6A))  //  'j'
+                    {
+                        //  SRCB change altered DEST element
+                        Return (0x72)
+                    }
+                    Else
+                    {
+                        //  DEST element is unknown value
+                        Return (0x73)
+                    }
+                }
+
+                //  verify altering DEST does NOT alter SRCB
+
+                Store (0x6B, Index (DEST, 4, )) //  DEST = "DeSkination buffer..."
+
+                Store (DerefOf (Index (DEST, 4, )), Local0)
+
+                If (LNotEqual (Local0, 0x6B))       //  'k'
+                {
+                    //  DEST element is unaltered
+                    Return (0x74)
+                }
+
+                Store (DerefOf (Index (SRCB, 2, )), Local0)
+
+                If (LNotEqual (Local0, 0x6F))       //  'o'
+                {   //  SRC element is altered
+                    If (LEqual (Local0, 0x6B))  //  'k'
+                    {
+                        //  DEST change altered SRCB element
+                        Return (0x75)
+                    }
+                    Else
+                    {
+                        //  SRCB element is unknown value
+                        Return (0x76)
+                    }
+                }
+
+                Store ("SRCB and DEST independent PASS", Debug)
+
+
+                // verify string can be written to Index target/destination
+                // Only FIRST byte is written
+
+                Store ("New Buff", Index (DEST, 2, ))   //  DEST = "DeNkination buffer..."
+
+                Store (DerefOf (Index (DEST, 2, )), Local0)
+
+                If (LNotEqual (Local0, 0x4E))       //  'N'
+                {
+                    //  DEST element is unaltered
+                    Return (0x81)
+                }
+
+                Store (DerefOf (Index (DEST, 6, )), Local0)
+
+                If (LNotEqual (Local0, 0x61))       //  'a'
+                {
+                    //  DEST element is unaltered
+                    Return (0x82)
+                }
+
+                Store (DerefOf (Index (DEST, 10, )), Local0)
+
+                If (LNotEqual (Local0, 0x6E))       //  'n'
+                {
+                    //  DEST element is unaltered
+                    Return (0x83)
+                }
+
+                Store ("Store(String,Index) PASS", Debug)
+
+
+                Return (0)  //  pass
+            }   //  B2IB:   store from Buffer into Index'ed Buffer
+
+            Method (FB2P)
+            {   //  FB2P:   store from Field Buffer into Index'ed Package
+                Name (DEST, Package (2) {})
+
+                //  initialize memory using 32-bit field elements
+                Store (0x01234567, \_SB.MEM.SMD0)
+                Store (0x89ABCDEF, \_SB.MEM.SMD1)
+                Store (0xFEDCBA98, \_SB.MEM.SMD2)
+                Store (0x76543210, \_SB.MEM.SMD3)
+
+                //  move greater than 64-bit buffers into DEST package
+                Store (\_SB.MEM.SME0, Index (DEST, 0))
+                Store (\_SB.MEM.SME1, Index (DEST, 1))
+
+                //  validate DEST contents
+                Store (DerefOf (Index (DEST, 0, )), Local0)
+                Store (DerefOf (Index (DEST, 1, )), Local1)
+
+                //  verify Local0 and Local1 are Buffers
+                Store (ObjectType (Local0), Local2)
+                if (LNotEqual (Local2, 3))  //  Buffer type is 3
+                {
+                    Return (0x11)
+                }
+
+                Store (ObjectType (Local1), Local3)
+                if (LNotEqual (Local3, 3))  //  Buffer type is 3
+                {
+                    Return (0x12)
+                }
+
+                //  validate DEST buffer contents
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0)), 0)), Local4)
+                If (LNotEqual (Local4, 0x67))
+                {
+                    Return (0x13)
+                }
+
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0)), 1)), Local4)
+                If (LNotEqual (Local4, 0x45))
+                {
+                    Return (0x14)
+                }
+
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0)), 4)), Local4)
+                If (LNotEqual (Local4, 0xEF))
+                {
+                    Return (0x15)
+                }
+
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0, )), 5, )), Local4)
+                If (LNotEqual (Local4, 0xCD))
+                {
+                    Return (0x16)
+                }
+
+                Store ("Store(Mem,PkgElement) PASS", Debug)
+
+
+                //  validate changing source \_SB.MEM.SMD* does not impact DEST
+                Store (0x12345678, \_SB.MEM.SMD0)
+
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0, )), 0, )), Local5)
+                If (LNotEqual (Local5, 0x67))
+                {
+                    Return (0x21)
+                }
+
+                Store (DerefOf (Index (DerefOf (Index (DEST, 0, )), 1, )), Local5)
+                If (LNotEqual (Local5, 0x45))
+                {
+                    Return (0x22)
+                }
+
+                //  validate changing DEST does not impact source \_SB.MEM.SMD*
+                Store (0x30, Index (DerefOf (Index (DEST, 0)), 0))
+
+                Store (DerefOf(Index (DerefOf (Index (DEST, 0)), 0)), Local5)
+                If (LNotEqual (Local5, 0x30))
+                {
+                    Return (0x23)
+                }
+
+                //
+                // This section was modified from the original iPCO code because
+                //  it attempted to compare two buffers.  This is not allowed until
+                //  ACPI v2.0, so the test has been modified to just check the
+                //  changed \_SB.MEM.SMD0
+                //
+                Store (\_SB.MEM.SMD0, Local5)
+
+                If(LNotEqual(Local5, 0x12345678))
+                {
+                    Return (0x24)
+                }
+
+                Store ("Mem and Pkg independent PASS", Debug)
+
+
+                Return (0)
+            }   //  FB2P:   store from Field Buffer into Index'ed Package
+
+            Method (TEST)
+            {
+                Store ("++++++++ IndexOp2 Test", Debug)
+
+                //  store _BIF package return value into Local0
+
+                Store (\_SB.BAT1._BIF, Local0)
+
+                //  save Local0 object type value into Local1
+                Store (ObjectType (Local0), Local1)
+
+                //  validate Local0 is a Package
+                If (LNotEqual (Local1, 4))  //  Package type is 4
+                {
+                    //  failure
+                    Return (2)
+                }
+
+                //  validate source and destination buffers are independent of each
+                //  of each other (i.e., changing one's contents does not change
+                //  other's contents) using B2IB (store from Buffer into Index'ed
+                //  Buffer) and FB2P (store from Field Buffer into Index'ed Package)
+
+                //  call B2IB (store from Buffer into Index'ed Buffer)
+                Store (B2IB, Local2)    //  Local2 is B2IB return value
+
+                //  save Local2 object type value into Local3
+                Store (ObjectType (Local2), Local3)
+
+                //  validate Local2 is a Number
+                If (LNotEqual (Local3, 1))  //  Number type is 1
+                {
+                    //  failure
+                    Return (4)
+                }
+
+                //  zero indicates pass, non-zero is an error code
+                If (LNotEqual (Local2, 0))
+                {
+                    //  return B2IB error code
+                    Return (Local2)
+                }
+
+                //  call FB2P (store from Field Buffer into Index'ed Package)
+                Store (FB2P, Local2)    //  Local2 is FB2P return value
+
+                //  save Local2 object type value into Local3
+                Store (ObjectType (Local2), Local3)
+
+                //  validate Local2 is a Number
+                If (LNotEqual (Local3, 1))  //  Number type is 1
+                {
+                    //  failure
+                    Return (5)
+                }
+
+                //  zero indicates pass, non-zero is an error code
+                If (LNotEqual (Local2, 0))
+                {
+                    //  return FB2P error code
+                    Return (Local2)
+                }
+
+
+                Return (0)
+            }   //  TEST
+        }   //  IDX2:   Test device name
+    }   //  _SB system bus
+
+//
+// test SizeOf.asl
+//
+//  Test for SizeOf
+//      test cases include following SizeOf arguments:
+//          buffer, buffer field;
+//          control method argument, control method local variable;
+//          control method return values;
+//          direct string, string;
+//          package;
+//          buffer, package, and string package elements
+//
+// MTL NOTE: This test has been modified to remove any SizeOf(Index(Buff,...
+//  calls because it is not legal to perform a SizeOf operation on a Buffer Field.
+//  This test has also been extended to test additional Package element sizes.
+//
+    Device (SIZO)
+    {
+        //  SAR0 control method validates SizeOf(Arg)
+        //      SAR0 should only be called by SARG
+        Method (SAR0, 2)
+        //  Arg0    object to determine size of
+        //  Arg1    expected Arg length
+        {   //  SAR0:   SizeOf(Arg) test control method
+            //  Local0  Arg0 length
+            //  Local1  Local0 object type
+
+            //  Store first string size (Arg0) into Local7
+            Store (SizeOf (Arg0), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+                {   Return (0x21)   }
+
+            //  If strings are not of equal size, return error code
+            If (LNotEqual (Local0, Arg1))
+                {   Return (0x22)   }
+
+            Return (0)
+        }   //  SAR0:   SizeOf(Arg) test control method
+
+        Method (SARG)
+        {   //  SARG:   SizeOf(Arg) test control method
+            Name (BUFR, Buffer (12) {}) //  uninitialized Buffer
+            Name (BUF1, Buffer() {0x01, 0x02, 0x03, 0x04, 0x05})
+            Name (PKG0, Package (4) {}) //  uninitialized Package
+            Name (STR0, "String")
+            Name (PKG1, Package (4)
+            {
+                BUFR,
+                "String2",
+                STR0,
+                PKG0
+            })  //  PKG1
+
+            Name (PKG2, Package (4)
+            {
+                Buffer (15) {},
+                "String 1",
+                Package (2) {}
+            })  //  PKG2
+
+            //  Namespace entry buffer reference
+            Store (SAR0 (BUFR, 12), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x23)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=BUFR) PASS", Debug)
+
+
+            //  Namespace entry package reference
+            Store (SAR0 (PKG0, 4), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x24)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=PKG0) PASS", Debug)
+
+
+            //  Namespace entry string reference
+            Store (SAR0 (STR0, 6), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x25)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=STR0) PASS", Debug)
+
+
+            //  direct string reference
+            Store (SAR0 ("String", 6), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x26)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=String) PASS", Debug)
+
+            Store (0x55, Index (BUF1, 2))
+
+            /****************************************************
+            //
+            // This section is commented because it is illegal to
+            //  perform a SizeOf operation on a Buffer Field
+            //
+            //  Namespace BufferField reference
+            Store (SAR0 (Index (BUFR, 2, ), 10), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+                {   Return (0x27)   }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+                {   Return (Local0) }
+
+            Store ("SizeOf(Arg=BufferField) PASS", Debug)
+            ****************************************************/
+
+            //  Namespace BufferPackageElement reference
+            //
+            Store (SAR0 (Index(PKG1, 0), 12), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x28)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=PackageBuffer NTE Reference Element) PASS", Debug)
+
+
+            //  Namespace StringPackageElement reference
+            Store (SAR0 (Index (PKG1, 1, ), 7), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x29)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package String Element) PASS", Debug)
+
+
+            //  Namespace StringPackageElement reference
+            Store (SAR0 (Index (PKG1, 2, ), 6), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x2A)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package String NTE Reference Element) PASS", Debug)
+
+
+            //  Namespace PackagePackageElement reference
+            Store (SAR0 (Index (PKG1, 3, ), 4), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x2B)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package Package NTE Reference Element) PASS", Debug)
+
+            // Package Buffer Element
+            Store (SAR0 (Index (PKG2, 0), 15), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x2B)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package Buffer Element) PASS", Debug)
+
+            // Package String Element
+            Store (SAR0 (Index (PKG2, 1), 8), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x2B)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package String Element) PASS", Debug)
+
+            // Package Package Element
+            Store (SAR0 (Index (PKG2, 2), 2), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x2B)
+            }
+
+            If (LNotEqual (Local0, 0))      //  Local0 is SAR0 return error code
+            {
+                Return (Local0)
+            }
+
+            Store ("SizeOf(Arg=Package Package Element) PASS", Debug)
+
+            Store ("SizeOf(Arg) PASS", Debug)
+
+            Return (0)
+        }   //  SARG:   SizeOf(Arg) test control method
+
+        Method (SBUF)
+        {   //  SBUF:   SizeOf(Buffer) test control method
+            Name (BUFR, Buffer (12) {})
+
+            //  store size of BUFR buffer into Local0
+            Store (SizeOf (BUFR), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x31)
+            }
+
+            If (LNotEqual (Local0, 12))     //  BUFR size is 12
+            {
+                Return (0x32)
+            }
+
+            Store ("SizeOf(BUFR) PASS", Debug)
+
+            Return (0)
+        }   //  SBUF:   SizeOf(Buffer) test control method
+
+
+        /****************************************************
+        //
+        // This section is commented because it is illegal to
+        //  perform a SizeOf operation on a Buffer Field
+        //
+        Method (SIND)
+        {   //  SIND:   SizeOf(Index(,,)) test control method
+            Name (BUFR, Buffer (12) {})
+
+            //  store size of Index(BUFR,2,) buffer into Local0
+            Store (SizeOf (Index (BUFR, 2, )), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x41)
+            }
+
+            If (LNotEqual (Local0, 10))     //  12 - 2 = 10
+            {
+                Return (0x42)
+            }
+
+            Store ("SizeOf(Index(BUFR,,)) PASS", Debug)
+
+            //  TBD:    strings and packages
+
+            Return (0)
+        }   //  SIND:   SizeOf(Index(,,)) test control method
+        ****************************************************/
+
+        Method (SLOC)
+        {   //  SLOC:   SizeOf(Local) test control method
+            Name (BUFR, Buffer (12) {}) //  uninitialized Buffer
+            Name (STR0, "String")
+            Name (PKG0, Package (4) {}) //  uninitialized Package
+
+
+            //  store BUFR Buffer into Local2
+            Store (BUFR, Local2)
+
+            //  store size of BUFR buffer into Local0
+            Store (SizeOf (Local2), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x51)
+            }
+
+            If (LNotEqual (Local0, 12)) //  BUFR size is 12
+            {
+                Return (0x52)
+            }
+
+            Store ("SizeOf(Local2=Buffer) PASS", Debug)
+
+
+            //  store STR0 string into Local2
+            Store (STR0, Local2)
+
+            //  store size of STR0 buffer into Local0
+            Store (SizeOf (Local2), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x53)
+            }
+
+            If (LNotEqual (Local0, 6))      //  STR0 size is 6
+            {
+                Return (0x54)
+            }
+
+            Store ("SizeOf(Local2=String) PASS", Debug)
+
+
+            //  store PKG0 Package into Local2
+            Store (PKG0, Local2)
+
+            //  store size of PKG0 buffer into Local0
+            Store (SizeOf (Local2), Local0)
+
+            //  save Local0 object type value into Local1
+            Store (ObjectType (Local0), Local1)
+
+            //  validate Local0 is a Number
+            If (LNotEqual (Local1, 1))      //  Number type is 1
+            {
+                Return (0x55)
+            }
+
+            If (LNotEqual (Local0, 4))      //  PKG0 size is 4
+            {
+                Return (0x56)
+            }
+
+            Store ("SizeOf(Local2=Package) PASS", Debug)
+
+
+            Return (0)
+        }   //  SLOC:   SizeOf(Local) test control method
+
+        Method (TEST)
+        {
+            Store ("++++++++ SizeOf Test", Debug)
+
+            //  Store current operating system string into Local0
+            Store (_OS, Local0)
+
+            Store (SizeOf (_OS), Local3)
+
+            //  save Local3 object type value into Local4
+            Store (ObjectType (Local3), Local4)
+
+            //  validate Local3 is a Number
+            If (LNotEqual (Local4, 1))  //  Number type is 1
+            {
+                //  failure
+                Return (0x61)
+            }
+
+            //  Store current operating system string into Local0
+            //  This verifies above SizeOf(_OS) did not corrupt ACPI namespace
+            Store (_OS, Local0)
+
+            //  Store SARG [Validate SizeOf(Arg)] return value into Local1
+            Store (SARG, Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+            {
+                //  failure
+                Return (0x62)
+            }
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+            {
+                //  return SARG error code
+                Return (Local1)
+            }
+
+
+            //  Store SBUF [Validate SizeOf(Buffer)] return value into Local1
+            Store (SBUF, Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+            {
+                //  failure
+                Return (0x63)
+            }
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+            {
+                //  return SBUF error code
+                Return (Local1)
+            }
+
+            /****************************************************
+            //
+            // This section is commented because it is illegal to
+            //  perform a SizeOf operation on a Buffer Field
+            //
+            //  Store SIND [verify SizeOf(Index(,,))] return value into Local1
+            Store (SIND, Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+            {
+                //  failure
+                Return (0x64)
+            }
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+            {
+                //  return SARG error code
+                Return (Local1)
+            }
+            ****************************************************/
+
+            //  Store SLOC [verify SizeOf(Local)] return value into Local1
+            Store (SLOC, Local1)
+
+            //  save Local1 object type value into Local2
+            Store (ObjectType (Local1), Local2)
+
+            //  validate Local1 is a Number
+            If (LNotEqual (Local2, 1))  //  Number type is 1
+            {
+                //  failure
+                Return (0x65)
+            }
+
+            //  zero indicates pass, non-zero is an error code
+            If (LNotEqual (Local1, 0))
+            {
+                //  return SLOC error code
+                Return (Local1)
+            }
+
+
+            //  TBD:    SizeOf (METH) -- where METH control method returns
+            //              buffer, BufferField, string, package, package element
+
+
+            Return (0)
+        }   //  TEST
+    }   //  SIZO
+
+//
+// test SmiShare.asl
+//
+    Scope (\_SB)    //  System Bus
+    {   //  _SB system bus
+        //  Declare an OpRegion in Memory starting at offset 0x400000 that is 10 bytes long
+        OperationRegion(RAM1, SystemMemory, 0x400000, 0xA)
+
+        Field (RAM1, AnyAcc, NoLock, Preserve)
+        {
+            BI1T, 1,        // Create some bits in memory to access
+            BI2T, 2,
+            BI3T, 3,
+            LST2, 2
+        }   //  End Field RAM1
+
+        Field (RAM1, WordAcc, NoLock, WriteAsZeros)
+        {
+            WRD, 16
+        }   //  End 2nd Field RAM1
+
+        Field (RAM1, ByteAcc, NoLock, WriteAsOnes)
+        {
+            BYTE, 8
+        }   //  End 3rd Field RAM1
+
+        Field (RAM1, ByteAcc, NoLock, Preserve)
+        {
+            SMIC, 8,
+            SMID, 8
+        }
+
+        Device (MBIT)
+        {
+            Method (_INI)
+            {
+                Store (0, BI1T)
+                Store (3, BI2T)
+                Store (7, BI3T)
+                Store (0, LST2)
+            }   //  End _INI Method
+        }   //  End Device MBIT
+
+        Device (MWRD)
+        {
+            Method (_INI)
+            {
+                Store (0, WRD)
+            }   //  End _INI Method
+        }   //  End Device MWRD
+
+        Device (MBYT)
+        {
+            Method (_INI)
+            {
+                Store (0, BYTE)
+                Store (0xC, SMIC)
+                Store (0xD, SMID)
+            }   //  End _INI Method
+        }   //  End Device MBYT
+
+    /*
+        //  Declare an OpRegion in Memory starting at offset 0x400000 that is 10 bytes long
+        OperationRegion(\RAM1, SystemMemory, 0x400000, 0xA)
+
+        Field (\RAM1, AnyAcc, NoLock, Preserve)
+        {
+            BI1T, 1,        // Create some bits in memory to access
+            BI2T, 2,
+            BI3T, 3,
+            LST2, 2
+        }   //  End Field RAM1
+
+        Field (\RAM1, WordAcc, NoLock, WriteAsZeros)
+        {
+            WRD, 16
+        }   //  End 2nd Field RAM1
+
+        Field (\RAM1, ByteAcc, NoLock, WriteAsOnes)
+        {
+            BYTE, 8
+        }   //  End 3rd Field RAM1
+
+        Field (\RAM1, ByteAcc, NoLock, Preserve)
+        {
+            SMIC, 8,
+            SMID, 8
+        }
+    */
+        Method (SMIX)
+        {
+            Return (BYTE)
+        }   //  End SMIX
+
+        Method (EVNT)
+        {
+            Store (SMIX, Local0)
+
+            If (And (Local0, 0x01))
+            {   Notify (\_SB_.SMIS, 0x21)}
+
+            If (And (Local0, 0x02))
+            {   Notify (\_SB_.SMIS, 0x22)}
+
+            If (And (Local0, 0x04))
+            {   Notify (\_SB_.SMIS, 0x24)}
+
+            If (And (Local0, 0x08))
+            {   Notify (\_SB_.SMIS, 0x28)}
+
+        }   //  End Method EVNT
+
+        Device (SMIS)
+        {
+            Method (BINK)
+            {
+                Store (0, Local0)               //  Zero out Local0
+
+                If (LNotEqual (SMID, 0xD))
+                {   Or (0x80, Local0, Local0)}
+
+                If (LNotEqual (SMIC, 0xC))
+                {   Or (0x40, Local0, Local0)}
+
+                If (LNotEqual (BYTE, 0))
+                {   Or (0x20, Local0, Local0)}
+
+                If (LNotEqual (WRD, 0))
+                {   Or (0x10, Local0, Local0)}
+
+                If (LNotEqual (LST2, 0))
+                {   Or (0x8, Local0, Local0)}
+
+                If (LNotEqual (BI3T, 0x7))
+                {   Or (0x4, Local0, Local0)}
+
+                If (LNotEqual (BI2T, 0x3))
+                {   Or (0x2, Local0, Local0)}
+
+                If (LNotEqual (BI1T, 0))
+                {   Or (0x1, Local0, Local0)}
+
+                Return (Local0)
+            }   //  End Method BINK
+
+            Method (TEST)
+            {
+                Store ("++++++++ SmiShare Test", Debug)
+
+                //  Expect EVNT to generate Notify value we just previously
+                //  stored in BYTE
+                Store (0x20, BYTE)
+                EVNT ()
+                Store (0x21, BYTE)
+                EVNT ()
+                Store (0x22, BYTE)
+                EVNT ()
+                Store (0x23, BYTE)
+                EVNT ()
+
+                Return (0)  //  pass
+            }   //  End Method TEST
+        }   //  Device SMIS
+
+        Device(CNDT)
+        {
+            Method(TEST)
+            {
+                If (ECOK)
+                {
+                    return("Broken")
+                }
+                Else
+                {
+                    return("Works")
+                }
+            }
+
+            Method(ECOK)
+            {
+                Return(0x0)
+            }
+        }
+
+    }   //  _SB system bus
+
+
+/* Test a very big buffer */
+
+    Name(WQAB, Buffer(6756)
+    {
+        0x46,0x4F,0x4D,0x42,0x01,0x00,0x00,0x00,
+        0x54,0x1A,0x00,0x00,0xBA,0xAD,0x00,0x00,
+        0x44,0x53,0x00,0x01,0x1A,0x7D,0xDA,0x54,
+        0x98,0xBD,0x92,0x00,0x01,0x06,0x18,0x42,
+        0x10,0x47,0x10,0x92,0x46,0x62,0x02,0x89,
+        0x80,0x90,0x18,0x18,0x14,0x81,0x85,0x00,
+        0x49,0x02,0x88,0xC4,0x41,0xE1,0x20,0xD4,
+        0x9F,0x40,0x7E,0x05,0x20,0x74,0x28,0x40,
+        0xA6,0x00,0x83,0x02,0x9C,0x22,0x88,0xA0,
+        0x57,0x01,0x36,0x05,0x98,0x14,0x60,0x51,
+        0x80,0x76,0x01,0x96,0x05,0xE8,0x16,0x20,
+        0x1D,0x96,0x88,0x04,0x47,0x89,0x01,0x47,
+        0xE9,0xC4,0x16,0x6E,0xD8,0xE0,0x85,0xA2,
+        0x68,0x06,0x51,0x12,0x94,0x8B,0x20,0x5D,
+        0x10,0x52,0x2E,0xC0,0x37,0x82,0x06,0x10,
+        0xA5,0x77,0x01,0xB6,0x05,0x98,0x86,0x27,
+        0xD2,0x20,0xE4,0x60,0x08,0x54,0xCE,0x80,
+        0x20,0x69,0x44,0x21,0x1E,0xA7,0x44,0x08,
+        0x0A,0x84,0x90,0xD4,0xF1,0xA0,0xA0,0x71,
+        0x88,0xAD,0xCE,0x46,0x93,0xA9,0x74,0x7E,
+        0x48,0x82,0x70,0xC6,0x2A,0x7E,0x3A,0x9A,
+        0xD0,0xD9,0x9C,0x60,0xE7,0x18,0x72,0x3C,
+        0x48,0xF4,0x20,0xB8,0x00,0x0F,0x1C,0x2C,
+        0x34,0x84,0x22,0x6B,0x80,0xC1,0x8C,0xDD,
+        0x63,0xB1,0x0B,0x4E,0x0A,0xEC,0x61,0xB3,
+        0x01,0x19,0xA2,0x24,0x38,0xD4,0x11,0xC0,
+        0x12,0x05,0x98,0x1F,0x87,0x0C,0x0F,0x95,
+        0x8C,0x25,0x24,0x1B,0xAB,0x87,0xC2,0xA5,
+        0x40,0x68,0x6C,0x27,0xED,0x19,0x45,0x2C,
+        0x79,0x4A,0x82,0x49,0xE0,0x51,0x44,0x36,
+        0x1A,0x27,0x28,0x1B,0x1A,0x25,0x03,0x42,
+        0x9E,0x05,0x58,0x07,0x26,0x04,0x76,0x2F,
+        0xC0,0x9A,0x00,0x73,0xB3,0x90,0xB1,0xB9,
+        0xE8,0xFF,0x0F,0x71,0xB0,0x31,0xDA,0x9A,
+        0xAE,0x90,0xC2,0xC4,0x88,0x12,0x2C,0x5E,
+        0xC5,0xC3,0x10,0xCA,0x93,0x42,0xA8,0x48,
+        0x95,0xA1,0x68,0xB4,0x51,0x2A,0x14,0xE0,
+        0x4C,0x80,0x30,0x5C,0x1D,0x03,0x82,0x46,
+        0x88,0x15,0x29,0x56,0xFB,0x83,0x20,0xF1,
+        0x2D,0x40,0x54,0x01,0xA2,0x48,0xA3,0x41,
+        0x9D,0x03,0x3C,0x5C,0x0F,0xF5,0xF0,0x3D,
+        0xF6,0x93,0x0C,0x72,0x90,0x67,0xF1,0xA8,
+        0x70,0x9C,0x06,0x49,0xE0,0x0B,0x80,0x4F,
+        0x08,0x1E,0x38,0xDE,0x35,0xA0,0x66,0x7C,
+        0xBC,0x4C,0x10,0x1C,0x6A,0x88,0x1E,0x68,
+        0xB8,0x13,0x38,0x44,0x06,0xE8,0x49,0x3D,
+        0x52,0x60,0x07,0x77,0x32,0xEF,0x01,0xAF,
+        0x0A,0xCD,0x5E,0x12,0x08,0xC1,0xF1,0xF8,
+        0x7E,0xC0,0x26,0x9C,0xC0,0xF2,0x07,0x81,
+        0x1A,0x99,0xA1,0x3D,0xCA,0xD3,0x8A,0x19,
+        0xF2,0x31,0xC1,0x04,0x16,0x0B,0x21,0x05,
+        0x10,0x1A,0x0F,0xF8,0x6F,0x00,0x8F,0x17,
+        0xBE,0x12,0xC4,0xF6,0x80,0x12,0x0C,0x0B,
+        0x21,0x23,0xAB,0xF0,0x78,0xE8,0x28,0x7C,
+        0x95,0x38,0x9C,0xD3,0x8A,0x67,0x82,0xE1,
+        0x20,0xF4,0x05,0x90,0x00,0x51,0xE7,0x0C,
+        0xD4,0x61,0xC1,0xE7,0x04,0x76,0x33,0x38,
+        0x83,0x47,0x00,0x8F,0xE4,0x84,0xFC,0x2B,
+        0xF1,0xC0,0xE0,0x03,0xE2,0xEF,0x1F,0xA7,
+        0xEC,0x11,0x9C,0xA9,0x01,0x7D,0x1C,0xF0,
+        0xFF,0x7F,0x28,0x7C,0x88,0x1E,0xDF,0x29,
+        0x1F,0xAF,0x4F,0x17,0x96,0x35,0x4E,0xE8,
+        0x77,0x08,0x9F,0x38,0x7C,0x64,0x71,0x44,
+        0x08,0x39,0x39,0x05,0xA0,0x81,0x4F,0xF7,
+        0xEC,0x22,0x9C,0xAE,0x27,0xE5,0x40,0xC3,
+        0xA0,0xE3,0x04,0xC7,0x79,0x00,0x1C,0xE3,
+        0x84,0x7F,0x2E,0x80,0x3F,0x40,0x7E,0xCA,
+        0x78,0xC5,0x48,0xE0,0x98,0x23,0x44,0x9F,
+        0x6B,0x3C,0x42,0x2C,0xFC,0x53,0x45,0xE1,
+        0x03,0x21,0x63,0x04,0x17,0xA0,0xC7,0x08,
+        0x7C,0x03,0x8E,0x11,0x7D,0x94,0xE0,0xEA,
+        0x0F,0x1A,0x74,0x80,0xB8,0xFF,0xFF,0x00,
+        0xE1,0x83,0x7A,0x80,0xC0,0x37,0xFA,0xD1,
+        0x03,0x3D,0x2E,0x8B,0x3E,0x0F,0xC8,0xF8,
+        0x89,0x46,0xF3,0xE2,0xA7,0x03,0x7E,0xF8,
+        0x00,0x0F,0xA8,0x87,0x84,0x03,0xC5,0x4C,
+        0x9B,0x83,0x3E,0xBB,0x1C,0x3A,0x76,0xB8,
+        0xE0,0x3F,0x81,0x80,0x4B,0xDE,0x21,0x0C,
+        0x14,0x23,0xC6,0x9F,0x83,0x7C,0x0A,0x03,
+        0xFF,0xFF,0xFF,0x14,0x06,0xFE,0xE1,0xF0,
+        0x20,0x4F,0x07,0x9F,0xB6,0xA8,0x74,0x18,
+        0xD4,0x81,0x0B,0xB0,0x32,0x89,0x08,0xCF,
+        0x12,0xB5,0x41,0xE8,0xD4,0xF0,0x36,0xF1,
+        0xB6,0xE5,0x5B,0x40,0x9C,0xD3,0xEC,0xED,
+        0xC0,0x45,0x30,0x22,0xD4,0x0C,0x45,0x4E,
+        0x5A,0x11,0x63,0x44,0x79,0xDC,0x32,0xCA,
+        0xDB,0xD6,0x0B,0x40,0xBC,0x13,0x7B,0xDE,
+        0x32,0x46,0xF0,0xC8,0x0F,0x5C,0x2C,0xC6,
+        0xEA,0xF5,0x5F,0xF3,0x81,0x0B,0x70,0xF6,
+        0xFF,0x3F,0x70,0x01,0x1C,0x0A,0x7A,0x18,
+        0x42,0x0F,0xC3,0x53,0x39,0x97,0x87,0xC8,
+        0x53,0x89,0x18,0x35,0x4C,0xD4,0x67,0x28,
+        0xDF,0x2D,0x7C,0x20,0x02,0xDF,0x99,0x0B,
+        0xF8,0xFD,0xFF,0x0F,0x44,0x70,0x8E,0x29,
+        0xB8,0x33,0x0D,0x78,0x7C,0xCE,0x40,0x20,
+        0xA7,0xE2,0x43,0x0D,0x60,0x41,0xF4,0x13,
+        0xC2,0x27,0x1A,0x2A,0x13,0x06,0x75,0xA8,
+        0x01,0xAC,0x5C,0x61,0x9E,0x46,0xCF,0xF9,
+        0x59,0xC6,0xA7,0x1A,0x1F,0x4A,0x8D,0x63,
+        0x88,0x97,0x99,0x87,0x1A,0x1F,0x0B,0x5E,
+        0x49,0x7D,0xA8,0x31,0x54,0x9C,0x87,0x1A,
+        0x0F,0x37,0x50,0xD4,0x37,0x9B,0x67,0x1B,
+        0xA3,0xC7,0xF7,0x0D,0xD5,0x10,0x0F,0x35,
+        0x4C,0xF2,0x4A,0x35,0x16,0x1F,0x6A,0xC0,
+        0xF1,0xFF,0x3F,0xD4,0x00,0xFC,0xFF,0xFF,
+        0x1F,0x6A,0x00,0x47,0x47,0x03,0x38,0x47,
+        0x46,0xDC,0xD1,0x00,0x5C,0x87,0x52,0xE0,
+        0x70,0x34,0x00,0x1E,0x47,0x21,0x30,0x5F,
+        0x68,0x7C,0x14,0x02,0x16,0xFF,0xFF,0xA3,
+        0x10,0xF8,0x65,0x9F,0x83,0x50,0x42,0x8F,
+        0x42,0x80,0xA0,0xDB,0xCF,0x53,0xC4,0xB3,
+        0x8F,0x2F,0x3F,0x0F,0x04,0x11,0x5E,0xF3,
+        0x7D,0x0A,0xF2,0x21,0xDF,0x47,0x21,0x06,
+        0x63,0x28,0x5F,0x83,0x7C,0x14,0x62,0x50,
+        0xAF,0x41,0xBE,0xEF,0x1B,0xE4,0xF1,0x22,
+        0x48,0xEC,0x67,0x02,0x1F,0x85,0x98,0xE8,
+        0xA3,0x10,0xA0,0xF0,0xFF,0x7F,0x14,0x02,
+        0xF8,0xFF,0xFF,0x3F,0x0A,0x01,0xCE,0x02,
+        0x1C,0x0D,0x40,0x37,0xAD,0x47,0x21,0xF0,
+        0xDE,0x59,0x4E,0xFB,0x04,0x7C,0x16,0x02,
+        0xCC,0xFE,0xFF,0xCF,0x42,0xC0,0xEC,0x28,
+        0x74,0x14,0x67,0xF9,0x2A,0xF4,0x04,0xF0,
+        0x02,0x10,0x23,0xCC,0x3B,0xD0,0x4B,0x26,
+        0xBB,0x8B,0x1B,0xE7,0xC9,0xE5,0x2C,0x9E,
+        0xC4,0x7D,0x09,0xF2,0x81,0xE2,0x59,0xC8,
+        0x50,0xA7,0x1B,0xF4,0x8D,0xDC,0x03,0x8B,
+        0x19,0x3F,0xC4,0xF3,0x90,0x21,0x9E,0x85,
+        0x00,0x76,0xFD,0xFF,0xCF,0x42,0x00,0xFF,
+        0xFF,0xFF,0x47,0x03,0xF8,0x2F,0x00,0x9F,
+        0x85,0x80,0xE7,0x09,0xE0,0x41,0xDB,0x67,
+        0x21,0x80,0x33,0x87,0xCB,0xF3,0x7F,0x05,
+        0x3A,0x96,0xF7,0x08,0xCF,0xFA,0x24,0x5F,
+        0x2F,0x3D,0xD3,0x87,0x82,0x67,0x21,0x86,
+        0x75,0x18,0x3E,0x0B,0x31,0x88,0x17,0x4D,
+        0x43,0xBC,0x70,0xFA,0x30,0xE0,0xFF,0x3F,
+        0x5E,0xE0,0x57,0x4E,0x03,0x05,0x09,0xF4,
+        0x2C,0x04,0x30,0xFE,0xFF,0x7F,0x16,0x02,
+        0xC8,0xB8,0x46,0x9D,0x85,0x80,0xE5,0x6D,
+        0xE5,0x19,0xDB,0xA7,0x95,0x04,0xFF,0xFF,
+        0x67,0x21,0xC0,0x41,0x2E,0x23,0x07,0x21,
+        0x4C,0xC4,0x87,0x83,0x8F,0x99,0x80,0x9E,
+        0x29,0xBE,0xB8,0x1B,0xE3,0x09,0xE0,0x45,
+        0xE2,0x31,0x93,0x1D,0x35,0x0D,0xF3,0x2C,
+        0x64,0xBC,0xB3,0x78,0x0D,0x78,0x82,0xF7,
+        0xE4,0x9F,0x85,0x18,0xD8,0x61,0x05,0x7B,
+        0x14,0x32,0xA8,0xC1,0x63,0x87,0x08,0x13,
+        0xE8,0x59,0x88,0xC5,0x7D,0xAE,0xE8,0x3C,
+        0xE1,0xB3,0x10,0xF0,0xFE,0xFF,0x9F,0x25,
+        0xE0,0x5E,0x0D,0x9E,0x85,0x00,0x13,0x87,
+        0x0D,0x9F,0x35,0xC0,0x33,0x7C,0x8F,0xEA,
+        0x1C,0x1E,0x8F,0x81,0x7F,0x56,0x1D,0xE7,
+        0x04,0x96,0x7B,0xD1,0xB2,0x71,0xA0,0xA1,
+        0x23,0xB2,0x3A,0x20,0x8D,0x0D,0x73,0x29,
+        0x89,0x7C,0x72,0x6C,0xD4,0x56,0x04,0xA7,
+        0x33,0x93,0x4F,0x00,0xD6,0x42,0x21,0x05,
+        0x34,0x1A,0x8B,0xE1,0x9D,0xF9,0xE8,0x44,
+        0x41,0x0C,0xE8,0xE3,0x90,0x6D,0x1C,0x0A,
+        0x50,0x7B,0xD1,0x14,0xC8,0x39,0x07,0xA3,
+        0x7F,0x76,0x74,0x36,0xBE,0x13,0x70,0x0D,
+        0x10,0x3A,0x25,0x18,0xDA,0x6A,0x04,0xFC,
+        0xFF,0x67,0x89,0x01,0x33,0xFE,0x53,0x8C,
+        0x09,0x7C,0x8E,0xC1,0x1F,0x0C,0xF0,0x03,
+        0x7F,0x31,0xA8,0xFA,0x5E,0xA0,0xFB,0x82,
+        0xD5,0xDD,0x64,0x20,0xCC,0xC8,0x04,0xF5,
+        0x9D,0x0E,0x40,0x01,0xE4,0x0B,0x81,0xCF,
+        0x51,0x0F,0x05,0x6C,0x22,0x21,0xC2,0x44,
+        0x33,0x3A,0x62,0xC2,0xA8,0xE8,0x13,0xA6,
+        0x20,0x9E,0xB0,0x63,0x4D,0x18,0x3D,0x13,
+        0x5F,0x74,0xD8,0x88,0x31,0x21,0xAE,0x1E,
+        0xD0,0x26,0x18,0xD4,0x97,0x22,0x58,0x43,
+        0xE6,0x63,0xF1,0x05,0x02,0x37,0x65,0x30,
+        0xCE,0x89,0x5D,0x13,0x7C,0xD9,0xC1,0xCD,
+        0x19,0x8C,0xF0,0x98,0xBB,0x18,0xBF,0x3A,
+        0x79,0x74,0xFC,0xA0,0xE0,0x1B,0x0E,0xC3,
+        0x7E,0x32,0xF3,0x8C,0xDE,0xCB,0x7C,0x8D,
+        0xC3,0xC0,0x7A,0xBC,0x1C,0xD6,0x68,0x61,
+        0x0F,0xED,0x3D,0xC4,0xFF,0xFF,0x43,0x8C,
+        0xCF,0x13,0xC6,0x08,0xEB,0xDB,0x0B,0x38,
+        0xEE,0x59,0xF0,0xEF,0x1A,0xE0,0xB9,0x84,
+        0xF8,0xAE,0x01,0x30,0xF0,0xFF,0x7F,0xD7,
+        0x00,0x4E,0xD7,0x04,0xDF,0x35,0x80,0xF7,
+        0xD0,0x7D,0xD7,0x00,0xAE,0xD9,0xEF,0x1A,
+        0xA8,0x63,0x80,0x15,0xDE,0x35,0xA0,0x5D,
+        0xD9,0xDE,0xD7,0x9E,0xB0,0xAC,0xE9,0xB2,
+        0x81,0x52,0x73,0xD9,0x00,0x14,0xFC,0xFF,
+        0x2F,0x1B,0x80,0x01,0x29,0x13,0x46,0x85,
+        0x9F,0x30,0x05,0xF1,0x84,0x1D,0xEC,0xB2,
+        0x01,0x8A,0x18,0x97,0x0D,0xD0,0x8F,0xED,
+        0x65,0x03,0x18,0xDC,0x13,0xF8,0x6D,0x03,
+        0x78,0x43,0xFA,0xB6,0x01,0xD6,0xFF,0xFF,
+        0x6D,0x03,0xAC,0xF9,0x6F,0x1B,0x28,0x0E,
+        0xAB,0xBC,0x6D,0x40,0x3C,0xC9,0x33,0x02,
+        0xAB,0xBA,0x6E,0xA0,0xF4,0x5C,0x37,0x00,
+        0x12,0x88,0x99,0x30,0x2A,0xFE,0x84,0x29,
+        0x88,0x27,0xEC,0x68,0xD7,0x0D,0x50,0x04,
+        0xB9,0x6E,0x80,0x7E,0x5E,0x09,0xFE,0xFF,
+        0xAF,0x1B,0xC0,0xE0,0xA2,0x80,0xB9,0x6F,
+        0x00,0x6F,0x58,0x7E,0xDF,0x00,0x7C,0xDC,
+        0xC4,0x31,0xF7,0x0D,0xC0,0xCC,0xFF,0xFF,
+        0xBE,0x01,0xB0,0xE7,0xA2,0x80,0xBB,0x6F,
+        0x00,0xEF,0x8B,0xB4,0xEF,0x1B,0x60,0xFE,
+        0xFF,0xDF,0x37,0xC0,0x28,0x6D,0xFD,0x1E,
+        0x1C,0x3D,0x21,0x78,0x7C,0xB8,0xFB,0xA5,
+        0xC7,0xE7,0xBB,0x39,0x38,0x06,0x79,0x8C,
+        0x87,0x76,0xC0,0xAF,0xEF,0x9E,0x98,0xEF,
+        0xE6,0xC0,0xFF,0x4C,0x70,0x3C,0x18,0x68,
+        0x1C,0x62,0xAB,0x97,0x06,0x72,0x34,0x38,
+        0x3F,0xDC,0x19,0x81,0x61,0x15,0x7F,0xF2,
+        0x47,0x38,0xC7,0xD0,0xD9,0xE1,0x20,0xB1,
+        0x83,0xE0,0xC1,0x56,0x6D,0x02,0x85,0x86,
+        0x50,0x14,0x18,0x14,0x8B,0x0F,0x18,0xF8,
+        0x61,0xB3,0xB3,0x00,0x93,0x04,0x87,0x3A,
+        0x02,0xF8,0x3E,0xD1,0xFC,0x38,0x74,0x37,
+        0x38,0x54,0x8F,0xE5,0xA1,0x80,0x9E,0x01,
+        0x71,0xC7,0x0C,0x32,0x69,0xCF,0x28,0xE2,
+        0x53,0xC2,0x29,0x85,0x49,0xE0,0xF3,0x03,
+        0x43,0xE3,0x04,0xAF,0x0D,0xA1,0xF9,0xFF,
+        0xFF,0xA4,0xC0,0x3C,0xDF,0x31,0x04,0x6C,
+        0x02,0xBB,0xBF,0x64,0xC8,0xDA,0xC0,0x75,
+        0x4B,0x32,0x44,0x6F,0x38,0xB2,0x85,0xA2,
+        0xE9,0x44,0x79,0xDF,0x88,0x62,0x67,0x08,
+        0xC2,0x88,0x12,0x2C,0xC8,0xA3,0x42,0xAC,
+        0x28,0x2F,0x05,0x46,0x88,0x18,0xE2,0x95,
+        0x23,0xD0,0x09,0x87,0x0F,0xF2,0xD8,0x14,
+        0xA7,0xFD,0x41,0x90,0x58,0x4F,0x02,0x8D,
+        0xC5,0x91,0x46,0x83,0x3A,0x07,0x78,0xB8,
+        0x3E,0xC4,0x78,0xF8,0x0F,0x21,0x06,0x39,
+        0xC8,0x73,0x7B,0x54,0x38,0x4E,0x5F,0x25,
+        0x4C,0xF0,0x02,0xE0,0x83,0x0A,0x1C,0xD7,
+        0x80,0x9A,0xF1,0x33,0x06,0x58,0x8E,0xE3,
+        0x3E,0xA9,0xC0,0x1D,0x8F,0xEF,0x07,0x6C,
+        0xC2,0x09,0x2C,0x7F,0x10,0xA8,0xE3,0x0C,
+        0x9F,0xE7,0x0B,0x8B,0x21,0x1F,0x13,0x4C,
+        0x60,0xB1,0x27,0x1B,0x3A,0x1E,0xF0,0xDF,
+        0x63,0x1E,0x2F,0x7C,0x32,0xF1,0x7C,0x4D,
+        0x30,0x22,0x84,0x9C,0x8C,0x07,0x7D,0x87,
+        0xC0,0x5C,0x6F,0xD8,0xB9,0x85,0x8B,0x3A,
+        0x68,0xA0,0x4E,0x0B,0x3E,0x28,0xB0,0x9B,
+        0x11,0xE6,0xB8,0xCE,0xCF,0x2A,0x60,0xF8,
+        0xFF,0x9F,0x55,0x60,0x8F,0x10,0xFE,0xED,
+        0xC1,0xF3,0xF2,0x95,0xE1,0xD5,0x21,0x81,
+        0x43,0x8E,0x10,0x3D,0x2E,0x8F,0x10,0x73,
+        0x3E,0xC2,0x0C,0x11,0x5C,0x67,0x01,0x70,
+        0x0C,0x11,0xF8,0x1C,0x70,0xC0,0x71,0x69,
+        0xE2,0x03,0xF5,0x01,0x07,0x70,0x70,0x4D,
+        0xC3,0x1D,0x70,0xC0,0x71,0x16,0x60,0xFF,
+        0xFF,0xC3,0x0D,0x2C,0x49,0x26,0x0E,0x23,
+        0x18,0x11,0x30,0x28,0x02,0x02,0xA4,0xB3,
+        0x80,0x0F,0x29,0x00,0x1F,0xAE,0x0C,0x0F,
+        0x29,0xD8,0x93,0x86,0x07,0x8E,0x1B,0x85,
+        0x07,0x8D,0x0B,0x30,0x68,0x7A,0xE2,0x80,
+        0x7F,0x4C,0xF0,0x19,0x05,0x1C,0xE3,0x06,
+        0xDF,0x2A,0x0C,0xFC,0xFF,0x3F,0x30,0xCC,
+        0xE1,0xC2,0x63,0x39,0x8A,0xA0,0x07,0x1E,
+        0xD4,0xF7,0x8C,0x33,0xF7,0x24,0x8F,0xD1,
+        0x51,0x0F,0x27,0xF4,0xE4,0x85,0x3B,0x57,
+        0xF9,0x0A,0x71,0x14,0x18,0xB8,0x77,0x29,
+        0x8F,0xCF,0x17,0x2B,0xC3,0x63,0x46,0xFB,
+        0x1E,0x72,0xD6,0x11,0x02,0xE2,0x2F,0x75,
+        0x6C,0xC0,0x60,0x39,0x18,0x00,0x87,0x01,
+        0xE3,0x13,0x0D,0x58,0x67,0x1B,0x3C,0xF4,
+        0x69,0x31,0xC4,0xE3,0x0B,0xFB,0x56,0x61,
+        0x82,0xEA,0x41,0x75,0x12,0xF4,0xD0,0xC0,
+        0x01,0xE8,0xA1,0xC1,0x3F,0xB9,0x90,0xFB,
+        0x2B,0x1D,0x82,0xB5,0xE2,0x69,0xDE,0x47,
+        0x1E,0xF3,0xDC,0xA2,0xBC,0x0D,0x3C,0x07,
+        0xF0,0xD3,0x82,0x87,0xE3,0x63,0x81,0xC7,
+        0xE9,0x4B,0x58,0x82,0xF7,0x1A,0x9F,0x6C,
+        0x1E,0x5C,0x58,0xB2,0x21,0xA0,0x06,0xEB,
+        0x21,0x60,0xA6,0x9A,0xC0,0x49,0x46,0x80,
+        0xCA,0x00,0xA1,0x1B,0xCB,0xE9,0x3E,0x8B,
+        0x84,0x38,0xCD,0x47,0x99,0xC7,0x02,0x8F,
+        0xF5,0xC1,0xC0,0xFF,0x7F,0xCD,0x23,0xD4,
+        0x7D,0xCD,0x33,0x7B,0x3A,0xC0,0xAC,0x22,
+        0xDC,0x7B,0xCE,0x1B,0x86,0xD1,0x9E,0x2D,
+        0x7C,0xCD,0x78,0xD6,0x34,0x42,0x38,0x76,
+        0x83,0xF3,0x48,0x8C,0xF0,0x82,0xC0,0x4E,
+        0x0C,0x0F,0x30,0xC6,0x39,0x79,0xC3,0xFA,
+        0xC2,0xCB,0x40,0x83,0x19,0xDB,0x97,0x01,
+        0x36,0x2A,0xDF,0x88,0xC0,0x97,0xFC,0x62,
+        0x00,0x65,0x16,0xBE,0x9E,0xF8,0xA0,0xC4,
+        0x2E,0x06,0x2C,0xE5,0xC5,0x00,0x54,0x37,
+        0x0C,0x5F,0x0C,0xE0,0x5F,0x89,0x5E,0x0C,
+        0xC0,0x70,0x71,0xF2,0x3D,0xC0,0x1E,0xEE,
+        0xA3,0x74,0x9C,0xBE,0xFD,0xBD,0x19,0xF8,
+        0x6C,0xC0,0x60,0x3C,0xC3,0x30,0xC6,0x08,
+        0xE3,0x51,0x86,0x31,0xC1,0xDC,0xB7,0x03,
+        0xE8,0x39,0x87,0x81,0x4A,0x78,0x3B,0x80,
+        0x72,0x0E,0xE8,0xF2,0x68,0x42,0x4F,0x01,
+        0x4F,0x07,0x3E,0x29,0x1A,0xA2,0xAF,0xB1,
+        0x0A,0x26,0x50,0xC4,0x07,0x0D,0x3E,0xB5,
+        0x28,0x3E,0x15,0x78,0x2D,0xCF,0x4E,0xE1,
+        0xE2,0x9C,0x89,0xA7,0x6A,0x38,0x03,0xBD,
+        0xE6,0x86,0x63,0xFF,0x7F,0x38,0xFC,0xA9,
+        0xE0,0x35,0x80,0x1D,0x24,0x3D,0x2D,0x23,
+        0xC2,0x38,0xA4,0x3C,0x32,0xF8,0xB6,0x18,
+        0xC7,0x90,0x0F,0x91,0xBE,0x13,0x18,0xF2,
+        0x21,0xEF,0x79,0xC7,0xC0,0xAF,0x08,0x71,
+        0x9E,0xB2,0x7C,0x67,0xF0,0x65,0x01,0x7C,
+        0x91,0x2E,0x0B,0x68,0x68,0x9F,0x64,0x7C,
+        0x41,0x30,0xEC,0x89,0xB3,0x00,0x77,0x05,
+        0x50,0x81,0xFA,0xAE,0x00,0xFF,0x42,0xF0,
+        0xAE,0x00,0x86,0x79,0xF9,0x56,0xC0,0x35,
+        0x1D,0x4A,0xD0,0x67,0x12,0x5F,0x17,0x70,
+        0x53,0x64,0xA9,0x8E,0x0A,0xD0,0x53,0x4C,
+        0x02,0x75,0x47,0xF7,0x51,0x01,0xC6,0x4D,
+        0xD9,0x07,0x54,0x76,0x5A,0x60,0x67,0x21,
+        0x76,0x1D,0xC1,0x5D,0x49,0x18,0xCA,0xB3,
+        0x81,0x2F,0x59,0xFC,0x70,0x00,0x03,0xDC,
+        0xB3,0x38,0xC4,0x08,0xB1,0xD9,0x81,0xEB,
+        0x75,0xD2,0x70,0x2F,0x44,0xEC,0xFF,0x7F,
+        0x32,0x00,0xE3,0x51,0x1B,0x1C,0x27,0x9D,
+        0xF0,0x91,0x9E,0x59,0xF8,0x49,0x19,0x30,
+        0x71,0xF2,0x03,0xE3,0xC9,0x1A,0xC6,0x00,
+        0xB8,0xBC,0x57,0x95,0x81,0xFC,0x43,0x90,
+        0x20,0x18,0xD4,0x29,0x19,0x38,0x1C,0xC5,
+        0x70,0xA7,0x64,0x78,0x50,0xF8,0xC3,0x00,
+        0xE6,0x46,0xE8,0x7B,0x82,0xA1,0xDE,0x93,
+        0x0E,0xE3,0x91,0xD0,0x04,0x3E,0x2D,0xC3,
+        0xFA,0xFF,0x9F,0x96,0x81,0xD5,0xB1,0xDD,
+        0x43,0xF6,0x59,0x01,0x77,0x76,0x80,0x3B,
+        0x3D,0x7E,0x7A,0x00,0x9C,0x00,0x3D,0x3D,
+        0x80,0xED,0xBC,0x01,0xF7,0x40,0x80,0x38,
+        0xFE,0xA3,0x82,0x5F,0x59,0x28,0x1C,0x3F,
+        0xB6,0xF3,0x63,0x09,0xEE,0x70,0xE0,0x23,
+        0x83,0x0F,0x90,0xB8,0xA1,0xF8,0x50,0x81,
+        0x3C,0x0B,0x80,0x62,0xF4,0x6C,0x04,0xEC,
+        0x06,0xF3,0xD2,0x12,0xE5,0xFF,0xFF,0xDE,
+        0xC0,0x4E,0x29,0xB8,0x83,0x00,0xF8,0x8E,
+        0x01,0xE0,0x1D,0x0C,0x97,0x35,0x66,0x94,
+        0x10,0x18,0x8D,0x19,0x77,0x08,0xE1,0x27,
+        0x02,0xDC,0x98,0x3D,0x6E,0x8F,0x19,0x77,
+        0x9C,0xE5,0xA3,0x7A,0xCA,0x08,0xE5,0x03,
+        0x07,0x3B,0x67,0xBC,0x11,0xF0,0xA1,0x03,
+        0x8F,0x03,0x0C,0xEE,0x48,0x01,0xC6,0xCB,
+        0x01,0x1B,0x3B,0xB8,0x83,0x90,0x53,0x20,
+        0x4B,0x87,0xD1,0xD8,0x71,0xB2,0x81,0x74,
+        0x8C,0xF1,0x21,0xD7,0x63,0xC7,0x0D,0xD6,
+        0x63,0xC7,0x1D,0x5F,0xB0,0xFF,0xFF,0xE3,
+        0x0B,0x18,0xC6,0xC0,0xC5,0x0F,0x03,0x7D,
+        0xF3,0xF3,0xE8,0x0C,0xEE,0x61,0xFB,0x04,
+        0x13,0xE3,0xF9,0x25,0xC4,0x23,0xCC,0x8B,
+        0x4B,0x84,0xA3,0x08,0xF2,0xE6,0x12,0xE7,
+        0xD5,0x20,0xCC,0x63,0x4B,0x94,0x10,0x11,
+        0x0E,0x26,0xCE,0x13,0x8C,0x11,0x0E,0x3C,
+        0x8A,0x21,0x22,0x9C,0x40,0x88,0x93,0x3E,
+        0xD9,0x20,0xE1,0x63,0x84,0x8D,0xF6,0x04,
+        0xC3,0xC7,0xC2,0xCF,0x2B,0x1E,0x3C,0x3F,
+        0xAD,0xF9,0x2E,0xE8,0xC9,0x9C,0xE3,0x43,
+        0x96,0xA7,0xF6,0x38,0xE9,0xC3,0x2C,0x6E,
+        0x50,0x0F,0x8E,0xEC,0xAE,0xE3,0xE3,0x35,
+        0xF6,0x14,0xE4,0x21,0xF0,0x13,0x81,0x2F,
+        0x88,0x9E,0xAC,0xEF,0x7A,0xEC,0x5E,0x66,
+        0x8C,0xEA,0xA7,0x80,0x3A,0xA6,0x9C,0xC1,
+        0x2B,0x04,0xBB,0xE7,0xF9,0x90,0xED,0xBB,
+        0x24,0x1B,0x05,0xEE,0x90,0xE0,0x33,0x12,
+        0x3F,0x55,0x78,0x18,0x1E,0x05,0x8C,0x19,
+        0xBC,0x23,0x1C,0x5A,0x88,0x03,0x7E,0xDF,
+        0x65,0x43,0x8D,0x71,0x7A,0x3E,0x7F,0xB0,
+        0x41,0xC0,0x87,0x3A,0x54,0x0F,0xF3,0xA8,
+        0x5E,0x0A,0x19,0xCE,0xD9,0xC1,0x1D,0x04,
+        0xF6,0xF8,0xE1,0x41,0xF0,0x9B,0x25,0x1F,
+        0x04,0x3B,0xDF,0xBC,0xC1,0x19,0xE4,0xFF,
+        0x7F,0x0C,0xB0,0xCF,0x54,0x3E,0x9A,0x20,
+        0x8E,0x80,0xE8,0xF3,0x87,0xC7,0xF0,0x26,
+        0xC7,0x87,0x83,0x3D,0x7A,0xE0,0x4E,0x22,
+        0x70,0x8F,0x5D,0x07,0xED,0x6B,0x9C,0x2F,
+        0x5A,0x30,0xEE,0x7B,0xCF,0x22,0xE0,0xC7,
+        0x78,0x6C,0x01,0xC7,0xA1,0x04,0xDC,0xC1,
+        0x8E,0x6B,0x1C,0x42,0x51,0x60,0x74,0x28,
+        0xC1,0xC5,0x00,0x12,0x8C,0x63,0x9C,0xD1,
+        0xD0,0x97,0x48,0x1F,0xD2,0xE0,0x0C,0x1A,
+        0xF6,0x3C,0x9F,0x50,0xB8,0x3D,0x01,0x8A,
+        0x4E,0x28,0x20,0xC3,0x7D,0x06,0xC1,0x9E,
+        0x10,0xF8,0x19,0x84,0xFD,0xFF,0x0F,0x8E,
+        0x1E,0xF7,0x7B,0xA3,0x4F,0x8D,0x6C,0xEE,
+        0x0F,0x01,0x27,0x70,0xEE,0xEC,0xD4,0x8C,
+        0x3B,0x33,0x60,0xCF,0x1F,0x1E,0x02,0x3F,
+        0x17,0x78,0xF8,0x1E,0x02,0x7E,0xF0,0x0F,
+        0xCC,0x06,0x07,0xE3,0x29,0xC2,0xD7,0x0E,
+        0x0E,0xCE,0x4F,0x03,0x06,0xE7,0xAF,0x50,
+        0x9F,0xE7,0x19,0x38,0xF6,0xD4,0xEB,0x7B,
+        0x87,0xE7,0xEB,0x43,0x05,0xFE,0xA6,0xE7,
+        0x43,0x05,0x38,0x0E,0x0F,0xFC,0xB0,0xC2,
+        0x86,0xF0,0x28,0x80,0x3F,0xB5,0xF8,0xF8,
+        0x17,0xE7,0x29,0x82,0xDD,0x46,0xB0,0x87,
+        0x0B,0xC0,0x51,0xB4,0xB3,0x18,0x2A,0xCC,
+        0x59,0x8C,0xFC,0xFF,0xCF,0x51,0xA8,0xB3,
+        0x18,0x3D,0x5C,0x00,0x2E,0x04,0x1F,0x0F,
+        0x40,0x73,0x10,0x78,0x5C,0xF0,0x85,0xE0,
+        0x48,0x0E,0xE4,0xE9,0x00,0xF0,0x19,0x4A,
+        0xC3,0xA1,0x09,0x13,0x03,0x06,0x75,0x3E,
+        0xF0,0x09,0xC5,0xC7,0x0E,0x7E,0x36,0xF0,
+        0x8D,0xDC,0x43,0xE5,0xA7,0x66,0x5F,0xF2,
+        0x11,0xE0,0x02,0x75,0xA0,0x61,0xA0,0x46,
+        0xE4,0x23,0xD2,0xFF,0xFF,0xB9,0x0D,0x1B,
+        0x60,0x68,0xF4,0x1C,0x0E,0xE3,0x80,0xEB,
+        0x73,0x38,0x76,0x40,0x3E,0x87,0xC3,0x3F,
+        0x47,0xC3,0x1F,0x1B,0x3B,0xDD,0xF3,0x81,
+        0xC1,0xBA,0x7E,0x63,0x06,0x06,0xB6,0x6F,
+        0x91,0x07,0x06,0x1C,0x51,0xCF,0xC6,0x57,
+        0x08,0x0F,0x0C,0x6C,0x80,0x1E,0x18,0xF0,
+        0x89,0x05,0x21,0x27,0x03,0x43,0x9D,0x32,
+        0x8C,0x1C,0xF3,0x89,0xC3,0xC3,0xF0,0xA1,
+        0x22,0xEA,0x33,0xC0,0x23,0x1E,0x1B,0x1B,
+        0xFB,0xFF,0x8F,0x0D,0x2C,0xC7,0x16,0x8F,
+        0x0D,0xFC,0x47,0x78,0xFC,0xD8,0xE0,0x8C,
+        0xE5,0xD1,0xC4,0x97,0x99,0x23,0x3B,0x8D,
+        0x33,0x7B,0x0D,0xF1,0xD1,0xEE,0xF1,0xDB,
+        0x63,0x03,0x97,0x85,0xB1,0x01,0xA5,0x90,
+        0x63,0x43,0x1F,0x52,0x7C,0x0A,0xB0,0x71,
+        0x54,0x32,0x0F,0x1F,0xAF,0x7C,0x62,0x38,
+        0xBA,0x20,0x6F,0xE8,0xBE,0x5C,0xF8,0x48,
+        0x63,0x30,0x5F,0x5A,0x7C,0x06,0xE5,0x43,
+        0x04,0xD7,0x57,0xC5,0x43,0x04,0x3E,0xA1,
+        0x86,0x88,0x1E,0xCF,0xFF,0xFF,0x11,0xCC,
+        0x43,0x64,0x43,0x03,0xAF,0x87,0xA1,0x01,
+        0xA5,0x98,0xC0,0x5E,0x85,0x87,0x46,0x4F,
+        0x3F,0x3E,0x04,0x30,0x08,0xDF,0x06,0xD8,
+        0x55,0xC0,0x57,0x21,0x83,0x24,0x18,0xE7,
+        0x64,0x41,0x07,0x07,0x8E,0x21,0x79,0x70,
+        0xF0,0x07,0xE3,0x21,0x70,0x60,0xCF,0xE0,
+        0xB9,0xE8,0x31,0xD8,0xA7,0x1D,0x9F,0x4A,
+        0xC0,0x77,0xE6,0x04,0xC7,0xE9,0x1D,0x7B,
+        0x29,0xF0,0x08,0x1E,0xAD,0x3C,0x02,0x7E,
+        0xB4,0x02,0x66,0xFF,0xFF,0xA3,0x15,0x30,
+        0x09,0x7A,0xE6,0xA4,0x03,0x77,0x34,0x18,
+        0xD4,0xD1,0x0A,0x5C,0x11,0xC0,0x75,0xDC,
+        0xF0,0xD1,0x02,0xCE,0x50,0x0F,0xDA,0x07,
+        0x65,0xCF,0xDA,0x97,0x21,0x76,0xB4,0x00,
+        0x97,0x89,0x43,0x08,0xD0,0x04,0x3E,0x89,
+        0x67,0xEF,0x43,0x03,0xB3,0x8A,0xA1,0x01,
+        0xA5,0xA3,0x01,0xEE,0x44,0x81,0xFD,0xFF,
+        0x9F,0x28,0x60,0xDE,0x30,0x70,0x07,0x0A,
+        0xC0,0xCD,0xE9,0xDB,0xE3,0xE2,0xD0,0x38,
+        0xC4,0xE7,0xA7,0x73,0xF6,0xD1,0xE8,0x4C,
+        0x71,0x67,0x11,0x30,0x9C,0x7D,0x11,0x8F,
+        0x18,0x03,0xF9,0x81,0x21,0x59,0x30,0x28,
+        0x16,0x0F,0xC5,0x07,0x03,0x0E,0xEC,0x23,
+        0x02,0x3B,0x17,0xB0,0x73,0xAD,0xE1,0xF8,
+        0x59,0xC0,0xA7,0x84,0xB7,0xA6,0x17,0x7B,
+        0x9F,0xD7,0x7D,0xD6,0x08,0xC9,0xCE,0xF4,
+        0x3E,0x89,0xE2,0x0E,0xA2,0x70,0x4E,0x9F,
+        0xE0,0x22,0xF0,0x65,0xDF,0xA3,0xE0,0xA7,
+        0x07,0xCF,0xF1,0x8D,0xC1,0xA7,0x07,0xE6,
+        0x7E,0xF8,0x9A,0xF1,0x33,0xC3,0xE3,0x43,
+        0x88,0x27,0xE2,0xDA,0xA6,0x20,0x5B,0x18,
+        0x42,0x09,0xF4,0xFF,0x8F,0x10,0xE5,0x6D,
+        0x20,0xCA,0x29,0x44,0x88,0x12,0xA4,0xB1,
+        0xC9,0x0B,0x35,0xCA,0xD9,0x45,0x6E,0x6D,
+        0xF6,0x82,0x0B,0x14,0x2A,0x66,0x9C,0x28,
+        0xEF,0x10,0xB1,0xDA,0x1F,0x04,0x91,0xF4,
+        0x32,0xD0,0x71,0xC9,0x91,0x0E,0x7D,0xE8,
+        0x61,0xFB,0x04,0x8C,0x3F,0x48,0xE2,0xAE,
+        0x2A,0x3E,0x28,0xF8,0x00,0x80,0x77,0x09,
+        0xA8,0x5B,0x9D,0xC7,0xED,0xF3,0x06,0xF8,
+        0xAF,0x17,0x58,0x82,0xF2,0x07,0x81,0x1A,
+        0x99,0xA1,0x3D,0xCC,0xB7,0x19,0x43,0xBE,
+        0x07,0x1C,0x16,0x3B,0x27,0xF9,0xF0,0x08,
+        0x1C,0x8E,0x01,0x4F,0x1B,0xBE,0x51,0x7B,
+        0xBE,0x3E,0x62,0x01,0x8E,0xFE,0xFF,0x47,
+        0x2C,0x30,0x9D,0xDF,0x7D,0x82,0x01,0xC7,
+        0xCD,0x82,0x9F,0x61,0x00,0x67,0x40,0xCF,
+        0x30,0x60,0x1F,0x2A,0x6E,0x08,0x5C,0xEE,
+        0x8A,0x28,0x90,0x05,0xC2,0xA0,0x0E,0xFD,
+        0xE4,0x08,0x42,0xCF,0x9C,0x70,0x86,0x72,
+        0xB2,0xBD,0x5F,0x1D,0xC8,0x2D,0xC2,0x43,
+        0x3D,0x8B,0xC7,0x04,0x76,0xDA,0x02,0x36,
+        0xFF,0xFF,0xE3,0x29,0xB0,0x98,0xF7,0xD3,
+        0x69,0x84,0x63,0x03,0xFB,0x71,0x0B,0x38,
+        0x1D,0xCC,0xE0,0xDC,0x7F,0xD8,0x2D,0x1A,
+        0x37,0x34,0xB0,0x0D,0xCC,0x43,0x03,0x3E,
+        0x27,0x47,0x30,0x9E,0x98,0xF8,0x55,0xE2,
+        0xE1,0x89,0x1F,0x43,0xC0,0xFA,0xFF,0x3F,
+        0x99,0x01,0xF6,0x84,0x1E,0xCB,0x50,0xD2,
+        0x4E,0x66,0x80,0xC0,0xFB,0xD8,0x3B,0xC3,
+        0x4B,0x83,0xE7,0x74,0xD2,0xCF,0x62,0x3E,
+        0x99,0x19,0x21,0x0A,0xBB,0x8F,0x19,0xAD,
+        0x37,0x14,0xCD,0x3C,0xE8,0x3B,0x99,0x51,
+        0x62,0x46,0x6A,0x0E,0x4C,0x48,0x11,0x0F,
+        0x27,0x4A,0x88,0x60,0xAF,0x13,0x6F,0x67,
+        0x4F,0x66,0x4C,0xD6,0xC9,0x0C,0x24,0xFF,
+        0xFF,0x93,0x19,0x98,0x5C,0x9F,0xCC,0x80,
+        0xCA,0x39,0x0A,0x7F,0x32,0x03,0x78,0x74,
+        0xC0,0xC2,0x9D,0xCC,0xC0,0xF2,0xFF,0x3F,
+        0xC4,0x00,0xCE,0xC7,0x0A,0x63,0x0C,0x3C,
+        0xDA,0xC1,0x0C,0x15,0xE6,0x6C,0x86,0x0E,
+        0x72,0x08,0xA1,0xC1,0x0E,0x21,0x50,0xE6,
+        0x72,0xA0,0xA7,0xF0,0x9A,0xE0,0x73,0x14,
+        0xD8,0x0F,0x67,0xC0,0xE1,0xD4,0x80,0x0F,
+        0x74,0xE2,0x42,0x8F,0xC2,0x23,0x0E,0x58,
+        0xFD,0xC0,0xC8,0xFF,0xFF,0x64,0x06,0x18,
+        0x78,0x6A,0xF8,0x40,0x82,0x63,0x31,0xEA,
+        0x1B,0xC4,0x21,0xBE,0x8D,0xF8,0xE8,0xFE,
+        0x6A,0xE2,0x4B,0x00,0xE6,0x42,0xE2,0xD3,
+        0x09,0xB3,0x70,0x38,0x03,0x5A,0x43,0x60,
+        0x57,0x26,0xCF,0x9C,0x0F,0xE1,0x6C,0x3C,
+        0x7A,0xDC,0xE9,0x04,0xDE,0x38,0x7C,0x3A,
+        0x01,0x5E,0x07,0x0C,0xCC,0x0C,0xC2,0x3F,
+        0x84,0xB0,0x21,0x9C,0xAA,0xC7,0x70,0xEE,
+        0xAF,0x38,0x3E,0x9D,0x80,0xF3,0xFF,0x7F,
+        0x62,0x03,0x0C,0x0A,0x7E,0x32,0xF8,0xB8,
+        0x46,0x25,0xC2,0xA0,0x8E,0xE6,0x80,0x7B,
+        0x98,0x27,0x36,0x26,0x6F,0xC5,0x1A,0x8B,
+        0x4F,0x6C,0x30,0xFF,0xFF,0x27,0x36,0x80,
+        0xD1,0x87,0x20,0xB0,0xFD,0xFF,0x0F,0x41,
+        0x60,0x1C,0xA0,0x0F,0x41,0x80,0x9B,0xD3,
+        0x09,0xEE,0xC4,0x07,0xB6,0x63,0x10,0x60,
+        0x6D,0xE8,0x3E,0x06,0x81,0xF9,0xFF,0x3F,
+        0x5A,0x98,0xA3,0xE0,0xC2,0x8E,0x7C,0x28,
+        0x29,0xA7,0x3E,0xB4,0x0C,0x20,0x69,0x38,
+        0xC9,0x01,0x9D,0xD3,0x3D,0x70,0x92,0x75,
+        0xEA,0x40,0x8F,0xC7,0xA0,0xAF,0x1C,0xBE,
+        0x12,0xF0,0x23,0x07,0x93,0x00,0xAA,0x41,
+        0xFA,0xCC,0x07,0x9C,0x8E,0x1C,0xE0,0x38,
+        0x26,0x05,0xC6,0xDE,0x0E,0xDE,0x22,0x3D,
+        0x89,0xA7,0xA1,0xE3,0x0C,0x51,0x38,0x26,
+        0x39,0x18,0x44,0x7A,0x95,0x62,0x03,0x7C,
+        0xAB,0xF1,0xD9,0xC8,0x07,0x10,0x78,0xE3,
+        0xF6,0xD8,0x61,0xFF,0xFF,0x0F,0x75,0xC0,
+        0x01,0xE2,0xA4,0xF8,0x21,0xC3,0x98,0x67,
+        0xC5,0x0F,0x75,0x80,0xF5,0x18,0x27,0x3A,
+        0x94,0xF0,0x43,0x1D,0x20,0xE8,0xFF,0x7F,
+        0xA8,0x03,0x86,0x38,0x6F,0x24,0xD1,0x1E,
+        0xEA,0x98,0xE8,0x43,0x1D,0x40,0xC8,0xFF,
+        0xFF,0xA1,0x0E,0x18,0x9E,0x87,0x00,0xAE,
+        0x9C,0xEF,0xC0,0x7C,0x22,0x02,0xEF,0xFF,
+        0xFF,0x7C,0x07,0xB8,0x1B,0x2D,0xCC,0x51,
+        0x70,0x41,0xAF,0x0E,0x03,0x51,0x09,0x30,
+        0x28,0x02,0xC7,0x5F,0x9B,0x60,0x1C,0xEA,
+        0x7C,0x87,0x3E,0x2F,0x78,0xD8,0x4F,0x05,
+        0x9E,0xC4,0xA9,0xFA,0x5A,0x70,0x14,0x4F,
+        0x00,0x3E,0xE1,0x01,0xFF,0xA1,0xC1,0x9A,
+        0x44,0xF1,0x43,0x03,0xF5,0x11,0xE4,0xFF,
+        0x7F,0x68,0xC0,0x28,0xEA,0xF9,0x06,0x7D,
+        0xCC,0xF2,0xD9,0x20,0xE6,0x0B,0x48,0x84,
+        0x07,0x10,0x5F,0x1F,0xD8,0x71,0xD2,0x67,
+        0xA0,0x40,0x51,0xDE,0x37,0xF8,0x09,0x07,
+        0x5C,0x83,0xF3,0x09,0x07,0xBC,0x87,0x23,
+        0x1F,0x4B,0xC0,0x77,0xD0,0x84,0x73,0x81,
+        0xF1,0x8D,0x8D,0x9D,0x06,0xC0,0x76,0x00,
+        0x06,0xDF,0x69,0x00,0x1C,0xC7,0x24,0x7E,
+        0x3A,0x04,0x13,0xCC,0xC1,0xBC,0x34,0xFB,
+        0xFF,0xEF,0xFD,0x94,0x43,0xCF,0x86,0x80,
+        0x75,0x49,0x07,0x43,0x94,0x88,0xB3,0x21,
+        0x20,0xFD,0xFF,0x7F,0x36,0xC4,0x20,0xC4,
+        0x09,0xFC,0x12,0xD1,0xDC,0xD9,0x90,0xAE,
+        0xD8,0x67,0x43,0x80,0xE1,0xFF,0xFF,0x23,
+        0x00,0xF6,0x7C,0x04,0x38,0x3D,0x64,0x83,
+        0xE7,0x14,0x08,0xE3,0xE4,0x03,0x38,0xFE,
+        0xFF,0x8F,0x15,0xE6,0x18,0x78,0xEA,0x97,
+        0x9B,0x8F,0x03,0x54,0xD4,0x2B,0xC2,0x30,
+        0x94,0xC5,0x87,0x05,0x1F,0x11,0xF8,0x61,
+        0xC1,0x23,0xA8,0x78,0x9C,0xF4,0x74,0xE3,
+        0x33,0x21,0x3B,0x24,0x38,0xFC,0x20,0xE9,
+        0x41,0x13,0x3C,0xE7,0x23,0x78,0xB7,0x1E,
+        0x38,0xA7,0x02,0xC0,0x4D,0xAE,0x27,0xA3,
+        0x4E,0x17,0x0E,0x70,0x8E,0x92,0x8D,0x63,
+        0x08,0xE5,0x70,0xCC,0xB7,0x87,0xA6,0xC9,
+        0x4E,0x56,0x30,0x63,0x41,0xEA,0x24,0xE0,
+        0x01,0x38,0x10,0x8C,0xB4,0x93,0x68,0x34,
+        0x86,0xB3,0x5A,0x18,0xC1,0x19,0xC4,0xC7,
+        0x11,0xE7,0x3A,0x19,0xA1,0x3F,0x07,0x3E,
+        0x15,0x61,0x82,0xDC,0x4B,0xE8,0xBC,0x7D,
+        0x37,0xE0,0x57,0x61,0x8F,0xC5,0xFF,0x7F,
+        0x60,0xDF,0x4E,0xC0,0x31,0x17,0xAB,0x01,
+        0x45,0x0D,0xC0,0x68,0x98,0x53,0xC0,0x53,
+        0x09,0xB8,0x82,0xCD,0x0D,0x7D,0x61,0xB1,
+        0xD6,0xA9,0xE8,0x14,0xF4,0x3E,0x70,0x70,
+        0xC0,0x63,0xF6,0x1E,0x1C,0x2C,0x34,0x0F,
+        0x0E,0x6C,0xD9,0x06,0x87,0x56,0x72,0x17,
+        0x21,0x87,0x0F,0xFC,0xEC,0x80,0x03,0xA0,
+        0x67,0x07,0x0B,0xC9,0xB3,0x03,0x9B,0xBE,
+        0xB3,0x08,0x28,0x70,0xFE,0xFF,0x11,0xDE,
+        0x3B,0x7C,0x6E,0x79,0xF6,0x60,0x63,0x78,
+        0x74,0x31,0x9A,0xD1,0xB9,0xA6,0xDB,0x04,
+        0x4A,0xC5,0x6D,0x82,0x82,0xF8,0x06,0xE0,
+        0x84,0x34,0xBA,0x75,0xE2,0x66,0x62,0xFC,
+        0x47,0x0C,0x1F,0x11,0x0E,0xE9,0x6C,0x4D,
+        0x30,0x0F,0xA4,0x9E,0x81,0xBE,0xB3,0xE1,
+        0x67,0x1F,0xF2,0xC1,0xC5,0xD3,0xF0,0xF5,
+        0x86,0xDC,0x3B,0xE8,0xB4,0x7D,0x66,0xC0,
+        0x1C,0x74,0x7D,0x9D,0x7A,0x83,0x27,0x57,
+        0x09,0xEA,0xE1,0x02,0x42,0x2F,0x34,0xBE,
+        0xDC,0x25,0x78,0xE0,0xF4,0xE9,0xEE,0xBD,
+        0x84,0x9D,0xF1,0x12,0xBC,0xE0,0x25,0x98,
+        0x77,0x10,0xA8,0x51,0x79,0x10,0x98,0xAB,
+        0x3C,0xCB,0x37,0x06,0x54,0xB2,0x8B,0x16,
+        0x3D,0xC3,0xBC,0xC3,0xF8,0x92,0xE0,0xEB,
+        0x87,0xCF,0x2D,0x5E,0xC0,0xEB,0x16,0x0C,
+        0x82,0x67,0xA0,0x57,0x17,0xDF,0xD9,0x0D,
+        0xFC,0x2A,0xF0,0x46,0x13,0x22,0x98,0x61,
+        0x0F,0xFF,0xDD,0xDD,0xA8,0xBE,0xE9,0x18,
+        0xEB,0x75,0xC4,0x23,0xE5,0xC7,0x96,0x03,
+        0x8A,0xF4,0xF2,0xE6,0x09,0xF8,0x2C,0xE3,
+        0x53,0xDD,0x49,0xF9,0x7A,0x68,0xF4,0x57,
+        0x08,0x1F,0x7E,0x8C,0xEC,0x73,0x0E,0x3B,
+        0xDF,0xB1,0x41,0x71,0xC4,0x07,0x86,0x97,
+        0x1A,0x4F,0x85,0x9D,0xBB,0x60,0x1C,0x1C,
+        0xD8,0xB1,0x08,0x73,0x7C,0x05,0xD7,0xC9,
+        0xE6,0xFF,0xFF,0xE4,0x00,0x6E,0x78,0xCC,
+        0xC1,0xD7,0xE7,0x0D,0xDF,0x0C,0x3C,0x2E,
+        0x7E,0xE4,0xF0,0x49,0xE3,0xA5,0xD3,0xD8,
+        0xA7,0xE9,0xA3,0xD1,0xCB,0x9B,0x4F,0x2F,
+        0x18,0x58,0x5F,0x1A,0x38,0xAC,0xD1,0xC2,
+        0x3E,0x06,0x9C,0xB9,0x2F,0x44,0xB8,0xC3,
+        0x23,0x58,0x00,0xF1,0xB7,0x92,0x47,0x0E,
+        0x4F,0xC0,0x80,0x4C,0xD3,0xBA,0x74,0x20,
+        0xE2,0xA7,0x3C,0x2B,0x5F,0x99,0x2E,0x43,
+        0x0C,0xE3,0xA9,0xF2,0xF1,0xC3,0xB3,0xF1,
+        0x51,0xC0,0xC7,0x28,0xCF,0xFC,0x8C,0x22,
+        0xBD,0x32,0x10,0x50,0x9D,0x88,0xB8,0x42,
+        0x18,0x89,0xA1,0xD1,0x9D,0x83,0xC7,0x1F,
+        0x22,0x05,0x31,0xA0,0x6F,0x2E,0xC0,0xF4,
+        0x4C,0x04,0x5C,0xFE,0xFF,0x37,0x17,0x80,
+        0xFF,0xFF,0xFF,0x9B,0x0B,0xE0,0xE6,0xFE,
+        0xE0,0x9B,0x0B,0x70,0x8D,0xB4,0x2A,0x7A,
+        0x61,0x77,0x08,0x18,0xD4,0x9D,0x1D,0x70,
+        0x78,0x2B,0x78,0x67,0x87,0xF5,0xFF,0xBF,
+        0xB3,0xC3,0xC3,0x8C,0x13,0xE5,0x85,0x21,
+        0xC6,0x3B,0x3B,0x0B,0xF0,0x26,0xD0,0x51,
+        0xC6,0x77,0x76,0x80,0x1F,0x67,0xD8,0x77,
+        0x69,0xF0,0x5E,0x75,0x81,0xF5,0xFF,0xFF,
+        0xAA,0x0B,0x3C,0x04,0xDF,0xA7,0x41,0x3E,
+        0x5E,0x30,0x8C,0x83,0x2B,0x27,0xA1,0xC7,
+        0x02,0x6B,0x85,0x41,0xDD,0xA9,0xC1,0xA5,
+        0x09,0x5C,0x17,0x5F,0x1F,0x6A,0x7C,0xA4,
+        0xC5,0x9F,0x2F,0x70,0x01,0x86,0x4C,0x4F,
+        0x65,0x30,0xAE,0x29,0x3E,0x95,0x61,0xEE,
+        0x0E,0x1E,0x90,0x8F,0x18,0xC0,0x67,0x15,
+        0x1E,0x18,0xEE,0xB4,0xE0,0x9B,0x92,0x41,
+        0xCF,0x31,0xA8,0x8F,0x3C,0x27,0xEF,0x7B,
+        0xC2,0xE3,0x84,0xA3,0x9E,0x83,0xE8,0xD8,
+        0xC0,0x71,0xDC,0xC0,0xFD,0xFF,0xC7,0x06,
+        0xEF,0x70,0x83,0x3B,0xE8,0xF8,0x62,0x70,
+        0x5C,0x18,0xB8,0xE7,0x02,0x0F,0xC3,0x37,
+        0x1D,0x8F,0x08,0x33,0xFE,0xD7,0x3F,0x23,
+        0x04,0xC4,0x5F,0x8C,0xD8,0x80,0xC1,0x78,
+        0x6B,0xF3,0xF5,0x0D,0x37,0x60,0x5F,0x1D,
+        0x7C,0xC1,0xF0,0x09,0xCC,0xE8,0x2F,0x30,
+        0x4F,0x62,0x3E,0x36,0x90,0x0B,0x1C,0x1D,
+        0x30,0x38,0x00,0x3D,0x60,0xF8,0x87,0x8B,
+        0x77,0x39,0x30,0x5C,0x05,0x7D,0x5C,0xF0,
+        0xB1,0xC7,0x8A,0xEE,0x72,0xE8,0x9B,0x9C,
+        0x61,0xE2,0x18,0xE2,0x0D,0x8C,0xDD,0x25,
+        0xC8,0x61,0x0E,0xEA,0x5D,0xC2,0x73,0xE0,
+        0x67,0x0B,0x9F,0xE0,0x7C,0xF3,0x09,0x71,
+        0xAA,0x8F,0x56,0xEF,0x01,0x3E,0x7A,0xBC,
+        0x77,0xF9,0xEC,0xC4,0x2E,0x02,0x3E,0x72,
+        0x19,0xC7,0xD3,0xF4,0x15,0xD0,0x43,0x36,
+        0xD8,0xAB,0x86,0x4F,0x60,0x3E,0xBA,0xE1,
+        0x8E,0x51,0x9E,0x89,0xA7,0xEF,0x3B,0x08,
+        0x3B,0x92,0x1C,0x75,0xA8,0x6B,0x7A,0x44,
+        0xF9,0xFF,0x9F,0xD0,0x81,0xF8,0xD6,0x06,
+        0xCE,0x68,0xF7,0x0F,0xF4,0x36,0x3D,0x32,
+        0xCC,0xD1,0x00,0xD6,0x25,0x04,0x5C,0x77,
+        0x0C,0x5F,0x42,0x80,0x4F,0xD0,0x4B,0x04,
+        0xFA,0x9A,0xE1,0xD1,0x3D,0x02,0x60,0xAE,
+        0x18,0xEC,0x58,0xE0,0xC3,0x86,0xAF,0x01,
+        0xEC,0x5E,0xE0,0x30,0xF7,0x08,0x50,0x81,
+        0x7A,0x78,0xF0,0xD5,0xDE,0x23,0x40,0x71,
+        0xB2,0xF4,0xA1,0xC1,0x03,0xB5,0xAA,0x33,
+        0x26,0x94,0x23,0x26,0x3F,0x9B,0xF9,0x26,
+        0x81,0xB9,0x5D,0xFA,0x26,0x01,0x37,0xCF,
+        0x2C,0x50,0x49,0x20,0xF4,0xFF,0xBF,0x49,
+        0xC0,0x85,0xE9,0xF2,0x32,0x43,0xE7,0x7F,
+        0xE0,0xBE,0xD5,0x79,0x84,0x3E,0x44,0x30,
+        0x94,0xF7,0x3C,0x9F,0xC2,0xF8,0x19,0xC2,
+        0x07,0x4C,0x76,0xA6,0xE0,0x67,0x4D,0xDC,
+        0x1D,0xC0,0x28,0x6F,0x9E,0x9E,0x00,0x3B,
+        0x7F,0x1A,0xF9,0xDD,0xE0,0x5D,0xC0,0xD3,
+        0xF7,0xBD,0x88,0x9F,0x28,0xC0,0x17,0xEC,
+        0x4E,0x07,0x05,0xFA,0x84,0x3C,0x22,0xA3,
+        0xFA,0x88,0xC0,0x2F,0x49,0x60,0x3C,0x92,
+        0xF8,0x40,0x01,0x84,0xEE,0x05,0xA8,0xD3,
+        0x07,0x47,0x3D,0xE3,0x17,0x54,0x63,0xBE,
+        0x5B,0x3D,0xC2,0x79,0x72,0x98,0xCB,0x01,
+        0x8B,0x73,0x4D,0x02,0xD5,0x71,0x97,0x8F,
+        0x0E,0xEE,0xB5,0x15,0xFB,0xFF,0x27,0x38,
+        0xB8,0x77,0x96,0x77,0x3E,0x43,0x79,0x90,
+        0xE0,0xBB,0xB6,0x82,0xE3,0xAA,0x06,0xE3,
+        0xD8,0xC2,0x2F,0x79,0x80,0x9D,0x61,0x71,
+        0xC1,0x7F,0x0F,0x03,0x51,0x89,0x30,0x28,
+        0x02,0xCB,0xBB,0xB7,0x52,0xF8,0x43,0x06,
+        0xE3,0x4D,0x81,0x4F,0x1A,0x3B,0x6A,0xE0,
+        0xFB,0xFF,0x1F,0x35,0xD8,0x86,0x8A,0xBB,
+        0x29,0x82,0x75,0xAA,0x98,0x21,0xF0,0x60,
+        0x0F,0x00,0x9F,0xAF,0x7C,0x06,0x50,0x14,
+        0x18,0xD4,0xA1,0x1D,0xCE,0x6D,0x18,0x70,
+        0x30,0x62,0xDC,0xA5,0x10,0xEE,0x94,0xDF,
+        0x51,0x62,0x3F,0x97,0xB3,0xE9,0xE2,0xAE,
+        0xE6,0x3E,0x9D,0xB0,0x0B,0x32,0x8C,0xB3,
+        0xC0,0x23,0xC0,0xAB,0x39,0xBF,0x20,0x3F,
+        0x17,0xBF,0x10,0x3C,0x26,0x85,0x78,0x53,
+        0x7A,0x25,0x36,0xC6,0x93,0x71,0x73,0xB7,
+        0x62,0x72,0xDE,0x79,0x41,0x36,0xC6,0xD1,
+        0x44,0x8C,0x72,0x6E,0x0F,0x03,0x91,0x5F,
+        0x90,0x7D,0x3F,0x79,0x21,0x88,0x18,0xCD,
+        0x10,0x41,0x9F,0x97,0x8D,0x15,0x28,0xDE,
+        0x0B,0x32,0x13,0xF8,0x56,0xD0,0xC1,0xC5,
+        0x17,0x64,0xEC,0xFF,0xFF,0x82,0x0C,0x30,
+        0xE2,0x64,0x04,0xF8,0x3C,0x71,0xE0,0xCE,
+        0x35,0x30,0xFE,0xFF,0x97,0x6A,0xD8,0x27,
+        0x1B,0xC0,0xD9,0xD0,0x7D,0xB2,0x01,0xF7,
+        0x68,0xE1,0x1D,0x4D,0x10,0x27,0x1B,0x0A,
+        0xE4,0xE0,0xEB,0xA2,0x70,0x3C,0xF4,0x49,
+        0x84,0x1E,0x9D,0x7C,0x94,0xC4,0x9D,0x19,
+        0x3C,0x91,0x77,0x16,0x8F,0xE2,0x65,0xD0,
+        0xF7,0x82,0x13,0x79,0x7D,0xB0,0x9C,0x63,
+        0x24,0xA8,0x46,0xE2,0xE3,0x03,0xFC,0xEB,
+        0x8B,0x8F,0x91,0xF0,0xF9,0xFC,0xC3,0xF2,
+        0x60,0x0C,0xF9,0xFF,0x7F,0x8A,0xC4,0x80,
+        0x3C,0xBB,0x3C,0x86,0xF0,0x0B,0x24,0xDC,
+        0xD3,0xCC,0x01,0x60,0x64,0x5D,0x1E,0xD1,
+        0x67,0x47,0x8E,0x11,0xD7,0x17,0x45,0x5F,
+        0x81,0x7D,0x10,0x38,0x9F,0xE7,0x44,0xB0,
+        0x8E,0x9A,0x1F,0x6D,0xF8,0xF8,0x39,0xF8,
+        0x5B,0xC1,0x03,0xA5,0x8F,0x45,0x21,0x1E,
+        0x91,0xF8,0x39,0x11,0x5C,0x26,0xCE,0x89,
+        0x40,0xE2,0xD0,0x0B,0xE3,0xB4,0x80,0x1B,
+        0x88,0xCF,0x94,0xD8,0x29,0x9F,0x08,0x3B,
+        0x97,0x60,0x46,0x07,0xAE,0xCB,0xBD,0x47,
+        0x07,0xFE,0x93,0x00,0x1E,0xEB,0xFF,0xFF,
+        0x78,0x07,0xBE,0x93,0xBA,0xEF,0x26,0xBE,
+        0xC8,0xF8,0x50,0xF4,0x7C,0x07,0xF8,0x0F,
+        0x77,0xB8,0x43,0xC5,0x39,0xDF,0x01,0xD2,
+        0xFE,0xFF,0xE7,0x3B,0x60,0x79,0xB6,0x7E,
+        0xBE,0x03,0xBB,0xC8,0xF3,0x1D,0x40,0xAC,
+        0xFF,0xFF,0xF9,0x0E,0xB0,0x73,0x46,0xC3,
+        0x9D,0xEF,0xC0,0x76,0xB4,0x01,0xCC,0x4D,
+        0xE3,0xD1,0x06,0xDC,0xC3,0x85,0x3D,0x0C,
+        0xAE,0xD0,0xA6,0x4F,0x8D,0x46,0xAD,0x1A,
+        0x94,0xA9,0x51,0xE6,0xFF,0xDF,0xA0,0x56,
+        0x9F,0x4A,0x8D,0x19,0xCB,0x0E,0xA5,0x80,
+        0x8F,0x0A,0x8D,0xCD,0xF2,0x28,0x04,0x62,
+        0x31,0xAF,0x06,0x81,0x38,0x2C,0x08,0x8D,
+        0xF4,0xCA,0x11,0x88,0x25,0x3F,0xFB,0x05,
+        0x62,0xB9,0x6F,0x06,0x81,0x38,0xE0,0x1B,
+        0x4C,0xE0,0xE4,0x61,0x25,0x70,0xF2,0x6E,
+        0x10,0x88,0x23,0x83,0x50,0xA1,0x3A,0x40,
+        0x58,0x4C,0x10,0x1A,0xCA,0x07,0x08,0x93,
+        0xFE,0x48,0x10,0x20,0x31,0x02,0xC2,0xC2,
+        0xBD,0xBF,0x04,0x62,0x69,0xEF,0x09,0x81,
+        0x58,0x88,0x15,0x10,0x16,0x17,0x84,0x86,
+        0xD3,0x02,0xC2,0x24,0x99,0x01,0x61,0x81,
+        0x40,0xA8,0x7C,0x35,0x20,0x4C,0xA4,0x1B,
+        0x40,0xBA,0x7A,0x81,0x38,0x88,0x1E,0x10,
+        0x26,0xC3,0x0F,0x08,0x0B,0x0D,0x42,0xA3,
+        0x3D,0x30,0x04,0x48,0x0C,0x81,0xB0,0xF8,
+        0x8E,0x40,0x98,0xF8,0x57,0x91,0x40,0x9C,
+        0xDF,0x12,0xC4,0x4D,0x69,0x88,0x35,0x01,
+        0x31,0x0D,0x9E,0x80,0x98,0x22,0x10,0x01,
+        0x39,0xF6,0xD3,0x43,0x40,0xD6,0x60,0x0A,
+        0x88,0x45,0x07,0x11,0x90,0x85,0xA8,0x02,
+        0x62,0x79,0x5D,0x01,0xB1,0xF0,0x20,0x02,
+        0x72,0xE6,0x97,0x9F,0x80,0xAC,0xE0,0xA5,
+        0xF3,0x10,0xC0,0xDE,0x10,0x81,0x48,0x72,
+        0x10,0x01,0x39,0xB0,0x2F,0x20,0x16,0x1F,
+        0x44,0x40,0xCE,0xFA,0x28,0x14,0x90,0x83,
+        0x83,0x68,0x10,0xE4,0x6B,0x26,0x20,0xA7,
+        0x07,0x11,0x10,0xF9,0x04,0x05,0x21,0x6A,
+        0xBD,0x81,0x30,0x3D,0x8F,0x42,0x0D,0x85,
+        0x80,0x50,0xE5,0xEA,0xCE,0x31,0x2C,0x07,
+        0x08,0xCD,0x05,0x22,0x30,0xAB,0x70,0x07,
+        0xC4,0x54,0x81,0x08,0xC8,0x09,0x80,0xC8,
+        0xFF,0x9F,0x60,0x2A,0x10,0x9A,0x12,0x8C,
+        0xEA,0x92,0x07,0xC4,0x12,0x80,0xD0,0x54,
+        0x20,0x34,0x25,0x88,0x00,0xAD,0xCA,0x1E,
+        0x10,0x53,0x0A,0x42,0x95,0x83,0xD0,0x74,
+        0x20,0x54,0xB6,0xBE,0xC3,0x02,0x05,0x11,
+        0x90,0xA3,0x83,0x50,0xE1,0xFE,0x40,0x98,
+        0xDE,0x97,0x86,0x00,0x9D,0x0E,0x44,0x40,
+        0x4E,0x0C,0x42,0x15,0x7C,0x32,0x82,0x10,
+        0xB1,0x20,0x54,0xC1,0x27,0x23,0x28,0xD1,
+        0xF2,0xB2,0x13,0x90,0xF5,0x81,0x50,0xBD,
+        0x20,0x02,0x73,0x36,0x20,0x9A,0x17,0x84,
+        0xE6,0x07,0xA3,0x5A,0x8D,0x02,0x31,0xFD,
+        0x20,0x34,0x0F,0x88,0xC0,0xAC,0xE0,0xF9,
+        0x71,0xC0,0x0C,0x84,0xAA,0x04,0x11,0x98,
+        0x73,0x01,0xD1,0xAC,0x20,0x34,0x3B,0x18,
+        0xD5,0xFE,0x0F,0xD1,0x00,0x08,0x08,0xCD,
+        0x07,0xA2,0xC3,0x00,0x79,0x96,0x09,0xC8,
+        0x1A,0x41,0xA8,0x66,0x10,0x81,0x39,0x27,
+        0x10,0xCD,0x0E,0x42,0x95,0xFD,0x4D,0x82,
+        0x91,0x8C,0x0F,0xD0,0x40,0x24,0x37,0x08,
+        0xD5,0xF1,0x0C,0x0A,0x46,0x74,0x83,0x08,
+        0xC8,0x59,0x40,0x68,0x36,0x30,0x9A,0x4C,
+        0xED,0x91,0x80,0xBA,0x05,0x61,0xE9,0x41,
+        0x68,0x3A,0xBB,0x83,0xA7,0x20,0x54,0x81,
+        0x5E,0x30,0xA6,0x19,0x44,0x87,0x05,0x02,
+        0x42,0x73,0x81,0x51,0x1D,0xAF,0x96,0x40,
+        0x44,0x1B,0x08,0xD5,0x0A,0xA2,0x81,0x93,
+        0x1F,0x53,0x10,0x92,0x14,0x84,0xFC,0xFF,
+        0x07,0xAA,0xC7,0x9C,0x40,0xAC,0xFA,0x5B,
+        0x25,0x50,0x27,0x01,0xA1,0xC9,0x40,0x74,
+        0x7C,0x20,0x0F,0xB8,0x83,0x64,0x20,0x54,
+        0x29,0x88,0xC0,0xAC,0xF4,0x63,0xA4,0x23,
+        0x05,0x51,0x7D,0xBC,0xA0,0x20,0x34,0xD1,
+        0x3B,0x2C,0x08,0x7B,0xB8,0x69,0xA8,0xE4,
+        0x59,0xA5,0xA1,0x12,0x10,0x9A,0x0D,0x44,
+        0xC7,0x04,0xF2,0xAA,0x79,0x4C,0x60,0x20,
+        0x54,0x2F,0x08,0xCD,0x01,0x42,0x13,0x83,
+        0x08,0xD4,0xA9,0xBF,0x37,0x1A,0x2A,0xF9,
+        0x5B,0x09,0xC4,0xCA,0x5E,0x69,0x02,0xB1,
+        0xDE,0xA7,0x4E,0x20,0xE6,0x1D,0x98,0xA9,
+        0x05,0xA1,0xEA,0x41,0x04,0xE6,0xB4,0x40,
+        0x54,0x81,0x78,0x10,0xA6,0x08,0x44,0x60,
+        0x4E,0x02,0x44,0xD3,0x81,0xD0,0xEC,0x60,
+        0x54,0xE7,0xA3,0x4D,0x40,0xD6,0x0E,0x42,
+        0xB3,0x80,0x08,0xCC,0x59,0x1E,0x69,0x02,
+        0xB1,0x92,0x2F,0x9D,0x0E,0x24,0x04,0x84,
+        0x26,0xD3,0x7F,0x68,0xA1,0x05,0x80,0x99,
+        0x84,0x04,0x20,0x4C,0x16,0x88,0x0E,0x27,
+        0xD6,0x08,0x22,0x40,0xC7,0x01,0xA3,0xD1,
+        0x40,0x68,0x5C,0x40,0x9A,0x1D,0x90,0x2A,
+        0x6D,0x00,0xC6,0x54,0x83,0xD0,0x24,0x20,
+        0x02,0x74,0x2C,0x10,0x01,0x5A,0x74,0x04,
+        0x30,0x16,0x01,0x84,0x46,0x05,0xA1,0xC9,
+        0x2A,0x80,0xB2,0x9C,0x20,0x1A,0x20,0xC9,
+        0x30,0x60,0x0A,0x42,0x33,0x81,0xD0,0x8C,
+        0x20,0x54,0x7C,0x07,0x10,0x16,0x04,0x84,
+        0x86,0x03,0xD1,0x00,0xFE,0xFF,0x8F,0x0C,
+        0x02,0xD1,0x00,0x9C,0x23,0xC4,0x61,0x85,
+        0x82,0xD0,0xF4,0x20,0x34,0x6C,0x09,0x50,
+        0x16,0x1D,0x44,0xC7,0x23,0x92,0x02,0x8C,
+        0x05,0x02,0xA1,0x31,0x41,0x68,0x6C,0x10,
+        0x1A,0x29,0x06,0x28,0x13,0x54,0xE3,0x50,
+        0x44,0x7B,0x80,0x31,0x99,0x20,0x54,0x36,
+        0x88,0xC0,0x1C,0x14,0x88,0x86,0x07,0xA1,
+        0x62,0x82,0x00,0x52,0x10,0x01,0x12,0x20,
+        0x1A,0x1E,0x84,0x8A,0x29,0x32,0x74,0x0A,
+        0x42,0x55,0x24,0x39,0x9A,0x50,0x10,0x1D,
+        0x4D,0x08,0x08,0xCD,0x07,0x46,0x75,0x35,
+        0x39,0x6E,0x50,0x10,0xAA,0x1D,0x84,0x06,
+        0x05,0xA1,0x39,0xA2,0x80,0xB2,0xEC,0x20,
+        0x02,0xB2,0x9E,0x2A,0x87,0x0A,0x0A,0x22,
+        0x30,0xA7,0x02,0xA2,0x49,0x41,0xA8,0x8E,
+        0x2C,0x47,0x0A,0x9A,0x06,0x84,0x25,0x06,
+        0xA1,0xC9,0xDA,0x80,0xB0,0x0C,0x75,0x0E,
+        0x24,0x14,0x84,0xE6,0x04,0xA1,0x4A,0xF2,
+        0x0C,0x8F,0x82,0xE8,0x38,0x42,0x80,0x68,
+        0x7A,0x10,0xAA,0xA6,0xCF,0x00,0x28,0x88,
+        0x06,0x40,0x40,0x68,0x4E,0x30,0xAA,0xA8,
+        0xD1,0xD1,0x84,0x82,0x50,0xDD,0x2F,0x4E,
+        0x81,0xF8,0xFF,0x0F,
+    })  // END MBUF
+
+} //end DefinitionBlock
+
diff --git a/tools/acpiexec/Makefile b/tools/acpiexec/Makefile
index f21adcbe88f..c9cf8a4b5b9 100644
--- a/tools/acpiexec/Makefile
+++ b/tools/acpiexec/Makefile
@@ -20,8 +20,10 @@ SRCS=	aetables.c aehandlers.c aeexec.c aemain.c \
 	../../disassembler/dmutils.c \
 	../../disassembler/dmwalk.c \
 	../../events/evevent.c \
-	../../events/evgpeblk.c \
 	../../events/evgpe.c \
+	../../events/evgpeblk.c \
+	../../events/evgpeinit.c \
+	../../events/evgpeutil.c \
 	../../events/evmisc.c \
 	../../events/evregion.c \
 	../../events/evrgnini.c \
diff --git a/tools/acpiexec/aeexec.c b/tools/acpiexec/aeexec.c
index 9aec80d9586..2e8790a5994 100644
--- a/tools/acpiexec/aeexec.c
+++ b/tools/acpiexec/aeexec.c
@@ -235,7 +235,6 @@ AfInstallGpeBlock (
     if (ACPI_SUCCESS (Status))
     {
         Status = AcpiInstallGpeBlock (Handle2, &BlockAddress, 7, 8);
-        Status = AcpiInstallGpeBlock (Handle2, &BlockAddress, 2, 0);
 
         AcpiInstallGpeHandler (Handle2, 8, ACPI_GPE_LEVEL_TRIGGERED, AeGpeHandler, NULL);
         AcpiEnableGpe (Handle2, 8, ACPI_GPE_TYPE_RUNTIME);
diff --git a/tools/acpiexec/aetables.c b/tools/acpiexec/aetables.c
index 4c3734a05cb..e1f61c6e61a 100644
--- a/tools/acpiexec/aetables.c
+++ b/tools/acpiexec/aetables.c
@@ -198,6 +198,34 @@ unsigned char Oem1Code[] =
     0x31,0x00,0x70,0x01,0x5F,0x58,0x54,0x32,  /* 00000030    "1.p._XT2" */
 };
 
+/* ASL source for this table is at the end of this file */
+
+unsigned char OemxCode[] =
+{
+    0x4F,0x45,0x4D,0x58,0xB0,0x00,0x00,0x00,  /* 00000000    "OEMX...." */
+    0x02,0x54,0x4D,0x79,0x4F,0x45,0x4D,0x00,  /* 00000008    ".TMyOEM." */
+    0x54,0x65,0x73,0x74,0x00,0x00,0x00,0x00,  /* 00000010    "Test...." */
+    0x32,0x04,0x00,0x00,0x49,0x4E,0x54,0x4C,  /* 00000018    "2...INTL" */
+    0x31,0x03,0x10,0x20,0x14,0x1D,0x5F,0x49,  /* 00000020    "1.. .._I" */
+    0x4E,0x49,0x00,0x70,0x0D,0x54,0x61,0x62,  /* 00000028    "NI.p.Tab" */
+    0x6C,0x65,0x20,0x4F,0x45,0x4D,0x58,0x20,  /* 00000030    "le OEMX " */
+    0x72,0x75,0x6E,0x6E,0x69,0x6E,0x67,0x00,  /* 00000038    "running." */
+    0x5B,0x31,0x10,0x22,0x5C,0x5F,0x47,0x50,  /* 00000040    "[1."\_GP" */
+    0x45,0x14,0x06,0x5F,0x45,0x30,0x37,0x00,  /* 00000048    "E.._E07." */
+    0x14,0x06,0x5F,0x45,0x32,0x32,0x00,0x14,  /* 00000050    ".._E22.." */
+    0x06,0x5F,0x4C,0x33,0x31,0x00,0x14,0x06,  /* 00000058    "._L31..." */
+    0x5F,0x4C,0x36,0x36,0x00,0x5B,0x82,0x10,  /* 00000060    "_L66.[.." */
+    0x4F,0x45,0x4D,0x31,0x08,0x5F,0x50,0x52,  /* 00000068    "OEM1._PR" */
+    0x57,0x12,0x05,0x02,0x0A,0x07,0x00,0x5B,  /* 00000070    "W......[" */
+    0x82,0x10,0x4F,0x45,0x4D,0x32,0x08,0x5F,  /* 00000078    "..OEM2._" */
+    0x50,0x52,0x57,0x12,0x05,0x02,0x0A,0x66,  /* 00000080    "PRW....f" */
+    0x00,0x10,0x26,0x5C,0x47,0x50,0x45,0x32,  /* 00000088    "..&\GPE2" */
+    0x14,0x06,0x5F,0x4C,0x30,0x31,0x00,0x14,  /* 00000090    ".._L01.." */
+    0x06,0x5F,0x45,0x30,0x37,0x00,0x08,0x5F,  /* 00000098    "._E07.._" */
+    0x50,0x52,0x57,0x12,0x0C,0x02,0x12,0x08,  /* 000000A0    "PRW....." */
+    0x02,0x5C,0x47,0x50,0x45,0x32,0x01,0x00   /* 000000A8    ".\GPE2.." */
+};
+
 /*
  * Example installable control method
  *
@@ -242,7 +270,7 @@ ACPI_TABLE_HEADER           LocalTEST;
 ACPI_TABLE_HEADER           LocalBADTABLE;
 ACPI_TABLE_RSDT             *LocalRSDT;
 
-#define BASE_RSDT_TABLES    6
+#define BASE_RSDT_TABLES    7
 #define BASE_RSDT_SIZE      (sizeof (ACPI_TABLE_RSDT) + ((BASE_RSDT_TABLES -1) * sizeof (UINT32)))
 
 #define ACPI_MAX_INIT_TABLES (32)
@@ -349,6 +377,10 @@ AeBuildLocalTables (
 
     LocalRSDT->TableOffsetEntry[5] = ACPI_PTR_TO_PHYSADDR (&Oem1Code);
 
+    /* Install the OEMx table to test LoadTable */
+
+    LocalRSDT->TableOffsetEntry[6] = ACPI_PTR_TO_PHYSADDR (&OemxCode);
+
     /*
      * Install the user tables. The DSDT must be installed in the FADT.
      * All other tables are installed directly into the RSDT.
@@ -583,3 +615,110 @@ AeLocalGetRootPointer (
 }
 
 
+#if 0
+/******************************************************************************
+ *
+ * DESCRIPTION: ASL tables that are used in RSDT/XSDT, also used to test
+ *              Load/LoadTable operators.
+ *
+ *****************************************************************************/
+
+DefinitionBlock ("", "OEMX", 2, "MyOEM", "Test", 0x00000432)
+{
+    External (GPE2, DeviceObj)
+
+    Method (_INI)
+    {
+        Store ("Table OEMX running", Debug)
+    }
+
+    Scope (\_GPE)
+    {
+        Method (_E07) {}
+        Method (_E22) {}
+        Method (_L31) {}
+        Method (_L66) {}
+    }
+
+    Device (OEM1)
+    {
+        Name (_PRW, Package(){7,0})
+    }
+    Device (OEM2)
+    {
+        Name (_PRW, Package(){0x66,0})
+    }
+
+    Scope (\GPE2)
+    {
+        Method (_L01) {}
+        Method (_E07) {}
+
+        Name (_PRW, Package() {Package() {\GPE2, 1}, 0})
+    }
+}
+
+/* Parent gr.asl file */
+
+DefinitionBlock ("", "DSDT", 2, "Intel", "Many", 0x00000001)
+{
+    Name (BUF1, Buffer()
+    {
+        0x4F,0x45,0x4D,0x58,0xB0,0x00,0x00,0x00,  /* 00000000    "OEMX...." */
+        0x02,0x54,0x4D,0x79,0x4F,0x45,0x4D,0x00,  /* 00000008    ".TMyOEM." */
+        0x54,0x65,0x73,0x74,0x00,0x00,0x00,0x00,  /* 00000010    "Test...." */
+        0x32,0x04,0x00,0x00,0x49,0x4E,0x54,0x4C,  /* 00000018    "2...INTL" */
+        0x31,0x03,0x10,0x20,0x14,0x1D,0x5F,0x49,  /* 00000020    "1.. .._I" */
+        0x4E,0x49,0x00,0x70,0x0D,0x54,0x61,0x62,  /* 00000028    "NI.p.Tab" */
+        0x6C,0x65,0x20,0x4F,0x45,0x4D,0x58,0x20,  /* 00000030    "le OEMX " */
+        0x72,0x75,0x6E,0x6E,0x69,0x6E,0x67,0x00,  /* 00000038    "running." */
+        0x5B,0x31,0x10,0x22,0x5C,0x5F,0x47,0x50,  /* 00000040    "[1."\_GP" */
+        0x45,0x14,0x06,0x5F,0x45,0x30,0x37,0x00,  /* 00000048    "E.._E07." */
+        0x14,0x06,0x5F,0x45,0x32,0x32,0x00,0x14,  /* 00000050    ".._E22.." */
+        0x06,0x5F,0x4C,0x33,0x31,0x00,0x14,0x06,  /* 00000058    "._L31..." */
+        0x5F,0x4C,0x36,0x36,0x00,0x5B,0x82,0x10,  /* 00000060    "_L66.[.." */
+        0x4F,0x45,0x4D,0x31,0x08,0x5F,0x50,0x52,  /* 00000068    "OEM1._PR" */
+        0x57,0x12,0x05,0x02,0x0A,0x07,0x00,0x5B,  /* 00000070    "W......[" */
+        0x82,0x10,0x4F,0x45,0x4D,0x32,0x08,0x5F,  /* 00000078    "..OEM2._" */
+        0x50,0x52,0x57,0x12,0x05,0x02,0x0A,0x66,  /* 00000080    "PRW....f" */
+        0x00,0x10,0x26,0x5C,0x47,0x50,0x45,0x32,  /* 00000088    "..&\GPE2" */
+        0x14,0x06,0x5F,0x4C,0x30,0x31,0x00,0x14,  /* 00000090    ".._L01.." */
+        0x06,0x5F,0x45,0x30,0x37,0x00,0x08,0x5F,  /* 00000098    "._E07.._" */
+        0x50,0x52,0x57,0x12,0x0C,0x02,0x12,0x08,  /* 000000A0    "PRW....." */
+        0x02,0x5C,0x47,0x50,0x45,0x32,0x01,0x00   /* 000000A8    ".\GPE2.." */
+    })
+
+    Name (HNDL, 0)
+    Method (LD)
+    {
+        Load (BUF1, HNDL)
+        Store ("Load operator, handle:", Debug)
+        Store (HNDL, Debug)
+    }
+
+    Method (MAIN, 0, NotSerialized)
+    {
+        Store ("Loading OEMX table", Debug)
+        Store (LoadTable ("OEMX", "MyOEM", "Test"), Debug)
+    }
+
+    Scope (\_GPE)
+    {
+        Method (_L08) {}
+        Method (_E08) {}
+        Method (_L0B) {}
+    }
+
+    Device (DEV0)
+    {
+        Name (_PRW, Package() {0x11, 0})
+    }
+
+    Device (\GPE2)
+    {
+        Method (_L00) {}
+    }
+}
+
+#endif
+
diff --git a/tools/acpisrc/astable.c b/tools/acpisrc/astable.c
index 51c7f3d00c7..5a96f06a8c8 100644
--- a/tools/acpisrc/astable.c
+++ b/tools/acpisrc/astable.c
@@ -310,7 +310,6 @@ ACPI_TYPED_IDENTIFIER_TABLE           AcpiIdentifiers[] = {
     {"ACPI_INTEGER_OVERLAY",                SRC_TYPE_STRUCT},
     {"ACPI_INTEGRITY_INFO",                 SRC_TYPE_STRUCT},
     {"ACPI_INTERFACE_INFO",                 SRC_TYPE_STRUCT},
-    {"ACPI_INTERNAL_RSDT",                  SRC_TYPE_STRUCT},
     {"ACPI_INTERPRETER_MODE",               SRC_TYPE_SIMPLE},
     {"ACPI_IO_ADDRESS",                     SRC_TYPE_SIMPLE},
     {"ACPI_IO_ATTRIBUTE",                   SRC_TYPE_STRUCT},
diff --git a/utilities/utcopy.c b/utilities/utcopy.c
index fe44083c598..c1fccaefe1b 100644
--- a/utilities/utcopy.c
+++ b/utilities/utcopy.c
@@ -797,6 +797,7 @@ AcpiUtCopySimpleObject (
     UINT16                  ReferenceCount;
     ACPI_OPERAND_OBJECT     *NextObject;
     ACPI_STATUS             Status;
+    ACPI_SIZE               CopySize;
 
 
     /* Save fields from destination that we don't want to overwrite */
@@ -804,10 +805,18 @@ AcpiUtCopySimpleObject (
     ReferenceCount = DestDesc->Common.ReferenceCount;
     NextObject = DestDesc->Common.NextObject;
 
-    /* Copy the entire source object over the destination object*/
+    /*
+     * Copy the entire source object over the destination object.
+     * Note: Source can be either an operand object or namespace node.
+     */
+    CopySize = sizeof (ACPI_OPERAND_OBJECT);
+    if (ACPI_GET_DESCRIPTOR_TYPE (SourceDesc) == ACPI_DESC_TYPE_NAMED)
+    {
+        CopySize = sizeof (ACPI_NAMESPACE_NODE);
+    }
 
-    ACPI_MEMCPY ((char *) DestDesc, (char *) SourceDesc,
-                    sizeof (ACPI_OPERAND_OBJECT));
+    ACPI_MEMCPY (ACPI_CAST_PTR (char, DestDesc),
+        ACPI_CAST_PTR (char, SourceDesc), CopySize);
 
     /* Restore the saved fields */
 
@@ -841,8 +850,7 @@ AcpiUtCopySimpleObject (
             /* Copy the actual buffer data */
 
             ACPI_MEMCPY (DestDesc->Buffer.Pointer,
-                    SourceDesc->Buffer.Pointer,
-                    SourceDesc->Buffer.Length);
+                SourceDesc->Buffer.Pointer, SourceDesc->Buffer.Length);
         }
         break;
 
@@ -864,7 +872,7 @@ AcpiUtCopySimpleObject (
             /* Copy the actual string data */
 
             ACPI_MEMCPY (DestDesc->String.Pointer, SourceDesc->String.Pointer,
-                         (ACPI_SIZE) SourceDesc->String.Length + 1);
+                (ACPI_SIZE) SourceDesc->String.Length + 1);
         }
         break;
 
diff --git a/utilities/utglobal.c b/utilities/utglobal.c
index 986dbeb58d8..34af1f5fc50 100644
--- a/utilities/utglobal.c
+++ b/utilities/utglobal.c
@@ -625,7 +625,7 @@ AcpiUtGetNodeName (
 
 static const char           *AcpiGbl_DescTypeNames[] =
 {
-    /* 00 */ "Invalid",
+    /* 00 */ "Not a Descriptor",
     /* 01 */ "Cached",
     /* 02 */ "State-Generic",
     /* 03 */ "State-Update",
@@ -656,7 +656,7 @@ AcpiUtGetDescriptorName (
 
     if (ACPI_GET_DESCRIPTOR_TYPE (Object) > ACPI_DESC_TYPE_MAX)
     {
-        return (ACPI_CAST_PTR (char, AcpiGbl_BadType));
+        return ("Not a Descriptor");
     }
 
     return (ACPI_CAST_PTR (char,
diff --git a/utilities/uttrack.c b/utilities/uttrack.c
index 37fdba1c446..ac45e9a3d88 100644
--- a/utilities/uttrack.c
+++ b/utilities/uttrack.c
@@ -644,6 +644,7 @@ AcpiUtDumpAllocations (
     ACPI_DEBUG_MEM_BLOCK    *Element;
     ACPI_DESCRIPTOR         *Descriptor;
     UINT32                  NumOutstanding = 0;
+    UINT8                   DescriptorType;
 
 
     ACPI_FUNCTION_TRACE (UtDumpAllocations);
@@ -663,43 +664,86 @@ AcpiUtDumpAllocations (
         if ((Element->Component & Component) &&
             ((Module == NULL) || (0 == ACPI_STRCMP (Module, Element->Module))))
         {
-            /* Ignore allocated objects that are in a cache */
-
             Descriptor = ACPI_CAST_PTR (ACPI_DESCRIPTOR, &Element->UserSpace);
-            if (ACPI_GET_DESCRIPTOR_TYPE (Descriptor) != ACPI_DESC_TYPE_CACHED)
+
+            if (Element->Size < sizeof (ACPI_COMMON_DESCRIPTOR))
             {
-                AcpiOsPrintf ("%p Len %04X %9.9s-%d [%s] ",
+                AcpiOsPrintf ("%p Length 0x%04X %9.9s-%d "
+                    "[Not a Descriptor - too small]\n",
                     Descriptor, Element->Size, Element->Module,
-                    Element->Line, AcpiUtGetDescriptorName (Descriptor));
-
-                /* Most of the elements will be Operand objects. */
-
-                switch (ACPI_GET_DESCRIPTOR_TYPE (Descriptor))
-                {
-                case ACPI_DESC_TYPE_OPERAND:
-                    AcpiOsPrintf ("%12.12s R%hd",
-                        AcpiUtGetTypeName (Descriptor->Object.Common.Type),
-                        Descriptor->Object.Common.ReferenceCount);
-                    break;
-
-                case ACPI_DESC_TYPE_PARSER:
-                    AcpiOsPrintf ("AmlOpcode %04hX",
-                        Descriptor->Op.Asl.AmlOpcode);
-                    break;
-
-                case ACPI_DESC_TYPE_NAMED:
-                    AcpiOsPrintf ("%4.4s",
-                        AcpiUtGetNodeName (&Descriptor->Node));
-                    break;
-
-                default:
-                    break;
-                }
-
-                AcpiOsPrintf ( "\n");
-                NumOutstanding++;
+                    Element->Line);
             }
+            else
+            {
+                /* Ignore allocated objects that are in a cache */
+
+                if (ACPI_GET_DESCRIPTOR_TYPE (Descriptor) != ACPI_DESC_TYPE_CACHED)
+                {
+                    AcpiOsPrintf ("%p Length 0x%04X %9.9s-%d [%s] ",
+                        Descriptor, Element->Size, Element->Module,
+                        Element->Line, AcpiUtGetDescriptorName (Descriptor));
+
+                    /* Validate the descriptor type using Type field and length */
+
+                    DescriptorType = 0; /* Not a valid descriptor type */
+
+                    switch (ACPI_GET_DESCRIPTOR_TYPE (Descriptor))
+                    {
+                    case ACPI_DESC_TYPE_OPERAND:
+                        if (Element->Size == sizeof (ACPI_DESC_TYPE_OPERAND))
+                        {
+                            DescriptorType = ACPI_DESC_TYPE_OPERAND;
+                        }
+                        break;
+
+                    case ACPI_DESC_TYPE_PARSER:
+                        if (Element->Size == sizeof (ACPI_DESC_TYPE_PARSER))
+                        {
+                            DescriptorType = ACPI_DESC_TYPE_PARSER;
+                        }
+                        break;
+
+                    case ACPI_DESC_TYPE_NAMED:
+                        if (Element->Size == sizeof (ACPI_DESC_TYPE_NAMED))
+                        {
+                            DescriptorType = ACPI_DESC_TYPE_NAMED;
+                        }
+                        break;
+
+                    default:
+                        break;
+                    }
+
+                    /* Display additional info for the major descriptor types */
+
+                    switch (DescriptorType)
+                    {
+                    case ACPI_DESC_TYPE_OPERAND:
+                        AcpiOsPrintf ("%12.12s  RefCount 0x%04X\n",
+                            AcpiUtGetTypeName (Descriptor->Object.Common.Type),
+                            Descriptor->Object.Common.ReferenceCount);
+                        break;
+
+                    case ACPI_DESC_TYPE_PARSER:
+                        AcpiOsPrintf ("AmlOpcode 0x%04hX\n",
+                            Descriptor->Op.Asl.AmlOpcode);
+                        break;
+
+                    case ACPI_DESC_TYPE_NAMED:
+                        AcpiOsPrintf ("%4.4s\n",
+                            AcpiUtGetNodeName (&Descriptor->Node));
+                        break;
+
+                    default:
+                        AcpiOsPrintf ( "\n");
+                        break;
+                    }
+                }
+            }
+
+            NumOutstanding++;
         }
+
         Element = Element->Next;
     }
 
@@ -709,13 +753,11 @@ AcpiUtDumpAllocations (
 
     if (!NumOutstanding)
     {
-        ACPI_INFO ((AE_INFO,
-            "No outstanding allocations"));
+        ACPI_INFO ((AE_INFO, "No outstanding allocations"));
     }
     else
     {
-        ACPI_ERROR ((AE_INFO,
-            "%d(0x%X) Outstanding allocations",
+        ACPI_ERROR ((AE_INFO, "%d(0x%X) Outstanding allocations",
             NumOutstanding, NumOutstanding));
     }
 

From 1228041bd5a73b740d8e2545ea9a6f38061cd888 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Wed, 28 Apr 2010 22:23:29 +0000
Subject: [PATCH 309/532] Don't assume that "resource" property is in metadata.

Reported by:	Mikolaj Golub <to.my.trociny@gmail.com>
MFC after:	3 days
---
 sbin/hastd/metadata.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sbin/hastd/metadata.c b/sbin/hastd/metadata.c
index 9bca66bbbe1..7a138e816dc 100644
--- a/sbin/hastd/metadata.c
+++ b/sbin/hastd/metadata.c
@@ -117,7 +117,7 @@ metadata_read(struct hast_resource *res, bool openrw)
 	}
 
 	str = nv_get_string(nv, "resource");
-	if (strcmp(str, res->hr_name) != 0) {
+	if (str != NULL && strcmp(str, res->hr_name) != 0) {
 		pjdlog_error("Provider %s is not part of resource %s.",
 		    res->hr_localpath, res->hr_name);
 		nv_free(nv);

From 06c117d1d131b959374ba1655d3cc4f78ebc8bac Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Wed, 28 Apr 2010 22:26:30 +0000
Subject: [PATCH 310/532] Use WEXITSTATUS() to obtain real exit code.

MFC after:	3 days
---
 sbin/hastd/hastd.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/sbin/hastd/hastd.c b/sbin/hastd/hastd.c
index 957885d7ac1..50bc9ffad81 100644
--- a/sbin/hastd/hastd.c
+++ b/sbin/hastd/hastd.c
@@ -294,9 +294,10 @@ listen_accept(void)
 				    "Waiting for worker process (pid=%u) failed",
 				    (unsigned int)res->hr_workerpid);
 				/* See above. */
-			} else if (status != 0) {
+			} else if (WEXITSTATUS(status) != 0) {
 				pjdlog_error("Worker process (pid=%u) exited ungracefully: status=%d.",
-				    (unsigned int)res->hr_workerpid, status);
+				    (unsigned int)res->hr_workerpid,
+				    WEXITSTATUS(status));
 				/* See above. */
 			} else {
 				pjdlog_debug(1,

From a16334d031e5f99a4c050428f8d6e9b3ef6deae5 Mon Sep 17 00:00:00 2001
From: Doug Barton <dougb@FreeBSD.org>
Date: Wed, 28 Apr 2010 22:29:17 +0000
Subject: [PATCH 311/532] In the case where named_chroot_autoupdate is NOT set,
 but named_chrootdir IS set, named-checkconf fails because it cannot find the
 conf file. Fix this by making checkconf a variable that includes "-t
 $named_chrootdir" as needed. Notice of the bug and suggested direction for
 the fix from [1].

Using required_files for named.conf is overkill ever since
I added the named-checkconf call, so rather than update the
logic to handle the case described above, remove it. This
also handles the case where named_chroot_autoupdate IS set
but the symlink doesn't exist yet.

PR:		conf/145904
Submitted by:	J R Matthews
---
 etc/rc.d/named | 14 ++++++++++----
 1 file changed, 10 insertions(+), 4 deletions(-)

diff --git a/etc/rc.d/named b/etc/rc.d/named
index 346b6fc9dd4..81db6463dd3 100755
--- a/etc/rc.d/named
+++ b/etc/rc.d/named
@@ -192,6 +192,13 @@ named_prestart()
 		$confgen_command
 	fi
 
+	local checkconf
+
+	checkconf="${command%/named}/named-checkconf"
+	if ! checkyesno named_chroot_autoupdate && [ -n "$named_chrootdir" ]; then
+		checkconf="$checkconf -t $named_chrootdir"
+	fi
+
 	# Create a forwarder configuration based on /etc/resolv.conf
 	if checkyesno named_auto_forward; then
 		if [ ! -s /etc/resolv.conf ]; then
@@ -201,7 +208,7 @@ named_prestart()
 			[ -s "${named_confdir}/auto_forward.conf" ] &&
 			    create_file ${named_confdir}/auto_forward.conf
 
-			${command%/named}/named-checkconf $named_conf ||
+			$checkconf $named_conf ||
 			    err 3 'named-checkconf for $named_conf failed'
 			return
 		fi
@@ -263,8 +270,7 @@ named_prestart()
 		    create_file ${named_confdir}/auto_forward.conf
 	fi
 
-	${command%/named}/named-checkconf $named_conf ||
-	    err 3 'named-checkconf for $named_conf failed'
+	$checkconf $named_conf || err 3 'named-checkconf for $named_conf failed'
 }
 
 load_rc_config $name
@@ -272,7 +278,7 @@ load_rc_config $name
 # Updating the following variables requires that rc.conf be loaded first
 #
 required_dirs="$named_chrootdir"	# if it is set, it must exist
-required_files="${named_conf:=/etc/namedb/named.conf}"
+
 pidfile="${named_pidfile:-/var/run/named/pid}"
 named_confdir="${named_chrootdir}${named_conf%/*}"
 

From 5abfc9c14530b3d43260b20627a78ddd25e17518 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Wed, 28 Apr 2010 22:39:47 +0000
Subject: [PATCH 312/532] Mark temporary issues as such.

MFC after:	3 days
---
 sbin/hastd/primary.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sbin/hastd/primary.c b/sbin/hastd/primary.c
index 09151546772..73f722b215e 100644
--- a/sbin/hastd/primary.c
+++ b/sbin/hastd/primary.c
@@ -480,7 +480,7 @@ init_remote(struct hast_resource *res, struct proto_conn **inp,
 
 	/* Prepare outgoing connection with remote node. */
 	if (proto_client(res->hr_remoteaddr, &out) < 0) {
-		primary_exit(EX_OSERR, "Unable to create connection to %s",
+		primary_exit(EX_TEMPFAIL, "Unable to create connection to %s",
 		    res->hr_remoteaddr);
 	}
 	/* Try to connect, but accept failure. */
@@ -739,7 +739,7 @@ hastd_primary(struct hast_resource *res)
 	pid = fork();
 	if (pid < 0) {
 		KEEP_ERRNO((void)pidfile_remove(pfh));
-		primary_exit(EX_OSERR, "Unable to fork");
+		primary_exit(EX_TEMPFAIL, "Unable to fork");
 	}
 
 	if (pid > 0) {

From 83a5671405e3914b3dd393cdcb96c3b8de73168b Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Wed, 28 Apr 2010 22:41:06 +0000
Subject: [PATCH 313/532] Restart worker thread only if the problem was
 temporary. In case of persistent problem we don't want to loop forever.

MFC after:	3 days
---
 sbin/hastd/hastd.c | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/sbin/hastd/hastd.c b/sbin/hastd/hastd.c
index 50bc9ffad81..7b2a8e255d2 100644
--- a/sbin/hastd/hastd.c
+++ b/sbin/hastd/hastd.c
@@ -140,9 +140,15 @@ child_exit(void)
 		proto_close(res->hr_ctrl);
 		res->hr_workerpid = 0;
 		if (res->hr_role == HAST_ROLE_PRIMARY) {
-			sleep(1);
-			pjdlog_info("Restarting worker process.");
-			hastd_primary(res);
+			if (WEXITSTATUS(status) == EX_TEMPFAIL) {
+				sleep(1);
+				pjdlog_info("Restarting worker process.");
+				hastd_primary(res);
+			} else {
+				res->hr_role = HAST_ROLE_INIT;
+				pjdlog_info("Changing resource role back to %s.",
+				    role2str(res->hr_role));
+			}
 		}
 		pjdlog_prefix_set("%s", "");
 	}

From cb8a84e08e92154e07608ad7efaadda77a8a8f9b Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Wed, 28 Apr 2010 23:05:42 +0000
Subject: [PATCH 314/532] Delete a diagnostic statement that is no longer
 useful from the experimental NFS client.

MFC after:	1 week
---
 sys/fs/nfsclient/nfs_clrpcops.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clrpcops.c b/sys/fs/nfsclient/nfs_clrpcops.c
index a84a3014abe..5d00102c970 100644
--- a/sys/fs/nfsclient/nfs_clrpcops.c
+++ b/sys/fs/nfsclient/nfs_clrpcops.c
@@ -1399,10 +1399,6 @@ nfsrpc_write(vnode_t vp, struct uio *uiop, int *iomode, u_char *verfp,
 		else
 			error = nfsrpc_writerpc(vp, uiop, iomode, verfp,
 			    newcred, &stateid, p, nap, attrflagp, stuff);
-if (error == NFSERR_BADSTATEID) {
-printf("st=0x%x 0x%x 0x%x\n",stateid.other[0],stateid.other[1],stateid.other[2]);
-nfscl_dumpstate(nmp, 1, 1, 0, 0);
-}
 		if (error == NFSERR_STALESTATEID)
 			nfscl_initiate_recovery(nmp->nm_clp);
 		if (lckp != NULL)

From 8583f92fdf265442bc56bd687e4b52a0adf2658a Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Wed, 28 Apr 2010 23:16:21 +0000
Subject: [PATCH 315/532] For the experimental NFS client, it should always
 flush dirty buffers before closing the NFSv4 opens, as the comment states.
 This patch deletes the call to nfscl_mustflush() which would return 0 for the
 case where a delegation still exists, which was incorrect and could cause
 crashes during recovery from an expired lease.

MFC after:	1 week
---
 sys/fs/nfsclient/nfs_clnode.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clnode.c b/sys/fs/nfsclient/nfs_clnode.c
index c1337427fd2..e36431fa0ec 100644
--- a/sys/fs/nfsclient/nfs_clnode.c
+++ b/sys/fs/nfsclient/nfs_clnode.c
@@ -198,8 +198,7 @@ ncl_inactive(struct vop_inactive_args *ap)
 		 * must be flushed before the close, so that the stateid is
 		 * available for the writes.
 		 */
-		if (nfscl_mustflush(vp))
-			(void) ncl_flush(vp, MNT_WAIT, NULL, ap->a_td, 1, 0);
+		(void) ncl_flush(vp, MNT_WAIT, NULL, ap->a_td, 1, 0);
 		(void) nfsrpc_close(vp, 1, ap->a_td);
 	}
 

From ff605594dda33c2c9f6ab5ddbf31bd67b98938b2 Mon Sep 17 00:00:00 2001
From: Maxim Sobolev <sobomax@FreeBSD.org>
Date: Thu, 29 Apr 2010 06:16:00 +0000
Subject: [PATCH 316/532] On certain chipsets AT keyboard controller isn't
 present and is emulated by BIOS using SMI interrupt. On those chipsets
 reading from the status port may be thousand times slower than usually.
 Sometimes this emilation is not working properly resulting in commands timing
 out and since we assume that inb() operation takes very little time to
 complete we need to adjust number of retries to keep waiting time within a
 designed limits (100ms). Measure time it takes to make read_status() call and
 adjust number of retries accordingly.

To keep it simple, use TSC to measure inb() performance and
keep it to amd64-only, since TSC may not available on older
CPUs.

Also enable detection of the AT controller absence on amd64.

Reviewed by:	jhb
MFC after:	1 month
---
 sys/dev/atkbdc/atkbdc.c      | 76 +++++++++++++++++++++++++++++-------
 sys/dev/atkbdc/atkbdc_ebus.c |  1 +
 sys/dev/atkbdc/atkbdc_isa.c  |  7 +++-
 sys/dev/atkbdc/atkbdcreg.h   |  1 +
 4 files changed, 69 insertions(+), 16 deletions(-)

diff --git a/sys/dev/atkbdc/atkbdc.c b/sys/dev/atkbdc/atkbdc.c
index df1f28e497d..f8e856a0efd 100644
--- a/sys/dev/atkbdc/atkbdc.c
+++ b/sys/dev/atkbdc/atkbdc.c
@@ -44,6 +44,10 @@ __FBSDID("$FreeBSD$");
 #include <machine/resource.h>
 #include <sys/rman.h>
 
+#if defined(__amd64__)
+#include <machine/clock.h>
+#endif
+
 #include <dev/atkbdc/atkbdcreg.h>
 
 #ifdef __sparc64__
@@ -153,7 +157,7 @@ atkbdc_configure(void)
 	bus_space_tag_t tag;
 	bus_space_handle_t h0;
 	bus_space_handle_t h1;
-#if defined(__i386__)
+#if defined(__i386__) || defined(__amd64__)
 	volatile int i;
 	register_t flags;
 #endif
@@ -222,7 +226,7 @@ atkbdc_configure(void)
 #endif
 #endif
 
-#if defined(__i386__)
+#if defined(__i386__) || defined(__amd64__)
 	/*
 	 * Check if we really have AT keyboard controller. Poll status
 	 * register until we get "all clear" indication. If no such
@@ -248,6 +252,11 @@ static int
 atkbdc_setup(atkbdc_softc_t *sc, bus_space_tag_t tag, bus_space_handle_t h0,
 	     bus_space_handle_t h1)
 {
+#if defined(__amd64__)
+	u_int64_t tscval[3], read_delay;
+	register_t flags;
+#endif
+
 	if (sc->ioh0 == 0) {	/* XXX */
 	    sc->command_byte = -1;
 	    sc->command_mask = 0;
@@ -264,6 +273,33 @@ atkbdc_setup(atkbdc_softc_t *sc, bus_space_tag_t tag, bus_space_handle_t h0,
 	sc->iot = tag;
 	sc->ioh0 = h0;
 	sc->ioh1 = h1;
+
+#if defined(__amd64__)
+	/*
+	 * On certain chipsets AT keyboard controller isn't present and is
+	 * emulated by BIOS using SMI interrupt. On those chipsets reading
+	 * from the status port may be thousand times slower than usually.
+	 * Sometimes this emilation is not working properly resulting in
+	 * commands timing our and since we assume that inb() operation 
+	 * takes very little time to complete we need to adjust number of
+	 * retries to keep waiting time within a designed limits (100ms).
+	 * Measure time it takes to make read_status() call and adjust
+	 * number of retries accordingly.
+	 */
+	flags = intr_disable();
+	tscval[0] = rdtsc();
+	read_status(sc);
+	tscval[1] = rdtsc();
+	DELAY(1000);
+	tscval[2] = rdtsc();
+	intr_restore(flags);
+	read_delay = tscval[1] - tscval[0];
+	read_delay /= (tscval[2] - tscval[1]) / 1000;
+	sc->retry = 100000 / ((KBDD_DELAYTIME * 2) + read_delay);
+#else
+	sc->retry = 5000;
+#endif
+
 	return 0;
 }
 
@@ -380,10 +416,12 @@ removeq(kqueue *q)
 static int
 wait_while_controller_busy(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 100msec at most */
-    int retry = 5000;
+    int retry;
     int f;
 
+    /* CPU will stay inside the loop for 100msec at most */
+    retry = kbdc->retry;
+
     while ((f = read_status(kbdc)) & KBDS_INPUT_BUFFER_FULL) {
 	if ((f & KBDS_BUFFER_FULL) == KBDS_KBD_BUFFER_FULL) {
 	    DELAY(KBDD_DELAYTIME);
@@ -406,10 +444,12 @@ wait_while_controller_busy(struct atkbdc_softc *kbdc)
 static int
 wait_for_data(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 200msec at most */
-    int retry = 10000;
+    int retry;
     int f;
 
+    /* CPU will stay inside the loop for 200msec at most */
+    retry = kbdc->retry * 2;
+
     while ((f = read_status(kbdc) & KBDS_ANY_BUFFER_FULL) == 0) {
         DELAY(KBDC_DELAYTIME);
         if (--retry < 0)
@@ -423,10 +463,12 @@ wait_for_data(struct atkbdc_softc *kbdc)
 static int
 wait_for_kbd_data(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 200msec at most */
-    int retry = 10000;
+    int retry;
     int f;
 
+    /* CPU will stay inside the loop for 200msec at most */
+    retry = kbdc->retry * 2;
+
     while ((f = read_status(kbdc) & KBDS_BUFFER_FULL)
 	    != KBDS_KBD_BUFFER_FULL) {
         if (f == KBDS_AUX_BUFFER_FULL) {
@@ -448,11 +490,13 @@ wait_for_kbd_data(struct atkbdc_softc *kbdc)
 static int
 wait_for_kbd_ack(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 200msec at most */
-    int retry = 10000;
+    int retry;
     int f;
     int b;
 
+    /* CPU will stay inside the loop for 200msec at most */
+    retry = kbdc->retry * 2;
+
     while (retry-- > 0) {
         if ((f = read_status(kbdc)) & KBDS_ANY_BUFFER_FULL) {
 	    DELAY(KBDD_DELAYTIME);
@@ -475,10 +519,12 @@ wait_for_kbd_ack(struct atkbdc_softc *kbdc)
 static int
 wait_for_aux_data(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 200msec at most */
-    int retry = 10000;
+    int retry;
     int f;
 
+    /* CPU will stay inside the loop for 200msec at most */
+    retry = kbdc->retry * 2;
+
     while ((f = read_status(kbdc) & KBDS_BUFFER_FULL)
 	    != KBDS_AUX_BUFFER_FULL) {
         if (f == KBDS_KBD_BUFFER_FULL) {
@@ -500,11 +546,13 @@ wait_for_aux_data(struct atkbdc_softc *kbdc)
 static int
 wait_for_aux_ack(struct atkbdc_softc *kbdc)
 {
-    /* CPU will stay inside the loop for 200msec at most */
-    int retry = 10000;
+    int retry;
     int f;
     int b;
 
+    /* CPU will stay inside the loop for 200msec at most */
+    retry = kbdc->retry * 2;
+
     while (retry-- > 0) {
         if ((f = read_status(kbdc)) & KBDS_ANY_BUFFER_FULL) {
 	    DELAY(KBDD_DELAYTIME);
diff --git a/sys/dev/atkbdc/atkbdc_ebus.c b/sys/dev/atkbdc/atkbdc_ebus.c
index 639203d548b..0dcb0a71e79 100644
--- a/sys/dev/atkbdc/atkbdc_ebus.c
+++ b/sys/dev/atkbdc/atkbdc_ebus.c
@@ -202,6 +202,7 @@ atkbdc_ebus_attach(device_t dev)
 		    "cannot determine command/data port resource\n");
 		return (ENXIO);
 	}
+	sc->retry = 5000;
 	sc->port0 = bus_alloc_resource(dev, SYS_RES_MEMORY, &rid, start, start,
 	    1, RF_ACTIVE);
 	if (sc->port0 == NULL) {
diff --git a/sys/dev/atkbdc/atkbdc_isa.c b/sys/dev/atkbdc/atkbdc_isa.c
index 975d2993571..10e172f4552 100644
--- a/sys/dev/atkbdc/atkbdc_isa.c
+++ b/sys/dev/atkbdc/atkbdc_isa.c
@@ -94,7 +94,7 @@ atkbdc_isa_probe(device_t dev)
 	u_long		count;
 	int		error;
 	int		rid;
-#if defined(__i386__)
+#if defined(__i386__) || defined(__amd64__)
 	bus_space_tag_t	tag;
 	bus_space_handle_t ioh1;
 	volatile int	i;
@@ -141,7 +141,7 @@ atkbdc_isa_probe(device_t dev)
 		return ENXIO;
 	}
 
-#if defined(__i386__)
+#if defined(__i386__) || defined(__amd64__)
 	/*
 	 * Check if we really have AT keyboard controller. Poll status
 	 * register until we get "all clear" indication. If no such
@@ -161,6 +161,8 @@ atkbdc_isa_probe(device_t dev)
 	if (i == 65535) {
 		bus_release_resource(dev, SYS_RES_IOPORT, 0, port0);
 		bus_release_resource(dev, SYS_RES_IOPORT, 1, port1);
+		if (bootverbose)
+			device_printf(dev, "AT keyboard controller not found\n");
 		return ENXIO;
 	}
 #endif
@@ -201,6 +203,7 @@ atkbdc_isa_attach(device_t dev)
 	}
 
 	rid = 0;
+	sc->retry = 5000;
 	sc->port0 = bus_alloc_resource_any(dev, SYS_RES_IOPORT, &rid,
 					   RF_ACTIVE);
 	if (sc->port0 == NULL)
diff --git a/sys/dev/atkbdc/atkbdcreg.h b/sys/dev/atkbdc/atkbdcreg.h
index 0715d9ece63..7ea26a66dd8 100644
--- a/sys/dev/atkbdc/atkbdcreg.h
+++ b/sys/dev/atkbdc/atkbdcreg.h
@@ -200,6 +200,7 @@ typedef struct atkbdc_softc {
     int lock;			/* FIXME: XXX not quite a semaphore... */
     kqueue kbd;			/* keyboard data queue */
     kqueue aux;			/* auxiliary data queue */
+    int retry;
 } atkbdc_softc_t; 
 
 enum kbdc_device_ivar {

From 9aac760c7db956424fa972f90efecf2fe2d41ff1 Mon Sep 17 00:00:00 2001
From: Yoshihiro Takahashi <nyan@FreeBSD.org>
Date: Thu, 29 Apr 2010 06:22:07 +0000
Subject: [PATCH 317/532] MFi386: revision 206553

  - Change printf() calls to uprintf() for sigreturn().
  - Normalize the messages to include both pid and thread name.
---
 sys/pc98/pc98/machdep.c | 12 ++++++++----
 1 file changed, 8 insertions(+), 4 deletions(-)

diff --git a/sys/pc98/pc98/machdep.c b/sys/pc98/pc98/machdep.c
index f470b5ef5ef..e0482e2198e 100644
--- a/sys/pc98/pc98/machdep.c
+++ b/sys/pc98/pc98/machdep.c
@@ -874,7 +874,8 @@ freebsd4_sigreturn(td, uap)
 		 * one less debugger trap, so allowing it is fairly harmless.
 		 */
 		if (!EFL_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF)) {
-			printf("freebsd4_sigreturn: eflags = 0x%x\n", eflags);
+			uprintf("pid %d (%s): freebsd4_sigreturn eflags = 0x%x\n",
+			    td->td_proc->p_pid, td->td_name, eflags);
 	    		return (EINVAL);
 		}
 
@@ -885,7 +886,8 @@ freebsd4_sigreturn(td, uap)
 		 */
 		cs = ucp->uc_mcontext.mc_cs;
 		if (!CS_SECURE(cs)) {
-			printf("freebsd4_sigreturn: cs = 0x%x\n", cs);
+			uprintf("pid %d (%s): freebsd4_sigreturn cs = 0x%x\n",
+			    td->td_proc->p_pid, td->td_name, cs);
 			ksiginfo_init_trap(&ksi);
 			ksi.ksi_signo = SIGBUS;
 			ksi.ksi_code = BUS_OBJERR;
@@ -986,7 +988,8 @@ sigreturn(td, uap)
 		 * one less debugger trap, so allowing it is fairly harmless.
 		 */
 		if (!EFL_SECURE(eflags & ~PSL_RF, regs->tf_eflags & ~PSL_RF)) {
-			printf("sigreturn: eflags = 0x%x\n", eflags);
+			uprintf("pid %d (%s): sigreturn eflags = 0x%x\n",
+			    td->td_proc->p_pid, td->td_name, eflags);
 	    		return (EINVAL);
 		}
 
@@ -997,7 +1000,8 @@ sigreturn(td, uap)
 		 */
 		cs = ucp->uc_mcontext.mc_cs;
 		if (!CS_SECURE(cs)) {
-			printf("sigreturn: cs = 0x%x\n", cs);
+			uprintf("pid %d (%s): sigreturn cs = 0x%x\n",
+			    td->td_proc->p_pid, td->td_name, cs);
 			ksiginfo_init_trap(&ksi);
 			ksi.ksi_signo = SIGBUS;
 			ksi.ksi_code = BUS_OBJERR;

From 9a9ae42a43422caf857b93276ab250a561cf72af Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:02:21 +0000
Subject: [PATCH 318/532] kern_ntptime: abstract time error check into a
 function

... to avoid code duplication

MFC after:	1 week
---
 sys/kern/kern_ntptime.c | 50 ++++++++++++++++++++++-------------------
 1 file changed, 27 insertions(+), 23 deletions(-)

diff --git a/sys/kern/kern_ntptime.c b/sys/kern/kern_ntptime.c
index 2d877404cf5..9a4630f8a09 100644
--- a/sys/kern/kern_ntptime.c
+++ b/sys/kern/kern_ntptime.c
@@ -198,22 +198,11 @@ static long pps_errcnt;			/* calibration errors */
 static void ntp_init(void);
 static void hardupdate(long offset);
 static void ntp_gettime1(struct ntptimeval *ntvp);
+static int ntp_is_time_error(void);
 
-static void
-ntp_gettime1(struct ntptimeval *ntvp)
+static int
+ntp_is_time_error(void)
 {
-	struct timespec atv;	/* nanosecond time */
-
-	GIANT_REQUIRED;
-
-	nanotime(&atv);
-	ntvp->time.tv_sec = atv.tv_sec;
-	ntvp->time.tv_nsec = atv.tv_nsec;
-	ntvp->maxerror = time_maxerror;
-	ntvp->esterror = time_esterror;
-	ntvp->tai = time_tai;
-	ntvp->time_state = time_state;
-
 	/*
 	 * Status word error decode. If any of these conditions occur,
 	 * an error is returned, instead of the status word. Most
@@ -243,6 +232,27 @@ ntp_gettime1(struct ntptimeval *ntvp)
 	 */
 	    (time_status & STA_PPSFREQ &&
 	    time_status & (STA_PPSWANDER | STA_PPSERROR)))
+		return (1);
+
+	return (0);
+}
+
+static void
+ntp_gettime1(struct ntptimeval *ntvp)
+{
+	struct timespec atv;	/* nanosecond time */
+
+	GIANT_REQUIRED;
+
+	nanotime(&atv);
+	ntvp->time.tv_sec = atv.tv_sec;
+	ntvp->time.tv_nsec = atv.tv_nsec;
+	ntvp->maxerror = time_maxerror;
+	ntvp->esterror = time_esterror;
+	ntvp->tai = time_tai;
+	ntvp->time_state = time_state;
+
+	if (ntp_is_time_error())
 		ntvp->time_state = TIME_ERROR;
 }
 
@@ -446,17 +456,11 @@ ntp_adjtime(struct thread *td, struct ntp_adjtime_args *uap)
 	 * Status word error decode. See comments in
 	 * ntp_gettime() routine.
 	 */
-	if ((time_status & (STA_UNSYNC | STA_CLOCKERR)) ||
-	    (time_status & (STA_PPSFREQ | STA_PPSTIME) &&
-	    !(time_status & STA_PPSSIGNAL)) ||
-	    (time_status & STA_PPSTIME &&
-	    time_status & STA_PPSJITTER) ||
-	    (time_status & STA_PPSFREQ &&
-	    time_status & (STA_PPSWANDER | STA_PPSERROR))) {
+	if (ntp_is_time_error())
 		td->td_retval[0] = TIME_ERROR;
-	} else {
+	else
 		td->td_retval[0] = time_state;
-	}
+
 done2:
 	mtx_unlock(&Giant);
 	return (error);

From 5c7e270fcd36ef2b0a71a9aebc2f8de0525914f9 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:02:46 +0000
Subject: [PATCH 319/532] periodically save system time to hardware time-of-day
 clock

This is done in kern_ntptime, perhaps not the best place.
This is done using resettodr().
Some features:
- make save period configurable via tunable and sysctl
- period of zero disables saving, setting a non-zero period re-enables
  it or reschedules it
- do saving only if system clock is ntp-synchronized
- save on shutdown

Discussed with:	des, Peter Jeremy <peterjeremy@acm.org>
X-Maybe:		save time near seconds boundary for better precision
MFC after:		2 weeks
---
 sys/kern/kern_ntptime.c | 65 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 65 insertions(+)

diff --git a/sys/kern/kern_ntptime.c b/sys/kern/kern_ntptime.c
index 9a4630f8a09..54c4b06b4f6 100644
--- a/sys/kern/kern_ntptime.c
+++ b/sys/kern/kern_ntptime.c
@@ -38,6 +38,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/param.h>
 #include <sys/systm.h>
 #include <sys/sysproto.h>
+#include <sys/eventhandler.h>
 #include <sys/kernel.h>
 #include <sys/priv.h>
 #include <sys/proc.h>
@@ -976,3 +977,67 @@ kern_adjtime(struct thread *td, struct timeval *delta, struct timeval *olddelta)
 	return (0);
 }
 
+static struct callout resettodr_callout;
+static int resettodr_period = 1800;
+
+static void
+periodic_resettodr(void *arg __unused)
+{
+
+	if (!ntp_is_time_error()) {
+		mtx_lock(&Giant);
+		resettodr();
+		mtx_unlock(&Giant);
+	}
+	if (resettodr_period > 0)
+		callout_schedule(&resettodr_callout, resettodr_period * hz);
+}
+
+static void
+shutdown_resettodr(void *arg __unused, int howto __unused)
+{
+
+	callout_drain(&resettodr_callout);
+	if (resettodr_period > 0 && !ntp_is_time_error()) {
+		mtx_lock(&Giant);
+		resettodr();
+		mtx_unlock(&Giant);
+	}
+}
+
+static int
+sysctl_resettodr_period(SYSCTL_HANDLER_ARGS)
+{
+	int error;
+
+	error = sysctl_handle_int(oidp, oidp->oid_arg1, oidp->oid_arg2, req);
+	if (error || !req->newptr)
+		return (error);
+	if (resettodr_period == 0)
+		callout_stop(&resettodr_callout);
+	else
+		callout_reset(&resettodr_callout, resettodr_period * hz,
+		    periodic_resettodr, NULL);
+	return (0);
+}
+
+SYSCTL_PROC(_machdep, OID_AUTO, rtc_save_period, CTLTYPE_INT|CTLFLAG_RW,
+	&resettodr_period, 1800, sysctl_resettodr_period, "I",
+	"Save system time to RTC with this period (in seconds)");
+TUNABLE_INT("machdep.rtc_save_period", &resettodr_period);
+
+static void
+start_periodic_resettodr(void *arg __unused)
+{
+
+	EVENTHANDLER_REGISTER(shutdown_pre_sync, shutdown_resettodr, NULL,
+	    SHUTDOWN_PRI_FIRST);
+	callout_init(&resettodr_callout, 1);
+	if (resettodr_period == 0)
+		return;
+	callout_reset(&resettodr_callout, resettodr_period * hz,
+	    periodic_resettodr, NULL);
+}
+
+SYSINIT(periodic_resettodr, SI_SUB_RUN_SCHEDULER, SI_ORDER_ANY - 1,
+	start_periodic_resettodr, NULL);

From 4f27c5edfe4a39248a2f54422d5de798b568a7f7 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:18:36 +0000
Subject: [PATCH 320/532] kern_ntptime: drop a comment that became stale after
 r207359

MFC after:	1 week
X-MFC after:	r207359
---
 sys/kern/kern_ntptime.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/sys/kern/kern_ntptime.c b/sys/kern/kern_ntptime.c
index 54c4b06b4f6..c427ac45134 100644
--- a/sys/kern/kern_ntptime.c
+++ b/sys/kern/kern_ntptime.c
@@ -453,10 +453,6 @@ ntp_adjtime(struct thread *td, struct ntp_adjtime_args *uap)
 	if (error)
 		goto done2;
 
-	/*
-	 * Status word error decode. See comments in
-	 * ntp_gettime() routine.
-	 */
 	if (ntp_is_time_error())
 		td->td_retval[0] = TIME_ERROR;
 	else

From e68d26fd5dbffdc0c64f711792f8adea0f5a73bc Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:55:51 +0000
Subject: [PATCH 321/532] Remove caddr_t casts.

Requested by:	bde
MFC after:	10 days
---
 sys/kern/kern_proc.c | 8 +++-----
 1 file changed, 3 insertions(+), 5 deletions(-)

diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index 5601ecf67cd..49a30971502 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -1084,11 +1084,9 @@ sysctl_out_proc_copyout(struct kinfo_proc *ki, struct sysctl_req *req)
 
 	if (req->flags & SCTL_MASK32) {
 		freebsd32_kinfo_proc_out(ki, &ki32);
-		error = SYSCTL_OUT(req, (caddr_t)&ki32,
-		    sizeof(struct kinfo_proc32));
+		error = SYSCTL_OUT(req, &ki32, sizeof(struct kinfo_proc32));
 	} else
-		error = SYSCTL_OUT(req, (caddr_t)ki,
-		    sizeof(struct kinfo_proc));
+		error = SYSCTL_OUT(req, ki, sizeof(struct kinfo_proc));
 	return (error);
 }
 #else
@@ -1096,7 +1094,7 @@ static int
 sysctl_out_proc_copyout(struct kinfo_proc *ki, struct sysctl_req *req)
 {
 
-	return (SYSCTL_OUT(req, (caddr_t)ki, sizeof(struct kinfo_proc)));
+	return (SYSCTL_OUT(req, ki, sizeof(struct kinfo_proc)));
 }
 #endif
 

From e86a87e97e26c059761438fcff9e13362dab21b6 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:57:25 +0000
Subject: [PATCH 322/532] In swap pager, do not free the non-requested pages
 from the run if they are wired. Kstack pages are wired, this change prepares
 swap pager for handling of long runs of kstack pages.

Noted and reviewed by:	alc
Tested by:	pho
MFC after:	2 weeks
---
 sys/vm/swap_pager.c | 14 +++++++++++---
 1 file changed, 11 insertions(+), 3 deletions(-)

diff --git a/sys/vm/swap_pager.c b/sys/vm/swap_pager.c
index f1d89d5e030..f47719bd1a1 100644
--- a/sys/vm/swap_pager.c
+++ b/sys/vm/swap_pager.c
@@ -378,6 +378,14 @@ static void swp_pager_meta_free(vm_object_t, vm_pindex_t, daddr_t);
 static void swp_pager_meta_free_all(vm_object_t);
 static daddr_t swp_pager_meta_ctl(vm_object_t, vm_pindex_t, int);
 
+static void
+swp_pager_free_nrpage(vm_page_t m)
+{
+
+	if (m->wire_count == 0)
+		vm_page_free(m);
+}
+
 /*
  * SWP_SIZECHECK() -	update swap_pager_full indication
  *	
@@ -1131,9 +1139,9 @@ swap_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 
 		vm_page_lock_queues();
 		for (k = 0; k < i; ++k)
-			vm_page_free(m[k]);
+			swp_pager_free_nrpage(m[k]);
 		for (k = j; k < count; ++k)
-			vm_page_free(m[k]);
+			swp_pager_free_nrpage(m[k]);
 		vm_page_unlock_queues();
 	}
 
@@ -1529,7 +1537,7 @@ swp_pager_async_iodone(struct buf *bp)
 				 */
 				m->valid = 0;
 				if (i != bp->b_pager.pg_reqpage)
-					vm_page_free(m);
+					swp_pager_free_nrpage(m);
 				else
 					vm_page_flash(m);
 				/*

From 6fb8c0c117d799bfb34f884996432dd8ee983625 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 29 Apr 2010 09:59:16 +0000
Subject: [PATCH 323/532] When doing kstack swapin, read as much pages in one
 run as possible.

Suggested and reviewed by:	alc (previous version)
Tested by:	pho
MFC after:	2 weeks
---
 sys/vm/vm_glue.c | 36 +++++++++++++++++++++++++-----------
 1 file changed, 25 insertions(+), 11 deletions(-)

diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index dbd5065d8b0..4eeaa4dd47f 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -538,23 +538,37 @@ static void
 vm_thread_swapin(struct thread *td)
 {
 	vm_object_t ksobj;
-	vm_page_t m, ma[KSTACK_MAX_PAGES];
-	int i, pages, rv;
+	vm_page_t ma[KSTACK_MAX_PAGES];
+	int i, j, k, pages, rv;
 
 	pages = td->td_kstack_pages;
 	ksobj = td->td_kstack_obj;
 	VM_OBJECT_LOCK(ksobj);
-	for (i = 0; i < pages; i++) {
-		m = vm_page_grab(ksobj, i, VM_ALLOC_NORMAL | VM_ALLOC_RETRY |
+	for (i = 0; i < pages; i++)
+		ma[i] = vm_page_grab(ksobj, i, VM_ALLOC_NORMAL | VM_ALLOC_RETRY |
 		    VM_ALLOC_WIRED);
-		if (m->valid != VM_PAGE_BITS_ALL) {
-			rv = vm_pager_get_pages(ksobj, &m, 1, 0);
+	for (i = 0; i < pages; i++) {
+		if (ma[i]->valid != VM_PAGE_BITS_ALL) {
+			KASSERT(ma[i]->oflags & VPO_BUSY,
+			    ("lost busy 1"));
+			vm_object_pip_add(ksobj, 1);
+			for (j = i + 1; j < pages; j++) {
+				KASSERT(ma[j]->valid == VM_PAGE_BITS_ALL ||
+				    (ma[j]->oflags & VPO_BUSY),
+				    ("lost busy 2"));
+				if (ma[j]->valid == VM_PAGE_BITS_ALL)
+					break;
+			}
+			rv = vm_pager_get_pages(ksobj, ma + i, j - i, 0);
 			if (rv != VM_PAGER_OK)
-				panic("vm_thread_swapin: cannot get kstack for proc: %d", td->td_proc->p_pid);
-			m = vm_page_lookup(ksobj, i);
-		}
-		ma[i] = m;
-		vm_page_wakeup(m);
+	panic("vm_thread_swapin: cannot get kstack for proc: %d",
+				    td->td_proc->p_pid);
+			vm_object_pip_wakeup(ksobj);
+			for (k = i; k < j; k++)
+				ma[k] = vm_page_lookup(ksobj, k);
+			vm_page_wakeup(ma[i]);
+		} else if (ma[i]->oflags & VPO_BUSY)
+			vm_page_wakeup(ma[i]);
 	}
 	VM_OBJECT_UNLOCK(ksobj);
 	pmap_qenter(td->td_kstack, ma, pages);

From deb3b115e23af3fd3016c4105caf4a3f9b01c7a3 Mon Sep 17 00:00:00 2001
From: Andriy Gapon <avg@FreeBSD.org>
Date: Thu, 29 Apr 2010 10:04:00 +0000
Subject: [PATCH 324/532] ffs_vfsops: restore alphabetic order of options in
 ffs_opts

The order was not correct only for nfsv4acls.
("no" prefix is ignored)

MFC after:	1 week
---
 sys/ufs/ffs/ffs_vfsops.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/ufs/ffs/ffs_vfsops.c b/sys/ufs/ffs/ffs_vfsops.c
index e4033686324..2963514508e 100644
--- a/sys/ufs/ffs/ffs_vfsops.c
+++ b/sys/ufs/ffs/ffs_vfsops.c
@@ -126,8 +126,8 @@ static struct buf_ops ffs_ops = {
 
 static const char *ffs_opts[] = { "acls", "async", "noatime", "noclusterr",
     "noclusterw", "noexec", "export", "force", "from", "multilabel", 
-    "snapshot", "nosuid", "suiddir", "nosymfollow", "sync",
-    "union", "nfsv4acls", NULL };
+    "nfsv4acls", "snapshot", "nosuid", "suiddir", "nosymfollow", "sync",
+    "union", NULL };
 
 static int
 ffs_mount(struct mount *mp)

From 82cea7e6f3da448dfadd071ac81f4a0b1e41a356 Mon Sep 17 00:00:00 2001
From: "Bjoern A. Zeeb" <bz@FreeBSD.org>
Date: Thu, 29 Apr 2010 11:52:42 +0000
Subject: [PATCH 325/532] MFP4: @176978-176982, 176984, 176990-176994, 177441

"Whitspace" churn after the VIMAGE/VNET whirls.

Remove the need for some "init" functions within the network
stack, like pim6_init(), icmp_init() or significantly shorten
others like ip6_init() and nd6_init(), using static initialization
again where possible and formerly missed.

Move (most) variables back to the place they used to be before the
container structs and VIMAGE_GLOABLS (before r185088) and try to
reduce the diff to stable/7 and earlier as good as possible,
to help out-of-tree consumers to update from 6.x or 7.x to 8 or 9.

This also removes some header file pollution for putatively
static global variables.

Revert VIMAGE specific changes in ipfilter::ip_auth.c, that are
no longer needed.

Reviewed by:	jhb
Discussed with:	rwatson
Sponsored by:	The FreeBSD Foundation
Sponsored by:	CK Software GmbH
MFC after:	6 days
---
 sys/contrib/ipfilter/netinet/ip_auth.c |   5 --
 sys/net/if.c                           |   6 +-
 sys/net/if_gif.c                       |  20 -----
 sys/net/if_gif.h                       |   4 -
 sys/net/route.c                        |  16 ++--
 sys/netinet/icmp6.h                    |   1 -
 sys/netinet/if_ether.c                 |  10 +--
 sys/netinet/in.c                       |  11 ++-
 sys/netinet/in_gif.c                   |   2 +
 sys/netinet/in_proto.c                 |   1 -
 sys/netinet/in_rmx.c                   |  21 ++---
 sys/netinet/ip_icmp.c                  |  59 +++++---------
 sys/netinet/ip_icmp.h                  |   1 -
 sys/netinet/ip_input.c                 |  72 ++++++++---------
 sys/netinet/ip_var.h                   |  13 ++-
 sys/netinet/tcp_hostcache.c            |   4 +-
 sys/netinet/tcp_input.c                |  39 +++++----
 sys/netinet/tcp_output.c               |  21 ++---
 sys/netinet/tcp_reass.c                |  21 ++---
 sys/netinet/tcp_sack.c                 |  17 ++--
 sys/netinet/tcp_subr.c                 |  90 ++++++---------------
 sys/netinet/tcp_syncache.c             |  19 ++---
 sys/netinet/tcp_timewait.c             |   8 +-
 sys/netinet/tcp_var.h                  |  56 ++-----------
 sys/netinet/udp_usrreq.c               |   7 +-
 sys/netinet/udp_var.h                  |  10 +--
 sys/netinet6/frag6.c                   |   5 +-
 sys/netinet6/icmp6.c                   |  21 ++---
 sys/netinet6/in6_gif.c                 |   9 +++
 sys/netinet6/in6_ifattach.c            |  11 ++-
 sys/netinet6/in6_proto.c               |  84 ++++++++++++-------
 sys/netinet6/in6_rmx.c                 |  23 +++---
 sys/netinet6/in6_src.c                 |   7 +-
 sys/netinet6/in6_var.h                 |   9 +--
 sys/netinet6/ip6_input.c               | 107 +++----------------------
 sys/netinet6/ip6_mroute.c              |  19 +----
 sys/netinet6/ip6_var.h                 |  53 ++++++------
 sys/netinet6/nd6.c                     |  77 +++++-------------
 sys/netinet6/nd6.h                     |   9 +--
 sys/netinet6/nd6_nbr.c                 |   8 +-
 sys/netinet6/nd6_rtr.c                 |  12 +--
 sys/netinet6/raw_ip6.c                 |   4 +-
 sys/netinet6/scope6.c                  |  12 ++-
 sys/netipsec/ah_var.h                  |   5 +-
 sys/netipsec/esp_var.h                 |   3 +-
 sys/netipsec/ipcomp_var.h              |   3 +-
 sys/netipsec/ipip_var.h                |   3 +-
 sys/netipsec/ipsec.c                   |   1 +
 sys/netipsec/ipsec.h                   |  52 ++++++------
 sys/netipsec/ipsec6.h                  |  11 +--
 sys/netipsec/key.c                     |  21 ++---
 sys/netipsec/xform_esp.c               |   3 +-
 52 files changed, 418 insertions(+), 688 deletions(-)

diff --git a/sys/contrib/ipfilter/netinet/ip_auth.c b/sys/contrib/ipfilter/netinet/ip_auth.c
index 19a12f47525..42d29d4f2a2 100644
--- a/sys/contrib/ipfilter/netinet/ip_auth.c
+++ b/sys/contrib/ipfilter/netinet/ip_auth.c
@@ -70,11 +70,6 @@ struct file;
 #include <netinet/in.h>
 #include <netinet/in_systm.h>
 #include <netinet/ip.h>
-#if !defined(_KERNEL) && defined(__FreeBSD_version) && \
-    __FreeBSD_version >= 800049
-# define V_ip_do_randomid	ip_do_randomid
-# define V_ip_id		ip_id
-#endif
 #if !defined(_KERNEL) && !defined(__osf__) && !defined(__sgi)
 # define	KERNEL
 # define	_KERNEL
diff --git a/sys/net/if.c b/sys/net/if.c
index 98c8afa4a2b..1acfb24dc51 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -163,9 +163,11 @@ static void	if_detach_internal(struct ifnet *, int);
 extern void	nd6_setmtu(struct ifnet *);
 #endif
 
+VNET_DEFINE(int, if_index);
+int	ifqmaxlen = IFQ_MAXLEN;
 VNET_DEFINE(struct ifnethead, ifnet);	/* depend on static init XXX */
 VNET_DEFINE(struct ifgrouphead, ifg_head);
-VNET_DEFINE(int, if_index);
+
 static VNET_DEFINE(int, if_indexlim) = 8;
 
 /* Table of ifnet by index. */
@@ -174,8 +176,6 @@ static VNET_DEFINE(struct ifindex_entry *, ifindex_table);
 #define	V_if_indexlim		VNET(if_indexlim)
 #define	V_ifindex_table		VNET(ifindex_table)
 
-int	ifqmaxlen = IFQ_MAXLEN;
-
 /*
  * The global network interface list (V_ifnet) and related state (such as
  * if_index, if_indexlim, and ifindex_table) are protected by an sxlock and
diff --git a/sys/net/if_gif.c b/sys/net/if_gif.c
index 22ef6abc8ee..ae0ff4e3369 100644
--- a/sys/net/if_gif.c
+++ b/sys/net/if_gif.c
@@ -94,20 +94,9 @@
  */
 static struct mtx gif_mtx;
 static MALLOC_DEFINE(M_GIF, "gif", "Generic Tunnel Interface");
-
 static VNET_DEFINE(LIST_HEAD(, gif_softc), gif_softc_list);
-
 #define	V_gif_softc_list	VNET(gif_softc_list)
 
-#ifdef INET
-VNET_DEFINE(int, ip_gif_ttl) = GIF_TTL;
-#define	V_ip_gif_ttl		VNET(ip_gif_ttl)
-#endif
-#ifdef INET6
-VNET_DEFINE(int, ip6_gif_hlim) = GIF_HLIM;
-#define	V_ip6_gif_hlim		VNET(ip6_gif_hlim)
-#endif
-
 void	(*ng_gif_input_p)(struct ifnet *ifp, struct mbuf **mp, int af);
 void	(*ng_gif_input_orphan_p)(struct ifnet *ifp, struct mbuf *m, int af);
 void	(*ng_gif_attach_p)(struct ifnet *ifp);
@@ -135,19 +124,11 @@ SYSCTL_NODE(_net_link, IFT_GIF, gif, CTLFLAG_RW, 0,
  */
 #define MAX_GIF_NEST 1
 #endif
-
 static VNET_DEFINE(int, max_gif_nesting) = MAX_GIF_NEST;
 #define	V_max_gif_nesting	VNET(max_gif_nesting)
-
 SYSCTL_VNET_INT(_net_link_gif, OID_AUTO, max_nesting, CTLFLAG_RW,
     &VNET_NAME(max_gif_nesting), 0, "Max nested tunnels");
 
-#ifdef INET6
-SYSCTL_DECL(_net_inet6_ip6);
-SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_GIF_HLIM, gifhlim, CTLFLAG_RW,
-    &VNET_NAME(ip6_gif_hlim), 0, "");
-#endif
-
 /*
  * By default, we disallow creation of multiple tunnels between the same
  * pair of addresses.  Some applications require this functionality so
@@ -159,7 +140,6 @@ static VNET_DEFINE(int, parallel_tunnels) = 1;
 static VNET_DEFINE(int, parallel_tunnels) = 0;
 #endif
 #define	V_parallel_tunnels	VNET(parallel_tunnels)
-
 SYSCTL_VNET_INT(_net_link_gif, OID_AUTO, parallel_tunnels, CTLFLAG_RW,
     &VNET_NAME(parallel_tunnels), 0, "Allow parallel tunnels?");
 
diff --git a/sys/net/if_gif.h b/sys/net/if_gif.h
index 18b7831a45b..5d0db6d1433 100644
--- a/sys/net/if_gif.h
+++ b/sys/net/if_gif.h
@@ -118,10 +118,6 @@ int gif_ioctl(struct ifnet *, u_long, caddr_t);
 int gif_set_tunnel(struct ifnet *, struct sockaddr *, struct sockaddr *);
 void gif_delete_tunnel(struct ifnet *);
 int gif_encapcheck(const struct mbuf *, int, int, void *);
-
-VNET_DECLARE(int, ip_gif_ttl);
-#define	V_ip_gif_ttl		VNET(ip_gif_ttl)
-
 #endif /* _KERNEL */
 
 #define GIFGOPTS	_IOWR('i', 150, struct ifreq)
diff --git a/sys/net/route.c b/sys/net/route.c
index e500ed13033..b45361e7168 100644
--- a/sys/net/route.c
+++ b/sys/net/route.c
@@ -88,16 +88,15 @@ SYSCTL_INT(_net, OID_AUTO, add_addr_allfibs, CTLFLAG_RW,
     &rt_add_addr_allfibs, 0, "");
 TUNABLE_INT("net.add_addr_allfibs", &rt_add_addr_allfibs);
 
-VNET_DEFINE(struct radix_node_head *, rt_tables);
-static VNET_DEFINE(uma_zone_t, rtzone);		/* Routing table UMA zone. */
-VNET_DEFINE(int, rttrash);		/* routes not in table but not freed */
 VNET_DEFINE(struct rtstat, rtstat);
-
-#define	V_rt_tables	VNET(rt_tables)
-#define	V_rtzone	VNET(rtzone)
-#define	V_rttrash	VNET(rttrash)
 #define	V_rtstat	VNET(rtstat)
 
+VNET_DEFINE(struct radix_node_head *, rt_tables);
+#define	V_rt_tables	VNET(rt_tables)
+
+VNET_DEFINE(int, rttrash);		/* routes not in table but not freed */
+#define	V_rttrash	VNET(rttrash)
+
 
 /* compare two sockaddr structures */
 #define	sa_equal(a1, a2) (bcmp((a1), (a2), (a1)->sa_len) == 0)
@@ -114,6 +113,9 @@ VNET_DEFINE(struct rtstat, rtstat);
  */
 #define RNTORT(p)	((struct rtentry *)(p))
 
+static VNET_DEFINE(uma_zone_t, rtzone);		/* Routing table UMA zone. */
+#define	V_rtzone	VNET(rtzone)
+
 #if 0
 /* default fib for tunnels to use */
 u_int tunnel_fib = 0;
diff --git a/sys/netinet/icmp6.h b/sys/netinet/icmp6.h
index c3a8ac7fcac..5faae7c1f2d 100644
--- a/sys/netinet/icmp6.h
+++ b/sys/netinet/icmp6.h
@@ -652,7 +652,6 @@ struct	rtentry;
 struct	rttimer;
 struct	in6_multi;
 # endif
-void	icmp6_init(void);
 void	icmp6_paramerror(struct mbuf *, int);
 void	icmp6_error(struct mbuf *, int, int, int);
 void	icmp6_error2(struct mbuf *, int, int, int, struct ifnet *);
diff --git a/sys/netinet/if_ether.c b/sys/netinet/if_ether.c
index 25fba9f10eb..08a162ff911 100644
--- a/sys/netinet/if_ether.c
+++ b/sys/netinet/if_ether.c
@@ -82,16 +82,15 @@ SYSCTL_DECL(_net_link_ether);
 SYSCTL_NODE(_net_link_ether, PF_INET, inet, CTLFLAG_RW, 0, "");
 SYSCTL_NODE(_net_link_ether, PF_ARP, arp, CTLFLAG_RW, 0, "");
 
-VNET_DEFINE(int, useloopback) = 1;	/* use loopback interface for
-					 * local traffic */
-
 /* timer values */
 static VNET_DEFINE(int, arpt_keep) = (20*60);	/* once resolved, good for 20
 						 * minutes */
+static VNET_DEFINE(int, arp_maxtries) = 5;
+VNET_DEFINE(int, useloopback) = 1;	/* use loopback interface for
+					 * local traffic */
+static VNET_DEFINE(int, arp_proxyall) = 0;
 static VNET_DEFINE(int, arpt_down) = 20;      /* keep incomplete entries for
 					       * 20 seconds */
-static VNET_DEFINE(int, arp_maxtries) = 5;
-static VNET_DEFINE(int, arp_proxyall);
 static VNET_DEFINE(struct arpstat, arpstat);  /* ARP statistics, see if_arp.h */
 
 #define	V_arpt_keep		VNET(arpt_keep)
@@ -103,7 +102,6 @@ static VNET_DEFINE(struct arpstat, arpstat);  /* ARP statistics, see if_arp.h */
 SYSCTL_VNET_INT(_net_link_ether_inet, OID_AUTO, max_age, CTLFLAG_RW,
 	&VNET_NAME(arpt_keep), 0,
 	"ARP entry lifetime in seconds");
-
 SYSCTL_VNET_INT(_net_link_ether_inet, OID_AUTO, maxtries, CTLFLAG_RW,
 	&VNET_NAME(arp_maxtries), 0,
 	"ARP resolution attempts before returning error");
diff --git a/sys/netinet/in.c b/sys/netinet/in.c
index 0a0fe472590..96bfa0ea67d 100644
--- a/sys/netinet/in.c
+++ b/sys/netinet/in.c
@@ -77,20 +77,19 @@ static int	in_ifinit(struct ifnet *,
 static void	in_purgemaddrs(struct ifnet *);
 
 static VNET_DEFINE(int, subnetsarelocal);
-static VNET_DEFINE(int, sameprefixcarponly);
-VNET_DECLARE(struct inpcbinfo, ripcbinfo);
-
 #define	V_subnetsarelocal		VNET(subnetsarelocal)
-#define	V_sameprefixcarponly		VNET(sameprefixcarponly)
-#define	V_ripcbinfo			VNET(ripcbinfo)
-
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, subnets_are_local, CTLFLAG_RW,
 	&VNET_NAME(subnetsarelocal), 0,
 	"Treat all subnets as directly connected");
+static VNET_DEFINE(int, sameprefixcarponly);
+#define	V_sameprefixcarponly		VNET(sameprefixcarponly)
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, same_prefix_carp_only, CTLFLAG_RW,
 	&VNET_NAME(sameprefixcarponly), 0,
 	"Refuse to create same prefixes on different interfaces");
 
+VNET_DECLARE(struct inpcbinfo, ripcbinfo);
+#define	V_ripcbinfo			VNET(ripcbinfo)
+
 /*
  * Return 1 if an internet address is for a ``local'' host
  * (one to which we have a connection).  If subnetsarelocal
diff --git a/sys/netinet/in_gif.c b/sys/netinet/in_gif.c
index 44b9961d44b..6c60390aa6a 100644
--- a/sys/netinet/in_gif.c
+++ b/sys/netinet/in_gif.c
@@ -85,6 +85,8 @@ struct protosw in_gif_protosw = {
 	.pr_usrreqs =		&rip_usrreqs
 };
 
+VNET_DEFINE(int, ip_gif_ttl) = GIF_TTL;
+#define	V_ip_gif_ttl		VNET(ip_gif_ttl)
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_GIF_TTL, gifttl, CTLFLAG_RW,
 	&VNET_NAME(ip_gif_ttl), 0, "");
 
diff --git a/sys/netinet/in_proto.c b/sys/netinet/in_proto.c
index d9cab845aee..6d72fbe3da3 100644
--- a/sys/netinet/in_proto.c
+++ b/sys/netinet/in_proto.c
@@ -208,7 +208,6 @@ struct protosw inetsw[] = {
 	.pr_flags =		PR_ATOMIC|PR_ADDR|PR_LASTHDR,
 	.pr_input =		icmp_input,
 	.pr_ctloutput =		rip_ctloutput,
-	.pr_init =		icmp_init,
 	.pr_usrreqs =		&rip_usrreqs
 },
 {
diff --git a/sys/netinet/in_rmx.c b/sys/netinet/in_rmx.c
index 6516277d0c4..5b8e0f619c8 100644
--- a/sys/netinet/in_rmx.c
+++ b/sys/netinet/in_rmx.c
@@ -131,22 +131,22 @@ in_matroute(void *v_arg, struct radix_node_head *head)
 	return rn;
 }
 
-static VNET_DEFINE(int, rtq_reallyold);
-static VNET_DEFINE(int, rtq_minreallyold);
-static VNET_DEFINE(int, rtq_toomany);
-
+static VNET_DEFINE(int, rtq_reallyold) = 60*60; /* one hour is "really old" */
 #define	V_rtq_reallyold		VNET(rtq_reallyold)
-#define	V_rtq_minreallyold	VNET(rtq_minreallyold)
-#define	V_rtq_toomany		VNET(rtq_toomany)
-
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_RTEXPIRE, rtexpire, CTLFLAG_RW,
     &VNET_NAME(rtq_reallyold), 0,
     "Default expiration time on dynamically learned routes");
 
+/* never automatically crank down to less */
+static VNET_DEFINE(int, rtq_minreallyold) = 10;
+#define	V_rtq_minreallyold	VNET(rtq_minreallyold)
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_RTMINEXPIRE, rtminexpire, CTLFLAG_RW,
     &VNET_NAME(rtq_minreallyold), 0,
     "Minimum time to attempt to hold onto dynamically learned routes");
 
+/* 128 cached routes is "too many" */
+static VNET_DEFINE(int, rtq_toomany) = 128;
+#define	V_rtq_toomany		VNET(rtq_toomany)
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_RTMAXCACHE, rtmaxcache, CTLFLAG_RW,
     &VNET_NAME(rtq_toomany), 0,
     "Upper limit on dynamically learned routes");
@@ -239,7 +239,7 @@ in_rtqkill(struct radix_node *rn, void *rock)
 }
 
 #define RTQ_TIMEOUT	60*10	/* run no less than once every ten minutes */
-static VNET_DEFINE(int, rtq_timeout);
+static VNET_DEFINE(int, rtq_timeout) = RTQ_TIMEOUT;
 static VNET_DEFINE(struct callout, rtq_timer);
 
 #define	V_rtq_timeout		VNET(rtq_timeout)
@@ -362,11 +362,6 @@ in_inithead(void **head, int off)
 	if (off == 0)		/* XXX MRT  see above */
 		return 1;	/* only do the rest for a real routing table */
 
-	V_rtq_reallyold = 60*60; /* one hour is "really old" */
-	V_rtq_minreallyold = 10; /* never automatically crank down to less */
-	V_rtq_toomany = 128;	 /* 128 cached routes is "too many" */
-	V_rtq_timeout = RTQ_TIMEOUT;
-
 	rnh = *head;
 	rnh->rnh_addaddr = in_addroute;
 	rnh->rnh_matchaddr = in_matroute;
diff --git a/sys/netinet/ip_icmp.c b/sys/netinet/ip_icmp.c
index b13bc7ca827..8dc08d9336c 100644
--- a/sys/netinet/ip_icmp.c
+++ b/sys/netinet/ip_icmp.c
@@ -75,65 +75,60 @@ __FBSDID("$FreeBSD$");
  * routines to turnaround packets back to the originator, and
  * host table maintenance routines.
  */
-
 VNET_DEFINE(struct icmpstat, icmpstat);
-static VNET_DEFINE(int, icmpmaskrepl);
-static VNET_DEFINE(u_int, icmpmaskfake);
-static VNET_DEFINE(int, drop_redirect);
-static VNET_DEFINE(int, log_redirect);
-static VNET_DEFINE(int, icmplim);
-static VNET_DEFINE(int, icmplim_output);
-static VNET_DEFINE(char, reply_src[IFNAMSIZ]);
-static VNET_DEFINE(int, icmp_rfi);
-static VNET_DEFINE(int, icmp_quotelen);
-static VNET_DEFINE(int, icmpbmcastecho);
-
-#define	V_icmpmaskrepl			VNET(icmpmaskrepl)
-#define	V_icmpmaskfake			VNET(icmpmaskfake)
-#define	V_drop_redirect			VNET(drop_redirect)
-#define	V_log_redirect			VNET(log_redirect)
-#define	V_icmplim			VNET(icmplim)
-#define	V_icmplim_output		VNET(icmplim_output)
-#define	V_reply_src			VNET(reply_src)
-#define	V_icmp_rfi			VNET(icmp_rfi)
-#define	V_icmp_quotelen			VNET(icmp_quotelen)
-#define	V_icmpbmcastecho		VNET(icmpbmcastecho)
-
 SYSCTL_VNET_STRUCT(_net_inet_icmp, ICMPCTL_STATS, stats, CTLFLAG_RW,
 	&VNET_NAME(icmpstat), icmpstat, "");
 
+static VNET_DEFINE(int, icmpmaskrepl) = 0;
+#define	V_icmpmaskrepl			VNET(icmpmaskrepl)
 SYSCTL_VNET_INT(_net_inet_icmp, ICMPCTL_MASKREPL, maskrepl, CTLFLAG_RW,
 	&VNET_NAME(icmpmaskrepl), 0,
 	"Reply to ICMP Address Mask Request packets.");
 
+static VNET_DEFINE(u_int, icmpmaskfake) = 0;
+#define	V_icmpmaskfake			VNET(icmpmaskfake)
 SYSCTL_VNET_UINT(_net_inet_icmp, OID_AUTO, maskfake, CTLFLAG_RW,
 	&VNET_NAME(icmpmaskfake), 0,
 	"Fake reply to ICMP Address Mask Request packets.");
 
+static VNET_DEFINE(int, drop_redirect) = 0;
+#define	V_drop_redirect			VNET(drop_redirect)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, drop_redirect, CTLFLAG_RW,
 	&VNET_NAME(drop_redirect), 0,
 	"Ignore ICMP redirects");
 
+static VNET_DEFINE(int, log_redirect) = 0;
+#define	V_log_redirect			VNET(log_redirect)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, log_redirect, CTLFLAG_RW,
 	&VNET_NAME(log_redirect), 0,
 	"Log ICMP redirects to the console");
 
+static VNET_DEFINE(int, icmplim) = 200;
+#define	V_icmplim			VNET(icmplim)
 SYSCTL_VNET_INT(_net_inet_icmp, ICMPCTL_ICMPLIM, icmplim, CTLFLAG_RW,
 	&VNET_NAME(icmplim), 0,
 	"Maximum number of ICMP responses per second");
 
+static VNET_DEFINE(int, icmplim_output) = 1;
+#define	V_icmplim_output		VNET(icmplim_output)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, icmplim_output, CTLFLAG_RW,
 	&VNET_NAME(icmplim_output), 0,
 	"Enable rate limiting of ICMP responses");
 
+static VNET_DEFINE(char, reply_src[IFNAMSIZ]);
+#define	V_reply_src			VNET(reply_src)
 SYSCTL_VNET_STRING(_net_inet_icmp, OID_AUTO, reply_src, CTLFLAG_RW,
 	&VNET_NAME(reply_src), IFNAMSIZ,
 	"icmp reply source for non-local packets.");
 
+static VNET_DEFINE(int, icmp_rfi) = 0;
+#define	V_icmp_rfi			VNET(icmp_rfi)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, reply_from_interface, CTLFLAG_RW,
 	&VNET_NAME(icmp_rfi), 0,
 	"ICMP reply from incoming interface for non-local packets");
 
+static VNET_DEFINE(int, icmp_quotelen) = 8;
+#define	V_icmp_quotelen			VNET(icmp_quotelen)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, quotelen, CTLFLAG_RW,
 	&VNET_NAME(icmp_quotelen), 0,
 	"Number of bytes from original packet to quote in ICMP reply");
@@ -141,7 +136,8 @@ SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, quotelen, CTLFLAG_RW,
 /*
  * ICMP broadcast echo sysctl
  */
-
+static VNET_DEFINE(int, icmpbmcastecho) = 0;
+#define	V_icmpbmcastecho		VNET(icmpbmcastecho)
 SYSCTL_VNET_INT(_net_inet_icmp, OID_AUTO, bmcastecho, CTLFLAG_RW,
 	&VNET_NAME(icmpbmcastecho), 0,
 	"");
@@ -156,21 +152,6 @@ static void	icmp_send(struct mbuf *, struct mbuf *);
 
 extern	struct protosw inetsw[];
 
-void
-icmp_init(void)
-{
-
-	V_icmpmaskrepl = 0;
-	V_icmpmaskfake = 0;
-	V_drop_redirect = 0;
-	V_log_redirect = 0;
-	V_icmplim = 200;
-	V_icmplim_output = 1;
-	V_icmp_rfi = 0;
-	V_icmp_quotelen = 8;
-	V_icmpbmcastecho = 0;
-}
-
 /*
  * Kernel module interface for updating icmpstat.  The argument is an index
  * into icmpstat treated as an array of u_long.  While this encodes the
diff --git a/sys/netinet/ip_icmp.h b/sys/netinet/ip_icmp.h
index e4ee7f7402c..9cabdb58c57 100644
--- a/sys/netinet/ip_icmp.h
+++ b/sys/netinet/ip_icmp.h
@@ -208,7 +208,6 @@ struct icmp {
 #ifdef _KERNEL
 void	icmp_error(struct mbuf *, int, int, uint32_t, int);
 void	icmp_input(struct mbuf *, int);
-void	icmp_init(void);
 int	ip_next_mtu(int, int);
 #endif
 
diff --git a/sys/netinet/ip_input.c b/sys/netinet/ip_input.c
index fdef6457b3e..a17907c397c 100644
--- a/sys/netinet/ip_input.c
+++ b/sys/netinet/ip_input.c
@@ -89,66 +89,40 @@ __FBSDID("$FreeBSD$");
 CTASSERT(sizeof(struct ip) == 20);
 #endif
 
-static VNET_DEFINE(int, ipsendredirects) = 1;	/* XXX */
-static VNET_DEFINE(int, ip_checkinterface);
-static VNET_DEFINE(int, ip_keepfaith);
-static VNET_DEFINE(int, ip_sendsourcequench);
-
-#define	V_ipsendredirects	VNET(ipsendredirects)
-#define	V_ip_checkinterface	VNET(ip_checkinterface)
-#define	V_ip_keepfaith		VNET(ip_keepfaith)
-#define	V_ip_sendsourcequench	VNET(ip_sendsourcequench)
-
-VNET_DEFINE(int, ip_defttl) = IPDEFTTL;
-VNET_DEFINE(int, ip_do_randomid);
-VNET_DEFINE(int, ipforwarding);
-
-VNET_DEFINE(struct in_ifaddrhead, in_ifaddrhead);  /* first inet address */
-VNET_DEFINE(struct in_ifaddrhashhead *, in_ifaddrhashtbl); /* inet addr hash table  */
-VNET_DEFINE(u_long, in_ifaddrhmask);		/* mask for hash table */
-VNET_DEFINE(struct ipstat, ipstat);
-
-static VNET_DEFINE(int, ip_rsvp_on);
-VNET_DEFINE(struct socket *, ip_rsvpd);
-VNET_DEFINE(int, rsvp_on);
-
-#define	V_ip_rsvp_on		VNET(ip_rsvp_on)
-
-static VNET_DEFINE(TAILQ_HEAD(ipqhead, ipq), ipq[IPREASS_NHASH]);
-static VNET_DEFINE(int, maxnipq);  /* Administrative limit on # reass queues. */
-static VNET_DEFINE(int, maxfragsperpacket);
-static VNET_DEFINE(int, nipq);			/* Total # of reass queues */
-
-#define	V_ipq			VNET(ipq)
-#define	V_maxnipq		VNET(maxnipq)
-#define	V_maxfragsperpacket	VNET(maxfragsperpacket)
-#define	V_nipq			VNET(nipq)
-
-VNET_DEFINE(int, ipstealth);
-
 struct	rwlock in_ifaddr_lock;
 RW_SYSINIT(in_ifaddr_lock, &in_ifaddr_lock, "in_ifaddr_lock");
 
+VNET_DEFINE(int, rsvp_on);
+
+VNET_DEFINE(int, ipforwarding);
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_FORWARDING, forwarding, CTLFLAG_RW,
     &VNET_NAME(ipforwarding), 0,
     "Enable IP forwarding between interfaces");
 
+static VNET_DEFINE(int, ipsendredirects) = 1;	/* XXX */
+#define	V_ipsendredirects	VNET(ipsendredirects)
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_SENDREDIRECTS, redirect, CTLFLAG_RW,
     &VNET_NAME(ipsendredirects), 0,
     "Enable sending IP redirects");
 
+VNET_DEFINE(int, ip_defttl) = IPDEFTTL;
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_DEFTTL, ttl, CTLFLAG_RW,
     &VNET_NAME(ip_defttl), 0,
     "Maximum TTL on IP packets");
 
+static VNET_DEFINE(int, ip_keepfaith);
+#define	V_ip_keepfaith		VNET(ip_keepfaith)
 SYSCTL_VNET_INT(_net_inet_ip, IPCTL_KEEPFAITH, keepfaith, CTLFLAG_RW,
     &VNET_NAME(ip_keepfaith), 0,
     "Enable packet capture for FAITH IPv4->IPv6 translater daemon");
 
+static VNET_DEFINE(int, ip_sendsourcequench);
+#define	V_ip_sendsourcequench	VNET(ip_sendsourcequench)
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, sendsourcequench, CTLFLAG_RW,
     &VNET_NAME(ip_sendsourcequench), 0,
     "Enable the transmission of source quench packets");
 
+VNET_DEFINE(int, ip_do_randomid);
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, random_id, CTLFLAG_RW,
     &VNET_NAME(ip_do_randomid), 0,
     "Assign random ip_id values");
@@ -166,6 +140,8 @@ SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, random_id, CTLFLAG_RW,
  * to the loopback interface instead of the interface where the
  * packets for those addresses are received.
  */
+static VNET_DEFINE(int, ip_checkinterface);
+#define	V_ip_checkinterface	VNET(ip_checkinterface)
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, check_interface, CTLFLAG_RW,
     &VNET_NAME(ip_checkinterface), 0,
     "Verify packet arrives on correct interface");
@@ -182,16 +158,22 @@ static struct netisr_handler ip_nh = {
 extern	struct domain inetdomain;
 extern	struct protosw inetsw[];
 u_char	ip_protox[IPPROTO_MAX];
+VNET_DEFINE(struct in_ifaddrhead, in_ifaddrhead);  /* first inet address */
+VNET_DEFINE(struct in_ifaddrhashhead *, in_ifaddrhashtbl); /* inet addr hash table  */
+VNET_DEFINE(u_long, in_ifaddrhmask);		/* mask for hash table */
 
+VNET_DEFINE(struct ipstat, ipstat);
 SYSCTL_VNET_STRUCT(_net_inet_ip, IPCTL_STATS, stats, CTLFLAG_RW,
     &VNET_NAME(ipstat), ipstat,
     "IP statistics (struct ipstat, netinet/ip_var.h)");
 
 static VNET_DEFINE(uma_zone_t, ipq_zone);
-#define	V_ipq_zone		VNET(ipq_zone)
-
+static VNET_DEFINE(TAILQ_HEAD(ipqhead, ipq), ipq[IPREASS_NHASH]);
 static struct mtx ipqlock;
 
+#define	V_ipq_zone		VNET(ipq_zone)
+#define	V_ipq			VNET(ipq)
+
 #define	IPQ_LOCK()	mtx_lock(&ipqlock)
 #define	IPQ_UNLOCK()	mtx_unlock(&ipqlock)
 #define	IPQ_LOCK_INIT()	mtx_init(&ipqlock, "ipqlock", NULL, MTX_DEF)
@@ -201,10 +183,16 @@ static void	maxnipq_update(void);
 static void	ipq_zone_change(void *);
 static void	ip_drain_locked(void);
 
+static VNET_DEFINE(int, maxnipq);  /* Administrative limit on # reass queues. */
+static VNET_DEFINE(int, nipq);			/* Total # of reass queues */
+#define	V_maxnipq		VNET(maxnipq)
+#define	V_nipq			VNET(nipq)
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, fragpackets, CTLFLAG_RD,
     &VNET_NAME(nipq), 0,
     "Current number of IPv4 fragment reassembly queue entries");
 
+static VNET_DEFINE(int, maxfragsperpacket);
+#define	V_maxfragsperpacket	VNET(maxfragsperpacket)
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, maxfragsperpacket, CTLFLAG_RW,
     &VNET_NAME(maxfragsperpacket), 0,
     "Maximum number of IPv4 fragments allowed per packet");
@@ -217,6 +205,7 @@ SYSCTL_INT(_net_inet_ip, IPCTL_DEFMTU, mtu, CTLFLAG_RW,
 #endif
 
 #ifdef IPSTEALTH
+VNET_DEFINE(int, ipstealth);
 SYSCTL_VNET_INT(_net_inet_ip, OID_AUTO, stealth, CTLFLAG_RW,
     &VNET_NAME(ipstealth), 0,
     "IP stealth mode, no TTL decrementation on forwarding");
@@ -1740,6 +1729,11 @@ makedummy:
  * locking.  This code remains in ip_input.c as ip_mroute.c is optionally
  * compiled.
  */
+static VNET_DEFINE(int, ip_rsvp_on);
+VNET_DEFINE(struct socket *, ip_rsvpd);
+
+#define	V_ip_rsvp_on		VNET(ip_rsvp_on)
+
 int
 ip_rsvp_init(struct socket *so)
 {
diff --git a/sys/netinet/ip_var.h b/sys/netinet/ip_var.h
index 389ad6ee316..4415001fc27 100644
--- a/sys/netinet/ip_var.h
+++ b/sys/netinet/ip_var.h
@@ -181,9 +181,13 @@ VNET_DECLARE(int, ipforwarding);		/* ip forwarding */
 #ifdef IPSTEALTH
 VNET_DECLARE(int, ipstealth);			/* stealth forwarding */
 #endif
-VNET_DECLARE(int, rsvp_on);
+extern u_char	ip_protox[];
 VNET_DECLARE(struct socket *, ip_rsvpd);	/* reservation protocol daemon*/
 VNET_DECLARE(struct socket *, ip_mrouter);	/* multicast routing daemon */
+extern int	(*legal_vif_num)(int);
+extern u_long	(*ip_mcast_src)(int);
+VNET_DECLARE(int, rsvp_on);
+extern struct	pr_usrreqs rip_usrreqs;
 
 #define	V_ipstat		VNET(ipstat)
 #define	V_ip_id			VNET(ip_id)
@@ -192,14 +196,9 @@ VNET_DECLARE(struct socket *, ip_mrouter);	/* multicast routing daemon */
 #ifdef IPSTEALTH
 #define	V_ipstealth		VNET(ipstealth)
 #endif
-#define	V_rsvp_on		VNET(rsvp_on)
 #define	V_ip_rsvpd		VNET(ip_rsvpd)
 #define	V_ip_mrouter		VNET(ip_mrouter)
-
-extern u_char	ip_protox[];
-extern int	(*legal_vif_num)(int);
-extern u_long	(*ip_mcast_src)(int);
-extern struct	pr_usrreqs rip_usrreqs;
+#define	V_rsvp_on		VNET(rsvp_on)
 
 void	inp_freemoptions(struct ip_moptions *);
 int	inp_getmoptions(struct inpcb *, struct sockopt *);
diff --git a/sys/netinet/tcp_hostcache.c b/sys/netinet/tcp_hostcache.c
index a0c4012f21f..d20adb7567d 100644
--- a/sys/netinet/tcp_hostcache.c
+++ b/sys/netinet/tcp_hostcache.c
@@ -107,9 +107,9 @@ __FBSDID("$FreeBSD$");
 #define TCP_HOSTCACHE_PRUNE		5*60	/* every 5 minutes */
 
 static VNET_DEFINE(struct tcp_hostcache, tcp_hostcache);
-static VNET_DEFINE(struct callout, tcp_hc_callout);
-
 #define	V_tcp_hostcache		VNET(tcp_hostcache)
+
+static VNET_DEFINE(struct callout, tcp_hc_callout);
 #define	V_tcp_hc_callout	VNET(tcp_hc_callout)
 
 static struct hc_metrics *tcp_hc_lookup(struct in_conninfo *);
diff --git a/sys/netinet/tcp_input.c b/sys/netinet/tcp_input.c
index 0254cff0b47..cbba9cd4ed0 100644
--- a/sys/netinet/tcp_input.c
+++ b/sys/netinet/tcp_input.c
@@ -99,20 +99,6 @@ __FBSDID("$FreeBSD$");
 static const int tcprexmtthresh = 3;
 
 VNET_DEFINE(struct tcpstat, tcpstat);
-VNET_DEFINE(int, blackhole);
-VNET_DEFINE(int, tcp_delack_enabled);
-VNET_DEFINE(int, drop_synfin);
-VNET_DEFINE(int, tcp_do_rfc3042);
-VNET_DEFINE(int, tcp_do_rfc3390);
-VNET_DEFINE(int, tcp_do_ecn);
-VNET_DEFINE(int, tcp_ecn_maxretries);
-VNET_DEFINE(int, tcp_insecure_rst);
-VNET_DEFINE(int, tcp_do_autorcvbuf);
-VNET_DEFINE(int, tcp_autorcvbuf_inc);
-VNET_DEFINE(int, tcp_autorcvbuf_max);
-VNET_DEFINE(int, tcp_do_rfc3465);
-VNET_DEFINE(int, tcp_abc_l_var);
-
 SYSCTL_VNET_STRUCT(_net_inet_tcp, TCPCTL_STATS, stats, CTLFLAG_RW,
     &VNET_NAME(tcpstat), tcpstat,
     "TCP statistics (struct tcpstat, netinet/tcp_var.h)");
@@ -122,56 +108,79 @@ SYSCTL_INT(_net_inet_tcp, OID_AUTO, log_in_vain, CTLFLAG_RW,
     &tcp_log_in_vain, 0,
     "Log all incoming TCP segments to closed ports");
 
+VNET_DEFINE(int, blackhole) = 0;
+#define	V_blackhole		VNET(blackhole)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, blackhole, CTLFLAG_RW,
     &VNET_NAME(blackhole), 0,
     "Do not send RST on segments to closed ports");
 
+VNET_DEFINE(int, tcp_delack_enabled) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, delayed_ack, CTLFLAG_RW,
     &VNET_NAME(tcp_delack_enabled), 0,
     "Delay ACK to try and piggyback it onto a data packet");
 
+VNET_DEFINE(int, drop_synfin) = 0;
+#define	V_drop_synfin		VNET(drop_synfin)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, drop_synfin, CTLFLAG_RW,
     &VNET_NAME(drop_synfin), 0,
     "Drop TCP packets with SYN+FIN set");
 
+VNET_DEFINE(int, tcp_do_rfc3042) = 1;
+#define	V_tcp_do_rfc3042	VNET(tcp_do_rfc3042)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, rfc3042, CTLFLAG_RW,
     &VNET_NAME(tcp_do_rfc3042), 0,
     "Enable RFC 3042 (Limited Transmit)");
 
+VNET_DEFINE(int, tcp_do_rfc3390) = 1;
+#define	V_tcp_do_rfc3390	VNET(tcp_do_rfc3390)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, rfc3390, CTLFLAG_RW,
     &VNET_NAME(tcp_do_rfc3390), 0,
     "Enable RFC 3390 (Increasing TCP's Initial Congestion Window)");
 
+VNET_DEFINE(int, tcp_do_rfc3465) = 1;
+#define	V_tcp_do_rfc3465	VNET(tcp_do_rfc3465)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, rfc3465, CTLFLAG_RW,
     &VNET_NAME(tcp_do_rfc3465), 0,
     "Enable RFC 3465 (Appropriate Byte Counting)");
 
+VNET_DEFINE(int, tcp_abc_l_var) = 2;
+#define	V_tcp_abc_l_var		VNET(tcp_abc_l_var)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, abc_l_var, CTLFLAG_RW,
     &VNET_NAME(tcp_abc_l_var), 2,
     "Cap the max cwnd increment during slow-start to this number of segments");
 
 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, ecn, CTLFLAG_RW, 0, "TCP ECN");
 
+VNET_DEFINE(int, tcp_do_ecn) = 0;
 SYSCTL_VNET_INT(_net_inet_tcp_ecn, OID_AUTO, enable, CTLFLAG_RW,
     &VNET_NAME(tcp_do_ecn), 0,
     "TCP ECN support");
 
+VNET_DEFINE(int, tcp_ecn_maxretries) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp_ecn, OID_AUTO, maxretries, CTLFLAG_RW,
     &VNET_NAME(tcp_ecn_maxretries), 0,
     "Max retries before giving up on ECN");
 
+VNET_DEFINE(int, tcp_insecure_rst) = 0;
+#define	V_tcp_insecure_rst	VNET(tcp_insecure_rst)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, insecure_rst, CTLFLAG_RW,
     &VNET_NAME(tcp_insecure_rst), 0,
     "Follow the old (insecure) criteria for accepting RST packets");
 
+VNET_DEFINE(int, tcp_do_autorcvbuf) = 1;
+#define	V_tcp_do_autorcvbuf	VNET(tcp_do_autorcvbuf)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, recvbuf_auto, CTLFLAG_RW,
     &VNET_NAME(tcp_do_autorcvbuf), 0,
     "Enable automatic receive buffer sizing");
 
+VNET_DEFINE(int, tcp_autorcvbuf_inc) = 16*1024;
+#define	V_tcp_autorcvbuf_inc	VNET(tcp_autorcvbuf_inc)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, recvbuf_inc, CTLFLAG_RW,
     &VNET_NAME(tcp_autorcvbuf_inc), 0,
     "Incrementor step size of automatic receive buffer");
 
+VNET_DEFINE(int, tcp_autorcvbuf_max) = 256*1024;
+#define	V_tcp_autorcvbuf_max	VNET(tcp_autorcvbuf_max)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, recvbuf_max, CTLFLAG_RW,
     &VNET_NAME(tcp_autorcvbuf_max), 0,
     "Max size of automatic receive buffer");
@@ -181,8 +190,8 @@ SYSCTL_INT(_net_inet_tcp, OID_AUTO, read_locking, CTLFLAG_RW,
     &tcp_read_locking, 0, "Enable read locking strategy");
 
 VNET_DEFINE(struct inpcbhead, tcb);
-VNET_DEFINE(struct inpcbinfo, tcbinfo);
 #define	tcb6	tcb  /* for KAME src sync over BSD*'s */
+VNET_DEFINE(struct inpcbinfo, tcbinfo);
 
 static void	 tcp_dooptions(struct tcpopt *, u_char *, int, int);
 static void	 tcp_do_segment(struct mbuf *, struct tcphdr *,
diff --git a/sys/netinet/tcp_output.c b/sys/netinet/tcp_output.c
index f9d1b6371e8..50973c4d455 100644
--- a/sys/netinet/tcp_output.c
+++ b/sys/netinet/tcp_output.c
@@ -87,43 +87,46 @@ __FBSDID("$FreeBSD$");
 extern struct mbuf *m_copypack();
 #endif
 
-VNET_DEFINE(int, path_mtu_discovery);
-VNET_DEFINE(int, ss_fltsz);
-VNET_DEFINE(int, ss_fltsz_local);
-VNET_DEFINE(int, tcp_do_newreno);
-VNET_DEFINE(int, tcp_do_tso);
-VNET_DEFINE(int, tcp_do_autosndbuf);
-VNET_DEFINE(int, tcp_autosndbuf_inc);
-VNET_DEFINE(int, tcp_autosndbuf_max);
-
+VNET_DEFINE(int, path_mtu_discovery) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, path_mtu_discovery, CTLFLAG_RW,
 	&VNET_NAME(path_mtu_discovery), 1,
 	"Enable Path MTU Discovery");
 
+VNET_DEFINE(int, ss_fltsz) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, slowstart_flightsize, CTLFLAG_RW,
 	&VNET_NAME(ss_fltsz), 1,
 	"Slow start flight size");
 
+VNET_DEFINE(int, ss_fltsz_local) = 4;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, local_slowstart_flightsize,
 	CTLFLAG_RW, &VNET_NAME(ss_fltsz_local), 1,
 	"Slow start flight size for local networks");
 
+VNET_DEFINE(int, tcp_do_newreno) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, newreno, CTLFLAG_RW,
 	&VNET_NAME(tcp_do_newreno), 0,
 	"Enable NewReno Algorithms");
 
+VNET_DEFINE(int, tcp_do_tso) = 1;
+#define	V_tcp_do_tso		VNET(tcp_do_tso)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, tso, CTLFLAG_RW,
 	&VNET_NAME(tcp_do_tso), 0,
 	"Enable TCP Segmentation Offload");
 
+VNET_DEFINE(int, tcp_do_autosndbuf) = 1;
+#define	V_tcp_do_autosndbuf	VNET(tcp_do_autosndbuf)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, sendbuf_auto, CTLFLAG_RW,
 	&VNET_NAME(tcp_do_autosndbuf), 0,
 	"Enable automatic send buffer sizing");
 
+VNET_DEFINE(int, tcp_autosndbuf_inc) = 8*1024;
+#define	V_tcp_autosndbuf_inc	VNET(tcp_autosndbuf_inc)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, sendbuf_inc, CTLFLAG_RW,
 	&VNET_NAME(tcp_autosndbuf_inc), 0,
 	"Incrementor step size of automatic send buffer");
 
+VNET_DEFINE(int, tcp_autosndbuf_max) = 256*1024;
+#define	V_tcp_autosndbuf_max	VNET(tcp_autosndbuf_max)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, sendbuf_max, CTLFLAG_RW,
 	&VNET_NAME(tcp_autosndbuf_max), 0,
 	"Max size of automatic send buffer");
diff --git a/sys/netinet/tcp_reass.c b/sys/netinet/tcp_reass.c
index cd7eb1ce39f..bfb53d3a742 100644
--- a/sys/netinet/tcp_reass.c
+++ b/sys/netinet/tcp_reass.c
@@ -74,30 +74,28 @@ __FBSDID("$FreeBSD$");
 #include <netinet/tcp_debug.h>
 #endif /* TCPDEBUG */
 
-static VNET_DEFINE(int, tcp_reass_maxseg);
-VNET_DEFINE(int, tcp_reass_qsize);
-static VNET_DEFINE(int, tcp_reass_maxqlen);
-static VNET_DEFINE(int, tcp_reass_overflows);
-
-#define	V_tcp_reass_maxseg		VNET(tcp_reass_maxseg)
-#define	V_tcp_reass_maxqlen		VNET(tcp_reass_maxqlen)
-#define	V_tcp_reass_overflows		VNET(tcp_reass_overflows)
-
 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, reass, CTLFLAG_RW, 0,
     "TCP Segment Reassembly Queue");
 
+static VNET_DEFINE(int, tcp_reass_maxseg) = 0;
+#define	V_tcp_reass_maxseg		VNET(tcp_reass_maxseg)
 SYSCTL_VNET_INT(_net_inet_tcp_reass, OID_AUTO, maxsegments, CTLFLAG_RDTUN,
     &VNET_NAME(tcp_reass_maxseg), 0,
     "Global maximum number of TCP Segments in Reassembly Queue");
 
+VNET_DEFINE(int, tcp_reass_qsize) = 0;
 SYSCTL_VNET_INT(_net_inet_tcp_reass, OID_AUTO, cursegments, CTLFLAG_RD,
     &VNET_NAME(tcp_reass_qsize), 0,
     "Global number of TCP Segments currently in Reassembly Queue");
 
+static VNET_DEFINE(int, tcp_reass_maxqlen) = 48;
+#define	V_tcp_reass_maxqlen		VNET(tcp_reass_maxqlen)
 SYSCTL_VNET_INT(_net_inet_tcp_reass, OID_AUTO, maxqlen, CTLFLAG_RW,
     &VNET_NAME(tcp_reass_maxqlen), 0,
     "Maximum number of TCP Segments per individual Reassembly Queue");
 
+static VNET_DEFINE(int, tcp_reass_overflows) = 0;
+#define	V_tcp_reass_overflows		VNET(tcp_reass_overflows)
 SYSCTL_VNET_INT(_net_inet_tcp_reass, OID_AUTO, overflows, CTLFLAG_RD,
     &VNET_NAME(tcp_reass_overflows), 0,
     "Global number of TCP Segment Reassembly Queue Overflows");
@@ -117,11 +115,6 @@ void
 tcp_reass_init(void)
 {
 
-	V_tcp_reass_maxseg = 0;
-	V_tcp_reass_qsize = 0;
-	V_tcp_reass_maxqlen = 48;
-	V_tcp_reass_overflows = 0;
-
 	V_tcp_reass_maxseg = nmbclusters / 16;
 	TUNABLE_INT_FETCH("net.inet.tcp.reass.maxsegments",
 	    &V_tcp_reass_maxseg);
diff --git a/sys/netinet/tcp_sack.c b/sys/netinet/tcp_sack.c
index 06fb50269fa..737c2b2af09 100644
--- a/sys/netinet/tcp_sack.c
+++ b/sys/netinet/tcp_sack.c
@@ -123,29 +123,28 @@ __FBSDID("$FreeBSD$");
 #include <machine/in_cksum.h>
 
 VNET_DECLARE(struct uma_zone *, sack_hole_zone);
-VNET_DEFINE(int, tcp_do_sack);
-VNET_DEFINE(int, tcp_sack_maxholes);
-VNET_DEFINE(int, tcp_sack_globalmaxholes);
-VNET_DEFINE(int, tcp_sack_globalholes);
-
 #define	V_sack_hole_zone		VNET(sack_hole_zone)
-#define	V_tcp_do_sack			VNET(tcp_do_sack)
-#define	V_tcp_sack_maxholes		VNET(tcp_sack_maxholes)
-#define	V_tcp_sack_globalmaxholes	VNET(tcp_sack_globalmaxholes)
-#define	V_tcp_sack_globalholes		VNET(tcp_sack_globalholes)
 
 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, sack, CTLFLAG_RW, 0, "TCP SACK");
+VNET_DEFINE(int, tcp_do_sack) = 1;
+#define	V_tcp_do_sack			VNET(tcp_do_sack)
 SYSCTL_VNET_INT(_net_inet_tcp_sack, OID_AUTO, enable, CTLFLAG_RW,
     &VNET_NAME(tcp_do_sack), 0, "Enable/Disable TCP SACK support");
 
+VNET_DEFINE(int, tcp_sack_maxholes) = 128;
+#define	V_tcp_sack_maxholes		VNET(tcp_sack_maxholes)
 SYSCTL_VNET_INT(_net_inet_tcp_sack, OID_AUTO, maxholes, CTLFLAG_RW,
     &VNET_NAME(tcp_sack_maxholes), 0,
     "Maximum number of TCP SACK holes allowed per connection");
 
+VNET_DEFINE(int, tcp_sack_globalmaxholes) = 65536;
+#define	V_tcp_sack_globalmaxholes	VNET(tcp_sack_globalmaxholes)
 SYSCTL_VNET_INT(_net_inet_tcp_sack, OID_AUTO, globalmaxholes, CTLFLAG_RW,
     &VNET_NAME(tcp_sack_globalmaxholes), 0, 
     "Global maximum number of TCP SACK holes");
 
+VNET_DEFINE(int, tcp_sack_globalholes) = 0;
+#define	V_tcp_sack_globalholes		VNET(tcp_sack_globalholes)
 SYSCTL_VNET_INT(_net_inet_tcp_sack, OID_AUTO, globalholes, CTLFLAG_RD,
     &VNET_NAME(tcp_sack_globalholes), 0,
     "Global number of TCP SACK holes currently allocated");
diff --git a/sys/netinet/tcp_subr.c b/sys/netinet/tcp_subr.c
index 9ec434c0c34..43ed1e4e750 100644
--- a/sys/netinet/tcp_subr.c
+++ b/sys/netinet/tcp_subr.c
@@ -111,28 +111,10 @@ __FBSDID("$FreeBSD$");
 
 #include <security/mac/mac_framework.h>
 
-VNET_DEFINE(int, tcp_mssdflt);
+VNET_DEFINE(int, tcp_mssdflt) = TCP_MSS;
 #ifdef INET6
-VNET_DEFINE(int, tcp_v6mssdflt);
+VNET_DEFINE(int, tcp_v6mssdflt) = TCP6_MSS;
 #endif
-VNET_DEFINE(int, tcp_minmss);
-VNET_DEFINE(int, tcp_do_rfc1323);
-
-static VNET_DEFINE(int, icmp_may_rst);
-static VNET_DEFINE(int, tcp_isn_reseed_interval);
-static VNET_DEFINE(int, tcp_inflight_enable);
-static VNET_DEFINE(int, tcp_inflight_rttthresh);
-static VNET_DEFINE(int, tcp_inflight_min);
-static VNET_DEFINE(int, tcp_inflight_max);
-static VNET_DEFINE(int, tcp_inflight_stab);
-
-#define	V_icmp_may_rst			VNET(icmp_may_rst)
-#define	V_tcp_isn_reseed_interval	VNET(tcp_isn_reseed_interval)
-#define	V_tcp_inflight_enable		VNET(tcp_inflight_enable)
-#define	V_tcp_inflight_rttthresh	VNET(tcp_inflight_rttthresh)
-#define	V_tcp_inflight_min		VNET(tcp_inflight_min)
-#define	V_tcp_inflight_max		VNET(tcp_inflight_max)
-#define	V_tcp_inflight_stab		VNET(tcp_inflight_stab)
 
 static int
 sysctl_net_inet_tcp_mss_check(SYSCTL_HANDLER_ARGS)
@@ -194,10 +176,12 @@ vnet_sysctl_msec_to_ticks(SYSCTL_HANDLER_ARGS)
  * with packet generation and sending. Set to zero to disable MINMSS
  * checking. This setting prevents us from sending too small packets.
  */
+VNET_DEFINE(int, tcp_minmss) = TCP_MINMSS;
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, minmss, CTLFLAG_RW,
      &VNET_NAME(tcp_minmss), 0,
     "Minmum TCP Maximum Segment Size");
 
+VNET_DEFINE(int, tcp_do_rfc1323) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp, TCPCTL_DO_RFC1323, rfc1323, CTLFLAG_RW,
     &VNET_NAME(tcp_do_rfc1323), 0,
     "Enable rfc1323 (high performance TCP) extensions");
@@ -217,10 +201,14 @@ SYSCTL_INT(_net_inet_tcp, OID_AUTO, do_tcpdrain, CTLFLAG_RW, &do_tcpdrain, 0,
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, pcbcount, CTLFLAG_RD,
     &VNET_NAME(tcbinfo.ipi_count), 0, "Number of active PCBs");
 
+static VNET_DEFINE(int, icmp_may_rst) = 1;
+#define	V_icmp_may_rst			VNET(icmp_may_rst)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, icmp_may_rst, CTLFLAG_RW,
     &VNET_NAME(icmp_may_rst), 0,
     "Certain ICMP unreachable messages may abort connections in SYN_SENT");
 
+static VNET_DEFINE(int, tcp_isn_reseed_interval) = 0;
+#define	V_tcp_isn_reseed_interval	VNET(tcp_isn_reseed_interval)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, isn_reseed_interval, CTLFLAG_RW,
     &VNET_NAME(tcp_isn_reseed_interval), 0,
     "Seconds between reseeding of ISN secret");
@@ -233,6 +221,8 @@ SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, isn_reseed_interval, CTLFLAG_RW,
 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, inflight, CTLFLAG_RW, 0,
     "TCP inflight data limiting");
 
+static VNET_DEFINE(int, tcp_inflight_enable) = 1;
+#define	V_tcp_inflight_enable		VNET(tcp_inflight_enable)
 SYSCTL_VNET_INT(_net_inet_tcp_inflight, OID_AUTO, enable, CTLFLAG_RW,
     &VNET_NAME(tcp_inflight_enable), 0,
     "Enable automatic TCP inflight data limiting");
@@ -242,19 +232,27 @@ SYSCTL_INT(_net_inet_tcp_inflight, OID_AUTO, debug, CTLFLAG_RW,
     &tcp_inflight_debug, 0,
     "Debug TCP inflight calculations");
 
+static VNET_DEFINE(int, tcp_inflight_rttthresh);
+#define	V_tcp_inflight_rttthresh	VNET(tcp_inflight_rttthresh)
 SYSCTL_VNET_PROC(_net_inet_tcp_inflight, OID_AUTO, rttthresh,
     CTLTYPE_INT|CTLFLAG_RW, &VNET_NAME(tcp_inflight_rttthresh), 0,
     vnet_sysctl_msec_to_ticks, "I",
     "RTT threshold below which inflight will deactivate itself");
 
+static VNET_DEFINE(int, tcp_inflight_min) = 6144;
+#define	V_tcp_inflight_min		VNET(tcp_inflight_min)
 SYSCTL_VNET_INT(_net_inet_tcp_inflight, OID_AUTO, min, CTLFLAG_RW,
     &VNET_NAME(tcp_inflight_min), 0,
     "Lower-bound for TCP inflight window");
 
+static VNET_DEFINE(int, tcp_inflight_max) = TCP_MAXWIN << TCP_MAX_WINSHIFT;
+#define	V_tcp_inflight_max		VNET(tcp_inflight_max)
 SYSCTL_VNET_INT(_net_inet_tcp_inflight, OID_AUTO, max, CTLFLAG_RW,
     &VNET_NAME(tcp_inflight_max), 0,
     "Upper-bound for TCP inflight window");
 
+static VNET_DEFINE(int, tcp_inflight_stab) = 20;
+#define	V_tcp_inflight_stab		VNET(tcp_inflight_stab)
 SYSCTL_VNET_INT(_net_inet_tcp_inflight, OID_AUTO, stab, CTLFLAG_RW,
     &VNET_NAME(tcp_inflight_stab), 0,
     "Inflight Algorithm Stabilization 20 = 2 packets");
@@ -329,53 +327,6 @@ tcp_init(void)
 {
 	int hashsize;
 
-	V_blackhole = 0;
-	V_tcp_delack_enabled = 1;
-	V_drop_synfin = 0;
-	V_tcp_do_rfc3042 = 1;
-	V_tcp_do_rfc3390 = 1;
-	V_tcp_do_ecn = 0;
-	V_tcp_ecn_maxretries = 1;
-	V_tcp_insecure_rst = 0;
-	V_tcp_do_autorcvbuf = 1;
-	V_tcp_autorcvbuf_inc = 16*1024;
-	V_tcp_autorcvbuf_max = 256*1024;
-	V_tcp_do_rfc3465 = 1;
-	V_tcp_abc_l_var = 2;
-
-	V_tcp_mssdflt = TCP_MSS;
-#ifdef INET6
-	V_tcp_v6mssdflt = TCP6_MSS;
-#endif
-	V_tcp_minmss = TCP_MINMSS;
-	V_tcp_do_rfc1323 = 1;
-	V_icmp_may_rst = 1;
-	V_tcp_isn_reseed_interval = 0;
-	V_tcp_inflight_enable = 1;
-	V_tcp_inflight_min = 6144;
-	V_tcp_inflight_max = TCP_MAXWIN << TCP_MAX_WINSHIFT;
-	V_tcp_inflight_stab = 20;
-
-	V_path_mtu_discovery = 1;
-	V_ss_fltsz = 1;
-	V_ss_fltsz_local = 4;
-	V_tcp_do_newreno = 1;
-	V_tcp_do_tso = 1;
-	V_tcp_do_autosndbuf = 1;
-	V_tcp_autosndbuf_inc = 8*1024;
-	V_tcp_autosndbuf_max = 256*1024;
-
-	V_nolocaltimewait = 0;
-
-	V_tcp_do_sack = 1;
-	V_tcp_sack_maxholes = 128;
-	V_tcp_sack_globalmaxholes = 65536;
-	V_tcp_sack_globalholes = 0;
-
-	V_tcp_inflight_rttthresh = TCPTV_INFLIGHT_RTTTHRESH;
-
-	TUNABLE_INT_FETCH("net.inet.tcp.sack.enable", &V_tcp_do_sack);
-
 	hashsize = TCBHASHSIZE;
 	TUNABLE_INT_FETCH("net.inet.tcp.tcbhashsize", &hashsize);
 	if (!powerof2(hashsize)) {
@@ -385,16 +336,21 @@ tcp_init(void)
 	in_pcbinfo_init(&V_tcbinfo, "tcp", &V_tcb, hashsize, hashsize,
 	    "tcp_inpcb", tcp_inpcb_init, NULL, UMA_ZONE_NOFREE);
 
+	V_tcp_inflight_rttthresh = TCPTV_INFLIGHT_RTTTHRESH;
+
 	/*
 	 * These have to be type stable for the benefit of the timers.
 	 */
 	V_tcpcb_zone = uma_zcreate("tcpcb", sizeof(struct tcpcb_mem),
 	    NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
 	uma_zone_set_max(V_tcpcb_zone, maxsockets);
+
 	tcp_tw_init();
 	syncache_init();
 	tcp_hc_init();
 	tcp_reass_init();
+
+	TUNABLE_INT_FETCH("net.inet.tcp.sack.enable", &V_tcp_do_sack);
 	V_sack_hole_zone = uma_zcreate("sackhole", sizeof(struct sackhole),
 	    NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, UMA_ZONE_NOFREE);
 
diff --git a/sys/netinet/tcp_syncache.c b/sys/netinet/tcp_syncache.c
index 4d9d487b503..c47493c2df3 100644
--- a/sys/netinet/tcp_syncache.c
+++ b/sys/netinet/tcp_syncache.c
@@ -97,19 +97,14 @@ __FBSDID("$FreeBSD$");
 
 #include <security/mac/mac_framework.h>
 
-static VNET_DEFINE(struct tcp_syncache, tcp_syncache);
-static VNET_DEFINE(int, tcp_syncookies);
-static VNET_DEFINE(int, tcp_syncookiesonly);
-VNET_DEFINE(int, tcp_sc_rst_sock_fail);
-
-#define	V_tcp_syncache			VNET(tcp_syncache)
+static VNET_DEFINE(int, tcp_syncookies) = 1;
 #define	V_tcp_syncookies		VNET(tcp_syncookies)
-#define	V_tcp_syncookiesonly		VNET(tcp_syncookiesonly)
-
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, syncookies, CTLFLAG_RW,
     &VNET_NAME(tcp_syncookies), 0,
     "Use TCP SYN cookies if the syncache overflows");
 
+static VNET_DEFINE(int, tcp_syncookiesonly) = 0;
+#define	V_tcp_syncookiesonly		VNET(tcp_syncookiesonly)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, syncookies_only, CTLFLAG_RW,
     &VNET_NAME(tcp_syncookiesonly), 0,
     "Use only TCP SYN cookies");
@@ -148,6 +143,9 @@ static struct syncache
 #define TCP_SYNCACHE_HASHSIZE		512
 #define TCP_SYNCACHE_BUCKETLIMIT	30
 
+static VNET_DEFINE(struct tcp_syncache, tcp_syncache);
+#define	V_tcp_syncache			VNET(tcp_syncache)
+
 SYSCTL_NODE(_net_inet_tcp, OID_AUTO, syncache, CTLFLAG_RW, 0, "TCP SYN cache");
 
 SYSCTL_VNET_INT(_net_inet_tcp_syncache, OID_AUTO, bucketlimit, CTLFLAG_RDTUN,
@@ -170,6 +168,7 @@ SYSCTL_VNET_INT(_net_inet_tcp_syncache, OID_AUTO, rexmtlimit, CTLFLAG_RW,
     &VNET_NAME(tcp_syncache.rexmt_limit), 0,
     "Limit on SYN/ACK retransmissions");
 
+VNET_DEFINE(int, tcp_sc_rst_sock_fail) = 1;
 SYSCTL_VNET_INT(_net_inet_tcp_syncache, OID_AUTO, rst_on_sock_fail,
     CTLFLAG_RW, &VNET_NAME(tcp_sc_rst_sock_fail), 0,
     "Send reset on socket allocation failure");
@@ -224,10 +223,6 @@ syncache_init(void)
 {
 	int i;
 
-	V_tcp_syncookies = 1;
-	V_tcp_syncookiesonly = 0;
-	V_tcp_sc_rst_sock_fail = 1;
-
 	V_tcp_syncache.cache_count = 0;
 	V_tcp_syncache.hashsize = TCP_SYNCACHE_HASHSIZE;
 	V_tcp_syncache.bucket_limit = TCP_SYNCACHE_BUCKETLIMIT;
diff --git a/sys/netinet/tcp_timewait.c b/sys/netinet/tcp_timewait.c
index 6e5b013ae98..42df4fe9463 100644
--- a/sys/netinet/tcp_timewait.c
+++ b/sys/netinet/tcp_timewait.c
@@ -92,6 +92,8 @@ __FBSDID("$FreeBSD$");
 
 #include <security/mac/mac_framework.h>
 
+static VNET_DEFINE(uma_zone_t, tcptw_zone);
+#define	V_tcptw_zone			VNET(tcptw_zone)
 static int	maxtcptw;
 
 /*
@@ -100,11 +102,7 @@ static int	maxtcptw;
  * queue pointers in each tcptw structure, are protected using the global
  * tcbinfo lock, which must be held over queue iteration and modification.
  */
-static VNET_DEFINE(uma_zone_t, tcptw_zone);
 static VNET_DEFINE(TAILQ_HEAD(, tcptw), twq_2msl);
-VNET_DEFINE(int, nolocaltimewait);
-
-#define	V_tcptw_zone			VNET(tcptw_zone)
 #define	V_twq_2msl			VNET(twq_2msl)
 
 static void	tcp_tw_2msl_reset(struct tcptw *, int);
@@ -149,6 +147,8 @@ SYSCTL_PROC(_net_inet_tcp, OID_AUTO, maxtcptw, CTLTYPE_INT|CTLFLAG_RW,
     &maxtcptw, 0, sysctl_maxtcptw, "IU",
     "Maximum number of compressed TCP TIME_WAIT entries");
 
+VNET_DEFINE(int, nolocaltimewait) = 0;
+#define	V_nolocaltimewait	VNET(nolocaltimewait)
 SYSCTL_VNET_INT(_net_inet_tcp, OID_AUTO, nolocaltimewait, CTLFLAG_RW,
     &VNET_NAME(nolocaltimewait), 0,
     "Do not create compressed TCP TIME_WAIT entries for local connections");
diff --git a/sys/netinet/tcp_var.h b/sys/netinet/tcp_var.h
index 3a59eee6c2e..5811439dcd6 100644
--- a/sys/netinet/tcp_var.h
+++ b/sys/netinet/tcp_var.h
@@ -42,12 +42,12 @@
  * Kernel variables for tcp.
  */
 VNET_DECLARE(int, tcp_do_rfc1323);
+#define	V_tcp_do_rfc1323	VNET(tcp_do_rfc1323)
+
 VNET_DECLARE(int, tcp_reass_qsize);
 VNET_DECLARE(struct uma_zone *, tcp_reass_zone);
-#define	V_tcp_do_rfc1323	VNET(tcp_do_rfc1323)
 #define	V_tcp_reass_qsize	VNET(tcp_reass_qsize)
 #define	V_tcp_reass_zone	VNET(tcp_reass_zone)
-
 #endif /* _KERNEL */
 
 /* TCP segment queue entry */
@@ -558,11 +558,10 @@ SYSCTL_DECL(_net_inet_tcp_sack);
 MALLOC_DECLARE(M_TCPLOG);
 #endif
 
-extern	int tcp_log_in_vain;
-
 VNET_DECLARE(struct inpcbhead, tcb);		/* queue of active tcpcb's */
 VNET_DECLARE(struct inpcbinfo, tcbinfo);
 VNET_DECLARE(struct tcpstat, tcpstat);		/* tcp statistics */
+extern	int tcp_log_in_vain;
 VNET_DECLARE(int, tcp_mssdflt);	/* XXX */
 VNET_DECLARE(int, tcp_minmss);
 VNET_DECLARE(int, tcp_delack_enabled);
@@ -570,7 +569,6 @@ VNET_DECLARE(int, tcp_do_newreno);
 VNET_DECLARE(int, path_mtu_discovery);
 VNET_DECLARE(int, ss_fltsz);
 VNET_DECLARE(int, ss_fltsz_local);
-
 #define	V_tcb			VNET(tcb)
 #define	V_tcbinfo		VNET(tcbinfo)
 #define	V_tcpstat		VNET(tcpstat)
@@ -582,55 +580,13 @@ VNET_DECLARE(int, ss_fltsz_local);
 #define	V_ss_fltsz		VNET(ss_fltsz)
 #define	V_ss_fltsz_local	VNET(ss_fltsz_local)
 
-VNET_DECLARE(int, blackhole);
-VNET_DECLARE(int, drop_synfin);
-VNET_DECLARE(int, tcp_do_rfc3042);
-VNET_DECLARE(int, tcp_do_rfc3390);
-VNET_DECLARE(int, tcp_insecure_rst);
-VNET_DECLARE(int, tcp_do_autorcvbuf);
-VNET_DECLARE(int, tcp_autorcvbuf_inc);
-VNET_DECLARE(int, tcp_autorcvbuf_max);
-VNET_DECLARE(int, tcp_do_rfc3465);
-VNET_DECLARE(int, tcp_abc_l_var);
-
-#define	V_blackhole		VNET(blackhole)
-#define	V_drop_synfin		VNET(drop_synfin)
-#define	V_tcp_do_rfc3042	VNET(tcp_do_rfc3042)
-#define	V_tcp_do_rfc3390	VNET(tcp_do_rfc3390)
-#define	V_tcp_insecure_rst	VNET(tcp_insecure_rst)
-#define	V_tcp_do_autorcvbuf	VNET(tcp_do_autorcvbuf)
-#define	V_tcp_autorcvbuf_inc	VNET(tcp_autorcvbuf_inc)
-#define	V_tcp_autorcvbuf_max	VNET(tcp_autorcvbuf_max)
-#define	V_tcp_do_rfc3465	VNET(tcp_do_rfc3465)
-#define	V_tcp_abc_l_var		VNET(tcp_abc_l_var)
-
-VNET_DECLARE(int, tcp_do_tso);
-VNET_DECLARE(int, tcp_do_autosndbuf);
-VNET_DECLARE(int, tcp_autosndbuf_inc);
-VNET_DECLARE(int, tcp_autosndbuf_max);
-
-#define	V_tcp_do_tso		VNET(tcp_do_tso)
-#define	V_tcp_do_autosndbuf	VNET(tcp_do_autosndbuf)
-#define	V_tcp_autosndbuf_inc	VNET(tcp_autosndbuf_inc)
-#define	V_tcp_autosndbuf_max	VNET(tcp_autosndbuf_max)
-
-VNET_DECLARE(int, nolocaltimewait);
-
-#define	V_nolocaltimewait	VNET(nolocaltimewait)
-
 VNET_DECLARE(int, tcp_do_sack);			/* SACK enabled/disabled */
-VNET_DECLARE(int, tcp_sack_maxholes);
-VNET_DECLARE(int, tcp_sack_globalmaxholes);
-VNET_DECLARE(int, tcp_sack_globalholes);
 VNET_DECLARE(int, tcp_sc_rst_sock_fail);	/* RST on sock alloc failure */
+#define	V_tcp_do_sack		VNET(tcp_do_sack)
+#define	V_tcp_sc_rst_sock_fail	VNET(tcp_sc_rst_sock_fail)
+
 VNET_DECLARE(int, tcp_do_ecn);			/* TCP ECN enabled/disabled */
 VNET_DECLARE(int, tcp_ecn_maxretries);
-
-#define	V_tcp_do_sack		VNET(tcp_do_sack)
-#define	V_tcp_sack_maxholes	VNET(tcp_sack_maxholes)
-#define	V_tcp_sack_globalmaxholes	VNET(tcp_sack_globalmaxholes)
-#define	V_tcp_sack_globalholes	VNET(tcp_sack_globalholes)
-#define	V_tcp_sc_rst_sock_fail	VNET(tcp_sc_rst_sock_fail)
 #define	V_tcp_do_ecn		VNET(tcp_do_ecn)
 #define	V_tcp_ecn_maxretries	VNET(tcp_ecn_maxretries)
 
diff --git a/sys/netinet/udp_usrreq.c b/sys/netinet/udp_usrreq.c
index 0d8e04deaf6..45410387c17 100644
--- a/sys/netinet/udp_usrreq.c
+++ b/sys/netinet/udp_usrreq.c
@@ -94,8 +94,6 @@ __FBSDID("$FreeBSD$");
  * Per RFC 768, August, 1980.
  */
 
-VNET_DEFINE(int, udp_blackhole);
-
 /*
  * BSD 4.2 defaulted the udp checksum to be off.  Turning off udp checksums
  * removes the only data integrity mechanism for packets and malformed
@@ -110,6 +108,7 @@ int	udp_log_in_vain = 0;
 SYSCTL_INT(_net_inet_udp, OID_AUTO, log_in_vain, CTLFLAG_RW,
     &udp_log_in_vain, 0, "Log all incoming UDP packets");
 
+VNET_DEFINE(int, udp_blackhole) = 0;
 SYSCTL_VNET_INT(_net_inet_udp, OID_AUTO, blackhole, CTLFLAG_RW,
     &VNET_NAME(udp_blackhole), 0,
     "Do not send port unreachables for refused connects");
@@ -133,14 +132,13 @@ SYSCTL_ULONG(_net_inet_udp, UDPCTL_RECVSPACE, recvspace, CTLFLAG_RW,
 VNET_DEFINE(struct inpcbhead, udb);		/* from udp_var.h */
 VNET_DEFINE(struct inpcbinfo, udbinfo);
 static VNET_DEFINE(uma_zone_t, udpcb_zone);
-VNET_DEFINE(struct udpstat, udpstat);		/* from udp_var.h */
-
 #define	V_udpcb_zone			VNET(udpcb_zone)
 
 #ifndef UDBHASHSIZE
 #define	UDBHASHSIZE	128
 #endif
 
+VNET_DEFINE(struct udpstat, udpstat);		/* from udp_var.h */
 SYSCTL_VNET_STRUCT(_net_inet_udp, UDPCTL_STATS, stats, CTLFLAG_RW,
     &VNET_NAME(udpstat), udpstat,
     "UDP statistics (struct udpstat, netinet/udp_var.h)");
@@ -179,7 +177,6 @@ void
 udp_init(void)
 {
 
-	V_udp_blackhole = 0;
 	in_pcbinfo_init(&V_udbinfo, "udp", &V_udb, UDBHASHSIZE, UDBHASHSIZE,
 	    "udp_inpcb", udp_inpcb_init, NULL, UMA_ZONE_NOFREE);
 	V_udpcb_zone = uma_zcreate("udpcb", sizeof(struct udpcb),
diff --git a/sys/netinet/udp_var.h b/sys/netinet/udp_var.h
index b8d994cc3a6..aa7ee470c8b 100644
--- a/sys/netinet/udp_var.h
+++ b/sys/netinet/udp_var.h
@@ -129,19 +129,17 @@ void	kmod_udpstat_inc(int statnum);
 SYSCTL_DECL(_net_inet_udp);
 
 extern struct pr_usrreqs	udp_usrreqs;
-
 VNET_DECLARE(struct inpcbhead, udb);
 VNET_DECLARE(struct inpcbinfo, udbinfo);
-VNET_DECLARE(struct udpstat, udpstat);
-VNET_DECLARE(int, udp_blackhole);
-
 #define	V_udb			VNET(udb)
 #define	V_udbinfo		VNET(udbinfo)
-#define	V_udpstat		VNET(udpstat)
-#define	V_udp_blackhole		VNET(udp_blackhole)
 
 extern u_long			udp_sendspace;
 extern u_long			udp_recvspace;
+VNET_DECLARE(struct udpstat, udpstat);
+VNET_DECLARE(int, udp_blackhole);
+#define	V_udpstat		VNET(udpstat)
+#define	V_udp_blackhole		VNET(udp_blackhole)
 extern int			udp_log_in_vain;
 
 int		 udp_newudpcb(struct inpcb *);
diff --git a/sys/netinet6/frag6.c b/sys/netinet6/frag6.c
index 8900f7d08b6..1523133d3db 100644
--- a/sys/netinet6/frag6.c
+++ b/sys/netinet6/frag6.c
@@ -106,16 +106,17 @@ void
 frag6_init(void)
 {
 
-	V_ip6q.ip6q_next = V_ip6q.ip6q_prev = &V_ip6q;
 	V_ip6_maxfragpackets = nmbclusters / 4;
 	V_ip6_maxfrags = nmbclusters / 4;
+	V_ip6q.ip6q_next = V_ip6q.ip6q_prev = &V_ip6q;
 
 	if (!IS_DEFAULT_VNET(curvnet))
 		return;
 
-	IP6Q_LOCK_INIT();
 	EVENTHANDLER_REGISTER(nmbclusters_change,
 	    frag6_change, NULL, EVENTHANDLER_PRI_ANY);
+
+	IP6Q_LOCK_INIT();
 }
 
 /*
diff --git a/sys/netinet6/icmp6.c b/sys/netinet6/icmp6.c
index 57f8d32c59a..87df9c369c8 100644
--- a/sys/netinet6/icmp6.c
+++ b/sys/netinet6/icmp6.c
@@ -113,23 +113,22 @@ __FBSDID("$FreeBSD$");
 
 extern struct domain inet6domain;
 
+VNET_DEFINE(struct icmp6stat, icmp6stat);
+
 VNET_DECLARE(struct inpcbinfo, ripcbinfo);
 VNET_DECLARE(struct inpcbhead, ripcb);
 VNET_DECLARE(int, icmp6errppslim);
+static VNET_DEFINE(int, icmp6errpps_count) = 0;
+static VNET_DEFINE(struct timeval, icmp6errppslim_last);
 VNET_DECLARE(int, icmp6_nodeinfo);
 
 #define	V_ripcbinfo			VNET(ripcbinfo)
 #define	V_ripcb				VNET(ripcb)
 #define	V_icmp6errppslim		VNET(icmp6errppslim)
+#define	V_icmp6errpps_count		VNET(icmp6errpps_count)
+#define	V_icmp6errppslim_last		VNET(icmp6errppslim_last)
 #define	V_icmp6_nodeinfo		VNET(icmp6_nodeinfo)
 
-VNET_DEFINE(struct icmp6stat, icmp6stat);
-static VNET_DEFINE(int, icmp6errpps_count);
-static VNET_DEFINE(struct timeval, icmp6errppslim_last);
-
-#define	V_icmp6errpps_count	VNET(icmp6errpps_count)
-#define	V_icmp6errppslim_last	VNET(icmp6errppslim_last)
-
 static void icmp6_errcount(struct icmp6errstat *, int, int);
 static int icmp6_rip6_input(struct mbuf **, int);
 static int icmp6_ratelimit(const struct in6_addr *, const int, const int);
@@ -144,14 +143,6 @@ static int ni6_store_addrs __P((struct icmp6_nodeinfo *, struct icmp6_nodeinfo *
 				struct ifnet *, int));
 static int icmp6_notify_error(struct mbuf **, int, int, int);
 
-
-void
-icmp6_init(void)
-{
-
-	V_icmp6errpps_count = 0;
-}
-
 /*
  * Kernel module interface for updating icmp6stat.  The argument is an index
  * into icmp6stat treated as an array of u_quad_t.  While this encodes the
diff --git a/sys/netinet6/in6_gif.c b/sys/netinet6/in6_gif.c
index a481706f53d..e786836459f 100644
--- a/sys/netinet6/in6_gif.c
+++ b/sys/netinet6/in6_gif.c
@@ -41,8 +41,10 @@ __FBSDID("$FreeBSD$");
 #include <sys/sockio.h>
 #include <sys/mbuf.h>
 #include <sys/errno.h>
+#include <sys/kernel.h>
 #include <sys/queue.h>
 #include <sys/syslog.h>
+#include <sys/sysctl.h>
 #include <sys/protosw.h>
 #include <sys/malloc.h>
 
@@ -69,6 +71,13 @@ __FBSDID("$FreeBSD$");
 
 #include <net/if_gif.h>
 
+VNET_DEFINE(int, ip6_gif_hlim) = GIF_HLIM;
+#define	V_ip6_gif_hlim			VNET(ip6_gif_hlim)
+
+SYSCTL_DECL(_net_inet6_ip6);
+SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_GIF_HLIM, gifhlim, CTLFLAG_RW,
+    &VNET_NAME(ip6_gif_hlim), 0, "");
+
 static int gif_validate6(const struct ip6_hdr *, struct gif_softc *,
 			 struct ifnet *);
 
diff --git a/sys/netinet6/in6_ifattach.c b/sys/netinet6/in6_ifattach.c
index 5041ee2f233..363d7bedeb6 100644
--- a/sys/netinet6/in6_ifattach.c
+++ b/sys/netinet6/in6_ifattach.c
@@ -67,10 +67,15 @@ __FBSDID("$FreeBSD$");
 #include <netinet6/mld6_var.h>
 #include <netinet6/scope6_var.h>
 
-VNET_DEFINE(unsigned long, in6_maxmtu);
-VNET_DEFINE(int, ip6_auto_linklocal);
-VNET_DEFINE(struct callout, in6_tmpaddrtimer_ch);
+VNET_DEFINE(unsigned long, in6_maxmtu) = 0;
 
+#ifdef IP6_AUTO_LINKLOCAL
+VNET_DEFINE(int, ip6_auto_linklocal) = IP6_AUTO_LINKLOCAL;
+#else
+VNET_DEFINE(int, ip6_auto_linklocal) = 1;	/* enabled by default */
+#endif
+
+VNET_DEFINE(struct callout, in6_tmpaddrtimer_ch);
 #define	V_in6_tmpaddrtimer_ch		VNET(in6_tmpaddrtimer_ch)
 
 VNET_DECLARE(struct inpcbinfo, ripcbinfo);
diff --git a/sys/netinet6/in6_proto.c b/sys/netinet6/in6_proto.c
index 3289e57a274..51f5187598e 100644
--- a/sys/netinet6/in6_proto.c
+++ b/sys/netinet6/in6_proto.c
@@ -234,7 +234,6 @@ struct ip6protosw inet6sw[] = {
 	.pr_output =		rip6_output,
 	.pr_ctlinput =		rip6_ctlinput,
 	.pr_ctloutput =		rip6_ctloutput,
-	.pr_init =		icmp6_init,
 	.pr_fasttimo =		icmp6_fasttimo,
 	.pr_slowtimo =		icmp6_slowtimo,
 	.pr_usrreqs =		&rip6_usrreqs
@@ -378,25 +377,44 @@ VNET_DOMAIN_SET(inet6);
 /*
  * Internet configuration info
  */
-VNET_DEFINE(int, ip6_forwarding);
-VNET_DEFINE(int, ip6_sendredirects);
-VNET_DEFINE(int, ip6_defhlim);
-VNET_DEFINE(int, ip6_defmcasthlim);
-VNET_DEFINE(int, ip6_accept_rtadv);
-VNET_DEFINE(int, ip6_maxfragpackets);
-VNET_DEFINE(int, ip6_maxfrags);
-VNET_DEFINE(int, ip6_log_interval);
-VNET_DEFINE(int, ip6_hdrnestlimit);
-VNET_DEFINE(int, ip6_dad_count);
-VNET_DEFINE(int, ip6_auto_flowlabel);
-VNET_DEFINE(int, ip6_use_deprecated);
-VNET_DEFINE(int, ip6_rr_prune);
-VNET_DEFINE(int, ip6_mcast_pmtu);
-VNET_DEFINE(int, ip6_v6only);
-VNET_DEFINE(int, ip6_keepfaith);
-VNET_DEFINE(time_t, ip6_log_time);
-VNET_DEFINE(int, ip6stealth);
-VNET_DEFINE(int, nd6_onlink_ns_rfc4861);
+#ifndef	IPV6FORWARDING
+#ifdef GATEWAY6
+#define	IPV6FORWARDING	1	/* forward IP6 packets not for us */
+#else
+#define	IPV6FORWARDING	0	/* don't forward IP6 packets not for us */
+#endif /* GATEWAY6 */
+#endif /* !IPV6FORWARDING */
+
+#ifndef	IPV6_SENDREDIRECTS
+#define	IPV6_SENDREDIRECTS	1
+#endif
+
+VNET_DEFINE(int, ip6_forwarding) = IPV6FORWARDING;	/* act as router? */
+VNET_DEFINE(int, ip6_sendredirects) = IPV6_SENDREDIRECTS;
+VNET_DEFINE(int, ip6_defhlim) = IPV6_DEFHLIM;
+VNET_DEFINE(int, ip6_defmcasthlim) = IPV6_DEFAULT_MULTICAST_HOPS;
+VNET_DEFINE(int, ip6_accept_rtadv) = 0;
+VNET_DEFINE(int, ip6_maxfragpackets);	/* initialized in frag6.c:frag6_init() */
+VNET_DEFINE(int, ip6_maxfrags);		/* initialized in frag6.c:frag6_init() */
+VNET_DEFINE(int, ip6_log_interval) = 5;
+VNET_DEFINE(int, ip6_hdrnestlimit) = 15;/* How many header options will we
+					 * process? */
+VNET_DEFINE(int, ip6_dad_count) = 1;	/* DupAddrDetectionTransmits */
+VNET_DEFINE(int, ip6_auto_flowlabel) = 1;
+VNET_DEFINE(int, ip6_use_deprecated) = 1;/* allow deprecated addr
+					 * (RFC2462 5.5.4) */
+VNET_DEFINE(int, ip6_rr_prune) = 5;	/* router renumbering prefix
+					 * walk list every 5 sec. */
+VNET_DEFINE(int, ip6_mcast_pmtu) = 0;	/* enable pMTU discovery for multicast? */
+VNET_DEFINE(int, ip6_v6only) = 1;
+
+VNET_DEFINE(int, ip6_keepfaith) = 0;
+VNET_DEFINE(time_t, ip6_log_time) = (time_t)0L;
+#ifdef IPSTEALTH
+VNET_DEFINE(int, ip6stealth) = 0;
+#endif
+VNET_DEFINE(int, nd6_onlink_ns_rfc4861) = 0;/* allow 'on-link' nd6 NS
+					     * (RFC 4861) */
 
 /* icmp6 */
 /*
@@ -404,26 +422,31 @@ VNET_DEFINE(int, nd6_onlink_ns_rfc4861);
  * XXX: what if we don't define INET? Should we define pmtu6_expire
  * or so? (jinmei@kame.net 19990310)
  */
-VNET_DEFINE(int, pmtu_expire);
-VNET_DEFINE(int, pmtu_probe);
+VNET_DEFINE(int, pmtu_expire) = 60*10;
+VNET_DEFINE(int, pmtu_probe) = 60*2;
 
 /* raw IP6 parameters */
 /*
  * Nominal space allocated to a raw ip socket.
  */
-VNET_DEFINE(u_long, rip6_sendspace);
-VNET_DEFINE(u_long, rip6_recvspace);
+#define	RIPV6SNDQ	8192
+#define	RIPV6RCVQ	8192
+
+VNET_DEFINE(u_long, rip6_sendspace) = RIPV6SNDQ;
+VNET_DEFINE(u_long, rip6_recvspace) = RIPV6RCVQ;
 
 /* ICMPV6 parameters */
-VNET_DEFINE(int, icmp6_rediraccept);
-VNET_DEFINE(int, icmp6_redirtimeout);
-VNET_DEFINE(int, icmp6errppslim);
+VNET_DEFINE(int, icmp6_rediraccept) = 1;/* accept and process redirects */
+VNET_DEFINE(int, icmp6_redirtimeout) = 10 * 60;	/* 10 minutes */
+VNET_DEFINE(int, icmp6errppslim) = 100;		/* 100pps */
 /* control how to respond to NI queries */
-VNET_DEFINE(int, icmp6_nodeinfo);
+VNET_DEFINE(int, icmp6_nodeinfo) =
+    (ICMP6_NODEINFO_FQDNOK|ICMP6_NODEINFO_NODEADDROK);
 
 /* UDP on IP6 parameters */
-VNET_DEFINE(int, udp6_sendspace);
-VNET_DEFINE(int, udp6_recvspace);
+VNET_DEFINE(int, udp6_sendspace) = 9216;/* really max datagram size */
+VNET_DEFINE(int, udp6_recvspace) = 40 * (1024 + sizeof(struct sockaddr_in6));
+					/* 40 1K datagrams */
 
 /*
  * sysctl related items.
@@ -571,7 +594,6 @@ SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_MAXNUDHINT, nd6_maxnudhint,
 	CTLFLAG_RW, &VNET_NAME(nd6_maxnudhint), 0, "");
 SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_DEBUG, nd6_debug, CTLFLAG_RW,
 	&VNET_NAME(nd6_debug), 0, "");
-
 SYSCTL_VNET_INT(_net_inet6_icmp6, ICMPV6CTL_ND6_ONLINKNSRFC4861,
 	nd6_onlink_ns_rfc4861, CTLFLAG_RW, &VNET_NAME(nd6_onlink_ns_rfc4861),
 	0, "Accept 'on-link' nd6 NS in compliance with RFC 4861.");
diff --git a/sys/netinet6/in6_rmx.c b/sys/netinet6/in6_rmx.c
index 1ae04c3f112..8927a176df9 100644
--- a/sys/netinet6/in6_rmx.c
+++ b/sys/netinet6/in6_rmx.c
@@ -204,20 +204,21 @@ in6_matroute(void *v_arg, struct radix_node_head *head)
 
 SYSCTL_DECL(_net_inet6_ip6);
 
-static VNET_DEFINE(int, rtq_reallyold6);
-static VNET_DEFINE(int, rtq_minreallyold6);
-static VNET_DEFINE(int, rtq_toomany6);
-
+static VNET_DEFINE(int, rtq_reallyold6) = 60*60;
+	/* one hour is ``really old'' */
 #define	V_rtq_reallyold6		VNET(rtq_reallyold6)
-#define	V_rtq_minreallyold6		VNET(rtq_minreallyold6)
-#define	V_rtq_toomany6			VNET(rtq_toomany6)
-
 SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RTEXPIRE, rtexpire, CTLFLAG_RW,
     &VNET_NAME(rtq_reallyold6) , 0, "");
 
+static VNET_DEFINE(int, rtq_minreallyold6) = 10;
+	/* never automatically crank down to less */
+#define	V_rtq_minreallyold6		VNET(rtq_minreallyold6)
 SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RTMINEXPIRE, rtminexpire, CTLFLAG_RW,
     &VNET_NAME(rtq_minreallyold6) , 0, "");
 
+static VNET_DEFINE(int, rtq_toomany6) = 128;
+	/* 128 cached routes is ``too many'' */
+#define	V_rtq_toomany6			VNET(rtq_toomany6)
 SYSCTL_VNET_INT(_net_inet6_ip6, IPV6CTL_RTMAXCACHE, rtmaxcache, CTLFLAG_RW,
     &VNET_NAME(rtq_toomany6) , 0, "");
 
@@ -277,7 +278,7 @@ in6_rtqkill(struct radix_node *rn, void *rock)
 }
 
 #define RTQ_TIMEOUT	60*10	/* run no less than once every ten minutes */
-static VNET_DEFINE(int, rtq_timeout6);
+static VNET_DEFINE(int, rtq_timeout6) = RTQ_TIMEOUT;
 static VNET_DEFINE(struct callout, rtq_timer6);
 
 #define	V_rtq_timeout6			VNET(rtq_timeout6)
@@ -346,7 +347,6 @@ struct mtuex_arg {
 	struct radix_node_head *rnh;
 	time_t nextstop;
 };
-
 static VNET_DEFINE(struct callout, rtq_mtutimer);
 #define	V_rtq_mtutimer			VNET(rtq_mtutimer)
 
@@ -422,11 +422,6 @@ in6_inithead(void **head, int off)
 	if (off == 0)		/* See above */
 		return 1;	/* only do the rest for the real thing */
 
-	V_rtq_reallyold6 = 60*60; /* one hour is ``really old'' */
-	V_rtq_minreallyold6 = 10; /* never automatically crank down to less */
-	V_rtq_toomany6 = 128;	  /* 128 cached routes is ``too many'' */
-	V_rtq_timeout6 = RTQ_TIMEOUT;
-
 	rnh = *head;
 	KASSERT(rnh == rt_tables_get_rnh(0, AF_INET6), ("rnh?"));
 	rnh->rnh_addaddr = in6_addroute;
diff --git a/sys/netinet6/in6_src.c b/sys/netinet6/in6_src.c
index e6c2cd83726..49bc715452a 100644
--- a/sys/netinet6/in6_src.c
+++ b/sys/netinet6/in6_src.c
@@ -122,12 +122,11 @@ static struct sx addrsel_sxlock;
 #define	ADDRSEL_XUNLOCK()	sx_xunlock(&addrsel_sxlock)
 
 #define ADDR_LABEL_NOTAPP (-1)
-
 static VNET_DEFINE(struct in6_addrpolicy, defaultaddrpolicy);
-VNET_DEFINE(int, ip6_prefer_tempaddr);
-
 #define	V_defaultaddrpolicy		VNET(defaultaddrpolicy)
 
+VNET_DEFINE(int, ip6_prefer_tempaddr) = 0;
+
 static int selectroute __P((struct sockaddr_in6 *, struct ip6_pktopts *,
 	struct ip6_moptions *, struct route_in6 *, struct ifnet **,
 	struct rtentry **, int));
@@ -952,8 +951,6 @@ void
 addrsel_policy_init(void)
 {
 
-	V_ip6_prefer_tempaddr = 0;
-
 	init_policy_queue();
 
 	/* initialize the "last resort" policy */
diff --git a/sys/netinet6/in6_var.h b/sys/netinet6/in6_var.h
index d0a54e06a36..00342fde762 100644
--- a/sys/netinet6/in6_var.h
+++ b/sys/netinet6/in6_var.h
@@ -487,12 +487,7 @@ struct	in6_rrenumreq {
 
 #ifdef _KERNEL
 VNET_DECLARE(struct in6_ifaddrhead, in6_ifaddrhead);
-VNET_DECLARE(struct icmp6stat, icmp6stat);
-VNET_DECLARE(unsigned long, in6_maxmtu);
-
 #define	V_in6_ifaddrhead		VNET(in6_ifaddrhead)
-#define	V_icmp6stat			VNET(icmp6stat)
-#define	V_in6_maxmtu			VNET(in6_maxmtu)
 
 extern struct rwlock in6_ifaddr_lock;
 #define	IN6_IFADDR_LOCK_ASSERT(	)	rw_assert(&in6_ifaddr_lock, RA_LOCKED)
@@ -503,6 +498,8 @@ extern struct rwlock in6_ifaddr_lock;
 #define	IN6_IFADDR_WLOCK_ASSERT()	rw_assert(&in6_ifaddr_lock, RA_WLOCKED)
 #define	IN6_IFADDR_WUNLOCK()		rw_wunlock(&in6_ifaddr_lock)
 
+VNET_DECLARE(struct icmp6stat, icmp6stat);
+#define	V_icmp6stat			VNET(icmp6stat)
 #define in6_ifstat_inc(ifp, tag) \
 do {								\
 	if (ifp)						\
@@ -511,6 +508,8 @@ do {								\
 
 extern struct in6_addr zeroin6_addr;
 extern u_char inet6ctlerrmap[];
+VNET_DECLARE(unsigned long, in6_maxmtu);
+#define	V_in6_maxmtu			VNET(in6_maxmtu)
 #endif /* _KERNEL */
 
 /*
diff --git a/sys/netinet6/ip6_input.c b/sys/netinet6/ip6_input.c
index c244f374b92..42f935143aa 100644
--- a/sys/netinet6/ip6_input.c
+++ b/sys/netinet6/ip6_input.c
@@ -116,6 +116,7 @@ __FBSDID("$FreeBSD$");
 extern struct domain inet6domain;
 
 u_char ip6_protox[IPPROTO_MAX];
+VNET_DEFINE(struct in6_ifaddrhead, in6_ifaddrhead);
 
 static struct netisr_handler ip6_nh = {
 	.nh_name = "ip6",
@@ -124,36 +125,16 @@ static struct netisr_handler ip6_nh = {
 	.nh_policy = NETISR_POLICY_FLOW,
 };
 
-VNET_DEFINE(struct in6_ifaddrhead, in6_ifaddrhead);
-VNET_DEFINE(struct ip6stat, ip6stat);
-
 VNET_DECLARE(struct callout, in6_tmpaddrtimer_ch);
-VNET_DECLARE(int, dad_init);
-VNET_DECLARE(int, pmtu_expire);
-VNET_DECLARE(int, pmtu_probe);
-VNET_DECLARE(u_long, rip6_sendspace);
-VNET_DECLARE(u_long, rip6_recvspace);
-VNET_DECLARE(int, icmp6errppslim);
-VNET_DECLARE(int, icmp6_nodeinfo);
-VNET_DECLARE(int, udp6_sendspace);
-VNET_DECLARE(int, udp6_recvspace);
-
 #define	V_in6_tmpaddrtimer_ch		VNET(in6_tmpaddrtimer_ch)
-#define	V_dad_init			VNET(dad_init)
-#define	V_pmtu_expire			VNET(pmtu_expire)
-#define	V_pmtu_probe			VNET(pmtu_probe)
-#define	V_rip6_sendspace		VNET(rip6_sendspace)
-#define	V_rip6_recvspace		VNET(rip6_recvspace)
-#define	V_icmp6errppslim		VNET(icmp6errppslim)
-#define	V_icmp6_nodeinfo		VNET(icmp6_nodeinfo)
-#define	V_udp6_sendspace		VNET(udp6_sendspace)
-#define	V_udp6_recvspace		VNET(udp6_recvspace)
+
+VNET_DEFINE(struct pfil_head, inet6_pfil_hook);
+
+VNET_DEFINE(struct ip6stat, ip6stat);
 
 struct rwlock in6_ifaddr_lock;
 RW_SYSINIT(in6_ifaddr_lock, &in6_ifaddr_lock, "in6_ifaddr_lock");
 
-VNET_DEFINE (struct pfil_head, inet6_pfil_hook);
-
 static void ip6_init2(void *);
 static struct ip6aux *ip6_setdstifaddr(struct mbuf *, struct in6_ifaddr *);
 static int ip6_hopopts_input(u_int32_t *, u_int32_t *, struct mbuf **, int *);
@@ -171,82 +152,11 @@ ip6_init(void)
 	struct ip6protosw *pr;
 	int i;
 
-	V_in6_maxmtu = 0;
-#ifdef IP6_AUTO_LINKLOCAL
-	V_ip6_auto_linklocal = IP6_AUTO_LINKLOCAL;
-#else
-	V_ip6_auto_linklocal = 1;	/* enabled by default */
-#endif
 	TUNABLE_INT_FETCH("net.inet6.ip6.auto_linklocal",
 	    &V_ip6_auto_linklocal);
 
-#ifndef IPV6FORWARDING
-#ifdef GATEWAY6
-#define IPV6FORWARDING	1	/* forward IP6 packets not for us */
-#else
-#define IPV6FORWARDING	0	/* don't forward IP6 packets not for us */
-#endif /* GATEWAY6 */
-#endif /* !IPV6FORWARDING */
-
-#ifndef IPV6_SENDREDIRECTS
-#define IPV6_SENDREDIRECTS	1
-#endif
-
-	V_ip6_forwarding = IPV6FORWARDING; /* act as router? */
-	V_ip6_sendredirects = IPV6_SENDREDIRECTS;
-	V_ip6_defhlim = IPV6_DEFHLIM;
-	V_ip6_defmcasthlim = IPV6_DEFAULT_MULTICAST_HOPS;
-	V_ip6_accept_rtadv = 0;
-	V_ip6_log_interval = 5;
-	V_ip6_hdrnestlimit = 15; /* How many header options will we process? */
-	V_ip6_dad_count = 1;	 /* DupAddrDetectionTransmits */
-	V_ip6_auto_flowlabel = 1;
-	V_ip6_use_deprecated = 1;/* allow deprecated addr (RFC2462 5.5.4) */
-	V_ip6_rr_prune = 5;	 /* router renumbering prefix
-                                  * walk list every 5 sec. */
-	V_ip6_mcast_pmtu = 0;	 /* enable pMTU discovery for multicast? */
-	V_ip6_v6only = 1;
-	V_ip6_keepfaith = 0;
-	V_ip6_log_time = (time_t)0L;
-#ifdef IPSTEALTH
-	V_ip6stealth = 0;
-#endif
-	V_nd6_onlink_ns_rfc4861 = 0; /* allow 'on-link' nd6 NS (RFC 4861) */
-
-	V_pmtu_expire = 60*10;
-	V_pmtu_probe = 60*2;
-
-	/* raw IP6 parameters */
-	/*
-	 * Nominal space allocated to a raw ip socket.
-	 */
-#define RIPV6SNDQ	8192
-#define RIPV6RCVQ	8192
-	V_rip6_sendspace = RIPV6SNDQ;
-	V_rip6_recvspace = RIPV6RCVQ;
-
-	/* ICMPV6 parameters */
-	V_icmp6_rediraccept = 1;	/* accept and process redirects */
-	V_icmp6_redirtimeout = 10 * 60;	/* 10 minutes */
-	V_icmp6errppslim = 100;		/* 100pps */
-	/* control how to respond to NI queries */
-	V_icmp6_nodeinfo = (ICMP6_NODEINFO_FQDNOK|ICMP6_NODEINFO_NODEADDROK);
-
-	/* UDP on IP6 parameters */
-	V_udp6_sendspace = 9216;	/* really max datagram size */
-	V_udp6_recvspace = 40 * (1024 + sizeof(struct sockaddr_in6));
-					/* 40 1K datagrams */
-	V_dad_init = 0;
-
 	TAILQ_INIT(&V_in6_ifaddrhead);
 
-	scope6_init();
-	addrsel_policy_init();
-	nd6_init();
-	frag6_init();
-
-	V_ip6_desync_factor = arc4random() % MAX_TEMP_DESYNC_FACTOR;
-
 	/* Initialize packet filter hooks. */
 	V_inet6_pfil_hook.ph_type = PFIL_TYPE_AF;
 	V_inet6_pfil_hook.ph_af = AF_INET6;
@@ -254,6 +164,13 @@ ip6_init(void)
 		printf("%s: WARNING: unable to register pfil hook, "
 			"error %d\n", __func__, i);
 
+	scope6_init();
+	addrsel_policy_init();
+	nd6_init();
+	frag6_init();
+
+	V_ip6_desync_factor = arc4random() % MAX_TEMP_DESYNC_FACTOR;
+
 	/* Skip global initialization stuff for non-default instances. */
 	if (!IS_DEFAULT_VNET(curvnet))
 		return;
diff --git a/sys/netinet6/ip6_mroute.c b/sys/netinet6/ip6_mroute.c
index 54960620d28..0c1ff78bbda 100644
--- a/sys/netinet6/ip6_mroute.c
+++ b/sys/netinet6/ip6_mroute.c
@@ -130,7 +130,6 @@ static MALLOC_DEFINE(M_MRTABLE6, "mf6c", "multicast forwarding cache entry");
 
 static int	ip6_mdq(struct mbuf *, struct ifnet *, struct mf6c *);
 static void	phyint_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
-static void	pim6_init(void);
 static int	register_send(struct ip6_hdr *, struct mif6 *, struct mbuf *);
 static int	set_pim6(int *);
 static int	socket_send(struct socket *, struct mbuf *,
@@ -148,12 +147,11 @@ static const struct ip6protosw in6_pim_protosw = {
 	.pr_input =		pim6_input,
 	.pr_output =		rip6_output,
 	.pr_ctloutput =		rip6_ctloutput,
-	.pr_init =		pim6_init,
 	.pr_usrreqs =		&rip6_usrreqs
 };
 static int pim6_encapcheck(const struct mbuf *, int, int, void *);
 
-static VNET_DEFINE(int, ip6_mrouter_ver);
+static VNET_DEFINE(int, ip6_mrouter_ver) = 0;
 #define	V_ip6_mrouter_ver	VNET(ip6_mrouter_ver)
 
 SYSCTL_DECL(_net_inet6);
@@ -212,7 +210,7 @@ static struct mtx mif6_mtx;
 #define	MIF6_LOCK_DESTROY()	mtx_destroy(&mif6_mtx)
 
 #ifdef MRT6DEBUG
-static VNET_DEFINE(u_int, mrt6debug);		/* debug level */
+static VNET_DEFINE(u_int, mrt6debug) = 0;	/* debug level */
 #define	V_mrt6debug		VNET(mrt6debug)
 #define DEBUG_MFC	0x02
 #define DEBUG_FORWARD	0x04
@@ -338,15 +336,6 @@ int X_ip6_mrouter_set(struct socket *, struct sockopt *);
 int X_ip6_mrouter_get(struct socket *, struct sockopt *);
 int X_mrt6_ioctl(u_long, caddr_t);
 
-static void
-pim6_init(void)
-{
-
-#ifdef MRT6DEBUG
-	V_mrt6debug = 0;	/* debug level */
-#endif
-}
-
 /*
  * Handle MRT setsockopt commands to modify the multicast routing tables.
  */
@@ -533,11 +522,7 @@ static int
 ip6_mrouter_init(struct socket *so, int v, int cmd)
 {
 
-	V_ip6_mrouter_ver = 0;
-
 #ifdef MRT6DEBUG
-	V_mrt6debug = 0;
-
 	if (V_mrt6debug)
 		log(LOG_DEBUG,
 		    "ip6_mrouter_init: so_type = %d, pr_protocol = %d\n",
diff --git a/sys/netinet6/ip6_var.h b/sys/netinet6/ip6_var.h
index a0a0f3a18d8..3e91a79d0b9 100644
--- a/sys/netinet6/ip6_var.h
+++ b/sys/netinet6/ip6_var.h
@@ -295,12 +295,20 @@ VNET_DECLARE(struct ip6stat, ip6stat);	/* statistics */
 VNET_DECLARE(int, ip6_defhlim);		/* default hop limit */
 VNET_DECLARE(int, ip6_defmcasthlim);	/* default multicast hop limit */
 VNET_DECLARE(int, ip6_forwarding);	/* act as router? */
-VNET_DECLARE(int, ip6_gif_hlim);	/* Hop limit for gif encap packet */
 VNET_DECLARE(int, ip6_use_deprecated);	/* allow deprecated addr as source */
 VNET_DECLARE(int, ip6_rr_prune);	/* router renumbering prefix
 					 * walk list every 5 sec.    */
 VNET_DECLARE(int, ip6_mcast_pmtu);	/* enable pMTU discovery for multicast? */
 VNET_DECLARE(int, ip6_v6only);
+#define	V_ip6stat			VNET(ip6stat)
+#define	V_ip6_defhlim			VNET(ip6_defhlim)
+#define	V_ip6_defmcasthlim		VNET(ip6_defmcasthlim)
+#define	V_ip6_forwarding		VNET(ip6_forwarding)
+#define	V_ip6_use_deprecated		VNET(ip6_use_deprecated)
+#define	V_ip6_rr_prune			VNET(ip6_rr_prune)
+#define	V_ip6_mcast_pmtu		VNET(ip6_mcast_pmtu)
+#define	V_ip6_v6only			VNET(ip6_v6only)
+
 VNET_DECLARE(struct socket *, ip6_mrouter);	/* multicast routing daemon */
 VNET_DECLARE(int, ip6_sendredirects);	/* send IP redirects when forwarding? */
 VNET_DECLARE(int, ip6_maxfragpackets);	/* Maximum packets in reassembly
@@ -314,31 +322,6 @@ VNET_DECLARE(time_t, ip6_log_time);
 VNET_DECLARE(int, ip6_hdrnestlimit);	/* upper limit of # of extension
 					 * headers */
 VNET_DECLARE(int, ip6_dad_count);	/* DupAddrDetectionTransmits */
-
-VNET_DECLARE(int, ip6_auto_flowlabel);
-VNET_DECLARE(int, ip6_auto_linklocal);
-
-VNET_DECLARE(int, ip6_use_tempaddr);	/* Whether to use temporary addresses */
-VNET_DECLARE(int, ip6_prefer_tempaddr);	/* Whether to prefer temporary
-					 * addresses in the source address
-					 * selection */
-
-#ifdef IPSTEALTH
-VNET_DECLARE(int, ip6stealth);
-#endif
-
-VNET_DECLARE(int, ip6_use_defzone);	/* Whether to use the default scope
-					 * zone when unspecified */
-
-#define	V_ip6stat			VNET(ip6stat)
-#define	V_ip6_defhlim			VNET(ip6_defhlim)
-#define	V_ip6_defmcasthlim		VNET(ip6_defmcasthlim)
-#define	V_ip6_forwarding		VNET(ip6_forwarding)
-#define	V_ip6_gif_hlim			VNET(ip6_gif_hlim)
-#define	V_ip6_use_deprecated		VNET(ip6_use_deprecated)
-#define	V_ip6_rr_prune			VNET(ip6_rr_prune)
-#define	V_ip6_mcast_pmtu		VNET(ip6_mcast_pmtu)
-#define	V_ip6_v6only			VNET(ip6_v6only)
 #define	V_ip6_mrouter			VNET(ip6_mrouter)
 #define	V_ip6_sendredirects		VNET(ip6_sendredirects)
 #define	V_ip6_maxfragpackets		VNET(ip6_maxfragpackets)
@@ -349,17 +332,29 @@ VNET_DECLARE(int, ip6_use_defzone);	/* Whether to use the default scope
 #define	V_ip6_log_time			VNET(ip6_log_time)
 #define	V_ip6_hdrnestlimit		VNET(ip6_hdrnestlimit)
 #define	V_ip6_dad_count			VNET(ip6_dad_count)
+
+VNET_DECLARE(int, ip6_auto_flowlabel);
+VNET_DECLARE(int, ip6_auto_linklocal);
 #define	V_ip6_auto_flowlabel		VNET(ip6_auto_flowlabel)
 #define	V_ip6_auto_linklocal		VNET(ip6_auto_linklocal)
+
+VNET_DECLARE(int, ip6_use_tempaddr);	/* Whether to use temporary addresses */
+VNET_DECLARE(int, ip6_prefer_tempaddr);	/* Whether to prefer temporary
+					 * addresses in the source address
+					 * selection */
 #define	V_ip6_use_tempaddr		VNET(ip6_use_tempaddr)
 #define	V_ip6_prefer_tempaddr		VNET(ip6_prefer_tempaddr)
-#ifdef IPSTEALTH
-#define	V_ip6stealth			VNET(ip6stealth)
-#endif
+
+VNET_DECLARE(int, ip6_use_defzone);	/* Whether to use the default scope
+					 * zone when unspecified */
 #define	V_ip6_use_defzone		VNET(ip6_use_defzone)
 
 VNET_DECLARE (struct pfil_head, inet6_pfil_hook);	/* packet filter hooks */
 #define	V_inet6_pfil_hook	VNET(inet6_pfil_hook)
+#ifdef IPSTEALTH
+VNET_DECLARE(int, ip6stealth);
+#define	V_ip6stealth			VNET(ip6stealth)
+#endif
 
 extern struct	pr_usrreqs rip6_usrreqs;
 struct sockopt;
diff --git a/sys/netinet6/nd6.c b/sys/netinet6/nd6.c
index a0ef2040de1..7a54226cc63 100644
--- a/sys/netinet6/nd6.c
+++ b/sys/netinet6/nd6.c
@@ -82,21 +82,31 @@ __FBSDID("$FreeBSD$");
 
 #define SIN6(s) ((struct sockaddr_in6 *)s)
 
-VNET_DEFINE(int, nd6_prune);
-VNET_DEFINE(int, nd6_delay);
-VNET_DEFINE(int, nd6_umaxtries);
-VNET_DEFINE(int, nd6_mmaxtries);
-VNET_DEFINE(int, nd6_useloopback);
-VNET_DEFINE(int, nd6_gctimer);
+/* timer values */
+VNET_DEFINE(int, nd6_prune)	= 1;	/* walk list every 1 seconds */
+VNET_DEFINE(int, nd6_delay)	= 5;	/* delay first probe time 5 second */
+VNET_DEFINE(int, nd6_umaxtries)	= 3;	/* maximum unicast query */
+VNET_DEFINE(int, nd6_mmaxtries)	= 3;	/* maximum multicast query */
+VNET_DEFINE(int, nd6_useloopback) = 1;	/* use loopback interface for
+					 * local traffic */
+VNET_DEFINE(int, nd6_gctimer)	= (60 * 60 * 24); /* 1 day: garbage
+					 * collection timer */
 
 /* preventing too many loops in ND option parsing */
-static VNET_DEFINE(int, nd6_maxndopt);
-VNET_DEFINE(int, nd6_maxnudhint);
-static VNET_DEFINE(int, nd6_maxqueuelen);
+static VNET_DEFINE(int, nd6_maxndopt) = 10; /* max # of ND options allowed */
+
+VNET_DEFINE(int, nd6_maxnudhint) = 0;	/* max # of subsequent upper
+					 * layer hints */
+static VNET_DEFINE(int, nd6_maxqueuelen) = 1; /* max pkts cached in unresolved
+					 * ND entries */
 #define	V_nd6_maxndopt			VNET(nd6_maxndopt)
 #define	V_nd6_maxqueuelen		VNET(nd6_maxqueuelen)
 
-VNET_DEFINE(int, nd6_debug);
+#ifdef ND6_DEBUG
+VNET_DEFINE(int, nd6_debug) = 1;
+#else
+VNET_DEFINE(int, nd6_debug) = 0;
+#endif
 
 /* for debugging? */
 #if 0
@@ -106,7 +116,7 @@ static int nd6_inuse, nd6_allocated;
 VNET_DEFINE(struct nd_drhead, nd_defrouter);
 VNET_DEFINE(struct nd_prhead, nd_prefix);
 
-VNET_DEFINE(int, nd6_recalc_reachtm_interval);
+VNET_DEFINE(int, nd6_recalc_reachtm_interval) = ND6_RECALC_REACHTM_INTERVAL;
 #define	V_nd6_recalc_reachtm_interval	VNET(nd6_recalc_reachtm_interval)
 
 static struct sockaddr_in6 all1_sa;
@@ -125,56 +135,13 @@ static VNET_DEFINE(struct callout, nd6_slowtimo_ch);
 
 VNET_DEFINE(struct callout, nd6_timer_ch);
 
-VNET_DECLARE(int, dad_ignore_ns);
-VNET_DECLARE(int, dad_maxtry);
-#define	V_dad_ignore_ns			VNET(dad_ignore_ns)
-#define	V_dad_maxtry			VNET(dad_maxtry)
-
 void
 nd6_init(void)
 {
 	int i;
 
-	V_nd6_prune	= 1;	/* walk list every 1 seconds */
-	V_nd6_delay	= 5;	/* delay first probe time 5 second */
-	V_nd6_umaxtries	= 3;	/* maximum unicast query */
-	V_nd6_mmaxtries	= 3;	/* maximum multicast query */
-	V_nd6_useloopback = 1;	/* use loopback interface for local traffic */
-	V_nd6_gctimer	= (60 * 60 * 24); /* 1 day: garbage collection timer */
-
-	/* preventing too many loops in ND option parsing */
-	V_nd6_maxndopt = 10;	/* max # of ND options allowed */
-
-	V_nd6_maxnudhint = 0;	/* max # of subsequent upper layer hints */
-	V_nd6_maxqueuelen = 1;	/* max pkts cached in unresolved ND entries */
-
-#ifdef ND6_DEBUG
-	V_nd6_debug = 1;
-#else
-	V_nd6_debug = 0;
-#endif
-
-	V_nd6_recalc_reachtm_interval = ND6_RECALC_REACHTM_INTERVAL;
-
-	V_dad_ignore_ns = 0;	/* ignore NS in DAD - specwise incorrect*/
-	V_dad_maxtry = 15;	/* max # of *tries* to transmit DAD packet */
-
-	/*
-	 * XXX just to get this to compile KMM
-	 */
-#ifdef notyet
-	V_llinfo_nd6.ln_next = &V_llinfo_nd6;
-	V_llinfo_nd6.ln_prev = &V_llinfo_nd6;
-#endif
 	LIST_INIT(&V_nd_prefix);
 
-	V_ip6_use_tempaddr = 0;
-	V_ip6_temp_preferred_lifetime = DEF_TEMP_PREFERRED_LIFETIME;
-	V_ip6_temp_valid_lifetime = DEF_TEMP_VALID_LIFETIME;
-	V_ip6_temp_regen_advance = TEMPADDR_REGEN_ADVANCE;
-
-	V_ip6_desync_factor = 0;
-
 	all1_sa.sin6_family = AF_INET6;
 	all1_sa.sin6_len = sizeof(struct sockaddr_in6);
 	for (i = 0; i < sizeof(all1_sa.sin6_addr); i++)
@@ -182,13 +149,13 @@ nd6_init(void)
 
 	/* initialization of the default router list */
 	TAILQ_INIT(&V_nd_defrouter);
+
 	/* start timer */
 	callout_init(&V_nd6_slowtimo_ch, 0);
 	callout_reset(&V_nd6_slowtimo_ch, ND6_SLOWTIMER_INTERVAL * hz,
 	    nd6_slowtimo, curvnet);
 }
 
-
 #ifdef VIMAGE
 void
 nd6_destroy()
diff --git a/sys/netinet6/nd6.h b/sys/netinet6/nd6.h
index ff8faa2acfa..abcfcb79803 100644
--- a/sys/netinet6/nd6.h
+++ b/sys/netinet6/nd6.h
@@ -330,8 +330,6 @@ VNET_DECLARE(struct nd_drhead, nd_defrouter);
 VNET_DECLARE(struct nd_prhead, nd_prefix);
 VNET_DECLARE(int, nd6_debug);
 VNET_DECLARE(int, nd6_onlink_ns_rfc4861);
-VNET_DECLARE(struct callout, nd6_timer_ch);
-
 #define	V_nd6_prune			VNET(nd6_prune)
 #define	V_nd6_delay			VNET(nd6_delay)
 #define	V_nd6_umaxtries			VNET(nd6_umaxtries)
@@ -343,6 +341,10 @@ VNET_DECLARE(struct callout, nd6_timer_ch);
 #define	V_nd_prefix			VNET(nd_prefix)
 #define	V_nd6_debug			VNET(nd6_debug)
 #define	V_nd6_onlink_ns_rfc4861		VNET(nd6_onlink_ns_rfc4861)
+
+#define nd6log(x)	do { if (V_nd6_debug) log x; } while (/*CONSTCOND*/ 0)
+
+VNET_DECLARE(struct callout, nd6_timer_ch);
 #define	V_nd6_timer_ch			VNET(nd6_timer_ch)
 
 /* nd6_rtr.c */
@@ -351,15 +353,12 @@ VNET_DECLARE(int, ip6_desync_factor);	/* seconds */
 VNET_DECLARE(u_int32_t, ip6_temp_preferred_lifetime); /* seconds */
 VNET_DECLARE(u_int32_t, ip6_temp_valid_lifetime); /* seconds */
 VNET_DECLARE(int, ip6_temp_regen_advance); /* seconds */
-
 #define	V_nd6_defifindex		VNET(nd6_defifindex)
 #define	V_ip6_desync_factor		VNET(ip6_desync_factor)
 #define	V_ip6_temp_preferred_lifetime	VNET(ip6_temp_preferred_lifetime)
 #define	V_ip6_temp_valid_lifetime	VNET(ip6_temp_valid_lifetime)
 #define	V_ip6_temp_regen_advance	VNET(ip6_temp_regen_advance)
 
-#define nd6log(x)	do { if (V_nd6_debug) log x; } while (/*CONSTCOND*/ 0)
-
 union nd_opts {
 	struct nd_opt_hdr *nd_opt_array[8];	/* max = target address list */
 	struct {
diff --git a/sys/netinet6/nd6_nbr.c b/sys/netinet6/nd6_nbr.c
index f9061d36f8f..a50925be5dc 100644
--- a/sys/netinet6/nd6_nbr.c
+++ b/sys/netinet6/nd6_nbr.c
@@ -89,9 +89,8 @@ static void nd6_dad_ns_output(struct dadq *, struct ifaddr *);
 static void nd6_dad_ns_input(struct ifaddr *);
 static void nd6_dad_na_input(struct ifaddr *);
 
-VNET_DEFINE(int, dad_ignore_ns);
-VNET_DEFINE(int, dad_maxtry);
-
+VNET_DEFINE(int, dad_ignore_ns) = 0;	/* ignore NS in DAD - specwise incorrect*/
+VNET_DEFINE(int, dad_maxtry) = 15;	/* max # of *tries* to transmit DAD packet */
 #define	V_dad_ignore_ns			VNET(dad_ignore_ns)
 #define	V_dad_maxtry			VNET(dad_maxtry)
 
@@ -1124,9 +1123,8 @@ struct dadq {
 };
 
 static VNET_DEFINE(TAILQ_HEAD(, dadq), dadq);
+VNET_DEFINE(int, dad_init) = 0;
 #define	V_dadq				VNET(dadq)
-
-VNET_DEFINE(int, dad_init);
 #define	V_dad_init			VNET(dad_init)
 
 static struct dadq *
diff --git a/sys/netinet6/nd6_rtr.c b/sys/netinet6/nd6_rtr.c
index 74f4954bf0f..19ec989fb22 100644
--- a/sys/netinet6/nd6_rtr.c
+++ b/sys/netinet6/nd6_rtr.c
@@ -90,14 +90,16 @@ VNET_DECLARE(int, nd6_recalc_reachtm_interval);
 #define	V_nd6_recalc_reachtm_interval	VNET(nd6_recalc_reachtm_interval)
 
 static VNET_DEFINE(struct ifnet *, nd6_defifp);
+VNET_DEFINE(int, nd6_defifindex);
 #define	V_nd6_defifp			VNET(nd6_defifp)
 
-VNET_DEFINE(int, nd6_defifindex);
-VNET_DEFINE(int, ip6_use_tempaddr);
+VNET_DEFINE(int, ip6_use_tempaddr) = 0;
+
 VNET_DEFINE(int, ip6_desync_factor);
-VNET_DEFINE(u_int32_t, ip6_temp_preferred_lifetime);
-VNET_DEFINE(u_int32_t, ip6_temp_valid_lifetime);
-VNET_DEFINE(int, ip6_temp_regen_advance);
+VNET_DEFINE(u_int32_t, ip6_temp_preferred_lifetime) = DEF_TEMP_PREFERRED_LIFETIME;
+VNET_DEFINE(u_int32_t, ip6_temp_valid_lifetime) = DEF_TEMP_VALID_LIFETIME;
+
+VNET_DEFINE(int, ip6_temp_regen_advance) = TEMPADDR_REGEN_ADVANCE;
 
 /* RTPREF_MEDIUM has to be 0! */
 #define RTPREF_HIGH	1
diff --git a/sys/netinet6/raw_ip6.c b/sys/netinet6/raw_ip6.c
index 6052b24e470..939aa6a5f98 100644
--- a/sys/netinet6/raw_ip6.c
+++ b/sys/netinet6/raw_ip6.c
@@ -119,11 +119,11 @@ VNET_DECLARE(struct inpcbinfo, ripcbinfo);
 #define	V_ripcb				VNET(ripcb)
 #define	V_ripcbinfo			VNET(ripcbinfo)
 
-VNET_DEFINE(struct rip6stat, rip6stat);
-
 extern u_long	rip_sendspace;
 extern u_long	rip_recvspace;
 
+VNET_DEFINE(struct rip6stat, rip6stat);
+
 /*
  * Hooks for multicast routing. They all default to NULL, so leave them not
  * initialized and rely on BSS being set to 0.
diff --git a/sys/netinet6/scope6.c b/sys/netinet6/scope6.c
index cced5e304ac..8189d87c5bc 100644
--- a/sys/netinet6/scope6.c
+++ b/sys/netinet6/scope6.c
@@ -50,6 +50,11 @@ __FBSDID("$FreeBSD$");
 #include <netinet6/ip6_var.h>
 #include <netinet6/scope6_var.h>
 
+#ifdef ENABLE_DEFAULT_SCOPE
+VNET_DEFINE(int, ip6_use_defzone) = 1;
+#else
+VNET_DEFINE(int, ip6_use_defzone) = 0;
+#endif
 
 /*
  * The scope6_lock protects the global sid default stored in
@@ -62,8 +67,6 @@ static struct mtx scope6_lock;
 #define	SCOPE6_LOCK_ASSERT()	mtx_assert(&scope6_lock, MA_OWNED)
 
 static VNET_DEFINE(struct scope6_id, sid_default);
-VNET_DEFINE(int, ip6_use_defzone);
-
 #define	V_sid_default			VNET(sid_default)
 
 #define SID(ifp) \
@@ -73,11 +76,6 @@ void
 scope6_init(void)
 {
 
-#ifdef ENABLE_DEFAULT_SCOPE
-	V_ip6_use_defzone = 1;
-#else
-	V_ip6_use_defzone = 0;
-#endif
 	bzero(&V_sid_default, sizeof(V_sid_default));
 
 	if (!IS_DEFAULT_VNET(curvnet))
diff --git a/sys/netipsec/ah_var.h b/sys/netipsec/ah_var.h
index b031a987b60..6145dba9997 100644
--- a/sys/netipsec/ah_var.h
+++ b/sys/netipsec/ah_var.h
@@ -72,10 +72,11 @@ struct ahstat {
 
 #ifdef _KERNEL
 VNET_DECLARE(int, ah_enable);
-#define	V_ah_enable		VNET(ah_enable)
 VNET_DECLARE(int, ah_cleartos);
-#define	V_ah_cleartos		VNET(ah_cleartos)
 VNET_DECLARE(struct ahstat, ahstat);
+
+#define	V_ah_enable		VNET(ah_enable)
+#define	V_ah_cleartos		VNET(ah_cleartos)
 #define	V_ahstat		VNET(ahstat)
 #endif /* _KERNEL */
 #endif /*_NETIPSEC_AH_VAR_H_*/
diff --git a/sys/netipsec/esp_var.h b/sys/netipsec/esp_var.h
index dc1aad4ca44..477dcbf5b39 100644
--- a/sys/netipsec/esp_var.h
+++ b/sys/netipsec/esp_var.h
@@ -73,8 +73,9 @@ struct espstat {
 
 #ifdef _KERNEL
 VNET_DECLARE(int, esp_enable);
-#define	V_esp_enable	VNET(esp_enable)
 VNET_DECLARE(struct espstat, espstat);
+
+#define	V_esp_enable	VNET(esp_enable)
 #define	V_espstat	VNET(espstat)
 #endif /* _KERNEL */
 #endif /*_NETIPSEC_ESP_VAR_H_*/
diff --git a/sys/netipsec/ipcomp_var.h b/sys/netipsec/ipcomp_var.h
index 78f18488221..c99a3be7d90 100644
--- a/sys/netipsec/ipcomp_var.h
+++ b/sys/netipsec/ipcomp_var.h
@@ -66,8 +66,9 @@ struct ipcompstat {
 
 #ifdef _KERNEL
 VNET_DECLARE(int, ipcomp_enable);
-#define	V_ipcomp_enable		VNET(ipcomp_enable)
 VNET_DECLARE(struct ipcompstat, ipcompstat);
+
+#define	V_ipcomp_enable		VNET(ipcomp_enable)
 #define	V_ipcompstat		VNET(ipcompstat)
 #endif /* _KERNEL */
 #endif /*_NETIPSEC_IPCOMP_VAR_H_*/
diff --git a/sys/netipsec/ipip_var.h b/sys/netipsec/ipip_var.h
index a6e33e8b435..3c8c3974d60 100644
--- a/sys/netipsec/ipip_var.h
+++ b/sys/netipsec/ipip_var.h
@@ -60,8 +60,9 @@ struct ipipstat
 
 #ifdef _KERNEL
 VNET_DECLARE(int, ipip_allow);
-#define	V_ipip_allow		VNET(ipip_allow)
 VNET_DECLARE(struct ipipstat, ipipstat);
+
+#define	V_ipip_allow		VNET(ipip_allow)
 #define	V_ipipstat		VNET(ipipstat)
 #endif /* _KERNEL */
 #endif /* _NETINET_IPIP_H_ */
diff --git a/sys/netipsec/ipsec.c b/sys/netipsec/ipsec.c
index 5ee4bbb8f40..1f48dd6c3da 100644
--- a/sys/netipsec/ipsec.c
+++ b/sys/netipsec/ipsec.c
@@ -102,6 +102,7 @@ VNET_DEFINE(int, ipsec_debug) = 1;
 #else
 VNET_DEFINE(int, ipsec_debug) = 0;
 #endif
+
 /* NB: name changed so netstat doesn't use it. */
 VNET_DEFINE(struct ipsecstat, ipsec4stat);
 VNET_DEFINE(int, ip4_ah_offsetmask) = 0;	/* maybe IP_DF? */
diff --git a/sys/netipsec/ipsec.h b/sys/netipsec/ipsec.h
index 92539b9cf32..090255aafb2 100644
--- a/sys/netipsec/ipsec.h
+++ b/sys/netipsec/ipsec.h
@@ -334,39 +334,39 @@ struct ipsec_history {
 
 VNET_DECLARE(int, ipsec_debug);
 #define	V_ipsec_debug		VNET(ipsec_debug)
-VNET_DECLARE(struct ipsecstat, ipsec4stat);
-#define	V_ipsec4stat		VNET(ipsec4stat)
-VNET_DECLARE(int, ip4_ah_offsetmask);
-#define	V_ip4_ah_offsetmask	VNET(ip4_ah_offsetmask)
-VNET_DECLARE(int, ip4_ipsec_dfbit);
-#define	V_ip4_ipsec_dfbit	VNET(ip4_ipsec_dfbit)
-VNET_DECLARE(int, ip4_esp_trans_deflev);
-#define	V_ip4_esp_trans_deflev	VNET(ip4_esp_trans_deflev)
-VNET_DECLARE(int, ip4_esp_net_deflev);
-#define	V_ip4_esp_net_deflev	VNET(ip4_esp_net_deflev)
-VNET_DECLARE(int, ip4_ah_trans_deflev);
-#define	V_ip4_ah_trans_deflev	VNET(ip4_ah_trans_deflev)
-VNET_DECLARE(int, ip4_ah_net_deflev);
-#define	V_ip4_ah_net_deflev	VNET(ip4_ah_net_deflev)
-VNET_DECLARE(struct secpolicy, ip4_def_policy);
-#define	V_ip4_def_policy	VNET(ip4_def_policy)
-VNET_DECLARE(int, ip4_ipsec_ecn);
-#define	V_ip4_ipsec_ecn		VNET(ip4_ipsec_ecn)
-VNET_DECLARE(int, ip4_esp_randpad);
-#define	V_ip4_esp_randpad	VNET(ip4_esp_randpad)
-
-VNET_DECLARE(int, crypto_support);
-#define	V_crypto_support	VNET(crypto_support)
-
-extern int ip4_ah_cleartos;
 
 #ifdef REGRESSION
 VNET_DECLARE(int, ipsec_replay);
-#define	V_ipsec_replay		VNET(ipsec_replay)
 VNET_DECLARE(int, ipsec_integrity);
+
+#define	V_ipsec_replay		VNET(ipsec_replay)
 #define	V_ipsec_integrity	VNET(ipsec_integrity)
 #endif
 
+VNET_DECLARE(struct ipsecstat, ipsec4stat);
+VNET_DECLARE(struct secpolicy, ip4_def_policy);
+VNET_DECLARE(int, ip4_esp_trans_deflev);
+VNET_DECLARE(int, ip4_esp_net_deflev);
+VNET_DECLARE(int, ip4_ah_trans_deflev);
+VNET_DECLARE(int, ip4_ah_net_deflev);
+VNET_DECLARE(int, ip4_ah_offsetmask);
+VNET_DECLARE(int, ip4_ipsec_dfbit);
+VNET_DECLARE(int, ip4_ipsec_ecn);
+VNET_DECLARE(int, ip4_esp_randpad);
+VNET_DECLARE(int, crypto_support);
+
+#define	V_ipsec4stat		VNET(ipsec4stat)
+#define	V_ip4_def_policy	VNET(ip4_def_policy)
+#define	V_ip4_esp_trans_deflev	VNET(ip4_esp_trans_deflev)
+#define	V_ip4_esp_net_deflev	VNET(ip4_esp_net_deflev)
+#define	V_ip4_ah_trans_deflev	VNET(ip4_ah_trans_deflev)
+#define	V_ip4_ah_net_deflev	VNET(ip4_ah_net_deflev)
+#define	V_ip4_ah_offsetmask	VNET(ip4_ah_offsetmask)
+#define	V_ip4_ipsec_dfbit	VNET(ip4_ipsec_dfbit)
+#define	V_ip4_ipsec_ecn		VNET(ip4_ipsec_ecn)
+#define	V_ip4_esp_randpad	VNET(ip4_esp_randpad)
+#define	V_crypto_support	VNET(crypto_support)
+
 #define ipseclog(x)	do { if (V_ipsec_debug) log x; } while (0)
 /* for openbsd compatibility */
 #define	DPRINTF(x)	do { if (V_ipsec_debug) printf x; } while (0)
diff --git a/sys/netipsec/ipsec6.h b/sys/netipsec/ipsec6.h
index 030113f1413..c004220e2bf 100644
--- a/sys/netipsec/ipsec6.h
+++ b/sys/netipsec/ipsec6.h
@@ -42,16 +42,17 @@
 
 #ifdef _KERNEL
 VNET_DECLARE(struct ipsecstat, ipsec6stat);
-#define	V_ipsec6stat		VNET(ipsec6stat)
 VNET_DECLARE(int, ip6_esp_trans_deflev);
-#define	V_ip6_esp_trans_deflev	VNET(ip6_esp_trans_deflev)
 VNET_DECLARE(int, ip6_esp_net_deflev);
-#define	V_ip6_esp_net_deflev	VNET(ip6_esp_net_deflev)
 VNET_DECLARE(int, ip6_ah_trans_deflev);
-#define	V_ip6_ah_trans_deflev	VNET(ip6_ah_trans_deflev)
 VNET_DECLARE(int, ip6_ah_net_deflev);
-#define	V_ip6_ah_net_deflev	VNET(ip6_ah_net_deflev)
 VNET_DECLARE(int, ip6_ipsec_ecn);
+
+#define	V_ipsec6stat		VNET(ipsec6stat)
+#define	V_ip6_esp_trans_deflev	VNET(ip6_esp_trans_deflev)
+#define	V_ip6_esp_net_deflev	VNET(ip6_esp_net_deflev)
+#define	V_ip6_ah_trans_deflev	VNET(ip6_ah_trans_deflev)
+#define	V_ip6_ah_net_deflev	VNET(ip6_ah_net_deflev)
 #define	V_ip6_ipsec_ecn		VNET(ip6_ipsec_ecn)
 
 struct inpcb;
diff --git a/sys/netipsec/key.c b/sys/netipsec/key.c
index 2b17fb0d91b..521976812eb 100644
--- a/sys/netipsec/key.c
+++ b/sys/netipsec/key.c
@@ -114,27 +114,27 @@
 
 VNET_DEFINE(u_int32_t, key_debug_level) = 0;
 static VNET_DEFINE(u_int, key_spi_trycnt) = 1000;
-#define	V_key_spi_trycnt	VNET(key_spi_trycnt)
 static VNET_DEFINE(u_int32_t, key_spi_minval) = 0x100;
-#define	V_key_spi_minval	VNET(key_spi_minval)
 static VNET_DEFINE(u_int32_t, key_spi_maxval) = 0x0fffffff;	/* XXX */
-#define	V_key_spi_maxval	VNET(key_spi_maxval)
 static VNET_DEFINE(u_int32_t, policy_id) = 0;
-#define	V_policy_id		VNET(policy_id)
 /*interval to initialize randseed,1(m)*/
 static VNET_DEFINE(u_int, key_int_random) = 60;
-#define	V_key_int_random	VNET(key_int_random)
 /* interval to expire acquiring, 30(s)*/
 static VNET_DEFINE(u_int, key_larval_lifetime) = 30;
-#define	V_key_larval_lifetime	VNET(key_larval_lifetime)
 /* counter for blocking SADB_ACQUIRE.*/
 static VNET_DEFINE(int, key_blockacq_count) = 10;
-#define	V_key_blockacq_count	VNET(key_blockacq_count)
 /* lifetime for blocking SADB_ACQUIRE.*/
 static VNET_DEFINE(int, key_blockacq_lifetime) = 20;
-#define	V_key_blockacq_lifetime	VNET(key_blockacq_lifetime)
 /* preferred old sa rather than new sa.*/
 static VNET_DEFINE(int, key_preferred_oldsa) = 1;
+#define	V_key_spi_trycnt	VNET(key_spi_trycnt)
+#define	V_key_spi_minval	VNET(key_spi_minval)
+#define	V_key_spi_maxval	VNET(key_spi_maxval)
+#define	V_policy_id		VNET(policy_id)
+#define	V_key_int_random	VNET(key_int_random)
+#define	V_key_larval_lifetime	VNET(key_larval_lifetime)
+#define	V_key_blockacq_count	VNET(key_blockacq_count)
+#define	V_key_blockacq_lifetime	VNET(key_blockacq_lifetime)
 #define	V_key_preferred_oldsa	VNET(key_preferred_oldsa)
 
 static VNET_DEFINE(u_int32_t, acq_seq) = 0;
@@ -270,10 +270,11 @@ static const int maxsize[] = {
 };
 
 static VNET_DEFINE(int, ipsec_esp_keymin) = 256;
-#define	V_ipsec_esp_keymin	VNET(ipsec_esp_keymin)
 static VNET_DEFINE(int, ipsec_esp_auth) = 0;
-#define	V_ipsec_esp_auth	VNET(ipsec_esp_auth)
 static VNET_DEFINE(int, ipsec_ah_keymin) = 128;
+
+#define	V_ipsec_esp_keymin	VNET(ipsec_esp_keymin)
+#define	V_ipsec_esp_auth	VNET(ipsec_esp_auth)
 #define	V_ipsec_ah_keymin	VNET(ipsec_ah_keymin)
 
 #ifdef SYSCTL_DECL
diff --git a/sys/netipsec/xform_esp.c b/sys/netipsec/xform_esp.c
index dbacd771475..6186b5961a5 100644
--- a/sys/netipsec/xform_esp.c
+++ b/sys/netipsec/xform_esp.c
@@ -85,8 +85,7 @@ SYSCTL_VNET_INT(_net_inet_esp, OID_AUTO,
 SYSCTL_VNET_STRUCT(_net_inet_esp, IPSECCTL_STATS,
 	stats,		CTLFLAG_RD,	&VNET_NAME(espstat),	espstat, "");
 
-/* max iv length over all algorithms */
-static VNET_DEFINE(int, esp_max_ivlen) = 0;
+static VNET_DEFINE(int, esp_max_ivlen);	/* max iv length over all algorithms */
 #define	V_esp_max_ivlen	VNET(esp_max_ivlen)
 
 static int esp_input_cb(struct cryptop *op);

From 4f43f86e0b4563362ed338e4cc23420ba425ad9f Mon Sep 17 00:00:00 2001
From: Rui Paulo <rpaulo@FreeBSD.org>
Date: Thu, 29 Apr 2010 15:19:11 +0000
Subject: [PATCH 326/532] Decrement rxa_pad due to the addition of rxa_private.

---
 sys/net80211/ieee80211_ht.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/net80211/ieee80211_ht.h b/sys/net80211/ieee80211_ht.h
index 21e7ff631d8..9cdc1ea9a4d 100644
--- a/sys/net80211/ieee80211_ht.h
+++ b/sys/net80211/ieee80211_ht.h
@@ -143,7 +143,7 @@ struct ieee80211_rx_ampdu {
 	int		rxa_nframes;	/* frames since ADDBA */
 	struct mbuf *rxa_m[IEEE80211_AGGR_BAWMAX];
 	void		*rxa_private;
-	uint64_t	rxa_pad[4];
+	uint64_t	rxa_pad[3];
 };
 
 void	ieee80211_ht_attach(struct ieee80211com *);

From 5571414ca8f80a1bb0a19d3299f5959b98cc1591 Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 29 Apr 2010 15:36:32 +0000
Subject: [PATCH 327/532] Fix a problem where hastd will stuck in recv(2) after
 sending request to secondary, which died between send(2) and recv(2). Do it
 by adding timeout to recv(2) for primary incoming and outgoing sockets and
 secondary outgoing socket.

Reported by:	Mikolaj Golub <to.my.trociny@gmail.com>
Tested by:	Mikolaj Golub <to.my.trociny@gmail.com>
MFC after:	3 days
---
 sbin/hastd/hast.conf.5    |  7 +++++++
 sbin/hastd/hast.h         |  3 +++
 sbin/hastd/hastd.c        |  4 ++++
 sbin/hastd/parse.y        | 32 +++++++++++++++++++++++++++++++-
 sbin/hastd/primary.c      |  6 ++++++
 sbin/hastd/proto.c        | 26 ++++++++++++++++++++++++++
 sbin/hastd/proto.h        |  1 +
 sbin/hastd/proto_common.c |  4 ++--
 sbin/hastd/secondary.c    |  6 ++++++
 sbin/hastd/token.l        |  1 +
 10 files changed, 87 insertions(+), 3 deletions(-)

diff --git a/sbin/hastd/hast.conf.5 b/sbin/hastd/hast.conf.5
index 5734ee8ef3d..1ccd47941c6 100644
--- a/sbin/hastd/hast.conf.5
+++ b/sbin/hastd/hast.conf.5
@@ -58,6 +58,7 @@ file is following:
 control <addr>
 listen <addr>
 replication <mode>
+timeout <seconds>
 
 on <node> {
 	# Node section
@@ -76,6 +77,7 @@ resource <name> {
 	replication <mode>
 	name <name>
 	local <path>
+	timeout <seconds>
 
 	on <node> {
 		# Resource-node section
@@ -194,6 +196,11 @@ The
 .Ic async
 replication mode is currently not implemented.
 .El
+.It Ic timeout Aq seconds
+.Pp
+Connection timeout in seconds.
+The default value is
+.Va 5 .
 .It Ic name Aq name
 .Pp
 GEOM provider name that will appear as
diff --git a/sbin/hastd/hast.h b/sbin/hastd/hast.h
index c5220b53be9..2230afb20d6 100644
--- a/sbin/hastd/hast.h
+++ b/sbin/hastd/hast.h
@@ -75,6 +75,7 @@
 #define	HIO_DELETE		3
 #define	HIO_FLUSH		4
 
+#define	HAST_TIMEOUT	5
 #define	HAST_CONFIG	"/etc/hast.conf"
 #define	HAST_CONTROL	"/var/run/hastctl"
 #define	HASTD_PORT	8457
@@ -148,6 +149,8 @@ struct hast_resource {
 	/* Token to verify both in and out connection are coming from
 	   the same node (not necessarily from the same address). */
 	unsigned char hr_token[HAST_TOKEN_SIZE];
+	/* Connection timeout. */
+	int	hr_timeout;
 
 	/* Resource unique identifier. */
 	uint64_t hr_resuid;
diff --git a/sbin/hastd/hastd.c b/sbin/hastd/hastd.c
index 7b2a8e255d2..7b1e2289fd0 100644
--- a/sbin/hastd/hastd.c
+++ b/sbin/hastd/hastd.c
@@ -187,6 +187,10 @@ listen_accept(void)
 	proto_remote_address(conn, raddr, sizeof(raddr));
 	pjdlog_info("Connection from %s to %s.", laddr, raddr);
 
+	/* Error in setting timeout is not critical, but why should it fail? */
+	if (proto_timeout(conn, HAST_TIMEOUT) < 0)
+		pjdlog_errno(LOG_WARNING, "Unable to set connection timeout");
+
 	nvin = nvout = nverr = NULL;
 
 	/*
diff --git a/sbin/hastd/parse.y b/sbin/hastd/parse.y
index 67553208ac6..840a8448951 100644
--- a/sbin/hastd/parse.y
+++ b/sbin/hastd/parse.y
@@ -58,6 +58,7 @@ static bool mynode;
 static char depth0_control[HAST_ADDRSIZE];
 static char depth0_listen[HAST_ADDRSIZE];
 static int depth0_replication;
+static int depth0_timeout;
 
 static char depth1_provname[PATH_MAX];
 static char depth1_localpath[PATH_MAX];
@@ -115,6 +116,7 @@ yy_config_parse(const char *config)
 	curres = NULL;
 	mynode = false;
 
+	depth0_timeout = HAST_TIMEOUT;
 	depth0_replication = HAST_REPLICATION_MEMSYNC;
 	strlcpy(depth0_control, HAST_CONTROL, sizeof(depth0_control));
 	strlcpy(depth0_listen, HASTD_LISTEN, sizeof(depth0_listen));
@@ -154,6 +156,13 @@ yy_config_parse(const char *config)
 			 */
 			curres->hr_replication = depth0_replication;
 		}
+		if (curres->hr_timeout == -1) {
+			/*
+			 * Timeout is not set at resource-level.
+			 * Use global or default setting.
+			 */
+			curres->hr_timeout = depth0_timeout;
+		}
 	}
 
 	return (&lconfig);
@@ -171,7 +180,7 @@ yy_config_free(struct hastd_config *config)
 }
 %}
 
-%token CONTROL LISTEN PORT REPLICATION EXTENTSIZE RESOURCE NAME LOCAL REMOTE ON
+%token CONTROL LISTEN PORT REPLICATION TIMEOUT EXTENTSIZE RESOURCE NAME LOCAL REMOTE ON
 %token FULLSYNC MEMSYNC ASYNC
 %token NUM STR OB CB
 
@@ -200,6 +209,8 @@ statement:
 	|
 	replication_statement
 	|
+	timeout_statement
+	|
 	node_statement
 	|
 	resource_statement
@@ -281,6 +292,22 @@ replication_type:
 	ASYNC		{ $$ = HAST_REPLICATION_ASYNC; }
 	;
 
+timeout_statement:	TIMEOUT NUM
+	{
+		switch (depth) {
+		case 0:
+			depth0_timeout = $2;
+			break;
+		case 1:
+			if (curres != NULL)
+				curres->hr_timeout = $2;
+			break;
+		default:
+			assert(!"timeout at wrong depth level");
+		}
+	}
+	;
+
 node_statement:		ON node_start OB node_entries CB
 	{
 		mynode = false;
@@ -389,6 +416,7 @@ resource_start:	STR
 		curres->hr_role = HAST_ROLE_INIT;
 		curres->hr_previous_role = HAST_ROLE_INIT;
 		curres->hr_replication = -1;
+		curres->hr_timeout = -1;
 		curres->hr_provname[0] = '\0';
 		curres->hr_localpath[0] = '\0';
 		curres->hr_localfd = -1;
@@ -405,6 +433,8 @@ resource_entries:
 resource_entry:
 	replication_statement
 	|
+	timeout_statement
+	|
 	name_statement
 	|
 	local_statement
diff --git a/sbin/hastd/primary.c b/sbin/hastd/primary.c
index 73f722b215e..9f2b2c79b49 100644
--- a/sbin/hastd/primary.c
+++ b/sbin/hastd/primary.c
@@ -489,6 +489,9 @@ init_remote(struct hast_resource *res, struct proto_conn **inp,
 		    res->hr_remoteaddr);
 		goto close;
 	}
+	/* Error in setting timeout is not critical, but why should it fail? */
+	if (proto_timeout(out, res->hr_timeout) < 0)
+		pjdlog_errno(LOG_WARNING, "Unable to set connection timeout");
 	/*
 	 * First handshake step.
 	 * Setup outgoing connection with remote node.
@@ -552,6 +555,9 @@ init_remote(struct hast_resource *res, struct proto_conn **inp,
 		    res->hr_remoteaddr);
 		goto close;
 	}
+	/* Error in setting timeout is not critical, but why should it fail? */
+	if (proto_timeout(in, res->hr_timeout) < 0)
+		pjdlog_errno(LOG_WARNING, "Unable to set connection timeout");
 	nvout = nv_alloc();
 	nv_add_string(nvout, res->hr_name, "resource");
 	nv_add_uint8_array(nvout, res->hr_token, sizeof(res->hr_token),
diff --git a/sbin/hastd/proto.c b/sbin/hastd/proto.c
index 103f20ce9fd..531e7e5fc05 100644
--- a/sbin/hastd/proto.c
+++ b/sbin/hastd/proto.c
@@ -30,7 +30,9 @@
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
+#include <sys/types.h>
 #include <sys/queue.h>
+#include <sys/socket.h>
 
 #include <assert.h>
 #include <errno.h>
@@ -247,6 +249,30 @@ proto_remote_address(const struct proto_conn *conn, char *addr, size_t size)
 	conn->pc_proto->hp_remote_address(conn->pc_ctx, addr, size);
 }
 
+int
+proto_timeout(const struct proto_conn *conn, int timeout)
+{
+	struct timeval tv;
+	int fd;
+
+	assert(conn != NULL);
+	assert(conn->pc_magic == PROTO_CONN_MAGIC);
+	assert(conn->pc_proto != NULL);
+
+	fd = proto_descriptor(conn);
+	if (fd < 0)
+		return (-1);
+
+	tv.tv_sec = timeout;
+	tv.tv_usec = 0;
+	if (setsockopt(fd, SOL_SOCKET, SO_SNDTIMEO, &tv, sizeof(tv)) < 0)
+		return (-1);
+	if (setsockopt(fd, SOL_SOCKET, SO_RCVTIMEO, &tv, sizeof(tv)) < 0)
+		return (-1);
+
+	return (0);
+}
+
 void
 proto_close(struct proto_conn *conn)
 {
diff --git a/sbin/hastd/proto.h b/sbin/hastd/proto.h
index cb196d8da86..13248bad649 100644
--- a/sbin/hastd/proto.h
+++ b/sbin/hastd/proto.h
@@ -49,6 +49,7 @@ void proto_local_address(const struct proto_conn *conn, char *addr,
     size_t size);
 void proto_remote_address(const struct proto_conn *conn, char *addr,
     size_t size);
+int proto_timeout(const struct proto_conn *conn, int timeout);
 void proto_close(struct proto_conn *conn);
 
 #endif	/* !_PROTO_H_ */
diff --git a/sbin/hastd/proto_common.c b/sbin/hastd/proto_common.c
index 22102d86e82..131d30e19ef 100644
--- a/sbin/hastd/proto_common.c
+++ b/sbin/hastd/proto_common.c
@@ -58,7 +58,7 @@ proto_common_send(int fd, const unsigned char *data, size_t size)
 		if (done == 0)
 			return (ENOTCONN);
 		else if (done < 0) {
-			if (errno == EAGAIN)
+			if (errno == EINTR)
 				continue;
 			return (errno);
 		}
@@ -76,7 +76,7 @@ proto_common_recv(int fd, unsigned char *data, size_t size)
 
 	do {
 		done = recv(fd, data, size, MSG_WAITALL);
-	} while (done == -1 && errno == EAGAIN);
+	} while (done == -1 && errno == EINTR);
 	if (done == 0)
 		return (ENOTCONN);
 	else if (done < 0)
diff --git a/sbin/hastd/secondary.c b/sbin/hastd/secondary.c
index 6af95b5d69e..b189b7e2a2b 100644
--- a/sbin/hastd/secondary.c
+++ b/sbin/hastd/secondary.c
@@ -337,6 +337,12 @@ hastd_secondary(struct hast_resource *res, struct nv *nvin)
 
 	setproctitle("%s (secondary)", res->hr_name);
 
+	/* Error in setting timeout is not critical, but why should it fail? */
+	if (proto_timeout(res->hr_remotein, 0) < 0)
+		pjdlog_errno(LOG_WARNING, "Unable to set connection timeout");
+	if (proto_timeout(res->hr_remoteout, res->hr_timeout) < 0)
+		pjdlog_errno(LOG_WARNING, "Unable to set connection timeout");
+
 	init_local(res);
 	init_remote(res, nvin);
 	init_environment();
diff --git a/sbin/hastd/token.l b/sbin/hastd/token.l
index 7b80384e28b..e5d4ca10df2 100644
--- a/sbin/hastd/token.l
+++ b/sbin/hastd/token.l
@@ -48,6 +48,7 @@ control			{ DP; return CONTROL; }
 listen			{ DP; return LISTEN; }
 port			{ DP; return PORT; }
 replication		{ DP; return REPLICATION; }
+timeout			{ DP; return TIMEOUT; }
 resource		{ DP; return RESOURCE; }
 name			{ DP; return NAME; }
 local			{ DP; return LOCAL; }

From c6ddcbe009b2952a3e34fa6e139f666209aa1e3e Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 29 Apr 2010 15:42:24 +0000
Subject: [PATCH 328/532] - Check if the worker process was killed by signal
 and restart it. - Improve logging.

Pointed out by:	Garrett Cooper <yanefbsd@gmail.com>
MFC after:	3 days
---
 sbin/hastd/hastd.c | 44 ++++++++++++++++++++++++++------------------
 1 file changed, 26 insertions(+), 18 deletions(-)

diff --git a/sbin/hastd/hastd.c b/sbin/hastd/hastd.c
index 7b1e2289fd0..27b9bba4064 100644
--- a/sbin/hastd/hastd.c
+++ b/sbin/hastd/hastd.c
@@ -107,6 +107,22 @@ g_gate_load(void)
 	}
 }
 
+static void
+child_exit_log(unsigned int pid, int status)
+{
+
+	if (WIFEXITED(status) && WEXITSTATUS(status) == 0) {
+		pjdlog_debug(1, "Worker process exited gracefully (pid=%u).",
+		    pid);
+	} else if (WIFSIGNALED(status)) {
+		pjdlog_error("Worker process killed (pid=%u, signal=%d).",
+		    pid, WTERMSIG(status));
+	} else {
+		pjdlog_error("Worker process exited ungracefully (pid=%u, exitcode=%d).",
+		    pid, WIFEXITED(status) ? WEXITSTATUS(status) : -1);
+	}
+}
+
 static void
 child_exit(void)
 {
@@ -129,18 +145,17 @@ child_exit(void)
 		}
 		pjdlog_prefix_set("[%s] (%s) ", res->hr_name,
 		    role2str(res->hr_role));
-		if (WEXITSTATUS(status) == 0) {
-			pjdlog_debug(1,
-			    "Worker process exited gracefully (pid=%u).",
-			    (unsigned int)pid);
-		} else {
-			pjdlog_error("Worker process failed (pid=%u, status=%d).",
-			    (unsigned int)pid, WEXITSTATUS(status));
-		}
+		child_exit_log(pid, status);
 		proto_close(res->hr_ctrl);
 		res->hr_workerpid = 0;
 		if (res->hr_role == HAST_ROLE_PRIMARY) {
-			if (WEXITSTATUS(status) == EX_TEMPFAIL) {
+			/*
+			 * Restart child process if it was killed by signal
+			 * or exited because of temporary problem.
+			 */
+			if (WIFSIGNALED(status) ||
+			    (WIFEXITED(status) &&
+			     WEXITSTATUS(status) == EX_TEMPFAIL)) {
 				sleep(1);
 				pjdlog_info("Restarting worker process.");
 				hastd_primary(res);
@@ -300,19 +315,12 @@ listen_accept(void)
 			/* Wait for it to exit. */
 			else if ((pid = waitpid(res->hr_workerpid,
 			    &status, 0)) != res->hr_workerpid) {
+				/* We can only log the problem. */
 				pjdlog_errno(LOG_ERR,
 				    "Waiting for worker process (pid=%u) failed",
 				    (unsigned int)res->hr_workerpid);
-				/* See above. */
-			} else if (WEXITSTATUS(status) != 0) {
-				pjdlog_error("Worker process (pid=%u) exited ungracefully: status=%d.",
-				    (unsigned int)res->hr_workerpid,
-				    WEXITSTATUS(status));
-				/* See above. */
 			} else {
-				pjdlog_debug(1,
-				    "Worker process (pid=%u) exited gracefully.",
-				    (unsigned int)res->hr_workerpid);
+				child_exit_log(res->hr_workerpid, status);
 			}
 			res->hr_workerpid = 0;
 		} else if (res->hr_remotein != NULL) {

From 1332aaf9eddd39c77e093d5205ce57ac5ab2edd2 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 29 Apr 2010 15:47:31 +0000
Subject: [PATCH 329/532] MFamd64/i386 r207205   Clearing a page table entry's
 accessed bit and setting the page's   PG_REFERENCED flag in pmap_protect()
 can't really be justified, so   don't do it.  Moreover, on ia64, don't set
 the page's dirty field   unless pmap_protect() is removing write access.

---
 sys/ia64/ia64/pmap.c       | 16 +++++-----------
 sys/sparc64/sparc64/pmap.c |  9 +++------
 sys/sun4v/sun4v/pmap.c     | 15 ++++-----------
 3 files changed, 12 insertions(+), 28 deletions(-)

diff --git a/sys/ia64/ia64/pmap.c b/sys/ia64/ia64/pmap.c
index a7c47eff842..91b2e07fdc4 100644
--- a/sys/ia64/ia64/pmap.c
+++ b/sys/ia64/ia64/pmap.c
@@ -1450,19 +1450,13 @@ pmap_protect(pmap_t pmap, vm_offset_t sva, vm_offset_t eva, vm_prot_t prot)
 		if (pmap_prot(pte) == prot)
 			continue;
 
-		if (pmap_managed(pte)) {
-			vm_offset_t pa = pmap_ppn(pte);
+		if ((prot & VM_PROT_WRITE) == 0 &&
+		    pmap_managed(pte) && pmap_dirty(pte)) {
+			vm_paddr_t pa = pmap_ppn(pte);
 			vm_page_t m = PHYS_TO_VM_PAGE(pa);
 
-			if (pmap_dirty(pte)) {
-				vm_page_dirty(m);
-				pmap_clear_dirty(pte);
-			}
-
-			if (pmap_accessed(pte)) {
-				vm_page_flag_set(m, PG_REFERENCED);
-				pmap_clear_accessed(pte);
-			}
+			vm_page_dirty(m);
+			pmap_clear_dirty(pte);
 		}
 
 		if (prot & VM_PROT_EXECUTE)
diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index d97900c4afd..22ae860b675 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -1273,13 +1273,10 @@ pmap_protect_tte(struct pmap *pm, struct pmap *pm2, struct tte *tp,
 	u_long data;
 	vm_page_t m;
 
-	data = atomic_clear_long(&tp->tte_data, TD_REF | TD_SW | TD_W);
-	if ((data & TD_PV) != 0) {
+	data = atomic_clear_long(&tp->tte_data, TD_SW | TD_W);
+	if ((data & (TD_PV | TD_W)) == (TD_PV | TD_W)) {
 		m = PHYS_TO_VM_PAGE(TD_PA(data));
-		if ((data & TD_REF) != 0)
-			vm_page_flag_set(m, PG_REFERENCED);
-		if ((data & TD_W) != 0)
-			vm_page_dirty(m);
+		vm_page_dirty(m);
 	}
 	return (1);
 }
diff --git a/sys/sun4v/sun4v/pmap.c b/sys/sun4v/sun4v/pmap.c
index 2633b8eb76f..cbd8106e18f 100644
--- a/sys/sun4v/sun4v/pmap.c
+++ b/sys/sun4v/sun4v/pmap.c
@@ -1824,17 +1824,10 @@ pmap_protect(pmap_t pmap, vm_offset_t sva, vm_offset_t eva, vm_prot_t prot)
 		if (!anychanged && (otte_data & VTD_W))
 			anychanged = 1;
 		
-		if (otte_data & VTD_MANAGED) {
-			m = NULL;
-
-			if (otte_data & VTD_REF) {
-				m = PHYS_TO_VM_PAGE(TTE_GET_PA(otte_data));
-				vm_page_flag_set(m, PG_REFERENCED);
-			}
-			if (otte_data & VTD_W) {
-				m = PHYS_TO_VM_PAGE(TTE_GET_PA(otte_data));
-				vm_page_dirty(m);
-			}
+		if ((otte_data & (VTD_MANAGED | VTD_W)) == (VTD_MANAGED |
+		    VTD_W)) {
+			m = PHYS_TO_VM_PAGE(TTE_GET_PA(otte_data));
+			vm_page_dirty(m);
 		} 
 	}
 

From 82bfb965d1df947e0f3e0ce4605ac9fc1f18b1ed Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 29 Apr 2010 16:18:45 +0000
Subject: [PATCH 330/532] Simplify the inner loop of
 vm_pageout_object_deactivate_pages().  Rather than checking each page for
 PG_UNMANAGED, check the vm object's type. Only OBJT_PHYS can have unmanaged
 pages.  Eliminate a pointless counter. The vm object is locked, that lock is
 never released by the inner loop, and the set of pages contained by the vm
 object is not changed by the inner loop.  Therefore, the counter serves no
 purpose.

---
 sys/vm/vm_pageout.c | 11 ++++-------
 1 file changed, 4 insertions(+), 7 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 735beee4b8b..258da8b10fc 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -508,17 +508,16 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 {
 	vm_object_t backing_object, object;
 	vm_page_t p, next;
-	int actcount, rcount, remove_mode;
+	int actcount, remove_mode;
 
 	VM_OBJECT_LOCK_ASSERT(first_object, MA_OWNED);
 	if (first_object->type == OBJT_DEVICE ||
-	    first_object->type == OBJT_SG ||
-	    first_object->type == OBJT_PHYS)
+	    first_object->type == OBJT_SG)
 		return;
 	for (object = first_object;; object = backing_object) {
 		if (pmap_resident_count(pmap) <= desired)
 			goto unlock_return;
-		if (object->paging_in_progress)
+		if (object->type == OBJT_PHYS || object->paging_in_progress)
 			goto unlock_return;
 
 		remove_mode = 0;
@@ -527,10 +526,9 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 		/*
 		 * scan the objects entire memory queue
 		 */
-		rcount = object->resident_page_count;
 		p = TAILQ_FIRST(&object->memq);
 		vm_page_lock_queues();
-		while (p && (rcount-- > 0)) {
+		while (p != NULL) {
 			if (pmap_resident_count(pmap) <= desired) {
 				vm_page_unlock_queues();
 				goto unlock_return;
@@ -541,7 +539,6 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 			    p->hold_count != 0 ||
 			    p->busy != 0 ||
 			    (p->oflags & VPO_BUSY) ||
-			    (p->flags & PG_UNMANAGED) ||
 			    !pmap_page_exists_quick(pmap, p)) {
 				p = next;
 				continue;

From 9c2851d29a3b7ad24d63e3c84ed6c54df04afb30 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 17:28:07 +0000
Subject: [PATCH 331/532] Preserve unknown bits of RX MAC control register when
 driver programs RX filter configuration. It seems RX MAC control register is
 one of key registers to get various offloading features as well as
 performance. Blindly clearing unrelated bits can result in unexpected
 results.

Tested by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c | 31 ++++++++++++++++---------------
 1 file changed, 16 insertions(+), 15 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 6c737bd3f42..49f6c0e673f 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -453,8 +453,9 @@ sge_rxfilter(struct sge_softc *sc)
 	SGE_LOCK_ASSERT(sc);
 
 	ifp = sc->sge_ifp;
-	hashes[0] = hashes[1] = 0;
-	rxfilt = AcceptMyPhys;
+	rxfilt = CSR_READ_2(sc, RxMacControl);
+	rxfilt &= ~(AcceptBroadcast | AcceptAllPhys | AcceptMulticast);
+	rxfilt |= AcceptMyPhys;
 	if ((ifp->if_flags & IFF_BROADCAST) != 0)
 		rxfilt |= AcceptBroadcast;
 	if ((ifp->if_flags & (IFF_PROMISC | IFF_ALLMULTI)) != 0) {
@@ -463,20 +464,20 @@ sge_rxfilter(struct sge_softc *sc)
 		rxfilt |= AcceptMulticast;
 		hashes[0] = 0xFFFFFFFF;
 		hashes[1] = 0xFFFFFFFF;
-		goto done;
+	} else {
+		rxfilt |= AcceptMulticast;
+		hashes[0] = hashes[1] = 0;
+		/* Now program new ones. */
+		if_maddr_rlock(ifp);
+		TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
+			if (ifma->ifma_addr->sa_family != AF_LINK)
+				continue;
+			crc = ether_crc32_be(LLADDR((struct sockaddr_dl *)
+			    ifma->ifma_addr), ETHER_ADDR_LEN);
+			hashes[crc >> 31] |= 1 << ((crc >> 26) & 0x1f);
+		}
+		if_maddr_runlock(ifp);
 	}
-	rxfilt |= AcceptMulticast;
-	/* Now program new ones. */
-	if_maddr_rlock(ifp);
-	TAILQ_FOREACH(ifma, &ifp->if_multiaddrs, ifma_link) {
-		if (ifma->ifma_addr->sa_family != AF_LINK)
-			continue;
-		crc = ether_crc32_be(LLADDR((struct sockaddr_dl *)
-		    ifma->ifma_addr), ETHER_ADDR_LEN);
-		hashes[crc >> 31] |= 1 << ((crc >> 26) & 0x1f);
-	}
-	if_maddr_runlock(ifp);
-done:
 	CSR_WRITE_2(sc, RxMacControl, rxfilt | 0x02);
 	CSR_WRITE_4(sc, RxHashTable, hashes[0]);
 	CSR_WRITE_4(sc, RxHashTable2, hashes[1]);

From 60a1a1542e082ca7fce670848197697f709f0273 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 17:30:21 +0000
Subject: [PATCH 332/532] Remove wrong link state chage.

---
 sys/dev/sge/if_sge.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 49f6c0e673f..736ac6a356d 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1591,7 +1591,6 @@ sge_ifmedia_upd(struct ifnet *ifp)
 	sc = ifp->if_softc;
 	SGE_LOCK(sc);
 	mii = device_get_softc(sc->sge_miibus);
-	sc->sge_flags &= ~SGE_FLAG_LINK;
 	if (mii->mii_instance) {
 		struct mii_softc *miisc;
 		LIST_FOREACH(miisc, &mii->mii_phys, mii_list)

From 7f20f021e1f17cdc9d6428636c61c9bfcf898359 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 17:34:01 +0000
Subject: [PATCH 333/532] Explicitly marks SiS190 to differentiate it from
 SiS191.

---
 sys/dev/sge/if_sge.c    | 2 +-
 sys/dev/sge/if_sgereg.h | 1 +
 2 files changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 736ac6a356d..813f063550b 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -572,7 +572,7 @@ sge_attach(device_t dev)
 	}
 	sc->sge_rev = pci_get_revid(dev);
 	if (pci_get_device(dev) == SIS_DEVICEID_190)
-		sc->sge_flags |= SGE_FLAG_FASTETHER;
+		sc->sge_flags |= SGE_FLAG_FASTETHER | SGE_FLAG_SIS190;
 	/* Reset the adapter. */
 	sge_reset(sc);
 
diff --git a/sys/dev/sge/if_sgereg.h b/sys/dev/sge/if_sgereg.h
index 29253e0f892..074886e1fa0 100644
--- a/sys/dev/sge/if_sgereg.h
+++ b/sys/dev/sge/if_sgereg.h
@@ -331,6 +331,7 @@ struct sge_softc {
 	int			sge_timer;
 	int			sge_flags;
 #define	SGE_FLAG_FASTETHER	0x0001
+#define	SGE_FLAG_SIS190		0x0002
 #define	SGE_FLAG_RGMII		0x0010
 #define	SGE_FLAG_SPEED_1000	0x2000
 #define	SGE_FLAG_FDX		0x4000

From d1c5ee803087a715365add86000d5e05c29554ca Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 18:00:42 +0000
Subject: [PATCH 334/532] Enable FCS stripping and padding 10 bytes bit of RX
 MAC control register. Due to lack of SiS190 controller, I'm not sure whether
 this is also applicable to SiS190 so this feature is only activated on SiS191
 controller. The controller can pad 10 bytes before DMAing a received frame to
 RX buffer and received bytes include the padded bytes. This padding is very
 useful on strict-alignment architectures because driver does not have to copy
 received frame to align IP header on 4 bytes boundary. It also gives better
 RX performance on non-strict alignment architectures. Special thanks to xclin
 to give me valuable register information. Without his enthusiastic trial and
 errors this wouldn't be even possible.

While I'm here tighten validity check of received frame. Controller
clears RDS_CRCOK bit when it received bad CRC frames. xclin found
that using loop back testing.

Tested by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c    | 33 ++++++++++++++++++++++++++++-----
 sys/dev/sge/if_sgereg.h |  4 ++++
 2 files changed, 32 insertions(+), 5 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 813f063550b..d30f45697f1 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1144,7 +1144,8 @@ sge_rxeof(struct sge_softc *sc)
 		if ((rxinfo & RDC_OWN) != 0)
 			break;
 		rxstat = le32toh(cur_rx->sge_sts_size);
-		if (SGE_RX_ERROR(rxstat) != 0 || SGE_RX_NSEGS(rxstat) != 1) {
+		if ((rxstat & RDS_CRCOK) == 0 || SGE_RX_ERROR(rxstat) != 0 ||
+		    SGE_RX_NSEGS(rxstat) != 1) {
 			/* XXX We don't support multi-segment frames yet. */
 #ifdef SGE_SHOW_ERRORS
 			device_printf(sc->sge_dev, "Rx error : 0x%b\n", rxstat,
@@ -1177,11 +1178,23 @@ sge_rxeof(struct sge_softc *sc)
 		/*
 		 * TODO : VLAN hardware tag stripping.
 		 */
-		m->m_pkthdr.len = m->m_len =
-		    SGE_RX_BYTES(rxstat) - ETHER_CRC_LEN;
+		if ((sc->sge_flags & SGE_FLAG_SIS190) == 0) {
+			/*
+			 * Account for 10bytes auto padding which is used
+			 * to align IP header on 32bit boundary.  Also note,
+			 * CRC bytes is automatically removed by the
+			 * hardware.
+			 */
+			m->m_data += SGE_RX_PAD_BYTES;
+			m->m_pkthdr.len = m->m_len = SGE_RX_BYTES(rxstat) -
+			    SGE_RX_PAD_BYTES;
+		} else {
+			m->m_pkthdr.len = m->m_len = SGE_RX_BYTES(rxstat) -
+			    ETHER_CRC_LEN;
 #ifndef __NO_STRICT_ALIGNMENT
-		sge_fixup_rx(m);
+			sge_fixup_rx(m);
 #endif
+		}
 		m->m_pkthdr.rcvif = ifp;
 		ifp->if_ipackets++;
 		SGE_UNLOCK(sc);
@@ -1503,6 +1516,7 @@ sge_init_locked(struct sge_softc *sc)
 {
 	struct ifnet *ifp;
 	struct mii_data *mii;
+	uint16_t rxfilt;
 	int i;
 
 	SGE_LOCK_ASSERT(sc);
@@ -1535,10 +1549,19 @@ sge_init_locked(struct sge_softc *sc)
 	CSR_WRITE_4(sc, RxWakeOnLan, 0);
 	CSR_WRITE_4(sc, RxWakeOnLanData, 0);
 	/* Allow receiving VLAN frames. */
-	CSR_WRITE_2(sc, RxMPSControl, ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN);
+	if ((sc->sge_flags & SGE_FLAG_SIS190) == 0)
+		CSR_WRITE_2(sc, RxMPSControl,
+		    ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN + SGE_RX_PAD_BYTES);
+	else
+		CSR_WRITE_2(sc, RxMPSControl, ETHER_MAX_LEN + ETHER_VLAN_ENCAP_LEN);
 
 	for (i = 0; i < ETHER_ADDR_LEN; i++)
 		CSR_WRITE_1(sc, RxMacAddr + i, IF_LLADDR(ifp)[i]);
+	/* Configure RX MAC. */
+	rxfilt = 0;
+	if ((sc->sge_flags & SGE_FLAG_SIS190) == 0)
+		rxfilt |= RXMAC_STRIP_FCS | RXMAC_PAD_ENB;
+	CSR_WRITE_2(sc, RxMacControl, rxfilt);
 	sge_rxfilter(sc);
 
 	/* Initialize default speed/duplex information. */
diff --git a/sys/dev/sge/if_sgereg.h b/sys/dev/sge/if_sgereg.h
index 074886e1fa0..e8cf5140b6e 100644
--- a/sys/dev/sge/if_sgereg.h
+++ b/sys/dev/sge/if_sgereg.h
@@ -137,6 +137,10 @@
 #define	AcceptAllPhys		0x0100
 #define	AcceptErr		0x0020
 #define	AcceptRunt		0x0010
+#define	RXMAC_STRIP_FCS		0x0010
+#define	RXMAC_PAD_ENB		0x0004
+
+#define	SGE_RX_PAD_BYTES	10
 
 /* Station control register. */
 #define	SC_LOOPBACK		0x80000000

From c186cf13eb76ca850a32b68f35688cd87105dc46 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 18:14:14 +0000
Subject: [PATCH 335/532] Enable VLAN hardware tag insertion/stripping. Due to
 lack of SiS190 controller, I'm not sure whether this is also applicable to
 SiS190 so this feature is only activated on SiS191 controller. In theory,
 controller reinitialization is not needed when VLAN tag configuration is
 changed, but xclin said controller was not stable whenever toggling VLAN tag
 bit. To address that, sge(4) reinitialize controller for VLAN configuration
 which seems to work as expected. VLAN tag information for TX/RX descriptor
 and configure bit of RxMacControl register was found by xclin.

Submitted by:	xclin <xclin <> cs dot nctu dot edu dot tw > (initial version)
Tested by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c    | 58 ++++++++++++++++++++++++++++++++++++++---
 sys/dev/sge/if_sgereg.h |  8 ++++--
 2 files changed, 60 insertions(+), 6 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index d30f45697f1..37476a287ae 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -137,6 +137,7 @@ static int	sge_get_mac_addr_eeprom(struct sge_softc *, uint8_t *);
 static uint16_t	sge_read_eeprom(struct sge_softc *, int);
 
 static void	sge_rxfilter(struct sge_softc *);
+static void	sge_setvlan(struct sge_softc *);
 static void	sge_reset(struct sge_softc *);
 static int	sge_list_rx_init(struct sge_softc *);
 static int	sge_list_rx_free(struct sge_softc *);
@@ -483,6 +484,25 @@ sge_rxfilter(struct sge_softc *sc)
 	CSR_WRITE_4(sc, RxHashTable2, hashes[1]);
 }
 
+static void
+sge_setvlan(struct sge_softc *sc)
+{
+	struct ifnet *ifp;
+	uint16_t rxfilt;
+
+	SGE_LOCK_ASSERT(sc);
+
+	ifp = sc->sge_ifp;
+	if ((ifp->if_capabilities & IFCAP_VLAN_HWTAGGING) == 0)
+		return;
+	rxfilt = CSR_READ_2(sc, RxMacControl);
+	if ((ifp->if_capenable & IFCAP_VLAN_HWTAGGING) != 0)
+		rxfilt |= RXMAC_STRIP_VLAN;
+	else
+		rxfilt &= ~RXMAC_STRIP_VLAN;
+	CSR_WRITE_2(sc, RxMacControl, rxfilt);
+}
+
 static void
 sge_reset(struct sge_softc *sc)
 {
@@ -619,6 +639,9 @@ sge_attach(device_t dev)
 	ether_ifattach(ifp, eaddr);
 
 	/* VLAN setup. */
+	if ((sc->sge_flags & SGE_FLAG_SIS190) == 0)
+		ifp->if_capabilities |= IFCAP_VLAN_HWTAGGING |
+		    IFCAP_VLAN_HWCSUM;
 	ifp->if_capabilities |= IFCAP_VLAN_MTU;
 	ifp->if_capenable = ifp->if_capabilities;
 	/* Tell the upper layer(s) we support long frames. */
@@ -1175,9 +1198,12 @@ sge_rxeof(struct sge_softc *sc)
 				m->m_pkthdr.csum_data = 0xffff;
 			}
 		}
-		/*
-		 * TODO : VLAN hardware tag stripping.
-		 */
+		/* Check for VLAN tagged frame. */
+		if ((ifp->if_capenable & IFCAP_VLAN_HWTAGGING) != 0 &&
+		    (rxstat & RDS_VLAN) != 0) {
+			m->m_pkthdr.ether_vtag = rxinfo & RDC_VLAN_MASK;
+			m->m_flags |= M_VLANTAG;
+		}
 		if ((sc->sge_flags & SGE_FLAG_SIS190) == 0) {
 			/*
 			 * Account for 10bytes auto padding which is used
@@ -1422,6 +1448,11 @@ sge_encap(struct sge_softc *sc, struct mbuf **m_head)
 	desc->sge_flags = htole32(txsegs[0].ds_len);
 	if (prod == SGE_TX_RING_CNT - 1)
 		desc->sge_flags |= htole32(RING_END);
+	/* Configure VLAN. */
+	if(((*m_head)->m_flags & M_VLANTAG) != 0) {
+		cflags |= (*m_head)->m_pkthdr.ether_vtag;
+		desc->sge_sts_size |= htole32(TDS_INS_VLAN);
+	}
 	desc->sge_cmdsts = htole32(TDC_DEF | TDC_CRC | TDC_PAD | cflags);
 #if 1
 	if ((sc->sge_flags & SGE_FLAG_SPEED_1000) != 0)
@@ -1563,6 +1594,7 @@ sge_init_locked(struct sge_softc *sc)
 		rxfilt |= RXMAC_STRIP_FCS | RXMAC_PAD_ENB;
 	CSR_WRITE_2(sc, RxMacControl, rxfilt);
 	sge_rxfilter(sc);
+	sge_setvlan(sc);
 
 	/* Initialize default speed/duplex information. */
 	if ((sc->sge_flags & SGE_FLAG_FASTETHER) == 0)
@@ -1653,7 +1685,7 @@ sge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 	struct sge_softc *sc;
 	struct ifreq *ifr;
 	struct mii_data *mii;
-	int error = 0, mask;
+	int error = 0, mask, reinit;
 
 	sc = ifp->if_softc;
 	ifr = (struct ifreq *)data;
@@ -1675,6 +1707,7 @@ sge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 		break;
 	case SIOCSIFCAP:
 		SGE_LOCK(sc);
+		reinit = 0;
 		mask = ifr->ifr_reqcap ^ ifp->if_capenable;
 		if ((mask & IFCAP_TXCSUM) != 0 &&
 		    (ifp->if_capabilities & IFCAP_TXCSUM) != 0) {
@@ -1687,7 +1720,24 @@ sge_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 		if ((mask & IFCAP_RXCSUM) != 0 &&
 		    (ifp->if_capabilities & IFCAP_RXCSUM) != 0)
 			ifp->if_capenable ^= IFCAP_RXCSUM;
+		if ((mask & IFCAP_VLAN_HWCSUM) != 0 &&
+		    (ifp->if_capabilities & IFCAP_VLAN_HWCSUM) != 0)
+			ifp->if_capenable ^= IFCAP_VLAN_HWCSUM;
+		if ((mask & IFCAP_VLAN_HWTAGGING) != 0 &&
+		    (ifp->if_capabilities & IFCAP_VLAN_HWTAGGING) != 0) {
+			/*
+			 * Due to unknown reason, toggling VLAN hardware
+			 * tagging require interface reinitialization.
+			 */
+			ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
+			reinit = 1;
+		}
+		if (reinit > 0 && (ifp->if_drv_flags & IFF_DRV_RUNNING) != 0) {
+			ifp->if_drv_flags &= ~IFF_DRV_RUNNING;
+			sge_init_locked(sc);
+		}
 		SGE_UNLOCK(sc);
+		VLAN_CAPABILITIES(ifp);
 		break;
 	case SIOCADDMULTI:
 	case SIOCDELMULTI:
diff --git a/sys/dev/sge/if_sgereg.h b/sys/dev/sge/if_sgereg.h
index e8cf5140b6e..722dd3c490d 100644
--- a/sys/dev/sge/if_sgereg.h
+++ b/sys/dev/sge/if_sgereg.h
@@ -137,6 +137,7 @@
 #define	AcceptAllPhys		0x0100
 #define	AcceptErr		0x0020
 #define	AcceptRunt		0x0010
+#define	RXMAC_STRIP_VLAN	0x0020
 #define	RXMAC_STRIP_FCS		0x0010
 #define	RXMAC_PAD_ENB		0x0004
 
@@ -187,12 +188,14 @@
 #define	TDC_COL			0x00040000
 #define	TDC_CRC			0x00020000
 #define	TDC_PAD			0x00010000
+#define	TDC_VLAN_MASK		0x0000FFFF
 
 #define	SGE_TX_INTR_FRAMES	32
 
 /*
  * TX descriptor status bits.
  */
+#define	TDS_INS_VLAN		0x80000000
 #define	TDS_OWC			0x00080000
 #define	TDS_ABT			0x00040000
 #define	TDS_FIFO		0x00020000
@@ -219,11 +222,12 @@
 #define	RDC_UCAST		0x00040000
 #define	RDC_CRCOFF		0x00020000
 #define	RDC_PREADD		0x00010000
+#define	RDC_VLAN_MASK		0x0000FFFF
 
 /*
  * RX descriptor status bits
  */
-#define	RDS_TAGON		0x80000000
+#define	RDS_VLAN		0x80000000
 #define	RDS_DESCS		0x3f000000
 #define	RDS_ABORT		0x00800000
 #define	RDS_SHORT		0x00400000
@@ -240,7 +244,7 @@
 #define	RX_ERR_BITS 		"\20"					\
 				"\21CRCOK\22COLON\23NIBON\24OVRUN"	\
 				"\25MIIER\26LIMIT\27SHORT\30ABORT"	\
-				"\40TAGON"
+				"\40VLAN"
 
 #define	RING_END		0x80000000
 #define	SGE_RX_BYTES(x)		((x) & 0xFFFF)

From 53ac9f49f061d77781b923577371ee9af2b0b822 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Thu, 29 Apr 2010 18:21:27 +0000
Subject: [PATCH 336/532] Now sge(4) supports hardware VLAN tag
 insertion/stripping.

---
 share/man/man4/sge.4  | 10 +++++-----
 share/man/man4/vlan.4 |  2 +-
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/share/man/man4/sge.4 b/share/man/man4/sge.4
index a3d057d386c..9ef179c96ff 100644
--- a/share/man/man4/sge.4
+++ b/share/man/man4/sge.4
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd April 14, 2010
+.Dd April 29, 2010
 .Dt SGE 4
 .Os
 .Sh NAME
@@ -53,11 +53,11 @@ controllers and SiS191 Fast/Gigabit Ethernet controllers.
 .Pp
 All LOMs supported by the
 .Nm
-driver have TCP/UDP/IP checksum offload for transmit and receive.
+driver have TCP/UDP/IP checksum offload for transmit and receive,
+hardware VLAN tag stripping/insertion features.
 Due to lack of documentation more offloading features like TCP
-segmentation offload (TSO), hardware VLAN tag stripping/insertion
-features, Wake On Lan (WOL), Jumbo frame and an interrupt moderation
-mechanism are not supported yet.
+segmentation offload (TSO), Wake On Lan (WOL), Jumbo frame and an
+interrupt moderation mechanism are not supported yet.
 .Pp
 The
 .Nm
diff --git a/share/man/man4/vlan.4 b/share/man/man4/vlan.4
index f7f7807a86d..17fff6a3dc5 100644
--- a/share/man/man4/vlan.4
+++ b/share/man/man4/vlan.4
@@ -134,6 +134,7 @@ in the hardware is limited to the following devices:
 .Xr msk 4 ,
 .Xr nge 4 ,
 .Xr re 4 ,
+.Xr sge 4 ,
 .Xr stge 4 ,
 .Xr ti 4 ,
 .Xr txp 4 ,
@@ -172,7 +173,6 @@ natively:
 .Xr nve 4 ,
 .Xr rl 4 ,
 .Xr sf 4 ,
-.Xr sge 4 ,
 .Xr sis 4 ,
 .Xr sk 4 ,
 .Xr ste 4 ,

From 976d54c3f703f37f313a8cdebf1808caa46050a1 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 29 Apr 2010 18:32:14 +0000
Subject: [PATCH 337/532] bwn(4) will first appear in FreeBSD 8.1-RELEASE.

MFC after:	1 week
---
 share/man/man4/bwn.4 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/share/man/man4/bwn.4 b/share/man/man4/bwn.4
index 1293c36ca1f..b75f9d03bab 100644
--- a/share/man/man4/bwn.4
+++ b/share/man/man4/bwn.4
@@ -123,7 +123,7 @@ ifconfig wlan create wlandev bwn0 ssid my_net \e
 The
 .Nm
 driver first appeared in
-.Fx 8.0 .
+.Fx 8.1 .
 .Sh AUTHORS
 .An -nosplit
 The

From 8bbe8f73937101fb4e966deea8a55f37707e0667 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 29 Apr 2010 18:37:56 +0000
Subject: [PATCH 338/532] Add FreeBSD 8.1 to known list as it's being
 referenced by a couple of manpages already.

MFC after:	3 days
---
 gnu/usr.bin/groff/tmac/mdoc.local | 1 +
 1 file changed, 1 insertion(+)

diff --git a/gnu/usr.bin/groff/tmac/mdoc.local b/gnu/usr.bin/groff/tmac/mdoc.local
index 4c83635c503..a3073ff7d8d 100644
--- a/gnu/usr.bin/groff/tmac/mdoc.local
+++ b/gnu/usr.bin/groff/tmac/mdoc.local
@@ -75,6 +75,7 @@
 .ds doc-operating-system-FreeBSD-7.2    7.2
 .ds doc-operating-system-FreeBSD-7.3    7.3
 .ds doc-operating-system-FreeBSD-8.0    8.0
+.ds doc-operating-system-FreeBSD-8.1    8.1
 .ds doc-operating-system-FreeBSD-9.0    9.0
 .
 .\" Definitions not (yet) in doc-syms

From d92714eea8a0734e23e5974815c183a9cbdf30ad Mon Sep 17 00:00:00 2001
From: Pawel Jakub Dawidek <pjd@FreeBSD.org>
Date: Thu, 29 Apr 2010 21:55:20 +0000
Subject: [PATCH 339/532] Default connection timeout is way too long. To make
 it shorter we have to make socket non-blocking, connect() and if we get
 EINPROGRESS, we have to wait using select(). Very complex, but I know no
 other way to define connection timeout for a given socket.

Reported by:	hiroshi@soupacific.com
MFC after:	3 days
---
 sbin/hastd/proto_tcp4.c | 78 +++++++++++++++++++++++++++++++++++++++--
 1 file changed, 75 insertions(+), 3 deletions(-)

diff --git a/sbin/hastd/proto_tcp4.c b/sbin/hastd/proto_tcp4.c
index 2fba9967e40..5af82d53185 100644
--- a/sbin/hastd/proto_tcp4.c
+++ b/sbin/hastd/proto_tcp4.c
@@ -37,6 +37,7 @@ __FBSDID("$FreeBSD$");
 
 #include <assert.h>
 #include <errno.h>
+#include <fcntl.h>
 #include <netdb.h>
 #include <stdbool.h>
 #include <stdint.h>
@@ -47,6 +48,7 @@ __FBSDID("$FreeBSD$");
 #include "hast.h"
 #include "pjdlog.h"
 #include "proto_impl.h"
+#include "subr.h"
 
 #define	TCP4_CTX_MAGIC	0x7c441c
 struct tcp4_ctx {
@@ -222,18 +224,88 @@ static int
 tcp4_connect(void *ctx)
 {
 	struct tcp4_ctx *tctx = ctx;
+	struct timeval tv;
+	fd_set fdset;
+	socklen_t esize;
+	int error, flags, ret;
 
 	assert(tctx != NULL);
 	assert(tctx->tc_magic == TCP4_CTX_MAGIC);
 	assert(tctx->tc_side == TCP4_SIDE_CLIENT);
 	assert(tctx->tc_fd >= 0);
 
-	if (connect(tctx->tc_fd, (struct sockaddr *)&tctx->tc_sin,
-	    sizeof(tctx->tc_sin)) < 0) {
+	flags = fcntl(tctx->tc_fd, F_GETFL);
+	if (flags == -1) {
+		KEEP_ERRNO(pjdlog_common(LOG_DEBUG, 1, errno,
+		    "fcntl(F_GETFL) failed"));
+		return (errno);
+	}
+	/*
+	 * We make socket non-blocking so we have decided about connection
+	 * timeout.
+	 */
+	flags |= O_NONBLOCK;
+	if (fcntl(tctx->tc_fd, F_SETFL, flags) == -1) {
+		KEEP_ERRNO(pjdlog_common(LOG_DEBUG, 1, errno,
+		    "fcntl(F_SETFL, O_NONBLOCK) failed"));
 		return (errno);
 	}
 
-	return (0);
+	if (connect(tctx->tc_fd, (struct sockaddr *)&tctx->tc_sin,
+	    sizeof(tctx->tc_sin)) == 0) {
+		error = 0;
+		goto done;
+	}
+	if (errno != EINPROGRESS) {
+		error = errno;
+		pjdlog_common(LOG_DEBUG, 1, errno, "connect() failed");
+		goto done;
+	}
+	/*
+	 * Connection can't be established immediately, let's wait
+	 * for HAST_TIMEOUT seconds.
+	 */
+	tv.tv_sec = HAST_TIMEOUT;
+	tv.tv_usec = 0;
+again:
+	FD_ZERO(&fdset);
+	FD_SET(tctx->tc_fd, &fdset); 
+	ret = select(tctx->tc_fd + 1, NULL, &fdset, NULL, &tv);
+	if (ret == 0) {
+		error = ETIMEDOUT;
+		goto done;
+	} else if (ret == -1) {
+		if (errno == EINTR)
+			goto again;
+		error = errno;
+		pjdlog_common(LOG_DEBUG, 1, errno, "select() failed");
+		goto done;
+	}
+	assert(ret > 0);
+	assert(FD_ISSET(tctx->tc_fd, &fdset));
+	esize = sizeof(error);
+	if (getsockopt(tctx->tc_fd, SOL_SOCKET, SO_ERROR, &error,
+	    &esize) == -1) {
+		error = errno;
+		pjdlog_common(LOG_DEBUG, 1, errno,
+		    "getsockopt(SO_ERROR) failed");
+		goto done;
+	}
+	if (error != 0) {
+		pjdlog_common(LOG_DEBUG, 1, error,
+		    "getsockopt(SO_ERROR) returned error");
+		goto done;
+	}
+	error = 0;
+done:
+	flags &= ~O_NONBLOCK;
+	if (fcntl(tctx->tc_fd, F_SETFL, flags) == -1) {
+		if (error == 0)
+			error = errno;
+		pjdlog_common(LOG_DEBUG, 1, errno,
+		    "fcntl(F_SETFL, ~O_NONBLOCK) failed");
+	}
+	return (error);
 }
 
 static int

From ae575eab2c55356fceab8749cd4726c432bba62a Mon Sep 17 00:00:00 2001
From: David Christensen <davidch@FreeBSD.org>
Date: Thu, 29 Apr 2010 22:00:57 +0000
Subject: [PATCH 340/532] - Pass flow control settings back to bce(4).

MFC after:	Two weeks
---
 sys/dev/mii/brgphy.c | 10 ++++------
 1 file changed, 4 insertions(+), 6 deletions(-)

diff --git a/sys/dev/mii/brgphy.c b/sys/dev/mii/brgphy.c
index 97a423cd27e..a733597fe76 100644
--- a/sys/dev/mii/brgphy.c
+++ b/sys/dev/mii/brgphy.c
@@ -685,16 +685,15 @@ brgphy_status(struct mii_softc *sc)
 
 	}
 
-#if 0
-	/* Todo: Change bge/bce to use these settings. */
+	/* Todo: Change bge to use these settings. */
 
-	/* Fetch flow control settings from the PHY */
+	/* Fetch flow control settings from the copper PHY. */
 	if ((sc->mii_flags & MIIF_HAVEFIBER) == 0) {
-		/* Set FLAG0 is RX is enabled and FLAG1 if TX is enabled */
+		/* Set FLAG0 if RX is enabled and FLAG1 if TX is enabled */
 		if ((anar & BRGPHY_ANAR_PC) && (anlpar & BRGPHY_ANLPAR_PC)) {
 			mii->mii_media_active |= IFM_FLAG0 | IFM_FLAG1;
 		} else if (!(anar & BRGPHY_ANAR_PC) && (anlpar & BRGPHY_ANAR_ASP) &&
-		    (anlpar & BRPHY_ANLPAR_PC) && (anlpar & BRGPHY_ANLPAR_ASP)) {
+		    (anlpar & BRGPHY_ANLPAR_PC) && (anlpar & BRGPHY_ANLPAR_ASP)) {
 			mii->mii_media_active |= IFM_FLAG1;
 		} else if ((anar & BRGPHY_ANAR_PC) && (anar & BRGPHY_ANAR_ASP) &&
 		    !(anlpar & BRGPHY_ANLPAR_PC) && (anlpar & BRGPHY_ANLPAR_ASP)) {
@@ -703,7 +702,6 @@ brgphy_status(struct mii_softc *sc)
 	}
 
 	/* Todo: Add support for fiber settings too. */
-#endif
 
 
 brgphy_status_exit:

From 8cfa1186ed9da66025e648084b294b149d6d84d3 Mon Sep 17 00:00:00 2001
From: Andrew Thompson <thompsa@FreeBSD.org>
Date: Thu, 29 Apr 2010 22:51:18 +0000
Subject: [PATCH 341/532] usb_controller.c requires opt_ddb.h

---
 sys/modules/usb/usb/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/modules/usb/usb/Makefile b/sys/modules/usb/usb/Makefile
index 5b0104e82ae..32a21075005 100644
--- a/sys/modules/usb/usb/Makefile
+++ b/sys/modules/usb/usb/Makefile
@@ -31,7 +31,7 @@ S=     ${.CURDIR}/../../..
 
 KMOD=	usb
 SRCS=	bus_if.h device_if.h usb_if.h usb_if.c vnode_if.h \
-	opt_usb.h opt_bus.h \
+	opt_usb.h opt_bus.h opt_ddb.h \
 	usbdevs.h usbdevs_data.h \
 	usb_busdma.c usb_controller.c usb_compat_linux.c usb_core.c usb_debug.c \
 	usb_dev.c usb_device.c usb_dynamic.c usb_error.c usb_generic.c \

From 106b2e2ff2025f21388ba446ce5af0fe944db767 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 30 Apr 2010 00:34:00 +0000
Subject: [PATCH 342/532] Both RX_GMF_LP_THR and RX_GMF_UP_THR must be 16 bits
 register. If it is 8bits register then RX FIFO size can't exceed 2KB which is
 not true for almost all Yukon II controller.

---
 sys/dev/msk/if_msk.c    | 4 ++--
 sys/dev/msk/if_mskreg.h | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index c083e1e10f6..9b924352668 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -3822,9 +3822,9 @@ msk_init_locked(struct msk_if_softc *sc_if)
 
 	if ((sc_if->msk_flags & MSK_FLAG_RAMBUF) == 0) {
 		/* Set Rx Pause threshould. */
-		CSR_WRITE_1(sc, MR_ADDR(sc_if->msk_port, RX_GMF_LP_THR),
+		CSR_WRITE_2(sc, MR_ADDR(sc_if->msk_port, RX_GMF_LP_THR),
 		    MSK_ECU_LLPP);
-		CSR_WRITE_1(sc, MR_ADDR(sc_if->msk_port, RX_GMF_UP_THR),
+		CSR_WRITE_2(sc, MR_ADDR(sc_if->msk_port, RX_GMF_UP_THR),
 		    MSK_ECU_ULPP);
 		/* Configure store-and-forward for Tx. */
 		msk_set_tx_stfwd(sc_if);
diff --git a/sys/dev/msk/if_mskreg.h b/sys/dev/msk/if_mskreg.h
index 34cac676305..293ca29f542 100644
--- a/sys/dev/msk/if_mskreg.h
+++ b/sys/dev/msk/if_mskreg.h
@@ -621,8 +621,8 @@
 #define RX_GMF_FL_MSK	0x0c4c	/* 32 bit Rx GMAC FIFO Flush Mask */
 #define RX_GMF_FL_THR	0x0c50	/* 32 bit Rx GMAC FIFO Flush Threshold */
 #define RX_GMF_TR_THR	0x0c54	/* 32 bit Rx Truncation Threshold (Yukon-2) */
-#define	RX_GMF_UP_THR	0x0c58	/*  8 bit Rx Upper Pause Thr (Yukon-EC_U) */
-#define	RX_GMF_LP_THR	0x0c5a	/*  8 bit Rx Lower Pause Thr (Yukon-EC_U) */
+#define	RX_GMF_UP_THR	0x0c58	/* 16 bit Rx Upper Pause Thr (Yukon-EC_U) */
+#define	RX_GMF_LP_THR	0x0c5a	/* 16 bit Rx Lower Pause Thr (Yukon-EC_U) */
 #define RX_GMF_VLAN	0x0c5c	/* 32 bit Rx VLAN Type Register (Yukon-2) */
 #define RX_GMF_WP	0x0c60	/* 32 bit Rx GMAC FIFO Write Pointer */
 #define RX_GMF_WLEV	0x0c68	/* 32 bit Rx GMAC FIFO Write Level */

From 2965a4531505c497a6bbc6f2974499a520c75394 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 00:46:43 +0000
Subject: [PATCH 343/532] On Alan's advice, rather than do a wholesale
 conversion on a single architecture from page queue lock to a hashed array of
 page locks (based on a patch by Jeff Roberson), I've implemented page lock
 support in the MI code and have only moved vm_page's hold_count out from
 under page queue mutex to page lock. This changes pmap_extract_and_hold on
 all pmaps.

Supported by: Bitgravity Inc.

Discussed with: alc, jeffr, and kib
---
 UPDATING                    |  8 ++++
 sys/amd64/amd64/pmap.c      | 17 ++++---
 sys/amd64/include/pmap.h    |  2 +
 sys/amd64/include/vmparam.h |  4 ++
 sys/arm/arm/pmap.c          | 13 +++---
 sys/arm/include/pmap.h      |  2 +
 sys/dev/drm/via_dmablit.c   |  2 +
 sys/i386/i386/pmap.c        | 11 ++++-
 sys/i386/include/pmap.h     |  3 ++
 sys/i386/xen/pmap.c         | 11 ++++-
 sys/ia64/ia64/pmap.c        |  8 +++-
 sys/ia64/include/pmap.h     |  2 +
 sys/kern/kern_exec.c        |  8 ++--
 sys/kern/subr_witness.c     |  9 ++++
 sys/kern/sys_pipe.c         | 12 ++---
 sys/kern/sys_process.c      |  8 ++--
 sys/kern/uipc_cow.c         |  4 +-
 sys/kern/vfs_bio.c          | 13 +++---
 sys/mips/include/pmap.h     |  2 +
 sys/mips/mips/pmap.c        | 11 +++--
 sys/net/bpf_zerocopy.c      |  2 +
 sys/powerpc/aim/mmu_oea.c   | 10 +++--
 sys/powerpc/aim/mmu_oea64.c |  9 +++-
 sys/powerpc/booke/pmap.c    | 11 +++--
 sys/powerpc/include/pmap.h  |  2 +
 sys/sparc64/include/pmap.h  |  2 +
 sys/sparc64/sparc64/pmap.c  | 14 +++---
 sys/sun4v/include/pmap.h    |  2 +
 sys/sun4v/sun4v/pmap.c      |  8 +++-
 sys/sys/param.h             |  2 +-
 sys/vm/device_pager.c       | 19 +++++---
 sys/vm/sg_pager.c           |  9 ++--
 sys/vm/swap_pager.c         | 28 +++++++++---
 sys/vm/uma_core.c           |  2 +
 sys/vm/vm_contig.c          |  2 +
 sys/vm/vm_fault.c           | 62 ++++++++++++++++++++++++--
 sys/vm/vm_glue.c            | 14 ++++--
 sys/vm/vm_kern.c            |  2 +
 sys/vm/vm_mmap.c            |  2 +
 sys/vm/vm_object.c          | 89 ++++++++++++++++++++++++++++++++-----
 sys/vm/vm_page.c            | 71 ++++++++++++++++++++++++++++-
 sys/vm/vm_page.h            | 29 +++++++++++-
 sys/vm/vm_pageout.c         | 86 ++++++++++++++++++++++++++++++-----
 sys/vm/vm_param.h           |  8 ++++
 sys/vm/vnode_pager.c        | 85 ++++++++++++++++++++++++-----------
 45 files changed, 587 insertions(+), 133 deletions(-)

diff --git a/UPDATING b/UPDATING
index 553e96185a8..937f9177e44 100644
--- a/UPDATING
+++ b/UPDATING
@@ -22,6 +22,14 @@ NOTE TO PEOPLE WHO THINK THAT FreeBSD 9.x IS SLOW:
 	machines to maximize performance.  (To disable malloc debugging, run
 	ln -s aj /etc/malloc.conf.)
 
+
+20100429:
+	'vm_page's are now hashed by physical address to an array of mutexes.
+	Currently this is only used to serialize access to hold_count. Over 
+	time the page queue mutex will be peeled away. This changes the size
+	of pmap on every architecture. And requires all callers of vm_page_hold
+	and vm_page_unhold to be updated. 
+ 
 20100402:
 	WITH_CTF can now be specified in src.conf (not recommended, there
 	are some problems with static executables), make.conf (would also
diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index d07292dd2c2..045b634a980 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -793,7 +793,6 @@ static u_long pmap_pdpe_demotions;
 SYSCTL_ULONG(_vm_pmap_pdpe, OID_AUTO, demotions, CTLFLAG_RD,
     &pmap_pdpe_demotions, 0, "1GB page demotions");
 
-
 /***************************************************
  * Low level helper routines.....
  ***************************************************/
@@ -1200,15 +1199,20 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 {
 	pd_entry_t pde, *pdep;
 	pt_entry_t pte;
+	vm_paddr_t pa;
 	vm_page_t m;
 
+	pa = 0;
 	m = NULL;
-	vm_page_lock_queues();
 	PMAP_LOCK(pmap);
+retry:
 	pdep = pmap_pde(pmap, va);
 	if (pdep != NULL && (pde = *pdep)) {
 		if (pde & PG_PS) {
 			if ((pde & PG_RW) || (prot & VM_PROT_WRITE) == 0) {
+				if (vm_page_pa_tryrelock(pmap, (pde & PG_PS_FRAME) |
+				       (va & PDRMASK), &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE((pde & PG_PS_FRAME) |
 				    (va & PDRMASK));
 				vm_page_hold(m);
@@ -1217,12 +1221,14 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 			pte = *pmap_pde_to_pte(pdep, va);
 			if ((pte & PG_V) &&
 			    ((pte & PG_RW) || (prot & VM_PROT_WRITE) == 0)) {
+				if (vm_page_pa_tryrelock(pmap, pte & PG_FRAME, &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE(pte & PG_FRAME);
 				vm_page_hold(m);
 			}
 		}
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
@@ -3143,9 +3149,8 @@ pmap_enter(pmap_t pmap, vm_offset_t va, vm_prot_t access, vm_page_t m,
 	 * In the case that a page table page is not
 	 * resident, we are creating it here.
 	 */
-	if (va < VM_MAXUSER_ADDRESS) {
+	if (va < VM_MAXUSER_ADDRESS)
 		mpte = pmap_allocpte(pmap, va, M_WAITOK);
-	}
 
 	pde = pmap_pde(pmap, va);
 	if (pde != NULL && (*pde & PG_V) != 0) {
@@ -3393,7 +3398,7 @@ pmap_enter_object(pmap_t pmap, vm_offset_t start, vm_offset_t end,
 			    mpte);
 		m = TAILQ_NEXT(m, listq);
 	}
- 	PMAP_UNLOCK(pmap);
+	PMAP_UNLOCK(pmap);
 }
 
 /*
diff --git a/sys/amd64/include/pmap.h b/sys/amd64/include/pmap.h
index 11a5628082f..6f0b188e9b6 100644
--- a/sys/amd64/include/pmap.h
+++ b/sys/amd64/include/pmap.h
@@ -245,6 +245,8 @@ struct pmap {
 	pml4_entry_t		*pm_pml4;	/* KVA of level 4 page table */
 	TAILQ_HEAD(,pv_chunk)	pm_pvchunk;	/* list of mappings in pmap */
 	u_int			pm_active;	/* active on cpus */
+	uint32_t		pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 	/* spare u_int here due to padding */
 	struct pmap_statistics	pm_stats;	/* pmap statistics */
 	vm_page_t		pm_root;	/* spare page table pages */
diff --git a/sys/amd64/include/vmparam.h b/sys/amd64/include/vmparam.h
index 0b5004c6818..6dbe3713767 100644
--- a/sys/amd64/include/vmparam.h
+++ b/sys/amd64/include/vmparam.h
@@ -145,6 +145,10 @@
 #define	VM_LEVEL_0_ORDER	9
 #endif
 
+#ifdef	SMP
+#define	PA_LOCK_COUNT	256
+#endif
+
 /*
  * Virtual addresses of things.  Derived from the page directory and
  * page table indexes from pmap.h for precision.
diff --git a/sys/arm/arm/pmap.c b/sys/arm/arm/pmap.c
index 366b43fe3e1..8fe6424a89a 100644
--- a/sys/arm/arm/pmap.c
+++ b/sys/arm/arm/pmap.c
@@ -3740,13 +3740,14 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 	struct l2_dtable *l2;
 	pd_entry_t l1pd;
 	pt_entry_t *ptep, pte;
-	vm_paddr_t pa;
+	vm_paddr_t pa, paddr;
 	vm_page_t m = NULL;
 	u_int l1idx;
 	l1idx = L1_IDX(va);
+	paddr = 0;
 
-	vm_page_lock_queues();
  	PMAP_LOCK(pmap);
+retry:
 	l1pd = pmap->pm_l1->l1_kva[l1idx];
 	if (l1pte_section_p(l1pd)) {
 		/*
@@ -3758,6 +3759,8 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 			pa = (l1pd & L1_SUP_FRAME) | (va & L1_SUP_OFFSET);
 		else
 			pa = (l1pd & L1_S_FRAME) | (va & L1_S_OFFSET);
+		if (vm_page_pa_tryrelock(pmap, pa & PG_FRAME, &paddr))
+			goto retry;
 		if (l1pd & L1_S_PROT_W || (prot & VM_PROT_WRITE) == 0) {
 			m = PHYS_TO_VM_PAGE(pa);
 			vm_page_hold(m);
@@ -3774,7 +3777,6 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 		if (l2 == NULL ||
 		    (ptep = l2->l2_bucket[L2_BUCKET(l1idx)].l2b_kva) == NULL) {
 		 	PMAP_UNLOCK(pmap);
-			vm_page_unlock_queues();
 			return (NULL);
 		}
 
@@ -3783,7 +3785,6 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 
 		if (pte == 0) {
 		 	PMAP_UNLOCK(pmap);
-			vm_page_unlock_queues();
 			return (NULL);
 		}
 		if (pte & L2_S_PROT_W || (prot & VM_PROT_WRITE) == 0) {
@@ -3796,13 +3797,15 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 				pa = (pte & L2_S_FRAME) | (va & L2_S_OFFSET);
 				break;
 			}
+			if (vm_page_pa_tryrelock(pmap, pa & PG_FRAME, &paddr))
+				goto retry;		
 			m = PHYS_TO_VM_PAGE(pa);
 			vm_page_hold(m);
 		}
 	}
 
  	PMAP_UNLOCK(pmap);
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(paddr);
 	return (m);
 }
 
diff --git a/sys/arm/include/pmap.h b/sys/arm/include/pmap.h
index 8ee7bac27e2..287c4c1206a 100644
--- a/sys/arm/include/pmap.h
+++ b/sys/arm/include/pmap.h
@@ -134,6 +134,8 @@ struct	pmap {
 	struct l1_ttable	*pm_l1;
 	struct l2_dtable	*pm_l2[L2_SIZE];
 	pd_entry_t		*pm_pdir;	/* KVA of page directory */
+	uint32_t		pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 	int			pm_active;	/* active on cpus */
 	struct pmap_statistics	pm_stats;	/* pmap statictics */
 	TAILQ_HEAD(,pv_entry)	pm_pvlist;	/* list of mappings in pmap */
diff --git a/sys/dev/drm/via_dmablit.c b/sys/dev/drm/via_dmablit.c
index ea449f1f992..3a66b261574 100644
--- a/sys/dev/drm/via_dmablit.c
+++ b/sys/dev/drm/via_dmablit.c
@@ -248,10 +248,12 @@ via_lock_all_dma_pages(drm_via_sg_info_t *vsg,  drm_via_dmablit_t *xfer)
 		    (vm_offset_t)xfer->mem_addr + IDX_TO_OFF(i), VM_PROT_RW);
 		if (m == NULL)
 			break;
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_wire(m);
 		vm_page_unhold(m);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		vsg->pages[i] = m;
 	}
 	vsg->state = dr_via_pages_locked;
diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c
index d8b9686e738..661fade94e3 100644
--- a/sys/i386/i386/pmap.c
+++ b/sys/i386/i386/pmap.c
@@ -1346,14 +1346,19 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 	pd_entry_t pde;
 	pt_entry_t pte;
 	vm_page_t m;
+	vm_paddr_t pa;
 
+	pa = 0;
 	m = NULL;
-	vm_page_lock_queues();
 	PMAP_LOCK(pmap);
+retry:
 	pde = *pmap_pde(pmap, va);
 	if (pde != 0) {
 		if (pde & PG_PS) {
 			if ((pde & PG_RW) || (prot & VM_PROT_WRITE) == 0) {
+				if (vm_page_pa_tryrelock(pmap, (pde & PG_PS_FRAME) |
+				       (va & PDRMASK), &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE((pde & PG_PS_FRAME) |
 				    (va & PDRMASK));
 				vm_page_hold(m);
@@ -1363,13 +1368,15 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 			pte = *pmap_pte_quick(pmap, va);
 			if (pte != 0 &&
 			    ((pte & PG_RW) || (prot & VM_PROT_WRITE) == 0)) {
+				if (vm_page_pa_tryrelock(pmap, pte & PG_FRAME, &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE(pte & PG_FRAME);
 				vm_page_hold(m);
 			}
 			sched_unpin();
 		}
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/i386/include/pmap.h b/sys/i386/include/pmap.h
index ae7d79d83fb..f45a9dfd679 100644
--- a/sys/i386/include/pmap.h
+++ b/sys/i386/include/pmap.h
@@ -420,11 +420,14 @@ struct pmap {
 	u_int			pm_active;	/* active on cpus */
 	struct pmap_statistics	pm_stats;	/* pmap statistics */
 	LIST_ENTRY(pmap) 	pm_list;	/* List of all pmaps */
+	uint32_t		pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 #ifdef PAE
 	pdpt_entry_t		*pm_pdpt;	/* KVA of page director pointer
 						   table */
 #endif
 	vm_page_t		pm_root;	/* spare page table pages */
+
 };
 
 typedef struct pmap	*pmap_t;
diff --git a/sys/i386/xen/pmap.c b/sys/i386/xen/pmap.c
index 18619c3b0bd..ec96ef0a41b 100644
--- a/sys/i386/xen/pmap.c
+++ b/sys/i386/xen/pmap.c
@@ -1219,14 +1219,19 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 	pd_entry_t pde;
 	pt_entry_t pte;
 	vm_page_t m;
+	vm_paddr_t pa;
 
+	pa = 0;
 	m = NULL;
-	vm_page_lock_queues();
 	PMAP_LOCK(pmap);
+retry:
 	pde = PT_GET(pmap_pde(pmap, va));
 	if (pde != 0) {
 		if (pde & PG_PS) {
 			if ((pde & PG_RW) || (prot & VM_PROT_WRITE) == 0) {
+				if (vm_page_pa_tryrelock(pmap, (pde & PG_PS_FRAME) |
+				       (va & PDRMASK), &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE((pde & PG_PS_FRAME) |
 				    (va & PDRMASK));
 				vm_page_hold(m);
@@ -1238,13 +1243,15 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 				PT_SET_MA(PADDR1, 0);
 			if ((pte & PG_V) &&
 			    ((pte & PG_RW) || (prot & VM_PROT_WRITE) == 0)) {
+				if (vm_page_pa_tryrelock(pmap, pte & PG_FRAME, &pa))
+					goto retry;
 				m = PHYS_TO_VM_PAGE(pte & PG_FRAME);
 				vm_page_hold(m);
 			}
 			sched_unpin();
 		}
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/ia64/ia64/pmap.c b/sys/ia64/ia64/pmap.c
index 91b2e07fdc4..7cc18c15289 100644
--- a/sys/ia64/ia64/pmap.c
+++ b/sys/ia64/ia64/pmap.c
@@ -1028,18 +1028,22 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 	struct ia64_lpte *pte;
 	pmap_t oldpmap;
 	vm_page_t m;
+	vm_paddr_t pa;
 
+	pa = 0;
 	m = NULL;
-	vm_page_lock_queues();
 	PMAP_LOCK(pmap);
 	oldpmap = pmap_switch(pmap);
+retry:
 	pte = pmap_find_vhpt(va);
 	if (pte != NULL && pmap_present(pte) &&
 	    (pmap_prot(pte) & prot) == prot) {
 		m = PHYS_TO_VM_PAGE(pmap_ppn(pte));
+		if (vm_page_pa_tryrelock(pmap, pmap_ppn(pte), &pa))
+			goto retry;
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	pmap_switch(oldpmap);
 	PMAP_UNLOCK(pmap);
 	return (m);
diff --git a/sys/ia64/include/pmap.h b/sys/ia64/include/pmap.h
index 44079c88129..ff059fd5563 100644
--- a/sys/ia64/include/pmap.h
+++ b/sys/ia64/include/pmap.h
@@ -77,6 +77,8 @@ struct pmap {
 	TAILQ_HEAD(,pv_entry)	pm_pvlist;	/* list of mappings in pmap */
 	u_int32_t		pm_rid[5];	/* base RID for pmap */
 	struct pmap_statistics	pm_stats;	/* pmap statistics */
+	uint32_t		pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 };
 
 typedef struct pmap	*pmap_t;
diff --git a/sys/kern/kern_exec.c b/sys/kern/kern_exec.c
index ed22519f693..bb92972fbe6 100644
--- a/sys/kern/kern_exec.c
+++ b/sys/kern/kern_exec.c
@@ -957,9 +957,9 @@ exec_map_first_page(imgp)
 			return (EIO);
 		}
 	}
-	vm_page_lock_queues();
+	vm_page_lock(ma[0]);
 	vm_page_hold(ma[0]);
-	vm_page_unlock_queues();
+	vm_page_unlock(ma[0]);
 	vm_page_wakeup(ma[0]);
 	VM_OBJECT_UNLOCK(object);
 
@@ -979,9 +979,9 @@ exec_unmap_first_page(imgp)
 		m = sf_buf_page(imgp->firstpage);
 		sf_buf_free(imgp->firstpage);
 		imgp->firstpage = NULL;
-		vm_page_lock_queues();
+		vm_page_lock(m);
 		vm_page_unhold(m);
-		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
 }
 
diff --git a/sys/kern/subr_witness.c b/sys/kern/subr_witness.c
index 5b7d5651750..ef1bc398dfd 100644
--- a/sys/kern/subr_witness.c
+++ b/sys/kern/subr_witness.c
@@ -596,6 +596,15 @@ static struct witness_order_list_entry order_lists[] = {
 	{ "vnode interlock", &lock_class_mtx_sleep },
 	{ "cdev", &lock_class_mtx_sleep },
 	{ NULL, NULL },
+	/*
+	 * VM
+	 * 
+	 */
+	{ "vm object", &lock_class_mtx_sleep },
+	{ "page lock", &lock_class_mtx_sleep },
+	{ "vm page queue mutex", &lock_class_mtx_sleep },
+	{ "pmap", &lock_class_mtx_sleep },
+	{ NULL, NULL },
 	/*
 	 * kqueue/VFS interaction
 	 */
diff --git a/sys/kern/sys_pipe.c b/sys/kern/sys_pipe.c
index e098648a669..7130c263e48 100644
--- a/sys/kern/sys_pipe.c
+++ b/sys/kern/sys_pipe.c
@@ -773,10 +773,12 @@ pipe_build_write_buffer(wpipe, uio)
 		 */
 	race:
 		if (vm_fault_quick((caddr_t)addr, VM_PROT_READ) < 0) {
-			vm_page_lock_queues();
-			for (j = 0; j < i; j++)
+			
+			for (j = 0; j < i; j++) {
+				vm_page_lock(wpipe->pipe_map.ms[j]);
 				vm_page_unhold(wpipe->pipe_map.ms[j]);
-			vm_page_unlock_queues();
+				vm_page_unlock(wpipe->pipe_map.ms[j]);
+			}
 			return (EFAULT);
 		}
 		wpipe->pipe_map.ms[i] = pmap_extract_and_hold(pmap, addr,
@@ -816,11 +818,11 @@ pipe_destroy_write_buffer(wpipe)
 	int i;
 
 	PIPE_LOCK_ASSERT(wpipe, MA_OWNED);
-	vm_page_lock_queues();
 	for (i = 0; i < wpipe->pipe_map.npages; i++) {
+		vm_page_lock(wpipe->pipe_map.ms[i]);
 		vm_page_unhold(wpipe->pipe_map.ms[i]);
+		vm_page_unlock(wpipe->pipe_map.ms[i]);
 	}
-	vm_page_unlock_queues();
 	wpipe->pipe_map.npages = 0;
 }
 
diff --git a/sys/kern/sys_process.c b/sys/kern/sys_process.c
index f5671d9c608..d8cc4f01d2d 100644
--- a/sys/kern/sys_process.c
+++ b/sys/kern/sys_process.c
@@ -328,9 +328,9 @@ proc_rwmem(struct proc *p, struct uio *uio)
 		/*
 		 * Hold the page in memory.
 		 */
-		vm_page_lock_queues();
+		vm_page_lock(m);
 		vm_page_hold(m);
-		vm_page_unlock_queues();
+		vm_page_unlock(m);
 
 		/*
 		 * We're done with tmap now.
@@ -349,9 +349,9 @@ proc_rwmem(struct proc *p, struct uio *uio)
 		/*
 		 * Release the page.
 		 */
-		vm_page_lock_queues();
+		vm_page_lock(m);
 		vm_page_unhold(m);
-		vm_page_unlock_queues();
+		vm_page_unlock(m);
 
 	} while (error == 0 && uio->uio_resid > 0);
 
diff --git a/sys/kern/uipc_cow.c b/sys/kern/uipc_cow.c
index 52988dd7e79..5c0dcd230ee 100644
--- a/sys/kern/uipc_cow.c
+++ b/sys/kern/uipc_cow.c
@@ -128,10 +128,12 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	/* 
 	 * set up COW
 	 */
+	vm_page_lock(pp);
 	vm_page_lock_queues();
 	if (vm_page_cowsetup(pp) != 0) {
 		vm_page_unhold(pp);
 		vm_page_unlock_queues();
+		vm_page_unlock(pp);
 		return (0);
 	}
 
@@ -141,7 +143,7 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	vm_page_wire(pp);
 	vm_page_unhold(pp);
 	vm_page_unlock_queues();
-
+	vm_page_unlock(pp);
 	/*
 	 * Allocate an sf buf
 	 */
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 156b676cf6b..ea846a5bf5d 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -3860,12 +3860,12 @@ vmapbuf(struct buf *bp)
 retry:
 		if (vm_fault_quick(addr >= bp->b_data ? addr : bp->b_data,
 		    prot) < 0) {
-			vm_page_lock_queues();
 			for (i = 0; i < pidx; ++i) {
+				vm_page_lock(bp->b_pages[i]);
 				vm_page_unhold(bp->b_pages[i]);
+				vm_page_unlock(bp->b_pages[i]);
 				bp->b_pages[i] = NULL;
 			}
-			vm_page_unlock_queues();
 			return(-1);
 		}
 		m = pmap_extract_and_hold(pmap, (vm_offset_t)addr, prot);
@@ -3896,11 +3896,12 @@ vunmapbuf(struct buf *bp)
 
 	npages = bp->b_npages;
 	pmap_qremove(trunc_page((vm_offset_t)bp->b_data), npages);
-	vm_page_lock_queues();
-	for (pidx = 0; pidx < npages; pidx++)
+	for (pidx = 0; pidx < npages; pidx++) {
+		vm_page_lock(bp->b_pages[pidx]);
 		vm_page_unhold(bp->b_pages[pidx]);
-	vm_page_unlock_queues();
-
+		vm_page_unlock(bp->b_pages[pidx]);
+	}
+	
 	bp->b_data = bp->b_saveaddr;
 }
 
diff --git a/sys/mips/include/pmap.h b/sys/mips/include/pmap.h
index 80772d97ee9..f4df6cae01c 100644
--- a/sys/mips/include/pmap.h
+++ b/sys/mips/include/pmap.h
@@ -88,6 +88,8 @@ struct pmap {
 	pd_entry_t *pm_segtab;	/* KVA of segment table */
 	TAILQ_HEAD(, pv_entry) pm_pvlist;	/* list of mappings in
 						 * pmap */
+	uint32_t	pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int		pm_retries;
 	int pm_active;		/* active on cpus */
 	struct {
 		u_int32_t asid:ASID_BITS;	/* TLB address space tag */
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index e7c323938dc..826177fb70a 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -147,7 +147,6 @@ unsigned pmap_max_asid;		/* max ASID supported by the system */
 
 #define	PMAP_ASID_RESERVED	0
 
-
 vm_offset_t kernel_vm_end;
 
 static struct tlb tlbstash[MAXCPU][MIPS_MAX_TLB_ENTRIES];
@@ -710,18 +709,22 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 {
 	pt_entry_t pte;
 	vm_page_t m;
+	vm_paddr_t pa;
 
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;
 	PMAP_LOCK(pmap);
-
+retry:
 	pte = *pmap_pte(pmap, va);
 	if (pte != 0 && pmap_pte_v(&pte) &&
 	    ((pte & PTE_RW) || (prot & VM_PROT_WRITE) == 0)) {
+		if (vm_page_pa_tryrelock(pmap, mips_tlbpfn_to_paddr(pte), &pa))
+			goto retry;
+
 		m = PHYS_TO_VM_PAGE(mips_tlbpfn_to_paddr(pte));
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/net/bpf_zerocopy.c b/sys/net/bpf_zerocopy.c
index a1dd923a4f5..3cf4b8d9a51 100644
--- a/sys/net/bpf_zerocopy.c
+++ b/sys/net/bpf_zerocopy.c
@@ -168,10 +168,12 @@ zbuf_sfbuf_get(struct vm_map *map, vm_offset_t uaddr)
 	    VM_PROT_WRITE);
 	if (pp == NULL)
 		return (NULL);
+	vm_page_lock(pp);
 	vm_page_lock_queues();
 	vm_page_wire(pp);
 	vm_page_unhold(pp);
 	vm_page_unlock_queues();
+	vm_page_unlock(pp);
 	sf = sf_buf_alloc(pp, SFB_NOWAIT);
 	if (sf == NULL) {
 		zbuf_page_free(pp);
diff --git a/sys/powerpc/aim/mmu_oea.c b/sys/powerpc/aim/mmu_oea.c
index 24e7b4e1d40..95936ed01b1 100644
--- a/sys/powerpc/aim/mmu_oea.c
+++ b/sys/powerpc/aim/mmu_oea.c
@@ -1241,18 +1241,22 @@ moea_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 {
 	struct	pvo_entry *pvo;
 	vm_page_t m;
-        
+        vm_paddr_t pa;
+
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;
 	PMAP_LOCK(pmap);
+retry:
 	pvo = moea_pvo_find_va(pmap, va & ~ADDR_POFF, NULL);
 	if (pvo != NULL && (pvo->pvo_pte.pte.pte_hi & PTE_VALID) &&
 	    ((pvo->pvo_pte.pte.pte_lo & PTE_PP) == PTE_RW ||
 	     (prot & VM_PROT_WRITE) == 0)) {
+		if (vm_page_pa_tryrelock(pmap, pvo->pvo_pte.pte.pte_lo & PTE_RPGN, &pa))
+			goto retry;
 		m = PHYS_TO_VM_PAGE(pvo->pvo_pte.pte.pte_lo & PTE_RPGN);
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/powerpc/aim/mmu_oea64.c b/sys/powerpc/aim/mmu_oea64.c
index 2571587733a..0483b41c4ba 100644
--- a/sys/powerpc/aim/mmu_oea64.c
+++ b/sys/powerpc/aim/mmu_oea64.c
@@ -1374,18 +1374,23 @@ moea64_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 {
 	struct	pvo_entry *pvo;
 	vm_page_t m;
+        vm_paddr_t pa;
         
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;
 	PMAP_LOCK(pmap);
+retry:
 	pvo = moea64_pvo_find_va(pmap, va & ~ADDR_POFF, NULL);
 	if (pvo != NULL && (pvo->pvo_pte.lpte.pte_hi & LPTE_VALID) &&
 	    ((pvo->pvo_pte.lpte.pte_lo & LPTE_PP) == LPTE_RW ||
 	     (prot & VM_PROT_WRITE) == 0)) {
+		if (vm_page_pa_tryrelock(pmap,
+			pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN, &pa))
+			goto retry;
 		m = PHYS_TO_VM_PAGE(pvo->pvo_pte.lpte.pte_lo & LPTE_RPGN);
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/powerpc/booke/pmap.c b/sys/powerpc/booke/pmap.c
index 549eaaa94bb..0b0fce4485e 100644
--- a/sys/powerpc/booke/pmap.c
+++ b/sys/powerpc/booke/pmap.c
@@ -2034,11 +2034,12 @@ mmu_booke_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va,
 	pte_t *pte;
 	vm_page_t m;
 	uint32_t pte_wbit;
-
+	vm_paddr_t pa;
+	
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;	
 	PMAP_LOCK(pmap);
-
+retry:
 	pte = pte_find(mmu, pmap, va);
 	if ((pte != NULL) && PTE_ISVALID(pte)) {
 		if (pmap == kernel_pmap)
@@ -2047,12 +2048,14 @@ mmu_booke_extract_and_hold(mmu_t mmu, pmap_t pmap, vm_offset_t va,
 			pte_wbit = PTE_UW;
 
 		if ((pte->flags & pte_wbit) || ((prot & VM_PROT_WRITE) == 0)) {
+			if (vm_page_pa_tryrelock(pmap, PTE_PA(pte), &pa))
+				goto retry;
 			m = PHYS_TO_VM_PAGE(PTE_PA(pte));
 			vm_page_hold(m);
 		}
 	}
 
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 	return (m);
 }
diff --git a/sys/powerpc/include/pmap.h b/sys/powerpc/include/pmap.h
index a23052ea975..a49bd6ce8ae 100644
--- a/sys/powerpc/include/pmap.h
+++ b/sys/powerpc/include/pmap.h
@@ -88,6 +88,8 @@ struct	pmap {
 	struct	mtx	pm_mtx;
 	u_int		pm_sr[16];
 	u_int		pm_active;
+	uint32_t	pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int		pm_retries;
 	u_int		pm_context;
 
 	struct pmap	*pmap_phys;
diff --git a/sys/sparc64/include/pmap.h b/sys/sparc64/include/pmap.h
index 83c81904b50..388f7512dd4 100644
--- a/sys/sparc64/include/pmap.h
+++ b/sys/sparc64/include/pmap.h
@@ -62,6 +62,8 @@ struct pmap {
 	struct	tte *pm_tsb;
 	vm_object_t pm_tsb_obj;
 	u_int	pm_active;
+	uint32_t	pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 	u_int	pm_context[MAXCPU];
 	struct	pmap_statistics pm_stats;
 };
diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index 22ae860b675..5d69c13e758 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -694,13 +694,17 @@ pmap_extract_and_hold(pmap_t pm, vm_offset_t va, vm_prot_t prot)
 {
 	struct tte *tp;
 	vm_page_t m;
+	vm_paddr_t pa;
 
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;
+	PMAP_LOCK(pm);
+retry:
 	if (pm == kernel_pmap) {
 		if (va >= VM_MIN_DIRECT_ADDRESS) {
 			tp = NULL;
 			m = PHYS_TO_VM_PAGE(TLB_DIRECT_TO_PHYS(va));
+			(void)vm_page_pa_tryrelock(pm, TLB_DIRECT_TO_PHYS(va), &pa);
 			vm_page_hold(m);
 		} else {
 			tp = tsb_kvtotte(va);
@@ -708,17 +712,17 @@ pmap_extract_and_hold(pmap_t pm, vm_offset_t va, vm_prot_t prot)
 				tp = NULL;
 		}
 	} else {
-		PMAP_LOCK(pm);
 		tp = tsb_tte_lookup(pm, va);
 	}
 	if (tp != NULL && ((tp->tte_data & TD_SW) ||
 	    (prot & VM_PROT_WRITE) == 0)) {
+		if (vm_page_pa_tryrelock(pm, TTE_GET_PA(tp), &pa))
+			goto retry;
 		m = PHYS_TO_VM_PAGE(TTE_GET_PA(tp));
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
-	if (pm != kernel_pmap)
-		PMAP_UNLOCK(pm);
+	PA_UNLOCK_COND(pa);
+	PMAP_UNLOCK(pm);
 	return (m);
 }
 
diff --git a/sys/sun4v/include/pmap.h b/sys/sun4v/include/pmap.h
index 90ae4c45560..6acebb40d62 100644
--- a/sys/sun4v/include/pmap.h
+++ b/sys/sun4v/include/pmap.h
@@ -75,6 +75,8 @@ struct pmap {
 	struct tte_hash        *pm_hash;
 	TAILQ_HEAD(,pv_entry)	pm_pvlist;	/* list of mappings in pmap */
 	struct hv_tsb_info      pm_tsb;
+	uint32_t		pm_gen_count;	/* generation count (pmap lock dropped) */
+	u_int			pm_retries;
 	pmap_cpumask_t          pm_active;      /* mask of cpus currently using pmap */
 	pmap_cpumask_t          pm_tlbactive;   /* mask of cpus that have used this pmap */
 	struct	pmap_statistics pm_stats;
diff --git a/sys/sun4v/sun4v/pmap.c b/sys/sun4v/sun4v/pmap.c
index cbd8106e18f..0c8442133b2 100644
--- a/sys/sun4v/sun4v/pmap.c
+++ b/sys/sun4v/sun4v/pmap.c
@@ -1275,17 +1275,21 @@ pmap_extract_and_hold(pmap_t pmap, vm_offset_t va, vm_prot_t prot)
 {
 	tte_t tte_data;
 	vm_page_t m;
+	vm_paddr_t pa;
 
 	m = NULL;
-	vm_page_lock_queues();
+	pa = 0;
 	PMAP_LOCK(pmap);
+retry:	
 	tte_data = tte_hash_lookup(pmap->pm_hash, va);
 	if (tte_data != 0 && 
 	    ((tte_data & VTD_SW_W) || (prot & VM_PROT_WRITE) == 0)) {
+		if (vm_page_pa_tryrelock(pmap, TTE_GET_PA(tte_data), &pa))
+			goto retry;
 		m = PHYS_TO_VM_PAGE(TTE_GET_PA(tte_data));
 		vm_page_hold(m);
 	}
-	vm_page_unlock_queues();
+	PA_UNLOCK_COND(pa);
 	PMAP_UNLOCK(pmap);
 
 	return (m);
diff --git a/sys/sys/param.h b/sys/sys/param.h
index 9f87ee2c21c..11c15ab1e84 100644
--- a/sys/sys/param.h
+++ b/sys/sys/param.h
@@ -58,7 +58,7 @@
  *		in the range 5 to 9.
  */
 #undef __FreeBSD_version
-#define __FreeBSD_version 900010	/* Master, propagated to newvers */
+#define __FreeBSD_version 900011	/* Master, propagated to newvers */
 
 #ifndef LOCORE
 #include <sys/types.h>
diff --git a/sys/vm/device_pager.c b/sys/vm/device_pager.c
index 9002e775bb9..996f740c0ee 100644
--- a/sys/vm/device_pager.c
+++ b/sys/vm/device_pager.c
@@ -251,12 +251,16 @@ dev_pager_getpages(object, m, count, reqpage)
 		VM_OBJECT_LOCK(object);
 		dev_pager_updatefake(page, paddr, memattr);
 		if (count > 1) {
-			vm_page_lock_queues();
+
 			for (i = 0; i < count; i++) {
-				if (i != reqpage)
+				if (i != reqpage) {
+					vm_page_lock(m[i]);
+					vm_page_lock_queues();		
 					vm_page_free(m[i]);
+					vm_page_unlock_queues();
+					vm_page_unlock(m[i]);
+				}
 			}
-			vm_page_unlock_queues();
 		}
 	} else {
 		/*
@@ -266,10 +270,13 @@ dev_pager_getpages(object, m, count, reqpage)
 		page = dev_pager_getfake(paddr, memattr);
 		VM_OBJECT_LOCK(object);
 		TAILQ_INSERT_TAIL(&object->un_pager.devp.devp_pglist, page, pageq);
-		vm_page_lock_queues();
-		for (i = 0; i < count; i++)
+		for (i = 0; i < count; i++) {
+			vm_page_lock(m[i]);
+			vm_page_lock_queues();
 			vm_page_free(m[i]);
-		vm_page_unlock_queues();
+			vm_page_unlock_queues();
+			vm_page_unlock(m[i]);
+		}
 		vm_page_insert(page, object, offset);
 		m[reqpage] = page;
 	}
diff --git a/sys/vm/sg_pager.c b/sys/vm/sg_pager.c
index a17fe82c2ef..845401a4bbd 100644
--- a/sys/vm/sg_pager.c
+++ b/sys/vm/sg_pager.c
@@ -198,10 +198,13 @@ sg_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 	TAILQ_INSERT_TAIL(&object->un_pager.sgp.sgp_pglist, page, pageq);
 
 	/* Free the original pages and insert this fake page into the object. */
-	vm_page_lock_queues();
-	for (i = 0; i < count; i++)
+	for (i = 0; i < count; i++) {
+		vm_page_lock(m[i]);
+		vm_page_lock_queues();
 		vm_page_free(m[i]);
-	vm_page_unlock_queues();
+		vm_page_unlock_queues();
+		vm_page_unlock(m[i]);
+	}
 	vm_page_insert(page, object, offset);
 	m[reqpage] = page;
 	page->valid = VM_PAGE_BITS_ALL;
diff --git a/sys/vm/swap_pager.c b/sys/vm/swap_pager.c
index f47719bd1a1..6017a52d4c7 100644
--- a/sys/vm/swap_pager.c
+++ b/sys/vm/swap_pager.c
@@ -1137,12 +1137,21 @@ swap_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 	if (0 < i || j < count) {
 		int k;
 
-		vm_page_lock_queues();
-		for (k = 0; k < i; ++k)
+		
+		for (k = 0; k < i; ++k) {
+			vm_page_lock(m[k]);
+			vm_page_lock_queues();
 			swp_pager_free_nrpage(m[k]);
-		for (k = j; k < count; ++k)
+			vm_page_unlock_queues();
+			vm_page_unlock(m[k]);
+		}
+		for (k = j; k < count; ++k) {
+			vm_page_lock(m[k]);
+			vm_page_lock_queues();
 			swp_pager_free_nrpage(m[k]);
-		vm_page_unlock_queues();
+			vm_page_unlock_queues();
+			vm_page_unlock(m[k]);
+		}
 	}
 
 	/*
@@ -1497,7 +1506,7 @@ swp_pager_async_iodone(struct buf *bp)
 		object = bp->b_pages[0]->object;
 		VM_OBJECT_LOCK(object);
 	}
-	vm_page_lock_queues();
+
 	/*
 	 * cleanup pages.  If an error occurs writing to swap, we are in
 	 * very serious trouble.  If it happens to be a disk error, though,
@@ -1509,6 +1518,8 @@ swp_pager_async_iodone(struct buf *bp)
 	for (i = 0; i < bp->b_npages; ++i) {
 		vm_page_t m = bp->b_pages[i];
 
+		vm_page_lock(m);
+		vm_page_lock_queues();
 		m->oflags &= ~VPO_SWAPINPROG;
 
 		if (bp->b_ioflags & BIO_ERROR) {
@@ -1605,8 +1616,9 @@ swp_pager_async_iodone(struct buf *bp)
 			if (vm_page_count_severe())
 				vm_page_try_to_cache(m);
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 
 	/*
 	 * adjust pip.  NOTE: the original parent may still have its own
@@ -1702,10 +1714,12 @@ swp_pager_force_pagein(vm_object_t object, vm_pindex_t pindex)
 	m = vm_page_grab(object, pindex, VM_ALLOC_NORMAL|VM_ALLOC_RETRY);
 	if (m->valid == VM_PAGE_BITS_ALL) {
 		vm_object_pip_subtract(object, 1);
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_activate(m);
 		vm_page_dirty(m);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		vm_page_wakeup(m);
 		vm_pager_page_unswapped(m);
 		return;
@@ -1714,10 +1728,12 @@ swp_pager_force_pagein(vm_object_t object, vm_pindex_t pindex)
 	if (swap_pager_getpages(object, &m, 1, 0) != VM_PAGER_OK)
 		panic("swap_pager_force_pagein: read from swap failed");/*XXX*/
 	vm_object_pip_subtract(object, 1);
+	vm_page_lock(m);
 	vm_page_lock_queues();
 	vm_page_dirty(m);
 	vm_page_dontneed(m);
 	vm_page_unlock_queues();
+	vm_page_unlock(m);
 	vm_page_wakeup(m);
 	vm_pager_page_unswapped(m);
 }
diff --git a/sys/vm/uma_core.c b/sys/vm/uma_core.c
index a2d5633bd51..00a8276d36f 100644
--- a/sys/vm/uma_core.c
+++ b/sys/vm/uma_core.c
@@ -1022,10 +1022,12 @@ obj_alloc(uma_zone_t zone, int bytes, u_int8_t *flags, int wait)
 			while (pages != startpages) {
 				pages--;
 				p = TAILQ_LAST(&object->memq, pglist);
+				vm_page_lock(p);
 				vm_page_lock_queues();
 				vm_page_unwire(p, 0);
 				vm_page_free(p);
 				vm_page_unlock_queues();
+				vm_page_unlock(p);
 			}
 			retkva = 0;
 			goto done;
diff --git a/sys/vm/vm_contig.c b/sys/vm/vm_contig.c
index 78d7e280836..7220055cd25 100644
--- a/sys/vm/vm_contig.c
+++ b/sys/vm/vm_contig.c
@@ -257,9 +257,11 @@ retry:
 				i -= PAGE_SIZE;
 				m = vm_page_lookup(object, OFF_TO_IDX(offset +
 				    i));
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				vm_page_free(m);
 				vm_page_unlock_queues();
+				vm_page_unlock(m);
 			}
 			VM_OBJECT_UNLOCK(object);
 			vm_map_delete(map, addr, addr + size);
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 058cbb09666..de7491518c2 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -137,9 +137,11 @@ release_page(struct faultstate *fs)
 {
 
 	vm_page_wakeup(fs->m);
+	vm_page_lock(fs->m);
 	vm_page_lock_queues();
 	vm_page_deactivate(fs->m);
 	vm_page_unlock_queues();
+	vm_page_unlock(fs->m);
 	fs->m = NULL;
 }
 
@@ -161,9 +163,11 @@ unlock_and_deallocate(struct faultstate *fs)
 	VM_OBJECT_UNLOCK(fs->object);
 	if (fs->object != fs->first_object) {
 		VM_OBJECT_LOCK(fs->first_object);
+		vm_page_lock(fs->first_m);
 		vm_page_lock_queues();
 		vm_page_free(fs->first_m);
 		vm_page_unlock_queues();
+		vm_page_unlock(fs->first_m);
 		vm_object_pip_wakeup(fs->first_object);
 		VM_OBJECT_UNLOCK(fs->first_object);
 		fs->first_m = NULL;
@@ -305,12 +309,14 @@ RetryFault:;
 			 * removes the page from the backing object, 
 			 * which is not what we want.
 			 */
+			vm_page_lock(fs.m);
 			vm_page_lock_queues();
 			if ((fs.m->cow) && 
 			    (fault_type & VM_PROT_WRITE) &&
 			    (fs.object == fs.first_object)) {
 				vm_page_cowfault(fs.m);
 				vm_page_unlock_queues();
+				vm_page_unlock(fs.m);
 				unlock_and_deallocate(&fs);
 				goto RetryFault;
 			}
@@ -333,12 +339,15 @@ RetryFault:;
 			 */
 			if ((fs.m->oflags & VPO_BUSY) || fs.m->busy) {
 				vm_page_unlock_queues();
+				vm_page_unlock(fs.m);
 				VM_OBJECT_UNLOCK(fs.object);
 				if (fs.object != fs.first_object) {
 					VM_OBJECT_LOCK(fs.first_object);
+					vm_page_lock(fs.first_m);
 					vm_page_lock_queues();
 					vm_page_free(fs.first_m);
 					vm_page_unlock_queues();
+					vm_page_unlock(fs.first_m);
 					vm_object_pip_wakeup(fs.first_object);
 					VM_OBJECT_UNLOCK(fs.first_object);
 					fs.first_m = NULL;
@@ -358,6 +367,7 @@ RetryFault:;
 			}
 			vm_pageq_remove(fs.m);
 			vm_page_unlock_queues();
+			vm_page_unlock(fs.m);
 
 			/*
 			 * Mark page busy for other processes, and the 
@@ -481,17 +491,25 @@ readrest:
 						continue;
 					if (!are_queues_locked) {
 						are_queues_locked = TRUE;
+						vm_page_lock(mt);
+						vm_page_lock_queues();
+					} else {
+						vm_page_unlock_queues();
+						vm_page_lock(mt);
 						vm_page_lock_queues();
 					}
 					if (mt->hold_count ||
-						mt->wire_count) 
+					    mt->wire_count) {
+						vm_page_unlock(mt);
 						continue;
+					}
 					pmap_remove_all(mt);
 					if (mt->dirty) {
 						vm_page_deactivate(mt);
 					} else {
 						vm_page_cache(mt);
 					}
+					vm_page_unlock(mt);
 				}
 				if (are_queues_locked)
 					vm_page_unlock_queues();
@@ -623,17 +641,21 @@ vnode_locked:
 			 */
 			if (((fs.map != kernel_map) && (rv == VM_PAGER_ERROR)) ||
 				(rv == VM_PAGER_BAD)) {
+				vm_page_lock(fs.m);
 				vm_page_lock_queues();
 				vm_page_free(fs.m);
 				vm_page_unlock_queues();
+				vm_page_unlock(fs.m);
 				fs.m = NULL;
 				unlock_and_deallocate(&fs);
 				return ((rv == VM_PAGER_ERROR) ? KERN_FAILURE : KERN_PROTECTION_FAILURE);
 			}
 			if (fs.object != fs.first_object) {
+				vm_page_lock(fs.m);
 				vm_page_lock_queues();
 				vm_page_free(fs.m);
 				vm_page_unlock_queues();
+				vm_page_unlock(fs.m);
 				fs.m = NULL;
 				/*
 				 * XXX - we cannot just fall out at this
@@ -746,18 +768,24 @@ vnode_locked:
 				 * We don't chase down the shadow chain
 				 */
 			    fs.object == fs.first_object->backing_object) {
+				vm_page_lock(fs.first_m);
 				vm_page_lock_queues();
 				/*
 				 * get rid of the unnecessary page
 				 */
 				vm_page_free(fs.first_m);
+				vm_page_unlock_queues();
+				vm_page_unlock(fs.first_m);
 				/*
 				 * grab the page and put it into the 
 				 * process'es object.  The page is 
 				 * automatically made dirty.
 				 */
+				vm_page_lock(fs.m);
+				vm_page_lock_queues();
 				vm_page_rename(fs.m, fs.first_object, fs.first_pindex);
 				vm_page_unlock_queues();
+				vm_page_unlock(fs.m);
 				vm_page_busy(fs.m);
 				fs.first_m = fs.m;
 				fs.m = NULL;
@@ -770,10 +798,17 @@ vnode_locked:
 				fs.first_m->valid = VM_PAGE_BITS_ALL;
 				if (wired && (fault_flags &
 				    VM_FAULT_CHANGE_WIRING) == 0) {
+					vm_page_lock(fs.first_m);
 					vm_page_lock_queues();
 					vm_page_wire(fs.first_m);
+					vm_page_unlock_queues();
+					vm_page_unlock(fs.first_m);
+					
+					vm_page_lock(fs.m);
+					vm_page_lock_queues();
 					vm_page_unwire(fs.m, FALSE);
 					vm_page_unlock_queues();
+					vm_page_unlock(fs.m);
 				}
 				/*
 				 * We no longer need the old page or object.
@@ -923,6 +958,7 @@ vnode_locked:
 	if ((fault_flags & VM_FAULT_CHANGE_WIRING) == 0 && wired == 0)
 		vm_fault_prefault(fs.map->pmap, vaddr, fs.entry);
 	VM_OBJECT_LOCK(fs.object);
+	vm_page_lock(fs.m);
 	vm_page_lock_queues();
 
 	/*
@@ -938,6 +974,7 @@ vnode_locked:
 		vm_page_activate(fs.m);
 	}
 	vm_page_unlock_queues();
+	vm_page_unlock(fs.m);
 	vm_page_wakeup(fs.m);
 
 	/*
@@ -1014,9 +1051,11 @@ vm_fault_prefault(pmap_t pmap, vm_offset_t addra, vm_map_entry_t entry)
 		}
 		if (m->valid == VM_PAGE_BITS_ALL &&
 		    (m->flags & PG_FICTITIOUS) == 0) {
+			vm_page_lock(m);
 			vm_page_lock_queues();
 			pmap_enter_quick(pmap, addr, m, entry->protection);
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
 		}
 		VM_OBJECT_UNLOCK(lobject);
 	}
@@ -1092,9 +1131,11 @@ vm_fault_unwire(vm_map_t map, vm_offset_t start, vm_offset_t end,
 		if (pa != 0) {
 			pmap_change_wiring(pmap, va, FALSE);
 			if (!fictitious) {
+				vm_page_lock(PHYS_TO_VM_PAGE(pa));
 				vm_page_lock_queues();
 				vm_page_unwire(PHYS_TO_VM_PAGE(pa), 1);
 				vm_page_unlock_queues();
+				vm_page_unlock(PHYS_TO_VM_PAGE(pa));
 			}
 		}
 	}
@@ -1237,13 +1278,26 @@ vm_fault_copy_entry(vm_map_t dst_map, vm_map_t src_map,
 		 * Mark it no longer busy, and put it on the active list.
 		 */
 		VM_OBJECT_LOCK(dst_object);
-		vm_page_lock_queues();
+		
 		if (upgrade) {
+			vm_page_lock(src_m);
+			vm_page_lock_queues();
 			vm_page_unwire(src_m, 0);
+			vm_page_unlock_queues();
+			vm_page_lock(src_m);
+
+			vm_page_lock(dst_m);
+			vm_page_lock_queues();
 			vm_page_wire(dst_m);
-		} else
+			vm_page_unlock_queues();
+			vm_page_lock(dst_m);
+		} else {
+			vm_page_lock(dst_m);
+			vm_page_lock_queues();
 			vm_page_activate(dst_m);
-		vm_page_unlock_queues();
+			vm_page_unlock_queues();
+			vm_page_lock(dst_m);
+		}
 		vm_page_wakeup(dst_m);
 	}
 	VM_OBJECT_UNLOCK(dst_object);
diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index 4eeaa4dd47f..288c5d72adc 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -257,16 +257,18 @@ vm_imgact_hold_page(vm_object_t object, vm_ooffset_t offset)
 		if (m == NULL)
 			goto out;
 		if (rv != VM_PAGER_OK) {
+			vm_page_lock(m);
 			vm_page_lock_queues();
 			vm_page_free(m);
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
 			m = NULL;
 			goto out;
 		}
 	}
-	vm_page_lock_queues();
+	vm_page_lock(m);
 	vm_page_hold(m);
-	vm_page_unlock_queues();
+	vm_page_unlock(m);
 	vm_page_wakeup(m);
 out:
 	VM_OBJECT_UNLOCK(object);
@@ -300,9 +302,9 @@ vm_imgact_unmap_page(struct sf_buf *sf)
 	m = sf_buf_page(sf);
 	sf_buf_free(sf);
 	sched_unpin();
-	vm_page_lock_queues();
+	vm_page_lock(m);
 	vm_page_unhold(m);
-	vm_page_unlock_queues();
+	vm_page_unlock(m);
 }
 
 void
@@ -434,10 +436,12 @@ vm_thread_stack_dispose(vm_object_t ksobj, vm_offset_t ks, int pages)
 		m = vm_page_lookup(ksobj, i);
 		if (m == NULL)
 			panic("vm_thread_dispose: kstack already missing?");
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_free(m);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(ksobj);
 	vm_object_deallocate(ksobj);
@@ -524,9 +528,11 @@ vm_thread_swapout(struct thread *td)
 		if (m == NULL)
 			panic("vm_thread_swapout: kstack already missing?");
 		vm_page_dirty(m);
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(ksobj);
 }
diff --git a/sys/vm/vm_kern.c b/sys/vm/vm_kern.c
index 739d289b613..ead6d0d71a6 100644
--- a/sys/vm/vm_kern.c
+++ b/sys/vm/vm_kern.c
@@ -380,10 +380,12 @@ retry:
 				i -= PAGE_SIZE;
 				m = vm_page_lookup(kmem_object,
 						   OFF_TO_IDX(offset + i));
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				vm_page_unwire(m, 0);
 				vm_page_free(m);
 				vm_page_unlock_queues();
+				vm_page_unlock(m);
 			}
 			VM_OBJECT_UNLOCK(kmem_object);
 			vm_map_delete(map, addr, addr + size);
diff --git a/sys/vm/vm_mmap.c b/sys/vm/vm_mmap.c
index 88ed3d5655b..f9b3db3c6ff 100644
--- a/sys/vm/vm_mmap.c
+++ b/sys/vm/vm_mmap.c
@@ -866,6 +866,7 @@ RestartScan:
 				 */
 				if (m != NULL && m->valid != 0) {
 					mincoreinfo = MINCORE_INCORE;
+					vm_page_lock(m);
 					vm_page_lock_queues();
 					if (m->dirty ||
 						pmap_is_modified(m))
@@ -874,6 +875,7 @@ RestartScan:
 					    pmap_is_referenced(m))
 						mincoreinfo |= MINCORE_REFERENCED_OTHER;
 					vm_page_unlock_queues();
+					vm_page_unlock(m);
 				}
 				VM_OBJECT_UNLOCK(current->object.vm_object);
 			}
diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index c25ab775794..47ef9737151 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -717,19 +717,21 @@ vm_object_terminate(vm_object_t object)
 	 * removes them from paging queues. Don't free wired pages, just
 	 * remove them from the object. 
 	 */
-	vm_page_lock_queues();
 	while ((p = TAILQ_FIRST(&object->memq)) != NULL) {
 		KASSERT(!p->busy && (p->oflags & VPO_BUSY) == 0,
 			("vm_object_terminate: freeing busy page %p "
 			"p->busy = %d, p->oflags %x\n", p, p->busy, p->oflags));
+		vm_page_lock(p);
+		vm_page_lock_queues();
 		if (p->wire_count == 0) {
 			vm_page_free(p);
 			cnt.v_pfree++;
 		} else {
 			vm_page_remove(p);
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(p);
 	}
-	vm_page_unlock_queues();
 
 #if VM_NRESERVLEVEL > 0
 	if (__predict_false(!LIST_EMPTY(&object->rvq)))
@@ -789,7 +791,6 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 		tend = end;
 	}
 
-	vm_page_lock_queues();
 	/*
 	 * If the caller is smart and only msync()s a range he knows is
 	 * dirty, we may be able to avoid an object scan.  This results in
@@ -818,8 +819,12 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 				++tscan;
 				continue;
 			}
+			vm_page_lock(p);
+			vm_page_lock_queues();
 			vm_page_test_dirty(p);
 			if (p->dirty == 0) {
+				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				if (--scanlimit == 0)
 					break;
 				++tscan;
@@ -830,6 +835,8 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 			 * this is a nosync page, we can't continue.
 			 */
 			if ((flags & OBJPC_NOSYNC) && (p->oflags & VPO_NOSYNC)) {
+				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				if (--scanlimit == 0)
 					break;
 				++tscan;
@@ -842,6 +849,8 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 			 * page (i.e. had to sleep).
 			 */
 			tscan += vm_object_page_collect_flush(object, p, curgeneration, pagerflags);
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 		}
 
 		/*
@@ -871,8 +880,13 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 		p->oflags |= VPO_CLEANCHK;
 		if ((flags & OBJPC_NOSYNC) && (p->oflags & VPO_NOSYNC))
 			clearobjflags = 0;
-		else
+		else {
+			vm_page_lock(p);
+			vm_page_lock_queues();
 			pmap_remove_write(p);
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
+		}
 	}
 
 	if (clearobjflags && (tstart == 0) && (tend == object->size))
@@ -895,8 +909,12 @@ again:
 			continue;
 		}
 
+		vm_page_lock(p);
+		vm_page_lock_queues();
 		vm_page_test_dirty(p);
 		if (p->dirty == 0) {
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			p->oflags &= ~VPO_CLEANCHK;
 			continue;
 		}
@@ -907,28 +925,37 @@ again:
 		 * not cleared in this case so we do not have to set them.
 		 */
 		if ((flags & OBJPC_NOSYNC) && (p->oflags & VPO_NOSYNC)) {
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			p->oflags &= ~VPO_CLEANCHK;
 			continue;
 		}
 
 		n = vm_object_page_collect_flush(object, p,
 			curgeneration, pagerflags);
-		if (n == 0)
+		if (n == 0) {
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			goto rescan;
+		}
 
-		if (object->generation != curgeneration)
+		if (object->generation != curgeneration) {
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			goto rescan;
+		}
 
 		/*
 		 * Try to optimize the next page.  If we can't we pick up
 		 * our (random) scan where we left off.
 		 */
 		if (msync_flush_flags & MSYNC_FLUSH_SOFTSEQ) {
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			if ((p = vm_page_lookup(object, pi + n)) != NULL)
 				goto again;
 		}
 	}
-	vm_page_unlock_queues();
 #if 0
 	VOP_FSYNC(vp, (pagerflags & VM_PAGER_PUT_SYNC)?MNT_WAIT:0, curproc);
 #endif
@@ -951,8 +978,10 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 	vm_page_t ma[vm_pageout_page_count];
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(p, MA_OWNED);
 	pi = p->pindex;
 	while (vm_page_sleep_if_busy(p, TRUE, "vpcwai")) {
+		vm_page_lock(p);
 		vm_page_lock_queues();
 		if (object->generation != curgeneration) {
 			return(0);
@@ -968,11 +997,16 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 				 (tp->oflags & VPO_CLEANCHK) == 0) ||
 				(tp->busy != 0))
 				break;
+			vm_page_unlock_queues();
+			vm_page_lock(tp);
+			vm_page_lock_queues();
 			vm_page_test_dirty(tp);
 			if (tp->dirty == 0) {
+				vm_page_unlock(tp);
 				tp->oflags &= ~VPO_CLEANCHK;
 				break;
 			}
+			vm_page_unlock(tp);
 			maf[ i - 1 ] = tp;
 			maxf++;
 			continue;
@@ -992,11 +1026,16 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 					 (tp->oflags & VPO_CLEANCHK) == 0) ||
 					(tp->busy != 0))
 					break;
+				vm_page_unlock_queues();
+				vm_page_lock(tp);
+				vm_page_lock_queues();
 				vm_page_test_dirty(tp);
 				if (tp->dirty == 0) {
+					vm_page_unlock(tp);
 					tp->oflags &= ~VPO_CLEANCHK;
 					break;
 				}
+				vm_page_unlock(tp);
 				mab[ i - 1 ] = tp;
 				maxb++;
 				continue;
@@ -1022,7 +1061,11 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 	vm_pageout_flush(ma, runlen, pagerflags);
 	for (i = 0; i < runlen; i++) {
 		if (ma[i]->dirty) {
+			vm_page_unlock_queues();
+			vm_page_lock(ma[i]);
+			vm_page_lock_queues();
 			pmap_remove_write(ma[i]);
+			vm_page_unlock(ma[i]);
 			ma[i]->oflags |= VPO_CLEANCHK;
 
 			/*
@@ -1195,9 +1238,11 @@ shadowlookup:
 		/*
 		 * If the page is not in a normal state, skip it.
 		 */
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		if (m->hold_count != 0 || m->wire_count != 0) {
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
 			goto unlock_tobject;
 		}
 		if ((m->oflags & VPO_BUSY) || m->busy) {
@@ -1209,6 +1254,7 @@ shadowlookup:
 				 */
 				vm_page_flag_set(m, PG_REFERENCED);
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
 			if (object != tobject)
 				VM_OBJECT_UNLOCK(object);
 			m->oflags |= VPO_WANTED;
@@ -1243,6 +1289,7 @@ shadowlookup:
 			vm_page_dontneed(m);
 		}
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		if (advise == MADV_FREE && tobject->type == OBJT_SWAP)
 			swap_pager_freespace(tobject, tpindex, 1);
 unlock_tobject:
@@ -1405,7 +1452,6 @@ retry:
 				m = TAILQ_NEXT(m, listq);
 		}
 	}
-	vm_page_lock_queues();
 	for (; m != NULL && (idx = m->pindex - offidxstart) < size;
 	    m = m_next) {
 		m_next = TAILQ_NEXT(m, listq);
@@ -1425,11 +1471,14 @@ retry:
 			VM_OBJECT_LOCK(new_object);
 			goto retry;
 		}
+		vm_page_lock(m);
+		vm_page_lock_queues();
 		vm_page_rename(m, new_object, idx);
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		/* page automatically made dirty by rename and cache handled */
 		vm_page_busy(m);
 	}
-	vm_page_unlock_queues();
 	if (orig_object->type == OBJT_SWAP) {
 		/*
 		 * swap_pager_copy() can sleep, in which case the orig_object's
@@ -1597,6 +1646,7 @@ vm_object_backing_scan(vm_object_t object, int op)
 				 * Page is out of the parent object's range, we 
 				 * can simply destroy it. 
 				 */
+				vm_page_lock(p);
 				vm_page_lock_queues();
 				KASSERT(!pmap_page_is_mapped(p),
 				    ("freeing mapped page %p", p));
@@ -1605,6 +1655,7 @@ vm_object_backing_scan(vm_object_t object, int op)
 				else
 					vm_page_remove(p);
 				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				p = next;
 				continue;
 			}
@@ -1621,6 +1672,7 @@ vm_object_backing_scan(vm_object_t object, int op)
 				 *
 				 * Leave the parent's page alone
 				 */
+				vm_page_lock(p);
 				vm_page_lock_queues();
 				KASSERT(!pmap_page_is_mapped(p),
 				    ("freeing mapped page %p", p));
@@ -1629,6 +1681,7 @@ vm_object_backing_scan(vm_object_t object, int op)
 				else
 					vm_page_remove(p);
 				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				p = next;
 				continue;
 			}
@@ -1648,9 +1701,11 @@ vm_object_backing_scan(vm_object_t object, int op)
 			 * If the page was mapped to a process, it can remain 
 			 * mapped through the rename.
 			 */
+			vm_page_lock(p);
 			vm_page_lock_queues();
 			vm_page_rename(p, object, new_pindex);
 			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			/* page automatically made dirty by rename */
 		}
 		p = next;
@@ -1915,7 +1970,7 @@ again:
 				p = TAILQ_NEXT(p, listq);
 		}
 	}
-	vm_page_lock_queues();
+
 	/*
 	 * Assert: the variable p is either (1) the page with the
 	 * least pindex greater than or equal to the parameter pindex
@@ -1934,6 +1989,8 @@ again:
 		 * cannot be freed.  They can, however, be invalidated
 		 * if "clean_only" is FALSE.
 		 */
+		vm_page_lock(p);
+		vm_page_lock_queues();
 		if ((wirings = p->wire_count) != 0 &&
 		    (wirings = pmap_page_wired_mappings(p)) != p->wire_count) {
 			/* Fictitious pages do not have managed mappings. */
@@ -1945,6 +2002,8 @@ again:
 				p->valid = 0;
 				vm_page_undirty(p);
 			}
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			continue;
 		}
 		if (vm_page_sleep_if_busy(p, TRUE, "vmopar"))
@@ -1953,16 +2012,20 @@ again:
 		    ("vm_object_page_remove: page %p is fictitious", p));
 		if (clean_only && p->valid) {
 			pmap_remove_write(p);
-			if (p->dirty)
+			if (p->dirty) {
+				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				continue;
+			}
 		}
 		pmap_remove_all(p);
 		/* Account for removal of managed, wired mappings. */
 		if (wirings != 0)
 			p->wire_count -= wirings;
 		vm_page_free(p);
+		vm_page_unlock_queues();
+		vm_page_unlock(p);
 	}
-	vm_page_unlock_queues();
 	vm_object_pip_wakeup(object);
 skipmemq:
 	if (__predict_false(object->cache != NULL))
@@ -1997,9 +2060,11 @@ vm_object_populate(vm_object_t object, vm_pindex_t start, vm_pindex_t end)
 			if (m == NULL)
 				break;
 			if (rv != VM_PAGER_OK) {
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				vm_page_free(m);
 				vm_page_unlock_queues();
+				vm_page_unlock(m);
 				break;
 			}
 		}
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index 7c149c9bc88..4c0d3850669 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -115,6 +115,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/vnode.h>
 
 #include <vm/vm.h>
+#include <vm/pmap.h>
 #include <vm/vm_param.h>
 #include <vm/vm_kern.h>
 #include <vm/vm_object.h>
@@ -129,6 +130,24 @@ __FBSDID("$FreeBSD$");
 
 #include <machine/md_var.h>
 
+#if defined(__amd64__) || defined (__i386__) 
+extern struct sysctl_oid_list sysctl__vm_pmap_children;
+#else
+SYSCTL_NODE(_vm, OID_AUTO, pmap, CTLFLAG_RD, 0, "VM/pmap parameters");
+#endif
+
+static uint64_t pmap_tryrelock_calls;
+SYSCTL_QUAD(_vm_pmap, OID_AUTO, tryrelock_calls, CTLFLAG_RD,
+    &pmap_tryrelock_calls, 0, "Number of tryrelock calls");
+
+static int pmap_tryrelock_restart;
+SYSCTL_INT(_vm_pmap, OID_AUTO, tryrelock_restart, CTLFLAG_RD,
+    &pmap_tryrelock_restart, 0, "Number of tryrelock restarts");
+
+static int pmap_tryrelock_race;
+SYSCTL_INT(_vm_pmap, OID_AUTO, tryrelock_race, CTLFLAG_RD,
+    &pmap_tryrelock_race, 0, "Number of tryrelock pmap race cases");
+
 /*
  *	Associated with page of user-allocatable memory is a
  *	page structure.
@@ -138,6 +157,8 @@ struct vpgqueues vm_page_queues[PQ_COUNT];
 struct vpglocks vm_page_queue_lock;
 struct vpglocks vm_page_queue_free_lock;
 
+struct vpglocks	pa_lock[PA_LOCK_COUNT] __aligned(CACHE_LINE_SIZE);
+
 vm_page_t vm_page_array = 0;
 int vm_page_array_size = 0;
 long first_page = 0;
@@ -157,6 +178,43 @@ CTASSERT(sizeof(u_long) >= 8);
 #endif
 #endif
 
+/*
+ * Try to acquire a physical address lock while a pmap is locked.  If we
+ * fail to trylock we unlock and lock the pmap directly and cache the
+ * locked pa in *locked.  The caller should then restart their loop in case
+ * the virtual to physical mapping has changed.
+ */
+int
+vm_page_pa_tryrelock(pmap_t pmap, vm_paddr_t pa, vm_paddr_t *locked)
+{
+	vm_paddr_t lockpa;
+	uint32_t gen_count;
+
+	gen_count = pmap->pm_gen_count;
+	atomic_add_long((volatile long *)&pmap_tryrelock_calls, 1);
+	lockpa = *locked;
+	*locked = pa;
+	if (lockpa) {
+		PA_LOCK_ASSERT(lockpa, MA_OWNED);
+		if (PA_LOCKPTR(pa) == PA_LOCKPTR(lockpa))
+			return (0);
+		PA_UNLOCK(lockpa);
+	}
+	if (PA_TRYLOCK(pa))
+		return (0);
+	PMAP_UNLOCK(pmap);
+	atomic_add_int((volatile int *)&pmap_tryrelock_restart, 1);
+	PA_LOCK(pa);
+	PMAP_LOCK(pmap);
+
+	if (pmap->pm_gen_count != gen_count + 1) {
+		pmap->pm_retries++;
+		atomic_add_int((volatile int *)&pmap_tryrelock_race, 1);
+		return (EAGAIN);
+	}
+	return (0);
+}
+
 /*
  *	vm_set_page_size:
  *
@@ -271,6 +329,11 @@ vm_page_startup(vm_offset_t vaddr)
 	mtx_init(&vm_page_queue_free_mtx, "vm page queue free mutex", NULL,
 	    MTX_DEF);
 
+	/* Setup page locks. */
+	for (i = 0; i < PA_LOCK_COUNT; i++)
+		mtx_init(&pa_lock[i].data, "page lock", NULL,
+		    MTX_DEF | MTX_RECURSE | MTX_DUPOK);
+
 	/*
 	 * Initialize the queue headers for the hold queue, the active queue,
 	 * and the inactive queue.
@@ -489,7 +552,7 @@ void
 vm_page_hold(vm_page_t mem)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(mem, MA_OWNED);
         mem->hold_count++;
 }
 
@@ -497,7 +560,7 @@ void
 vm_page_unhold(vm_page_t mem)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(mem, MA_OWNED);
 	--mem->hold_count;
 	KASSERT(mem->hold_count >= 0, ("vm_page_unhold: hold count < 0!!!"));
 	if (mem->hold_count == 0 && VM_PAGE_INQUEUE2(mem, PQ_HOLD))
@@ -542,10 +605,13 @@ vm_page_sleep(vm_page_t m, const char *msg)
 {
 
 	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
+	if (!mtx_owned(vm_page_lockptr(m)))
+		vm_page_lock(m);
 	if (!mtx_owned(&vm_page_queue_mtx))
 		vm_page_lock_queues();
 	vm_page_flag_set(m, PG_REFERENCED);
 	vm_page_unlock_queues();
+	vm_page_unlock(m);
 
 	/*
 	 * It's possible that while we sleep, the page will get
@@ -1425,6 +1491,7 @@ vm_page_free_toq(vm_page_t m)
 		panic("vm_page_free: freeing wired page");
 	}
 	if (m->hold_count != 0) {
+		vm_page_lock_assert(m, MA_OWNED);
 		m->flags &= ~PG_ZERO;
 		vm_page_enqueue(PQ_HOLD, m);
 	} else {
diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index 662af98be83..35a81f8a8d7 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -177,9 +177,35 @@ struct vpglocks {
 } __aligned(CACHE_LINE_SIZE);
 
 extern struct vpglocks vm_page_queue_free_lock;
+extern struct vpglocks pa_lock[];
+
+#if defined(__arm__)
+#define	PDRSHIFT	PDR_SHIFT
+#elif !defined(PDRSHIFT)
+#define PDRSHIFT	21
+#endif
+
+#define	pa_index(pa)	((pa) >> PDRSHIFT)
+#define	PA_LOCKPTR(pa)	&pa_lock[pa_index((pa)) % PA_LOCK_COUNT].data
+#define	PA_LOCKOBJPTR(pa)	((struct lock_object *)PA_LOCKPTR((pa)))
+#define	PA_LOCK(pa)	mtx_lock(PA_LOCKPTR(pa))
+#define	PA_TRYLOCK(pa)	mtx_trylock(PA_LOCKPTR(pa))
+#define	PA_UNLOCK(pa)	mtx_unlock(PA_LOCKPTR(pa))
+#define	PA_UNLOCK_COND(pa) 			\
+	do {		   			\
+		if (pa) 			\
+			PA_UNLOCK(pa);		\
+	} while (0)
+
+#define	PA_LOCK_ASSERT(pa, a)	mtx_assert(PA_LOCKPTR(pa), (a))
+
+#define	vm_page_lockptr(m)	(PA_LOCKPTR(VM_PAGE_TO_PHYS((m))))
+#define	vm_page_lock(m)		mtx_lock(vm_page_lockptr((m)))
+#define	vm_page_unlock(m)	mtx_unlock(vm_page_lockptr((m)))
+#define	vm_page_trylock(m)	mtx_trylock(vm_page_lockptr((m)))
+#define	vm_page_lock_assert(m, a)	mtx_assert(vm_page_lockptr((m)), (a))
 
 #define	vm_page_queue_free_mtx	vm_page_queue_free_lock.data
-
 /*
  * These are the flags defined for vm_page.
  *
@@ -324,6 +350,7 @@ void vm_page_dontneed(vm_page_t);
 void vm_page_deactivate (vm_page_t);
 void vm_page_insert (vm_page_t, vm_object_t, vm_pindex_t);
 vm_page_t vm_page_lookup (vm_object_t, vm_pindex_t);
+int vm_page_pa_tryrelock(pmap_t, vm_paddr_t, vm_paddr_t *);
 void vm_page_remove (vm_page_t);
 void vm_page_rename (vm_page_t, vm_object_t, vm_pindex_t);
 void vm_page_requeue(vm_page_t m);
diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 258da8b10fc..95517dec5c3 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -252,7 +252,9 @@ vm_pageout_fallback_object_lock(vm_page_t m, vm_page_t *next)
 	TAILQ_INSERT_AFTER(&vm_page_queues[queue].pl,
 			   m, &marker, pageq);
 	vm_page_unlock_queues();
+	vm_page_unlock(m);
 	VM_OBJECT_LOCK(object);
+	vm_page_lock(m);
 	vm_page_lock_queues();
 
 	/* Page queue might have changed. */
@@ -275,8 +277,7 @@ vm_pageout_fallback_object_lock(vm_page_t m, vm_page_t *next)
  * late and we cannot do anything that will mess with the page.
  */
 static int
-vm_pageout_clean(m)
-	vm_page_t m;
+vm_pageout_clean(vm_page_t m)
 {
 	vm_object_t object;
 	vm_page_t mc[2*vm_pageout_page_count];
@@ -284,7 +285,8 @@ vm_pageout_clean(m)
 	int ib, is, page_base;
 	vm_pindex_t pindex = m->pindex;
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_NOTOWNED);
+	vm_page_lock(m);
 	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
 
 	/*
@@ -301,6 +303,7 @@ vm_pageout_clean(m)
 	 */
 	if ((m->hold_count != 0) ||
 	    ((m->busy != 0) || (m->oflags & VPO_BUSY))) {
+		vm_page_unlock(m);
 		return 0;
 	}
 
@@ -347,13 +350,19 @@ more:
 			ib = 0;
 			break;
 		}
+		vm_page_lock(p);
+		vm_page_lock_queues();
 		vm_page_test_dirty(p);
 		if (p->dirty == 0 ||
 		    p->queue != PQ_INACTIVE ||
 		    p->hold_count != 0) {	/* may be undergoing I/O */
+			vm_page_unlock(p);
+			vm_page_unlock_queues();
 			ib = 0;
 			break;
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(p);
 		mc[--page_base] = p;
 		++pageout_count;
 		++ib;
@@ -374,12 +383,18 @@ more:
 		if ((p->oflags & VPO_BUSY) || p->busy) {
 			break;
 		}
+		vm_page_lock(p);
+		vm_page_lock_queues();
 		vm_page_test_dirty(p);
 		if (p->dirty == 0 ||
 		    p->queue != PQ_INACTIVE ||
 		    p->hold_count != 0) {	/* may be undergoing I/O */
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			break;
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(p);
 		mc[page_base + pageout_count] = p;
 		++pageout_count;
 		++is;
@@ -393,6 +408,7 @@ more:
 	if (ib && pageout_count < vm_pageout_page_count)
 		goto more;
 
+	vm_page_unlock(m);
 	/*
 	 * we allow reads during pageouts...
 	 */
@@ -416,7 +432,6 @@ vm_pageout_flush(vm_page_t *mc, int count, int flags)
 	int numpagedout = 0;
 	int i;
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	/*
 	 * Initiate I/O.  Bump the vm_page_t->busy counter and
@@ -433,17 +448,21 @@ vm_pageout_flush(vm_page_t *mc, int count, int flags)
 		    ("vm_pageout_flush: partially invalid page %p index %d/%d",
 			mc[i], i, count));
 		vm_page_io_start(mc[i]);
+		vm_page_lock(mc[i]);
+		vm_page_lock_queues();
 		pmap_remove_write(mc[i]);
+		vm_page_unlock(mc[i]);
+		vm_page_unlock_queues();
 	}
-	vm_page_unlock_queues();
 	vm_object_pip_add(object, count);
 
 	vm_pager_put_pages(object, mc, count, flags, pageout_status);
 
-	vm_page_lock_queues();
 	for (i = 0; i < count; i++) {
 		vm_page_t mt = mc[i];
 
+		vm_page_lock(mt);
+		vm_page_lock_queues();
 		KASSERT(pageout_status[i] == VM_PAGER_PEND ||
 		    (mt->flags & PG_WRITEABLE) == 0,
 		    ("vm_pageout_flush: page %p is not write protected", mt));
@@ -485,6 +504,8 @@ vm_pageout_flush(vm_page_t *mc, int count, int flags)
 			if (vm_page_count_severe())
 				vm_page_try_to_cache(mt);
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(mt);
 	}
 	return numpagedout;
 }
@@ -527,10 +548,12 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 		 * scan the objects entire memory queue
 		 */
 		p = TAILQ_FIRST(&object->memq);
-		vm_page_lock_queues();
 		while (p != NULL) {
+			vm_page_lock(p);
+			vm_page_lock_queues();
 			if (pmap_resident_count(pmap) <= desired) {
 				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				goto unlock_return;
 			}
 			next = TAILQ_NEXT(p, listq);
@@ -540,6 +563,8 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 			    p->busy != 0 ||
 			    (p->oflags & VPO_BUSY) ||
 			    !pmap_page_exists_quick(pmap, p)) {
+				vm_page_unlock_queues();
+				vm_page_unlock(p);
 				p = next;
 				continue;
 			}
@@ -573,9 +598,10 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 			} else if (p->queue == PQ_INACTIVE) {
 				pmap_remove_all(p);
 			}
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			p = next;
 		}
-		vm_page_unlock_queues();
 		if ((backing_object = object->backing_object) == NULL)
 			goto unlock_return;
 		VM_OBJECT_LOCK(backing_object);
@@ -742,7 +768,6 @@ rescan0:
 		}
 
 		next = TAILQ_NEXT(m, pageq);
-		object = m->object;
 
 		/*
 		 * skip marker pages
@@ -758,18 +783,26 @@ rescan0:
 			addl_page_shortage++;
 			continue;
 		}
+
+		if (!vm_page_trylock(m) || (object = m->object) == NULL) {
+			addl_page_shortage++;
+			continue;
+		}
+
 		/*
 		 * Don't mess with busy pages, keep in the front of the
 		 * queue, most likely are being paged out.
 		 */
 		if (!VM_OBJECT_TRYLOCK(object) &&
 		    (!vm_pageout_fallback_object_lock(m, &next) ||
-		     m->hold_count != 0)) {
+			m->hold_count != 0)) {
 			VM_OBJECT_UNLOCK(object);
+			vm_page_unlock(m);
 			addl_page_shortage++;
 			continue;
 		}
 		if (m->busy || (m->oflags & VPO_BUSY)) {
+			vm_page_unlock(m);
 			VM_OBJECT_UNLOCK(object);
 			addl_page_shortage++;
 			continue;
@@ -798,6 +831,7 @@ rescan0:
 			vm_page_activate(m);
 			VM_OBJECT_UNLOCK(object);
 			m->act_count += (actcount + ACT_ADVANCE);
+			vm_page_unlock(m);
 			continue;
 		}
 
@@ -813,6 +847,7 @@ rescan0:
 			vm_page_activate(m);
 			VM_OBJECT_UNLOCK(object);
 			m->act_count += (actcount + ACT_ADVANCE + 1);
+			vm_page_unlock(m);
 			continue;
 		}
 
@@ -898,6 +933,7 @@ rescan0:
 			 * Those objects are in a "rundown" state.
 			 */
 			if (!swap_pageouts_ok || (object->flags & OBJ_DEAD)) {
+				vm_page_unlock(m);
 				VM_OBJECT_UNLOCK(object);
 				vm_page_requeue(m);
 				continue;
@@ -937,6 +973,8 @@ rescan0:
 			 * of time.
 			 */
 			if (object->type == OBJT_VNODE) {
+				vm_page_unlock_queues();
+				vm_page_unlock(m);
 				vp = object->handle;
 				if (vp->v_type == VREG &&
 				    vn_start_write(vp, &mp, V_NOWAIT) != 0) {
@@ -944,11 +982,11 @@ rescan0:
 					++pageout_lock_miss;
 					if (object->flags & OBJ_MIGHTBEDIRTY)
 						vnodes_skipped++;
+					vm_page_lock_queues();
 					goto unlock_and_continue;
 				}
 				KASSERT(mp != NULL,
 				    ("vp %p with NULL v_mount", vp));
-				vm_page_unlock_queues();
 				vm_object_reference_locked(object);
 				VM_OBJECT_UNLOCK(object);
 				vfslocked = VFS_LOCK_GIANT(vp->v_mount);
@@ -963,6 +1001,7 @@ rescan0:
 					goto unlock_and_continue;
 				}
 				VM_OBJECT_LOCK(object);
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				/*
 				 * The page might have been moved to another
@@ -973,6 +1012,7 @@ rescan0:
 				if (VM_PAGE_GETQUEUE(m) != PQ_INACTIVE ||
 				    m->object != object ||
 				    TAILQ_NEXT(m, pageq) != &marker) {
+					vm_page_unlock(m);
 					if (object->flags & OBJ_MIGHTBEDIRTY)
 						vnodes_skipped++;
 					goto unlock_and_continue;
@@ -985,6 +1025,7 @@ rescan0:
 				 * statistics are more correct if we don't.
 				 */
 				if (m->busy || (m->oflags & VPO_BUSY)) {
+					vm_page_unlock(m);
 					goto unlock_and_continue;
 				}
 
@@ -993,12 +1034,14 @@ rescan0:
 				 * be undergoing I/O, so skip it
 				 */
 				if (m->hold_count) {
+					vm_page_unlock(m);
 					vm_page_requeue(m);
 					if (object->flags & OBJ_MIGHTBEDIRTY)
 						vnodes_skipped++;
 					goto unlock_and_continue;
 				}
 			}
+			vm_page_unlock(m);
 
 			/*
 			 * If a page is dirty, then it is either being washed
@@ -1010,11 +1053,14 @@ rescan0:
 			 * the (future) cleaned page.  Otherwise we could wind
 			 * up laundering or cleaning too many pages.
 			 */
+			vm_page_unlock_queues();
 			if (vm_pageout_clean(m) != 0) {
 				--page_shortage;
 				--maxlaunder;
 			}
+			vm_page_lock_queues();
 unlock_and_continue:
+			vm_page_lock_assert(m, MA_NOTOWNED);
 			VM_OBJECT_UNLOCK(object);
 			if (mp != NULL) {
 				vm_page_unlock_queues();
@@ -1028,8 +1074,10 @@ unlock_and_continue:
 			next = TAILQ_NEXT(&marker, pageq);
 			TAILQ_REMOVE(&vm_page_queues[PQ_INACTIVE].pl,
 				     &marker, pageq);
+			vm_page_lock_assert(m, MA_NOTOWNED);
 			continue;
 		}
+		vm_page_unlock(m);
 		VM_OBJECT_UNLOCK(object);
 	}
 
@@ -1048,6 +1096,7 @@ unlock_and_continue:
 	 */
 	pcount = cnt.v_active_count;
 	m = TAILQ_FIRST(&vm_page_queues[PQ_ACTIVE].pl);
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 
 	while ((m != NULL) && (pcount-- > 0) && (page_shortage > 0)) {
 
@@ -1060,9 +1109,14 @@ unlock_and_continue:
 			m = next;
 			continue;
 		}
+		if (!vm_page_trylock(m) || (object = m->object) == NULL) {
+			m = next;
+			continue;
+		}
 		if (!VM_OBJECT_TRYLOCK(object) &&
 		    !vm_pageout_fallback_object_lock(m, &next)) {
 			VM_OBJECT_UNLOCK(object);
+			vm_page_unlock(m);
 			m = next;
 			continue;
 		}
@@ -1073,6 +1127,7 @@ unlock_and_continue:
 		if ((m->busy != 0) ||
 		    (m->oflags & VPO_BUSY) ||
 		    (m->hold_count != 0)) {
+			vm_page_unlock(m);
 			VM_OBJECT_UNLOCK(object);
 			vm_page_requeue(m);
 			m = next;
@@ -1132,6 +1187,7 @@ unlock_and_continue:
 				vm_page_requeue(m);
 			}
 		}
+		vm_page_unlock(m);
 		VM_OBJECT_UNLOCK(object);
 		m = next;
 	}
@@ -1317,9 +1373,15 @@ vm_pageout_page_stats()
 			m = next;
 			continue;
 		}
+		vm_page_lock_assert(m, MA_NOTOWNED);
+		if (vm_page_trylock(m) == 0 || (object = m->object) == NULL) {
+			m = next;
+			continue;
+		}
 		if (!VM_OBJECT_TRYLOCK(object) &&
 		    !vm_pageout_fallback_object_lock(m, &next)) {
 			VM_OBJECT_UNLOCK(object);
+			vm_page_unlock(m);
 			m = next;
 			continue;
 		}
@@ -1330,6 +1392,7 @@ vm_pageout_page_stats()
 		if ((m->busy != 0) ||
 		    (m->oflags & VPO_BUSY) ||
 		    (m->hold_count != 0)) {
+			vm_page_unlock(m);
 			VM_OBJECT_UNLOCK(object);
 			vm_page_requeue(m);
 			m = next;
@@ -1366,6 +1429,7 @@ vm_pageout_page_stats()
 				vm_page_requeue(m);
 			}
 		}
+		vm_page_unlock(m);
 		VM_OBJECT_UNLOCK(object);
 		m = next;
 	}
diff --git a/sys/vm/vm_param.h b/sys/vm/vm_param.h
index 2ff26036353..c404989cd25 100644
--- a/sys/vm/vm_param.h
+++ b/sys/vm/vm_param.h
@@ -126,6 +126,14 @@ struct xswdev {
 #define	KERN_NOT_RECEIVER	7
 #define	KERN_NO_ACCESS		8
 
+#ifndef PA_LOCK_COUNT
+#ifdef SMP
+#define	PA_LOCK_COUNT	32
+#else
+#define PA_LOCK_COUNT	1
+#endif	/* !SMP */
+#endif	/* !PA_LOCK_COUNT */
+
 #ifndef ASSEMBLER
 #ifdef _KERNEL
 #define num_pages(x) \
diff --git a/sys/vm/vnode_pager.c b/sys/vm/vnode_pager.c
index aedc794c519..eb21c60a7c3 100644
--- a/sys/vm/vnode_pager.c
+++ b/sys/vm/vnode_pager.c
@@ -429,9 +429,11 @@ vnode_pager_setsize(vp, nsize)
 			 * bits.  This would prevent bogus_page
 			 * replacement from working properly.
 			 */
+			vm_page_lock(m);
 			vm_page_lock_queues();
 			vm_page_clear_dirty(m, base, PAGE_SIZE - base);
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
 		} else if ((nsize & PAGE_MASK) &&
 		    __predict_false(object->cache != NULL)) {
 			vm_page_cache_free(object, OFF_TO_IDX(nsize),
@@ -719,11 +721,15 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 	error = VOP_BMAP(vp, foff / bsize, &bo, &reqblock, NULL, NULL);
 	if (error == EOPNOTSUPP) {
 		VM_OBJECT_LOCK(object);
-		vm_page_lock_queues();
+		
 		for (i = 0; i < count; i++)
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(m[i]);
+				vm_page_lock_queues();
 				vm_page_free(m[i]);
-		vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[i]);
+			}
 		PCPU_INC(cnt.v_vnodein);
 		PCPU_INC(cnt.v_vnodepgsin);
 		error = vnode_pager_input_old(object, m[reqpage]);
@@ -731,11 +737,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		return (error);
 	} else if (error != 0) {
 		VM_OBJECT_LOCK(object);
-		vm_page_lock_queues();
 		for (i = 0; i < count; i++)
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(m[i]);
+				vm_page_lock_queues();
 				vm_page_free(m[i]);
-		vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[i]);
+			}
 		VM_OBJECT_UNLOCK(object);
 		return (VM_PAGER_ERROR);
 
@@ -747,11 +756,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 	} else if ((PAGE_SIZE / bsize) > 1 &&
 	    (vp->v_mount->mnt_stat.f_type != nfs_mount_type)) {
 		VM_OBJECT_LOCK(object);
-		vm_page_lock_queues();
 		for (i = 0; i < count; i++)
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(m[i]);
+				vm_page_lock_queues();
 				vm_page_free(m[i]);
-		vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[i]);
+			}
 		VM_OBJECT_UNLOCK(object);
 		PCPU_INC(cnt.v_vnodein);
 		PCPU_INC(cnt.v_vnodepgsin);
@@ -765,11 +777,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 	 */
 	VM_OBJECT_LOCK(object);
 	if (m[reqpage]->valid == VM_PAGE_BITS_ALL) {
-		vm_page_lock_queues();
 		for (i = 0; i < count; i++)
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(m[i]);
+				vm_page_lock_queues();
 				vm_page_free(m[i]);
-		vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[i]);
+			}
 		VM_OBJECT_UNLOCK(object);
 		return VM_PAGER_OK;
 	} else if (reqblock == -1) {
@@ -777,11 +792,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		KASSERT(m[reqpage]->dirty == 0,
 		    ("vnode_pager_generic_getpages: page %p is dirty", m));
 		m[reqpage]->valid = VM_PAGE_BITS_ALL;
-		vm_page_lock_queues();
 		for (i = 0; i < count; i++)
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(m[i]);
+				vm_page_lock_queues();
 				vm_page_free(m[i]);
-		vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[i]);
+			}
 		VM_OBJECT_UNLOCK(object);
 		return (VM_PAGER_OK);
 	}
@@ -800,11 +818,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		if (vnode_pager_addr(vp, IDX_TO_OFF(m[i]->pindex), &firstaddr,
 		    &runpg) != 0) {
 			VM_OBJECT_LOCK(object);
-			vm_page_lock_queues();
 			for (; i < count; i++)
-				if (i != reqpage)
+				if (i != reqpage) {
+					vm_page_lock(m[i]);
+					vm_page_lock_queues();
 					vm_page_free(m[i]);
-			vm_page_unlock_queues();
+					vm_page_unlock_queues();
+					vm_page_unlock(m[i]);
+				}
 			VM_OBJECT_UNLOCK(object);
 			return (VM_PAGER_ERROR);
 		}
@@ -818,9 +839,11 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 				    (object->un_pager.vnp.vnp_size >> 32),
 				    (uintmax_t)object->un_pager.vnp.vnp_size);
 			}
+			vm_page_lock(m[i]);
 			vm_page_lock_queues();
 			vm_page_free(m[i]);
 			vm_page_unlock_queues();
+			vm_page_unlock(m[i]);
 			VM_OBJECT_UNLOCK(object);
 			runend = i + 1;
 			first = runend;
@@ -829,18 +852,24 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		runend = i + runpg;
 		if (runend <= reqpage) {
 			VM_OBJECT_LOCK(object);
-			vm_page_lock_queues();
-			for (j = i; j < runend; j++)
+			for (j = i; j < runend; j++) {
+				vm_page_lock(m[j]);
+				vm_page_lock_queues();
 				vm_page_free(m[j]);
-			vm_page_unlock_queues();
+				vm_page_unlock_queues();
+				vm_page_unlock(m[j]);
+			}
 			VM_OBJECT_UNLOCK(object);
 		} else {
 			if (runpg < (count - first)) {
 				VM_OBJECT_LOCK(object);
-				vm_page_lock_queues();
-				for (i = first + runpg; i < count; i++)
+				for (i = first + runpg; i < count; i++) {
+					vm_page_lock(m[i]);
+					vm_page_lock_queues();
 					vm_page_free(m[i]);
-				vm_page_unlock_queues();
+					vm_page_unlock_queues();
+					vm_page_unlock(m[i]);
+				}
 				VM_OBJECT_UNLOCK(object);
 				count = first + runpg;
 			}
@@ -931,13 +960,14 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 	relpbuf(bp, &vnode_pbuf_freecnt);
 
 	VM_OBJECT_LOCK(object);
-	vm_page_lock_queues();
 	for (i = 0, tfoff = foff; i < count; i++, tfoff = nextoff) {
 		vm_page_t mt;
 
 		nextoff = tfoff + PAGE_SIZE;
 		mt = m[i];
 
+		vm_page_lock(mt);
+		vm_page_lock_queues();
 		if (nextoff <= object->un_pager.vnp.vnp_size) {
 			/*
 			 * Read filled up entire page.
@@ -989,8 +1019,9 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 				vm_page_free(mt);
 			}
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(mt);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(object);
 	if (error) {
 		printf("vnode_pager_getpages: I/O read error\n");
@@ -1113,10 +1144,12 @@ vnode_pager_generic_putpages(vp, m, bytecount, flags, rtvals)
 			maxsize = object->un_pager.vnp.vnp_size - poffset;
 			ncount = btoc(maxsize);
 			if ((pgoff = (int)maxsize & PAGE_MASK) != 0) {
+				vm_page_lock(m[ncount - 1]);
 				vm_page_lock_queues();
 				vm_page_clear_dirty(m[ncount - 1], pgoff,
 					PAGE_SIZE - pgoff);
 				vm_page_unlock_queues();
+				vm_page_unlock(m[ncount - 1]);
 			}
 		} else {
 			maxsize = 0;

From 7c409214bffeac18800ff4040d73bcc5ab19bcac Mon Sep 17 00:00:00 2001
From: David Christensen <davidch@FreeBSD.org>
Date: Fri, 30 Apr 2010 02:35:46 +0000
Subject: [PATCH 344/532] - Enable flow control. - Print device details only
 when verbose boot is enabled. - Add debug output for shared memory access. -
 Add debug statistics (checksum offload & VLAN frame counters). - Modify TX
 path to update consumer index for each frame completed   rather than updating
 the consumer index only once for a group of   frames to improve small packet
 performance. - Print driver/firmware pulse messages only when verbose boot  
 is enabled. - Add debug sysctl to clear statistics. - Fix more style(9)
 violations.

MFC after:	2 weeks
---
 sys/dev/bce/if_bce.c    | 1131 ++++++++++++++++++++++-----------------
 sys/dev/bce/if_bcereg.h |  608 +++++++++++----------
 2 files changed, 955 insertions(+), 784 deletions(-)

diff --git a/sys/dev/bce/if_bce.c b/sys/dev/bce/if_bce.c
index 2ad5ab2d37f..6608dbde2c1 100644
--- a/sys/dev/bce/if_bce.c
+++ b/sys/dev/bce/if_bce.c
@@ -303,7 +303,7 @@ static void bce_dump_txbd		(struct bce_softc *,
 static void bce_dump_rxbd		(struct bce_softc *,
     int, struct rx_bd *);
 #ifdef BCE_JUMBO_HDRSPLIT
-static void bce_dump_pgbd		(struct bce_softc *, 
+static void bce_dump_pgbd		(struct bce_softc *,
     int, struct rx_bd *);
 #endif
 static void bce_dump_l2fhdr		(struct bce_softc *,
@@ -368,7 +368,7 @@ static int  bce_nvram_write		(struct bce_softc *, u32, u8 *, int);
 /****************************************************************************/
 static void bce_get_media		(struct bce_softc *);
 static void bce_init_media		(struct bce_softc *);
-static void bce_dma_map_addr		(void *, 
+static void bce_dma_map_addr		(void *,
     bus_dma_segment_t *, int, int);
 static int  bce_dma_alloc		(device_t);
 static void bce_dma_free		(struct bce_softc *);
@@ -379,7 +379,7 @@ static void bce_release_resources	(struct bce_softc *);
 /****************************************************************************/
 static int  bce_fw_sync			(struct bce_softc *, u32);
 static void bce_load_rv2p_fw		(struct bce_softc *, u32 *, u32, u32);
-static void bce_load_cpu_fw		(struct bce_softc *, 
+static void bce_load_cpu_fw		(struct bce_softc *,
     struct cpu_reg *, struct fw_info *);
 static void bce_start_cpu		(struct bce_softc *, struct cpu_reg *);
 static void bce_halt_cpu		(struct bce_softc *, struct cpu_reg *);
@@ -401,21 +401,21 @@ static int  bce_blockinit 		(struct bce_softc *);
 static int  bce_init_tx_chain		(struct bce_softc *);
 static void bce_free_tx_chain		(struct bce_softc *);
 
-static int  bce_get_rx_buf		(struct bce_softc *, 
+static int  bce_get_rx_buf		(struct bce_softc *,
     struct mbuf *, u16 *, u16 *, u32 *);
 static int  bce_init_rx_chain		(struct bce_softc *);
 static void bce_fill_rx_chain		(struct bce_softc *);
 static void bce_free_rx_chain		(struct bce_softc *);
 
 #ifdef BCE_JUMBO_HDRSPLIT
-static int  bce_get_pg_buf		(struct bce_softc *, 
+static int  bce_get_pg_buf		(struct bce_softc *,
     struct mbuf *, u16 *, u16 *);
 static int  bce_init_pg_chain		(struct bce_softc *);
 static void bce_fill_pg_chain		(struct bce_softc *);
 static void bce_free_pg_chain		(struct bce_softc *);
 #endif
 
-static struct mbuf *bce_tso_setup	(struct bce_softc *, 
+static struct mbuf *bce_tso_setup	(struct bce_softc *,
     struct mbuf **, u16 *);
 static int  bce_tx_encap		(struct bce_softc *, struct mbuf **);
 static void bce_start_locked		(struct ifnet *);
@@ -566,7 +566,7 @@ bce_probe(device_t dev)
 
 			/* Print out the device identity. */
 			snprintf(descbuf, BCE_DEVDESC_MAX, "%s (%c%d)",
-			    t->bce_name, (((pci_read_config(dev, 
+			    t->bce_name, (((pci_read_config(dev,
 			    PCIR_REVID, 4) & 0xf0) >> 4) + 'A'),
 			    (pci_read_config(dev, PCIR_REVID, 4) & 0xf));
 
@@ -593,57 +593,60 @@ bce_probe(device_t dev)
 static void
 bce_print_adapter_info(struct bce_softc *sc)
 {
-    int i = 0;
+	int i = 0;
 
 	DBENTER(BCE_VERBOSE_LOAD);
 
-	BCE_PRINTF("ASIC (0x%08X); ", sc->bce_chipid);
-	printf("Rev (%c%d); ", ((BCE_CHIP_ID(sc) & 0xf000) >> 12) + 'A',
-	    ((BCE_CHIP_ID(sc) & 0x0ff0) >> 4));
+	if (bootverbose) {
+		BCE_PRINTF("ASIC (0x%08X); ", sc->bce_chipid);
+		printf("Rev (%c%d); ", ((BCE_CHIP_ID(sc) & 0xf000) >>
+		    12) + 'A', ((BCE_CHIP_ID(sc) & 0x0ff0) >> 4));
 
-	/* Bus info. */
-	if (sc->bce_flags & BCE_PCIE_FLAG) {
-		printf("Bus (PCIe x%d, ", sc->link_width);
-		switch (sc->link_speed) {
-		case 1: printf("2.5Gbps); "); break;
-		case 2:	printf("5Gbps); "); break;
-		default: printf("Unknown link speed); ");
+
+		/* Bus info. */
+		if (sc->bce_flags & BCE_PCIE_FLAG) {
+			printf("Bus (PCIe x%d, ", sc->link_width);
+			switch (sc->link_speed) {
+			case 1: printf("2.5Gbps); "); break;
+			case 2:	printf("5Gbps); "); break;
+			default: printf("Unknown link speed); ");
+			}
+		} else {
+			printf("Bus (PCI%s, %s, %dMHz); ",
+			    ((sc->bce_flags & BCE_PCIX_FLAG) ? "-X" : ""),
+			    ((sc->bce_flags & BCE_PCI_32BIT_FLAG) ?
+			    "32-bit" : "64-bit"), sc->bus_speed_mhz);
 		}
-	} else {
-		printf("Bus (PCI%s, %s, %dMHz); ",
-		    ((sc->bce_flags & BCE_PCIX_FLAG) ? "-X" : ""),
-		    ((sc->bce_flags & BCE_PCI_32BIT_FLAG) ? 
-		    "32-bit" : "64-bit"), sc->bus_speed_mhz);
-	}
 
-	/* Firmware version and device features. */
-	printf("B/C (%s); Flags (", sc->bce_bc_ver);
+		/* Firmware version and device features. */
+		printf("B/C (%s); Flags (", sc->bce_bc_ver);
 
-#ifdef BCE_JUMBO_HDRSPLIT
-	printf("SPLT");
-	i++;
-#endif
+	#ifdef BCE_JUMBO_HDRSPLIT
+		printf("SPLT");
+		i++;
+	#endif
 
-	if (sc->bce_flags & BCE_USING_MSI_FLAG) {
-		if (i > 0) printf("|");
-		printf("MSI"); i++;
-	}
+		if (sc->bce_flags & BCE_USING_MSI_FLAG) {
+			if (i > 0) printf("|");
+			printf("MSI"); i++;
+		}
 
-	if (sc->bce_flags & BCE_USING_MSIX_FLAG) {
-		if (i > 0) printf("|");
-		printf("MSI-X"); i++;
-	}
+		if (sc->bce_flags & BCE_USING_MSIX_FLAG) {
+			if (i > 0) printf("|");
+			printf("MSI-X"); i++;
+		}
 
-	if (sc->bce_phy_flags & BCE_PHY_2_5G_CAPABLE_FLAG) {
-		if (i > 0) printf("|");
-		printf("2.5G"); i++;
-	}
+		if (sc->bce_phy_flags & BCE_PHY_2_5G_CAPABLE_FLAG) {
+			if (i > 0) printf("|");
+			printf("2.5G"); i++;
+		}
 
-	if (sc->bce_flags & BCE_MFW_ENABLE_FLAG) {
-		if (i > 0) printf("|");
-		printf("MFW); MFW (%s)\n", sc->bce_mfw_ver);
-	} else {
-		printf(")\n");
+		if (sc->bce_flags & BCE_MFW_ENABLE_FLAG) {
+			if (i > 0) printf("|");
+			printf("MFW); MFW (%s)\n", sc->bce_mfw_ver);
+		} else {
+			printf(")\n");
+		}
 	}
 
 	DBEXIT(BCE_VERBOSE_LOAD);
@@ -785,13 +788,13 @@ bce_attach(device_t dev)
 		(bce_msi_enable >= 1) && (sc->bce_msi_count == 0)) {
 		sc->bce_msi_count = 1;
 		if ((error = pci_alloc_msi(dev, &sc->bce_msi_count)) != 0) {
-			BCE_PRINTF("%s(%d): MSI allocation failed! error = %d\n",
-				__FILE__, __LINE__, error);
+			BCE_PRINTF("%s(%d): MSI allocation failed! "
+			    "error = %d\n", __FILE__, __LINE__, error);
 			sc->bce_msi_count = 0;
 			pci_release_msi(dev);
 		} else {
-			DBPRINT(sc, BCE_INFO_LOAD, "%s(): Using MSI interrupt.\n",
-				__FUNCTION__);
+			DBPRINT(sc, BCE_INFO_LOAD, "%s(): Using MSI "
+			    "interrupt.\n", __FUNCTION__);
 			sc->bce_flags |= BCE_USING_MSI_FLAG;
 			if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 				(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716))
@@ -848,10 +851,11 @@ bce_attach(device_t dev)
 	case BCE_CHIP_ID_5709_B0:
 	case BCE_CHIP_ID_5709_B1:
 	case BCE_CHIP_ID_5709_B2:
-		BCE_PRINTF("%s(%d): Unsupported controller revision (%c%d)!\n",
-		    __FILE__, __LINE__,
-		    (((pci_read_config(dev, PCIR_REVID, 4) & 0xf0) >> 4) + 'A'),
-		    (pci_read_config(dev, PCIR_REVID, 4) & 0xf));
+		BCE_PRINTF("%s(%d): Unsupported controller "
+		    "revision (%c%d)!\n", __FILE__, __LINE__,
+		    (((pci_read_config(dev, PCIR_REVID, 4) &
+		    0xf0) >> 4) + 'A'), (pci_read_config(dev,
+		    PCIR_REVID, 4) & 0xf));
 		rc = ENODEV;
 		goto bce_attach_fail;
 	}
@@ -1072,19 +1076,19 @@ bce_attach(device_t dev)
 	ifp = sc->bce_ifp = if_alloc(IFT_ETHER);
 	if (ifp == NULL) {
 		BCE_PRINTF("%s(%d): Interface allocation failed!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 		rc = ENXIO;
 		goto bce_attach_fail;
 	}
 
 	/* Initialize the ifnet interface. */
-	ifp->if_softc        = sc;
+	ifp->if_softc	= sc;
 	if_initname(ifp, device_get_name(dev), device_get_unit(dev));
-	ifp->if_flags        = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
-	ifp->if_ioctl        = bce_ioctl;
-	ifp->if_start        = bce_start;
-	ifp->if_init         = bce_init;
-	ifp->if_mtu          = ETHERMTU;
+	ifp->if_flags	= IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
+	ifp->if_ioctl	= bce_ioctl;
+	ifp->if_start	= bce_start;
+	ifp->if_init	= bce_init;
+	ifp->if_mtu	= ETHERMTU;
 
 	if (bce_tso_enable) {
 		ifp->if_hwassist = BCE_IF_HWASSIST | CSUM_TSO;
@@ -1095,7 +1099,7 @@ bce_attach(device_t dev)
 		ifp->if_capabilities = BCE_IF_CAPABILITIES;
 	}
 
-	ifp->if_capenable    = ifp->if_capabilities;
+	ifp->if_capenable = ifp->if_capabilities;
 
 	/*
 	 * Assume standard mbuf sizes for buffer allocation.
@@ -1105,16 +1109,17 @@ bce_attach(device_t dev)
 #ifdef BCE_JUMBO_HDRSPLIT
 	sc->rx_bd_mbuf_alloc_size = MHLEN;
 	/* Make sure offset is 16 byte aligned for hardware. */
-	sc->rx_bd_mbuf_align_pad  = roundup2((MSIZE - MHLEN), 16) -
-		(MSIZE - MHLEN);
-	sc->rx_bd_mbuf_data_len   = sc->rx_bd_mbuf_alloc_size -
-		sc->rx_bd_mbuf_align_pad;
+	sc->rx_bd_mbuf_align_pad =
+	    roundup2((MSIZE - MHLEN), 16) - (MSIZE - MHLEN);
+	sc->rx_bd_mbuf_data_len = sc->rx_bd_mbuf_alloc_size -
+	    sc->rx_bd_mbuf_align_pad;
 	sc->pg_bd_mbuf_alloc_size = MCLBYTES;
 #else
 	sc->rx_bd_mbuf_alloc_size = MCLBYTES;
-	sc->rx_bd_mbuf_align_pad  = roundup2(MCLBYTES, 16) - MCLBYTES;
-	sc->rx_bd_mbuf_data_len   = sc->rx_bd_mbuf_alloc_size -
-		sc->rx_bd_mbuf_align_pad;
+	sc->rx_bd_mbuf_align_pad =
+	    roundup2(MCLBYTES, 16) - MCLBYTES;
+	sc->rx_bd_mbuf_data_len = sc->rx_bd_mbuf_alloc_size -
+	    sc->rx_bd_mbuf_align_pad;
 #endif
 
 	ifp->if_snd.ifq_drv_maxlen = USABLE_TX_BD;
@@ -1126,14 +1131,14 @@ bce_attach(device_t dev)
 	else
 		ifp->if_baudrate = IF_Mbps(1000);
 
-    /* Handle any special PHY initialization for SerDes PHYs. */
-    bce_init_media(sc);
+	/* Handle any special PHY initialization for SerDes PHYs. */
+	bce_init_media(sc);
 
 	/* MII child bus by probing the PHY. */
 	if (mii_phy_probe(dev, &sc->bce_miibus, bce_ifmedia_upd,
 		bce_ifmedia_sts)) {
 		BCE_PRINTF("%s(%d): No PHY found on child MII bus!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 		rc = ENXIO;
 		goto bce_attach_fail;
 	}
@@ -1155,7 +1160,7 @@ bce_attach(device_t dev)
 
 	if (rc) {
 		BCE_PRINTF("%s(%d): Failed to setup IRQ!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 		bce_detach(dev);
 		goto bce_attach_exit;
 	}
@@ -1396,6 +1401,9 @@ bce_reg_wr_ind(struct bce_softc *sc, u32 offset, u32 val)
 static void
 bce_shmem_wr(struct bce_softc *sc, u32 offset, u32 val)
 {
+	DBPRINT(sc, BCE_VERBOSE_FIRMWARE, "%s(): Writing 0x%08X  to  "
+	    "0x%08X\n",	__FUNCTION__, val, offset);
+
 	bce_reg_wr_ind(sc, sc->bce_shmem_base + offset, val);
 }
 
@@ -1411,7 +1419,12 @@ bce_shmem_wr(struct bce_softc *sc, u32 offset, u32 val)
 static u32
 bce_shmem_rd(struct bce_softc *sc, u32 offset)
 {
-	return (bce_reg_rd_ind(sc, sc->bce_shmem_base + offset));
+	u32 val = bce_reg_rd_ind(sc, sc->bce_shmem_base + offset);
+
+	DBPRINT(sc, BCE_VERBOSE_FIRMWARE, "%s(): Reading 0x%08X from "
+	    "0x%08X\n",	__FUNCTION__, val, offset);
+
+	return val;
 }
 
 
@@ -1430,9 +1443,9 @@ bce_ctx_rd(struct bce_softc *sc, u32 cid_addr, u32 ctx_offset)
 {
 	u32 idx, offset, retry_cnt = 5, val;
 
-	DBRUNIF((cid_addr > MAX_CID_ADDR || ctx_offset & 0x3 || cid_addr & CTX_MASK),
-		BCE_PRINTF("%s(): Invalid CID address: 0x%08X.\n",
-			__FUNCTION__, cid_addr));
+	DBRUNIF((cid_addr > MAX_CID_ADDR || ctx_offset & 0x3 ||
+	    cid_addr & CTX_MASK), BCE_PRINTF("%s(): Invalid CID "
+	    "address: 0x%08X.\n", __FUNCTION__, cid_addr));
 
 	offset = ctx_offset + cid_addr;
 
@@ -1450,8 +1463,8 @@ bce_ctx_rd(struct bce_softc *sc, u32 cid_addr, u32 ctx_offset)
 
 		if (val & BCE_CTX_CTX_CTRL_READ_REQ)
 			BCE_PRINTF("%s(%d); Unable to read CTX memory: "
-				"cid_addr = 0x%08X, offset = 0x%08X!\n",
-				__FILE__, __LINE__, cid_addr, ctx_offset);
+			    "cid_addr = 0x%08X, offset = 0x%08X!\n",
+			    __FILE__, __LINE__, cid_addr, ctx_offset);
 
 		val = REG_RD(sc, BCE_CTX_CTX_DATA);
 	} else {
@@ -1487,7 +1500,7 @@ bce_ctx_wr(struct bce_softc *sc, u32 cid_addr, u32 ctx_offset, u32 ctx_val)
 
 	DBRUNIF((cid_addr > MAX_CID_ADDR || ctx_offset & 0x3 || cid_addr & CTX_MASK),
 		BCE_PRINTF("%s(): Invalid CID address: 0x%08X.\n",
-			__FUNCTION__, cid_addr));
+		    __FUNCTION__, cid_addr));
 
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
@@ -1504,8 +1517,8 @@ bce_ctx_wr(struct bce_softc *sc, u32 cid_addr, u32 ctx_offset, u32 ctx_val)
 
 		if (val & BCE_CTX_CTX_CTRL_WRITE_REQ)
 			BCE_PRINTF("%s(%d); Unable to write CTX memory: "
-				"cid_addr = 0x%08X, offset = 0x%08X!\n",
-				__FILE__, __LINE__, cid_addr, ctx_offset);
+			    "cid_addr = 0x%08X, offset = 0x%08X!\n",
+			    __FILE__, __LINE__, cid_addr, ctx_offset);
 
 	} else {
 		REG_WR(sc, BCE_CTX_DATA_ADR, offset);
@@ -1706,54 +1719,73 @@ bce_miibus_statchg(device_t dev)
 
 	val = REG_RD(sc, BCE_EMAC_MODE);
 	val &= ~(BCE_EMAC_MODE_PORT | BCE_EMAC_MODE_HALF_DUPLEX |
-		BCE_EMAC_MODE_MAC_LOOP | BCE_EMAC_MODE_FORCE_LINK |
-		BCE_EMAC_MODE_25G);
+	    BCE_EMAC_MODE_MAC_LOOP | BCE_EMAC_MODE_FORCE_LINK |
+	    BCE_EMAC_MODE_25G);
 
-	/* Set MII or GMII interface based on the speed negotiated by the PHY. */
+	/* Set MII or GMII interface based on the PHY speed. */
 	switch (IFM_SUBTYPE(mii->mii_media_active)) {
 	case IFM_10_T:
 		if (BCE_CHIP_NUM(sc) != BCE_CHIP_NUM_5706) {
-			DBPRINT(sc, BCE_INFO, "Enabling 10Mb interface.\n");
+			DBPRINT(sc, BCE_INFO_PHY,
+			    "Enabling 10Mb interface.\n");
 			val |= BCE_EMAC_MODE_PORT_MII_10;
 			break;
 		}
 		/* fall-through */
 	case IFM_100_TX:
-		DBPRINT(sc, BCE_INFO, "Enabling MII interface.\n");
+		DBPRINT(sc, BCE_INFO_PHY, "Enabling MII interface.\n");
 		val |= BCE_EMAC_MODE_PORT_MII;
 		break;
 	case IFM_2500_SX:
-		DBPRINT(sc, BCE_INFO, "Enabling 2.5G MAC mode.\n");
+		DBPRINT(sc, BCE_INFO_PHY, "Enabling 2.5G MAC mode.\n");
 		val |= BCE_EMAC_MODE_25G;
 		/* fall-through */
 	case IFM_1000_T:
 	case IFM_1000_SX:
-		DBPRINT(sc, BCE_INFO, "Enabling GMII interface.\n");
+		DBPRINT(sc, BCE_INFO_PHY, "Enabling GMII interface.\n");
 		val |= BCE_EMAC_MODE_PORT_GMII;
 		break;
 	default:
-		DBPRINT(sc, BCE_INFO, "Unknown speed, enabling default GMII "
-			"interface.\n");
+		DBPRINT(sc, BCE_INFO_PHY, "Unknown link speed, enabling "
+		    "default GMII interface.\n");
 		val |= BCE_EMAC_MODE_PORT_GMII;
 	}
 
-	/* Set half or full duplex based on the duplicity negotiated by the PHY. */
+	/* Set half or full duplex based on PHY settings. */
 	if ((mii->mii_media_active & IFM_GMASK) == IFM_HDX) {
-		DBPRINT(sc, BCE_INFO, "Setting Half-Duplex interface.\n");
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "Setting Half-Duplex interface.\n");
 		val |= BCE_EMAC_MODE_HALF_DUPLEX;
 	} else
-		DBPRINT(sc, BCE_INFO, "Setting Full-Duplex interface.\n");
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "Setting Full-Duplex interface.\n");
 
 	REG_WR(sc, BCE_EMAC_MODE, val);
 
-#if 0
-	/* ToDo: Enable flow control support in brgphy and bge. */
 	/* FLAG0 is set if RX is enabled and FLAG1 if TX is enabled */
-	if (mii->mii_media_active & IFM_FLAG0)
+ 	if (mii->mii_media_active & IFM_FLAG0) {
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "%s(): Enabling RX flow control.\n", __FUNCTION__);
 		BCE_SETBIT(sc, BCE_EMAC_RX_MODE, BCE_EMAC_RX_MODE_FLOW_EN);
-	if (mii->mii_media_active & IFM_FLAG1)
-		BCE_SETBIT(sc, BCE_EMAC_RX_MODE, BCE_EMAC_TX_MODE_FLOW_EN);
-#endif
+	} else {
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "%s(): Disabling RX flow control.\n", __FUNCTION__);
+		BCE_CLRBIT(sc, BCE_EMAC_RX_MODE, BCE_EMAC_RX_MODE_FLOW_EN);
+	}
+
+ 	if (mii->mii_media_active & IFM_FLAG1) {
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "%s(): Enabling TX flow control.\n", __FUNCTION__);
+		BCE_SETBIT(sc, BCE_EMAC_TX_MODE, BCE_EMAC_TX_MODE_FLOW_EN);
+		sc->bce_flags |= BCE_USING_TX_FLOW_CONTROL;
+	} else {
+		DBPRINT(sc, BCE_INFO_PHY,
+		    "%s(): Disabling TX flow control.\n", __FUNCTION__);
+		BCE_CLRBIT(sc, BCE_EMAC_TX_MODE, BCE_EMAC_TX_MODE_FLOW_EN);
+		sc->bce_flags &= ~BCE_USING_TX_FLOW_CONTROL;
+	}
+
+	/* ToDo: Update watermarks in bce_init_rx_context(). */
 
 	DBEXIT(BCE_VERBOSE_PHY);
 }
@@ -1926,8 +1958,8 @@ bce_enable_nvram_access(struct bce_softc *sc)
 
 	val = REG_RD(sc, BCE_NVM_ACCESS_ENABLE);
 	/* Enable both bits, even on read. */
-	REG_WR(sc, BCE_NVM_ACCESS_ENABLE,
-	       val | BCE_NVM_ACCESS_ENABLE_EN | BCE_NVM_ACCESS_ENABLE_WR_EN);
+	REG_WR(sc, BCE_NVM_ACCESS_ENABLE, val |
+	    BCE_NVM_ACCESS_ENABLE_EN | BCE_NVM_ACCESS_ENABLE_WR_EN);
 
 	DBEXIT(BCE_VERBOSE_NVRAM);
 }
@@ -1951,9 +1983,8 @@ bce_disable_nvram_access(struct bce_softc *sc)
 	val = REG_RD(sc, BCE_NVM_ACCESS_ENABLE);
 
 	/* Disable both bits, even after read. */
-	REG_WR(sc, BCE_NVM_ACCESS_ENABLE,
-		val & ~(BCE_NVM_ACCESS_ENABLE_EN |
-			BCE_NVM_ACCESS_ENABLE_WR_EN));
+	REG_WR(sc, BCE_NVM_ACCESS_ENABLE, val &
+	    ~(BCE_NVM_ACCESS_ENABLE_EN | BCE_NVM_ACCESS_ENABLE_WR_EN));
 
 	DBEXIT(BCE_VERBOSE_NVRAM);
 }
@@ -1983,7 +2014,7 @@ bce_nvram_erase_page(struct bce_softc *sc, u32 offset)
 
 	/* Build an erase command. */
 	cmd = BCE_NVM_COMMAND_ERASE | BCE_NVM_COMMAND_WR |
-	      BCE_NVM_COMMAND_DOIT;
+	    BCE_NVM_COMMAND_DOIT;
 
 	/*
 	 * Clear the DONE bit separately, set the NVRAM adress to erase,
@@ -2026,8 +2057,8 @@ bce_nvram_erase_page_exit:
 /*   0 on success and the 32 bit value read, positive value on failure.     */
 /****************************************************************************/
 static int
-bce_nvram_read_dword(struct bce_softc *sc, u32 offset, u8 *ret_val,
-							u32 cmd_flags)
+bce_nvram_read_dword(struct bce_softc *sc,
+    u32 offset, u8 *ret_val, u32 cmd_flags)
 {
 	u32 cmd;
 	int i, rc = 0;
@@ -2040,8 +2071,8 @@ bce_nvram_read_dword(struct bce_softc *sc, u32 offset, u8 *ret_val,
 	/* Calculate the offset for buffered flash if translation is used. */
 	if (sc->bce_flash_info->flags & BCE_NV_TRANSLATE) {
 		offset = ((offset / sc->bce_flash_info->page_size) <<
-			   sc->bce_flash_info->page_bits) +
-			  (offset % sc->bce_flash_info->page_size);
+		    sc->bce_flash_info->page_bits) +
+		    (offset % sc->bce_flash_info->page_size);
 	}
 
 	/*
@@ -2070,8 +2101,8 @@ bce_nvram_read_dword(struct bce_softc *sc, u32 offset, u8 *ret_val,
 
 	/* Check for errors. */
 	if (i >= NVRAM_TIMEOUT_COUNT) {
-		BCE_PRINTF("%s(%d): Timeout error reading NVRAM at offset 0x%08X!\n",
-			__FILE__, __LINE__, offset);
+		BCE_PRINTF("%s(%d): Timeout error reading NVRAM at "
+		    "offset 0x%08X!\n",	__FILE__, __LINE__, offset);
 		rc = EBUSY;
 	}
 
@@ -2106,8 +2137,8 @@ bce_nvram_write_dword(struct bce_softc *sc, u32 offset, u8 *val,
 	/* Calculate the offset for buffered flash if translation is used. */
 	if (sc->bce_flash_info->flags & BCE_NV_TRANSLATE) {
 		offset = ((offset / sc->bce_flash_info->page_size) <<
-			  sc->bce_flash_info->page_bits) +
-			 (offset % sc->bce_flash_info->page_size);
+		    sc->bce_flash_info->page_bits) +
+		    (offset % sc->bce_flash_info->page_size);
 	}
 
 	/*
@@ -2129,8 +2160,8 @@ bce_nvram_write_dword(struct bce_softc *sc, u32 offset, u8 *val,
 			break;
 	}
 	if (j >= NVRAM_TIMEOUT_COUNT) {
-		BCE_PRINTF("%s(%d): Timeout error writing NVRAM at offset 0x%08X\n",
-			__FILE__, __LINE__, offset);
+		BCE_PRINTF("%s(%d): Timeout error writing NVRAM at "
+		    "offset 0x%08X\n", __FILE__, __LINE__, offset);
 		rc = EBUSY;
 	}
 
@@ -2232,7 +2263,7 @@ bce_init_nvram(struct bce_softc *sc)
 	if (j == entry_count) {
 		sc->bce_flash_info = NULL;
 		BCE_PRINTF("%s(%d): Unknown Flash NVRAM found!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 		rc = ENODEV;
 	}
 
@@ -2246,8 +2277,8 @@ bce_init_nvram_get_flash_size:
 		sc->bce_flash_size = sc->bce_flash_info->total_size;
 
 	DBPRINT(sc, BCE_INFO_LOAD, "%s(): Found %s, size = 0x%08X\n",
-		__FUNCTION__, sc->bce_flash_info->name,
-		sc->bce_flash_info->total_size);
+	    __FUNCTION__, sc->bce_flash_info->name,
+	    sc->bce_flash_info->total_size);
 
 	DBEXIT(BCE_VERBOSE_NVRAM);
 	return rc;
@@ -2604,7 +2635,8 @@ bce_nvram_test(struct bce_softc *sc)
 	 * the magic value at offset 0.
 	 */
 	if ((rc = bce_nvram_read(sc, 0, data, 4)) != 0) {
-		BCE_PRINTF("%s(%d): Unable to read NVRAM!\n", __FILE__, __LINE__);
+		BCE_PRINTF("%s(%d): Unable to read NVRAM!\n",
+		    __FILE__, __LINE__);
 		goto bce_nvram_test_exit;
 	}
 
@@ -2615,9 +2647,9 @@ bce_nvram_test(struct bce_softc *sc)
     magic = bce_be32toh(buf[0]);
 	if (magic != BCE_NVRAM_MAGIC) {
 		rc = ENODEV;
-		BCE_PRINTF("%s(%d): Invalid NVRAM magic value! Expected: 0x%08X, "
-			"Found: 0x%08X\n",
-			__FILE__, __LINE__, BCE_NVRAM_MAGIC, magic);
+		BCE_PRINTF("%s(%d): Invalid NVRAM magic value! "
+		    "Expected: 0x%08X, Found: 0x%08X\n",
+		    __FILE__, __LINE__, BCE_NVRAM_MAGIC, magic);
 		goto bce_nvram_test_exit;
 	}
 
@@ -2626,26 +2658,27 @@ bce_nvram_test(struct bce_softc *sc)
 	 * configuration data.
 	 */
 	if ((rc = bce_nvram_read(sc, 0x100, data, BCE_NVRAM_SIZE)) != 0) {
-		BCE_PRINTF("%s(%d): Unable to read Manufacturing Information from "
-			"NVRAM!\n", __FILE__, __LINE__);
+		BCE_PRINTF("%s(%d): Unable to read manufacturing "
+		    "Information from  NVRAM!\n", __FILE__, __LINE__);
 		goto bce_nvram_test_exit;
 	}
 
 	csum = ether_crc32_le(data, 0x100);
 	if (csum != BCE_CRC32_RESIDUAL) {
 		rc = ENODEV;
-		BCE_PRINTF("%s(%d): Invalid Manufacturing Information NVRAM CRC! "
-			"Expected: 0x%08X, Found: 0x%08X\n",
-			__FILE__, __LINE__, BCE_CRC32_RESIDUAL, csum);
+		BCE_PRINTF("%s(%d): Invalid manufacturing information "
+		    "NVRAM CRC!	Expected: 0x%08X, Found: 0x%08X\n",
+		    __FILE__, __LINE__, BCE_CRC32_RESIDUAL, csum);
 		goto bce_nvram_test_exit;
 	}
 
 	csum = ether_crc32_le(data + 0x100, 0x100);
 	if (csum != BCE_CRC32_RESIDUAL) {
 		rc = ENODEV;
-		BCE_PRINTF("%s(%d): Invalid Feature Configuration Information "
-			"NVRAM CRC! Expected: 0x%08X, Found: 08%08X\n",
-			__FILE__, __LINE__, BCE_CRC32_RESIDUAL, csum);
+		BCE_PRINTF("%s(%d): Invalid feature configuration "
+		    "information NVRAM CRC! Expected: 0x%08X, "
+		    "Found: 08%08X\n", __FILE__, __LINE__,
+		    BCE_CRC32_RESIDUAL, csum);
 	}
 
 bce_nvram_test_exit:
@@ -2666,7 +2699,7 @@ bce_get_media(struct bce_softc *sc)
 {
 	u32 val;
 
-	DBENTER(BCE_VERBOSE);
+	DBENTER(BCE_VERBOSE_PHY);
 
 	/* Assume PHY address for copper controllers. */
 	sc->bce_phy_addr = 1;
@@ -2692,10 +2725,10 @@ bce_get_media(struct bce_softc *sc)
 		}
 
 		if (val & BCE_MISC_DUAL_MEDIA_CTRL_STRAP_OVERRIDE)
-			strap = (val & 
+			strap = (val &
 			    BCE_MISC_DUAL_MEDIA_CTRL_PHY_CTRL) >> 21;
 		else
-			strap = (val & 
+			strap = (val &
 			    BCE_MISC_DUAL_MEDIA_CTRL_PHY_CTRL_STRAP) >> 8;
 
 		if (pci_get_function(sc->bce_dev) == 0) {
@@ -2744,7 +2777,7 @@ bce_get_media(struct bce_softc *sc)
 
 			val = bce_shmem_rd(sc, BCE_SHARED_HW_CFG_CONFIG);
 			if (val & BCE_SHARED_HW_CFG_PHY_2_5G) {
-				sc->bce_phy_flags |= 
+				sc->bce_phy_flags |=
 				    BCE_PHY_2_5G_CAPABLE_FLAG;
 				DBPRINT(sc, BCE_INFO_LOAD, "Found 2.5Gb "
 				    "capable adapter\n");
@@ -2758,7 +2791,7 @@ bce_get_media_exit:
 	DBPRINT(sc, (BCE_INFO_LOAD | BCE_INFO_PHY),
 		"Using PHY address %d.\n", sc->bce_phy_addr);
 
-	DBEXIT(BCE_VERBOSE);
+	DBEXIT(BCE_VERBOSE_PHY);
 }
 
 
@@ -3056,7 +3089,9 @@ bce_dma_map_addr(void *arg, bus_dma_segment_t *segs, int nseg, int error)
 
 	/* Simulate a mapping failure. */
 	DBRUNIF(DB_RANDOMTRUE(dma_map_addr_failed_sim_control),
-		error = ENOMEM);
+	    error = ENOMEM);
+
+	/* ToDo: How to increment debug sim_count variable here? */
 
 	/* Check for an error and signal the caller that an error occurred. */
 	if (error) {
@@ -3154,7 +3189,7 @@ bce_dma_alloc(device_t dev)
 		goto bce_dma_alloc_exit;
 	}
 
-	DBPRINT(sc, BCE_INFO, "%s(): status_block_paddr = 0x%jX\n",
+	DBPRINT(sc, BCE_INFO_LOAD, "%s(): status_block_paddr = 0x%jX\n",
 	    __FUNCTION__, (uintmax_t) sc->status_block_paddr);
 
 	/*
@@ -3193,7 +3228,7 @@ bce_dma_alloc(device_t dev)
 		goto bce_dma_alloc_exit;
 	}
 
-	DBPRINT(sc, BCE_INFO, "%s(): stats_block_paddr = 0x%jX\n",
+	DBPRINT(sc, BCE_INFO_LOAD, "%s(): stats_block_paddr = 0x%jX\n",
 	    __FUNCTION__, (uintmax_t) sc->stats_block_paddr);
 
 	/* BCM5709 uses host memory as cache for context memory. */
@@ -3217,8 +3252,8 @@ bce_dma_alloc(device_t dev)
 		    BCE_DMA_BOUNDARY, sc->max_bus_addr,	BUS_SPACE_MAXADDR,
 		    NULL, NULL,	BCM_PAGE_SIZE, 1, BCM_PAGE_SIZE,
 		    0, NULL, NULL, &sc->ctx_tag)) {
-			BCE_PRINTF("%s(%d): Could not allocate CTX DMA tag!\n",
-			    __FILE__, __LINE__);
+			BCE_PRINTF("%s(%d): Could not allocate CTX "
+			    "DMA tag!\n", __FILE__, __LINE__);
 			rc = ENOMEM;
 			goto bce_dma_alloc_exit;
 		}
@@ -3248,8 +3283,9 @@ bce_dma_alloc(device_t dev)
 				goto bce_dma_alloc_exit;
 			}
 
-			DBPRINT(sc, BCE_INFO, "%s(): ctx_paddr[%d] = 0x%jX\n",
-			    __FUNCTION__, i, (uintmax_t) sc->ctx_paddr[i]);
+			DBPRINT(sc, BCE_INFO_LOAD, "%s(): ctx_paddr[%d] "
+			    "= 0x%jX\n", __FUNCTION__, i,
+			    (uintmax_t) sc->ctx_paddr[i]);
 		}
 	}
 
@@ -3262,15 +3298,15 @@ bce_dma_alloc(device_t dev)
 	    sc->max_bus_addr, BUS_SPACE_MAXADDR, NULL, NULL,
 	    BCE_TX_CHAIN_PAGE_SZ, 1, BCE_TX_CHAIN_PAGE_SZ, 0,
 	    NULL, NULL,	&sc->tx_bd_chain_tag)) {
-		BCE_PRINTF("%s(%d): Could not allocate TX descriptor chain "
-		    "DMA tag!\n", __FILE__, __LINE__);
+		BCE_PRINTF("%s(%d): Could not allocate TX descriptor "
+		    "chain DMA tag!\n", __FILE__, __LINE__);
 		rc = ENOMEM;
 		goto bce_dma_alloc_exit;
 	}
 
 	for (i = 0; i < TX_PAGES; i++) {
 
-		if(bus_dmamem_alloc(sc->tx_bd_chain_tag, 
+		if(bus_dmamem_alloc(sc->tx_bd_chain_tag,
 		    (void **)&sc->tx_bd_chain[i], BUS_DMA_NOWAIT,
 		    &sc->tx_bd_chain_map[i])) {
 			BCE_PRINTF("%s(%d): Could not allocate TX descriptor "
@@ -3291,8 +3327,9 @@ bce_dma_alloc(device_t dev)
 			goto bce_dma_alloc_exit;
 		}
 
-		DBPRINT(sc, BCE_INFO, "%s(): tx_bd_chain_paddr[%d] = 0x%jX\n",
-		    __FUNCTION__, i, (uintmax_t) sc->tx_bd_chain_paddr[i]);
+		DBPRINT(sc, BCE_INFO_LOAD, "%s(): tx_bd_chain_paddr[%d] = "
+		    "0x%jX\n", __FUNCTION__, i,
+		    (uintmax_t) sc->tx_bd_chain_paddr[i]);
 	}
 
 	/* Check the required size before mapping to conserve resources. */
@@ -3368,8 +3405,9 @@ bce_dma_alloc(device_t dev)
 			goto bce_dma_alloc_exit;
 		}
 
-		DBPRINT(sc, BCE_INFO, "%s(): rx_bd_chain_paddr[%d] = 0x%jX\n",
-		    __FUNCTION__, i, (uintmax_t) sc->rx_bd_chain_paddr[i]);
+		DBPRINT(sc, BCE_INFO_LOAD, "%s(): rx_bd_chain_paddr[%d] = "
+		    "0x%jX\n", __FUNCTION__, i,
+		    (uintmax_t) sc->rx_bd_chain_paddr[i]);
 	}
 
 	/*
@@ -3383,9 +3421,10 @@ bce_dma_alloc(device_t dev)
 #endif
 	max_segments = 1;
 
-	DBPRINT(sc, BCE_INFO, "%s(): Creating rx_mbuf_tag (max size = 0x%jX "
-	    "max segments = %d, max segment size = 0x%jX)\n", __FUNCTION__,
-	    (uintmax_t) max_size, max_segments, (uintmax_t) max_seg_size);
+	DBPRINT(sc, BCE_INFO_LOAD, "%s(): Creating rx_mbuf_tag "
+	    "(max size = 0x%jX max segments = %d, max segment "
+	    "size = 0x%jX)\n", __FUNCTION__, (uintmax_t) max_size,
+	     max_segments, (uintmax_t) max_seg_size);
 
 	if (bus_dma_tag_create(sc->parent_tag, 1, BCE_DMA_BOUNDARY,
 	    sc->max_bus_addr, BUS_SPACE_MAXADDR, NULL, NULL, max_size,
@@ -3429,7 +3468,7 @@ bce_dma_alloc(device_t dev)
 		    (void **)&sc->pg_bd_chain[i], BUS_DMA_NOWAIT,
 		    &sc->pg_bd_chain_map[i])) {
 			BCE_PRINTF("%s(%d): Could not allocate page "
-			    "descriptor chain DMA memory!\n", 
+			    "descriptor chain DMA memory!\n",
 			    __FILE__, __LINE__);
 			rc = ENOMEM;
 			goto bce_dma_alloc_exit;
@@ -3437,7 +3476,7 @@ bce_dma_alloc(device_t dev)
 
 		bzero((char *)sc->pg_bd_chain[i], BCE_PG_CHAIN_PAGE_SZ);
 
-		error = bus_dmamap_load(sc->pg_bd_chain_tag, 
+		error = bus_dmamap_load(sc->pg_bd_chain_tag,
 		    sc->pg_bd_chain_map[i], sc->pg_bd_chain[i],
 		    BCE_PG_CHAIN_PAGE_SZ, bce_dma_map_addr,
 		    &sc->pg_bd_chain_paddr[i], BUS_DMA_NOWAIT);
@@ -3449,8 +3488,9 @@ bce_dma_alloc(device_t dev)
 			goto bce_dma_alloc_exit;
 		}
 
-		DBPRINT(sc, BCE_INFO, "%s(): pg_bd_chain_paddr[%d] = 0x%jX\n",
-		    __FUNCTION__, i, (uintmax_t) sc->pg_bd_chain_paddr[i]);
+		DBPRINT(sc, BCE_INFO_LOAD, "%s(): pg_bd_chain_paddr[%d] = "
+		    "0x%jX\n", __FUNCTION__, i,
+		    (uintmax_t) sc->pg_bd_chain_paddr[i]);
 	}
 
 	/*
@@ -3524,7 +3564,7 @@ bce_release_resources(struct bce_softc *sc)
 
 	if (sc->bce_res_mem != NULL) {
 		DBPRINT(sc, BCE_INFO_RESET, "Releasing PCI memory.\n");
-		    bus_release_resource(dev, SYS_RES_MEMORY, PCIR_BAR(0), 
+		    bus_release_resource(dev, SYS_RES_MEMORY, PCIR_BAR(0),
 		    sc->bce_res_mem);
 	}
 
@@ -3582,7 +3622,7 @@ bce_fw_sync(struct bce_softc *sc, u32 msg_data)
 		DELAY(1000);
 	}
 
-	/* If we've timed out, tell the bootcode that we've stopped waiting. */
+	/* If we've timed out, tell bootcode that we've stopped waiting. */
 	if (((val & BCE_FW_MSG_ACK) != (msg_data & BCE_DRV_MSG_SEQ)) &&
 	    ((msg_data & BCE_DRV_MSG_DATA) != BCE_DRV_MSG_DATA_WAIT0)) {
 
@@ -4319,22 +4359,22 @@ bce_init_cpus(struct bce_softc *sc)
 		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 
 		if ((BCE_CHIP_REV(sc) == BCE_CHIP_REV_Ax)) {
-			bce_load_rv2p_fw(sc, bce_xi90_rv2p_proc1, 
-				sizeof(bce_xi90_rv2p_proc1), RV2P_PROC1);
-			bce_load_rv2p_fw(sc, bce_xi90_rv2p_proc2, 
-				sizeof(bce_xi90_rv2p_proc2), RV2P_PROC2);
+			bce_load_rv2p_fw(sc, bce_xi90_rv2p_proc1,
+			    sizeof(bce_xi90_rv2p_proc1), RV2P_PROC1);
+			bce_load_rv2p_fw(sc, bce_xi90_rv2p_proc2,
+			    sizeof(bce_xi90_rv2p_proc2), RV2P_PROC2);
 		} else {
-			bce_load_rv2p_fw(sc, bce_xi_rv2p_proc1, 
-				sizeof(bce_xi_rv2p_proc1), RV2P_PROC1);
-			bce_load_rv2p_fw(sc, bce_xi_rv2p_proc2, 
-				sizeof(bce_xi_rv2p_proc2), RV2P_PROC2);
+			bce_load_rv2p_fw(sc, bce_xi_rv2p_proc1,
+			    sizeof(bce_xi_rv2p_proc1), RV2P_PROC1);
+			bce_load_rv2p_fw(sc, bce_xi_rv2p_proc2,
+			    sizeof(bce_xi_rv2p_proc2), RV2P_PROC2);
 		}
 
 	} else {
-		bce_load_rv2p_fw(sc, bce_rv2p_proc1, 
-			sizeof(bce_rv2p_proc1),	RV2P_PROC1);
+		bce_load_rv2p_fw(sc, bce_rv2p_proc1,
+		    sizeof(bce_rv2p_proc1), RV2P_PROC1);
 		bce_load_rv2p_fw(sc, bce_rv2p_proc2,
-			sizeof(bce_rv2p_proc2),	RV2P_PROC2);
+		    sizeof(bce_rv2p_proc2), RV2P_PROC2);
 	}
 
 	bce_init_rxp_cpu(sc);
@@ -4373,7 +4413,7 @@ bce_init_ctx(struct bce_softc *sc)
 		 * in host memory so prepare the host memory
 		 * for access.
 		 */
-		val = BCE_CTX_COMMAND_ENABLED | 
+		val = BCE_CTX_COMMAND_ENABLED |
 		    BCE_CTX_COMMAND_MEM_INIT | (1 << 12);
 		val |= (BCM_PAGE_BITS - 8) << 16;
 		REG_WR(sc, BCE_CTX_COMMAND, val);
@@ -4406,7 +4446,7 @@ bce_init_ctx(struct bce_softc *sc)
 			/* Verify the context memory write was successful. */
 			for (j = 0; j < retry_cnt; j++) {
 				val = REG_RD(sc, BCE_CTX_HOST_PAGE_TBL_CTRL);
-				if ((val & 
+				if ((val &
 				    BCE_CTX_HOST_PAGE_TBL_CTRL_WRITE_REQ) == 0)
 					break;
 				DELAY(5);
@@ -4461,6 +4501,7 @@ bce_get_mac_addr(struct bce_softc *sc)
 	u32 mac_lo = 0, mac_hi = 0;
 
 	DBENTER(BCE_VERBOSE_RESET);
+
 	/*
 	 * The NetXtreme II bootcode populates various NIC
 	 * power-on and runtime configuration items in a
@@ -4475,7 +4516,7 @@ bce_get_mac_addr(struct bce_softc *sc)
 
 	if ((mac_lo == 0) && (mac_hi == 0)) {
 		BCE_PRINTF("%s(%d): Invalid Ethernet address!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 	} else {
 		sc->eaddr[0] = (u_char)(mac_hi >> 8);
 		sc->eaddr[1] = (u_char)(mac_hi >> 0);
@@ -4485,7 +4526,8 @@ bce_get_mac_addr(struct bce_softc *sc)
 		sc->eaddr[5] = (u_char)(mac_lo >> 0);
 	}
 
-	DBPRINT(sc, BCE_INFO_MISC, "Permanent Ethernet address = %6D\n", sc->eaddr, ":");
+	DBPRINT(sc, BCE_INFO_MISC, "Permanent Ethernet "
+	    "address = %6D\n", sc->eaddr, ":");
 	DBEXIT(BCE_VERBOSE_RESET);
 }
 
@@ -4505,14 +4547,15 @@ bce_set_mac_addr(struct bce_softc *sc)
 	/* ToDo: Add support for setting multiple MAC addresses. */
 
 	DBENTER(BCE_VERBOSE_RESET);
-	DBPRINT(sc, BCE_INFO_MISC, "Setting Ethernet address = %6D\n", sc->eaddr, ":");
+	DBPRINT(sc, BCE_INFO_MISC, "Setting Ethernet address = "
+	    "%6D\n", sc->eaddr, ":");
 
 	val = (mac_addr[0] << 8) | mac_addr[1];
 
 	REG_WR(sc, BCE_EMAC_MAC_MATCH0, val);
 
 	val = (mac_addr[2] << 24) | (mac_addr[3] << 16) |
-		(mac_addr[4] << 8) | mac_addr[5];
+	    (mac_addr[4] << 8) | mac_addr[5];
 
 	REG_WR(sc, BCE_EMAC_MAC_MATCH1, val);
 
@@ -4598,20 +4641,20 @@ bce_reset(struct bce_softc *sc, u32 reset_code)
 	DBENTER(BCE_VERBOSE_RESET);
 
 	DBPRINT(sc, BCE_VERBOSE_RESET, "%s(): reset_code = 0x%08X\n",
-		__FUNCTION__, reset_code);
+	    __FUNCTION__, reset_code);
 
 	/* Wait for pending PCI transactions to complete. */
 	REG_WR(sc, BCE_MISC_ENABLE_CLR_BITS,
-	       BCE_MISC_ENABLE_CLR_BITS_TX_DMA_ENABLE |
-	       BCE_MISC_ENABLE_CLR_BITS_DMA_ENGINE_ENABLE |
-	       BCE_MISC_ENABLE_CLR_BITS_RX_DMA_ENABLE |
-	       BCE_MISC_ENABLE_CLR_BITS_HOST_COALESCE_ENABLE);
+	    BCE_MISC_ENABLE_CLR_BITS_TX_DMA_ENABLE |
+	    BCE_MISC_ENABLE_CLR_BITS_DMA_ENGINE_ENABLE |
+	    BCE_MISC_ENABLE_CLR_BITS_RX_DMA_ENABLE |
+	    BCE_MISC_ENABLE_CLR_BITS_HOST_COALESCE_ENABLE);
 	val = REG_RD(sc, BCE_MISC_ENABLE_CLR_BITS);
 	DELAY(5);
 
 	/* Disable DMA */
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
-		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
+	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 		val = REG_RD(sc, BCE_MISC_NEW_CORE_CTL);
 		val &= ~BCE_MISC_NEW_CORE_CTL_DMA_ENABLE;
 		REG_WR(sc, BCE_MISC_NEW_CORE_CTL, val);
@@ -4634,26 +4677,26 @@ bce_reset(struct bce_softc *sc, u32 reset_code)
 
 	/* Chip reset. */
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
-		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
+	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 		REG_WR(sc, BCE_MISC_COMMAND, BCE_MISC_COMMAND_SW_RESET);
 		REG_RD(sc, BCE_MISC_COMMAND);
 		DELAY(5);
 
 		val = BCE_PCICFG_MISC_CONFIG_REG_WINDOW_ENA |
-		      BCE_PCICFG_MISC_CONFIG_TARGET_MB_WORD_SWAP;
+		    BCE_PCICFG_MISC_CONFIG_TARGET_MB_WORD_SWAP;
 
 		pci_write_config(sc->bce_dev, BCE_PCICFG_MISC_CONFIG, val, 4);
 	} else {
 		val = BCE_PCICFG_MISC_CONFIG_CORE_RST_REQ |
-			BCE_PCICFG_MISC_CONFIG_REG_WINDOW_ENA |
-			BCE_PCICFG_MISC_CONFIG_TARGET_MB_WORD_SWAP;
+		    BCE_PCICFG_MISC_CONFIG_REG_WINDOW_ENA |
+		    BCE_PCICFG_MISC_CONFIG_TARGET_MB_WORD_SWAP;
 		REG_WR(sc, BCE_PCICFG_MISC_CONFIG, val);
 
 		/* Allow up to 30us for reset to complete. */
 		for (i = 0; i < 10; i++) {
 			val = REG_RD(sc, BCE_PCICFG_MISC_CONFIG);
 			if ((val & (BCE_PCICFG_MISC_CONFIG_CORE_RST_REQ |
-				BCE_PCICFG_MISC_CONFIG_CORE_RST_BSY)) == 0) {
+			    BCE_PCICFG_MISC_CONFIG_CORE_RST_BSY)) == 0) {
 				break;
 			}
 			DELAY(10);
@@ -4661,9 +4704,9 @@ bce_reset(struct bce_softc *sc, u32 reset_code)
 
 		/* Check that reset completed successfully. */
 		if (val & (BCE_PCICFG_MISC_CONFIG_CORE_RST_REQ |
-			BCE_PCICFG_MISC_CONFIG_CORE_RST_BSY)) {
+		    BCE_PCICFG_MISC_CONFIG_CORE_RST_BSY)) {
 			BCE_PRINTF("%s(%d): Reset failed!\n",
-				__FILE__, __LINE__);
+			    __FILE__, __LINE__);
 			rc = EBUSY;
 			goto bce_reset_exit;
 		}
@@ -4673,7 +4716,7 @@ bce_reset(struct bce_softc *sc, u32 reset_code)
 	val = REG_RD(sc, BCE_PCI_SWAP_DIAG0);
 	if (val != 0x01020304) {
 		BCE_PRINTF("%s(%d): Byte swap is incorrect!\n",
-			__FILE__, __LINE__);
+		    __FILE__, __LINE__);
 		rc = ENODEV;
 		goto bce_reset_exit;
 	}
@@ -4685,8 +4728,8 @@ bce_reset(struct bce_softc *sc, u32 reset_code)
 	/* Wait for the firmware to finish its initialization. */
 	rc = bce_fw_sync(sc, BCE_DRV_MSG_DATA_WAIT1 | reset_code);
 	if (rc)
-		BCE_PRINTF("%s(%d): Firmware did not complete initialization!\n",
-			__FILE__, __LINE__);
+		BCE_PRINTF("%s(%d): Firmware did not complete "
+		    "initialization!\n", __FILE__, __LINE__);
 
 bce_reset_exit:
 	DBEXIT(BCE_VERBOSE_RESET);
@@ -4709,13 +4752,13 @@ bce_chipinit(struct bce_softc *sc)
 	 * channels and PCI clock compensation delay.
 	 */
 	val = BCE_DMA_CONFIG_DATA_BYTE_SWAP |
-	      BCE_DMA_CONFIG_DATA_WORD_SWAP |
+	    BCE_DMA_CONFIG_DATA_WORD_SWAP |
 #if BYTE_ORDER == BIG_ENDIAN
-	      BCE_DMA_CONFIG_CNTL_BYTE_SWAP |
+	    BCE_DMA_CONFIG_CNTL_BYTE_SWAP |
 #endif
-	      BCE_DMA_CONFIG_CNTL_WORD_SWAP |
-	      DMA_READ_CHANS << 12 |
-	      DMA_WRITE_CHANS << 16;
+	    BCE_DMA_CONFIG_CNTL_WORD_SWAP |
+	    DMA_READ_CHANS << 12 |
+	    DMA_WRITE_CHANS << 16;
 
 	val |= (0x2 << 20) | BCE_DMA_CONFIG_CNTL_PCI_COMP_DLY;
 
@@ -4765,7 +4808,7 @@ bce_chipinit(struct bce_softc *sc)
 
 	/* Enable bins used on the 5709. */
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
-		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
+	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 		val |= BCE_MQ_CONFIG_BIN_MQ_MODE;
 		if (BCE_CHIP_ID(sc) == BCE_CHIP_ID_5709_A1)
 			val |= BCE_MQ_CONFIG_HALT_DIS;
@@ -4927,7 +4970,7 @@ bce_blockinit(struct bce_softc *sc)
 	}
 
 	/* Allow bootcode to apply additional fixes before enabling MAC. */
-	rc = bce_fw_sync(sc, BCE_DRV_MSG_DATA_WAIT2 | 
+	rc = bce_fw_sync(sc, BCE_DRV_MSG_DATA_WAIT2 |
 	    BCE_DRV_MSG_CODE_RESET);
 
 	/* Enable link state change interrupt generation. */
@@ -4938,7 +4981,7 @@ bce_blockinit(struct bce_softc *sc)
 
 	/* Disable management frames (NC-SI) from flowing to the MCP. */
 	if (sc->bce_flags & BCE_MFW_ENABLE_FLAG) {
-		val = REG_RD(sc, BCE_RPM_MGMT_PKT_CTRL) & 
+		val = REG_RD(sc, BCE_RPM_MGMT_PKT_CTRL) &
 		    ~BCE_RPM_MGMT_PKT_CTRL_MGMT_EN;
 		REG_WR(sc, BCE_RPM_MGMT_PKT_CTRL, val);
 	}
@@ -4946,10 +4989,10 @@ bce_blockinit(struct bce_softc *sc)
 	/* Enable all remaining blocks in the MAC. */
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716))
-		REG_WR(sc, BCE_MISC_ENABLE_SET_BITS, 
+		REG_WR(sc, BCE_MISC_ENABLE_SET_BITS,
 		    BCE_MISC_ENABLE_DEFAULT_XI);
 	else
-		REG_WR(sc, BCE_MISC_ENABLE_SET_BITS, 
+		REG_WR(sc, BCE_MISC_ENABLE_SET_BITS,
 		    BCE_MISC_ENABLE_DEFAULT);
 
 	REG_RD(sc, BCE_MISC_ENABLE_SET_BITS);
@@ -4988,8 +5031,9 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 
 	/* Make sure the inputs are valid. */
 	DBRUNIF((*chain_prod > MAX_RX_BD),
-		BCE_PRINTF("%s(%d): RX producer out of range: 0x%04X > 0x%04X\n",
-		__FILE__, __LINE__, *chain_prod, (u16) MAX_RX_BD));
+	    BCE_PRINTF("%s(%d): RX producer out of range: "
+	    "0x%04X > 0x%04X\n", __FILE__, __LINE__,
+	    *chain_prod, (u16) MAX_RX_BD));
 
 	DBPRINT(sc, BCE_EXTREME_RECV, "%s(enter): prod = 0x%04X, "
 	    "chain_prod = 0x%04X, prod_bseq = 0x%08X\n", __FUNCTION__,
@@ -4997,8 +5041,9 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 
 	/* Update some debug statistic counters */
 	DBRUNIF((sc->free_rx_bd < sc->rx_low_watermark),
-		sc->rx_low_watermark = sc->free_rx_bd);
-	DBRUNIF((sc->free_rx_bd == sc->max_rx_bd), sc->rx_empty_count++);
+	    sc->rx_low_watermark = sc->free_rx_bd);
+	DBRUNIF((sc->free_rx_bd == sc->max_rx_bd),
+	    sc->rx_empty_count++);
 
 	/* Check whether this is a new mbuf allocation. */
 	if (m == NULL) {
@@ -5017,7 +5062,7 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 		if (sc->rx_bd_mbuf_alloc_size <= MCLBYTES)
 			m_new = m_getcl(M_DONTWAIT, MT_DATA, M_PKTHDR);
 		else
-			m_new = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR, 
+			m_new = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR,
 			    sc->rx_bd_mbuf_alloc_size);
 #endif
 
@@ -5049,8 +5094,8 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 
 	/* Handle any mapping errors. */
 	if (error) {
-		BCE_PRINTF("%s(%d): Error mapping mbuf into RX chain (%d)!\n",
-		    __FILE__, __LINE__, error);
+		BCE_PRINTF("%s(%d): Error mapping mbuf into RX "
+		    "chain (%d)!\n", __FILE__, __LINE__, error);
 
 		sc->dma_map_addr_rx_failed_count++;
 		m_freem(m_new);
@@ -5078,11 +5123,11 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 	sc->rx_mbuf_ptr[*chain_prod] = m_new;
 	sc->free_rx_bd -= nsegs;
 
-	DBRUNMSG(BCE_INSANE_RECV, 
+	DBRUNMSG(BCE_INSANE_RECV,
 	    bce_dump_rx_mbuf_chain(sc, debug_chain_prod, nsegs));
 
 	DBPRINT(sc, BCE_EXTREME_RECV, "%s(exit): prod = 0x%04X, "
-	    "chain_prod = 0x%04X, prod_bseq = 0x%08X\n", 
+	    "chain_prod = 0x%04X, prod_bseq = 0x%08X\n",
 	    __FUNCTION__, *prod, *chain_prod, *prod_bseq);
 
 bce_get_rx_buf_exit:
@@ -5116,8 +5161,9 @@ bce_get_pg_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 
 	/* Make sure the inputs are valid. */
 	DBRUNIF((*prod_idx > MAX_PG_BD),
-	    BCE_PRINTF("%s(%d): page producer out of range: 0x%04X > 0x%04X\n",
-	    __FILE__, __LINE__, *prod_idx, (u16) MAX_PG_BD));
+	    BCE_PRINTF("%s(%d): page producer out of range: "
+	    "0x%04X > 0x%04X\n", __FILE__, __LINE__,
+	    *prod_idx, (u16) MAX_PG_BD));
 
 	DBPRINT(sc, BCE_EXTREME_RECV, "%s(enter): prod = 0x%04X, "
 	    "chain_prod = 0x%04X\n", __FUNCTION__, *prod, *prod_idx);
@@ -5159,7 +5205,7 @@ bce_get_pg_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 	/* Map the mbuf cluster into device memory. */
 	map = sc->pg_mbuf_map[*prod_idx];
 	error = bus_dmamap_load(sc->pg_mbuf_tag, map, mtod(m_new, void *),
-	    sc->pg_bd_mbuf_alloc_size, bce_dma_map_addr, 
+	    sc->pg_bd_mbuf_alloc_size, bce_dma_map_addr,
 	    &busaddr, BUS_DMA_NOWAIT);
 
 	/* Handle any mapping errors. */
@@ -5191,7 +5237,7 @@ bce_get_pg_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 	sc->pg_mbuf_ptr[*prod_idx] = m_new;
 	sc->free_pg_bd--;
 
-	DBRUNMSG(BCE_INSANE_RECV, 
+	DBRUNMSG(BCE_INSANE_RECV,
 	    bce_dump_pg_mbuf_chain(sc, debug_prod_idx, 1));
 
 	DBPRINT(sc, BCE_EXTREME_RECV, "%s(exit): prod = 0x%04X, "
@@ -5222,19 +5268,19 @@ bce_init_tx_context(struct bce_softc *sc)
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 		(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 		/* Set the CID type to support an L2 connection. */
-		val = BCE_L2CTX_TX_TYPE_TYPE_L2_XI | 
+		val = BCE_L2CTX_TX_TYPE_TYPE_L2_XI |
 		    BCE_L2CTX_TX_TYPE_SIZE_L2_XI;
 		CTX_WR(sc, GET_CID_ADDR(TX_CID), BCE_L2CTX_TX_TYPE_XI, val);
 		val = BCE_L2CTX_TX_CMD_TYPE_TYPE_L2_XI | (8 << 16);
-		CTX_WR(sc, GET_CID_ADDR(TX_CID), 
+		CTX_WR(sc, GET_CID_ADDR(TX_CID),
 		    BCE_L2CTX_TX_CMD_TYPE_XI, val);
 
 		/* Point the hardware to the first page in the chain. */
 		val = BCE_ADDR_HI(sc->tx_bd_chain_paddr[0]);
-		CTX_WR(sc, GET_CID_ADDR(TX_CID), 
+		CTX_WR(sc, GET_CID_ADDR(TX_CID),
 		    BCE_L2CTX_TX_TBDR_BHADDR_HI_XI, val);
 		val = BCE_ADDR_LO(sc->tx_bd_chain_paddr[0]);
-		CTX_WR(sc, GET_CID_ADDR(TX_CID), 
+		CTX_WR(sc, GET_CID_ADDR(TX_CID),
 		    BCE_L2CTX_TX_TBDR_BHADDR_LO_XI, val);
 	} else {
 		/* Set the CID type to support an L2 connection. */
@@ -5245,10 +5291,10 @@ bce_init_tx_context(struct bce_softc *sc)
 
 		/* Point the hardware to the first page in the chain. */
 		val = BCE_ADDR_HI(sc->tx_bd_chain_paddr[0]);
-		CTX_WR(sc, GET_CID_ADDR(TX_CID), 
+		CTX_WR(sc, GET_CID_ADDR(TX_CID),
 		    BCE_L2CTX_TX_TBDR_BHADDR_HI, val);
 		val = BCE_ADDR_LO(sc->tx_bd_chain_paddr[0]);
-		CTX_WR(sc, GET_CID_ADDR(TX_CID), 
+		CTX_WR(sc, GET_CID_ADDR(TX_CID),
 		    BCE_L2CTX_TX_TBDR_BHADDR_LO, val);
 	}
 
@@ -5331,7 +5377,7 @@ bce_free_tx_chain(struct bce_softc *sc)
 	for (i = 0; i < TOTAL_TX_BD; i++) {
 		if (sc->tx_mbuf_ptr[i] != NULL) {
 			if (sc->tx_mbuf_map[i] != NULL)
-				bus_dmamap_sync(sc->tx_mbuf_tag, 
+				bus_dmamap_sync(sc->tx_mbuf_tag,
 				    sc->tx_mbuf_map[i],
 				    BUS_DMASYNC_POSTWRITE);
 			m_freem(sc->tx_mbuf_ptr[i]);
@@ -5349,7 +5395,7 @@ bce_free_tx_chain(struct bce_softc *sc)
 	/* Check if we lost any mbufs in the process. */
 	DBRUNIF((sc->debug_tx_mbuf_alloc),
 	    BCE_PRINTF("%s(%d): Memory leak! Lost %d mbufs "
-	    "from tx chain!\n",	__FILE__, __LINE__, 
+	    "from tx chain!\n",	__FILE__, __LINE__,
 	    sc->debug_tx_mbuf_alloc));
 
 	DBEXIT(BCE_VERBOSE_RESET | BCE_VERBOSE_SEND | BCE_VERBOSE_UNLOAD);
@@ -5385,9 +5431,22 @@ bce_init_rx_context(struct bce_softc *sc)
 	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 		u32 lo_water, hi_water;
 
-		lo_water = BCE_L2CTX_RX_LO_WATER_MARK_DEFAULT;
+		if (sc->bce_flags && BCE_USING_TX_FLOW_CONTROL) {
+			lo_water = BCE_L2CTX_RX_LO_WATER_MARK_DEFAULT;
+		} else {
+			lo_water = 0;
+		}
+
+		if (lo_water >= USABLE_RX_BD) {
+			lo_water = 0;
+		}
+
 		hi_water = USABLE_RX_BD / 4;
 
+		if (hi_water <= lo_water) {
+			lo_water = 0;
+		}
+
 		lo_water /= BCE_L2CTX_RX_LO_WATER_MARK_SCALE;
 		hi_water /= BCE_L2CTX_RX_HI_WATER_MARK_SCALE;
 
@@ -5395,11 +5454,12 @@ bce_init_rx_context(struct bce_softc *sc)
 			hi_water = 0xf;
 		else if (hi_water == 0)
 			lo_water = 0;
+
 		val |= (lo_water << BCE_L2CTX_RX_LO_WATER_MARK_SHIFT) |
-			(hi_water << BCE_L2CTX_RX_HI_WATER_MARK_SHIFT);
+		    (hi_water << BCE_L2CTX_RX_HI_WATER_MARK_SHIFT);
 	}
 
- 	CTX_WR(sc, GET_CID_ADDR(RX_CID), BCE_L2CTX_RX_CTX_TYPE, val);
+	CTX_WR(sc, GET_CID_ADDR(RX_CID), BCE_L2CTX_RX_CTX_TYPE, val);
 
 	/* Setup the MQ BIN mapping for l2_ctx_host_bseq. */
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
@@ -5453,9 +5513,9 @@ bce_init_rx_chain(struct bce_softc *sc)
 			j = i + 1;
 
 		/* Setup the chain page pointers. */
-		rxbd->rx_bd_haddr_hi = 
+		rxbd->rx_bd_haddr_hi =
 		    htole32(BCE_ADDR_HI(sc->rx_bd_chain_paddr[j]));
-		rxbd->rx_bd_haddr_lo = 
+		rxbd->rx_bd_haddr_lo =
 		    htole32(BCE_ADDR_LO(sc->rx_bd_chain_paddr[j]));
 	}
 
@@ -5517,14 +5577,14 @@ bce_fill_rx_chain(struct bce_softc *sc)
 
 	/* We should never end up pointing to a next page pointer. */
 	DBRUNIF(((prod & USABLE_RX_BD_PER_PAGE) == USABLE_RX_BD_PER_PAGE),
-		BCE_PRINTF("%s(): Invalid rx_prod value: 0x%04X\n",
-		__FUNCTION__, sc->rx_prod));
+	    BCE_PRINTF("%s(): Invalid rx_prod value: 0x%04X\n",
+	    __FUNCTION__, sc->rx_prod));
 
 	/* Write the mailbox and tell the chip about the waiting rx_bd's. */
-	REG_WR16(sc, MB_GET_CID_ADDR(RX_CID) + BCE_L2MQ_RX_HOST_BDIDX,
-		sc->rx_prod);
-	REG_WR(sc, MB_GET_CID_ADDR(RX_CID) + BCE_L2MQ_RX_HOST_BSEQ,
-		sc->rx_prod_bseq);
+	REG_WR16(sc, MB_GET_CID_ADDR(RX_CID) +
+	    BCE_L2MQ_RX_HOST_BDIDX, sc->rx_prod);
+	REG_WR(sc, MB_GET_CID_ADDR(RX_CID) +
+	    BCE_L2MQ_RX_HOST_BSEQ, sc->rx_prod_bseq);
 
 	DBEXIT(BCE_VERBOSE_RESET | BCE_EXTREME_RECV | BCE_VERBOSE_LOAD |
 	    BCE_VERBOSE_CTX);
@@ -5548,7 +5608,7 @@ bce_free_rx_chain(struct bce_softc *sc)
 	for (i = 0; i < TOTAL_RX_BD; i++) {
 		if (sc->rx_mbuf_ptr[i] != NULL) {
 			if (sc->rx_mbuf_map[i] != NULL)
-				bus_dmamap_sync(sc->rx_mbuf_tag, 
+				bus_dmamap_sync(sc->rx_mbuf_tag,
 				    sc->rx_mbuf_map[i],
 				    BUS_DMASYNC_POSTREAD);
 			m_freem(sc->rx_mbuf_ptr[i]);
@@ -5560,7 +5620,7 @@ bce_free_rx_chain(struct bce_softc *sc)
 	/* Clear each RX chain page. */
 	for (i = 0; i < RX_PAGES; i++)
 		if (sc->rx_bd_chain[i] != NULL) {
-			bzero((char *)sc->rx_bd_chain[i], 
+			bzero((char *)sc->rx_bd_chain[i],
 			    BCE_RX_CHAIN_PAGE_SZ);
 		}
 
@@ -5667,7 +5727,7 @@ bce_fill_pg_chain(struct bce_softc *sc)
 	u16 prod, prod_idx;
 
 	DBENTER(BCE_VERBOSE_RESET | BCE_EXTREME_RECV | BCE_VERBOSE_LOAD |
-		BCE_VERBOSE_CTX);
+	    BCE_VERBOSE_CTX);
 
 	/* Get the page chain prodcuer index. */
 	prod = sc->pg_prod;
@@ -5686,18 +5746,18 @@ bce_fill_pg_chain(struct bce_softc *sc)
 	sc->pg_prod = prod;
 
 	DBRUNIF(((prod & USABLE_RX_BD_PER_PAGE) == USABLE_RX_BD_PER_PAGE),
-		BCE_PRINTF("%s(): Invalid pg_prod value: 0x%04X\n",
-		__FUNCTION__, sc->pg_prod));
+	    BCE_PRINTF("%s(): Invalid pg_prod value: 0x%04X\n",
+	    __FUNCTION__, sc->pg_prod));
 
 	/*
 	 * Write the mailbox and tell the chip about
 	 * the new rx_bd's in the page chain.
 	 */
-	REG_WR16(sc, MB_GET_CID_ADDR(RX_CID) + BCE_L2MQ_RX_HOST_PG_BDIDX,
-		sc->pg_prod);
+	REG_WR16(sc, MB_GET_CID_ADDR(RX_CID) +
+	    BCE_L2MQ_RX_HOST_PG_BDIDX, sc->pg_prod);
 
 	DBEXIT(BCE_VERBOSE_RESET | BCE_EXTREME_RECV | BCE_VERBOSE_LOAD |
-		BCE_VERBOSE_CTX);
+	    BCE_VERBOSE_CTX);
 }
 
 
@@ -5718,8 +5778,9 @@ bce_free_pg_chain(struct bce_softc *sc)
 	for (i = 0; i < TOTAL_PG_BD; i++) {
 		if (sc->pg_mbuf_ptr[i] != NULL) {
 			if (sc->pg_mbuf_map[i] != NULL)
-				bus_dmamap_sync(sc->pg_mbuf_tag, sc->pg_mbuf_map[i],
-					BUS_DMASYNC_POSTREAD);
+				bus_dmamap_sync(sc->pg_mbuf_tag,
+				    sc->pg_mbuf_map[i],
+				    BUS_DMASYNC_POSTREAD);
 			m_freem(sc->pg_mbuf_ptr[i]);
 			sc->pg_mbuf_ptr[i] = NULL;
 			DBRUN(sc->debug_pg_mbuf_alloc--);
@@ -5734,8 +5795,8 @@ bce_free_pg_chain(struct bce_softc *sc)
 
 	/* Check if we lost any mbufs in the process. */
 	DBRUNIF((sc->debug_pg_mbuf_alloc),
-		BCE_PRINTF("%s(): Memory leak! Lost %d mbufs from page chain!\n",
-			__FUNCTION__, sc->debug_pg_mbuf_alloc));
+	    BCE_PRINTF("%s(): Memory leak! Lost %d mbufs from page chain!\n",
+	    __FUNCTION__, sc->debug_pg_mbuf_alloc));
 
 	DBEXIT(BCE_VERBOSE_RESET | BCE_VERBOSE_RECV | BCE_VERBOSE_UNLOAD);
 }
@@ -5776,7 +5837,7 @@ bce_ifmedia_upd_locked(struct ifnet *ifp)
 	struct bce_softc *sc = ifp->if_softc;
 	struct mii_data *mii;
 
-	DBENTER(BCE_VERBOSE);
+	DBENTER(BCE_VERBOSE_PHY);
 
 	BCE_LOCK_ASSERT(sc);
 
@@ -5794,7 +5855,7 @@ bce_ifmedia_upd_locked(struct ifnet *ifp)
 		mii_mediachg(mii);
 	}
 
-	DBEXIT(BCE_VERBOSE);
+	DBEXIT(BCE_VERBOSE_PHY);
 }
 
 
@@ -5810,7 +5871,7 @@ bce_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr)
 	struct bce_softc *sc = ifp->if_softc;
 	struct mii_data *mii;
 
-	DBENTER(BCE_VERBOSE);
+	DBENTER(BCE_VERBOSE_PHY);
 
 	BCE_LOCK(sc);
 
@@ -5822,7 +5883,7 @@ bce_ifmedia_sts(struct ifnet *ifp, struct ifmediareq *ifmr)
 
 	BCE_UNLOCK(sc);
 
-	DBEXIT(BCE_VERBOSE);
+	DBEXIT(BCE_VERBOSE_PHY);
 }
 
 
@@ -5921,7 +5982,7 @@ bce_rx_intr(struct bce_softc *sc)
 #endif
 
 	DBENTER(BCE_VERBOSE_RECV | BCE_VERBOSE_INTR);
-	DBRUN(sc->rx_interrupts++);
+	DBRUN(sc->interrupts_rx++);
 	DBPRINT(sc, BCE_EXTREME_RECV, "%s(enter): rx_prod = 0x%04X, "
 	    "rx_cons = 0x%04X, rx_prod_bseq = 0x%08X\n",
 	    __FUNCTION__, sc->rx_prod, sc->rx_cons, sc->rx_prod_bseq);
@@ -5943,14 +6004,16 @@ bce_rx_intr(struct bce_softc *sc)
 
 	/* Get working copies of the driver's view of the consumer indices. */
 	sw_rx_cons = sc->rx_cons;
+
 #ifdef BCE_JUMBO_HDRSPLIT
 	sw_pg_cons = sc->pg_cons;
 #endif
 
 	/* Update some debug statistics counters */
 	DBRUNIF((sc->free_rx_bd < sc->rx_low_watermark),
-		sc->rx_low_watermark = sc->free_rx_bd);
-	DBRUNIF((sc->free_rx_bd == sc->max_rx_bd), sc->rx_empty_count++);
+	    sc->rx_low_watermark = sc->free_rx_bd);
+	DBRUNIF((sc->free_rx_bd == sc->max_rx_bd),
+	    sc->rx_empty_count++);
 
 	/* Scan through the receive chain as long as there is work to do */
 	/* ToDo: Consider setting a limit on the number of packets processed. */
@@ -5962,7 +6025,7 @@ bce_rx_intr(struct bce_softc *sc)
 		sw_rx_cons_idx = RX_CHAIN_IDX(sw_rx_cons);
 
 		/* Unmap the mbuf from DMA space. */
-		bus_dmamap_sync(sc->rx_mbuf_tag, 
+		bus_dmamap_sync(sc->rx_mbuf_tag,
 		    sc->rx_mbuf_map[sw_rx_cons_idx],
 		    BUS_DMASYNC_POSTREAD);
 		bus_dmamap_unload(sc->rx_mbuf_tag,
@@ -5975,7 +6038,7 @@ bce_rx_intr(struct bce_softc *sc)
 		sc->free_rx_bd++;
 
 		if(m0 == NULL) {
-			DBPRINT(sc, BCE_EXTREME_RECV, 
+			DBPRINT(sc, BCE_EXTREME_RECV,
 			    "%s(): Oops! Empty mbuf pointer "
 			    "found in sc->rx_mbuf_ptr[0x%04X]!\n",
 			    __FUNCTION__, sw_rx_cons_idx);
@@ -5983,20 +6046,23 @@ bce_rx_intr(struct bce_softc *sc)
 		}
 
 		/*
-		 * Frames received on the NetXteme II are prepended	with an
-		 * l2_fhdr structure which provides status information about
-		 * the received frame (including VLAN tags and checksum info).
-		 * The frames are also automatically adjusted to align the IP
-		 * header (i.e. two null bytes are inserted before the Ethernet
-		 * header).  As a result the data DMA'd by the controller into
-		 * the mbuf is as follows:
-		 * 
+ 		 * Frames received on the NetXteme II are prepended
+ 		 * with an l2_fhdr structure which provides status
+ 		 * information about the received frame (including
+ 		 * VLAN tags and checksum info).  The frames are
+		 * also automatically adjusted to align the IP
+ 		 * header (i.e. two null bytes are inserted before
+ 		 * the Ethernet	header).  As a result the data
+ 		 * DMA'd by the controller into	the mbuf looks
+		 * like this:
+		 *
 		 * +---------+-----+---------------------+-----+
 		 * | l2_fhdr | pad | packet data         | FCS |
 		 * +---------+-----+---------------------+-----+
-		 * 
-		 * The l2_fhdr needs to be checked and skipped and the FCS needs
-		 * to be stripped before sending the packet up the stack.
+		 *
+ 		 * The l2_fhdr needs to be checked and skipped and
+ 		 * the FCS needs to be stripped before sending the
+		 * packet up the stack.
 		 */
 		l2fhdr  = mtod(m0, struct l2_fhdr *);
 
@@ -6100,7 +6166,7 @@ bce_rx_intr(struct bce_softc *sc)
 			m0->m_pkthdr.len = m0->m_len = pkt_len;
 		}
 #else
-        /* Set the total packet length. */
+		/* Set the total packet length. */
 		m0->m_pkthdr.len = m0->m_len = pkt_len;
 #endif
 
@@ -6115,12 +6181,11 @@ bce_rx_intr(struct bce_softc *sc)
 		     m_print(m0, 128));
 
 		DBRUNIF(DB_RANDOMTRUE(l2fhdr_error_sim_control),
-		    BCE_PRINTF("Simulating l2_fhdr status error.\n");
 		    sc->l2fhdr_error_sim_count++;
 		    status = status | L2_FHDR_ERRORS_PHY_DECODE);
 
 		/* Check the received frame for errors. */
-		if (status & (L2_FHDR_ERRORS_BAD_CRC | 
+		if (status & (L2_FHDR_ERRORS_BAD_CRC |
 		    L2_FHDR_ERRORS_PHY_DECODE | L2_FHDR_ERRORS_ALIGNMENT |
 		    L2_FHDR_ERRORS_TOO_SHORT  | L2_FHDR_ERRORS_GIANT_FRAME)) {
 
@@ -6144,12 +6209,12 @@ bce_rx_intr(struct bce_softc *sc)
 
 			/* Check for an IP datagram. */
 		 	if (!(status & L2_FHDR_STATUS_SPLIT) &&
-				(status & L2_FHDR_STATUS_IP_DATAGRAM)) {
+			    (status & L2_FHDR_STATUS_IP_DATAGRAM)) {
 				m0->m_pkthdr.csum_flags |= CSUM_IP_CHECKED;
-
+				DBRUN(sc->csum_offload_ip++);
 				/* Check if the IP checksum is valid. */
 				if ((l2fhdr->l2_fhdr_ip_xsum ^ 0xffff) == 0)
-					m0->m_pkthdr.csum_flags |= 
+					m0->m_pkthdr.csum_flags |=
 					    CSUM_IP_VALID;
 			}
 
@@ -6160,9 +6225,10 @@ bce_rx_intr(struct bce_softc *sc)
 				/* Check for a good TCP/UDP checksum. */
 				if ((status & (L2_FHDR_ERRORS_TCP_XSUM |
 				    L2_FHDR_ERRORS_UDP_XSUM)) == 0) {
+					DBRUN(sc->csum_offload_tcp_udp++);
 					m0->m_pkthdr.csum_data =
 					    l2fhdr->l2_fhdr_tcp_udp_xsum;
-					m0->m_pkthdr.csum_flags |= 
+					m0->m_pkthdr.csum_flags |=
 					    (CSUM_DATA_VALID
 					    | CSUM_PSEUDO_HDR);
 				}
@@ -6234,12 +6300,13 @@ bce_rx_int_next_rx:
 			hw_rx_cons = sc->hw_rx_cons = bce_get_hw_rx_cons(sc);
 	}
 
-	/* No new packets to process.  Refill the RX and page chains and exit. */
 #ifdef BCE_JUMBO_HDRSPLIT
+	/* No new packets.  Refill the page chain. */
 	sc->pg_cons = sw_pg_cons;
 	bce_fill_pg_chain(sc);
 #endif
 
+	/* No new packets.  Refill the RX chain. */
 	sc->rx_cons = sw_rx_cons;
 	bce_fill_rx_chain(sc);
 
@@ -6295,7 +6362,7 @@ bce_tx_intr(struct bce_softc *sc)
 	u16 hw_tx_cons, sw_tx_cons, sw_tx_chain_cons;
 
 	DBENTER(BCE_VERBOSE_SEND | BCE_VERBOSE_INTR);
-	DBRUN(sc->tx_interrupts++);
+	DBRUN(sc->interrupts_tx++);
 	DBPRINT(sc, BCE_EXTREME_SEND, "%s(enter): tx_prod = 0x%04X, "
 	    "tx_cons = 0x%04X, tx_prod_bseq = 0x%08X\n",
 	    __FUNCTION__, sc->tx_prod, sc->tx_cons, sc->tx_prod_bseq);
@@ -6306,7 +6373,7 @@ bce_tx_intr(struct bce_softc *sc)
 	hw_tx_cons = sc->hw_tx_cons = bce_get_hw_tx_cons(sc);
 	sw_tx_cons = sc->tx_cons;
 
-	/* Prevent speculative reads from getting ahead of the status block. */
+	/* Prevent speculative reads of the status block. */
 	bus_space_barrier(sc->bce_btag, sc->bce_bhandle, 0, 0,
 	    BUS_SPACE_BARRIER_READ);
 
@@ -6354,7 +6421,7 @@ bce_tx_intr(struct bce_softc *sc)
 
 			DBRUNMSG(BCE_INFO_SEND,
 			    BCE_PRINTF("%s(): Unloading map/freeing mbuf "
-			    "from tx_bd[0x%04X]\n", __FUNCTION__, 
+			    "from tx_bd[0x%04X]\n", __FUNCTION__,
 			    sw_tx_chain_cons));
 
 			/* Unmap the mbuf. */
@@ -6496,9 +6563,9 @@ bce_init_locked(struct bce_softc *sc)
 	 * size. Be generous on the receive if we have room.
 	 */
 #ifdef BCE_JUMBO_HDRSPLIT
-	if (ifp->if_mtu <= (sc->rx_bd_mbuf_data_len + 
+	if (ifp->if_mtu <= (sc->rx_bd_mbuf_data_len +
 	    sc->pg_bd_mbuf_alloc_size))
-		ether_mtu = sc->rx_bd_mbuf_data_len + 
+		ether_mtu = sc->rx_bd_mbuf_data_len +
 		    sc->pg_bd_mbuf_alloc_size;
 #else
 	if (ifp->if_mtu <= sc->rx_bd_mbuf_data_len)
@@ -6509,7 +6576,7 @@ bce_init_locked(struct bce_softc *sc)
 
 	ether_mtu += ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN + ETHER_CRC_LEN;
 
-	DBPRINT(sc, BCE_INFO_MISC, "%s(): setting h/w mtu = %d\n", 
+	DBPRINT(sc, BCE_INFO_MISC, "%s(): setting h/w mtu = %d\n",
 	    __FUNCTION__, ether_mtu);
 
 	/* Program the mtu, enabling jumbo frame support if necessary. */
@@ -6522,7 +6589,7 @@ bce_init_locked(struct bce_softc *sc)
 
 	DBPRINT(sc, BCE_INFO_LOAD,
 	    "%s(): rx_bd_mbuf_alloc_size = %d, rx_bce_mbuf_data_len = %d, "
-	    "rx_bd_mbuf_align_pad = %d\n", __FUNCTION__, 
+	    "rx_bd_mbuf_align_pad = %d\n", __FUNCTION__,
 	    sc->rx_bd_mbuf_alloc_size, sc->rx_bd_mbuf_data_len,
 	    sc->rx_bd_mbuf_align_pad);
 
@@ -6617,6 +6684,12 @@ bce_init(void *xsc)
 }
 
 
+/****************************************************************************/
+/* Modifies an mbuf for TSO on the hardware.                                */
+/*                                                                          */
+/* Returns:                                                                 */
+/*   Pointer to a modified mbuf.                                            */
+/****************************************************************************/
 static struct mbuf *
 bce_tso_setup(struct bce_softc *sc, struct mbuf **m_head, u16 *flags)
 {
@@ -6627,8 +6700,9 @@ bce_tso_setup(struct bce_softc *sc, struct mbuf **m_head, u16 *flags)
 	u16 etype;
 	int hdr_len, ip_hlen = 0, tcp_hlen = 0, ip_len = 0;
 
-	DBRUN(sc->requested_tso_frames++);
-	/* Controller requires to monify mbuf chains. */
+	DBRUN(sc->tso_frames_requested++);
+
+	/* Controller may modify mbuf chains. */
 	if (M_WRITABLE(*m_head) == 0) {
 		m = m_dup(*m_head, M_DONTWAIT);
 		m_freem(*m_head);
@@ -6639,6 +6713,7 @@ bce_tso_setup(struct bce_softc *sc, struct mbuf **m_head, u16 *flags)
 		}
 		*m_head = m;
 	}
+
 	/*
 	 * For TSO the controller needs two pieces of info,
 	 * the MSS and the IP+TCP options length.
@@ -6714,9 +6789,12 @@ bce_tso_setup(struct bce_softc *sc, struct mbuf **m_head, u16 *flags)
 
 	/* Set the LSO flag in the TX BD */
 	*flags |= TX_BD_FLAGS_SW_LSO;
+
 	/* Set the length of IP + TCP options (in 32 bit words) */
 	*flags |= (((ip_hlen + tcp_hlen - sizeof(struct ip) -
 	    sizeof(struct tcphdr)) >> 2) << 8);
+
+	DBRUN(sc->tso_frames_completed++);
 	return (*m_head);
 }
 
@@ -6743,22 +6821,24 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 #ifdef BCE_DEBUG
 	u16 debug_prod;
 #endif
+
 	int i, error, nsegs, rc = 0;
 
 	DBENTER(BCE_VERBOSE_SEND);
-	DBPRINT(sc, BCE_INFO_SEND,
-	    "%s(enter): tx_prod = 0x%04X, tx_chain_prod = %04X, "
-	    "tx_prod_bseq = 0x%08X\n",
-	    __FUNCTION__, sc->tx_prod, (u16) TX_CHAIN_IDX(sc->tx_prod),
-	    sc->tx_prod_bseq);
+
+	/* Make sure we have room in the TX chain. */
+	if (sc->used_tx_bd >= sc->max_tx_bd)
+		goto bce_tx_encap_exit;
 
 	/* Transfer any checksum offload flags to the bd. */
 	m0 = *m_head;
 	if (m0->m_pkthdr.csum_flags) {
 		if (m0->m_pkthdr.csum_flags & CSUM_TSO) {
 			m0 = bce_tso_setup(sc, m_head, &flags);
-			if (m0 == NULL)
+			if (m0 == NULL) {
+				DBRUN(sc->tso_frames_failed++);
 				goto bce_tx_encap_exit;
+			}
 			mss = htole16(m0->m_pkthdr.tso_segsz);
 		} else {
 			if (m0->m_pkthdr.csum_flags & CSUM_IP)
@@ -6785,8 +6865,7 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 
 	/* Check if the DMA mapping was successful */
 	if (error == EFBIG) {
-
-		sc->fragmented_mbuf_count++;
+		sc->mbuf_frag_count++;
 
 		/* Try to defrag the mbuf. */
 		m0 = m_collapse(*m_head, M_DONTWAIT, BCE_MAX_SEGMENTS);
@@ -6801,8 +6880,8 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 
 		/* Defrag was successful, try mapping again */
 		*m_head = m0;
-		error = bus_dmamap_load_mbuf_sg(sc->tx_mbuf_tag, map, m0,
-		    segs, &nsegs, BUS_DMA_NOWAIT);
+		error = bus_dmamap_load_mbuf_sg(sc->tx_mbuf_tag,
+		    map, m0, segs, &nsegs, BUS_DMA_NOWAIT);
 
 		/* Still getting an error after a defrag. */
 		if (error == ENOMEM) {
@@ -6864,9 +6943,11 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 		txbd= &sc->tx_bd_chain[TX_PAGE(chain_prod)]
 		    [TX_IDX(chain_prod)];
 
-		txbd->tx_bd_haddr_lo = htole32(BCE_ADDR_LO(segs[i].ds_addr));
-		txbd->tx_bd_haddr_hi = htole32(BCE_ADDR_HI(segs[i].ds_addr));
-		txbd->tx_bd_mss_nbytes = htole32(mss << 16) | 
+		txbd->tx_bd_haddr_lo =
+		    htole32(BCE_ADDR_LO(segs[i].ds_addr));
+		txbd->tx_bd_haddr_hi =
+		    htole32(BCE_ADDR_HI(segs[i].ds_addr));
+		txbd->tx_bd_mss_nbytes = htole32(mss << 16) |
 		    htole16(segs[i].ds_len);
 		txbd->tx_bd_vlan_tag = htole16(vlan_tag);
 		txbd->tx_bd_flags = htole16(flags);
@@ -6879,12 +6960,8 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 	/* Set the END flag on the last TX buffer descriptor. */
 	txbd->tx_bd_flags |= htole16(TX_BD_FLAGS_END);
 
-	DBRUNMSG(BCE_EXTREME_SEND, bce_dump_tx_chain(sc, debug_prod, nsegs));
-
-	DBPRINT(sc, BCE_INFO_SEND,
-	    "%s( end ): prod = 0x%04X, chain_prod = 0x%04X, "
-	    "prod_bseq = 0x%08X\n",
-	    __FUNCTION__, prod, chain_prod, prod_bseq);
+	DBRUNMSG(BCE_EXTREME_SEND,
+	    bce_dump_tx_chain(sc, debug_prod, nsegs));
 
 	/*
 	 * Ensure that the mbuf pointer for this transmission
@@ -6910,11 +6987,11 @@ bce_tx_encap(struct bce_softc *sc, struct mbuf **m_head)
 	sc->tx_prod = prod;
 	sc->tx_prod_bseq = prod_bseq;
 
-	DBPRINT(sc, BCE_INFO_SEND,
-	    "%s(exit): prod = 0x%04X, chain_prod = %04X, "
-	    "prod_bseq = 0x%08X\n", __FUNCTION__, 
-	    sc->tx_prod, (u16) TX_CHAIN_IDX(sc->tx_prod),
-	    sc->tx_prod_bseq);
+	/* Tell the chip about the waiting TX frames. */
+	REG_WR16(sc, MB_GET_CID_ADDR(TX_CID) +
+	    BCE_L2MQ_TX_HOST_BIDX, sc->tx_prod);
+	REG_WR(sc, MB_GET_CID_ADDR(TX_CID) +
+	    BCE_L2MQ_TX_HOST_BSEQ, sc->tx_prod_bseq);
 
 bce_tx_encap_exit:
 	DBEXIT(BCE_VERBOSE_SEND);
@@ -7006,23 +7083,6 @@ bce_start_locked(struct ifnet *ifp)
 	DBPRINT(sc, BCE_VERBOSE_SEND, "%s(): Inserted %d frames into "
 	    "send queue.\n", __FUNCTION__, count);
 
-	REG_WR(sc, BCE_MQ_COMMAND, REG_RD(sc, BCE_MQ_COMMAND) | 
-	    BCE_MQ_COMMAND_NO_MAP_ERROR);
-
-	/* Write the mailbox and tell the chip about the waiting tx_bd's. */
-	DBPRINT(sc, BCE_VERBOSE_SEND, "%s(): MB_GET_CID_ADDR(TX_CID) = "
-	    "0x%08X; BCE_L2MQ_TX_HOST_BIDX = 0x%08X, sc->tx_prod = 0x%04X\n",
-	    __FUNCTION__, MB_GET_CID_ADDR(TX_CID), 
-	    BCE_L2MQ_TX_HOST_BIDX, sc->tx_prod);
-	REG_WR16(sc, MB_GET_CID_ADDR(TX_CID) + 
-	    BCE_L2MQ_TX_HOST_BIDX, sc->tx_prod);
-
-	DBPRINT(sc, BCE_VERBOSE_SEND, "%s(): MB_GET_CID_ADDR(TX_CID) = "
-	    "0x%08X; BCE_L2MQ_TX_HOST_BSEQ = 0x%08X, sc->tx_prod_bseq = "
-	    "0x%04X\n",	__FUNCTION__, MB_GET_CID_ADDR(TX_CID), 
-	    BCE_L2MQ_TX_HOST_BSEQ, sc->tx_prod_bseq);
-	REG_WR(sc, MB_GET_CID_ADDR(TX_CID) + BCE_L2MQ_TX_HOST_BSEQ, sc->tx_prod_bseq);
-
 	/* Set the tx timeout. */
 	sc->watchdog_timer = BCE_TX_TIMEOUT;
 
@@ -7103,19 +7163,19 @@ bce_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 		/* No buffer allocation size changes are necessary. */
 #else
 		/* Recalculate our buffer allocation sizes. */
-		if ((ifp->if_mtu + ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN + 
+		if ((ifp->if_mtu + ETHER_HDR_LEN + ETHER_VLAN_ENCAP_LEN +
 		     ETHER_CRC_LEN) > MCLBYTES) {
 			sc->rx_bd_mbuf_alloc_size = MJUM9BYTES;
-			sc->rx_bd_mbuf_align_pad  = 
+			sc->rx_bd_mbuf_align_pad  =
 			    roundup2(MJUM9BYTES, 16) - MJUM9BYTES;
-			sc->rx_bd_mbuf_data_len = 
+			sc->rx_bd_mbuf_data_len =
 			    sc->rx_bd_mbuf_alloc_size -
 			    sc->rx_bd_mbuf_align_pad;
 		} else {
 			sc->rx_bd_mbuf_alloc_size = MCLBYTES;
-			sc->rx_bd_mbuf_align_pad  = 
+			sc->rx_bd_mbuf_align_pad  =
 			    roundup2(MCLBYTES, 16) - MCLBYTES;
-			sc->rx_bd_mbuf_data_len = 
+			sc->rx_bd_mbuf_data_len =
 			    sc->rx_bd_mbuf_alloc_size -
 			    sc->rx_bd_mbuf_align_pad;
 		}
@@ -7161,7 +7221,7 @@ bce_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 	/* Add/Delete multicast address */
 	case SIOCADDMULTI:
 	case SIOCDELMULTI:
-		DBPRINT(sc, BCE_VERBOSE_MISC, 
+		DBPRINT(sc, BCE_VERBOSE_MISC,
 		    "Received SIOCADDMULTI/SIOCDELMULTI\n");
 
 		BCE_LOCK(sc);
@@ -7174,7 +7234,7 @@ bce_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 	/* Set/Get Interface media */
 	case SIOCSIFMEDIA:
 	case SIOCGIFMEDIA:
-		DBPRINT(sc, BCE_VERBOSE_MISC, 
+		DBPRINT(sc, BCE_VERBOSE_MISC,
 		    "Received SIOCSIFMEDIA/SIOCGIFMEDIA\n");
 
 		mii = device_get_softc(sc->bce_miibus);
@@ -7185,7 +7245,7 @@ bce_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 	/* Set interface capability */
 	case SIOCSIFCAP:
 		mask = ifr->ifr_reqcap ^ ifp->if_capenable;
-		DBPRINT(sc, BCE_INFO_MISC, 
+		DBPRINT(sc, BCE_INFO_MISC,
 		    "Received SIOCSIFCAP = 0x%08X\n", (u32) mask);
 
 		/* Toggle the TX checksum capabilities enable flag. */
@@ -7270,7 +7330,7 @@ bce_watchdog(struct bce_softc *sc)
 		goto bce_watchdog_exit;
 
 	BCE_PRINTF("%s(%d): Watchdog timeout occurred, resetting!\n",
-		__FILE__, __LINE__);
+	    __FILE__, __LINE__);
 
 	DBRUNMSG(BCE_INFO,
 	    bce_dump_driver_state(sc);
@@ -7336,7 +7396,7 @@ bce_intr(void *xsc)
 	 * interrupt then there's nothing to do.
 	 */
 	if ((sc->status_block->status_idx == sc->last_status_idx) &&
-	    (REG_RD(sc, BCE_PCICFG_MISC_STATUS) & 
+	    (REG_RD(sc, BCE_PCICFG_MISC_STATUS) &
 	     BCE_PCICFG_MISC_STATUS_INTA_VALUE)) {
 		DBPRINT(sc, BCE_VERBOSE_INTR, "%s(): Spurious interrupt.\n",
 		    __FUNCTION__);
@@ -7361,17 +7421,17 @@ bce_intr(void *xsc)
 		    BCE_PRINTF("Simulating unexpected status attention "
 		    "bit set.");
 		    sc->unexpected_attention_sim_count++;
-		    status_attn_bits = status_attn_bits | 
+		    status_attn_bits = status_attn_bits |
 		    STATUS_ATTN_BITS_PARITY_ERROR);
 
 		/* Was it a link change interrupt? */
 		if ((status_attn_bits & STATUS_ATTN_BITS_LINK_STATE) !=
-		    (sc->status_block->status_attn_bits_ack & 
+		    (sc->status_block->status_attn_bits_ack &
 		     STATUS_ATTN_BITS_LINK_STATE)) {
 			bce_phy_intr(sc);
 
 			/* Clear transient updates during link state change. */
-			REG_WR(sc, BCE_HC_COMMAND, sc->hc_command | 
+			REG_WR(sc, BCE_HC_COMMAND, sc->hc_command |
 			    BCE_HC_COMMAND_COAL_NOW_WO_INT);
 			REG_RD(sc, BCE_HC_COMMAND);
 		}
@@ -7384,7 +7444,7 @@ bce_intr(void *xsc)
 			sc->unexpected_attention_count++;
 
 			BCE_PRINTF("%s(%d): Fatal attention detected: "
-			    "0x%08X\n",	__FILE__, __LINE__, 
+			    "0x%08X\n",	__FILE__, __LINE__,
 			    sc->status_block->status_attn_bits);
 
 			DBRUNMSG(BCE_FATAL,
@@ -7420,7 +7480,7 @@ bce_intr(void *xsc)
 		hw_rx_cons = bce_get_hw_rx_cons(sc);
 		hw_tx_cons = bce_get_hw_tx_cons(sc);
 
-		if ((hw_rx_cons == sc->hw_rx_cons) && 
+		if ((hw_rx_cons == sc->hw_rx_cons) &&
 		    (hw_tx_cons == sc->hw_tx_cons))
 			break;
 
@@ -7433,7 +7493,7 @@ bce_intr(void *xsc)
 	bce_enable_intr(sc, 0);
 
 	/* Handle any frames that arrived while handling the interrupt. */
-	if (ifp->if_drv_flags & IFF_DRV_RUNNING && 
+	if (ifp->if_drv_flags & IFF_DRV_RUNNING &&
 	    !IFQ_DRV_IS_EMPTY(&ifp->if_snd))
 		bce_start_locked(ifp);
 
@@ -7560,7 +7620,7 @@ bce_stats_update(struct bce_softc *sc)
 	 */
 	if (!(BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5706) &&
 	    !(BCE_CHIP_ID(sc) == BCE_CHIP_ID_5708_A0))
-		ifp->if_oerrors += 
+		ifp->if_oerrors +=
 		    (u_long) stats->stat_Dot3StatsCarrierSenseErrors;
 
 	/*
@@ -7799,25 +7859,28 @@ bce_pulse(void *xsc)
 	sc->bc_state = bce_shmem_rd(sc, BCE_BC_STATE_CONDITION);
 
 	/* Report whether the bootcode still knows the driver is running. */
-	if (sc->bce_drv_cardiac_arrest == FALSE) {
-		if (!(sc->bc_state & BCE_CONDITION_DRV_PRESENT)) {
-			sc->bce_drv_cardiac_arrest = TRUE;
-			BCE_PRINTF("%s(): Bootcode lost the driver pulse! "
-			    "(bc_state = 0x%08X)\n", __FUNCTION__, 
-			    sc->bc_state);
-		}
-	} else {
- 		/*
- 		 * Not supported by all bootcode versions.
- 		 * (v5.0.11+ and v5.2.1+)  Older bootcode
- 		 * will require the driver to reset the
- 		 * controller to clear this condition.
-		 */
-		if (sc->bc_state & BCE_CONDITION_DRV_PRESENT) {
-			sc->bce_drv_cardiac_arrest = FALSE;
-			BCE_PRINTF("%s(): Bootcode found the driver pulse! "
-			    "(bc_state = 0x%08X)\n", __FUNCTION__, 
-			    sc->bc_state);
+	if (bootverbose) {
+		if (sc->bce_drv_cardiac_arrest == FALSE) {
+			if (!(sc->bc_state & BCE_CONDITION_DRV_PRESENT)) {
+				sc->bce_drv_cardiac_arrest = TRUE;
+				BCE_PRINTF("%s(): Warning: bootcode "
+				    "thinks driver is absent! "
+				    "(bc_state = 0x%08X)\n",
+				    __FUNCTION__, sc->bc_state);
+			}
+		} else {
+			/*
+			 * Not supported by all bootcode versions.
+			 * (v5.0.11+ and v5.2.1+)  Older bootcode
+			 * will require the driver to reset the
+			 * controller to clear this condition.
+			 */
+			if (sc->bc_state & BCE_CONDITION_DRV_PRESENT) {
+				sc->bce_drv_cardiac_arrest = FALSE;
+				BCE_PRINTF("%s(): Bootcode found the "
+				    "driver pulse! (bc_state = 0x%08X)\n",
+				    __FUNCTION__, sc->bc_state);
+			}
 		}
 	}
 
@@ -7874,7 +7937,7 @@ bce_tick(void *xsc)
 	/* Check if the link has come up. */
 	if ((mii->mii_media_status & IFM_ACTIVE) &&
 	    (IFM_SUBTYPE(mii->mii_media_active) != IFM_NONE)) {
-		DBPRINT(sc, BCE_VERBOSE_MISC, 
+		DBPRINT(sc, BCE_VERBOSE_MISC,
 		    "%s(): Link up!\n", __FUNCTION__);
 		sc->bce_link_up = TRUE;
 		if ((IFM_SUBTYPE(mii->mii_media_active) == IFM_1000_T ||
@@ -8010,6 +8073,62 @@ bce_sysctl_stats_block(SYSCTL_HANDLER_ARGS)
 }
 
 
+/****************************************************************************/
+/* Allows the stat counters to be cleared without unloading/reloading the   */
+/* driver.                                                                  */
+/*                                                                          */
+/* Returns:                                                                 */
+/*   0 for success, positive value for failure.                             */
+/****************************************************************************/
+static int
+bce_sysctl_stats_clear(SYSCTL_HANDLER_ARGS)
+{
+	int error;
+	int result;
+	struct bce_softc *sc;
+
+	result = -1;
+	error = sysctl_handle_int(oidp, &result, 0, req);
+
+	if (error || !req->newptr)
+		return (error);
+
+	if (result == 1) {
+		sc = (struct bce_softc *)arg1;
+
+		/* Clear the internal H/W statistics counters. */
+		REG_WR(sc, BCE_HC_COMMAND, BCE_HC_COMMAND_CLR_STAT_NOW);
+
+		/* Reset the driver maintained statistics. */
+		sc->interrupts_rx =
+		    sc->interrupts_tx = 0;
+		sc->tso_frames_requested =
+		    sc->tso_frames_completed =
+		    sc->tso_frames_failed = 0;
+		sc->rx_empty_count =
+		    sc->tx_full_count = 0;
+		sc->rx_low_watermark = USABLE_RX_BD;
+		sc->tx_hi_watermark = 0;
+		sc->l2fhdr_error_count =
+		    sc->l2fhdr_error_sim_count = 0;
+		sc->mbuf_alloc_failed_count =
+		    sc->mbuf_alloc_failed_sim_count = 0;
+		sc->dma_map_addr_rx_failed_count =
+		    sc->dma_map_addr_tx_failed_count = 0;
+		sc->mbuf_frag_count = 0;
+		sc->csum_offload_tcp_udp =
+		   sc->csum_offload_ip = 0;
+		sc->vlan_tagged_frames_rcvd =
+		   sc->vlan_tagged_frames_stripped = 0;
+
+		/* Clear firmware maintained statistics. */
+		REG_WR_IND(sc, 0x120084, 0);
+	}
+
+	return error;
+}
+
+
 /****************************************************************************/
 /* Allows the bootcode state to be dumped through the sysctl interface.     */
 /*                                                                          */
@@ -8252,8 +8371,7 @@ static int
 bce_sysctl_dump_ctx(SYSCTL_HANDLER_ARGS)
 {
 	struct bce_softc *sc;
-	int error;
-	u16 result;
+	int error, result;
 
 	result = -1;
 	error = sysctl_handle_int(oidp, &result, 0, req);
@@ -8352,8 +8470,8 @@ bce_add_sysctls(struct bce_softc *sc)
 	    0, "Number of mbuf allocation failures");
 
 	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
-	    "fragmented_mbuf_count",
-	    CTLFLAG_RD, &sc->fragmented_mbuf_count,
+	    "mbuf_frag_count",
+	    CTLFLAG_RD, &sc->mbuf_frag_count,
 	    0, "Number of fragmented mbufs");
 
 #ifdef BCE_DEBUG
@@ -8367,7 +8485,7 @@ bce_add_sysctls(struct bce_softc *sc)
 	    "dma_map_addr_failed_sim_count",
 	    CTLFLAG_RD, &sc->dma_map_addr_failed_sim_count,
 	    0, "Number of simulated DMA mapping failures");
-	
+
 #endif
 
 	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
@@ -8424,18 +8542,48 @@ bce_add_sysctls(struct bce_softc *sc)
 	    0, "Number of times the TX chain was full");
 
 	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
-	    "requested_tso_frames",
-	    CTLFLAG_RD, &sc->requested_tso_frames,
-	    0, "Number of TSO frames received");
+	    "tso_frames_requested",
+	    CTLFLAG_RD, &sc->tso_frames_requested,
+	    0, "Number of TSO frames requested");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "tso_frames_completed",
+	    CTLFLAG_RD, &sc->tso_frames_completed,
+	    0, "Number of TSO frames completed");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "tso_frames_failed",
+	    CTLFLAG_RD, &sc->tso_frames_failed,
+	    0, "Number of TSO frames failed");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "csum_offload_ip",
+	    CTLFLAG_RD, &sc->csum_offload_ip,
+	    0, "Number of IP checksum offload frames");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "csum_offload_tcp_udp",
+	    CTLFLAG_RD, &sc->csum_offload_tcp_udp,
+	    0, "Number of TCP/UDP checksum offload frames");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "vlan_tagged_frames_rcvd",
+	    CTLFLAG_RD, &sc->vlan_tagged_frames_rcvd,
+	    0, "Number of VLAN tagged frames received");
+
+	SYSCTL_ADD_INT(ctx, children, OID_AUTO,
+	    "vlan_tagged_frames_stripped",
+	    CTLFLAG_RD, &sc->vlan_tagged_frames_stripped,
+	    0, "Number of VLAN tagged frames stripped");
 
 	SYSCTL_ADD_UINT(ctx, children, OID_AUTO,
-	    "rx_interrupts",
-	    CTLFLAG_RD, &sc->rx_interrupts,
+	    "interrupts_rx",
+	    CTLFLAG_RD, &sc->interrupts_rx,
 	    0, "Number of RX interrupts");
 
 	SYSCTL_ADD_UINT(ctx, children, OID_AUTO,
-	    "tx_interrupts",
-	    CTLFLAG_RD, &sc->tx_interrupts,
+	    "interrupts_tx",
+	    CTLFLAG_RD, &sc->interrupts_tx,
 	    0, "Number of TX interrupts");
 #endif
 
@@ -8730,12 +8878,17 @@ bce_add_sysctls(struct bce_softc *sc)
 	SYSCTL_ADD_PROC(ctx, children, OID_AUTO,
 	    "status_block", CTLTYPE_INT | CTLFLAG_RW,
 	    (void *)sc, 0,
-	    bce_sysctl_status_block, "I", "Status block");
+	    bce_sysctl_status_block, "I", "Dump status block");
 
 	SYSCTL_ADD_PROC(ctx, children, OID_AUTO,
 	    "stats_block", CTLTYPE_INT | CTLFLAG_RW,
 	    (void *)sc, 0,
-	    bce_sysctl_stats_block, "I", "Stats block");
+	    bce_sysctl_stats_block, "I", "Dump statistics block");
+
+	SYSCTL_ADD_PROC(ctx, children, OID_AUTO,
+	    "stats_clear", CTLTYPE_INT | CTLFLAG_RW,
+	    (void *)sc, 0,
+	    bce_sysctl_stats_clear, "I", "Clear statistics block");
 
 	SYSCTL_ADD_PROC(ctx, children, OID_AUTO,
 	    "bc_state", CTLTYPE_INT | CTLFLAG_RW,
@@ -8882,18 +9035,18 @@ bce_dump_enet(struct bce_softc *sc, struct mbuf *m)
 		switch (ip->ip_p) {
 		case IPPROTO_TCP:
 			th = (struct tcphdr *)((caddr_t)ip + (ip->ip_hl << 2));
-			BCE_PRINTF("-tcp: dest = %d, src = %d, hlen = %d bytes, "
-			    "flags = 0x%b, csum = 0x%04X\n",
-			    ntohs(th->th_dport), ntohs(th->th_sport), 
-			    (th->th_off << 2), th->th_flags, 
+			BCE_PRINTF("-tcp: dest = %d, src = %d, hlen = "
+			    "%d bytes, flags = 0x%b, csum = 0x%04X\n",
+			    ntohs(th->th_dport), ntohs(th->th_sport),
+			    (th->th_off << 2), th->th_flags,
 			    "\20\10CWR\07ECE\06URG\05ACK\04PSH\03RST"
 			    "\02SYN\01FIN", ntohs(th->th_sum));
 			break;
 		case IPPROTO_UDP:
 			uh = (struct udphdr *)((caddr_t)ip + (ip->ip_hl << 2));
 			BCE_PRINTF("-udp: dest = %d, src = %d, len = %d "
-			    "bytes, csum = 0x%04X\n", ntohs(uh->uh_dport), 
-			    ntohs(uh->uh_sport), ntohs(uh->uh_ulen), 
+			    "bytes, csum = 0x%04X\n", ntohs(uh->uh_dport),
+			    ntohs(uh->uh_sport), ntohs(uh->uh_ulen),
 			    ntohs(uh->uh_sum));
 			break;
 		case IPPROTO_ICMP:
@@ -8960,10 +9113,10 @@ bce_dump_mbuf(struct bce_softc *sc, struct mbuf *m)
 
 		if (mp->m_flags & M_PKTHDR) {
 			BCE_PRINTF("- m_pkthdr: len = %d, flags = 0x%b, "
-			    "csum_flags = %b\n", mp->m_pkthdr.len, 
+			    "csum_flags = %b\n", mp->m_pkthdr.len,
 			    mp->m_flags, "\20\12M_BCAST\13M_MCAST\14M_FRAG"
 			    "\15M_FIRSTFRAG\16M_LASTFRAG\21M_VLANTAG"
-			    "\22M_PROMISC\23M_NOFREE", 
+			    "\22M_PROMISC\23M_NOFREE",
 			    mp->m_pkthdr.csum_flags,
 			    "\20\1CSUM_IP\2CSUM_TCP\3CSUM_UDP\4CSUM_IP_FRAGS"
 			    "\5CSUM_FRAGMENT\6CSUM_TSO\11CSUM_IP_CHECKED"
@@ -8995,7 +9148,7 @@ bce_dump_mbuf(struct bce_softc *sc, struct mbuf *m)
 				printf("EXT_DISPOSABLE\n"); break;
 			case EXT_EXTREF:
 				printf("EXT_EXTREF\n"); break;
-			default:             
+			default:
 				printf("UNKNOWN\n");
 			}
 		}
@@ -9116,96 +9269,96 @@ bce_dump_txbd(struct bce_softc *sc, int idx, struct tx_bd *txbd)
 	else if ((idx & USABLE_TX_BD_PER_PAGE) == USABLE_TX_BD_PER_PAGE)
 		/* TX Chain page pointer. */
 		BCE_PRINTF("tx_bd[0x%04X]: haddr = 0x%08X:%08X, chain page "
-		    "pointer\n", idx, txbd->tx_bd_haddr_hi, 
+		    "pointer\n", idx, txbd->tx_bd_haddr_hi,
 		    txbd->tx_bd_haddr_lo);
 	else {
 		/* Normal tx_bd entry. */
 		BCE_PRINTF("tx_bd[0x%04X]: haddr = 0x%08X:%08X, "
 		    "mss_nbytes = 0x%08X, vlan tag = 0x%04X, flags = "
-		    "0x%04X (", idx, txbd->tx_bd_haddr_hi, 
-		    txbd->tx_bd_haddr_lo, txbd->tx_bd_mss_nbytes, 
+		    "0x%04X (", idx, txbd->tx_bd_haddr_hi,
+		    txbd->tx_bd_haddr_lo, txbd->tx_bd_mss_nbytes,
 		    txbd->tx_bd_vlan_tag, txbd->tx_bd_flags);
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_CONN_FAULT) {
-			if (i>0) 
-				printf("|"); 
-			printf("CONN_FAULT"); 
+			if (i>0)
+				printf("|");
+			printf("CONN_FAULT");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_TCP_UDP_CKSUM) {
-			if (i>0) 
-				printf("|"); 
-			printf("TCP_UDP_CKSUM"); 
+			if (i>0)
+				printf("|");
+			printf("TCP_UDP_CKSUM");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_IP_CKSUM) {
-			if (i>0) 
-				printf("|"); 
-			printf("IP_CKSUM"); 
+			if (i>0)
+				printf("|");
+			printf("IP_CKSUM");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_VLAN_TAG) {
-			if (i>0) 
-				printf("|"); 
-			printf("VLAN"); 
+			if (i>0)
+				printf("|");
+			printf("VLAN");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_COAL_NOW) {
-			if (i>0) 
-				printf("|"); 
-			printf("COAL_NOW"); 
+			if (i>0)
+				printf("|");
+			printf("COAL_NOW");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_DONT_GEN_CRC) {
-			if (i>0) 
-				printf("|"); 
-			printf("DONT_GEN_CRC"); 
+			if (i>0)
+				printf("|");
+			printf("DONT_GEN_CRC");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_START) {
-			if (i>0) 
-				printf("|"); 
-			printf("START"); 
+			if (i>0)
+				printf("|");
+			printf("START");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_END) {
-			if (i>0) 
-				printf("|"); 
-			printf("END"); 
+			if (i>0)
+				printf("|");
+			printf("END");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_SW_LSO) {
-			if (i>0) 
-				printf("|"); 
-			printf("LSO"); 
+			if (i>0)
+				printf("|");
+			printf("LSO");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_SW_OPTION_WORD) {
-			if (i>0) 
-				printf("|"); 
-			printf("SW_OPTION=%d", ((txbd->tx_bd_flags & 
+			if (i>0)
+				printf("|");
+			printf("SW_OPTION=%d", ((txbd->tx_bd_flags &
 			    TX_BD_FLAGS_SW_OPTION_WORD) >> 8)); i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_SW_FLAGS) {
-			if (i>0) 
-				printf("|"); 
-			printf("SW_FLAGS"); 
+			if (i>0)
+				printf("|");
+			printf("SW_FLAGS");
 			i++;
 		}
 
 		if (txbd->tx_bd_flags & TX_BD_FLAGS_SW_SNAP) {
-			if (i>0) 
-				printf("|"); 
+			if (i>0)
+				printf("|");
 			printf("SNAP)");
 		} else {
 			printf(")\n");
@@ -9229,13 +9382,13 @@ bce_dump_rxbd(struct bce_softc *sc, int idx, struct rx_bd *rxbd)
 	else if ((idx & USABLE_RX_BD_PER_PAGE) == USABLE_RX_BD_PER_PAGE)
 		/* RX Chain page pointer. */
 		BCE_PRINTF("rx_bd[0x%04X]: haddr = 0x%08X:%08X, chain page "
-		    "pointer\n", idx, rxbd->rx_bd_haddr_hi, 
+		    "pointer\n", idx, rxbd->rx_bd_haddr_hi,
 		    rxbd->rx_bd_haddr_lo);
 	else
 		/* Normal rx_bd entry. */
 		BCE_PRINTF("rx_bd[0x%04X]: haddr = 0x%08X:%08X, nbytes = "
-		    "0x%08X, flags = 0x%08X\n", idx, rxbd->rx_bd_haddr_hi, 
-		    rxbd->rx_bd_haddr_lo, rxbd->rx_bd_len, 
+		    "0x%08X, flags = 0x%08X\n", idx, rxbd->rx_bd_haddr_hi,
+		    rxbd->rx_bd_haddr_lo, rxbd->rx_bd_len,
 		    rxbd->rx_bd_flags);
 }
 
@@ -9325,57 +9478,57 @@ bce_dump_ctx(struct bce_softc *sc, u16 cid)
 		    "index\n", CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_NX_BDIDX));
 		BCE_PRINTF(" 0x%08X - (L2CTX_RX_HOST_PG_BDIDX) host page "
-		    "producer index\n", CTX_RD(sc, GET_CID_ADDR(cid), 
+		    "producer index\n", CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_HOST_PG_BDIDX));
 		BCE_PRINTF(" 0x%08X - (L2CTX_RX_PG_BUF_SIZE) host rx_bd/page "
-		    "buffer size\n", CTX_RD(sc, GET_CID_ADDR(cid), 
+		    "buffer size\n", CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_PG_BUF_SIZE));
 		BCE_PRINTF(" 0x%08X - (L2CTX_RX_NX_PG_BDHADDR_HI) h/w page "
-		    "chain address\n", CTX_RD(sc, GET_CID_ADDR(cid), 
+		    "chain address\n", CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_NX_PG_BDHADDR_HI));
 		BCE_PRINTF(" 0x%08X - (L2CTX_RX_NX_PG_BDHADDR_LO) h/w page "
-		    "chain address\n", CTX_RD(sc, GET_CID_ADDR(cid), 
+		    "chain address\n", CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_NX_PG_BDHADDR_LO));
 		BCE_PRINTF(" 0x%08X - (L2CTX_RX_NX_PG_BDIDX) h/w page "
-		    "consumer index\n",	CTX_RD(sc, GET_CID_ADDR(cid), 
+		    "consumer index\n",	CTX_RD(sc, GET_CID_ADDR(cid),
 		    BCE_L2CTX_RX_NX_PG_BDIDX));
 	} else if (cid == TX_CID) {
 		if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 		    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716)) {
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TYPE_XI) ctx type\n",
-			    CTX_RD(sc, GET_CID_ADDR(cid), 
+			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_TYPE_XI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_CMD_TX_TYPE_XI) ctx "
-			    "cmd\n", CTX_RD(sc, GET_CID_ADDR(cid), 
+			    "cmd\n", CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_CMD_TYPE_XI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TBDR_BDHADDR_HI_XI) "
-			    "h/w buffer descriptor address\n",	
-			    CTX_RD(sc, GET_CID_ADDR(cid), 
+			    "h/w buffer descriptor address\n",
+			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_TBDR_BHADDR_HI_XI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TBDR_BHADDR_LO_XI) "
-			    "h/w buffer	descriptor address\n", 
+			    "h/w buffer	descriptor address\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_TBDR_BHADDR_LO_XI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_HOST_BIDX_XI) "
-			    "host producer index\n", 
+			    "host producer index\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_HOST_BIDX_XI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_HOST_BSEQ_XI) "
-			    "host byte sequence\n", 
+			    "host byte sequence\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_HOST_BSEQ_XI));
 		} else {
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TYPE) ctx type\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid), BCE_L2CTX_TX_TYPE));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_CMD_TYPE) ctx cmd\n",
-			    CTX_RD(sc, GET_CID_ADDR(cid), 
+			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_CMD_TYPE));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TBDR_BDHADDR_HI) "
-			    "h/w buffer	descriptor address\n", 
+			    "h/w buffer	descriptor address\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_TBDR_BHADDR_HI));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_TBDR_BHADDR_LO) "
-			    "h/w buffer	descriptor address\n", 
+			    "h/w buffer	descriptor address\n",
 			    CTX_RD(sc, GET_CID_ADDR(cid),
 			    BCE_L2CTX_TX_TBDR_BHADDR_LO));
 			BCE_PRINTF(" 0x%08X - (L2CTX_TX_HOST_BIDX) host "
@@ -9605,8 +9758,8 @@ bce_dump_ftqs(struct bce_softc *sc)
 
 	if ((BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5709) ||
 	    (BCE_CHIP_NUM(sc) == BCE_CHIP_NUM_5716))
-		val = val | 
-		    (BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PCSQ_VALID_CNT_XI << 
+		val = val |
+		    (BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PCSQ_VALID_CNT_XI <<
 		     24);
 	REG_WR(sc, BCE_HC_STAT_GEN_SEL_0, val);
 
@@ -10020,11 +10173,11 @@ bce_dump_driver_state(struct bce_softc *sc)
 	BCE_PRINTF("         0x%08X - (sc->interrupts_generated) "
 	    "h/w intrs\n", sc->interrupts_generated);
 
-	BCE_PRINTF("         0x%08X - (sc->rx_interrupts) "
-	    "rx interrupts handled\n", sc->rx_interrupts);
+	BCE_PRINTF("         0x%08X - (sc->interrupts_rx) "
+	    "rx interrupts handled\n", sc->interrupts_rx);
 
-	BCE_PRINTF("         0x%08X - (sc->tx_interrupts) "
-	    "tx interrupts handled\n", sc->tx_interrupts);
+	BCE_PRINTF("         0x%08X - (sc->interrupts_tx) "
+	    "tx interrupts handled\n", sc->interrupts_tx);
 
 	BCE_PRINTF("         0x%08X - (sc->phy_interrupts) "
 	    "phy interrupts handled\n", sc->phy_interrupts);
@@ -10122,15 +10275,15 @@ bce_dump_hw_state(struct bce_softc *sc)
 	    val, BCE_MISC_ENABLE_STATUS_BITS);
 
 	val = REG_RD(sc, BCE_DMA_STATUS);
-	BCE_PRINTF("0x%08X - (0x%06X) dma_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) dma_status\n",
 	    val, BCE_DMA_STATUS);
 
 	val = REG_RD(sc, BCE_CTX_STATUS);
-	BCE_PRINTF("0x%08X - (0x%06X) ctx_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) ctx_status\n",
 	    val, BCE_CTX_STATUS);
 
 	val = REG_RD(sc, BCE_EMAC_STATUS);
-	BCE_PRINTF("0x%08X - (0x%06X) emac_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) emac_status\n",
 	    val, BCE_EMAC_STATUS);
 
 	val = REG_RD(sc, BCE_RPM_STATUS);
@@ -10139,16 +10292,16 @@ bce_dump_hw_state(struct bce_softc *sc)
 
 	/* ToDo: Create a #define for this constant. */
 	val = REG_RD(sc, 0x2004);
-	BCE_PRINTF("0x%08X - (0x%06X) rlup_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rlup_status\n",
 	    val, 0x2004);
 
 	val = REG_RD(sc, BCE_RV2P_STATUS);
-	BCE_PRINTF("0x%08X - (0x%06X) rv2p_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rv2p_status\n",
 	    val, BCE_RV2P_STATUS);
 
 	/* ToDo: Create a #define for this constant. */
 	val = REG_RD(sc, 0x2c04);
-	BCE_PRINTF("0x%08X - (0x%06X) rdma_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rdma_status\n",
 	    val, 0x2c04);
 
 	val = REG_RD(sc, BCE_TBDR_STATUS);
@@ -10156,7 +10309,7 @@ bce_dump_hw_state(struct bce_softc *sc)
 	    val, BCE_TBDR_STATUS);
 
 	val = REG_RD(sc, BCE_TDMA_STATUS);
-	BCE_PRINTF("0x%08X - (0x%06X) tdma_status\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) tdma_status\n",
 	    val, BCE_TDMA_STATUS);
 
 	val = REG_RD(sc, BCE_HC_STATUS);
@@ -10164,27 +10317,27 @@ bce_dump_hw_state(struct bce_softc *sc)
 	    val, BCE_HC_STATUS);
 
 	val = REG_RD_IND(sc, BCE_TXP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_state\n",
 	    val, BCE_TXP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_TPAT_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_state\n",
 	    val, BCE_TPAT_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_RXP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_state\n",
 	    val, BCE_RXP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_COM_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_state\n",
 	    val, BCE_COM_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_MCP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) mcp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) mcp_cpu_state\n",
 	    val, BCE_MCP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_CP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_state\n",
 	    val, BCE_CP_CPU_STATE);
 
 	BCE_PRINTF(
@@ -10306,15 +10459,15 @@ bce_dump_txp_state(struct bce_softc *sc, int regs)
 	BCE_PRINTF("Firmware version - %s\n", (char *) fw_version);
 
 	val = REG_RD_IND(sc, BCE_TXP_CPU_MODE);
-	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_mode\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_mode\n",
 	    val, BCE_TXP_CPU_MODE);
 
 	val = REG_RD_IND(sc, BCE_TXP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_state\n",
 	    val, BCE_TXP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_TXP_CPU_EVENT_MASK);
-	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_event_mask\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) txp_cpu_event_mask\n",
 	    val, BCE_TXP_CPU_EVENT_MASK);
 
 	if (regs) {
@@ -10327,8 +10480,8 @@ bce_dump_txp_state(struct bce_softc *sc, int regs)
 			/* Skip the big blank spaces */
 			if (i < 0x454000 && i > 0x5ffff)
 				BCE_PRINTF("0x%04X: 0x%08X 0x%08X "
-				    "0x%08X 0x%08X\n", i, 
-				    REG_RD_IND(sc, i), 
+				    "0x%08X 0x%08X\n", i,
+				    REG_RD_IND(sc, i),
 				    REG_RD_IND(sc, i + 0x4),
 				    REG_RD_IND(sc, i + 0x8),
 				    REG_RD_IND(sc, i + 0xC));
@@ -10366,15 +10519,15 @@ bce_dump_rxp_state(struct bce_softc *sc, int regs)
 	BCE_PRINTF("Firmware version - %s\n", (char *) fw_version);
 
 	val = REG_RD_IND(sc, BCE_RXP_CPU_MODE);
-	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_mode\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_mode\n",
 	    val, BCE_RXP_CPU_MODE);
 
 	val = REG_RD_IND(sc, BCE_RXP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_state\n",
 	    val, BCE_RXP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_RXP_CPU_EVENT_MASK);
-	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_event_mask\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) rxp_cpu_event_mask\n",
 	    val, BCE_RXP_CPU_EVENT_MASK);
 
 	if (regs) {
@@ -10387,10 +10540,10 @@ bce_dump_rxp_state(struct bce_softc *sc, int regs)
 			/* Skip the big blank sapces */
 			if (i < 0xc5400 && i > 0xdffff)
 				BCE_PRINTF("0x%04X: 0x%08X 0x%08X "
-				    "0x%08X 0x%08X\n", i, 
-				    REG_RD_IND(sc, i), 
+				    "0x%08X 0x%08X\n", i,
+				    REG_RD_IND(sc, i),
 				    REG_RD_IND(sc, i + 0x4),
-				    REG_RD_IND(sc, i + 0x8), 
+				    REG_RD_IND(sc, i + 0x8),
 				    REG_RD_IND(sc, i + 0xC));
 		}
 	}
@@ -10426,15 +10579,15 @@ bce_dump_tpat_state(struct bce_softc *sc, int regs)
 	BCE_PRINTF("Firmware version - %s\n", (char *) fw_version);
 
 	val = REG_RD_IND(sc, BCE_TPAT_CPU_MODE);
-	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_mode\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_mode\n",
 	    val, BCE_TPAT_CPU_MODE);
 
 	val = REG_RD_IND(sc, BCE_TPAT_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_state\n",
 	    val, BCE_TPAT_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_TPAT_CPU_EVENT_MASK);
-	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_event_mask\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) tpat_cpu_event_mask\n",
 	    val, BCE_TPAT_CPU_EVENT_MASK);
 
 	if (regs) {
@@ -10448,9 +10601,9 @@ bce_dump_tpat_state(struct bce_softc *sc, int regs)
 			if (i < 0x854000 && i > 0x9ffff)
 				BCE_PRINTF("0x%04X: 0x%08X 0x%08X "
 				    "0x%08X 0x%08X\n", i,
-				    REG_RD_IND(sc, i), 
+				    REG_RD_IND(sc, i),
 				    REG_RD_IND(sc, i + 0x4),
-				    REG_RD_IND(sc, i + 0x8), 
+				    REG_RD_IND(sc, i + 0x8),
 				    REG_RD_IND(sc, i + 0xC));
 		}
 	}
@@ -10486,11 +10639,11 @@ bce_dump_cp_state(struct bce_softc *sc, int regs)
 	BCE_PRINTF("Firmware version - %s\n", (char *) fw_version);
 
 	val = REG_RD_IND(sc, BCE_CP_CPU_MODE);
-	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_mode\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_mode\n",
 	    val, BCE_CP_CPU_MODE);
 
 	val = REG_RD_IND(sc, BCE_CP_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) cp_cpu_state\n",
 	    val, BCE_CP_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_CP_CPU_EVENT_MASK);
@@ -10507,10 +10660,10 @@ bce_dump_cp_state(struct bce_softc *sc, int regs)
 			/* Skip the big blank spaces */
 			if (i < 0x185400 && i > 0x19ffff)
 				BCE_PRINTF("0x%04X: 0x%08X 0x%08X "
-				    "0x%08X 0x%08X\n", i, 
-				    REG_RD_IND(sc, i), 
+				    "0x%08X 0x%08X\n", i,
+				    REG_RD_IND(sc, i),
 				    REG_RD_IND(sc, i + 0x4),
-				    REG_RD_IND(sc, i + 0x8), 
+				    REG_RD_IND(sc, i + 0x8),
 				    REG_RD_IND(sc, i + 0xC));
 		}
 	}
@@ -10546,11 +10699,11 @@ bce_dump_com_state(struct bce_softc *sc, int regs)
 	BCE_PRINTF("Firmware version - %s\n", (char *) fw_version);
 
 	val = REG_RD_IND(sc, BCE_COM_CPU_MODE);
-	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_mode\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_mode\n",
 	    val, BCE_COM_CPU_MODE);
 
 	val = REG_RD_IND(sc, BCE_COM_CPU_STATE);
-	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_state\n", 
+	BCE_PRINTF("0x%08X - (0x%06X) com_cpu_state\n",
 	    val, BCE_COM_CPU_STATE);
 
 	val = REG_RD_IND(sc, BCE_COM_CPU_EVENT_MASK);
@@ -10565,8 +10718,8 @@ bce_dump_com_state(struct bce_softc *sc, int regs)
 
 		for (int i = BCE_COM_CPU_MODE; i < 0x1053e8; i += 0x10) {
 			BCE_PRINTF("0x%04X: 0x%08X 0x%08X "
-			    "0x%08X 0x%08X\n", i, 
-			    REG_RD_IND(sc, i), 
+			    "0x%08X 0x%08X\n", i,
+			    REG_RD_IND(sc, i),
 			    REG_RD_IND(sc, i + 0x4),
 			    REG_RD_IND(sc, i + 0x8),
 			    REG_RD_IND(sc, i + 0xC));
@@ -10605,17 +10758,17 @@ bce_dump_rv2p_state(struct bce_softc *sc)
 	val = 0x00000001;
 	REG_WR_IND(sc, BCE_RV2P_PROC1_ADDR_CMD, val);
 	fw_ver_low = REG_RD_IND(sc, BCE_RV2P_INSTR_LOW);
-	fw_ver_high = REG_RD_IND(sc, BCE_RV2P_INSTR_HIGH) & 
+	fw_ver_high = REG_RD_IND(sc, BCE_RV2P_INSTR_HIGH) &
 	    BCE_RV2P_INSTR_HIGH_HIGH;
-	BCE_PRINTF("RV2P1 Firmware version - 0x%08X:0x%08X\n", 
+	BCE_PRINTF("RV2P1 Firmware version - 0x%08X:0x%08X\n",
 	    fw_ver_high, fw_ver_low);
 
 	val = 0x00000001;
 	REG_WR_IND(sc, BCE_RV2P_PROC2_ADDR_CMD, val);
 	fw_ver_low = REG_RD_IND(sc, BCE_RV2P_INSTR_LOW);
-	fw_ver_high = REG_RD_IND(sc, BCE_RV2P_INSTR_HIGH) & 
+	fw_ver_high = REG_RD_IND(sc, BCE_RV2P_INSTR_HIGH) &
 	    BCE_RV2P_INSTR_HIGH_HIGH;
-	BCE_PRINTF("RV2P2 Firmware version - 0x%08X:0x%08X\n", 
+	BCE_PRINTF("RV2P2 Firmware version - 0x%08X:0x%08X\n",
 	    fw_ver_high, fw_ver_low);
 
 	/* Resume the RV2P processors. */
diff --git a/sys/dev/bce/if_bcereg.h b/sys/dev/bce/if_bcereg.h
index 49bba6377f8..d6e7c242d63 100644
--- a/sys/dev/bce/if_bcereg.h
+++ b/sys/dev/bce/if_bcereg.h
@@ -284,8 +284,8 @@
 	"\02EnaPauseRcv"	\
 	"\01EnaPausXmit"
 
-/* 
- * Remove before release: 
+/*
+ * Remove before release:
  *
  * #define BCE_DEBUG
  * #define BCE_NVRAM_WRITE_SUPPORT
@@ -439,7 +439,7 @@
 	}
 
 /* Runs a particular command based on the code path. */
-#define DBRUNCP(cp, args...) 					\
+#define DBRUNCP(cp, args...)					\
 	if (BCE_CODE_PATH(cp)) { 				\
 		args; 						\
 	}
@@ -4116,7 +4116,7 @@ struct l2_fhdr {
 #define BCE_RDMA_FTQ_CTL_CUR_DEPTH			(0x3ffL<<22)
 
 
-	
+
 /*
  *  timer_reg definition
  *  offset: 0x4400
@@ -4476,107 +4476,107 @@ struct l2_fhdr {
  *  csch_reg definition
  *  offset: 0x4000
  */
-#define BCE_CSCH_COMMAND							0x00004000
-#define BCE_CSCH_CH_FTQ_CMD							0x000043f8
-#define BCE_CSCH_CH_FTQ_CTL							0x000043fc
-#define BCE_CSCH_CH_FTQ_CTL_MAX_DEPTH				(0x3ffL<<12)
-#define BCE_CSCH_CH_FTQ_CTL_CUR_DEPTH				(0x3ffL<<22)
+#define BCE_CSCH_COMMAND				0x00004000
+#define BCE_CSCH_CH_FTQ_CMD				0x000043f8
+#define BCE_CSCH_CH_FTQ_CTL				0x000043fc
+#define BCE_CSCH_CH_FTQ_CTL_MAX_DEPTH			(0x3ffL<<12)
+#define BCE_CSCH_CH_FTQ_CTL_CUR_DEPTH			(0x3ffL<<22)
 
 
 /*
  *  tbdr_reg definition
  *  offset: 0x5000
  */
-#define BCE_TBDR_COMMAND							0x00005000
-#define BCE_TBDR_COMMAND_ENABLE						(1L<<0)
-#define BCE_TBDR_COMMAND_SOFT_RST					(1L<<1)
-#define BCE_TBDR_COMMAND_MSTR_ABORT					(1L<<4)
+#define BCE_TBDR_COMMAND				0x00005000
+#define BCE_TBDR_COMMAND_ENABLE				(1L<<0)
+#define BCE_TBDR_COMMAND_SOFT_RST			(1L<<1)
+#define BCE_TBDR_COMMAND_MSTR_ABORT			(1L<<4)
 
-#define BCE_TBDR_STATUS								0x00005004
-#define BCE_TBDR_STATUS_DMA_WAIT					(1L<<0)
-#define BCE_TBDR_STATUS_FTQ_WAIT					(1L<<1)
-#define BCE_TBDR_STATUS_FIFO_OVERFLOW				(1L<<2)
-#define BCE_TBDR_STATUS_FIFO_UNDERFLOW				(1L<<3)
-#define BCE_TBDR_STATUS_SEARCHMISS_ERROR			(1L<<4)
-#define BCE_TBDR_STATUS_FTQ_ENTRY_CNT				(1L<<5)
-#define BCE_TBDR_STATUS_BURST_CNT					(1L<<6)
+#define BCE_TBDR_STATUS					0x00005004
+#define BCE_TBDR_STATUS_DMA_WAIT			(1L<<0)
+#define BCE_TBDR_STATUS_FTQ_WAIT			(1L<<1)
+#define BCE_TBDR_STATUS_FIFO_OVERFLOW			(1L<<2)
+#define BCE_TBDR_STATUS_FIFO_UNDERFLOW			(1L<<3)
+#define BCE_TBDR_STATUS_SEARCHMISS_ERROR		(1L<<4)
+#define BCE_TBDR_STATUS_FTQ_ENTRY_CNT			(1L<<5)
+#define BCE_TBDR_STATUS_BURST_CNT			(1L<<6)
 
-#define BCE_TBDR_CONFIG								0x00005008
-#define BCE_TBDR_CONFIG_MAX_BDS						(0xffL<<0)
-#define BCE_TBDR_CONFIG_SWAP_MODE					(1L<<8)
-#define BCE_TBDR_CONFIG_PRIORITY					(1L<<9)
+#define BCE_TBDR_CONFIG					0x00005008
+#define BCE_TBDR_CONFIG_MAX_BDS				(0xffL<<0)
+#define BCE_TBDR_CONFIG_SWAP_MODE			(1L<<8)
+#define BCE_TBDR_CONFIG_PRIORITY			(1L<<9)
 #define BCE_TBDR_CONFIG_CACHE_NEXT_PAGE_PTRS		(1L<<10)
-#define BCE_TBDR_CONFIG_PAGE_SIZE			 		(0xfL<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_256				(0L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_512				(1L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_1K				(2L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_2K				(3L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_4K				(4L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_8K				(5L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_16K				(6L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_32K				(7L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_64K				(8L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_128K				(9L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_256K				(10L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_512K				(11L<<24)
-#define BCE_TBDR_CONFIG_PAGE_SIZE_1M				(12L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE			(0xfL<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_256			(0L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_512			(1L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_1K			(2L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_2K			(3L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_4K			(4L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_8K			(5L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_16K			(6L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_32K			(7L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_64K			(8L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_128K			(9L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_256K			(10L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_512K			(11L<<24)
+#define BCE_TBDR_CONFIG_PAGE_SIZE_1M			(12L<<24)
 
-#define BCE_TBDR_DEBUG_VECT_PEEK					0x0000500c
-#define BCE_TBDR_DEBUG_VECT_PEEK_1_VALUE			(0x7ffL<<0)
-#define BCE_TBDR_DEBUG_VECT_PEEK_1_PEEK_EN			(1L<<11)
-#define BCE_TBDR_DEBUG_VECT_PEEK_1_SEL				(0xfL<<12)
-#define BCE_TBDR_DEBUG_VECT_PEEK_2_VALUE			(0x7ffL<<16)
-#define BCE_TBDR_DEBUG_VECT_PEEK_2_PEEK_EN			(1L<<27)
-#define BCE_TBDR_DEBUG_VECT_PEEK_2_SEL				(0xfL<<28)
+#define BCE_TBDR_DEBUG_VECT_PEEK			0x0000500c
+#define BCE_TBDR_DEBUG_VECT_PEEK_1_VALUE		(0x7ffL<<0)
+#define BCE_TBDR_DEBUG_VECT_PEEK_1_PEEK_EN		(1L<<11)
+#define BCE_TBDR_DEBUG_VECT_PEEK_1_SEL			(0xfL<<12)
+#define BCE_TBDR_DEBUG_VECT_PEEK_2_VALUE		(0x7ffL<<16)
+#define BCE_TBDR_DEBUG_VECT_PEEK_2_PEEK_EN		(1L<<27)
+#define BCE_TBDR_DEBUG_VECT_PEEK_2_SEL			(0xfL<<28)
 
-#define BCE_TBDR_FTQ_DATA							0x000053c0
-#define BCE_TBDR_FTQ_CMD							0x000053f8
-#define BCE_TBDR_FTQ_CMD_OFFSET						(0x3ffL<<0)
-#define BCE_TBDR_FTQ_CMD_WR_TOP						(1L<<10)
-#define BCE_TBDR_FTQ_CMD_WR_TOP_0					(0L<<10)
-#define BCE_TBDR_FTQ_CMD_WR_TOP_1					(1L<<10)
-#define BCE_TBDR_FTQ_CMD_SFT_RESET					(1L<<25)
-#define BCE_TBDR_FTQ_CMD_RD_DATA					(1L<<26)
-#define BCE_TBDR_FTQ_CMD_ADD_INTERVEN				(1L<<27)
-#define BCE_TBDR_FTQ_CMD_ADD_DATA					(1L<<28)
-#define BCE_TBDR_FTQ_CMD_INTERVENE_CLR				(1L<<29)
-#define BCE_TBDR_FTQ_CMD_POP						(1L<<30)
-#define BCE_TBDR_FTQ_CMD_BUSY						(1L<<31)
+#define BCE_TBDR_FTQ_DATA				0x000053c0
+#define BCE_TBDR_FTQ_CMD				0x000053f8
+#define BCE_TBDR_FTQ_CMD_OFFSET				(0x3ffL<<0)
+#define BCE_TBDR_FTQ_CMD_WR_TOP				(1L<<10)
+#define BCE_TBDR_FTQ_CMD_WR_TOP_0			(0L<<10)
+#define BCE_TBDR_FTQ_CMD_WR_TOP_1			(1L<<10)
+#define BCE_TBDR_FTQ_CMD_SFT_RESET			(1L<<25)
+#define BCE_TBDR_FTQ_CMD_RD_DATA			(1L<<26)
+#define BCE_TBDR_FTQ_CMD_ADD_INTERVEN			(1L<<27)
+#define BCE_TBDR_FTQ_CMD_ADD_DATA			(1L<<28)
+#define BCE_TBDR_FTQ_CMD_INTERVENE_CLR			(1L<<29)
+#define BCE_TBDR_FTQ_CMD_POP				(1L<<30)
+#define BCE_TBDR_FTQ_CMD_BUSY				(1L<<31)
 
-#define BCE_TBDR_FTQ_CTL							0x000053fc
-#define BCE_TBDR_FTQ_CTL_INTERVENE					(1L<<0)
-#define BCE_TBDR_FTQ_CTL_OVERFLOW					(1L<<1)
-#define BCE_TBDR_FTQ_CTL_FORCE_INTERVENE			(1L<<2)
-#define BCE_TBDR_FTQ_CTL_MAX_DEPTH					(0x3ffL<<12)
-#define BCE_TBDR_FTQ_CTL_CUR_DEPTH					(0x3ffL<<22)
+#define BCE_TBDR_FTQ_CTL				0x000053fc
+#define BCE_TBDR_FTQ_CTL_INTERVENE			(1L<<0)
+#define BCE_TBDR_FTQ_CTL_OVERFLOW			(1L<<1)
+#define BCE_TBDR_FTQ_CTL_FORCE_INTERVENE		(1L<<2)
+#define BCE_TBDR_FTQ_CTL_MAX_DEPTH			(0x3ffL<<12)
+#define BCE_TBDR_FTQ_CTL_CUR_DEPTH			(0x3ffL<<22)
 
 
 /*
  *  tdma_reg definition
  *  offset: 0x5c00
  */
-#define BCE_TDMA_COMMAND							0x00005c00
-#define BCE_TDMA_COMMAND_ENABLED					(1L<<0)
-#define BCE_TDMA_COMMAND_MASTER_ABORT				(1L<<4)
-#define BCE_TDMA_COMMAND_BAD_L2_LENGTH_ABORT		(1L<<7)
+#define BCE_TDMA_COMMAND				0x00005c00
+#define BCE_TDMA_COMMAND_ENABLED			 (1L<<0)
+#define BCE_TDMA_COMMAND_MASTER_ABORT			 (1L<<4)
+#define BCE_TDMA_COMMAND_BAD_L2_LENGTH_ABORT		 (1L<<7)
 
-#define BCE_TDMA_STATUS								0x00005c04
-#define BCE_TDMA_STATUS_DMA_WAIT					(1L<<0)
-#define BCE_TDMA_STATUS_PAYLOAD_WAIT				(1L<<1)
-#define BCE_TDMA_STATUS_PATCH_FTQ_WAIT				(1L<<2)
+#define BCE_TDMA_STATUS					0x00005c04
+#define BCE_TDMA_STATUS_DMA_WAIT			 (1L<<0)
+#define BCE_TDMA_STATUS_PAYLOAD_WAIT			 (1L<<1)
+#define BCE_TDMA_STATUS_PATCH_FTQ_WAIT			 (1L<<2)
 #define BCE_TDMA_STATUS_LOCK_WAIT			 (1L<<3)
 #define BCE_TDMA_STATUS_FTQ_ENTRY_CNT			 (1L<<16)
 #define BCE_TDMA_STATUS_BURST_CNT			 (1L<<17)
 
-#define BCE_TDMA_CONFIG				0x00005c08
-#define BCE_TDMA_CONFIG_ONE_DMA			 (1L<<0)
+#define BCE_TDMA_CONFIG					0x00005c08
+#define BCE_TDMA_CONFIG_ONE_DMA				 (1L<<0)
 #define BCE_TDMA_CONFIG_ONE_RECORD			 (1L<<1)
 #define BCE_TDMA_CONFIG_LIMIT_SZ			 (0xfL<<4)
 #define BCE_TDMA_CONFIG_LIMIT_SZ_64			 (0L<<4)
 #define BCE_TDMA_CONFIG_LIMIT_SZ_128			 (0x4L<<4)
 #define BCE_TDMA_CONFIG_LIMIT_SZ_256			 (0x6L<<4)
 #define BCE_TDMA_CONFIG_LIMIT_SZ_512			 (0x8L<<4)
-#define BCE_TDMA_CONFIG_LINE_SZ			 (0xfL<<8)
+#define BCE_TDMA_CONFIG_LINE_SZ				 (0xfL<<8)
 #define BCE_TDMA_CONFIG_LINE_SZ_64			 (0L<<8)
 #define BCE_TDMA_CONFIG_LINE_SZ_128			 (4L<<8)
 #define BCE_TDMA_CONFIG_LINE_SZ_256			 (6L<<8)
@@ -4612,7 +4612,7 @@ struct l2_fhdr {
 #define BCE_TDMA_DR_INTF_FSM_DR_BUF			 (0x7L<<12)
 #define BCE_TDMA_DR_INTF_FSM_DMAD			 (0x7L<<16)
 
-#define BCE_TDMA_DR_INTF_STATUS			0x00005c8c
+#define BCE_TDMA_DR_INTF_STATUS				0x00005c8c
 #define BCE_TDMA_DR_INTF_STATUS_HOLE_PHASE		 (0x7L<<0)
 #define BCE_TDMA_DR_INTF_STATUS_DATA_AVAIL		 (0x3L<<4)
 #define BCE_TDMA_DR_INTF_STATUS_SHIFT_ADDR		 (0x7L<<8)
@@ -4621,8 +4621,8 @@ struct l2_fhdr {
 
 #define BCE_TDMA_FTQ_DATA				0x00005fc0
 #define BCE_TDMA_FTQ_CMD				0x00005ff8
-#define BCE_TDMA_FTQ_CMD_OFFSET			 (0x3ffL<<0)
-#define BCE_TDMA_FTQ_CMD_WR_TOP			 (1L<<10)
+#define BCE_TDMA_FTQ_CMD_OFFSET				 (0x3ffL<<0)
+#define BCE_TDMA_FTQ_CMD_WR_TOP				 (1L<<10)
 #define BCE_TDMA_FTQ_CMD_WR_TOP_0			 (0L<<10)
 #define BCE_TDMA_FTQ_CMD_WR_TOP_1			 (1L<<10)
 #define BCE_TDMA_FTQ_CMD_SFT_RESET			 (1L<<25)
@@ -4645,7 +4645,7 @@ struct l2_fhdr {
  *  nvm_reg definition
  *  offset: 0x6400
  */
-#define BCE_NVM_COMMAND				0x00006400
+#define BCE_NVM_COMMAND					0x00006400
 #define BCE_NVM_COMMAND_RST				 (1L<<0)
 #define BCE_NVM_COMMAND_DONE				 (1L<<3)
 #define BCE_NVM_COMMAND_DOIT				 (1L<<4)
@@ -4694,23 +4694,23 @@ struct l2_fhdr {
 #define BCE_NVM_READ_NVM_READ_VALUE_SI			 (32L<<0)
 
 #define BCE_NVM_CFG1					0x00006414
-#define BCE_NVM_CFG1_FLASH_MODE			 (1L<<0)
+#define BCE_NVM_CFG1_FLASH_MODE				 (1L<<0)
 #define BCE_NVM_CFG1_BUFFER_MODE			 (1L<<1)
 #define BCE_NVM_CFG1_PASS_MODE				 (1L<<2)
 #define BCE_NVM_CFG1_BITBANG_MODE			 (1L<<3)
-#define BCE_NVM_CFG1_STATUS_BIT			 (0x7L<<4)
+#define BCE_NVM_CFG1_STATUS_BIT				 (0x7L<<4)
 #define BCE_NVM_CFG1_STATUS_BIT_FLASH_RDY		 (0L<<4)
 #define BCE_NVM_CFG1_STATUS_BIT_BUFFER_RDY		 (7L<<4)
 #define BCE_NVM_CFG1_SPI_CLK_DIV			 (0xfL<<7)
 #define BCE_NVM_CFG1_SEE_CLK_DIV			 (0x7ffL<<11)
 #define BCE_NVM_CFG1_PROTECT_MODE			 (1L<<24)
-#define BCE_NVM_CFG1_FLASH_SIZE			 (1L<<25)
+#define BCE_NVM_CFG1_FLASH_SIZE				 (1L<<25)
 #define BCE_NVM_CFG1_COMPAT_BYPASSS			 (1L<<31)
 
 #define BCE_NVM_CFG2					0x00006418
 #define BCE_NVM_CFG2_ERASE_CMD				 (0xffL<<0)
 #define BCE_NVM_CFG2_DUMMY				 (0xffL<<8)
-#define BCE_NVM_CFG2_STATUS_CMD			 (0xffL<<16)
+#define BCE_NVM_CFG2_STATUS_CMD				 (0xffL<<16)
 
 #define BCE_NVM_CFG3					0x0000641c
 #define BCE_NVM_CFG3_BUFFER_RD_CMD			 (0xffL<<0)
@@ -4727,10 +4727,10 @@ struct l2_fhdr {
 #define BCE_NVM_SW_ARB_ARB_REQ_CLR1			 (1L<<5)
 #define BCE_NVM_SW_ARB_ARB_REQ_CLR2			 (1L<<6)
 #define BCE_NVM_SW_ARB_ARB_REQ_CLR3			 (1L<<7)
-#define BCE_NVM_SW_ARB_ARB_ARB0			 (1L<<8)
-#define BCE_NVM_SW_ARB_ARB_ARB1			 (1L<<9)
-#define BCE_NVM_SW_ARB_ARB_ARB2			 (1L<<10)
-#define BCE_NVM_SW_ARB_ARB_ARB3			 (1L<<11)
+#define BCE_NVM_SW_ARB_ARB_ARB0				 (1L<<8)
+#define BCE_NVM_SW_ARB_ARB_ARB1				 (1L<<9)
+#define BCE_NVM_SW_ARB_ARB_ARB2				 (1L<<10)
+#define BCE_NVM_SW_ARB_ARB_ARB3				 (1L<<11)
 #define BCE_NVM_SW_ARB_REQ0				 (1L<<12)
 #define BCE_NVM_SW_ARB_REQ1				 (1L<<13)
 #define BCE_NVM_SW_ARB_REQ2				 (1L<<14)
@@ -4741,8 +4741,8 @@ struct l2_fhdr {
 #define BCE_NVM_ACCESS_ENABLE_WR_EN			 (1L<<1)
 
 #define BCE_NVM_WRITE1					0x00006428
-#define BCE_NVM_WRITE1_WREN_CMD			 (0xffL<<0)
-#define BCE_NVM_WRITE1_WRDI_CMD			 (0xffL<<8)
+#define BCE_NVM_WRITE1_WREN_CMD				 (0xffL<<0)
+#define BCE_NVM_WRITE1_WRDI_CMD				 (0xffL<<8)
 #define BCE_NVM_WRITE1_SR_DATA				 (0xffL<<16)
 
 
@@ -4753,7 +4753,7 @@ struct l2_fhdr {
 #define BCE_HC_COMMAND					0x00006800
 #define BCE_HC_COMMAND_ENABLE				 (1L<<0)
 #define BCE_HC_COMMAND_SKIP_ABORT			 (1L<<4)
-#define BCE_HC_COMMAND_COAL_NOW			 (1L<<16)
+#define BCE_HC_COMMAND_COAL_NOW			 	 (1L<<16)
 #define BCE_HC_COMMAND_COAL_NOW_WO_INT			 (1L<<17)
 #define BCE_HC_COMMAND_STATS_NOW			 (1L<<18)
 #define BCE_HC_COMMAND_FORCE_INT			 (0x3L<<19)
@@ -4769,7 +4769,7 @@ struct l2_fhdr {
 #define BCE_HC_STATUS_MASTER_ABORT			 (1L<<0)
 #define BCE_HC_STATUS_PARITY_ERROR_STATE		 (1L<<1)
 #define BCE_HC_STATUS_PCI_CLK_CNT_STAT			 (1L<<16)
-#define BCE_HC_STATUS_CORE_CLK_CNT_STAT		 (1L<<17)
+#define BCE_HC_STATUS_CORE_CLK_CNT_STAT			 (1L<<17)
 #define BCE_HC_STATUS_NUM_STATUS_BLOCKS_STAT		 (1L<<18)
 #define BCE_HC_STATUS_NUM_INT_GEN_STAT			 (1L<<19)
 #define BCE_HC_STATUS_NUM_INT_MBOX_WR_STAT		 (1L<<20)
@@ -4790,27 +4790,27 @@ struct l2_fhdr {
 #define BCE_HC_CONFIG_ONE_SHOT				 (1L<<17)
 #define BCE_HC_CONFIG_USE_INT_PARAM			 (1L<<18)
 #define BCE_HC_CONFIG_SET_MASK_AT_RD			 (1L<<19)
-#define BCE_HC_CONFIG_PER_COLLECT_LIMIT		 (0xfL<<20)
+#define BCE_HC_CONFIG_PER_COLLECT_LIMIT			 (0xfL<<20)
 #define BCE_HC_CONFIG_SB_ADDR_INC			 (0x7L<<24)
 #define BCE_HC_CONFIG_SB_ADDR_INC_64B			 (0L<<24)
 #define BCE_HC_CONFIG_SB_ADDR_INC_128B			 (1L<<24)
 #define BCE_HC_CONFIG_SB_ADDR_INC_256B			 (2L<<24)
 #define BCE_HC_CONFIG_SB_ADDR_INC_512B			 (3L<<24)
-#define BCE_HC_CONFIG_SB_ADDR_INC_1024B		 (4L<<24)
-#define BCE_HC_CONFIG_SB_ADDR_INC_2048B		 (5L<<24)
-#define BCE_HC_CONFIG_SB_ADDR_INC_4096B		 (6L<<24)
-#define BCE_HC_CONFIG_SB_ADDR_INC_8192B		 (7L<<24)
-#define BCE_HC_CONFIG_GEN_STAT_AVG_INTR		 (1L<<29)
+#define BCE_HC_CONFIG_SB_ADDR_INC_1024B			 (4L<<24)
+#define BCE_HC_CONFIG_SB_ADDR_INC_2048B			 (5L<<24)
+#define BCE_HC_CONFIG_SB_ADDR_INC_4096B			 (6L<<24)
+#define BCE_HC_CONFIG_SB_ADDR_INC_8192B			 (7L<<24)
+#define BCE_HC_CONFIG_GEN_STAT_AVG_INTR			 (1L<<29)
 #define BCE_HC_CONFIG_UNMASK_ALL			 (1L<<30)
 #define BCE_HC_CONFIG_TX_SEL				 (1L<<31)
 
-#define BCE_HC_ATTN_BITS_ENABLE			0x0000680c
+#define BCE_HC_ATTN_BITS_ENABLE				0x0000680c
 #define BCE_HC_STATUS_ADDR_L				0x00006810
 #define BCE_HC_STATUS_ADDR_H				0x00006814
 #define BCE_HC_STATISTICS_ADDR_L			0x00006818
 #define BCE_HC_STATISTICS_ADDR_H			0x0000681c
 #define BCE_HC_TX_QUICK_CONS_TRIP			0x00006820
-#define BCE_HC_TX_QUICK_CONS_TRIP_VALUE		 (0xffL<<0)
+#define BCE_HC_TX_QUICK_CONS_TRIP_VALUE			 (0xffL<<0)
 #define BCE_HC_TX_QUICK_CONS_TRIP_INT			 (0xffL<<16)
 
 #define BCE_HC_COMP_PROD_TRIP				0x00006824
@@ -4818,14 +4818,14 @@ struct l2_fhdr {
 #define BCE_HC_COMP_PROD_TRIP_INT			 (0xffL<<16)
 
 #define BCE_HC_RX_QUICK_CONS_TRIP			0x00006828
-#define BCE_HC_RX_QUICK_CONS_TRIP_VALUE		 (0xffL<<0)
+#define BCE_HC_RX_QUICK_CONS_TRIP_VALUE			 (0xffL<<0)
 #define BCE_HC_RX_QUICK_CONS_TRIP_INT			 (0xffL<<16)
 
-#define BCE_HC_RX_TICKS				0x0000682c
+#define BCE_HC_RX_TICKS					0x0000682c
 #define BCE_HC_RX_TICKS_VALUE				 (0x3ffL<<0)
 #define BCE_HC_RX_TICKS_INT				 (0x3ffL<<16)
 
-#define BCE_HC_TX_TICKS				0x00006830
+#define BCE_HC_TX_TICKS					0x00006830
 #define BCE_HC_TX_TICKS_VALUE				 (0x3ffL<<0)
 #define BCE_HC_TX_TICKS_INT				 (0x3ffL<<16)
 
@@ -4838,7 +4838,7 @@ struct l2_fhdr {
 #define BCE_HC_CMD_TICKS_INT				 (0x3ffL<<16)
 
 #define BCE_HC_PERIODIC_TICKS				0x0000683c
-#define BCE_HC_PERIODIC_TICKS_HC_PERIODIC_TICKS	 (0xffffL<<0)
+#define BCE_HC_PERIODIC_TICKS_HC_PERIODIC_TICKS		 (0xffffL<<0)
 #define BCE_HC_PERIODIC_TICKS_HC_INT_PERIODIC_TICKS	 (0xffffL<<16)
 
 #define BCE_HC_STAT_COLLECT_TICKS			0x00006840
@@ -4848,12 +4848,12 @@ struct l2_fhdr {
 #define BCE_HC_STATS_TICKS_HC_STAT_TICKS		 (0xffffL<<8)
 
 #define BCE_HC_STATS_INTERRUPT_STATUS			0x00006848
-#define BCE_HC_STATS_INTERRUPT_STATUS_SB_STATUS	 (0x1ffL<<0)
+#define BCE_HC_STATS_INTERRUPT_STATUS_SB_STATUS		 (0x1ffL<<0)
 #define BCE_HC_STATS_INTERRUPT_STATUS_INT_STATUS	 (0x1ffL<<16)
 
 #define BCE_HC_STAT_MEM_DATA				0x0000684c
 #define BCE_HC_STAT_GEN_SEL_0				0x00006850
-#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0		 (0x7fL<<0)
+#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0			 (0x7fL<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RXP_STAT0	 (0L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RXP_STAT1	 (1L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RXP_STAT2	 (2L<<0)
@@ -4921,9 +4921,9 @@ struct l2_fhdr {
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RLUPQ_VALID_CNT	 (66L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RXPQ_VALID_CNT	 (67L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RXPCQ_VALID_CNT	 (68L<<0)
-#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PPQ_VALID_CNT	 (69L<<0)
-#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PMQ_VALID_CNT	 (70L<<0)
-#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PTQ_VALID_CNT	 (71L<<0)
+#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PPQ_VALID_CNT (69L<<0)
+#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PMQ_VALID_CNT (70L<<0)
+#define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RV2PTQ_VALID_CNT (71L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_RDMAQ_VALID_CNT	 (72L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_TSCHQ_VALID_CNT	 (73L<<0)
 #define BCE_HC_STAT_GEN_SEL_0_GEN_SEL_0_TBDRQ_VALID_CNT	 (74L<<0)
@@ -6006,7 +6006,7 @@ struct l2_fhdr {
 #define BCE_TAS_FTQ_CTL_MAX_DEPTH			(0x3ffL<<12)
 #define BCE_TAS_FTQ_CTL_CUR_DEPTH			(0x3ffL<<22)
 
-	
+
 /*
  *  mcp_reg definition
  *  offset: 0x140000
@@ -6372,7 +6372,7 @@ struct fw_info {
 
 struct bce_softc
 {
-	/* Interface info */
+	/* Interface info.  Must be first!! */
 	struct ifnet		*bce_ifp;
 
 	/* Parent device handle */
@@ -6403,167 +6403,168 @@ struct bce_softc
 
 	/* Interrupt handler. */
 	driver_intr_t		*bce_intr;
-	void *bce_intrhand;
-	int  bce_irq_rid;
-	int  bce_msi_count;
+	void			*bce_intrhand;
+	int			bce_irq_rid;
+	int			bce_msi_count;
 
 	/* ASIC Chip ID. */
-	u32  bce_chipid;
+	u32			bce_chipid;
 
 	/* General controller flags. */
-	u32  bce_flags;
-#define BCE_PCIX_FLAG			0x00000001
-#define BCE_PCI_32BIT_FLAG 		0x00000002
-#define BCE_RESERVED_FLAG		0x00000004
-#define BCE_NO_WOL_FLAG			0x00000008
-#define BCE_USING_DAC_FLAG		0x00000010
-#define BCE_USING_MSI_FLAG 		0x00000020
-#define BCE_MFW_ENABLE_FLAG		0x00000040
-#define BCE_ONE_SHOT_MSI_FLAG		0x00000080
-#define BCE_USING_MSIX_FLAG		0x00000100
-#define BCE_PCIE_FLAG			0x00000200
+	u32			bce_flags;
+#define BCE_PCIX_FLAG				0x00000001
+#define BCE_PCI_32BIT_FLAG 			0x00000002
+#define BCE_RESERVED_FLAG			0x00000004
+#define BCE_NO_WOL_FLAG				0x00000008
+#define BCE_USING_DAC_FLAG			0x00000010
+#define BCE_USING_MSI_FLAG 			0x00000020
+#define BCE_MFW_ENABLE_FLAG			0x00000040
+#define BCE_ONE_SHOT_MSI_FLAG			0x00000080
+#define BCE_USING_MSIX_FLAG			0x00000100
+#define BCE_PCIE_FLAG				0x00000200
+#define BCE_USING_TX_FLOW_CONTROL		0x00000400
 
 	/* Controller capability flags. */
-	u32  bce_cap_flags;
-#define BCE_MSI_CAPABLE_FLAG		0x00000001
-#define BCE_MSIX_CAPABLE_FLAG		0x00000002
-#define BCE_PCIE_CAPABLE_FLAG		0x00000004
-#define BCE_PCIX_CAPABLE_FLAG		0x00000008
+	u32			bce_cap_flags;
+#define BCE_MSI_CAPABLE_FLAG			0x00000001
+#define BCE_MSIX_CAPABLE_FLAG			0x00000002
+#define BCE_PCIE_CAPABLE_FLAG			0x00000004
+#define BCE_PCIX_CAPABLE_FLAG			0x00000008
 
 	/* PHY specific flags. */
-	u32  bce_phy_flags;
-#define BCE_PHY_SERDES_FLAG		0x00000001
-#define BCE_PHY_CRC_FIX_FLAG		0x00000002
-#define BCE_PHY_PARALLEL_DETECT_FLAG	0x00000004
-#define BCE_PHY_2_5G_CAPABLE_FLAG	0x00000008
-#define BCE_PHY_INT_MODE_MASK_FLAG	0x00000300
-#define BCE_PHY_INT_MODE_AUTO_POLLING_FLAG 0x00000100
-#define BCE_PHY_INT_MODE_LINK_READY_FLAG 0x00000200
-#define BCE_PHY_IEEE_CLAUSE_45_FLAG	0x00000400
+	u32			bce_phy_flags;
+#define BCE_PHY_SERDES_FLAG			0x00000001
+#define BCE_PHY_CRC_FIX_FLAG			0x00000002
+#define BCE_PHY_PARALLEL_DETECT_FLAG		0x00000004
+#define BCE_PHY_2_5G_CAPABLE_FLAG		0x00000008
+#define BCE_PHY_INT_MODE_MASK_FLAG		0x00000300
+#define BCE_PHY_INT_MODE_AUTO_POLLING_FLAG	0x00000100
+#define BCE_PHY_INT_MODE_LINK_READY_FLAG	0x00000200
+#define BCE_PHY_IEEE_CLAUSE_45_FLAG		0x00000400
 
 	/* Values that need to be shared with the PHY driver. */
-	u32  bce_shared_hw_cfg;
-	u32  bce_port_hw_cfg;
+	u32			bce_shared_hw_cfg;
+	u32			bce_port_hw_cfg;
 
-	bus_addr_t max_bus_addr;
+	bus_addr_t		max_bus_addr;
 
 	/* PCI bus speed */
-	u16  bus_speed_mhz;
+	u16			bus_speed_mhz;
 
 	/* PCIe link width */
-	u16  link_width;
+	u16			link_width;
 
 	/* PCIe link speed */
-	u16  link_speed;
+	u16			link_speed;
 
 	/* Flash NVRAM settings */
-	struct flash_spec *bce_flash_info;
+	struct flash_spec	*bce_flash_info;
 
 	/* Flash NVRAM size */
-	u32  bce_flash_size;
+	u32			bce_flash_size;
 
 	/* Shared Memory base address */
-	u32  bce_shmem_base;		
+	u32			bce_shmem_base;
 
 	/* Name string */
-	char * bce_name;
+	char			*bce_name;
 
 	/* Tracks the version of bootcode firmware. */
-	char bce_bc_ver[32];
+	char			bce_bc_ver[32];
 
 	/* Tracks the version of management firmware. */
-	char bce_mfw_ver[32];
+	char			bce_mfw_ver[32];
 
-	/* 
+	/*
 	 * Tracks the state of the firmware.  0 = Running while any
 	 * other value indicates that the firmware is not responding.
 	 */
-	u16  bce_fw_timed_out;
+	u16			bce_fw_timed_out;
 
-	/* 
+	/*
 	 * An incrementing sequence used to coordinate messages passed
 	 * from the driver to the firmware.
-	 */ 
-	u16  bce_fw_wr_seq;
+	 */
+	u16			bce_fw_wr_seq;
 
-	/* 
+	/*
 	 * An incrementing sequence used to let the firmware know that
 	 * the driver is still operating.  Without the pulse, management
 	 * firmware such as IPMI or UMP will operate in OS absent state.
-	 */ 
-	u16  bce_fw_drv_pulse_wr_seq;
+	 */
+	u16			bce_fw_drv_pulse_wr_seq;
 
 	/* Tracks whether firmware has lost the driver's pulse. */
-	u16  bce_drv_cardiac_arrest;
+	u16			bce_drv_cardiac_arrest;
 
 	/* Ethernet MAC address. */
-	u_char eaddr[6];
+	u_char			eaddr[6];
 
-	/* 
+	/*
 	 * These setting are used by the host coalescing (HC) block to
 	 * to control how often the status block, statistics block and
 	 * interrupts are generated.
-	 */ 
-	u16  bce_tx_quick_cons_trip_int;
-	u16  bce_tx_quick_cons_trip;
-	u16  bce_rx_quick_cons_trip_int;
-	u16  bce_rx_quick_cons_trip;
-	u16  bce_tx_ticks_int;
-	u16  bce_tx_ticks;
-	u16  bce_rx_ticks_int;
-	u16  bce_rx_ticks;
-	u32  bce_stats_ticks;
+	 */
+	u16			bce_tx_quick_cons_trip_int;
+	u16			bce_tx_quick_cons_trip;
+	u16			bce_rx_quick_cons_trip_int;
+	u16			bce_rx_quick_cons_trip;
+	u16			bce_tx_ticks_int;
+	u16			bce_tx_ticks;
+	u16			bce_rx_ticks_int;
+	u16			bce_rx_ticks;
+	u32			bce_stats_ticks;
 
 	/* ToDo: Can these be removed? */
-	u16  bce_comp_prod_trip_int;
-	u16  bce_comp_prod_trip;
-	u16  bce_com_ticks_int;
-	u16  bce_com_ticks;
-	u16  bce_cmd_ticks_int;
-	u16  bce_cmd_ticks;
+	u16			bce_comp_prod_trip_int;
+	u16			bce_comp_prod_trip;
+	u16			bce_com_ticks_int;
+	u16			bce_com_ticks;
+	u16			bce_cmd_ticks_int;
+	u16			bce_cmd_ticks;
 
 	/* The address of the integrated PHY on the MII bus. */
-	int  bce_phy_addr;
+	int			bce_phy_addr;
 
 	/* The device handle for the MII bus child device. */
-	device_t bce_miibus;
+	device_t		bce_miibus;
 
 	/* Driver maintained TX chain pointers and byte counter. */
-	u16  rx_prod;
-	u16  rx_cons;
+	u16			rx_prod;
+	u16			rx_cons;
 
 	/* Counts the bytes used in the RX chain. */
-	u32  rx_prod_bseq;
-	u16  tx_prod;
-	u16  tx_cons;
+	u32			rx_prod_bseq;
+	u16			tx_prod;
+	u16			tx_cons;
 
 	/* Counts the bytes used in the TX chain. */
-	u32  tx_prod_bseq;
+	u32			tx_prod_bseq;
 
 #ifdef BCE_JUMBO_HDRSPLIT
-	u16  pg_prod;
-	u16  pg_cons;
+	u16			pg_prod;
+	u16			pg_cons;
 #endif
 
-	int  bce_link_up;
-	struct callout bce_tick_callout;
-	struct callout bce_pulse_callout;
+	int			bce_link_up;
+	struct			callout bce_tick_callout;
+	struct			callout bce_pulse_callout;
 
 	/* Ticks until chip reset */
-	int watchdog_timer;
+	int			watchdog_timer;
 
 	/* Frame size and mbuf allocation size for RX frames. */
-	u32  max_frame_size;
-	int  rx_bd_mbuf_alloc_size;
-	int  rx_bd_mbuf_data_len;
-	int  rx_bd_mbuf_align_pad;
+	u32			max_frame_size;
+	int			rx_bd_mbuf_alloc_size;
+	int			rx_bd_mbuf_data_len;
+	int			rx_bd_mbuf_align_pad;
 
 #ifdef BCE_JUMBO_HDRSPLIT
-	int  pg_bd_mbuf_alloc_size;
+	int			pg_bd_mbuf_alloc_size;
 #endif
 
 	/* Receive mode settings (i.e promiscuous, multicast, etc.). */
-	u32  rx_mode;
+	u32			rx_mode;
 
 	/* Bus tag for the bce controller. */
 	bus_dma_tag_t		parent_tag;
@@ -6595,9 +6596,9 @@ struct bce_softc
 	bus_addr_t		status_block_paddr;
 
 	/* Driver maintained status block values. */
-	u16  last_status_idx;
-	u16  hw_rx_cons;
-	u16  hw_tx_cons;
+	u16			last_status_idx;
+	u16			hw_rx_cons;
+	u16			hw_tx_cons;
 
 	/* H/W maintained statistics block. */
 	bus_dma_tag_t		stats_tag;
@@ -6606,7 +6607,7 @@ struct bce_softc
 	bus_addr_t		stats_block_paddr;
 
 	/* H/W maintained context block. */
-	int  ctx_pages;
+	int			ctx_pages;
 	bus_dma_tag_t		ctx_tag;
 
 	/* BCM5709/16 use host memory for context. */
@@ -6637,139 +6638,156 @@ struct bce_softc
 #endif
 
 	/* Track the number of buffer descriptors in use. */
-	u16 free_rx_bd;
-	u16 max_rx_bd;
-	u16 used_tx_bd;
-	u16 max_tx_bd;
+	u16			free_rx_bd;
+	u16			max_rx_bd;
+	u16			used_tx_bd;
+	u16			max_tx_bd;
 
 #ifdef BCE_JUMBO_HDRSPLIT
-	u16 free_pg_bd;
-	u16 max_pg_bd;
+	u16			free_pg_bd;
+	u16			max_pg_bd;
 #endif
 
 	/* Provides access to hardware statistics through sysctl. */
-	u64 stat_IfHCInOctets;
-	u64 stat_IfHCInBadOctets;
-	u64 stat_IfHCOutOctets;
-	u64 stat_IfHCOutBadOctets;
-	u64 stat_IfHCInUcastPkts;
-	u64 stat_IfHCInMulticastPkts;
-	u64 stat_IfHCInBroadcastPkts;
-	u64 stat_IfHCOutUcastPkts;
-	u64 stat_IfHCOutMulticastPkts;
-	u64 stat_IfHCOutBroadcastPkts;
+	u64			stat_IfHCInOctets;
+	u64			stat_IfHCInBadOctets;
+	u64			stat_IfHCOutOctets;
+	u64			stat_IfHCOutBadOctets;
+	u64			stat_IfHCInUcastPkts;
+	u64			stat_IfHCInMulticastPkts;
+	u64			stat_IfHCInBroadcastPkts;
+	u64			stat_IfHCOutUcastPkts;
+	u64			stat_IfHCOutMulticastPkts;
+	u64			stat_IfHCOutBroadcastPkts;
 
-	u32 stat_emac_tx_stat_dot3statsinternalmactransmiterrors;
-	u32 stat_Dot3StatsCarrierSenseErrors;
-	u32 stat_Dot3StatsFCSErrors;
-	u32 stat_Dot3StatsAlignmentErrors;
-	u32 stat_Dot3StatsSingleCollisionFrames;
-	u32 stat_Dot3StatsMultipleCollisionFrames;
-	u32 stat_Dot3StatsDeferredTransmissions;
-	u32 stat_Dot3StatsExcessiveCollisions;
-	u32 stat_Dot3StatsLateCollisions;
-	u32 stat_EtherStatsCollisions;
-	u32 stat_EtherStatsFragments;
-	u32 stat_EtherStatsJabbers;
-	u32 stat_EtherStatsUndersizePkts;
-	u32 stat_EtherStatsOversizePkts;
-	u32 stat_EtherStatsPktsRx64Octets;
-	u32 stat_EtherStatsPktsRx65Octetsto127Octets;
-	u32 stat_EtherStatsPktsRx128Octetsto255Octets;
-	u32 stat_EtherStatsPktsRx256Octetsto511Octets;
-	u32 stat_EtherStatsPktsRx512Octetsto1023Octets;
-	u32 stat_EtherStatsPktsRx1024Octetsto1522Octets;
-	u32 stat_EtherStatsPktsRx1523Octetsto9022Octets;
-	u32 stat_EtherStatsPktsTx64Octets;
-	u32 stat_EtherStatsPktsTx65Octetsto127Octets;
-	u32 stat_EtherStatsPktsTx128Octetsto255Octets;
-	u32 stat_EtherStatsPktsTx256Octetsto511Octets;
-	u32 stat_EtherStatsPktsTx512Octetsto1023Octets;
-	u32 stat_EtherStatsPktsTx1024Octetsto1522Octets;
-	u32 stat_EtherStatsPktsTx1523Octetsto9022Octets;
-	u32 stat_XonPauseFramesReceived;
-	u32 stat_XoffPauseFramesReceived;
-	u32 stat_OutXonSent;
-	u32 stat_OutXoffSent;
-	u32 stat_FlowControlDone;
-	u32 stat_MacControlFramesReceived;
-	u32 stat_XoffStateEntered;
-	u32 stat_IfInFramesL2FilterDiscards;
-	u32 stat_IfInRuleCheckerDiscards;
-	u32 stat_IfInFTQDiscards;
-	u32 stat_IfInMBUFDiscards;
-	u32 stat_IfInRuleCheckerP4Hit;
-	u32 stat_CatchupInRuleCheckerDiscards;
-	u32 stat_CatchupInFTQDiscards;
-	u32 stat_CatchupInMBUFDiscards;
-	u32 stat_CatchupInRuleCheckerP4Hit;
+	u32	stat_emac_tx_stat_dot3statsinternalmactransmiterrors;
+	u32			stat_Dot3StatsCarrierSenseErrors;
+	u32			stat_Dot3StatsFCSErrors;
+	u32			stat_Dot3StatsAlignmentErrors;
+	u32			stat_Dot3StatsSingleCollisionFrames;
+	u32			stat_Dot3StatsMultipleCollisionFrames;
+	u32			stat_Dot3StatsDeferredTransmissions;
+	u32			stat_Dot3StatsExcessiveCollisions;
+	u32			stat_Dot3StatsLateCollisions;
+	u32			stat_EtherStatsCollisions;
+	u32			stat_EtherStatsFragments;
+	u32			stat_EtherStatsJabbers;
+	u32			stat_EtherStatsUndersizePkts;
+	u32			stat_EtherStatsOversizePkts;
+	u32			stat_EtherStatsPktsRx64Octets;
+	u32			stat_EtherStatsPktsRx65Octetsto127Octets;
+	u32			stat_EtherStatsPktsRx128Octetsto255Octets;
+	u32			stat_EtherStatsPktsRx256Octetsto511Octets;
+	u32			stat_EtherStatsPktsRx512Octetsto1023Octets;
+	u32			stat_EtherStatsPktsRx1024Octetsto1522Octets;
+	u32			stat_EtherStatsPktsRx1523Octetsto9022Octets;
+	u32			stat_EtherStatsPktsTx64Octets;
+	u32			stat_EtherStatsPktsTx65Octetsto127Octets;
+	u32			stat_EtherStatsPktsTx128Octetsto255Octets;
+	u32			stat_EtherStatsPktsTx256Octetsto511Octets;
+	u32			stat_EtherStatsPktsTx512Octetsto1023Octets;
+	u32			stat_EtherStatsPktsTx1024Octetsto1522Octets;
+	u32			stat_EtherStatsPktsTx1523Octetsto9022Octets;
+	u32			stat_XonPauseFramesReceived;
+	u32			stat_XoffPauseFramesReceived;
+	u32			stat_OutXonSent;
+	u32			stat_OutXoffSent;
+	u32			stat_FlowControlDone;
+	u32			stat_MacControlFramesReceived;
+	u32			stat_XoffStateEntered;
+	u32			stat_IfInFramesL2FilterDiscards;
+	u32			stat_IfInRuleCheckerDiscards;
+	u32			stat_IfInFTQDiscards;
+	u32			stat_IfInMBUFDiscards;
+	u32			stat_IfInRuleCheckerP4Hit;
+	u32			stat_CatchupInRuleCheckerDiscards;
+	u32			stat_CatchupInFTQDiscards;
+	u32			stat_CatchupInMBUFDiscards;
+	u32			stat_CatchupInRuleCheckerP4Hit;
 
 	/* Provides access to certain firmware statistics. */
-	u32 com_no_buffers;
+	u32			com_no_buffers;
 
 	/* Recoverable failure counters. */
-	u32 mbuf_alloc_failed_count;
-	u32 fragmented_mbuf_count;
-	u32 unexpected_attention_count;
-	u32 l2fhdr_error_count;
-	u32 dma_map_addr_tx_failed_count;
-	u32 dma_map_addr_rx_failed_count;
+	u32			mbuf_alloc_failed_count;
+	u32			mbuf_frag_count;
+	u32			unexpected_attention_count;
+	u32			l2fhdr_error_count;
+	u32			dma_map_addr_tx_failed_count;
+	u32			dma_map_addr_rx_failed_count;
 
 	/* Host coalescing block command register */
-	u32 hc_command;
+	u32			hc_command;
 
 	/* Bootcode state */
-	u32 bc_state;
+	u32			bc_state;
 
 #ifdef BCE_DEBUG
 	/* Simulated recoverable failure counters. */
-	u32 mbuf_alloc_failed_sim_count;
-	u32 unexpected_attention_sim_count;
-	u32 l2fhdr_error_sim_count;
-	u32 dma_map_addr_failed_sim_count;
+	u32			mbuf_alloc_failed_sim_count;
+	u32			unexpected_attention_sim_count;
+	u32			l2fhdr_error_sim_count;
+	u32			dma_map_addr_failed_sim_count;
 
 	/* Track the number of enqueued mbufs. */
-	int debug_tx_mbuf_alloc;
-	int debug_rx_mbuf_alloc;
+	int			debug_tx_mbuf_alloc;
+	int			debug_rx_mbuf_alloc;
 
 #ifdef BCE_JUMBO_HDRSPLIT
-	int debug_pg_mbuf_alloc;
+	int			debug_pg_mbuf_alloc;
 #endif
 
 	/* Track how many and what type of interrupts are generated. */
-	u32 interrupts_generated;
-	u32 interrupts_handled;
-	u32 rx_interrupts;
-	u32 tx_interrupts;
-	u32 phy_interrupts;
+	u32			interrupts_generated;
+	u32			interrupts_handled;
+	u32			interrupts_rx;
+	u32			interrupts_tx;
+	u32			phy_interrupts;
 
 	/* Track interrupt time (25MHz clock). */
-	u64 rx_intr_time;
-	u64 tx_intr_time;
+	u64			rx_intr_time;
+	u64			tx_intr_time;
 
 	/* Lowest number of rx_bd's free. */
-	u32 rx_low_watermark;
+	u32			rx_low_watermark;
 
 	/* Number of times the RX chain was empty. */
-	u32 rx_empty_count;
+	u32			rx_empty_count;
 
 #ifdef BCE_JUMBO_HDRSPLIT
-
 	/* Lowest number of pages free. */
-	u32 pg_low_watermark;
+	u32			pg_low_watermark;
 
 	/* Number of times the page chain was empty. */
-	u32 pg_empty_count;
+	u32			pg_empty_count;
 #endif
 
 	/* Greatest number of tx_bd's used. */
-	u32 tx_hi_watermark;
+	u32			tx_hi_watermark;
 
 	/* Number of times the TX chain was full. */
-	u32 tx_full_count;
+	u32			tx_full_count;
 
-	/* Number of TSO frames enqueued. */
-	u32 requested_tso_frames;
+	/* Number of TSO frames requested. */
+	u32			tso_frames_requested;
+
+	/* Number of TSO frames completed. */
+	u32			tso_frames_completed;
+
+	/* Number of TSO frames failed. */
+	u32			tso_frames_failed;
+
+	/* Number of IP checksum offload frames.*/
+	u32			csum_offload_ip;
+
+	/* Number of TCP/UDP checksum offload frames.*/
+	u32			csum_offload_tcp_udp;
+
+	/* Number of VLAN tagged frames received. */
+	u32			vlan_tagged_frames_rcvd;
+
+	/* Number of VLAN tagged frames stripped. */
+	u32			vlan_tagged_frames_stripped;
 #endif
 };
 

From 6d74d042e3d6f8e0242d03ef849ce2579954470d Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 02:46:49 +0000
Subject: [PATCH 345/532] don't allow unsynchronized free in vm_page_unhold

---
 sys/vm/vm_page.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index 4c0d3850669..fb0950eb9d9 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -563,8 +563,11 @@ vm_page_unhold(vm_page_t mem)
 	vm_page_lock_assert(mem, MA_OWNED);
 	--mem->hold_count;
 	KASSERT(mem->hold_count >= 0, ("vm_page_unhold: hold count < 0!!!"));
-	if (mem->hold_count == 0 && VM_PAGE_INQUEUE2(mem, PQ_HOLD))
+	if (mem->hold_count == 0 && VM_PAGE_INQUEUE2(mem, PQ_HOLD)) {
+		vm_page_lock_queues();
 		vm_page_free_toq(mem);
+		vm_page_unlock_queues();
+	}
 }
 
 /*

From feb112c5526565905bce00be48ce80b2b9000af0 Mon Sep 17 00:00:00 2001
From: Alfred Perlstein <alfred@FreeBSD.org>
Date: Fri, 30 Apr 2010 03:10:53 +0000
Subject: [PATCH 346/532] Do not set IO_NODELOCKED while writing to vnodes as
 our consumers do not lock the vnodes.

Obtained from: Juniper Networks

Reviewed by:	jhb
---
 sys/kern/kern_gzio.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/sys/kern/kern_gzio.c b/sys/kern/kern_gzio.c
index 1f9c387dffa..c1b2b3f7025 100644
--- a/sys/kern/kern_gzio.c
+++ b/sys/kern/kern_gzio.c
@@ -166,7 +166,7 @@ gzFile gz_open (path, mode, vp)
              0 /*xflags*/, OS_CODE);
 
     if ((error = vn_rdwr(UIO_WRITE, s->file, buf, GZ_HEADER_LEN, s->outoff,
-                         UIO_SYSSPACE, IO_NODELOCKED|IO_UNIT, curproc->p_ucred,
+                         UIO_SYSSPACE, IO_UNIT, curproc->p_ucred,
                          NOCRED, &resid, curthread))) {
         s->outoff += GZ_HEADER_LEN - resid;
         return destroy(s), (gzFile)Z_NULL;
@@ -234,7 +234,7 @@ int ZEXPORT gzwrite (file, buf, len)
             s->stream.next_out = s->outbuf;
             vfslocked = VFS_LOCK_GIANT(s->file->v_mount);
             error = vn_rdwr_inchunks(UIO_WRITE, s->file, s->outbuf, Z_BUFSIZE,
-                        curoff, UIO_SYSSPACE, IO_NODELOCKED|IO_UNIT,
+                        curoff, UIO_SYSSPACE, IO_UNIT,
                         curproc->p_ucred, NOCRED, &resid, curthread);
             VFS_UNLOCK_GIANT(vfslocked);
             if (error) {
@@ -291,7 +291,7 @@ local int do_flush (file, flush)
         if (len != 0) {
             vfslocked = VFS_LOCK_GIANT(s->file->v_mount);
             error = vn_rdwr_inchunks(UIO_WRITE, s->file, s->outbuf, len, curoff,
-                        UIO_SYSSPACE, IO_NODELOCKED|IO_UNIT, curproc->p_ucred,
+                        UIO_SYSSPACE, IO_UNIT, curproc->p_ucred,
                         NOCRED, &resid, curthread);
             VFS_UNLOCK_GIANT(vfslocked);
 	    if (error) {
@@ -350,7 +350,7 @@ local void putU32 (s, x)
     xx = x;
 #endif
     vn_rdwr(UIO_WRITE, s->file, (caddr_t)&xx, sizeof(xx), curoff,
-            UIO_SYSSPACE, IO_NODELOCKED|IO_UNIT, curproc->p_ucred,
+            UIO_SYSSPACE, IO_UNIT, curproc->p_ucred,
             NOCRED, &resid, curthread);
     s->outoff += sizeof(xx) - resid;
 }

From fba6b1af2e946e86b8a3aac9c561b0717b9729bc Mon Sep 17 00:00:00 2001
From: Alfred Perlstein <alfred@FreeBSD.org>
Date: Fri, 30 Apr 2010 03:13:24 +0000
Subject: [PATCH 347/532] Don't leak core_buf or gzfile if doing a compressed
 core file and we hit an error condition.

Obtained from: Juniper Networks
---
 sys/kern/imgact_elf.c | 11 +++++++----
 1 file changed, 7 insertions(+), 4 deletions(-)

diff --git a/sys/kern/imgact_elf.c b/sys/kern/imgact_elf.c
index 236894f4b2d..c48e0f5f471 100644
--- a/sys/kern/imgact_elf.c
+++ b/sys/kern/imgact_elf.c
@@ -1088,8 +1088,10 @@ __elfN(coredump)(struct thread *td, struct vnode *vp, off_t limit, int flags)
 	hdrsize = 0;
 	__elfN(puthdr)(td, (void *)NULL, &hdrsize, seginfo.count);
 
-	if (hdrsize + seginfo.size >= limit)
-		return (EFAULT);
+	if (hdrsize + seginfo.size >= limit) {
+		error = EFAULT;
+		goto done;
+	}
 
 	/*
 	 * Allocate memory for building the header, fill it up,
@@ -1097,7 +1099,8 @@ __elfN(coredump)(struct thread *td, struct vnode *vp, off_t limit, int flags)
 	 */
 	hdr = malloc(hdrsize, M_TEMP, M_WAITOK);
 	if (hdr == NULL) {
-		return (EINVAL);
+		error = EINVAL;
+		goto done;
 	}
 	error = __elfN(corehdr)(td, vp, cred, seginfo.count, hdr, hdrsize,
 	    gzfile);
@@ -1125,8 +1128,8 @@ __elfN(coredump)(struct thread *td, struct vnode *vp, off_t limit, int flags)
 		    curproc->p_comm, error);
 	}
 
-#ifdef COMPRESS_USER_CORES
 done:
+#ifdef COMPRESS_USER_CORES
 	if (core_buf)
 		free(core_buf, M_TEMP);
 	if (gzfile)

From b7402d8269d7b80ecec75c2cd47678f3155ac9de Mon Sep 17 00:00:00 2001
From: Alfred Perlstein <alfred@FreeBSD.org>
Date: Fri, 30 Apr 2010 03:15:00 +0000
Subject: [PATCH 348/532] Avoid allocating MAXHOSTNAMELEN bytes on the stack in
 expand_name(), use the heap instead.

Obtained from: Juniper Networks

Reviewed by:	jhb
---
 sys/kern/kern_sig.c | 22 +++++++++++++++++++---
 1 file changed, 19 insertions(+), 3 deletions(-)

diff --git a/sys/kern/kern_sig.c b/sys/kern/kern_sig.c
index 706433d3604..789bb612529 100644
--- a/sys/kern/kern_sig.c
+++ b/sys/kern/kern_sig.c
@@ -3004,8 +3004,9 @@ expand_name(const char *name, uid_t uid, pid_t pid, struct thread *td,
 	char *temp;
 	size_t i;
 	int indexpos;
-	char hostname[MAXHOSTNAMELEN];
+	char *hostname;
 	
+	hostname = NULL;
 	format = corefilename;
 	temp = malloc(MAXPATHLEN, M_TEMP, M_NOWAIT | M_ZERO);
 	if (temp == NULL)
@@ -3021,8 +3022,21 @@ expand_name(const char *name, uid_t uid, pid_t pid, struct thread *td,
 				sbuf_putc(&sb, '%');
 				break;
 			case 'H':	/* hostname */
+				if (hostname == NULL) {
+					hostname = malloc(MAXHOSTNAMELEN,
+					    M_TEMP, M_NOWAIT);
+					if (hostname == NULL) {
+						log(LOG_ERR,
+						    "pid %ld (%s), uid (%lu): "
+						    "unable to alloc memory "
+						    "for corefile hostname\n",
+						    (long)pid, name,
+						    (u_long)uid);
+                                                goto nomem;
+                                        }
+                                }
 				getcredhostname(td->td_ucred, hostname,
-				    sizeof(hostname));
+				    MAXHOSTNAMELEN);
 				sbuf_printf(&sb, "%s", hostname);
 				break;
 			case 'I':       /* autoincrementing index */
@@ -3048,15 +3062,17 @@ expand_name(const char *name, uid_t uid, pid_t pid, struct thread *td,
 			sbuf_putc(&sb, format[i]);
 		}
 	}
+	free(hostname, M_TEMP);
 #ifdef COMPRESS_USER_CORES
 	if (compress) {
 		sbuf_printf(&sb, GZ_SUFFIX);
 	}
 #endif
 	if (sbuf_overflowed(&sb)) {
-		sbuf_delete(&sb);
 		log(LOG_ERR, "pid %ld (%s), uid (%lu): corename is too "
 		    "long\n", (long)pid, name, (u_long)uid);
+nomem:
+		sbuf_delete(&sb);
 		free(temp, M_TEMP);
 		return (NULL);
 	}

From 958d87cd86f7268a6074f4b224ec7a3164418cd9 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 03:26:12 +0000
Subject: [PATCH 349/532] merge 194209 in to the i386/xen pmap

requested by: alc@
---
 sys/i386/xen/pmap.c | 95 +++++++++++++++++++++++----------------------
 1 file changed, 48 insertions(+), 47 deletions(-)

diff --git a/sys/i386/xen/pmap.c b/sys/i386/xen/pmap.c
index ec96ef0a41b..1bd081f7fe2 100644
--- a/sys/i386/xen/pmap.c
+++ b/sys/i386/xen/pmap.c
@@ -289,6 +289,12 @@ SYSCTL_INT(_vm_pmap, OID_AUTO, pv_entry_max, CTLFLAG_RD, &pv_entry_max, 0,
 	"Max number of PV entries");
 SYSCTL_INT(_vm_pmap, OID_AUTO, shpgperproc, CTLFLAG_RD, &shpgperproc, 0,
 	"Page share factor per proc");
+SYSCTL_NODE(_vm_pmap, OID_AUTO, pde, CTLFLAG_RD, 0,
+    "2/4MB page mapping counters");
+
+static u_long pmap_pde_mappings;
+SYSCTL_ULONG(_vm_pmap_pde, OID_AUTO, mappings, CTLFLAG_RD,
+    &pmap_pde_mappings, 0, "2/4MB page mappings");
 
 static void	free_pv_entry(pmap_t pmap, pv_entry_t pv);
 static pv_entry_t get_pv_entry(pmap_t locked_pmap, int try);
@@ -3130,64 +3136,59 @@ pmap_object_init_pt(pmap_t pmap, vm_offset_t addr,
 		    vm_object_t object, vm_pindex_t pindex,
 		    vm_size_t size)
 {
+	pd_entry_t *pde;
+	vm_paddr_t pa, ptepa;
 	vm_page_t p;
+	int pat_mode;
 
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	KASSERT(object->type == OBJT_DEVICE || object->type == OBJT_SG,
 	    ("pmap_object_init_pt: non-device object"));
 	if (pseflag && 
-	    ((addr & (NBPDR - 1)) == 0) && ((size & (NBPDR - 1)) == 0)) {
-		int i;
-		vm_page_t m[1];
-		unsigned int ptepindex;
-		int npdes;
-		pd_entry_t ptepa;
-
-		PMAP_LOCK(pmap);
-		if (pmap->pm_pdir[ptepindex = (addr >> PDRSHIFT)])
-			goto out;
-		PMAP_UNLOCK(pmap);
-retry:
+	    (addr & (NBPDR - 1)) == 0 && (size & (NBPDR - 1)) == 0) {
+		if (!vm_object_populate(object, pindex, pindex + atop(size)))
+			return;
 		p = vm_page_lookup(object, pindex);
-		if (p != NULL) {
-			if (vm_page_sleep_if_busy(p, FALSE, "init4p"))
-				goto retry;
-		} else {
-			p = vm_page_alloc(object, pindex, VM_ALLOC_NORMAL);
-			if (p == NULL)
-				return;
-			m[0] = p;
-
-			if (vm_pager_get_pages(object, m, 1, 0) != VM_PAGER_OK) {
-				vm_page_lock_queues();
-				vm_page_free(p);
-				vm_page_unlock_queues();
-				return;
-			}
-
-			p = vm_page_lookup(object, pindex);
-			vm_page_wakeup(p);
-		}
-
+		KASSERT(p->valid == VM_PAGE_BITS_ALL,
+		    ("pmap_object_init_pt: invalid page %p", p));
+		pat_mode = p->md.pat_mode;
+		/*
+		 * Abort the mapping if the first page is not physically
+		 * aligned to a 2/4MB page boundary.
+		 */
 		ptepa = VM_PAGE_TO_PHYS(p);
 		if (ptepa & (NBPDR - 1))
 			return;
-
-		p->valid = VM_PAGE_BITS_ALL;
-
-		PMAP_LOCK(pmap);
-		pmap->pm_stats.resident_count += size >> PAGE_SHIFT;
-		npdes = size >> PDRSHIFT;
-		critical_enter();
-		for(i = 0; i < npdes; i++) {
-			PD_SET_VA(pmap, ptepindex,
-			    ptepa | PG_U | PG_M | PG_RW | PG_V | PG_PS, FALSE);
-			ptepa += NBPDR;
-			ptepindex += 1;
+		/*
+		 * Skip the first page.  Abort the mapping if the rest of
+		 * the pages are not physically contiguous or have differing
+		 * memory attributes.
+		 */
+		p = TAILQ_NEXT(p, listq);
+		for (pa = ptepa + PAGE_SIZE; pa < ptepa + size;
+		    pa += PAGE_SIZE) {
+			KASSERT(p->valid == VM_PAGE_BITS_ALL,
+			    ("pmap_object_init_pt: invalid page %p", p));
+			if (pa != VM_PAGE_TO_PHYS(p) ||
+			    pat_mode != p->md.pat_mode)
+				return;
+			p = TAILQ_NEXT(p, listq);
+		}
+		/* Map using 2/4MB pages. */
+		PMAP_LOCK(pmap);
+		for (pa = ptepa | pmap_cache_bits(pat_mode, 1); pa < ptepa +
+		    size; pa += NBPDR) {
+			pde = pmap_pde(pmap, addr);
+			if (*pde == 0) {
+				pde_store(pde, pa | PG_PS | PG_M | PG_A |
+				    PG_U | PG_RW | PG_V);
+				pmap->pm_stats.resident_count += NBPDR /
+				    PAGE_SIZE;
+				pmap_pde_mappings++;
+			}
+			/* Else continue on if the PDE is already valid. */
+			addr += NBPDR;
 		}
-		pmap_invalidate_all(pmap);
-		critical_exit();
-out:
 		PMAP_UNLOCK(pmap);
 	}
 }

From a6e09ef1c209ad765e5db1d2d16e09d4f7ee4e8f Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Fri, 30 Apr 2010 04:21:22 +0000
Subject: [PATCH 350/532]  - Use the path to the filesystem mountpoint to look
 up the statfs    structure so that we correctly reload.  Note that tunefs
 doesn't    properly detect the need to reload if the disk device is specified
    for a read-only mounted filesystem.  - Lessen the contiguity requirement
 for the journal so that it is more    likely to succeed.

---
 sbin/tunefs/tunefs.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/sbin/tunefs/tunefs.c b/sbin/tunefs/tunefs.c
index a10b35daec2..7987620bc59 100644
--- a/sbin/tunefs/tunefs.c
+++ b/sbin/tunefs/tunefs.c
@@ -280,9 +280,9 @@ main(int argc, char *argv[])
 	if (ufs_disk_fillout(&disk, special) == -1)
 		goto err;
 	if (disk.d_name != special) {
-		special = disk.d_name;
-		if (statfs(special, &stfs) == 0 &&
-		    strcmp(special, stfs.f_mntonname) == 0)
+		if (statfs(special, &stfs) != 0)
+			warn("Can't stat %s", special);
+		if (strcmp(special, stfs.f_mntonname) == 0)
 			active = 1;
 	}
 
@@ -546,7 +546,7 @@ journal_balloc(void)
 			 * Try to minimize fragmentation by requiring a minimum
 			 * number of blocks present.
 			 */
-			if (cgp->cg_cs.cs_nbfree > blocks / 8)
+			if (cgp->cg_cs.cs_nbfree > 128 * 1024 * 1024)
 				break;
 			if (contig == 0 && cgp->cg_cs.cs_nbfree)
 				break;

From 8016863c91a6c51579e2bbc04d86d9252d4882b4 Mon Sep 17 00:00:00 2001
From: Edward Tomasz Napierala <trasz@FreeBSD.org>
Date: Fri, 30 Apr 2010 07:09:13 +0000
Subject: [PATCH 351/532] Avoid undefined behaviour.

Reviewed by:	zec@
---
 sys/netgraph/ng_pipe.c | 19 +++++++++++--------
 1 file changed, 11 insertions(+), 8 deletions(-)

diff --git a/sys/netgraph/ng_pipe.c b/sys/netgraph/ng_pipe.c
index a094646e660..a4546248dd9 100644
--- a/sys/netgraph/ng_pipe.c
+++ b/sys/netgraph/ng_pipe.c
@@ -816,14 +816,17 @@ pipe_dequeue(struct hookinfo *hinfo, struct timeval *now) {
 		}
 
 		/* Randomly discard the frame, according to BER setting */
-		if (hinfo->cfg.ber && 
-		    ((oldrand = rand) ^ (rand = random())<<17) >=
-		    hinfo->ber_p[priv->overhead + m->m_pkthdr.len] ) {
-			hinfo->stats.out_disc_frames++;
-			hinfo->stats.out_disc_octets += m->m_pkthdr.len;
-			uma_zfree(ngp_zone, ngp_h);
-			m_freem(m);
-			continue;
+		if (hinfo->cfg.ber) {
+			oldrand = rand;
+			rand = random();
+			if (((oldrand ^ rand) << 17) >=
+			    hinfo->ber_p[priv->overhead + m->m_pkthdr.len]) {
+				hinfo->stats.out_disc_frames++;
+				hinfo->stats.out_disc_octets += m->m_pkthdr.len;
+				uma_zfree(ngp_zone, ngp_h);
+				m_freem(m);
+				continue;
+			}
 		}
 
 		/* Discard frame if outbound queue size limit exceeded */

From d8665eb1f6c066967cedf5120ecd07af1d285c78 Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Fri, 30 Apr 2010 07:48:29 +0000
Subject: [PATCH 352/532] Fix improper pool write throughput calculation.

OpenSolaris onnv revision:	9366:17553395a745

PR:		kern/146108
Approved by:	pjd, delphij (mentor)
Obtained from:	OpenSolaris, Bug ID 6817339
MFC after:	2 weeks
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_pool.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_pool.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_pool.c
index 03af3d1583c..58a79cac24b 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_pool.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dsl_pool.c
@@ -300,6 +300,7 @@ dsl_pool_sync(dsl_pool_t *dp, uint64_t txg)
 	tx = dmu_tx_create_assigned(dp, txg);
 
 	dp->dp_read_overhead = 0;
+	start = gethrtime();
 	zio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
 	while (ds = txg_list_remove(&dp->dp_dirty_datasets, txg)) {
 		if (!list_link_active(&ds->ds_synced_link))
@@ -310,7 +311,6 @@ dsl_pool_sync(dsl_pool_t *dp, uint64_t txg)
 	}
 	DTRACE_PROBE(pool_sync__1setup);
 
-	start = gethrtime();
 	err = zio_wait(zio);
 	write_time = gethrtime() - start;
 	ASSERT(err == 0);

From e2a751891fb250e505d5f2982b7df48bf134377d Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Fri, 30 Apr 2010 07:56:21 +0000
Subject: [PATCH 353/532] Report PMP absence using target 15, same as for
 precence (not a wildcard), to not confuse target ID checks at SIMs.

---
 sys/cam/ata/ata_xpt.c | 10 ++++++----
 1 file changed, 6 insertions(+), 4 deletions(-)

diff --git a/sys/cam/ata/ata_xpt.c b/sys/cam/ata/ata_xpt.c
index 10e61827ad7..bd748cd4f45 100644
--- a/sys/cam/ata/ata_xpt.c
+++ b/sys/cam/ata/ata_xpt.c
@@ -1118,13 +1118,13 @@ ata_scan_bus(struct cam_periph *periph, union ccb *request_ccb)
 		work_ccb = request_ccb;
 		/* Reuse the same CCB to query if a device was really found */
 		scan_info = (ata_scan_bus_info *)work_ccb->ccb_h.ppriv_ptr0;
-		/* Free the current request path- we're done with it. */
-		xpt_free_path(work_ccb->ccb_h.path);
 		/* If there is PMP... */
 		if ((scan_info->cpi->hba_inquiry & PI_SATAPM) &&
 		    (scan_info->counter == scan_info->cpi->max_target)) {
 			if (work_ccb->ccb_h.status == CAM_REQ_CMP) {
-				/* everything else willbe probed by it */
+				/* everything else will be probed by it */
+				/* Free the current request path- we're done with it. */
+				xpt_free_path(work_ccb->ccb_h.path);
 				goto done;
 			} else {
 				struct ccb_trans_settings cts;
@@ -1132,7 +1132,7 @@ ata_scan_bus(struct cam_periph *periph, union ccb *request_ccb)
 				/* Report SIM that PM is absent. */
 				bzero(&cts, sizeof(cts));
 				xpt_setup_ccb(&cts.ccb_h,
-				    scan_info->request_ccb->ccb_h.path, 1);
+				    work_ccb->ccb_h.path, CAM_PRIORITY_NONE);
 				cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
 				cts.type = CTS_TYPE_CURRENT_SETTINGS;
 				cts.xport_specific.sata.pm_present = 0;
@@ -1140,6 +1140,8 @@ ata_scan_bus(struct cam_periph *periph, union ccb *request_ccb)
 				xpt_action((union ccb *)&cts);
 			}
 		}
+		/* Free the current request path- we're done with it. */
+		xpt_free_path(work_ccb->ccb_h.path);
 		if (scan_info->counter ==
 		    ((scan_info->cpi->hba_inquiry & PI_SATAPM) ?
 		    0 : scan_info->cpi->max_target)) {

From 89f2835fc9689c5d897d0d069fd7e1d64bc52f27 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Fri, 30 Apr 2010 08:02:12 +0000
Subject: [PATCH 354/532] Add Target/LUN ID checks and deny access to targets
 1-14 when PMP absent. Enforce PMA bit clearing when PMP detached to avoid
 further scan timeouts.

---
 sys/dev/ahci/ahci.c | 69 ++++++++++++++++++++-------------------------
 1 file changed, 30 insertions(+), 39 deletions(-)

diff --git a/sys/dev/ahci/ahci.c b/sys/dev/ahci/ahci.c
index 5d399683945..34f1be426b1 100644
--- a/sys/dev/ahci/ahci.c
+++ b/sys/dev/ahci/ahci.c
@@ -2083,6 +2083,7 @@ ahci_start(device_t dev, int fbs)
 	}
 	/* Start operations on this channel */
 	cmd = ATA_INL(ch->r_mem, AHCI_P_CMD);
+	cmd &= ~AHCI_P_CMD_PMA;
 	ATA_OUTL(ch->r_mem, AHCI_P_CMD, cmd | AHCI_P_CMD_ST |
 	    (ch->pm_present ? AHCI_P_CMD_PMA : 0));
 }
@@ -2381,6 +2382,24 @@ ahci_sata_phy_reset(device_t dev)
 	return (1);
 }
 
+static int
+ahci_check_ids(device_t dev, union ccb *ccb)
+{
+	struct ahci_channel *ch = device_get_softc(dev);
+
+	if (ccb->ccb_h.target_id > ((ch->caps & AHCI_CAP_SPM) ? 15 : 0)) {
+		ccb->ccb_h.status = CAM_TID_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	if (ccb->ccb_h.target_lun != 0) {
+		ccb->ccb_h.status = CAM_LUN_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	return (0);
+}
+
 static void
 ahciaction(struct cam_sim *sim, union ccb *ccb)
 {
@@ -2396,9 +2415,12 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 	/* Common cases first */
 	case XPT_ATA_IO:	/* Execute the requested I/O operation */
 	case XPT_SCSI_IO:
-		if (ch->devices == 0) {
+		if (ahci_check_ids(dev, ccb))
+			return;
+		if (ch->devices == 0 ||
+		    (ch->pm_present == 0 &&
+		     ccb->ccb_h.target_id > 0 && ccb->ccb_h.target_id < 15)) {
 			ccb->ccb_h.status = CAM_SEL_TIMEOUT;
-			xpt_done(ccb);
 			break;
 		}
 		/* Check for command collision. */
@@ -2410,7 +2432,7 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 			return;
 		}
 		ahci_begin_transaction(dev, ccb);
-		break;
+		return;
 	case XPT_EN_LUN:		/* Enable LUN as a target */
 	case XPT_TARGET_IO:		/* Execute target I/O request */
 	case XPT_ACCEPT_TARGET_IO:	/* Accept Host Target Mode CDB */
@@ -2418,13 +2440,14 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 	case XPT_ABORT:			/* Abort the specified CCB */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_SET_TRAN_SETTINGS:
 	{
 		struct	ccb_trans_settings *cts = &ccb->cts;
 		struct	ahci_device *d; 
 
+		if (ahci_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -2442,7 +2465,6 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_ATAPI)
 			d->atapi = cts->xport_specific.sata.atapi;
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
 	case XPT_GET_TRAN_SETTINGS:
@@ -2452,6 +2474,8 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 		struct  ahci_device *d;
 		uint32_t status;
 
+		if (ahci_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -2487,48 +2511,16 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 		cts->xport_specific.sata.atapi = d->atapi;
 		cts->xport_specific.sata.valid |= CTS_SATA_VALID_ATAPI;
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
-#if 0
-	case XPT_CALC_GEOMETRY:
-	{
-		struct	  ccb_calc_geometry *ccg;
-		uint32_t size_mb;
-		uint32_t secs_per_cylinder;
-
-		ccg = &ccb->ccg;
-		size_mb = ccg->volume_size
-			/ ((1024L * 1024L) / ccg->block_size);
-		if (size_mb >= 1024 && (aha->extended_trans != 0)) {
-			if (size_mb >= 2048) {
-				ccg->heads = 255;
-				ccg->secs_per_track = 63;
-			} else {
-				ccg->heads = 128;
-				ccg->secs_per_track = 32;
-			}
-		} else {
-			ccg->heads = 64;
-			ccg->secs_per_track = 32;
-		}
-		secs_per_cylinder = ccg->heads * ccg->secs_per_track;
-		ccg->cylinders = ccg->volume_size / secs_per_cylinder;
-		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
-		break;
-	}
-#endif
 	case XPT_RESET_BUS:		/* Reset the specified SCSI bus */
 	case XPT_RESET_DEV:	/* Bus Device Reset the specified SCSI device */
 		ahci_reset(dev);
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	case XPT_TERM_IO:		/* Terminate the I/O process */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_PATH_INQ:		/* Path routing inquiry */
 	{
@@ -2564,14 +2556,13 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 		if (pci_get_devid(device_get_parent(dev)) == 0x43801002)
 			cpi->maxio = min(cpi->maxio, 128 * 512);
 		cpi->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
 	default:
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	}
+	xpt_done(ccb);
 }
 
 static void

From 81a75ae3655b9a074963eb27a010c1d284cb5d79 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Fri, 30 Apr 2010 08:22:47 +0000
Subject: [PATCH 355/532] Add Target/LUN ID checks and deny access to targets
 1-14 when PMP absent.

---
 sys/dev/siis/siis.c | 67 +++++++++++++++++++--------------------------
 1 file changed, 28 insertions(+), 39 deletions(-)

diff --git a/sys/dev/siis/siis.c b/sys/dev/siis/siis.c
index c9c157226fa..c60d8e47f69 100644
--- a/sys/dev/siis/siis.c
+++ b/sys/dev/siis/siis.c
@@ -1610,6 +1610,23 @@ siis_sata_connect(struct siis_channel *ch)
 	return (1);
 }
 
+static int
+siis_check_ids(device_t dev, union ccb *ccb)
+{
+
+	if (ccb->ccb_h.target_id > 15) {
+		ccb->ccb_h.status = CAM_TID_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	if (ccb->ccb_h.target_lun != 0) {
+		ccb->ccb_h.status = CAM_LUN_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	return (0);
+}
+
 static void
 siisaction(struct cam_sim *sim, union ccb *ccb)
 {
@@ -1626,9 +1643,12 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 	/* Common cases first */
 	case XPT_ATA_IO:	/* Execute the requested I/O operation */
 	case XPT_SCSI_IO:
-		if (ch->devices == 0) {
+		if (siis_check_ids(dev, ccb))
+			return;
+		if (ch->devices == 0 ||
+		    (ch->pm_present == 0 &&
+		     ccb->ccb_h.target_id > 0 && ccb->ccb_h.target_id < 15)) {
 			ccb->ccb_h.status = CAM_SEL_TIMEOUT;
-			xpt_done(ccb);
 			break;
 		}
 		/* Check for command collision. */
@@ -1640,7 +1660,7 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 			return;
 		}
 		siis_begin_transaction(dev, ccb);
-		break;
+		return;
 	case XPT_EN_LUN:		/* Enable LUN as a target */
 	case XPT_TARGET_IO:		/* Execute target I/O request */
 	case XPT_ACCEPT_TARGET_IO:	/* Accept Host Target Mode CDB */
@@ -1648,13 +1668,14 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 	case XPT_ABORT:			/* Abort the specified CCB */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_SET_TRAN_SETTINGS:
 	{
 		struct	ccb_trans_settings *cts = &ccb->cts;
 		struct	siis_device *d; 
 
+		if (siis_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -1677,7 +1698,6 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_TAGS)
 			d->atapi = cts->xport_specific.sata.atapi;
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
 	case XPT_GET_TRAN_SETTINGS:
@@ -1687,6 +1707,8 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 		struct  siis_device *d;
 		uint32_t status;
 
+		if (siis_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -1722,48 +1744,16 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 		cts->xport_specific.sata.atapi = d->atapi;
 		cts->xport_specific.sata.valid |= CTS_SATA_VALID_ATAPI;
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
-#if 0
-	case XPT_CALC_GEOMETRY:
-	{
-		struct	  ccb_calc_geometry *ccg;
-		uint32_t size_mb;
-		uint32_t secs_per_cylinder;
-
-		ccg = &ccb->ccg;
-		size_mb = ccg->volume_size
-			/ ((1024L * 1024L) / ccg->block_size);
-		if (size_mb >= 1024 && (aha->extended_trans != 0)) {
-			if (size_mb >= 2048) {
-				ccg->heads = 255;
-				ccg->secs_per_track = 63;
-			} else {
-				ccg->heads = 128;
-				ccg->secs_per_track = 32;
-			}
-		} else {
-			ccg->heads = 64;
-			ccg->secs_per_track = 32;
-		}
-		secs_per_cylinder = ccg->heads * ccg->secs_per_track;
-		ccg->cylinders = ccg->volume_size / secs_per_cylinder;
-		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
-		break;
-	}
-#endif
 	case XPT_RESET_BUS:		/* Reset the specified SCSI bus */
 	case XPT_RESET_DEV:	/* Bus Device Reset the specified SCSI device */
 		siis_reset(dev);
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	case XPT_TERM_IO:		/* Terminate the I/O process */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_PATH_INQ:		/* Path routing inquiry */
 	{
@@ -1790,14 +1780,13 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 		cpi->protocol_version = PROTO_VERSION_UNSPECIFIED;
 		cpi->ccb_h.status = CAM_REQ_CMP;
 		cpi->maxio = MAXPHYS;
-		xpt_done(ccb);
 		break;
 	}
 	default:
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	}
+	xpt_done(ccb);
 }
 
 static void

From 5b228a7a86e43281be5606a9105dae5284847dc5 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Fri, 30 Apr 2010 08:37:00 +0000
Subject: [PATCH 356/532] Add Target/LUN ID checks.

---
 sys/dev/ata/ata-all.c | 65 +++++++++++++++++--------------------------
 1 file changed, 26 insertions(+), 39 deletions(-)

diff --git a/sys/dev/ata/ata-all.c b/sys/dev/ata/ata-all.c
index 73dfa94fd83..5c6f0db714f 100644
--- a/sys/dev/ata/ata-all.c
+++ b/sys/dev/ata/ata-all.c
@@ -1430,6 +1430,24 @@ ata_cam_end_transaction(device_t dev, struct ata_request *request)
 		ata_reinit(dev);
 }
 
+static int
+ata_check_ids(device_t dev, union ccb *ccb)
+{
+	struct ata_channel *ch = device_get_softc(dev);
+
+	if (ccb->ccb_h.target_id > ((ch->flags & ATA_NO_SLAVE) ? 0 : 1)) {
+		ccb->ccb_h.status = CAM_TID_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	if (ccb->ccb_h.target_lun != 0) {
+		ccb->ccb_h.status = CAM_LUN_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	return (0);
+}
+
 static void
 ataaction(struct cam_sim *sim, union ccb *ccb)
 {
@@ -1445,10 +1463,11 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 	/* Common cases first */
 	case XPT_ATA_IO:	/* Execute the requested I/O operation */
 	case XPT_SCSI_IO:
+		if (ata_check_ids(dev, ccb))
+			return;
 		if ((ch->devices & ((ATA_ATA_MASTER | ATA_ATAPI_MASTER)
 		    << ccb->ccb_h.target_id)) == 0) {
 			ccb->ccb_h.status = CAM_SEL_TIMEOUT;
-			xpt_done(ccb);
 			break;
 		}
 		if (ch->running)
@@ -1467,11 +1486,10 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 				res->lba_mid = 0x14;
 			}
 			ccb->ccb_h.status = CAM_REQ_CMP;
-			xpt_done(ccb);
 			break;
 		}
 		ata_cam_begin_transaction(dev, ccb);
-		break;
+		return;
 	case XPT_EN_LUN:		/* Enable LUN as a target */
 	case XPT_TARGET_IO:		/* Execute target I/O request */
 	case XPT_ACCEPT_TARGET_IO:	/* Accept Host Target Mode CDB */
@@ -1479,13 +1497,14 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 	case XPT_ABORT:			/* Abort the specified CCB */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_SET_TRAN_SETTINGS:
 	{
 		struct	ccb_trans_settings *cts = &ccb->cts;
 		struct	ata_cam_device *d; 
 
+		if (ata_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -1520,7 +1539,6 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 				d->atapi = cts->xport_specific.ata.atapi;
 		}
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
 	case XPT_GET_TRAN_SETTINGS:
@@ -1528,6 +1546,8 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 		struct	ccb_trans_settings *cts = &ccb->cts;
 		struct  ata_cam_device *d;
 
+		if (ata_check_ids(dev, ccb))
+			return;
 		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
 			d = &ch->curr[ccb->ccb_h.target_id];
 		else
@@ -1567,48 +1587,16 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 			cts->xport_specific.ata.valid |= CTS_ATA_VALID_ATAPI;
 		}
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
-#if 0
-	case XPT_CALC_GEOMETRY:
-	{
-		struct	  ccb_calc_geometry *ccg;
-		uint32_t size_mb;
-		uint32_t secs_per_cylinder;
-
-		ccg = &ccb->ccg;
-		size_mb = ccg->volume_size
-			/ ((1024L * 1024L) / ccg->block_size);
-		if (size_mb >= 1024 && (aha->extended_trans != 0)) {
-			if (size_mb >= 2048) {
-				ccg->heads = 255;
-				ccg->secs_per_track = 63;
-			} else {
-				ccg->heads = 128;
-				ccg->secs_per_track = 32;
-			}
-		} else {
-			ccg->heads = 64;
-			ccg->secs_per_track = 32;
-		}
-		secs_per_cylinder = ccg->heads * ccg->secs_per_track;
-		ccg->cylinders = ccg->volume_size / secs_per_cylinder;
-		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
-		break;
-	}
-#endif
 	case XPT_RESET_BUS:		/* Reset the specified SCSI bus */
 	case XPT_RESET_DEV:	/* Bus Device Reset the specified SCSI device */
 		ata_reinit(dev);
 		ccb->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	case XPT_TERM_IO:		/* Terminate the I/O process */
 		/* XXX Implement */
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	case XPT_PATH_INQ:		/* Path routing inquiry */
 	{
@@ -1643,14 +1631,13 @@ ataaction(struct cam_sim *sim, union ccb *ccb)
 		cpi->protocol_version = PROTO_VERSION_UNSPECIFIED;
 		cpi->maxio = ch->dma.max_iosize ? ch->dma.max_iosize : DFLTPHYS;
 		cpi->ccb_h.status = CAM_REQ_CMP;
-		xpt_done(ccb);
 		break;
 	}
 	default:
 		ccb->ccb_h.status = CAM_REQ_INVALID;
-		xpt_done(ccb);
 		break;
 	}
+	xpt_done(ccb);
 }
 
 static void

From 6ee4d8696fee64731a078e2c69fabe4fb4f6c435 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Fri, 30 Apr 2010 08:57:03 +0000
Subject: [PATCH 357/532] Revert r198705. As scottl@ noticed,
 max_target/max_lun was intended to be only a hint for existing bus scanner.
 Some FC/SAS SIMs report fake values there, that are smaller then maximum
 supported IDs. In that case this check makes impossible manual scan outside
 hinted range.

For ATA/SATA SIMs respective check was instead implemented at SIM level.
Newer SCSI SIMs expected to have these checks at driver or firmware level.
Some older SCSI SIMs have no this check and the issues will get back there.
---
 sys/cam/cam_xpt.c | 29 +----------------------------
 1 file changed, 1 insertion(+), 28 deletions(-)

diff --git a/sys/cam/cam_xpt.c b/sys/cam/cam_xpt.c
index 1812edbf70e..ed05c40f729 100644
--- a/sys/cam/cam_xpt.c
+++ b/sys/cam/cam_xpt.c
@@ -467,34 +467,7 @@ xptioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flag, struct thread *td
 			ccb = xpt_alloc_ccb();
 
 			CAM_SIM_LOCK(bus->sim);
-			/* Ensure passed in target/lun supported on this bus. */
-			if ((inccb->ccb_h.target_id != CAM_TARGET_WILDCARD) ||
-			    (inccb->ccb_h.target_lun != CAM_LUN_WILDCARD)) {
-				if (xpt_create_path(&ccb->ccb_h.path,
-					    xpt_periph,
-					    inccb->ccb_h.path_id,
-					    CAM_TARGET_WILDCARD,
-					    CAM_LUN_WILDCARD) != CAM_REQ_CMP) {
-					error = EINVAL;
-					CAM_SIM_UNLOCK(bus->sim);
-					xpt_free_ccb(ccb);
-					break;
-				}
-				xpt_setup_ccb(&ccb->ccb_h, ccb->ccb_h.path,
-				    inccb->ccb_h.pinfo.priority);
-				ccb->ccb_h.func_code = XPT_PATH_INQ;
-				xpt_action(ccb);
-				xpt_free_path(ccb->ccb_h.path);
-				if ((inccb->ccb_h.target_id != CAM_TARGET_WILDCARD &&
-				    inccb->ccb_h.target_id > ccb->cpi.max_target) ||
-				    (inccb->ccb_h.target_lun != CAM_LUN_WILDCARD &&
-				    inccb->ccb_h.target_lun > ccb->cpi.max_lun)) {
-					error = EINVAL;
-					CAM_SIM_UNLOCK(bus->sim);
-					xpt_free_ccb(ccb);
-					break;
-				}
-			}
+
 			/*
 			 * Create a path using the bus, target, and lun the
 			 * user passed in.

From 63a88d586e0bd37eb1232105454878133062da99 Mon Sep 17 00:00:00 2001
From: Philip Paeps <philip@FreeBSD.org>
Date: Fri, 30 Apr 2010 15:12:30 +0000
Subject: [PATCH 358/532] Fix typo in usage message.  Add -n to the list of
 valid arguments.

MFC after:	1 day
---
 tools/tools/nanobsd/nanobsd.sh | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/tools/tools/nanobsd/nanobsd.sh b/tools/tools/nanobsd/nanobsd.sh
index ce518a5bdef..7afb047a3fd 100644
--- a/tools/tools/nanobsd/nanobsd.sh
+++ b/tools/tools/nanobsd/nanobsd.sh
@@ -682,12 +682,12 @@ pprint() {
 
 usage () {
 	(
-	echo "Usage: $0 [-bikqvw] [-c config_file]"
+	echo "Usage: $0 [-biknqvw] [-c config_file]"
 	echo "	-b	suppress builds (both kernel and world)"
 	echo "	-i	suppress disk image build"
 	echo "	-k	suppress buildkernel"
 	echo "	-n	add -DNO_CLEAN to buildworld, buildkernel, etc"
-	echo "	-q	make output more quite"
+	echo "	-q	make output more quiet"
 	echo "	-v	make output more verbose"
 	echo "	-w	suppress buildworld"
 	echo "	-c	specify config file"

From c7a0df65b1178a53ba72f4228569f9d917ca16f3 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 30 Apr 2010 15:22:52 +0000
Subject: [PATCH 359/532] MFamd64/i386 r207205   Clearing a page table entry's
 accessed bit and setting the page's   PG_REFERENCED flag in pmap_protect()
 can't really be justified, so   don't do it.

Additionally, two changes that make this pmap behave like the others do:

Change pmap_protect() such that it calls vm_page_dirty() only if the
page is managed.

Change pmap_remove_write() such that it doesn't clear a page table
entry's accessed bit.
---
 sys/powerpc/booke/pmap.c | 16 +++-------------
 1 file changed, 3 insertions(+), 13 deletions(-)

diff --git a/sys/powerpc/booke/pmap.c b/sys/powerpc/booke/pmap.c
index 0b0fce4485e..78e9afb5530 100644
--- a/sys/powerpc/booke/pmap.c
+++ b/sys/powerpc/booke/pmap.c
@@ -1915,16 +1915,11 @@ mmu_booke_protect(mmu_t mmu, pmap_t pmap, vm_offset_t sva, vm_offset_t eva,
 				tlb_miss_lock();
 
 				/* Handle modified pages. */
-				if (PTE_ISMODIFIED(pte))
+				if (PTE_ISMODIFIED(pte) && PTE_ISMANAGED(pte))
 					vm_page_dirty(m);
 
-				/* Referenced pages. */
-				if (PTE_ISREFERENCED(pte))
-					vm_page_flag_set(m, PG_REFERENCED);
-
 				tlb0_flush_entry(va);
-				pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED |
-				    PTE_REFERENCED);
+				pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED);
 
 				tlb_miss_unlock();
 				mtx_unlock_spin(&tlbivax_mutex);
@@ -1962,13 +1957,8 @@ mmu_booke_remove_write(mmu_t mmu, vm_page_t m)
 				if (PTE_ISMODIFIED(pte))
 					vm_page_dirty(m);
 
-				/* Referenced pages. */
-				if (PTE_ISREFERENCED(pte))
-					vm_page_flag_set(m, PG_REFERENCED);
-
 				/* Flush mapping from TLB0. */
-				pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED |
-				    PTE_REFERENCED);
+				pte->flags &= ~(PTE_UW | PTE_SW | PTE_MODIFIED);
 
 				tlb_miss_unlock();
 				mtx_unlock_spin(&tlbivax_mutex);

From e20e8c15588ea7f2876c56e99677ee00941ef83b Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Fri, 30 Apr 2010 16:20:14 +0000
Subject: [PATCH 360/532] Unlock page lock instead of recursively locking it.

---
 sys/vm/vm_fault.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index de7491518c2..d9cf63e9b11 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -1284,19 +1284,19 @@ vm_fault_copy_entry(vm_map_t dst_map, vm_map_t src_map,
 			vm_page_lock_queues();
 			vm_page_unwire(src_m, 0);
 			vm_page_unlock_queues();
-			vm_page_lock(src_m);
+			vm_page_unlock(src_m);
 
 			vm_page_lock(dst_m);
 			vm_page_lock_queues();
 			vm_page_wire(dst_m);
 			vm_page_unlock_queues();
-			vm_page_lock(dst_m);
+			vm_page_unlock(dst_m);
 		} else {
 			vm_page_lock(dst_m);
 			vm_page_lock_queues();
 			vm_page_activate(dst_m);
 			vm_page_unlock_queues();
-			vm_page_lock(dst_m);
+			vm_page_unlock(dst_m);
 		}
 		vm_page_wakeup(dst_m);
 	}

From 1dac2224197084100d4dbce5fd0f7713157d4c4e Mon Sep 17 00:00:00 2001
From: Zachary Loafman <zml@FreeBSD.org>
Date: Fri, 30 Apr 2010 16:29:05 +0000
Subject: [PATCH 361/532] Handle taskqueue_drain(9) correctly on a threaded
 taskqueue:

taskqueue_drain(9) will not correctly detect whether a task is
currently running.  The check is against a field in the taskqueue
struct, but for a threaded queue with more than one thread, multiple
threads can simultaneously be running a task, thus stomping over the
tq_running field.

Submitted by:       Matthew Fleming <matthew.fleming@isilon.com>
Reviewed by:        jhb
Approved by:        dfr (mentor)
---
 sys/kern/subr_taskqueue.c | 11 ++++++-----
 sys/sys/_task.h           | 16 +++++++++++-----
 sys/sys/taskqueue.h       |  1 +
 3 files changed, 18 insertions(+), 10 deletions(-)

diff --git a/sys/kern/subr_taskqueue.c b/sys/kern/subr_taskqueue.c
index 8405b3d642b..3b1826984db 100644
--- a/sys/kern/subr_taskqueue.c
+++ b/sys/kern/subr_taskqueue.c
@@ -51,7 +51,6 @@ struct taskqueue {
 	const char		*tq_name;
 	taskqueue_enqueue_fn	tq_enqueue;
 	void			*tq_context;
-	struct task		*tq_running;
 	struct mtx		tq_mutex;
 	struct thread		**tq_threads;
 	int			tq_tcount;
@@ -233,13 +232,13 @@ taskqueue_run(struct taskqueue *queue)
 		STAILQ_REMOVE_HEAD(&queue->tq_queue, ta_link);
 		pending = task->ta_pending;
 		task->ta_pending = 0;
-		queue->tq_running = task;
+		task->ta_flags |= TA_FLAGS_RUNNING;
 		TQ_UNLOCK(queue);
 
 		task->ta_func(task->ta_context, pending);
 
 		TQ_LOCK(queue);
-		queue->tq_running = NULL;
+		task->ta_flags &= ~TA_FLAGS_RUNNING;
 		wakeup(task);
 	}
 
@@ -256,14 +255,16 @@ taskqueue_drain(struct taskqueue *queue, struct task *task)
 {
 	if (queue->tq_spin) {		/* XXX */
 		mtx_lock_spin(&queue->tq_mutex);
-		while (task->ta_pending != 0 || task == queue->tq_running)
+		while (task->ta_pending != 0 ||
+		    (task->ta_flags & TA_FLAGS_RUNNING) != 0)
 			msleep_spin(task, &queue->tq_mutex, "-", 0);
 		mtx_unlock_spin(&queue->tq_mutex);
 	} else {
 		WITNESS_WARN(WARN_GIANTOK | WARN_SLEEPOK, NULL, __func__);
 
 		mtx_lock(&queue->tq_mutex);
-		while (task->ta_pending != 0 || task == queue->tq_running)
+		while (task->ta_pending != 0 ||
+		    (task->ta_flags & TA_FLAGS_RUNNING) != 0)
 			msleep(task, &queue->tq_mutex, PWAIT, "-", 0);
 		mtx_unlock(&queue->tq_mutex);
 	}
diff --git a/sys/sys/_task.h b/sys/sys/_task.h
index 2a51e1b07b9..781cd48cf65 100644
--- a/sys/sys/_task.h
+++ b/sys/sys/_task.h
@@ -36,15 +36,21 @@
  * taskqueue_run().  The first argument is taken from the 'ta_context'
  * field of struct task and the second argument is a count of how many
  * times the task was enqueued before the call to taskqueue_run().
+ *
+ * List of locks
+ * (c)	const after init
+ * (q)	taskqueue lock
  */
 typedef void task_fn_t(void *context, int pending);
 
 struct task {
-	STAILQ_ENTRY(task) ta_link;	/* link for queue */
-	u_short	ta_pending;		/* count times queued */
-	u_short	ta_priority;		/* Priority */
-	task_fn_t *ta_func;		/* task handler */
-	void	*ta_context;		/* argument for handler */
+	STAILQ_ENTRY(task) ta_link;	/* (q) link for queue */
+	u_int	ta_flags;		/* (q) state of this task */
+#define	TA_FLAGS_RUNNING	0x01
+	u_short	ta_pending;		/* (q) count times queued */
+	u_short	ta_priority;		/* (c) Priority */
+	task_fn_t *ta_func;		/* (c) task handler */
+	void	*ta_context;		/* (c) argument for handler */
 };
 
 #endif /* !_SYS__TASK_H_ */
diff --git a/sys/sys/taskqueue.h b/sys/sys/taskqueue.h
index bf2e4ee31d0..fa8897d6dce 100644
--- a/sys/sys/taskqueue.h
+++ b/sys/sys/taskqueue.h
@@ -75,6 +75,7 @@ void	taskqueue_thread_enqueue(void *context);
 	(task)->ta_priority = (priority);		\
 	(task)->ta_func = (func);			\
 	(task)->ta_context = (context);			\
+	(task)->ta_flags = 0;				\
 } while (0)
 
 /*

From ad30c76f940ed881a9d29f79fcac4bad8174d934 Mon Sep 17 00:00:00 2001
From: Randall Stewart <rrs@FreeBSD.org>
Date: Fri, 30 Apr 2010 17:12:20 +0000
Subject: [PATCH 362/532] Bug in the memory mapping module. The wrong physaddr
 was being used in the macro (1 should be used not 2)...

Obtained from:	JC
---
 sys/mips/mips/pmap.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index 826177fb70a..af9ef432037 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -236,7 +236,7 @@ caddr_t virtual_sys_start = (caddr_t)0;
 	sched_pin();							\
 	va1 = sysm->base;						\
 	va2 = sysm->base + PAGE_SIZE;					\
-	npte = mips_paddr_to_tlbpfn(phys2) |				\
+	npte = mips_paddr_to_tlbpfn(phys1) |				\
 	    PTE_RW | PTE_V | PTE_G | PTE_W | PTE_CACHE;			\
 	pte = pmap_pte(kernel_pmap, va1);				\
 	*pte = npte;							\

From 7c8db6fd169d3591cd2b406bd9ceff30c46a090a Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 30 Apr 2010 18:04:46 +0000
Subject: [PATCH 363/532] Disable non-ASF packet flushing on Yukon Extreme as
 vendor's driver does. Without this change, Yukon Extreme seems to generate
 lots of RX FIFO overruns even though controller has available RX buffers.
 These excessive RX FIFO overruns generated lots of pause frames which in turn
 killed devices plugged into switch. It seems there is still occasional RX
 frame corruption on Yukon Extreme but this change seems to fix the pause
 frame storm.

Reported by:	jhb
Tested by:	jhb
MFC after:	5 days
---
 sys/dev/msk/if_msk.c    | 5 +++++
 sys/dev/msk/if_mskreg.h | 2 ++
 2 files changed, 7 insertions(+)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index 9b924352668..fbae30baafc 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -3917,6 +3917,11 @@ msk_init_locked(struct msk_if_softc *sc_if)
 		msk_stop(sc_if);
 		return;
 	}
+	if (sc->msk_hw_id == CHIP_ID_YUKON_EX) {
+		/* Disable flushing of non-ASF packets. */
+		CSR_WRITE_4(sc, MR_ADDR(sc_if->msk_port, RX_GMF_CTRL_T),
+		    GMF_RX_MACSEC_FLUSH_OFF);
+	}
 
 	/* Configure interrupt handling. */
 	if (sc_if->msk_port == MSK_PORT_A) {
diff --git a/sys/dev/msk/if_mskreg.h b/sys/dev/msk/if_mskreg.h
index 293ca29f542..3e8ae472b56 100644
--- a/sys/dev/msk/if_mskreg.h
+++ b/sys/dev/msk/if_mskreg.h
@@ -1941,6 +1941,8 @@
 #define RX_TRUNC_OFF		BIT_26	/* disable packet truncation */
 #define RX_VLAN_STRIP_ON	BIT_25	/* enable  VLAN stripping */
 #define RX_VLAN_STRIP_OFF	BIT_24	/* disable VLAN stripping */
+#define GMF_RX_MACSEC_FLUSH_ON	BIT_23
+#define GMF_RX_MACSEC_FLUSH_OFF	BIT_22
 #define GMF_RX_OVER_ON		BIT_19	/* enable flushing on receive overrun */
 #define GMF_RX_OVER_OFF		BIT_18	/* disable flushing on receive overrun */
 #define GMF_ASF_RX_OVER_ON	BIT_17	/* enable flushing of ASF when overrun */

From e19bd6ee554a657e08406265d3d8e0ab66849caf Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 30 Apr 2010 18:58:55 +0000
Subject: [PATCH 364/532] Add basic support for Marvell 88E8059 Yukon Optima.

Tested by:	James LaLagna < jameslalagna <> gmail dot com >
MFC after:	5 days
---
 sys/dev/msk/if_msk.c    | 20 +++++++++++++++++---
 sys/dev/msk/if_mskreg.h |  6 ++++++
 2 files changed, 23 insertions(+), 3 deletions(-)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index fbae30baafc..e15ccfe738d 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -223,6 +223,8 @@ static struct msk_product {
 	    "Marvell Yukon 88E8072 Gigabit Ethernet" },
 	{ VENDORID_MARVELL, DEVICEID_MRVL_4380,
 	    "Marvell Yukon 88E8057 Gigabit Ethernet" },
+	{ VENDORID_MARVELL, DEVICEID_MRVL_4381,
+	    "Marvell Yukon 88E8059 Gigabit Ethernet" },
 	{ VENDORID_DLINK, DEVICEID_DLINK_DGE550SX,
 	    "D-Link 550SX Gigabit Ethernet" },
 	{ VENDORID_DLINK, DEVICEID_DLINK_DGE560SX,
@@ -239,7 +241,9 @@ static const char *model_name[] = {
         "Yukon FE",
         "Yukon FE+",
         "Yukon Supreme",
-        "Yukon Ultra 2"
+        "Yukon Ultra 2",
+        "Yukon Unknown",
+        "Yukon Optima",
 };
 
 static int mskc_probe(device_t);
@@ -1229,6 +1233,7 @@ msk_phy_power(struct msk_softc *sc, int mode)
 		case CHIP_ID_YUKON_EX:
 		case CHIP_ID_YUKON_FE_P:
 		case CHIP_ID_YUKON_UL_2:
+		case CHIP_ID_YUKON_OPT:
 			CSR_WRITE_2(sc, B0_CTST, Y2_HW_WOL_OFF);
 
 			/* Enable all clocks. */
@@ -1372,6 +1377,10 @@ mskc_reset(struct msk_softc *sc)
 			    GMC_BYP_MACSECRX_ON | GMC_BYP_MACSECTX_ON |
 			    GMC_BYP_RETR_ON);
 	}
+	if (sc->msk_hw_id == CHIP_ID_YUKON_OPT && sc->msk_hw_rev == 0) {
+		/* Disable PCIe PHY powerdown(reg 0x80, bit7). */
+		CSR_WRITE_4(sc, Y2_PEX_PHY_DATA, (0x0080 << 16) | 0x0080);
+	}
 	CSR_WRITE_1(sc, B2_TST_CTRL1, TST_CFG_WRITE_OFF);
 
 	/* LED On. */
@@ -1705,8 +1714,9 @@ mskc_attach(device_t dev)
 	sc->msk_hw_rev = (CSR_READ_1(sc, B2_MAC_CFG) >> 4) & 0x0f;
 	/* Bail out if chip is not recognized. */
 	if (sc->msk_hw_id < CHIP_ID_YUKON_XL ||
-	    sc->msk_hw_id > CHIP_ID_YUKON_UL_2 ||
-	    sc->msk_hw_id == CHIP_ID_YUKON_SUPR) {
+	    sc->msk_hw_id > CHIP_ID_YUKON_OPT ||
+	    sc->msk_hw_id == CHIP_ID_YUKON_SUPR ||
+	    sc->msk_hw_id == CHIP_ID_YUKON_UNKNOWN) {
 		device_printf(dev, "unknown device: id=0x%02x, rev=0x%02x\n",
 		    sc->msk_hw_id, sc->msk_hw_rev);
 		mtx_destroy(&sc->msk_mtx);
@@ -1819,6 +1829,10 @@ mskc_attach(device_t dev)
 		sc->msk_clock = 125;	/* 125 MHz */
 		sc->msk_pflags |= MSK_FLAG_JUMBO;
 		break;
+	case CHIP_ID_YUKON_OPT:
+		sc->msk_clock = 125;	/* 125 MHz */
+		sc->msk_pflags |= MSK_FLAG_JUMBO | MSK_FLAG_DESCV2;
+		break;
 	default:
 		sc->msk_clock = 156;	/* 156 MHz */
 		break;
diff --git a/sys/dev/msk/if_mskreg.h b/sys/dev/msk/if_mskreg.h
index 3e8ae472b56..3b05b9ba858 100644
--- a/sys/dev/msk/if_mskreg.h
+++ b/sys/dev/msk/if_mskreg.h
@@ -145,6 +145,7 @@
 #define DEVICEID_MRVL_436B	0x436B
 #define DEVICEID_MRVL_436C	0x436C
 #define DEVICEID_MRVL_4380	0x4380
+#define DEVICEID_MRVL_4381	0x4381
 
 /*
  * D-Link gigabit ethernet device ID
@@ -828,6 +829,9 @@
 #define Y2_IS_CHK_RX2	BIT_10	/* Descriptor error Rx 2 */
 #define Y2_IS_CHK_TXS2	BIT_9	/* Descriptor error TXS 2 */
 #define Y2_IS_CHK_TXA2	BIT_8	/* Descriptor error TXA 2 */
+#define Y2_IS_PSM_ACK	BIT_7	/* PSM Ack (Yukon Optima) */
+#define Y2_IS_PTP_TIST	BIT_6	/* PTP TIme Stamp (Yukon Optima) */
+#define Y2_IS_PHY_QLNK	BIT_5	/* PHY Quick Link (Yukon Optima) */
 #define Y2_IS_IRQ_PHY1	BIT_4	/* Interrupt from PHY 1 */
 #define Y2_IS_IRQ_MAC1	BIT_3	/* Interrupt from MAC 1 */
 #define Y2_IS_CHK_RX1	BIT_2	/* Descriptor error Rx 1 */
@@ -894,6 +898,8 @@
 #define CHIP_ID_YUKON_FE_P	0xb8 /* Chip ID for YUKON-2 FE+ */
 #define CHIP_ID_YUKON_SUPR	0xb9 /* Chip ID for YUKON-2 Supreme */
 #define CHIP_ID_YUKON_UL_2	0xba /* Chip ID for YUKON-2 Ultra 2 */
+#define CHIP_ID_YUKON_UNKNOWN	0xbb
+#define CHIP_ID_YUKON_OPT	0xbc /* Chip ID for YUKON-2 Optima */
 
 #define	CHIP_REV_YU_XL_A0	0 /* Chip Rev. for Yukon-2 A0 */
 #define	CHIP_REV_YU_XL_A1	1 /* Chip Rev. for Yukon-2 A1 */

From b9ddb8695d047dea7e17088699bee7db467d745e Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 30 Apr 2010 19:01:55 +0000
Subject: [PATCH 365/532] Add Marvell PHYG65G Gigabit PHY which is found on
 88E8059 Yukon Optima.

Tested by:	James LaLagna < jameslalagna <> gmail dot com >
MFC after:	5 days
---
 sys/dev/mii/e1000phy.c | 2 ++
 sys/dev/mii/miidevs    | 1 +
 2 files changed, 3 insertions(+)

diff --git a/sys/dev/mii/e1000phy.c b/sys/dev/mii/e1000phy.c
index b50eb07a651..f30f91ce7a0 100644
--- a/sys/dev/mii/e1000phy.c
+++ b/sys/dev/mii/e1000phy.c
@@ -112,6 +112,7 @@ static const struct mii_phydesc e1000phys[] = {
 	MII_PHY_DESC(MARVELL, E1116R),
 	MII_PHY_DESC(MARVELL, E1118),
 	MII_PHY_DESC(MARVELL, E3016),
+	MII_PHY_DESC(MARVELL, PHYG65G),
 	MII_PHY_DESC(xxMARVELL, E1000),
 	MII_PHY_DESC(xxMARVELL, E1011),
 	MII_PHY_DESC(xxMARVELL, E1000_3),
@@ -230,6 +231,7 @@ e1000phy_reset(struct mii_softc *sc)
 		case MII_MODEL_MARVELL_E1116:
 		case MII_MODEL_MARVELL_E1118:
 		case MII_MODEL_MARVELL_E1149:
+		case MII_MODEL_MARVELL_PHYG65G:
 			/* Disable energy detect mode. */
 			reg &= ~E1000_SCR_EN_DETECT_MASK;
 			reg |= E1000_SCR_AUTO_X_MODE;
diff --git a/sys/dev/mii/miidevs b/sys/dev/mii/miidevs
index 30cd09c4071..bbf668b8c45 100644
--- a/sys/dev/mii/miidevs
+++ b/sys/dev/mii/miidevs
@@ -247,6 +247,7 @@ model MARVELL E1116		0x0021 Marvell 88E1116 Gigabit PHY
 model MARVELL E1116R		0x0024 Marvell 88E1116R Gigabit PHY
 model MARVELL E1118		0x0022 Marvell 88E1118 Gigabit PHY
 model MARVELL E3016		0x0026 Marvell 88E3016 10/100 Fast Ethernet PHY
+model MARVELL PHYG65G		0x0027 Marvell PHYG65G Gigabit PHY
 model xxMARVELL E1000		0x0005 Marvell 88E1000 Gigabit PHY
 model xxMARVELL E1011		0x0002 Marvell 88E1011 Gigabit PHY
 model xxMARVELL E1000_3		0x0003 Marvell 88E1000 Gigabit PHY

From a26a72d90b17bd8ee48505cfa61c8d2ec4b3b02a Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 30 Apr 2010 19:05:01 +0000
Subject: [PATCH 366/532] Marvell 88E8059(Yukon Optima) is now supported.

---
 share/man/man4/msk.4 | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/share/man/man4/msk.4 b/share/man/man4/msk.4
index 7a842f5ecfd..7ea12ddc818 100644
--- a/share/man/man4/msk.4
+++ b/share/man/man4/msk.4
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd March 1, 2010
+.Dd April 30, 2010
 .Dt MSK 4
 .Os
 .Sh NAME
@@ -208,6 +208,8 @@ Marvell Yukon 88E8057 Gigabit Ethernet
 .It
 Marvell Yukon 88E8058 Gigabit Ethernet
 .It
+Marvell Yukon 88E8059 Gigabit Ethernet
+.It
 Marvell Yukon 88E8070 Gigabit Ethernet
 .It
 Marvell Yukon 88E8071 Gigabit Ethernet

From e8f263195d1ad707879d69cf6e8054239d240f4c Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 19:40:37 +0000
Subject: [PATCH 367/532] - don't check hold_count without the page lock held -
 don't leak the page lock if m->object is NULL   (assuming that that check
 will in fact even be valid when m->object is protected by the page lock)

---
 sys/vm/vm_pageout.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 95517dec5c3..73c51b3f583 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -775,16 +775,17 @@ rescan0:
 		if (m->flags & PG_MARKER)
 			continue;
 
-		/*
-		 * A held page may be undergoing I/O, so skip it.
-		 */
-		if (m->hold_count) {
-			vm_page_requeue(m);
+		if (!vm_page_trylock(m)) {
 			addl_page_shortage++;
 			continue;
 		}
 
-		if (!vm_page_trylock(m) || (object = m->object) == NULL) {
+		/*
+		 * A held page may be undergoing I/O, so skip it.
+		 */
+		if (m->hold_count || (object = m->object) == NULL) {
+			vm_page_unlock(m);
+			vm_page_requeue(m);
 			addl_page_shortage++;
 			continue;
 		}

From 524461f158c0b326faef715bd5767d3e4e76d999 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 30 Apr 2010 19:52:35 +0000
Subject: [PATCH 368/532] telnet: Fix infinite loop if local output generates
 SIGPIPE.

Instead of catching SIGPIPE and jumping out of the signal handler with
longjmp, ignore it and handle write errors to the local output by exiting
from there. I have changed the error message to mention the local output
instead of NetBSD's wrong "Connection closed by foreign host". Write errors
to the network were already handled by exiting immediately and this now
applies to EPIPE too.

The code assumed that SIGPIPE could only be generated by the network
connection; if it was generated by the local output, it would longjmp out of
the signal handler and write an error message which caused another SIGPIPE.

PR:		19773
Obtained from:	NetBSD
MFC after:	1 week
---
 contrib/telnet/telnet/commands.c |  3 +--
 contrib/telnet/telnet/externs.h  |  1 -
 contrib/telnet/telnet/network.c  |  2 +-
 contrib/telnet/telnet/sys_bsd.c  | 10 +---------
 contrib/telnet/telnet/telnet.c   |  1 -
 contrib/telnet/telnet/terminal.c | 16 ++++++++++++++--
 6 files changed, 17 insertions(+), 16 deletions(-)

diff --git a/contrib/telnet/telnet/commands.c b/contrib/telnet/telnet/commands.c
index a64d763b64a..c39b18744ca 100644
--- a/contrib/telnet/telnet/commands.c
+++ b/contrib/telnet/telnet/commands.c
@@ -2491,8 +2491,7 @@ tn(int argc, char *argv[])
 	env_export("USER");
     }
     (void) call(status, "status", "notmuch", 0);
-    if (setjmp(peerdied) == 0)
-	telnet(user);
+    telnet(user); 
     (void) NetClose(net);
     ExitString("Connection closed by foreign host.\n",1);
     /*NOTREACHED*/
diff --git a/contrib/telnet/telnet/externs.h b/contrib/telnet/telnet/externs.h
index cfa1aec372d..e07aebbdb82 100644
--- a/contrib/telnet/telnet/externs.h
+++ b/contrib/telnet/telnet/externs.h
@@ -233,7 +233,6 @@ extern void
     SetNetTrace(char *);	/* Function to change where debugging goes */
 
 extern jmp_buf
-    peerdied,
     toplevel;		/* For error conditions. */
 
 extern void
diff --git a/contrib/telnet/telnet/network.c b/contrib/telnet/telnet/network.c
index bb75d6309f9..b140d6050d2 100644
--- a/contrib/telnet/telnet/network.c
+++ b/contrib/telnet/telnet/network.c
@@ -158,7 +158,7 @@ netflush(void)
 	    perror(hostname);
 	    (void)NetClose(net);
 	    ring_clear_mark(&netoring);
-	    longjmp(peerdied, -1);
+	    ExitString("Connection closed by foreign host.\n", 1);
 	    /*NOTREACHED*/
 	}
 	n = 0;
diff --git a/contrib/telnet/telnet/sys_bsd.c b/contrib/telnet/telnet/sys_bsd.c
index 45ef1e2aa3c..9fba74feada 100644
--- a/contrib/telnet/telnet/sys_bsd.c
+++ b/contrib/telnet/telnet/sys_bsd.c
@@ -808,14 +808,6 @@ NetNonblockingIO(int fd, int onoff)
  * Various signal handling routines.
  */
 
-/* ARGSUSED */
-static SIG_FUNC_RET
-deadpeer(int sig __unused)
-{
-	setcommandmode();
-	longjmp(peerdied, -1);
-}
-
 /* ARGSUSED */
 SIG_FUNC_RET
 intr(int sig __unused)
@@ -884,7 +876,7 @@ sys_telnet_init(void)
 {
     (void) signal(SIGINT, intr);
     (void) signal(SIGQUIT, intr2);
-    (void) signal(SIGPIPE, deadpeer);
+    (void) signal(SIGPIPE, SIG_IGN);
 #ifdef	SIGWINCH
     (void) signal(SIGWINCH, sendwin);
 #endif
diff --git a/contrib/telnet/telnet/telnet.c b/contrib/telnet/telnet/telnet.c
index 28237bdf9d0..8c457cf613d 100644
--- a/contrib/telnet/telnet/telnet.c
+++ b/contrib/telnet/telnet/telnet.c
@@ -146,7 +146,6 @@ unsigned char telopt_environ = TELOPT_NEW_ENVIRON;
 #endif
 
 jmp_buf	toplevel;
-jmp_buf	peerdied;
 
 int	flushline;
 int	linemode;
diff --git a/contrib/telnet/telnet/terminal.c b/contrib/telnet/telnet/terminal.c
index 3dda1c5bf88..596c2b85559 100644
--- a/contrib/telnet/telnet/terminal.c
+++ b/contrib/telnet/telnet/terminal.c
@@ -111,7 +111,8 @@ init_terminal(void)
 }
 
 /*
- *		Send as much data as possible to the terminal.
+ *		Send as much data as possible to the terminal, else exits if
+ *		it encounters a permanent failure when writing to the tty.
  *
  *		Return value:
  *			-1: No useful work done, data waiting to go out.
@@ -152,8 +153,19 @@ ttyflush(int drop)
 	}
 	ring_consumed(&ttyoring, n);
     }
-    if (n < 0)
+    if (n < 0) {
+	if (errno == EAGAIN || errno == EINTR) {
+	    return -1;
+	} else {
+	    ring_consumed(&ttyoring, ring_full_count(&ttyoring));
+	    setconnmode(0);
+	    setcommandmode();
+	    NetClose(net);
+	    fprintf(stderr, "Write error on local output.\n");
+	    exit(1);
+	}
 	return -1;
+    }
     if (n == n0) {
 	if (n0)
 	    return -1;

From 6dd8b893d63e64f58a18a812d8749afed45607a9 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 21:20:14 +0000
Subject: [PATCH 369/532] - acquire the page lock in vm_contig_launder_page
 before checking page fields - release page queue lock before calling
 vm_pageout_flush

---
 sys/vm/vm_contig.c | 21 +++++++++++++++------
 1 file changed, 15 insertions(+), 6 deletions(-)

diff --git a/sys/vm/vm_contig.c b/sys/vm/vm_contig.c
index 7220055cd25..2c8b8b11c25 100644
--- a/sys/vm/vm_contig.c
+++ b/sys/vm/vm_contig.c
@@ -96,30 +96,33 @@ vm_contig_launder_page(vm_page_t m, vm_page_t *next)
 	vm_page_t m_tmp;
 	struct vnode *vp;
 	struct mount *mp;
-	int vfslocked;
+	int vfslocked, dirty;
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock(m);
+	vm_page_lock_queues();
 	object = m->object;
 	if (!VM_OBJECT_TRYLOCK(object) &&
 	    !vm_pageout_fallback_object_lock(m, next)) {
 		VM_OBJECT_UNLOCK(object);
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		return (EAGAIN);
 	}
 	if (vm_page_sleep_if_busy(m, TRUE, "vpctw0")) {
 		VM_OBJECT_UNLOCK(object);
-		vm_page_lock_queues();
 		return (EBUSY);
 	}
 	vm_page_test_dirty(m);
 	if (m->dirty == 0 && m->hold_count == 0)
 		pmap_remove_all(m);
-	if (m->dirty) {
+	if ((dirty = m->dirty) != 0) {
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		if ((object->flags & OBJ_DEAD) != 0) {
 			VM_OBJECT_UNLOCK(object);
 			return (EAGAIN);
 		}
 		if (object->type == OBJT_VNODE) {
-			vm_page_unlock_queues();
 			vp = object->handle;
 			vm_object_reference_locked(object);
 			VM_OBJECT_UNLOCK(object);
@@ -133,7 +136,6 @@ vm_contig_launder_page(vm_page_t m, vm_page_t *next)
 			VFS_UNLOCK_GIANT(vfslocked);
 			vm_object_deallocate(object);
 			vn_finished_write(mp);
-			vm_page_lock_queues();
 			return (0);
 		} else if (object->type == OBJT_SWAP ||
 			   object->type == OBJT_DEFAULT) {
@@ -144,6 +146,11 @@ vm_contig_launder_page(vm_page_t m, vm_page_t *next)
 		}
 	} else if (m->hold_count == 0)
 		vm_page_cache(m);
+
+	if (dirty == 0) {
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
+	}
 	VM_OBJECT_UNLOCK(object);
 	return (0);
 }
@@ -162,7 +169,9 @@ vm_contig_launder(int queue)
 
 		KASSERT(VM_PAGE_INQUEUE2(m, queue),
 		    ("vm_contig_launder: page %p's queue is not %d", m, queue));
+		vm_page_unlock_queues();
 		error = vm_contig_launder_page(m, &next);
+		vm_page_lock_queues();
 		if (error == 0)
 			return (TRUE);
 		if (error == EBUSY)

From ad0c05daf989df91c6d519efcde766c5c6bce734 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 21:21:21 +0000
Subject: [PATCH 370/532] don't call vm_pageout_flush with the page queue mutex
 held

Reported by: Michael Butler
---
 sys/vm/vm_object.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 47ef9737151..7e1cc2216ca 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1058,7 +1058,9 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 	}
 	runlen = maxb + maxf + 1;
 
+	vm_page_unlock_queues();
 	vm_pageout_flush(ma, runlen, pagerflags);
+	vm_page_lock_queues();
 	for (i = 0; i < runlen; i++) {
 		if (ma[i]->dirty) {
 			vm_page_unlock_queues();

From 7bec141b121da05bad35c061c1e9ab17f47c71d7 Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Fri, 30 Apr 2010 22:31:37 +0000
Subject: [PATCH 371/532] push up dropping of the page queue lock to avoid
 holding it in vm_pageout_flush

---
 sys/vm/vm_object.c  | 46 +++++++++++++++++----------------------------
 sys/vm/vm_pageout.c |  2 ++
 2 files changed, 19 insertions(+), 29 deletions(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 7e1cc2216ca..604aa3c9138 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -774,6 +774,7 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 	int pagerflags;
 	int curgeneration;
 
+	mtx_assert(&vm_page_queue_mtx, MA_NOTOWNED);
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	if ((object->flags & OBJ_MIGHTBEDIRTY) == 0)
 		return;
@@ -830,13 +831,13 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 				++tscan;
 				continue;
 			}
+			vm_page_unlock_queues();
+			vm_page_unlock(p);
 			/*
 			 * If we have been asked to skip nosync pages and 
 			 * this is a nosync page, we can't continue.
 			 */
 			if ((flags & OBJPC_NOSYNC) && (p->oflags & VPO_NOSYNC)) {
-				vm_page_unlock_queues();
-				vm_page_unlock(p);
 				if (--scanlimit == 0)
 					break;
 				++tscan;
@@ -849,8 +850,7 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 			 * page (i.e. had to sleep).
 			 */
 			tscan += vm_object_page_collect_flush(object, p, curgeneration, pagerflags);
-			vm_page_unlock_queues();
-			vm_page_unlock(p);
+
 		}
 
 		/*
@@ -860,7 +860,6 @@ vm_object_page_clean(vm_object_t object, vm_pindex_t start, vm_pindex_t end, int
 		 * return immediately.
 		 */
 		if (tscan >= tend && (tstart || tend < object->size)) {
-			vm_page_unlock_queues();
 			vm_object_clear_flag(object, OBJ_CLEANING);
 			return;
 		}
@@ -918,43 +917,33 @@ again:
 			p->oflags &= ~VPO_CLEANCHK;
 			continue;
 		}
-
+		vm_page_unlock_queues();
+		vm_page_unlock(p);
 		/*
 		 * If we have been asked to skip nosync pages and this is a
 		 * nosync page, skip it.  Note that the object flags were
 		 * not cleared in this case so we do not have to set them.
 		 */
 		if ((flags & OBJPC_NOSYNC) && (p->oflags & VPO_NOSYNC)) {
-			vm_page_unlock_queues();
-			vm_page_unlock(p);
 			p->oflags &= ~VPO_CLEANCHK;
 			continue;
 		}
 
 		n = vm_object_page_collect_flush(object, p,
 			curgeneration, pagerflags);
-		if (n == 0) {
-			vm_page_unlock_queues();
-			vm_page_unlock(p);
+		if (n == 0)
 			goto rescan;
-		}
 
-		if (object->generation != curgeneration) {
-			vm_page_unlock_queues();
-			vm_page_unlock(p);
+		if (object->generation != curgeneration)
 			goto rescan;
-		}
 
 		/*
 		 * Try to optimize the next page.  If we can't we pick up
 		 * our (random) scan where we left off.
 		 */
-		if (msync_flush_flags & MSYNC_FLUSH_SOFTSEQ) {
-			vm_page_unlock_queues();
-			vm_page_unlock(p);
+		if (msync_flush_flags & MSYNC_FLUSH_SOFTSEQ)
 			if ((p = vm_page_lookup(object, pi + n)) != NULL)
 				goto again;
-		}
 	}
 #if 0
 	VOP_FSYNC(vp, (pagerflags & VM_PAGER_PUT_SYNC)?MNT_WAIT:0, curproc);
@@ -977,12 +966,11 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 	vm_page_t mab[vm_pageout_page_count];
 	vm_page_t ma[vm_pageout_page_count];
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
-	vm_page_lock_assert(p, MA_OWNED);
+	mtx_assert(&vm_page_queue_mtx, MA_NOTOWNED);
+	vm_page_lock_assert(p, MA_NOTOWNED);
+	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	pi = p->pindex;
 	while (vm_page_sleep_if_busy(p, TRUE, "vpcwai")) {
-		vm_page_lock(p);
-		vm_page_lock_queues();
 		if (object->generation != curgeneration) {
 			return(0);
 		}
@@ -997,16 +985,17 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 				 (tp->oflags & VPO_CLEANCHK) == 0) ||
 				(tp->busy != 0))
 				break;
-			vm_page_unlock_queues();
 			vm_page_lock(tp);
 			vm_page_lock_queues();
 			vm_page_test_dirty(tp);
 			if (tp->dirty == 0) {
 				vm_page_unlock(tp);
+				vm_page_unlock_queues();
 				tp->oflags &= ~VPO_CLEANCHK;
 				break;
 			}
 			vm_page_unlock(tp);
+			vm_page_unlock_queues();
 			maf[ i - 1 ] = tp;
 			maxf++;
 			continue;
@@ -1026,15 +1015,16 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 					 (tp->oflags & VPO_CLEANCHK) == 0) ||
 					(tp->busy != 0))
 					break;
-				vm_page_unlock_queues();
 				vm_page_lock(tp);
 				vm_page_lock_queues();
 				vm_page_test_dirty(tp);
 				if (tp->dirty == 0) {
+					vm_page_unlock_queues();
 					vm_page_unlock(tp);
 					tp->oflags &= ~VPO_CLEANCHK;
 					break;
 				}
+				vm_page_unlock_queues();
 				vm_page_unlock(tp);
 				mab[ i - 1 ] = tp;
 				maxb++;
@@ -1058,15 +1048,13 @@ vm_object_page_collect_flush(vm_object_t object, vm_page_t p, int curgeneration,
 	}
 	runlen = maxb + maxf + 1;
 
-	vm_page_unlock_queues();
 	vm_pageout_flush(ma, runlen, pagerflags);
-	vm_page_lock_queues();
 	for (i = 0; i < runlen; i++) {
 		if (ma[i]->dirty) {
-			vm_page_unlock_queues();
 			vm_page_lock(ma[i]);
 			vm_page_lock_queues();
 			pmap_remove_write(ma[i]);
+			vm_page_unlock_queues();
 			vm_page_unlock(ma[i]);
 			ma[i]->oflags |= VPO_CLEANCHK;
 
diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 73c51b3f583..39a06f84939 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -433,6 +433,8 @@ vm_pageout_flush(vm_page_t *mc, int count, int flags)
 	int i;
 
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
+	mtx_assert(&vm_page_queue_mtx, MA_NOTOWNED);
+
 	/*
 	 * Initiate I/O.  Bump the vm_page_t->busy counter and
 	 * mark the pages read-only.

From e1e9ba3370d01ab5448943461a07d6d183a15bf1 Mon Sep 17 00:00:00 2001
From: Ed Schouten <ed@FreeBSD.org>
Date: Fri, 30 Apr 2010 22:33:49 +0000
Subject: [PATCH 372/532] Remove WNOHANG flag from wait3().

Because script(1) now reliably terminates when the TTY is closed, it may
be the case that the call to wait3() occurs just before the child
process exits. This causes error codes to be ignored.

Just change script(1) to use waitpid() instead of wait3(). This makes it
more portable and prevents the need for a loop, since waitpid() only
returns a specified process.

PR:		bin/146189
Tested by:	amdmi3@, older version
MFC after:	2 weeks
---
 usr.bin/script/script.c | 24 +++++++++---------------
 1 file changed, 9 insertions(+), 15 deletions(-)

diff --git a/usr.bin/script/script.c b/usr.bin/script/script.c
index 6c4e0ee4db3..a21785abbc4 100644
--- a/usr.bin/script/script.c
+++ b/usr.bin/script/script.c
@@ -219,23 +219,17 @@ usage(void)
 void
 finish(void)
 {
-	pid_t pid;
-	int die, e, status;
+	int e, status;
 
-	die = e = 0;
-	while ((pid = wait3(&status, WNOHANG, 0)) > 0)
-	        if (pid == child) {
-			die = 1;
-			if (WIFEXITED(status))
-				e = WEXITSTATUS(status);
-			else if (WIFSIGNALED(status))
-				e = WTERMSIG(status);
-			else /* can't happen */
-				e = 1;
-		}
-
-	if (die)
+	if (waitpid(child, &status, 0) == child) {
+		if (WIFEXITED(status))
+			e = WEXITSTATUS(status);
+		else if (WIFSIGNALED(status))
+			e = WTERMSIG(status);
+		else /* can't happen */
+			e = 1;
 		done(e);
+	}
 }
 
 void

From 2083bca542032e4eac1a32d5963a6c3396f335de Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Sat, 1 May 2010 00:36:40 +0000
Subject: [PATCH 373/532] Do not initialize mutex and return error if it cannot
 map memory.

---
 sys/compat/x86bios/x86bios.c | 26 +++++++++++++-------------
 1 file changed, 13 insertions(+), 13 deletions(-)

diff --git a/sys/compat/x86bios/x86bios.c b/sys/compat/x86bios/x86bios.c
index d5512fc43f3..defdbcadf86 100644
--- a/sys/compat/x86bios/x86bios.c
+++ b/sys/compat/x86bios/x86bios.c
@@ -563,15 +563,15 @@ x86bios_unmap_mem(void)
 	contigfree(x86bios_seg, X86BIOS_SEG_SIZE, M_DEVBUF);
 }
 
-static void
-x86bios_init(void *arg __unused)
+static int
+x86bios_init(void)
 {
 	int i;
 
-	mtx_init(&x86bios_lock, "x86bios lock", NULL, MTX_SPIN);
-
 	if (x86bios_map_mem() != 0)
-		return;
+		return (ENOMEM);
+
+	mtx_init(&x86bios_lock, "x86bios lock", NULL, MTX_SPIN);
 
 	x86bios_map = malloc(sizeof(*x86bios_map) * X86BIOS_PAGES, M_DEVBUF,
 	    M_WAITOK | M_ZERO);
@@ -600,10 +600,12 @@ x86bios_init(void *arg __unused)
 
 	for (i = 0; i < 256; i++)
 		x86bios_emu._x86emu_intrTab[i] = x86bios_emu_get_intr;
+
+	return (0);
 }
 
-static void
-x86bios_uninit(void *arg __unused)
+static int
+x86bios_uninit(void)
 {
 	vm_offset_t *map = x86bios_map;
 
@@ -618,6 +620,8 @@ x86bios_uninit(void *arg __unused)
 		x86bios_unmap_mem();
 
 	mtx_destroy(&x86bios_lock);
+
+	return (0);
 }
 
 static int
@@ -626,16 +630,12 @@ x86bios_modevent(module_t mod __unused, int type, void *data __unused)
 
 	switch (type) {
 	case MOD_LOAD:
-		x86bios_init(NULL);
-		break;
+		return (x86bios_init());
 	case MOD_UNLOAD:
-		x86bios_uninit(NULL);
-		break;
+		return (x86bios_uninit());
 	default:
 		return (ENOTSUP);
 	}
-
-	return (0);
 }
 
 static moduledata_t x86bios_mod = {

From 483191871b206c25ea7f6182e8b8a8acdda0af24 Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Sat, 1 May 2010 01:05:07 +0000
Subject: [PATCH 374/532] Reduce MD code further.  At least, it compiles on
 ia64 now (but it is not connected to build).  The idea/code was shamelessly
 taken from r207329.

---
 sys/compat/x86bios/x86bios.c | 114 ++++++++++++++++++++++++-----------
 1 file changed, 78 insertions(+), 36 deletions(-)

diff --git a/sys/compat/x86bios/x86bios.c b/sys/compat/x86bios/x86bios.c
index defdbcadf86..cb8b85e17ab 100644
--- a/sys/compat/x86bios/x86bios.c
+++ b/sys/compat/x86bios/x86bios.c
@@ -47,11 +47,15 @@ __FBSDID("$FreeBSD$");
 #include <dev/pci/pcireg.h>
 #include <dev/pci/pcivar.h>
 
-#include <machine/cpufunc.h>
+#include <machine/iodev.h>
 
 #include <vm/vm.h>
 #include <vm/pmap.h>
 
+#if defined(__amd64__) || defined(__i386__)
+#define	X86BIOS_NATIVE_ARCH
+#endif
+
 #define	X86BIOS_PAGE_SIZE	0x00001000	/* 4K */
 
 #define	X86BIOS_IVT_SIZE	0x00000500	/* 1K + 256 (BDA) */
@@ -236,27 +240,49 @@ x86bios_emu_inb(struct x86emu *emu, uint16_t port)
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return (0);
 
-	return (inb(port));
+	return (iodev_read_1(port));
 }
 
 static uint16_t
 x86bios_emu_inw(struct x86emu *emu, uint16_t port)
 {
+	uint16_t val;
 
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return (0);
 
-	return (inw(port));
+#ifndef X86BIOS_NATIVE_ARCH
+	if ((port & 1) != 0) {
+		val = iodev_read_1(port);
+		val |= iodev_read_1(port + 1) << 8;
+	} else
+#endif
+	val = iodev_read_2(port);
+
+	return (val);
 }
 
 static uint32_t
 x86bios_emu_inl(struct x86emu *emu, uint16_t port)
 {
+	uint32_t val;
 
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return (0);
 
-	return (inl(port));
+#ifndef X86BIOS_NATIVE_ARCH
+	if ((port & 1) != 0) {
+		val = iodev_read_1(port);
+		val |= iodev_read_2(port + 1) << 8;
+		val |= iodev_read_1(port + 3) << 24;
+	} else if ((port & 2) != 0) {
+		val = iodev_read_2(port);
+		val |= iodev_read_2(port + 2) << 16;
+	} else
+#endif
+	val = iodev_read_4(port);
+
+	return (val);
 }
 
 static void
@@ -268,7 +294,7 @@ x86bios_emu_outb(struct x86emu *emu, uint16_t port, uint8_t val)
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return;
 
-	outb(port, val);
+	iodev_write_1(port, val);
 }
 
 static void
@@ -278,7 +304,13 @@ x86bios_emu_outw(struct x86emu *emu, uint16_t port, uint16_t val)
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return;
 
-	outw(port, val);
+#ifndef X86BIOS_NATIVE_ARCH
+	if ((port & 1) != 0) {
+		iodev_write_1(port, val);
+		iodev_write_1(port + 1, val >> 8);
+	} else
+#endif
+	iodev_write_2(port, val);
 }
 
 static void
@@ -288,7 +320,17 @@ x86bios_emu_outl(struct x86emu *emu, uint16_t port, uint32_t val)
 	if (port >= 0x80 && port < 0x88) /* POST status register */
 		return;
 
-	outl(port, val);
+#ifndef X86BIOS_NATIVE_ARCH
+	if ((port & 1) != 0) {
+		iodev_write_1(port, val);
+		iodev_write_2(port + 1, val >> 8);
+		iodev_write_1(port + 3, val >> 24);
+	} else if ((port & 2) != 0) {
+		iodev_write_2(port, val);
+		iodev_write_2(port + 2, val >> 16);
+	} else
+#endif
+	iodev_write_4(port, val);
 }
 
 static void
@@ -484,45 +526,53 @@ x86bios_match_device(uint32_t offset, device_t dev)
 	return (1);
 }
 
-#if defined(__amd64__) || (defined(__i386__) && !defined(PC98))
-#define	PROBE_EBDA	1
+static __inline void
+x86bios_unmap_mem(void)
+{
+
+	if (x86bios_ivt != NULL)
+#ifdef X86BIOS_NATIVE_ARCH
+		pmap_unmapdev((vm_offset_t)x86bios_ivt, X86BIOS_IVT_SIZE);
 #else
-#define	PROBE_EBDA	0
+		free(x86bios_ivt, M_DEVBUF);
 #endif
+	if (x86bios_rom != NULL)
+		pmap_unmapdev((vm_offset_t)x86bios_rom, X86BIOS_ROM_SIZE);
+	if (x86bios_seg != NULL)
+		contigfree(x86bios_seg, X86BIOS_SEG_SIZE, M_DEVBUF);
+}
 
 static __inline int
 x86bios_map_mem(void)
 {
 
+#ifdef X86BIOS_NATIVE_ARCH
 	x86bios_ivt = pmap_mapbios(X86BIOS_IVT_BASE, X86BIOS_IVT_SIZE);
-	if (x86bios_ivt == NULL)
-		return (1);
 
-#if PROBE_EBDA
+#ifndef PC98
 	/* Probe EBDA via BDA. */
-	x86bios_rom_phys = *(uint16_t *)((vm_offset_t)x86bios_ivt + 0x40e);
-	x86bios_rom_phys = le16toh(x86bios_rom_phys) << 4;
+	x86bios_rom_phys = *(uint16_t *)((caddr_t)x86bios_ivt + 0x40e);
+	x86bios_rom_phys = x86bios_rom_phys << 4;
 	if (x86bios_rom_phys != 0 && x86bios_rom_phys < X86BIOS_ROM_BASE &&
 	    X86BIOS_ROM_BASE - x86bios_rom_phys <= 128 * 1024)
 		x86bios_rom_phys =
 		    rounddown(x86bios_rom_phys, X86BIOS_PAGE_SIZE);
 	else
 #endif
+#else
+	x86bios_ivt = malloc(X86BIOS_IVT_SIZE, M_DEVBUF, M_ZERO | M_WAITOK);
+#endif
+
 	x86bios_rom_phys = X86BIOS_ROM_BASE;
 	x86bios_rom = pmap_mapdev(x86bios_rom_phys, X86BIOS_ROM_SIZE);
-	if (x86bios_rom == NULL) {
-		pmap_unmapdev((vm_offset_t)x86bios_ivt, X86BIOS_IVT_SIZE);
-		return (1);
-	}
-#if PROBE_EBDA
+	if (x86bios_rom == NULL)
+		goto fail;
+#if defined(X86BIOS_NATIVE_ARCH) && !defined(PC98)
 	/* Change attribute for EBDA. */
 	if (x86bios_rom_phys < X86BIOS_ROM_BASE &&
 	    pmap_change_attr((vm_offset_t)x86bios_rom,
-	    X86BIOS_ROM_BASE - x86bios_rom_phys, PAT_WRITE_BACK) != 0) {
-		pmap_unmapdev((vm_offset_t)x86bios_ivt, X86BIOS_IVT_SIZE);
-		pmap_unmapdev((vm_offset_t)x86bios_rom, X86BIOS_ROM_SIZE);
-		return (1);
-	}
+	    X86BIOS_ROM_BASE - x86bios_rom_phys, PAT_WRITE_BACK) != 0)
+		goto fail;
 #endif
 
 	x86bios_seg = contigmalloc(X86BIOS_SEG_SIZE, M_DEVBUF, M_WAITOK,
@@ -537,12 +587,10 @@ x86bios_map_mem(void)
 		    (uint32_t)x86bios_seg_phys,
 		    X86BIOS_SEG_SIZE + (uint32_t)x86bios_seg_phys - 1,
 		    x86bios_seg);
-#if PROBE_EBDA
 		if (x86bios_rom_phys < X86BIOS_ROM_BASE)
 			printf("x86bios:  EBDA 0x%06x-0x%06x at %p\n",
 			    (uint32_t)x86bios_rom_phys, X86BIOS_ROM_BASE - 1,
 			    x86bios_rom);
-#endif
 		printf("x86bios:   ROM 0x%06x-0x%06x at %p\n",
 		    X86BIOS_ROM_BASE, X86BIOS_MEM_SIZE - X86BIOS_SEG_SIZE - 1,
 		    (void *)((vm_offset_t)x86bios_rom + X86BIOS_ROM_BASE -
@@ -550,17 +598,11 @@ x86bios_map_mem(void)
 	}
 
 	return (0);
-}
 
-#undef PROBE_EBDA
+fail:
+	x86bios_unmap_mem();
 
-static __inline void
-x86bios_unmap_mem(void)
-{
-
-	pmap_unmapdev((vm_offset_t)x86bios_ivt, X86BIOS_IVT_SIZE);
-	pmap_unmapdev((vm_offset_t)x86bios_rom, X86BIOS_ROM_SIZE);
-	contigfree(x86bios_seg, X86BIOS_SEG_SIZE, M_DEVBUF);
+	return (1);
 }
 
 static int

From 0ce3ba8cd58ea437906bcb4cf8c03a4e92a9e8ca Mon Sep 17 00:00:00 2001
From: Kip Macy <kmacy@FreeBSD.org>
Date: Sat, 1 May 2010 03:41:21 +0000
Subject: [PATCH 375/532] Update locking comment above vm_page:        -
 re-assign page queue lock "Q"        - assign page lock "P"        - update
 several uncommented fields        - observe that hold_count is now protected
 by the page lock "P"

---
 sys/vm/vm_page.h | 19 ++++++++++---------
 1 file changed, 10 insertions(+), 9 deletions(-)

diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index 35a81f8a8d7..51a896f9ae3 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -90,20 +90,21 @@
  *	and sundry status bits.
  *
  *	Fields in this structure are locked either by the lock on the
- *	object that the page belongs to (O) or by the lock on the page
- *	queues (P).
+ *	object that the page belongs to (O), its corresponding page lock (P),
+ *	or by the lock on the page queues (Q).
+ *	
  */
 
 TAILQ_HEAD(pglist, vm_page);
 
 struct vm_page {
-	TAILQ_ENTRY(vm_page) pageq;	/* queue info for FIFO queue or free list (P) */
+	TAILQ_ENTRY(vm_page) pageq;	/* queue info for FIFO queue or free list (Q) */
 	TAILQ_ENTRY(vm_page) listq;	/* pages in same object (O) 	*/
 	struct vm_page *left;		/* splay tree link (O)		*/
 	struct vm_page *right;		/* splay tree link (O)		*/
 
-	vm_object_t object;		/* which object am I in (O,P)*/
-	vm_pindex_t pindex;		/* offset into object (O,P) */
+	vm_object_t object;		/* which object am I in (O,Q)*/
+	vm_pindex_t pindex;		/* offset into object (O,Q) */
 	vm_paddr_t phys_addr;		/* physical address of page */
 	struct md_page md;		/* machine dependant stuff */
 	uint8_t	queue;			/* page queue index */
@@ -111,11 +112,11 @@ struct vm_page {
 	u_short	flags;			/* see below */
 	uint8_t	order;			/* index of the buddy queue */
 	uint8_t pool;
-	u_short cow;			/* page cow mapping count */
-	u_int wire_count;		/* wired down maps refs (P) */
-	short hold_count;		/* page hold count */
+	u_short cow;			/* page cow mapping count (Q) */
+	u_int wire_count;		/* wired down maps refs (Q) */
+	short hold_count;		/* page hold count (P) */
 	u_short oflags;			/* page flags (O) */
-	u_char	act_count;		/* page usage count */
+	u_char	act_count;		/* page usage count (Q) */
 	u_char	busy;			/* page busy count (O) */
 	/* NOTE that these must support one bit per DEV_BSIZE in a page!!! */
 	/* so, on normal X86 kernels, they must be at least 8 bits wide */

From 02b9c3f52f62a581757a389abfdd079bd8c67c15 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Sat, 1 May 2010 07:25:28 +0000
Subject: [PATCH 376/532] sparc64, and possibly other architectures, pads the
 length of the section holding the config file to sh_addralign bytes using
 NULs. This bogusly triggers an assert.  Break out of the loop when we hit an
 NUL within that many bytes of the end.

MFC after:	3 days
---
 usr.sbin/config/main.c | 18 +++++++++++-------
 1 file changed, 11 insertions(+), 7 deletions(-)

diff --git a/usr.sbin/config/main.c b/usr.sbin/config/main.c
index 5cffed2c641..e63e6ab8f56 100644
--- a/usr.sbin/config/main.c
+++ b/usr.sbin/config/main.c
@@ -670,7 +670,7 @@ kernconfdump(const char *file)
 	struct stat st;
 	FILE *fp, *pp;
 	int error, len, osz, r;
-	unsigned int i, off, size;
+	unsigned int i, off, size, t1, t2, align;
 	char *cmd, *o;
 
 	r = open(file, O_RDONLY);
@@ -689,8 +689,8 @@ kernconfdump(const char *file)
 	if (o == NULL)
 		err(EXIT_FAILURE, "Couldn't allocate memory");
 	/* ELF note section header. */
-	asprintf(&cmd, "/usr/bin/elfdump -c %s | grep -A 5 kern_conf"
-	    "| tail -2 | cut -d ' ' -f 2 | paste - - -", file);
+	asprintf(&cmd, "/usr/bin/elfdump -c %s | grep -A 8 kern_conf"
+	    "| tail -5 | cut -d ' ' -f 2 | paste - - - - -", file);
 	if (cmd == NULL)
 		errx(EXIT_FAILURE, "asprintf() failed");
 	pp = popen(cmd, "r");
@@ -699,24 +699,28 @@ kernconfdump(const char *file)
 	free(cmd);
 	len = fread(o, osz, 1, pp);
 	pclose(pp);
-	r = sscanf(o, "%d\t%d", &off, &size);
+	r = sscanf(o, "%d%d%d%d%d", &off, &size, &t1, &t2, &align);
 	free(o);
-	if (r != 2)
+	if (r != 5)
 		errx(EXIT_FAILURE, "File %s doesn't contain configuration "
 		    "file. Either unsupported, or not compiled with "
 		    "INCLUDE_CONFIG_FILE", file);
 	r = fseek(fp, off, SEEK_CUR);
 	if (r != 0)
 		err(EXIT_FAILURE, "fseek() failed");
-	for (i = 0; i < size - 1; i++) {
+	for (i = 0; i < size; i++) {
 		r = fgetc(fp);
 		if (r == EOF)
 			break;
 		/* 
 		 * If '\0' is present in the middle of the configuration
 		 * string, this means something very weird is happening.
-		 * Make such case very visible.
+		 * Make such case very visible.  However, some architectures
+		 * pad the length of the section with NULs to a multiple of
+		 * sh_addralign, allow a NUL in that part of the section.
 		 */
+		if (r == '\0' && (size - i) < align)
+			break;
 		assert(r != '\0' && ("Char present in the configuration "
 		    "string mustn't be equal to 0"));
 		fputc(r, stdout);

From 727c1288976ba54b63eb2789711b1f94798180af Mon Sep 17 00:00:00 2001
From: Edwin Groothuis <edwin@FreeBSD.org>
Date: Sat, 1 May 2010 09:05:06 +0000
Subject: [PATCH 377/532] Improve usage of tunefs:

Document -j switch in usage() to reflect recent SUJ work.

Submitted by:   Alastair Hogge
MFC after:      1 week
---
 sbin/tunefs/tunefs.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/sbin/tunefs/tunefs.c b/sbin/tunefs/tunefs.c
index 7987620bc59..dc2492fc229 100644
--- a/sbin/tunefs/tunefs.c
+++ b/sbin/tunefs/tunefs.c
@@ -1007,10 +1007,11 @@ out:
 void
 usage(void)
 {
-	fprintf(stderr, "%s\n%s\n%s\n%s\n",
+	fprintf(stderr, "%s\n%s\n%s\n%s\n%s\n",
 "usage: tunefs [-A] [-a enable | disable] [-e maxbpg] [-f avgfilesize]",
-"              [-J enable | disable ] [-L volname] [-l enable | disable]",
-"              [-m minfree] [-N enable | disable] [-n enable | disable]",
+"              [-J enable | disable] [-j enable | disable]", 
+"              [-L volname] [-l enable | disable] [-m minfree]",
+"              [-N enable | disable] [-n enable | disable]",
 "              [-o space | time] [-p] [-s avgfpdir] special | filesystem");
 	exit(2);
 }

From 6a5baa54fd90486341e67f9d532447e4c5c090b1 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sat, 1 May 2010 13:15:35 +0000
Subject: [PATCH 378/532] Remove debugging code that was not used once since
 commit.

Suggested by:	bde
MFC after:	1 week
---
 sys/amd64/amd64/trap.c | 86 +-----------------------------------------
 1 file changed, 1 insertion(+), 85 deletions(-)

diff --git a/sys/amd64/amd64/trap.c b/sys/amd64/amd64/trap.c
index 8acde03560c..f3dba9459ed 100644
--- a/sys/amd64/amd64/trap.c
+++ b/sys/amd64/amd64/trap.c
@@ -172,52 +172,6 @@ SYSCTL_INT(_machdep, OID_AUTO, prot_fault_translation, CTLFLAG_RW,
 
 extern char *syscallnames[];
 
-/* #define DEBUG 1 */
-#ifdef DEBUG
-static void
-report_seg_fault(const char *segn, struct trapframe *frame)
-{
-	struct proc_ldt *pldt;
-	struct trapframe *pf;
-
-	pldt = curproc->p_md.md_ldt;
-	printf("%d: %s load fault %lx %p %d\n",
-	    curproc->p_pid, segn, frame->tf_err,
-	    pldt != NULL ? pldt->ldt_base : NULL,
-	    pldt != NULL ? pldt->ldt_refcnt : 0);
-	kdb_backtrace();
-	pf = (struct trapframe *)frame->tf_rsp;
-	printf("rdi %lx\n", pf->tf_rdi);
-	printf("rsi %lx\n", pf->tf_rsi);
-	printf("rdx %lx\n", pf->tf_rdx);
-	printf("rcx %lx\n", pf->tf_rcx);
-	printf("r8  %lx\n", pf->tf_r8);
-	printf("r9  %lx\n", pf->tf_r9);
-	printf("rax %lx\n", pf->tf_rax);
-	printf("rbx %lx\n", pf->tf_rbx);
-	printf("rbp %lx\n", pf->tf_rbp);
-	printf("r10 %lx\n", pf->tf_r10);
-	printf("r11 %lx\n", pf->tf_r11);
-	printf("r12 %lx\n", pf->tf_r12);
-	printf("r13 %lx\n", pf->tf_r13);
-	printf("r14 %lx\n", pf->tf_r14);
-	printf("r15 %lx\n", pf->tf_r15);
-	printf("fs  %x\n", pf->tf_fs);
-	printf("gs  %x\n", pf->tf_gs);
-	printf("es  %x\n", pf->tf_es);
-	printf("ds  %x\n", pf->tf_ds);
-	printf("tno %x\n", pf->tf_trapno);
-	printf("adr %lx\n", pf->tf_addr);
-	printf("flg %x\n", pf->tf_flags);
-	printf("err %lx\n", pf->tf_err);
-	printf("rip %lx\n", pf->tf_rip);
-	printf("cs  %lx\n", pf->tf_cs);
-	printf("rfl %lx\n", pf->tf_rflags);
-	printf("rsp %lx\n", pf->tf_rsp);
-	printf("ss  %lx\n", pf->tf_ss);
-}
-#endif
-
 /*
  * Exception, fault, and trap interface to the FreeBSD kernel.
  * This common code is called from assembly language IDT gate entry
@@ -314,9 +268,7 @@ trap(struct trapframe *frame)
 			 */
 			printf("kernel trap %d with interrupts disabled\n",
 			    type);
-#ifdef DEBUG
-			report_seg_fault("hlt", frame);
-#endif
+
 			/*
 			 * We shouldn't enable interrupts while holding a
 			 * spin lock or servicing an NMI.
@@ -535,33 +487,21 @@ trap(struct trapframe *frame)
 				goto out;
 			}
 			if (frame->tf_rip == (long)ld_ds) {
-#ifdef DEBUG
-				report_seg_fault("ds", frame);
-#endif
 				frame->tf_rip = (long)ds_load_fault;
 				frame->tf_ds = _udatasel;
 				goto out;
 			}
 			if (frame->tf_rip == (long)ld_es) {
-#ifdef DEBUG
-				report_seg_fault("es", frame);
-#endif
 				frame->tf_rip = (long)es_load_fault;
 				frame->tf_es = _udatasel;
 				goto out;
 			}
 			if (frame->tf_rip == (long)ld_fs) {
-#ifdef DEBUG
-				report_seg_fault("fs", frame);
-#endif
 				frame->tf_rip = (long)fs_load_fault;
 				frame->tf_fs = _ufssel;
 				goto out;
 			}
 			if (frame->tf_rip == (long)ld_gs) {
-#ifdef DEBUG
-				report_seg_fault("gs", frame);
-#endif
 				frame->tf_rip = (long)gs_load_fault;
 				frame->tf_gs = _ugssel;
 				goto out;
@@ -667,30 +607,6 @@ trap(struct trapframe *frame)
 	ksi.ksi_addr = (void *)addr;
 	trapsignal(td, &ksi);
 
-#ifdef DEBUG
-{
-	register_t rg,rgk, rf;
-
-	if (type <= MAX_TRAP_MSG) {
-		uprintf("fatal process exception: %s",
-			trap_msg[type]);
-		if ((type == T_PAGEFLT) || (type == T_PROTFLT))
-			uprintf(", fault VA = 0x%lx", frame->tf_addr);
-		uprintf("\n");
-	}
-	rf = rdmsr(0xc0000100);
-	rg = rdmsr(0xc0000101);
-	rgk = rdmsr(0xc0000102);
-	uprintf("pid %d TRAP %d rip %lx err %lx addr %lx cs %lx ss %lx ds %x "
-		"es %x fs %x fsbase %lx %lx gs %x gsbase %lx %lx %lx\n",
-		curproc->p_pid, type, frame->tf_rip, frame->tf_err,
-		frame->tf_addr,
-		frame->tf_cs, frame->tf_ss, frame->tf_ds, frame->tf_es,
-		frame->tf_fs, td->td_pcb->pcb_fsbase, rf,
-		frame->tf_gs, td->td_pcb->pcb_gsbase, rg, rgk);
-}
-#endif
-
 user:
 	userret(td, frame);
 	mtx_assert(&Giant, MA_NOTOWNED);

From 3087dc40a9128137f46cb4faedcec83b7ed62646 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sat, 1 May 2010 14:46:17 +0000
Subject: [PATCH 379/532] Extract thread_lock()/ruxagg()/thread_unlock()
 fragment into utility function ruxagg_tlock(). Convert the definition of
 kern_getrusage() to ANSI C.

Submitted by:	Alexander Krizhanovsky <ak natsys-lab com>
MFC after:	1 week
---
 sys/kern/kern_resource.c | 27 ++++++++++++++-------------
 1 file changed, 14 insertions(+), 13 deletions(-)

diff --git a/sys/kern/kern_resource.c b/sys/kern/kern_resource.c
index f8678395904..a3ed75d3583 100644
--- a/sys/kern/kern_resource.c
+++ b/sys/kern/kern_resource.c
@@ -76,6 +76,7 @@ static void	calcru1(struct proc *p, struct rusage_ext *ruxp,
 		    struct timeval *up, struct timeval *sp);
 static int	donice(struct thread *td, struct proc *chgp, int n);
 static struct uidinfo *uilookup(uid_t uid);
+static void	ruxagg_tlock(struct proc *p, struct thread *td);
 
 /*
  * Resource controls and accounting.
@@ -629,9 +630,7 @@ lim_cb(void *arg)
 		return;
 	PROC_SLOCK(p);
 	FOREACH_THREAD_IN_PROC(p, td) {
-		thread_lock(td);
-		ruxagg(&p->p_rux, td);
-		thread_unlock(td);
+		ruxagg_tlock(p, td);
 	}
 	PROC_SUNLOCK(p);
 	if (p->p_rux.rux_runtime > p->p_cpulimit * cpu_tickrate()) {
@@ -842,9 +841,7 @@ calcru(struct proc *p, struct timeval *up, struct timeval *sp)
 	FOREACH_THREAD_IN_PROC(p, td) {
 		if (td->td_incruntime == 0)
 			continue;
-		thread_lock(td);
-		ruxagg(&p->p_rux, td);
-		thread_unlock(td);
+		ruxagg_tlock(p, td);
 	}
 	calcru1(p, &p->p_rux, up, sp);
 }
@@ -945,10 +942,7 @@ getrusage(td, uap)
 }
 
 int
-kern_getrusage(td, who, rup)
-	struct thread *td;
-	int who;
-	struct rusage *rup;
+kern_getrusage(struct thread *td, int who, struct rusage *rup)
 {
 	struct proc *p;
 	int error;
@@ -1022,6 +1016,15 @@ ruxagg(struct rusage_ext *rux, struct thread *td)
 	td->td_sticks = 0;
 }
 
+static void
+ruxagg_tlock(struct proc *p, struct thread *td)
+{
+
+	thread_lock(td);
+	ruxagg(&p->p_rux, td);
+	thread_unlock(td);
+}
+
 /*
  * Update the rusage_ext structure and fetch a valid aggregate rusage
  * for proc p if storage for one is supplied.
@@ -1036,9 +1039,7 @@ rufetch(struct proc *p, struct rusage *ru)
 	*ru = p->p_ru;
 	if (p->p_numthreads > 0)  {
 		FOREACH_THREAD_IN_PROC(p, td) {
-			thread_lock(td);
-			ruxagg(&p->p_rux, td);
-			thread_unlock(td);
+			ruxagg_tlock(p, td);
 			rucollect(ru, &td->td_ru);
 		}
 	}

From bc391cb2ce90b59d981a2148eb64779640cec9cb Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Sat, 1 May 2010 16:36:14 +0000
Subject: [PATCH 380/532] The Atheros AR71xx CPUs, when paired with the AR5212
 parts, has a bug that generates a fatal bus trap.  Normally, the chips are
 setup to do 128 byte DMA bursts, but when on this CPU, they can only safely
 due 4-byte DMA bursts due to this bug.  Details of the exact nature of the
 bug are sketchy, but some can be found at
 https://forum.openwrt.org/viewtopic.php?pid=70060 on pages 4, 5 and 6.
 There's a small performance penalty associated with this workaround, so it is
 only enabled when needed on the Atheros AR71xx platforms.

Unfortunately, this condition is impossible to detect at runtime
without MIPS specific ifdefs.  Rather than cast an overly-broad net
like Linux/OpenWRT dues (which enables this workaround all the time on
MIPS32 platforms), we put this option in the kernel for just the
affected machines.  Sam didn't like this aspect of the patch when he
reviewed it, and I'd love to hear sane proposals on how to fix it :)

Reviewed by:	sam@
---
 sys/conf/NOTES                            | 9 +++++++++
 sys/conf/options                          | 1 +
 sys/dev/ath/ath_hal/ar5212/ar5212_reset.c | 8 ++++++++
 3 files changed, 18 insertions(+)

diff --git a/sys/conf/NOTES b/sys/conf/NOTES
index f0cc9a0bc10..cc1964d0458 100644
--- a/sys/conf/NOTES
+++ b/sys/conf/NOTES
@@ -2091,6 +2091,15 @@ device		ath_hal		# pci/cardbus chip support
 #device		ath_rf5413
 #device		ath_ar5416	# AR5416 chips
 options 	AH_SUPPORT_AR5416	# enable AR5416 tx/rx descriptors
+# All of the AR5212 parts have a problem when paired with the AR71xx
+# CPUS.  These parts have a bug that triggers a fatal bus error on the AR71xx
+# only.  Details of the exact nature of the bug are sketchy, but some can be
+# found at https://forum.openwrt.org/viewtopic.php?pid=70060 on pages 4, 5 and
+# 6.  This option enables this workaround.  There is a performance penalty
+# for this work around, but without it things don't work at all.  The DMA
+# from the card usually bursts 128 bytes, but on the affected CPUs, only
+# 4 are safe.
+options	   	AH_RXCFG_SDMAMW_4BYTES
 #device		ath_ar9160	# AR9160 chips
 #device		ath_ar9280	# AR9280 chips
 #device		ath_ar9285	# AR9285 chips
diff --git a/sys/conf/options b/sys/conf/options
index fabc0bf772f..8f1dcaaa90b 100644
--- a/sys/conf/options
+++ b/sys/conf/options
@@ -771,6 +771,7 @@ AH_PRIVATE_DIAG		opt_ah.h
 AH_NEED_DESC_SWAP	opt_ah.h
 AH_USE_INIPDGAIN	opt_ah.h
 AH_MAXCHAN		opt_ah.h
+AH_RXCFG_SDMAMW_4BYTES	opt_ah.h
 
 # options for the Broadcom BCM43xx driver (bwi)
 BWI_DEBUG		opt_bwi.h
diff --git a/sys/dev/ath/ath_hal/ar5212/ar5212_reset.c b/sys/dev/ath/ath_hal/ar5212/ar5212_reset.c
index fc937ea149c..8e6341a04c7 100644
--- a/sys/dev/ath/ath_hal/ar5212/ar5212_reset.c
+++ b/sys/dev/ath/ath_hal/ar5212/ar5212_reset.c
@@ -283,6 +283,14 @@ ar5212Reset(struct ath_hal *ah, HAL_OPMODE opmode,
 	regWrites = ath_hal_ini_write(ah, &ahp->ah_ini_modes, modesIndex, 0);
 	regWrites = write_common(ah, &ahp->ah_ini_common, bChannelChange,
 		regWrites);
+#ifdef AH_RXCFG_SDMAMW_4BYTES
+	/*
+	 * Nala doesn't work with 128 byte bursts on pb42(hydra) (ar71xx),
+	 * use 4 instead.  Enabling it on all platforms would hurt performance,
+	 * so we only enable it on the ones that are affected by it.
+	 */
+	OS_REG_WRITE(ah, AR_RXCFG, 0);
+#endif
 	ahp->ah_rfHal->writeRegs(ah, modesIndex, freqIndex, regWrites);
 
 	OS_MARK(ah, AH_MARK_RESET_LINE, __LINE__);

From 7e005410db8e796b2662121fcec50bb5c4f2a34e Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Sat, 1 May 2010 16:38:40 +0000
Subject: [PATCH 381/532] Put the -current debugging options back into AR71XX.

---
 sys/mips/conf/AR71XX | 10 +++++-----
 1 file changed, 5 insertions(+), 5 deletions(-)

diff --git a/sys/mips/conf/AR71XX b/sys/mips/conf/AR71XX
index 4ee6aab8a1d..5e9499b2720 100644
--- a/sys/mips/conf/AR71XX
+++ b/sys/mips/conf/AR71XX
@@ -29,11 +29,11 @@ options		_KPOSIX_PRIORITY_SCHEDULING #Posix P1003_1B real-time extensions
 
 # options		NFS_LEGACYRPC
 # Debugging for use in -current
-# options		DEADLKRES
-# options		INVARIANTS
-# options		INVARIANT_SUPPORT
-# options		WITNESS
-# options		WITNESS_SKIPSPIN
+options		DEADLKRES
+options		INVARIANTS
+options		INVARIANT_SUPPORT
+options		WITNESS
+options		WITNESS_SKIPSPIN
 options         FFS                     #Berkeley Fast Filesystem
 options         SOFTUPDATES             #Enable FFS soft updates support
 options         UFS_ACL                 #Support for access control lists

From a932985f50a9ca5176e1e02a1a1c2b5b5bb95433 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Sat, 1 May 2010 16:39:46 +0000
Subject: [PATCH 382/532] Enable AH_RXCFG_SDMAMW_4BYTES option.  See NOTES file
 for why this is workaround (WAR) is needed.

---
 sys/mips/conf/AR71XX | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/mips/conf/AR71XX b/sys/mips/conf/AR71XX
index 5e9499b2720..12c73daa5b8 100644
--- a/sys/mips/conf/AR71XX
+++ b/sys/mips/conf/AR71XX
@@ -62,6 +62,7 @@ device		ath             # Atheros pci/cardbus NIC's
 options 	ATH_DEBUG
 device		ath_hal
 option		AH_SUPPORT_AR5416
+option		AH_RXCFG_SDMAMW_4BYTES	# See NOTES for details of this WAR
 device		ath_rate_sample
 
 device		mii

From a83baab6e44131d87926395d6b006d32bc4f704c Mon Sep 17 00:00:00 2001
From: Marko Zec <zec@FreeBSD.org>
Date: Sat, 1 May 2010 18:34:50 +0000
Subject: [PATCH 383/532] Remove a redundant variable assignment.

Reviewed by:	bz, rwatson
MFC after:	3 days
---
 sys/kern/uipc_mbuf.c | 3 +--
 1 file changed, 1 insertion(+), 2 deletions(-)

diff --git a/sys/kern/uipc_mbuf.c b/sys/kern/uipc_mbuf.c
index bc0e88d0a47..85e0e395528 100644
--- a/sys/kern/uipc_mbuf.c
+++ b/sys/kern/uipc_mbuf.c
@@ -948,9 +948,8 @@ m_adj(struct mbuf *mp, int req_len)
 				len = 0;
 			}
 		}
-		m = mp;
 		if (mp->m_flags & M_PKTHDR)
-			m->m_pkthdr.len -= (req_len - len);
+			mp->m_pkthdr.len -= (req_len - len);
 	} else {
 		/*
 		 * Trim from tail.  Scan the mbuf chain,

From b97e003f422ad719016644fea22e9439ad3fe722 Mon Sep 17 00:00:00 2001
From: Ed Maste <emaste@FreeBSD.org>
Date: Sat, 1 May 2010 18:56:45 +0000
Subject: [PATCH 384/532] Search beyond the first 1/8th of inodes.

Submitted by:	jeff
---
 lib/libufs/cgroup.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/libufs/cgroup.c b/lib/libufs/cgroup.c
index b0cb4a7d63d..8ab33f7c7c9 100644
--- a/lib/libufs/cgroup.c
+++ b/lib/libufs/cgroup.c
@@ -145,7 +145,7 @@ cgialloc(struct uufsd *disk)
 	fs = &disk->d_fs;
 	cgp = &disk->d_cg;
 	inosused = cg_inosused(cgp);
-	for (ino = 0; ino < fs->fs_ipg / NBBY; ino++)
+	for (ino = 0; ino < fs->fs_ipg; ino++)
 		if (isclr(inosused, ino))
 			goto gotit;
 	return (0);

From 9ccdc9600e5bc44b42857799e0e949850139c722 Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Sat, 1 May 2010 19:53:15 +0000
Subject: [PATCH 385/532] Change description of tunable group vfs.zfs.txg to be
 more understandable.

Approved by:	pjd, delphij (mentor)
MFC after:	3 days
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
index 040e4d70fc0..143b4b87b08 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
@@ -40,7 +40,8 @@ int zfs_txg_timeout = 30;	/* max seconds worth of delta per txg */
 extern int zfs_txg_synctime;
 
 SYSCTL_DECL(_vfs_zfs);
-SYSCTL_NODE(_vfs_zfs, OID_AUTO, txg, CTLFLAG_RW, 0, "ZFS TXG");
+SYSCTL_NODE(_vfs_zfs, OID_AUTO, txg, CTLFLAG_RW, 0,
+    "ZFS transaction groups (TXG)");
 TUNABLE_INT("vfs.zfs.txg.timeout", &zfs_txg_timeout);
 SYSCTL_INT(_vfs_zfs_txg, OID_AUTO, timeout, CTLFLAG_RDTUN, &zfs_txg_timeout, 0,
     "Maximum seconds worth of delta per txg");

From df04ddbaa6f84fca9ed18ef224cff2ea412d7cce Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Sat, 1 May 2010 20:44:37 +0000
Subject: [PATCH 386/532] Add sysctl and loader tunable
 vfs.zfs.txg.write_limit_override. This tunable improves fine-tuning of ZFS
 write throttling.

PR:		kern/146108
Suggested by:	Nikolay Denev <ndenev at gmail.com>
Approved by:	pjd, delphij (mentor)
MFC after:	2 weeks
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
index 143b4b87b08..8650fa117e1 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/txg.c
@@ -38,6 +38,7 @@ static void txg_quiesce_thread(void *arg);
 
 int zfs_txg_timeout = 30;	/* max seconds worth of delta per txg */
 extern int zfs_txg_synctime;
+extern uint64_t zfs_write_limit_override;
 
 SYSCTL_DECL(_vfs_zfs);
 SYSCTL_NODE(_vfs_zfs, OID_AUTO, txg, CTLFLAG_RW, 0,
@@ -48,6 +49,11 @@ SYSCTL_INT(_vfs_zfs_txg, OID_AUTO, timeout, CTLFLAG_RDTUN, &zfs_txg_timeout, 0,
 TUNABLE_INT("vfs.zfs.txg.synctime", &zfs_txg_synctime);
 SYSCTL_INT(_vfs_zfs_txg, OID_AUTO, synctime, CTLFLAG_RDTUN, &zfs_txg_synctime,
     0, "Target seconds to sync a txg");
+TUNABLE_QUAD("vfs.zfs.txg.write_limit_override", &zfs_write_limit_override);
+SYSCTL_QUAD(_vfs_zfs_txg, OID_AUTO, write_limit_override, CTLFLAG_RW,
+    &zfs_write_limit_override, 0,
+    "Override maximum size of a txg to this size in bytes, "
+    "value of 0 means don't override");
 
 /*
  * Prepare the txg subsystem.

From aa1b887b418e1e4e05e2651b0e39302baf7ce9f1 Mon Sep 17 00:00:00 2001
From: Ryan Stone <rstone@FreeBSD.org>
Date: Sat, 1 May 2010 21:59:06 +0000
Subject: [PATCH 387/532] When configuring hwpmc to use the EXT_SNOOP event,
 only send a default cachestate qualifier on the Atom processor.  Other Intel
 processors do not accept a cachestate qualifier and currently hwpmc will
 return EINVAL if you try to use the EXT_SNOOP event on those processors

Approved by:	jkoshy (mentor)
MFC after:	2 weeks
---
 lib/libpmc/libpmc.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/lib/libpmc/libpmc.c b/lib/libpmc/libpmc.c
index c440aa8ee73..8922dd8c0e0 100644
--- a/lib/libpmc/libpmc.c
+++ b/lib/libpmc/libpmc.c
@@ -737,9 +737,16 @@ iap_allocate_pmc(enum pmc_event pe, char *ctrspec,
 	case PMC_EV_IAP_EVENT_40H: /* Core */
 	case PMC_EV_IAP_EVENT_41H: /* Core */
 	case PMC_EV_IAP_EVENT_42H: /* Core, Core2, Atom */
-	case PMC_EV_IAP_EVENT_77H: /* Core */
 		if (cachestate == 0)
 			cachestate = (0xF << 8);
+		break;
+	case PMC_EV_IAP_EVENT_77H: /* Atom */
+		/* IAP_EVENT_77H only accepts a cachestate qualifier on the
+		 * Atom processor
+		 */
+		if(cpu_info.pm_cputype == PMC_CPU_INTEL_ATOM && cachestate == 0)
+			cachestate = (0xF << 8);
+	    break;
 	default:
 		break;
 	}

From da2195255971eb5f04fedad2456f852ff5e8e8c2 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 1 May 2010 22:00:28 +0000
Subject: [PATCH 388/532] pathchk: Add the new POSIX -P option.

This option checks for empty pathnames and components starting with '-'.
Our -p option also checks for the latter, which remains the case.

MFC after:	1 week
---
 usr.bin/pathchk/pathchk.1 | 14 ++++++++++++--
 usr.bin/pathchk/pathchk.c | 18 ++++++++++++++----
 2 files changed, 26 insertions(+), 6 deletions(-)

diff --git a/usr.bin/pathchk/pathchk.1 b/usr.bin/pathchk/pathchk.1
index e8639559cf4..fb50e84c4b7 100644
--- a/usr.bin/pathchk/pathchk.1
+++ b/usr.bin/pathchk/pathchk.1
@@ -27,7 +27,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd May 21, 2002
+.Dd May 1, 2010
 .Dt PATHCHK 1
 .Os
 .Sh NAME
@@ -35,7 +35,7 @@
 .Nd check pathnames
 .Sh SYNOPSIS
 .Nm
-.Op Fl p
+.Op Fl pP
 .Ar pathname ...
 .Sh DESCRIPTION
 The
@@ -95,6 +95,16 @@ No component may start with the hyphen
 .Pq Ql \&-
 character.
 .El
+.It Fl P
+In addition to the default or
+.Fl p
+checks, write a diagnostic for each argument that:
+.Bl -bullet
+.It
+Is empty.
+.It
+Contains a component that starts with a hyphen.
+.El
 .El
 .Sh EXIT STATUS
 .Ex -std
diff --git a/usr.bin/pathchk/pathchk.c b/usr.bin/pathchk/pathchk.c
index dd9768a497e..2d8fa595d32 100644
--- a/usr.bin/pathchk/pathchk.c
+++ b/usr.bin/pathchk/pathchk.c
@@ -51,6 +51,7 @@ static int	 portable(const char *);
 static void	 usage(void);
 
 static int	 pflag;			/* Perform portability checks */
+static int	 Pflag;			/* Check for empty paths, leading '-' */
 
 int
 main(int argc, char *argv[])
@@ -58,11 +59,14 @@ main(int argc, char *argv[])
 	int ch, rval;
 	const char *arg;
 
-	while ((ch = getopt(argc, argv, "p")) > 0) {
+	while ((ch = getopt(argc, argv, "pP")) > 0) {
 		switch (ch) {
 		case 'p':
 			pflag = 1;
 			break;
+		case 'P':
+			Pflag = 1;
+			break;
 		default:
 			usage();
 			/*NOTREACHED*/
@@ -102,6 +106,15 @@ check(const char *path)
 
 	p = pathd;
 
+	if (Pflag && *p == '\0') {
+		warnx("%s: empty pathname", path);
+		goto bad;
+	}
+	if ((Pflag || pflag) && (*p == '-' || strstr(p, "/-") != NULL)) {
+		warnx("%s: contains a component starting with '-'", path);
+		goto bad;
+	}
+
 	if (!pflag) {
 		errno = 0;
 		namemax = pathconf(*p == '/' ? "/" : ".", _PC_NAME_MAX);
@@ -182,9 +195,6 @@ portable(const char *path)
 	    "0123456789._-";
 	long s;
 
-	if (*path == '-')
-		return (*path);
-
 	s = strspn(path, charset);
 	if (path[s] != '\0')
 		return (path[s]);

From 04001891bbbea3ca38b5395259af8f8876a2abc6 Mon Sep 17 00:00:00 2001
From: Ryan Stone <rstone@FreeBSD.org>
Date: Sat, 1 May 2010 22:04:58 +0000
Subject: [PATCH 389/532] When configuring a system-wide couting PMC, hwpmc was
 incorrectly logging process mappings for that PMC.  Nothing ever reads pmc
 logs out of a counting PMC, so the log buffers were leaked when the PMC was
 deconfigured.  The process mappings are only useful for sampling PMCs anyway,
 so only log the mappings if the PMC is a sampling PMC.

This bug would cause allocating sample-mode PMCs to fail with ENOMEM after allocating several counting-mode PMCs.

Approved by:	jkoshy (mentor)
MFC after:	2 weeks
---
 sys/dev/hwpmc/hwpmc_mod.c | 18 +++++++++---------
 1 file changed, 9 insertions(+), 9 deletions(-)

diff --git a/sys/dev/hwpmc/hwpmc_mod.c b/sys/dev/hwpmc/hwpmc_mod.c
index e33b431e7e3..a19c85ec379 100644
--- a/sys/dev/hwpmc/hwpmc_mod.c
+++ b/sys/dev/hwpmc/hwpmc_mod.c
@@ -2675,16 +2675,16 @@ pmc_start(struct pmc *pm)
 			PMCDBG(PMC,OPS,1, "po=%p in global list", po);
 		}
 		po->po_sscount++;
-	}
 
-	/*
-	 * Log mapping information for all existing processes in the
-	 * system.  Subsequent mappings are logged as they happen;
-	 * see pmc_process_mmap().
-	 */
-	if (po->po_logprocmaps == 0) {
-		pmc_log_all_process_mappings(po);
-		po->po_logprocmaps = 1;
+		/*
+		 * Log mapping information for all existing processes in the
+		 * system.  Subsequent mappings are logged as they happen;
+		 * see pmc_process_mmap().
+		 */
+		if (po->po_logprocmaps == 0) {
+			pmc_log_all_process_mappings(po);
+			po->po_logprocmaps = 1;
+		}
 	}
 
 	/*

From 8f3bd46654badc4318169b43e04d367ba23b1e5c Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 1 May 2010 22:10:45 +0000
Subject: [PATCH 390/532] pathchk(1): Fix the example so it allows arbitrary
 pathnames.

Spaces and various other characters in pathnames are not passed through
literally by xargs in its default mode. Instead, use find . -exec ... {} +

Although the -- argument is not strictly required here, add it anyway to
avoid surprises when modifying the code to find -f -somedir ...

MFC after:	1 week
---
 usr.bin/pathchk/pathchk.1 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.bin/pathchk/pathchk.1 b/usr.bin/pathchk/pathchk.1
index fb50e84c4b7..931f82f47e6 100644
--- a/usr.bin/pathchk/pathchk.1
+++ b/usr.bin/pathchk/pathchk.1
@@ -114,7 +114,7 @@ other
 .Tn POSIX
 systems:
 .Pp
-.Dl "find . -print | xargs pathchk -p"
+.Dl "find . -exec pathchk -p -- {} +"
 .Sh SEE ALSO
 .Xr getconf 1 ,
 .Xr pathconf 2 ,

From 9b55fc042996c9c9ed8fab6e0eea9239937919b7 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 01:25:03 +0000
Subject: [PATCH 391/532] Correct an error of omission in r206819.  If
 VMFS_TLB_ALIGNED_SPACE is specified to vm_map_find(), then retry the
 vm_map_findspace() if vm_map_insert() fails because the aligned space is
 already partly used.

Reported by:	Neel Natu
---
 sys/vm/vm_map.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/sys/vm/vm_map.c b/sys/vm/vm_map.c
index 1d22fa6b116..2a57e33b6e5 100644
--- a/sys/vm/vm_map.c
+++ b/sys/vm/vm_map.c
@@ -1412,7 +1412,11 @@ vm_map_find(vm_map_t map, vm_object_t object, vm_ooffset_t offset,
 		}
 		result = vm_map_insert(map, object, offset, start, start +
 		    length, prot, max, cow);
-	} while (result == KERN_NO_SPACE && find_space == VMFS_ALIGNED_SPACE);
+	} while (result == KERN_NO_SPACE && (find_space == VMFS_ALIGNED_SPACE
+#ifdef VMFS_TLB_ALIGNED_SPACE
+	    || find_space == VMFS_TLB_ALIGNED_SPACE
+#endif
+	    ));
 	vm_map_unlock(map);
 	return (result);
 }

From 8caae13f6879c2080a6c5ec8b56d48b6bee564f7 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Sun, 2 May 2010 04:16:39 +0000
Subject: [PATCH 392/532] Add xpt_schedule_dev_sendq() call, lost at r203108.
 It is not needed in usual operation, but required in some conditions to make
 queue running after being shrinked.

MFC after:	3 days
---
 sys/cam/cam_xpt.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/cam/cam_xpt.c b/sys/cam/cam_xpt.c
index ed05c40f729..5d83539c899 100644
--- a/sys/cam/cam_xpt.c
+++ b/sys/cam/cam_xpt.c
@@ -4874,6 +4874,8 @@ camisr_runqueue(void *V_queue)
 			if ((dev->flags & CAM_DEV_TAG_AFTER_COUNT) != 0
 			 && (--dev->tag_delay_count == 0))
 				xpt_start_tags(ccb_h->path);
+			if (!device_is_send_queued(dev))
+				xpt_schedule_dev_sendq(ccb_h->path->bus, dev);
 		}
 
 		if (ccb_h->status & CAM_RELEASE_SIMQ) {

From f930c0db49a2049e7b92ec49b7736546839ed0a3 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Sun, 2 May 2010 11:36:27 +0000
Subject: [PATCH 393/532] Add -d and -f arguments to `camcontrol cmd`, to
 execute DMA ATA commands.

---
 sbin/camcontrol/camcontrol.8 |  8 +++++++-
 sbin/camcontrol/camcontrol.c | 16 ++++++++++++++--
 2 files changed, 21 insertions(+), 3 deletions(-)

diff --git a/sbin/camcontrol/camcontrol.8 b/sbin/camcontrol/camcontrol.8
index d1e9587ab95..1a075b3624d 100644
--- a/sbin/camcontrol/camcontrol.8
+++ b/sbin/camcontrol/camcontrol.8
@@ -27,7 +27,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd January 20, 2010
+.Dd May 2, 2010
 .Dt CAMCONTROL 8
 .Os
 .Sh NAME
@@ -123,6 +123,8 @@
 .Op generic args
 .Aq Fl a Ar cmd Op args
 .Aq Fl c Ar cmd Op args
+.Op Fl d
+.Op Fl f
 .Op Fl i Ar len Ar fmt
 .Bk -words
 .Op Fl o Ar len Ar fmt Op args
@@ -530,6 +532,10 @@ lba_high_exp, features_exp, sector_count, sector_count_exp).
 .It Fl c Ar cmd Op args
 This specifies the SCSI CDB.
 SCSI CDBs may be 6, 10, 12 or 16 bytes.
+.It Fl d
+Specifies DMA protocol to be used for ATA command.
+.It Fl f
+Specifies FPDMA (NCQ) protocol to be used for ATA command.
 .It Fl i Ar len Ar fmt
 This specifies the amount of data to read, and how it should be displayed.
 If the format is
diff --git a/sbin/camcontrol/camcontrol.c b/sbin/camcontrol/camcontrol.c
index 8e0605f6de2..9136cbe3469 100644
--- a/sbin/camcontrol/camcontrol.c
+++ b/sbin/camcontrol/camcontrol.c
@@ -123,7 +123,7 @@ struct camcontrol_opts {
 };
 
 #ifndef MINIMALISTIC
-static const char scsicmd_opts[] = "a:c:i:o:r";
+static const char scsicmd_opts[] = "a:c:dfi:o:r";
 static const char readdefect_opts[] = "f:GP";
 static const char negotiate_opts[] = "acD:M:O:qR:T:UW:";
 #endif
@@ -2184,6 +2184,8 @@ scsicmd(struct cam_device *device, int argc, char **argv, char *combinedopt,
 	int c, data_bytes = 0;
 	int cdb_len = 0;
 	int atacmd_len = 0;
+	int dmacmd = 0;
+	int fpdmacmd = 0;
 	int need_res = 0;
 	char *datastr = NULL, *tstr, *resstr = NULL;
 	int error = 0;
@@ -2246,6 +2248,12 @@ scsicmd(struct cam_device *device, int argc, char **argv, char *combinedopt,
 			 */
 			optind += hook.got;
 			break;
+		case 'd':
+			dmacmd = 1;
+			break;
+		case 'f':
+			fpdmacmd = 1;
+			break;
 		case 'i':
 			if (arglist & CAM_ARG_CMD_OUT) {
 				warnx("command must either be "
@@ -2422,6 +2430,10 @@ scsicmd(struct cam_device *device, int argc, char **argv, char *combinedopt,
 		bcopy(atacmd, &ccb->ataio.cmd.command, atacmd_len);
 		if (need_res)
 			ccb->ataio.cmd.flags |= CAM_ATAIO_NEEDRESULT;
+		if (dmacmd)
+			ccb->ataio.cmd.flags |= CAM_ATAIO_DMA;
+		if (fpdmacmd)
+			ccb->ataio.cmd.flags |= CAM_ATAIO_FPDMA;
 
 		cam_fill_ataio(&ccb->ataio,
 		      /*retries*/ retry_count,
@@ -4353,7 +4365,7 @@ usage(int verbose)
 "                              [-P pagectl][-e | -b][-d]\n"
 "        camcontrol cmd        [dev_id][generic args]\n"
 "                              <-a cmd [args] | -c cmd [args]>\n"
-"                              [-i len fmt|-o len fmt [args]] [-r fmt]\n"
+"                              [-d] [-f] [-i len fmt|-o len fmt [args]] [-r fmt]\n"
 "        camcontrol debug      [-I][-P][-T][-S][-X][-c]\n"
 "                              <all|bus[:target[:lun]]|off>\n"
 "        camcontrol tags       [dev_id][generic args] [-N tags] [-q] [-v]\n"

From da6808c111755c13b9444a9838ce2415644a5d3f Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Sun, 2 May 2010 12:07:47 +0000
Subject: [PATCH 394/532] Make SATA XPT negotiate and enable some additional
 SATA features, such as:  - device initiated power management (some devices
 support only this way);  - Automatic Partial to Slumber Transition (more
 power saving);  - DMA auto-activation (expected to slightly improve
 performance). More features could be added later, when hardware supports.

---
 sbin/camcontrol/camcontrol.c |   4 +
 sys/cam/ata/ata_pmp.c        |  20 ++++-
 sys/cam/ata/ata_xpt.c        | 137 +++++++++++++++++++++++++++++
 sys/cam/cam_ccb.h            |   9 ++
 sys/dev/ahci/ahci.c          | 166 ++++++++++++++++++++---------------
 sys/dev/ahci/ahci.h          |   1 +
 sys/dev/siis/siis.c          |  12 +++
 sys/dev/siis/siis.h          |   1 +
 8 files changed, 277 insertions(+), 73 deletions(-)

diff --git a/sbin/camcontrol/camcontrol.c b/sbin/camcontrol/camcontrol.c
index 9136cbe3469..c39d7ba3704 100644
--- a/sbin/camcontrol/camcontrol.c
+++ b/sbin/camcontrol/camcontrol.c
@@ -2855,6 +2855,10 @@ cts_print(struct cam_device *device, struct ccb_trans_settings *cts)
 			fprintf(stdout, "%sNumber of tags: %d\n", pathstr,
 				sata->tags);
 		}
+		if ((sata->valid & CTS_SATA_VALID_CAPS) != 0) {
+			fprintf(stdout, "%sSATA capabilities: %08x\n", pathstr,
+				sata->caps);
+		}
 	}
 	if (cts->protocol == PROTO_SCSI) {
 		struct ccb_trans_settings_scsi *scsi=
diff --git a/sys/cam/ata/ata_pmp.c b/sys/cam/ata/ata_pmp.c
index b60a4a7596b..0835a2bf228 100644
--- a/sys/cam/ata/ata_pmp.c
+++ b/sys/cam/ata/ata_pmp.c
@@ -101,6 +101,7 @@ struct pmp_softc {
 	int			events;
 #define PMP_EV_RESET	1
 #define PMP_EV_RESCAN	2
+	u_int			caps;
 	struct task		sysctl_task;
 	struct sysctl_ctx_list	sysctl_ctx;
 	struct sysctl_oid	*sysctl_tree;
@@ -457,6 +458,14 @@ pmpstart(struct cam_periph *periph, union ccb *start_ccb)
 		ata_pm_read_cmd(ataio, 2, 15);
 		break;
 	case PMP_STATE_PRECONFIG:
+		/* Get/update host SATA capabilities. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, periph->path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_CURRENT_SETTINGS;
+		xpt_action((union ccb *)&cts);
+		if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			softc->caps = cts.xport_specific.sata.caps;
 		cam_fill_ataio(ataio,
 		      pmp_retry_count,
 		      pmpdone,
@@ -644,14 +653,16 @@ pmpdone(struct cam_periph *periph, union ccb *done_ccb)
 		    (done_ccb->ataio.res.lba_mid << 16) +
 		    (done_ccb->ataio.res.lba_low << 8) +
 		    done_ccb->ataio.res.sector_count;
-		if ((res & 0xf0f) == 0x103 && (res & 0x0f0) != 0) {
+		if (((res & 0xf0f) == 0x103 && (res & 0x0f0) != 0) ||
+		    (res & 0x600) != 0) {
 			if (bootverbose) {
 				printf("%s%d: port %d status: %08x\n",
 				    periph->periph_name, periph->unit_number,
 				    softc->pm_step, res);
 			}
-			/* Report device speed. */
-			if (xpt_create_path(&dpath, periph,
+			/* Report device speed if it is online. */
+			if ((res & 0xf0f) == 0x103 &&
+			    xpt_create_path(&dpath, periph,
 			    xpt_path_path_id(periph->path),
 			    softc->pm_step, 0) == CAM_REQ_CMP) {
 				bzero(&cts, sizeof(cts));
@@ -660,6 +671,9 @@ pmpdone(struct cam_periph *periph, union ccb *done_ccb)
 				cts.type = CTS_TYPE_CURRENT_SETTINGS;
 				cts.xport_specific.sata.revision = (res & 0x0f0) >> 4;
 				cts.xport_specific.sata.valid = CTS_SATA_VALID_REVISION;
+				cts.xport_specific.sata.caps = softc->caps &
+				    (CTS_SATA_CAPS_H_PMREQ | CTS_SATA_CAPS_H_DMAAA);
+				cts.xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
 				xpt_action((union ccb *)&cts);
 				xpt_free_path(dpath);
 			}
diff --git a/sys/cam/ata/ata_xpt.c b/sys/cam/ata/ata_xpt.c
index bd748cd4f45..478b0060ce4 100644
--- a/sys/cam/ata/ata_xpt.c
+++ b/sys/cam/ata/ata_xpt.c
@@ -88,6 +88,9 @@ typedef enum {
 	PROBE_IDENTIFY,
 	PROBE_SPINUP,
 	PROBE_SETMODE,
+	PROBE_SETPM,
+	PROBE_SETAPST,
+	PROBE_SETDMAAA,
 	PROBE_SET_MULTI,
 	PROBE_INQUIRY,
 	PROBE_FULL_INQUIRY,
@@ -101,6 +104,9 @@ static char *probe_action_text[] = {
 	"PROBE_IDENTIFY",
 	"PROBE_SPINUP",
 	"PROBE_SETMODE",
+	"PROBE_SETPM",
+	"PROBE_SETAPST",
+	"PROBE_SETDMAAA",
 	"PROBE_SET_MULTI",
 	"PROBE_INQUIRY",
 	"PROBE_FULL_INQUIRY",
@@ -132,6 +138,7 @@ typedef struct {
 	uint32_t	pm_prv;
 	int		restart;
 	int		spinup;
+	u_int		caps;
 	struct cam_periph *periph;
 } probe_softc;
 
@@ -393,6 +400,45 @@ negotiate:
 		ata_28bit_cmd(ataio, ATA_SETFEATURES, ATA_SF_SETXFER, 0, mode);
 		break;
 	}
+	case PROBE_SETPM:
+		cam_fill_ataio(ataio,
+		    1,
+		    probedone,
+		    CAM_DIR_NONE,
+		    0,
+		    NULL,
+		    0,
+		    30*1000);
+		ata_28bit_cmd(ataio, ATA_SETFEATURES,
+		    (softc->caps & CTS_SATA_CAPS_H_PMREQ) ? 0x10 : 0x90,
+		    0, 0x03);
+		break;
+	case PROBE_SETAPST:
+		cam_fill_ataio(ataio,
+		    1,
+		    probedone,
+		    CAM_DIR_NONE,
+		    0,
+		    NULL,
+		    0,
+		    30*1000);
+		ata_28bit_cmd(ataio, ATA_SETFEATURES,
+		    (softc->caps & CTS_SATA_CAPS_H_APST) ? 0x10 : 0x90,
+		    0, 0x07);
+		break;
+	case PROBE_SETDMAAA:
+		cam_fill_ataio(ataio,
+		    1,
+		    probedone,
+		    CAM_DIR_NONE,
+		    0,
+		    NULL,
+		    0,
+		    30*1000);
+		ata_28bit_cmd(ataio, ATA_SETFEATURES,
+		    (softc->caps & CTS_SATA_CAPS_H_DMAAA) ? 0x10 : 0x90,
+		    0, 0x02);
+		break;
 	case PROBE_SET_MULTI:
 	{
 		u_int sectors, bytecount;
@@ -685,6 +731,7 @@ probedone(struct cam_periph *periph, union ccb *done_ccb)
 	probe_softc *softc;
 	struct cam_path *path;
 	u_int32_t  priority;
+	u_int caps;
 	int found = 1;
 
 	CAM_DEBUG(done_ccb->ccb_h.path, CAM_DEBUG_TRACE, ("probedone\n"));
@@ -879,6 +926,67 @@ noerror:
 		xpt_schedule(periph, priority);
 		return;
 	case PROBE_SETMODE:
+		if (path->device->transport != XPORT_SATA)
+			goto notsata;
+		/* Set supported bits. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_CURRENT_SETTINGS;
+		xpt_action((union ccb *)&cts);
+		if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			caps = cts.xport_specific.sata.caps & CTS_SATA_CAPS_H;
+		else
+			caps = 0;
+		if (ident_buf->satacapabilities != 0xffff) {
+			if (ident_buf->satacapabilities & ATA_SUPPORT_IFPWRMNGTRCV)
+				caps |= CTS_SATA_CAPS_D_PMREQ;
+			if (ident_buf->satacapabilities & ATA_SUPPORT_HAPST)
+				caps |= CTS_SATA_CAPS_D_APST;
+		}
+		/* Mask unwanted bits. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_USER_SETTINGS;
+		xpt_action((union ccb *)&cts);
+		if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			caps &= cts.xport_specific.sata.caps;
+		/* Store result to SIM. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_CURRENT_SETTINGS;
+		cts.xport_specific.sata.caps = caps;
+		cts.xport_specific.sata.valid = CTS_SATA_VALID_CAPS;
+		xpt_action((union ccb *)&cts);
+		softc->caps = caps;
+		if (ident_buf->satasupport & ATA_SUPPORT_IFPWRMNGT) {
+			PROBE_SET_ACTION(softc, PROBE_SETPM);
+			xpt_release_ccb(done_ccb);
+			xpt_schedule(periph, priority);
+			return;
+		}
+		/* FALLTHROUGH */
+	case PROBE_SETPM:
+		if (ident_buf->satacapabilities != 0xffff &&
+		    ident_buf->satacapabilities & ATA_SUPPORT_DAPST) {
+			PROBE_SET_ACTION(softc, PROBE_SETAPST);
+			xpt_release_ccb(done_ccb);
+			xpt_schedule(periph, priority);
+			return;
+		}
+		/* FALLTHROUGH */
+	case PROBE_SETAPST:
+		if (ident_buf->satasupport & ATA_SUPPORT_AUTOACTIVATE) {
+			PROBE_SET_ACTION(softc, PROBE_SETDMAAA);
+			xpt_release_ccb(done_ccb);
+			xpt_schedule(periph, priority);
+			return;
+		}
+		/* FALLTHROUGH */
+	case PROBE_SETDMAAA:
+notsata:
 		if (path->device->protocol == PROTO_ATA) {
 			PROBE_SET_ACTION(softc, PROBE_SET_MULTI);
 		} else {
@@ -964,6 +1072,35 @@ noerror:
 		snprintf(ident_buf->revision, sizeof(ident_buf->revision),
 		    "%04x", softc->pm_prv);
 		path->device->flags |= CAM_DEV_IDENTIFY_DATA_VALID;
+		/* Set supported bits. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_CURRENT_SETTINGS;
+		xpt_action((union ccb *)&cts);
+		if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			caps = cts.xport_specific.sata.caps & CTS_SATA_CAPS_H;
+		else
+			caps = 0;
+		/* All PMPs must support PM requests. */
+		caps |= CTS_SATA_CAPS_D_PMREQ;
+		/* Mask unwanted bits. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_GET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_USER_SETTINGS;
+		xpt_action((union ccb *)&cts);
+		if (cts.xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			caps &= cts.xport_specific.sata.caps;
+		/* Store result to SIM. */
+		bzero(&cts, sizeof(cts));
+		xpt_setup_ccb(&cts.ccb_h, path, CAM_PRIORITY_NONE);
+		cts.ccb_h.func_code = XPT_SET_TRAN_SETTINGS;
+		cts.type = CTS_TYPE_CURRENT_SETTINGS;
+		cts.xport_specific.sata.caps = caps;
+		cts.xport_specific.sata.valid = CTS_SATA_VALID_CAPS;
+		xpt_action((union ccb *)&cts);
+		softc->caps = caps;
 		if (periph->path->device->flags & CAM_DEV_UNCONFIGURED) {
 			path->device->flags &= ~CAM_DEV_UNCONFIGURED;
 			xpt_acquire_device(path->device);
diff --git a/sys/cam/cam_ccb.h b/sys/cam/cam_ccb.h
index 2810c545f71..c5ac35e9846 100644
--- a/sys/cam/cam_ccb.h
+++ b/sys/cam/cam_ccb.h
@@ -837,12 +837,21 @@ struct ccb_trans_settings_sata {
 #define	CTS_SATA_VALID_PM		0x08
 #define	CTS_SATA_VALID_TAGS		0x10
 #define	CTS_SATA_VALID_ATAPI		0x20
+#define	CTS_SATA_VALID_CAPS		0x40
 	int		mode;		/* Legacy PATA mode */
 	u_int 		bytecount;	/* Length of PIO transaction */
 	int		revision;	/* SATA revision */
 	u_int 		pm_present;	/* PM is present (XPT->SIM) */
 	u_int 		tags;		/* Number of allowed tags */
 	u_int 		atapi;		/* Length of ATAPI CDB */
+	u_int 		caps;		/* Device and host SATA caps. */
+#define	CTS_SATA_CAPS_H			0x0000ffff
+#define	CTS_SATA_CAPS_H_PMREQ		0x00000001
+#define	CTS_SATA_CAPS_H_APST		0x00000002
+#define	CTS_SATA_CAPS_H_DMAAA		0x00000010 /* Auto-activation */
+#define	CTS_SATA_CAPS_D			0xffff0000
+#define	CTS_SATA_CAPS_D_PMREQ		0x00010000
+#define	CTS_SATA_CAPS_D_APST		0x00020000
 };
 
 /* Get/Set transfer rate/width/disconnection/tag queueing settings */
diff --git a/sys/dev/ahci/ahci.c b/sys/dev/ahci/ahci.c
index 34f1be426b1..60fb85153a8 100644
--- a/sys/dev/ahci/ahci.c
+++ b/sys/dev/ahci/ahci.c
@@ -111,6 +111,7 @@ static struct {
 #define AHCI_Q_EDGEIS	64
 #define AHCI_Q_SATA2	128
 #define AHCI_Q_NOBSYRES	256
+#define AHCI_Q_NOAA	512
 } ahci_ids[] = {
 	{0x43801002, 0x00, "ATI IXP600",	0},
 	{0x43901002, 0x00, "ATI IXP700",	0},
@@ -167,75 +168,75 @@ static struct {
 	{0x614511ab, 0x00, "Marvell 88SX6145",	AHCI_Q_NOFORCE|AHCI_Q_4CH|AHCI_Q_EDGEIS},
 	{0x91231b4b, 0x11, "Marvell 88SE912x",	AHCI_Q_NOBSYRES},
 	{0x91231b4b, 0x00, "Marvell 88SE912x",	AHCI_Q_EDGEIS|AHCI_Q_SATA2|AHCI_Q_NOBSYRES},
-	{0x044c10de, 0x00, "NVIDIA MCP65",	0},
-	{0x044d10de, 0x00, "NVIDIA MCP65",	0},
-	{0x044e10de, 0x00, "NVIDIA MCP65",	0},
-	{0x044f10de, 0x00, "NVIDIA MCP65",	0},
-	{0x045c10de, 0x00, "NVIDIA MCP65",	0},
-	{0x045d10de, 0x00, "NVIDIA MCP65",	0},
-	{0x045e10de, 0x00, "NVIDIA MCP65",	0},
-	{0x045f10de, 0x00, "NVIDIA MCP65",	0},
-	{0x055010de, 0x00, "NVIDIA MCP67",	0},
-	{0x055110de, 0x00, "NVIDIA MCP67",	0},
-	{0x055210de, 0x00, "NVIDIA MCP67",	0},
-	{0x055310de, 0x00, "NVIDIA MCP67",	0},
-	{0x055410de, 0x00, "NVIDIA MCP67",	0},
-	{0x055510de, 0x00, "NVIDIA MCP67",	0},
-	{0x055610de, 0x00, "NVIDIA MCP67",	0},
-	{0x055710de, 0x00, "NVIDIA MCP67",	0},
-	{0x055810de, 0x00, "NVIDIA MCP67",	0},
-	{0x055910de, 0x00, "NVIDIA MCP67",	0},
-	{0x055A10de, 0x00, "NVIDIA MCP67",	0},
-	{0x055B10de, 0x00, "NVIDIA MCP67",	0},
-	{0x058410de, 0x00, "NVIDIA MCP67",	0},
-	{0x07f010de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f110de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f210de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f310de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f410de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f510de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f610de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f710de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f810de, 0x00, "NVIDIA MCP73",	0},
-	{0x07f910de, 0x00, "NVIDIA MCP73",	0},
-	{0x07fa10de, 0x00, "NVIDIA MCP73",	0},
-	{0x07fb10de, 0x00, "NVIDIA MCP73",	0},
-	{0x0ad010de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad110de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad210de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad310de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad410de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad510de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad610de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad710de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad810de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ad910de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ada10de, 0x00, "NVIDIA MCP77",	0},
-	{0x0adb10de, 0x00, "NVIDIA MCP77",	0},
-	{0x0ab410de, 0x00, "NVIDIA MCP79",	0},
-	{0x0ab510de, 0x00, "NVIDIA MCP79",	0},
-	{0x0ab610de, 0x00, "NVIDIA MCP79",	0},
-	{0x0ab710de, 0x00, "NVIDIA MCP79",	0},
-	{0x0ab810de, 0x00, "NVIDIA MCP79",	0},
-	{0x0ab910de, 0x00, "NVIDIA MCP79",	0},
-	{0x0aba10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0abb10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0abc10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0abd10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0abe10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0abf10de, 0x00, "NVIDIA MCP79",	0},
-	{0x0d8410de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8510de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8610de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8710de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8810de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8910de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8a10de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8b10de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8c10de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8d10de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8e10de, 0x00, "NVIDIA MCP89",	0},
-	{0x0d8f10de, 0x00, "NVIDIA MCP89",	0},
+	{0x044c10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x044d10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x044e10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x044f10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x045c10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x045d10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x045e10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x045f10de, 0x00, "NVIDIA MCP65",	AHCI_Q_NOAA},
+	{0x055010de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055110de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055210de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055310de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055410de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055510de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055610de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055710de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055810de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055910de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055A10de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x055B10de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x058410de, 0x00, "NVIDIA MCP67",	AHCI_Q_NOAA},
+	{0x07f010de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f110de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f210de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f310de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f410de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f510de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f610de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f710de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f810de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07f910de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07fa10de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x07fb10de, 0x00, "NVIDIA MCP73",	AHCI_Q_NOAA},
+	{0x0ad010de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad110de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad210de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad310de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad410de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad510de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad610de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad710de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad810de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ad910de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ada10de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0adb10de, 0x00, "NVIDIA MCP77",	AHCI_Q_NOAA},
+	{0x0ab410de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0ab510de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0ab610de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0ab710de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0ab810de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0ab910de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0aba10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0abb10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0abc10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0abd10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0abe10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0abf10de, 0x00, "NVIDIA MCP79",	AHCI_Q_NOAA},
+	{0x0d8410de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8510de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8610de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8710de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8810de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8910de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8a10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8b10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8c10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8d10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8e10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
+	{0x0d8f10de, 0x00, "NVIDIA MCP89",	AHCI_Q_NOAA},
 	{0x33491106, 0x00, "VIA VT8251",	0},
 	{0x62871106, 0x00, "VIA VT8251",	0},
 	{0x11841039, 0x00, "SiS 966",		0},
@@ -860,7 +861,14 @@ ahci_ch_attach(device_t dev)
 		ch->user[i].mode = 0;
 		ch->user[i].bytecount = 8192;
 		ch->user[i].tags = ch->numslots;
+		ch->user[i].caps = 0;
 		ch->curr[i] = ch->user[i];
+		if (ch->pm_level) {
+			ch->user[i].caps = CTS_SATA_CAPS_H_PMREQ |
+			    CTS_SATA_CAPS_H_APST |
+			    CTS_SATA_CAPS_D_PMREQ | CTS_SATA_CAPS_D_APST;
+		}
+		ch->user[i].caps |= CTS_SATA_CAPS_H_DMAAA;
 	}
 	rid = ch->unit;
 	if (!(ch->r_mem = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
@@ -1960,7 +1968,8 @@ ahci_end_transaction(struct ahci_slot *slot, enum ahci_err_type et)
 	    et != AHCI_ERR_TIMEOUT)
 		ahci_rearm_timeout(dev);
 	/* Start PM timer. */
-	if (ch->numrslots == 0 && ch->pm_level > 3) {
+	if (ch->numrslots == 0 && ch->pm_level > 3 &&
+	    (ch->curr[ch->pm_present ? 15 : 0].caps & CTS_SATA_CAPS_D_PMREQ)) {
 		callout_schedule(&ch->pm_timer,
 		    (ch->pm_level == 4) ? hz / 1000 : hz / 8);
 	}
@@ -2464,6 +2473,8 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 			ch->pm_present = cts->xport_specific.sata.pm_present;
 		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_ATAPI)
 			d->atapi = cts->xport_specific.sata.atapi;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			d->caps = cts->xport_specific.sata.caps;
 		ccb->ccb_h.status = CAM_REQ_CMP;
 		break;
 	}
@@ -2496,9 +2507,24 @@ ahciaction(struct cam_sim *sim, union ccb *ccb)
 				cts->xport_specific.sata.valid |=
 				    CTS_SATA_VALID_REVISION;
 			}
+			cts->xport_specific.sata.caps = d->caps & CTS_SATA_CAPS_D;
+			if (ch->pm_level) {
+				if (ch->caps & (AHCI_CAP_PSC | AHCI_CAP_SSC))
+					cts->xport_specific.sata.caps |= CTS_SATA_CAPS_H_PMREQ;
+				if (ch->caps2 & AHCI_CAP2_APST)
+					cts->xport_specific.sata.caps |= CTS_SATA_CAPS_H_APST;
+			}
+			if ((ch->caps & AHCI_CAP_SNCQ) &&
+			    (ch->quirks & AHCI_Q_NOAA) == 0)
+				cts->xport_specific.sata.caps |= CTS_SATA_CAPS_H_DMAAA;
+			cts->xport_specific.sata.caps &=
+			    ch->user[ccb->ccb_h.target_id].caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
 		} else {
 			cts->xport_specific.sata.revision = d->revision;
 			cts->xport_specific.sata.valid |= CTS_SATA_VALID_REVISION;
+			cts->xport_specific.sata.caps = d->caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
 		}
 		cts->xport_specific.sata.mode = d->mode;
 		cts->xport_specific.sata.valid |= CTS_SATA_VALID_MODE;
diff --git a/sys/dev/ahci/ahci.h b/sys/dev/ahci/ahci.h
index d4c73c9cfe5..1473962eb5b 100644
--- a/sys/dev/ahci/ahci.h
+++ b/sys/dev/ahci/ahci.h
@@ -372,6 +372,7 @@ struct ahci_device {
 	u_int			bytecount;
 	u_int			atapi;
 	u_int			tags;
+	u_int			caps;
 };
 
 /* structure describing an ATA channel */
diff --git a/sys/dev/siis/siis.c b/sys/dev/siis/siis.c
index c60d8e47f69..937d04cfa1d 100644
--- a/sys/dev/siis/siis.c
+++ b/sys/dev/siis/siis.c
@@ -448,6 +448,8 @@ siis_ch_attach(device_t dev)
 		ch->user[i].bytecount = 8192;
 		ch->user[i].tags = SIIS_MAX_SLOTS;
 		ch->curr[i] = ch->user[i];
+		if (ch->pm_level)
+			ch->user[i].caps = CTS_SATA_CAPS_H_PMREQ;
 	}
 	mtx_init(&ch->mtx, "SIIS channel lock", NULL, MTX_DEF);
 	rid = ch->unit;
@@ -1697,6 +1699,8 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 		}
 		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_TAGS)
 			d->atapi = cts->xport_specific.sata.atapi;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			d->caps = cts->xport_specific.sata.caps;
 		ccb->ccb_h.status = CAM_REQ_CMP;
 		break;
 	}
@@ -1729,9 +1733,17 @@ siisaction(struct cam_sim *sim, union ccb *ccb)
 				cts->xport_specific.sata.valid |=
 				    CTS_SATA_VALID_REVISION;
 			}
+			cts->xport_specific.sata.caps = d->caps & CTS_SATA_CAPS_D;
+			if (ch->pm_level)
+				cts->xport_specific.sata.caps |= CTS_SATA_CAPS_H_PMREQ;
+			cts->xport_specific.sata.caps &=
+			    ch->user[ccb->ccb_h.target_id].caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
 		} else {
 			cts->xport_specific.sata.revision = d->revision;
 			cts->xport_specific.sata.valid |= CTS_SATA_VALID_REVISION;
+			cts->xport_specific.sata.caps = d->caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
 		}
 		cts->xport_specific.sata.mode = d->mode;
 		cts->xport_specific.sata.valid |= CTS_SATA_VALID_MODE;
diff --git a/sys/dev/siis/siis.h b/sys/dev/siis/siis.h
index 92b4e26d13a..53176287d31 100644
--- a/sys/dev/siis/siis.h
+++ b/sys/dev/siis/siis.h
@@ -358,6 +358,7 @@ struct siis_device {
 	u_int			bytecount;
 	u_int			atapi;
 	u_int			tags;
+	u_int			caps;
 };
 
 /* structure describing an ATA channel */

From 7e9aef235a685d309e9287d47513511a09fd7da8 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Sun, 2 May 2010 12:08:15 +0000
Subject: [PATCH 395/532] Add a hack for SPARC64 V CPUs, which set some
 undocumented bits in the first data word.

---
 sys/sparc64/sparc64/interrupt.S | 7 +++++--
 1 file changed, 5 insertions(+), 2 deletions(-)

diff --git a/sys/sparc64/sparc64/interrupt.S b/sys/sparc64/sparc64/interrupt.S
index 75a97275301..b71a05da98d 100644
--- a/sys/sparc64/sparc64/interrupt.S
+++ b/sys/sparc64/sparc64/interrupt.S
@@ -83,8 +83,11 @@ ENTRY(intr_vector)
 	 * The 2nd word points to code to execute and the 3rd is an argument
 	 * to pass.  Jump to it.
 	 */
-	brnz,a,pt %g3, 1f
-	 nop
+	brnz,pt %g3, 1f
+	/*
+	 * NB: Zeus CPUs set some undocumented bits in the first data word.
+	 */
+	 and	%g3, IV_MAX - 1, %g3
 	jmpl	%g4, %g0
 	 nop
 	/* NOTREACHED */

From 205367f9bcb541eeca05f663bd2ca612bc6333ea Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Sun, 2 May 2010 14:46:05 +0000
Subject: [PATCH 396/532] Enable PCI busmastering explicitly to be sure.

---
 sys/dev/ahci/ahci.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/dev/ahci/ahci.c b/sys/dev/ahci/ahci.c
index 60fb85153a8..1aa719e044b 100644
--- a/sys/dev/ahci/ahci.c
+++ b/sys/dev/ahci/ahci.c
@@ -340,6 +340,7 @@ ahci_attach(device_t dev)
 		rman_fini(&ctlr->sc_iomem);
 		return (error);
 	}
+	pci_enable_busmaster(dev);
 	/* Reset controller */
 	if ((error = ahci_ctlr_reset(dev)) != 0) {
 		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);

From f6c8c187d4c03431d69b785ed2a71e18821fe570 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 16:44:06 +0000
Subject: [PATCH 397/532] This change addresses the race condition that was
 introduced by the previous revision, r207450, to this file.  Specifically,
 between dropping the page queues lock in vm_contig_launder() and reacquiring
 it in vm_contig_launder_page(), the page may be removed from the active or
 inactive queue.  It could be wired, freed, cached, etc.  None of which
 vm_contig_launder_page() is prepared for.

Reviewed by:	kib, kmacy
---
 sys/vm/vm_contig.c | 30 ++++++++++++++++--------------
 1 file changed, 16 insertions(+), 14 deletions(-)

diff --git a/sys/vm/vm_contig.c b/sys/vm/vm_contig.c
index 2c8b8b11c25..e6a9b78c4a0 100644
--- a/sys/vm/vm_contig.c
+++ b/sys/vm/vm_contig.c
@@ -96,33 +96,33 @@ vm_contig_launder_page(vm_page_t m, vm_page_t *next)
 	vm_page_t m_tmp;
 	struct vnode *vp;
 	struct mount *mp;
-	int vfslocked, dirty;
+	int vfslocked;
 
-	vm_page_lock(m);
-	vm_page_lock_queues();
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	object = m->object;
 	if (!VM_OBJECT_TRYLOCK(object) &&
 	    !vm_pageout_fallback_object_lock(m, next)) {
-		VM_OBJECT_UNLOCK(object);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
+		VM_OBJECT_UNLOCK(object);
 		return (EAGAIN);
 	}
 	if (vm_page_sleep_if_busy(m, TRUE, "vpctw0")) {
 		VM_OBJECT_UNLOCK(object);
+		vm_page_lock_queues();
 		return (EBUSY);
 	}
 	vm_page_test_dirty(m);
 	if (m->dirty == 0 && m->hold_count == 0)
 		pmap_remove_all(m);
-	if ((dirty = m->dirty) != 0) {
-		vm_page_unlock_queues();
+	if (m->dirty != 0) {
 		vm_page_unlock(m);
 		if ((object->flags & OBJ_DEAD) != 0) {
 			VM_OBJECT_UNLOCK(object);
 			return (EAGAIN);
 		}
 		if (object->type == OBJT_VNODE) {
+			vm_page_unlock_queues();
 			vp = object->handle;
 			vm_object_reference_locked(object);
 			VM_OBJECT_UNLOCK(object);
@@ -136,19 +136,20 @@ vm_contig_launder_page(vm_page_t m, vm_page_t *next)
 			VFS_UNLOCK_GIANT(vfslocked);
 			vm_object_deallocate(object);
 			vn_finished_write(mp);
+			vm_page_lock_queues();
 			return (0);
 		} else if (object->type == OBJT_SWAP ||
 			   object->type == OBJT_DEFAULT) {
+			vm_page_unlock_queues();
 			m_tmp = m;
 			vm_pageout_flush(&m_tmp, 1, VM_PAGER_PUT_SYNC);
 			VM_OBJECT_UNLOCK(object);
+			vm_page_lock_queues();
 			return (0);
 		}
-	} else if (m->hold_count == 0)
-		vm_page_cache(m);
-
-	if (dirty == 0) {
-		vm_page_unlock_queues();
+	} else {
+		if (m->hold_count == 0)
+			vm_page_cache(m);
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(object);
@@ -167,11 +168,12 @@ vm_contig_launder(int queue)
 		if ((m->flags & PG_MARKER) != 0)
 			continue;
 
+		if (!vm_page_trylock(m))
+			continue;
 		KASSERT(VM_PAGE_INQUEUE2(m, queue),
 		    ("vm_contig_launder: page %p's queue is not %d", m, queue));
-		vm_page_unlock_queues();
 		error = vm_contig_launder_page(m, &next);
-		vm_page_lock_queues();
+		vm_page_lock_assert(m, MA_NOTOWNED);
 		if (error == 0)
 			return (TRUE);
 		if (error == EBUSY)

From b88b6c9d80ed2e1ad85a6b928e2c9346a41aced6 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 17:33:46 +0000
Subject: [PATCH 398/532] It makes no sense for vm_page_sleep_if_busy()'s
 helper, vm_page_sleep(), to unconditionally set PG_REFERENCED on a page
 before sleeping.  In many cases, it's perfectly ok for the page to disappear,
 i.e., be reclaimed by the page daemon, before the caller to vm_page_sleep()
 is reawakened. Instead, we now explicitly set PG_REFERENCED in those cases
 where having the page persist until the caller is awakened is clearly
 desirable.  Note, however, that setting PG_REFERENCED on the page is still
 only a hint, and not a guarantee that the page should persist.

---
 sys/fs/tmpfs/tmpfs_vnops.c | 30 +++++++++++++++++++++++++++---
 sys/kern/vfs_bio.c         | 11 ++++++++++-
 sys/vm/vm_fault.c          |  6 ++++++
 sys/vm/vm_page.c           | 25 ++++++++++++++++---------
 4 files changed, 59 insertions(+), 13 deletions(-)

diff --git a/sys/fs/tmpfs/tmpfs_vnops.c b/sys/fs/tmpfs/tmpfs_vnops.c
index c17efff9d5f..3d3755d0528 100644
--- a/sys/fs/tmpfs/tmpfs_vnops.c
+++ b/sys/fs/tmpfs/tmpfs_vnops.c
@@ -516,8 +516,16 @@ tmpfs_mappedread(vm_object_t vobj, vm_object_t tobj, size_t len, struct uio *uio
 lookupvpg:
 	if (((m = vm_page_lookup(vobj, idx)) != NULL) &&
 	    vm_page_is_valid(m, offset, tlen)) {
-		if (vm_page_sleep_if_busy(m, FALSE, "tmfsmr"))
+		if ((m->oflags & VPO_BUSY) != 0) {
+			/*
+			 * Reference the page before unlocking and sleeping so
+			 * that the page daemon is less likely to reclaim it.  
+			 */
+			vm_page_lock_queues();
+			vm_page_flag_set(m, PG_REFERENCED);
+			vm_page_sleep(m, "tmfsmr");
 			goto lookupvpg;
+		}
 		vm_page_busy(m);
 		VM_OBJECT_UNLOCK(vobj);
 		error = uiomove_fromphys(&m, offset, tlen, uio);
@@ -526,8 +534,16 @@ lookupvpg:
 		VM_OBJECT_UNLOCK(vobj);
 		return	(error);
 	} else if (m != NULL && uio->uio_segflg == UIO_NOCOPY) {
-		if (vm_page_sleep_if_busy(m, FALSE, "tmfsmr"))
+		if ((m->oflags & VPO_BUSY) != 0) {
+			/*
+			 * Reference the page before unlocking and sleeping so
+			 * that the page daemon is less likely to reclaim it.  
+			 */
+			vm_page_lock_queues();
+			vm_page_flag_set(m, PG_REFERENCED);
+			vm_page_sleep(m, "tmfsmr");
 			goto lookupvpg;
+		}
 		vm_page_busy(m);
 		VM_OBJECT_UNLOCK(vobj);
 		sched_pin();
@@ -627,8 +643,16 @@ tmpfs_mappedwrite(vm_object_t vobj, vm_object_t tobj, size_t len, struct uio *ui
 lookupvpg:
 	if (((vpg = vm_page_lookup(vobj, idx)) != NULL) &&
 	    vm_page_is_valid(vpg, offset, tlen)) {
-		if (vm_page_sleep_if_busy(vpg, FALSE, "tmfsmw"))
+		if ((vpg->oflags & VPO_BUSY) != 0) {
+			/*
+			 * Reference the page before unlocking and sleeping so
+			 * that the page daemon is less likely to reclaim it.  
+			 */
+			vm_page_lock_queues();
+			vm_page_flag_set(vpg, PG_REFERENCED);
+			vm_page_sleep(vpg, "tmfsmw");
 			goto lookupvpg;
+		}
 		vm_page_busy(vpg);
 		vm_page_lock_queues();
 		vm_page_undirty(vpg);
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index ea846a5bf5d..dbcd6e85a83 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -3024,8 +3024,17 @@ allocbuf(struct buf *bp, int size)
 				 *  vm_fault->getpages->cluster_read->allocbuf
 				 *
 				 */
-				if (vm_page_sleep_if_busy(m, FALSE, "pgtblk"))
+				if ((m->oflags & VPO_BUSY) != 0) {
+					/*
+					 * Reference the page before unlocking
+					 * and sleeping so that the page daemon
+					 * is less likely to reclaim it.  
+					 */
+					vm_page_lock_queues();
+					vm_page_flag_set(m, PG_REFERENCED);
+					vm_page_sleep(m, "pgtblk");
 					continue;
+				}
 
 				/*
 				 * We have a good page.
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index d9cf63e9b11..5ee4ab439a0 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -338,6 +338,12 @@ RetryFault:;
 			 * to pmap it.
 			 */
 			if ((fs.m->oflags & VPO_BUSY) || fs.m->busy) {
+				/*
+				 * Reference the page before unlocking and
+				 * sleeping so that the page daemon is less
+				 * likely to reclaim it. 
+				 */
+				vm_page_flag_set(fs.m, PG_REFERENCED);
 				vm_page_unlock_queues();
 				vm_page_unlock(fs.m);
 				VM_OBJECT_UNLOCK(fs.object);
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index fb0950eb9d9..e73586a6007 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -599,7 +599,7 @@ vm_page_free_zero(vm_page_t m)
 /*
  *	vm_page_sleep:
  *
- *	Sleep and release the page queues lock.
+ *	Sleep and release the page and page queues locks.
  *
  *	The object containing the given page must be locked.
  */
@@ -608,13 +608,10 @@ vm_page_sleep(vm_page_t m, const char *msg)
 {
 
 	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
-	if (!mtx_owned(vm_page_lockptr(m)))
-		vm_page_lock(m);
-	if (!mtx_owned(&vm_page_queue_mtx))
-		vm_page_lock_queues();
-	vm_page_flag_set(m, PG_REFERENCED);
-	vm_page_unlock_queues();
-	vm_page_unlock(m);
+	if (mtx_owned(&vm_page_queue_mtx))
+		vm_page_unlock_queues();
+	if (mtx_owned(vm_page_lockptr(m)))
+		vm_page_unlock(m);
 
 	/*
 	 * It's possible that while we sleep, the page will get
@@ -1896,7 +1893,17 @@ vm_page_grab(vm_object_t object, vm_pindex_t pindex, int allocflags)
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 retrylookup:
 	if ((m = vm_page_lookup(object, pindex)) != NULL) {
-		if (vm_page_sleep_if_busy(m, TRUE, "pgrbwt")) {
+		if ((m->oflags & VPO_BUSY) != 0 || m->busy != 0) {
+			if ((allocflags & VM_ALLOC_RETRY) != 0) {
+				/*
+				 * Reference the page before unlocking and
+				 * sleeping so that the page daemon is less
+				 * likely to reclaim it. 
+				 */
+				vm_page_lock_queues();
+				vm_page_flag_set(m, PG_REFERENCED);
+			}
+			vm_page_sleep(m, "pgrbwt");
 			if ((allocflags & VM_ALLOC_RETRY) == 0)
 				return (NULL);
 			goto retrylookup;

From ac800a8490f6c36e0489feadf4e9318111f595dc Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 18:09:33 +0000
Subject: [PATCH 399/532] Correct an error in r207410: Remove an unlock of a
 lock that is no longer held.

---
 sys/vm/vm_object.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 604aa3c9138..b5f246c0bec 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1454,7 +1454,6 @@ retry:
 		 * not be changed by this operation.
 		 */
 		if ((m->oflags & VPO_BUSY) || m->busy) {
-			vm_page_unlock_queues();
 			VM_OBJECT_UNLOCK(new_object);
 			m->oflags |= VPO_WANTED;
 			msleep(m, VM_OBJECT_MTX(orig_object), PVM, "spltwt", 0);

From a9ef85b7922a20c4a5b2ef26d3e8c0aa4814791a Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Sun, 2 May 2010 19:05:57 +0000
Subject: [PATCH 400/532] Remove a soft member which was never used.

Approved by:	marcel
---
 sys/dev/quicc/quicc_bfe.h | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/quicc/quicc_bfe.h b/sys/dev/quicc/quicc_bfe.h
index 9ba622c6794..5c589146ac4 100644
--- a/sys/dev/quicc/quicc_bfe.h
+++ b/sys/dev/quicc/quicc_bfe.h
@@ -50,7 +50,6 @@ struct quicc_softc {
 	u_int		sc_clock;
 
 	int		sc_fastintr:1;
-	int		sc_leaving:1;
 	int		sc_polled:1;
 };
 

From f6ffc3c26bf8738c43c84b0161aa50d6e27232e1 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Sun, 2 May 2010 19:07:19 +0000
Subject: [PATCH 401/532] Remove redundant checking of sc_leaving (uart_intr()
 already handles this).

Approved by:	marcel
---
 sys/dev/uart/uart_dev_ns8250.c      | 2 +-
 sys/mips/cavium/uart_dev_oct16550.c | 5 +----
 2 files changed, 2 insertions(+), 5 deletions(-)

diff --git a/sys/dev/uart/uart_dev_ns8250.c b/sys/dev/uart/uart_dev_ns8250.c
index c01fd615009..b93e254d519 100644
--- a/sys/dev/uart/uart_dev_ns8250.c
+++ b/sys/dev/uart/uart_dev_ns8250.c
@@ -604,7 +604,7 @@ ns8250_bus_ipend(struct uart_softc *sc)
 	if (ipend == 0)
 		ns8250_clrint(bas);
 	uart_unlock(sc->sc_hwmtx);
-	return ((sc->sc_leaving) ? 0 : ipend);
+	return (ipend);
 }
 
 static int
diff --git a/sys/mips/cavium/uart_dev_oct16550.c b/sys/mips/cavium/uart_dev_oct16550.c
index 3e6d7c7cce7..5c6e11ee59c 100644
--- a/sys/mips/cavium/uart_dev_oct16550.c
+++ b/sys/mips/cavium/uart_dev_oct16550.c
@@ -644,12 +644,9 @@ oct16550_bus_ipend(struct uart_softc *sc)
         if (ipend)	octeon_led_run_wheel(&where1, 6 + device_get_unit(sc->sc_dev));
 #endif
 
-	return ((sc->sc_leaving) ? 0 : ipend);
+	return (ipend);
 }
 
-
-
-
 static int
 oct16550_bus_param (struct uart_softc *sc, int baudrate, int databits,
     int stopbits, int parity)

From 139a0de7f1391b9fb7c2ca99afc50c1323ec9f77 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 19:10:27 +0000
Subject: [PATCH 402/532] Properly synchronize access to the page's hold_count
 in vfs_vmio_release().

Reviewed by:	kib
---
 sys/kern/vfs_bio.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index dbcd6e85a83..10489985388 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -1563,7 +1563,6 @@ vfs_vmio_release(struct buf *bp)
 	vm_page_t m;
 
 	VM_OBJECT_LOCK(bp->b_bufobj->bo_object);
-	vm_page_lock_queues();
 	for (i = 0; i < bp->b_npages; i++) {
 		m = bp->b_pages[i];
 		bp->b_pages[i] = NULL;
@@ -1571,16 +1570,16 @@ vfs_vmio_release(struct buf *bp)
 		 * In order to keep page LRU ordering consistent, put
 		 * everything on the inactive queue.
 		 */
+		vm_page_lock(m);
+		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		/*
 		 * We don't mess with busy pages, it is
 		 * the responsibility of the process that
 		 * busied the pages to deal with them.
 		 */
-		if ((m->oflags & VPO_BUSY) || (m->busy != 0))
-			continue;
-			
-		if (m->wire_count == 0) {
+		if ((m->oflags & VPO_BUSY) == 0 && m->busy == 0 &&
+		    m->wire_count == 0) {
 			/*
 			 * Might as well free the page if we can and it has
 			 * no valid data.  We also free the page if the
@@ -1595,8 +1594,9 @@ vfs_vmio_release(struct buf *bp)
 				vm_page_try_to_cache(m);
 			}
 		}
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(bp->b_bufobj->bo_object);
 	pmap_qremove(trunc_page((vm_offset_t) bp->b_data), bp->b_npages);
 	

From a0b8e597e5b0575c09efb04177e21842f907e42e Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Sun, 2 May 2010 19:25:22 +0000
Subject: [PATCH 403/532] Lock the page around hold_count access.

Reviewed by:	alc
---
 sys/kern/uipc_syscalls.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/sys/kern/uipc_syscalls.c b/sys/kern/uipc_syscalls.c
index a14be72064e..3271235a6f9 100644
--- a/sys/kern/uipc_syscalls.c
+++ b/sys/kern/uipc_syscalls.c
@@ -2108,6 +2108,7 @@ retry_space:
 				mbstat.sf_iocnt++;
 			}
 			if (error) {
+				vm_page_lock(pg);
 				vm_page_lock_queues();
 				vm_page_unwire(pg, 0);
 				/*
@@ -2121,6 +2122,7 @@ retry_space:
 					vm_page_free(pg);
 				}
 				vm_page_unlock_queues();
+				vm_page_unlock(pg);
 				VM_OBJECT_UNLOCK(obj);
 				if (error == EAGAIN)
 					error = 0;	/* not a real error */

From dd48af360fdbbb9552f9fc6de7abe50d68ad5331 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Sun, 2 May 2010 19:28:30 +0000
Subject: [PATCH 404/532] Import mvs(4) - Marvell
 88SX50XX/88SX60XX/88SX70XX/SoC SATA controllers driver for CAM ATA subsystem.
 This driver supports same hardware as atamarvell, ataadaptec and atamvsata
 drivers from ata(4), but provides many additional features, such as NCQ, PMP,
 etc.

---
 share/man/man4/Makefile  |    1 +
 share/man/man4/mvs.4     |  176 +++
 sys/arm/mv/files.mv      |    1 +
 sys/conf/NOTES           |    2 +
 sys/conf/files           |    3 +
 sys/conf/kmod.mk         |    2 +-
 sys/dev/mvs/mvs.c        | 2173 ++++++++++++++++++++++++++++++++++++++
 sys/dev/mvs/mvs.h        |  650 ++++++++++++
 sys/dev/mvs/mvs_if.m     |   34 +
 sys/dev/mvs/mvs_pci.c    |  507 +++++++++
 sys/dev/mvs/mvs_soc.c    |  437 ++++++++
 sys/modules/Makefile     |    1 +
 sys/modules/mvs/Makefile |   10 +
 13 files changed, 3996 insertions(+), 1 deletion(-)
 create mode 100644 share/man/man4/mvs.4
 create mode 100644 sys/dev/mvs/mvs.c
 create mode 100644 sys/dev/mvs/mvs.h
 create mode 100644 sys/dev/mvs/mvs_if.m
 create mode 100644 sys/dev/mvs/mvs_pci.c
 create mode 100644 sys/dev/mvs/mvs_soc.c
 create mode 100644 sys/modules/mvs/Makefile

diff --git a/share/man/man4/Makefile b/share/man/man4/Makefile
index c71a09740e7..b1282daf9bb 100644
--- a/share/man/man4/Makefile
+++ b/share/man/man4/Makefile
@@ -220,6 +220,7 @@ MAN=	aac.4 \
 	msk.4 \
 	mtio.4 \
 	multicast.4 \
+	mvs.4 \
 	mwl.4 \
 	mwlfw.4 \
 	mxge.4 \
diff --git a/share/man/man4/mvs.4 b/share/man/man4/mvs.4
new file mode 100644
index 00000000000..599933500e2
--- /dev/null
+++ b/share/man/man4/mvs.4
@@ -0,0 +1,176 @@
+.\" Copyright (c) 2009 Alexander Motin <mav@FreeBSD.org>
+.\" All rights reserved.
+.\"
+.\" Redistribution and use in source and binary forms, with or without
+.\" modification, are permitted provided that the following conditions
+.\" are met:
+.\" 1. Redistributions of source code must retain the above copyright
+.\"    notice, this list of conditions and the following disclaimer.
+.\" 2. Redistributions in binary form must reproduce the above copyright
+.\"    notice, this list of conditions and the following disclaimer in the
+.\"    documentation and/or other materials provided with the distribution.
+.\" 3. The name of the author may not be used to endorse or promote products
+.\"    derived from this software without specific prior written permission.
+.\"
+.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+.\" IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+.\" OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+.\" IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+.\" INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+.\" NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+.\" DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+.\" THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+.\" (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+.\" THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+.\"
+.\" $FreeBSD$
+.\"
+.Dd April 27, 2010
+.Dt MVS 4
+.Os
+.Sh NAME
+.Nm mvs
+.Nd Marvell Serial ATA Host Controller driver
+.Sh SYNOPSIS
+To compile this driver into the kernel,
+place the following lines in your
+kernel configuration file:
+.Bd -ragged -offset indent
+.Cd "device pci"
+.Cd "device scbus"
+.Cd "device mvs"
+.Ed
+.Pp
+Alternatively, to load the driver as a
+module at boot time, place the following line in
+.Xr loader.conf 5 :
+.Bd -literal -offset indent
+mvs_load="YES"
+.Ed
+.Pp
+The following tunables are settable from the
+.Xr loader 8 :
+.Bl -ohang
+.It Va hint.mvs. Ns Ar X Ns Va .msi
+controls Message Signaled Interrupts (MSI) usage by the specified controller.
+.It Va hint.mvs. Ns Ar X Ns Va .ccc
+controls Command Completion Coalescing (CCC) usage by the specified controller.
+Non-zero value enables CCC and defines maximum time (in us), request can wait
+for interrupt.
+CCC reduces number of context switches on systems with many parallel requests,
+but it can decrease disk performance on some workloads due to additional
+command latency.
+.It Va hint.mvs. Ns Ar X Ns Va .cccc
+defines number of completed commands for CCC, which trigger interrupt without
+waiting for specified coalescing timeout.
+.It Va hint.mvs. Ns Ar X Ns Va .pm_level
+controls SATA interface Power Management for the specified channel,
+allowing some power to be saved at the cost of additional command
+latency.
+Possible values:
+.Bl -tag -compact
+.It 0
+interface Power Management is disabled (default);
+.It 1
+device is allowed to initiate PM state change, host is passive;
+.It 4
+driver initiates PARTIAL PM state transition 1ms after port becomes idle;
+.It 5
+driver initiates SLUMBER PM state transition 125ms after port becomes idle.
+.El
+.Pp
+Note that interface Power Management is not compatible with
+device presence detection.
+A manual bus reset is needed on device hot-plug.
+.It Va hint.mvs. Ns Ar X Ns Va .sata_rev
+setting to nonzero value limits maximum SATA revision (speed).
+Values 1, 2 and 3 are respectively 1.5, 3 and 6Gbps.
+.El
+.Sh DESCRIPTION
+This driver provides the
+.Xr CAM 4
+subsystem with native access to the
+.Tn SATA
+ports of several generations (Gen-I/II/IIe) of Marvell SATA controllers.
+Each SATA port found is represented to CAM as a separate bus with one
+target, or, if HBA supports Port Multipliers (Gen-II/IIe), 16 targets.
+Most of the bus-management details are handled by the SATA-specific
+transport of CAM.
+Connected ATA disks are handled by the ATA protocol disk peripheral driver
+.Xr ada 4 .
+ATAPI devices are handled by the SCSI protocol peripheral drivers
+.Xr cd 4 ,
+.Xr da 4 ,
+.Xr sa 4 ,
+etc.
+.Pp
+Driver features include support for Serial ATA and ATAPI devices,
+Port Multipliers (including FIS-based switching, when supported),
+hardware command queues (up to 31 command per port),
+Native Command Queuing, SATA interface Power Management, device hot-plug
+and Message Signaled Interrupts.
+.Pp
+Same hardware is also supported by atamarvell and ataadaptec drivers from
+.Xr ata 4
+subsystem.
+If both drivers are loaded at the same time, this one will be
+given precedence as the more functional of the two.
+.Sh HARDWARE
+The
+.Nm
+driver supports the following controllers:
+.Bl -tag -compact
+.It Gen-I (SATA 1.5Gbps):
+.Bl -bullet -compact
+.It
+88SX5040
+.It
+88SX5041
+.It
+88SX5080
+.It
+88SX5081
+.El
+.It Gen-II (SATA 3Gbps, NCQ, PMP):
+.Bl -bullet -compact
+.It
+88SX6040
+.It
+88SX6041 (including Adaptec 1420SA)
+.It
+88SX6080
+.It
+88SX6081
+.El
+.It Gen-IIe (SATA 3Gbps, NCQ, PMP with FBS):
+.Bl -bullet -compact
+.It
+88SX6042
+.It
+88SX7042 (including Adaptec 1430SA)
+.It
+88F5182 SoC
+.It
+88F6281 SoC
+.It
+MV78100 SoC
+.El
+.El
+Note, that this hardware supports command queueing and FIS-based switching
+only for ATA DMA commands. ATAPI and non-DMA ATA commands executed one by one
+for each port.
+.Pp
+.Sh SEE ALSO
+.Xr ada 4 ,
+.Xr ata 4 ,
+.Xr cam 4 ,
+.Xr cd 4 ,
+.Xr da 4 ,
+.Xr sa 4
+.Sh HISTORY
+The
+.Nm
+driver first appeared in
+.Fx 9.0 .
+.Sh AUTHORS
+.An Alexander Motin Aq mav@FreeBSD.org .
diff --git a/sys/arm/mv/files.mv b/sys/arm/mv/files.mv
index 36be745f147..d7352b79b14 100644
--- a/sys/arm/mv/files.mv
+++ b/sys/arm/mv/files.mv
@@ -30,6 +30,7 @@ arm/mv/timer.c			standard
 arm/mv/twsi.c			optional	iicbus
 
 dev/mge/if_mge.c		optional	mge
+dev/mvs/mvs_soc.c		optional	mvs
 dev/uart/uart_bus_mbus.c	optional	uart
 dev/uart/uart_cpu_mv.c		optional	uart
 dev/uart/uart_dev_ns8250.c	optional	uart
diff --git a/sys/conf/NOTES b/sys/conf/NOTES
index cc1964d0458..1ec82a9c83c 100644
--- a/sys/conf/NOTES
+++ b/sys/conf/NOTES
@@ -1660,12 +1660,14 @@ device		twe		# 3ware ATA RAID
 # Serial ATA host controllers:
 #
 # ahci: Advanced Host Controller Interface (AHCI) compatible
+# mvs:  Marvell 88SX50XX/88SX60XX/88SX70XX/SoC controllers
 # siis: SiliconImage SiI3124/SiI3132/SiI3531 controllers
 #
 # These drivers are part of cam(4) subsystem. They supersede less featured
 # ata(4) subsystem drivers, supporting same hardware.
 
 device		ahci
+device		mvs
 device		siis
 
 #
diff --git a/sys/conf/files b/sys/conf/files
index f4bb0621102..a485f2f0796 100644
--- a/sys/conf/files
+++ b/sys/conf/files
@@ -1293,6 +1293,9 @@ dev/mpt/mpt_pci.c		optional mpt pci
 dev/mpt/mpt_raid.c		optional mpt
 dev/mpt/mpt_user.c		optional mpt
 dev/msk/if_msk.c		optional msk inet
+dev/mvs/mvs.c			optional mvs
+dev/mvs/mvs_if.m		standard
+dev/mvs/mvs_pci.c		optional mvs pci
 dev/mwl/if_mwl.c		optional mwl
 dev/mwl/if_mwl_pci.c		optional mwl pci
 dev/mwl/mwlhal.c		optional mwl
diff --git a/sys/conf/kmod.mk b/sys/conf/kmod.mk
index 18bfbb12b3b..c6413d7c425 100644
--- a/sys/conf/kmod.mk
+++ b/sys/conf/kmod.mk
@@ -342,7 +342,7 @@ MFILES?= dev/acpica/acpi_if.m dev/acpi_support/acpi_wmi_if.m \
 	dev/agp/agp_if.m dev/ata/ata_if.m dev/eisa/eisa_if.m \
 	dev/iicbus/iicbb_if.m dev/iicbus/iicbus_if.m \
 	dev/mmc/mmcbr_if.m dev/mmc/mmcbus_if.m \
-	dev/mii/miibus_if.m dev/ofw/ofw_bus_if.m \
+	dev/mii/miibus_if.m dev/mvs/mvs_if.m dev/ofw/ofw_bus_if.m \
 	dev/pccard/card_if.m dev/pccard/power_if.m dev/pci/pci_if.m \
 	dev/pci/pcib_if.m dev/ppbus/ppbus_if.m dev/smbus/smbus_if.m \
 	dev/sound/pcm/ac97_if.m dev/sound/pcm/channel_if.m \
diff --git a/sys/dev/mvs/mvs.c b/sys/dev/mvs/mvs.c
new file mode 100644
index 00000000000..038e9f21072
--- /dev/null
+++ b/sys/dev/mvs/mvs.c
@@ -0,0 +1,2173 @@
+/*-
+ * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer,
+ *    without modification, immediately at the beginning of the file.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/module.h>
+#include <sys/systm.h>
+#include <sys/kernel.h>
+#include <sys/ata.h>
+#include <sys/bus.h>
+#include <sys/endian.h>
+#include <sys/malloc.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <vm/uma.h>
+#include <machine/stdarg.h>
+#include <machine/resource.h>
+#include <machine/bus.h>
+#include <sys/rman.h>
+#include "mvs.h"
+
+#include <cam/cam.h>
+#include <cam/cam_ccb.h>
+#include <cam/cam_sim.h>
+#include <cam/cam_xpt_sim.h>
+#include <cam/cam_debug.h>
+
+/* local prototypes */
+static int mvs_ch_suspend(device_t dev);
+static int mvs_ch_resume(device_t dev);
+static void mvs_dmainit(device_t dev);
+static void mvs_dmasetupc_cb(void *xsc, bus_dma_segment_t *segs, int nsegs, int error);
+static void mvs_dmafini(device_t dev);
+static void mvs_slotsalloc(device_t dev);
+static void mvs_slotsfree(device_t dev);
+static void mvs_setup_edma_queues(device_t dev);
+static void mvs_set_edma_mode(device_t dev, enum mvs_edma_mode mode);
+static void mvs_ch_pm(void *arg);
+static void mvs_ch_intr_locked(void *data);
+static void mvs_ch_intr(void *data);
+static void mvs_reset(device_t dev);
+static void mvs_softreset(device_t dev, union ccb *ccb);
+
+static int mvs_sata_connect(struct mvs_channel *ch);
+static int mvs_sata_phy_reset(device_t dev);
+static int mvs_wait(device_t dev, u_int s, u_int c, int t);
+static void mvs_tfd_read(device_t dev, union ccb *ccb);
+static void mvs_tfd_write(device_t dev, union ccb *ccb);
+static void mvs_legacy_intr(device_t dev);
+static void mvs_crbq_intr(device_t dev);
+static void mvs_begin_transaction(device_t dev, union ccb *ccb);
+static void mvs_legacy_execute_transaction(struct mvs_slot *slot);
+static void mvs_timeout(struct mvs_slot *slot);
+static void mvs_dmasetprd(void *arg, bus_dma_segment_t *segs, int nsegs, int error);
+static void mvs_requeue_frozen(device_t dev);
+static void mvs_execute_transaction(struct mvs_slot *slot);
+static void mvs_end_transaction(struct mvs_slot *slot, enum mvs_err_type et);
+
+static void mvs_issue_read_log(device_t dev);
+static void mvs_process_read_log(device_t dev, union ccb *ccb);
+
+static void mvsaction(struct cam_sim *sim, union ccb *ccb);
+static void mvspoll(struct cam_sim *sim);
+
+MALLOC_DEFINE(M_MVS, "MVS driver", "MVS driver data buffers");
+
+static int
+mvs_ch_probe(device_t dev)
+{
+
+	device_set_desc_copy(dev, "Marvell SATA channel");
+	return (0);
+}
+
+static int
+mvs_ch_attach(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(device_get_parent(dev));
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct cam_devq *devq;
+	int rid, error, i, sata_rev = 0;
+
+	ch->dev = dev;
+	ch->unit = (intptr_t)device_get_ivars(dev);
+	ch->quirks = ctlr->quirks;
+	mtx_init(&ch->mtx, "MVS channel lock", NULL, MTX_DEF);
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "pm_level", &ch->pm_level);
+	if (ch->pm_level > 3)
+		callout_init_mtx(&ch->pm_timer, &ch->mtx, 0);
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "sata_rev", &sata_rev);
+	for (i = 0; i < 16; i++) {
+		ch->user[i].revision = sata_rev;
+		ch->user[i].mode = 0;
+		ch->user[i].bytecount = (ch->quirks & MVS_Q_GENIIE) ? 8192 : 2048;
+		ch->user[i].tags = MVS_MAX_SLOTS;
+		ch->curr[i] = ch->user[i];
+		if (ch->pm_level) {
+			ch->user[i].caps = CTS_SATA_CAPS_H_PMREQ |
+			    CTS_SATA_CAPS_H_APST |
+			    CTS_SATA_CAPS_D_PMREQ | CTS_SATA_CAPS_D_APST;
+		}
+	}
+	rid = ch->unit;
+	if (!(ch->r_mem = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
+	    &rid, RF_ACTIVE)))
+		return (ENXIO);
+	mvs_dmainit(dev);
+	mvs_slotsalloc(dev);
+	mvs_ch_resume(dev);
+	mtx_lock(&ch->mtx);
+	rid = ATA_IRQ_RID;
+	if (!(ch->r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ,
+	    &rid, RF_SHAREABLE | RF_ACTIVE))) {
+		device_printf(dev, "Unable to map interrupt\n");
+		error = ENXIO;
+		goto err0;
+	}
+	if ((bus_setup_intr(dev, ch->r_irq, ATA_INTR_FLAGS, NULL,
+	    mvs_ch_intr_locked, dev, &ch->ih))) {
+		device_printf(dev, "Unable to setup interrupt\n");
+		error = ENXIO;
+		goto err1;
+	}
+	/* Create the device queue for our SIM. */
+	devq = cam_simq_alloc(MVS_MAX_SLOTS - 1);
+	if (devq == NULL) {
+		device_printf(dev, "Unable to allocate simq\n");
+		error = ENOMEM;
+		goto err1;
+	}
+	/* Construct SIM entry */
+	ch->sim = cam_sim_alloc(mvsaction, mvspoll, "mvsch", ch,
+	    device_get_unit(dev), &ch->mtx,
+	    2, (ch->quirks & MVS_Q_GENI) ? 0 : MVS_MAX_SLOTS - 1,
+	    devq);
+	if (ch->sim == NULL) {
+		cam_simq_free(devq);
+		device_printf(dev, "unable to allocate sim\n");
+		error = ENOMEM;
+		goto err1;
+	}
+	if (xpt_bus_register(ch->sim, dev, 0) != CAM_SUCCESS) {
+		device_printf(dev, "unable to register xpt bus\n");
+		error = ENXIO;
+		goto err2;
+	}
+	if (xpt_create_path(&ch->path, /*periph*/NULL, cam_sim_path(ch->sim),
+	    CAM_TARGET_WILDCARD, CAM_LUN_WILDCARD) != CAM_REQ_CMP) {
+		device_printf(dev, "unable to create path\n");
+		error = ENXIO;
+		goto err3;
+	}
+	if (ch->pm_level > 3) {
+		callout_reset(&ch->pm_timer,
+		    (ch->pm_level == 4) ? hz / 1000 : hz / 8,
+		    mvs_ch_pm, dev);
+	}
+	mtx_unlock(&ch->mtx);
+	return (0);
+
+err3:
+	xpt_bus_deregister(cam_sim_path(ch->sim));
+err2:
+	cam_sim_free(ch->sim, /*free_devq*/TRUE);
+err1:
+	bus_release_resource(dev, SYS_RES_IRQ, ATA_IRQ_RID, ch->r_irq);
+err0:
+	bus_release_resource(dev, SYS_RES_MEMORY, ch->unit, ch->r_mem);
+	mtx_unlock(&ch->mtx);
+	mtx_destroy(&ch->mtx);
+	return (error);
+}
+
+static int
+mvs_ch_detach(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	mtx_lock(&ch->mtx);
+	xpt_async(AC_LOST_DEVICE, ch->path, NULL);
+	xpt_free_path(ch->path);
+	xpt_bus_deregister(cam_sim_path(ch->sim));
+	cam_sim_free(ch->sim, /*free_devq*/TRUE);
+	mtx_unlock(&ch->mtx);
+
+	if (ch->pm_level > 3)
+		callout_drain(&ch->pm_timer);
+	bus_teardown_intr(dev, ch->r_irq, ch->ih);
+	bus_release_resource(dev, SYS_RES_IRQ, ATA_IRQ_RID, ch->r_irq);
+
+	mvs_ch_suspend(dev);
+	mvs_slotsfree(dev);
+	mvs_dmafini(dev);
+
+	bus_release_resource(dev, SYS_RES_MEMORY, ch->unit, ch->r_mem);
+	mtx_destroy(&ch->mtx);
+	return (0);
+}
+
+static int
+mvs_ch_suspend(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	/* Stop EDMA */
+	mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+	/* Disable port interrupts. */
+	ATA_OUTL(ch->r_mem, EDMA_IEM, 0);
+	return (0);
+}
+
+static int
+mvs_ch_resume(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint32_t reg;
+
+	/* Disable port interrupts */
+	ATA_OUTL(ch->r_mem, EDMA_IEM, 0);
+	/* Stop EDMA */
+	ch->curr_mode = MVS_EDMA_UNKNOWN;
+	mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+	/* Clear and configure FIS interrupts. */
+	ATA_OUTL(ch->r_mem, SATA_FISIC, 0);
+	reg = ATA_INL(ch->r_mem, SATA_FISC);
+	reg |= SATA_FISC_FISWAIT4HOSTRDYEN_B1;
+	ATA_OUTL(ch->r_mem, SATA_FISC, reg);
+	reg = ATA_INL(ch->r_mem, SATA_FISIM);
+	reg |= SATA_FISC_FISWAIT4HOSTRDYEN_B1;
+	ATA_OUTL(ch->r_mem, SATA_FISC, reg);
+	/* Clear SATA error register. */
+	ATA_OUTL(ch->r_mem, SATA_SE, 0xffffffff);
+	/* Clear any outstanding error interrupts. */
+	ATA_OUTL(ch->r_mem, EDMA_IEC, 0);
+	/* Unmask all error interrupts */
+	ATA_OUTL(ch->r_mem, EDMA_IEM, ~EDMA_IE_TRANSIENT);
+	return (0);
+}
+
+struct mvs_dc_cb_args {
+	bus_addr_t maddr;
+	int error;
+};
+
+static void
+mvs_dmainit(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct mvs_dc_cb_args dcba;
+
+	/* EDMA command request area. */
+	if (bus_dma_tag_create(bus_get_dma_tag(dev), 1024, 0,
+	    BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR,
+	    NULL, NULL, MVS_WORKRQ_SIZE, 1, MVS_WORKRQ_SIZE,
+	    0, NULL, NULL, &ch->dma.workrq_tag))
+		goto error;
+	if (bus_dmamem_alloc(ch->dma.workrq_tag, (void **)&ch->dma.workrq, 0,
+	    &ch->dma.workrq_map))
+		goto error;
+	if (bus_dmamap_load(ch->dma.workrq_tag, ch->dma.workrq_map, ch->dma.workrq,
+	    MVS_WORKRQ_SIZE, mvs_dmasetupc_cb, &dcba, 0) || dcba.error) {
+		bus_dmamem_free(ch->dma.workrq_tag, ch->dma.workrq, ch->dma.workrq_map);
+		goto error;
+	}
+	ch->dma.workrq_bus = dcba.maddr;
+	/* EDMA command response area. */
+	if (bus_dma_tag_create(bus_get_dma_tag(dev), 256, 0,
+	    BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR,
+	    NULL, NULL, MVS_WORKRP_SIZE, 1, MVS_WORKRP_SIZE,
+	    0, NULL, NULL, &ch->dma.workrp_tag))
+		goto error;
+	if (bus_dmamem_alloc(ch->dma.workrp_tag, (void **)&ch->dma.workrp, 0,
+	    &ch->dma.workrp_map))
+		goto error;
+	if (bus_dmamap_load(ch->dma.workrp_tag, ch->dma.workrp_map, ch->dma.workrp,
+	    MVS_WORKRP_SIZE, mvs_dmasetupc_cb, &dcba, 0) || dcba.error) {
+		bus_dmamem_free(ch->dma.workrp_tag, ch->dma.workrp, ch->dma.workrp_map);
+		goto error;
+	}
+	ch->dma.workrp_bus = dcba.maddr;
+	/* Data area. */
+	if (bus_dma_tag_create(bus_get_dma_tag(dev), 2, MVS_EPRD_MAX,
+	    BUS_SPACE_MAXADDR, BUS_SPACE_MAXADDR,
+	    NULL, NULL,
+	    MVS_SG_ENTRIES * PAGE_SIZE * MVS_MAX_SLOTS,
+	    MVS_SG_ENTRIES, MVS_EPRD_MAX,
+	    0, busdma_lock_mutex, &ch->mtx, &ch->dma.data_tag)) {
+		goto error;
+	}
+	return;
+
+error:
+	device_printf(dev, "WARNING - DMA initialization failed\n");
+	mvs_dmafini(dev);
+}
+
+static void
+mvs_dmasetupc_cb(void *xsc, bus_dma_segment_t *segs, int nsegs, int error)
+{
+	struct mvs_dc_cb_args *dcba = (struct mvs_dc_cb_args *)xsc;
+
+	if (!(dcba->error = error))
+		dcba->maddr = segs[0].ds_addr;
+}
+
+static void
+mvs_dmafini(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	if (ch->dma.data_tag) {
+		bus_dma_tag_destroy(ch->dma.data_tag);
+		ch->dma.data_tag = NULL;
+	}
+	if (ch->dma.workrp_bus) {
+		bus_dmamap_unload(ch->dma.workrp_tag, ch->dma.workrp_map);
+		bus_dmamem_free(ch->dma.workrp_tag, ch->dma.workrp, ch->dma.workrp_map);
+		ch->dma.workrp_bus = 0;
+		ch->dma.workrp_map = NULL;
+		ch->dma.workrp = NULL;
+	}
+	if (ch->dma.workrp_tag) {
+		bus_dma_tag_destroy(ch->dma.workrp_tag);
+		ch->dma.workrp_tag = NULL;
+	}
+	if (ch->dma.workrq_bus) {
+		bus_dmamap_unload(ch->dma.workrq_tag, ch->dma.workrq_map);
+		bus_dmamem_free(ch->dma.workrq_tag, ch->dma.workrq, ch->dma.workrq_map);
+		ch->dma.workrq_bus = 0;
+		ch->dma.workrq_map = NULL;
+		ch->dma.workrq = NULL;
+	}
+	if (ch->dma.workrq_tag) {
+		bus_dma_tag_destroy(ch->dma.workrq_tag);
+		ch->dma.workrq_tag = NULL;
+	}
+}
+
+static void
+mvs_slotsalloc(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int i;
+
+	/* Alloc and setup command/dma slots */
+	bzero(ch->slot, sizeof(ch->slot));
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		struct mvs_slot *slot = &ch->slot[i];
+
+		slot->dev = dev;
+		slot->slot = i;
+		slot->state = MVS_SLOT_EMPTY;
+		slot->ccb = NULL;
+		callout_init_mtx(&slot->timeout, &ch->mtx, 0);
+
+		if (bus_dmamap_create(ch->dma.data_tag, 0, &slot->dma.data_map))
+			device_printf(ch->dev, "FAILURE - create data_map\n");
+	}
+}
+
+static void
+mvs_slotsfree(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int i;
+
+	/* Free all dma slots */
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		struct mvs_slot *slot = &ch->slot[i];
+
+		callout_drain(&slot->timeout);
+		if (slot->dma.data_map) {
+			bus_dmamap_destroy(ch->dma.data_tag, slot->dma.data_map);
+			slot->dma.data_map = NULL;
+		}
+	}
+}
+
+static void
+mvs_setup_edma_queues(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint64_t work;
+
+	/* Requests queue. */
+	work = ch->dma.workrq_bus;
+	ATA_OUTL(ch->r_mem, EDMA_REQQBAH, work >> 32);
+	ATA_OUTL(ch->r_mem, EDMA_REQQIP, work & 0xffffffff);
+	ATA_OUTL(ch->r_mem, EDMA_REQQOP, work & 0xffffffff);
+	bus_dmamap_sync(ch->dma.workrq_tag, ch->dma.workrq_map, BUS_DMASYNC_PREWRITE);
+	/* Reponses queue. */
+	bzero(ch->dma.workrp, 256);
+	work = ch->dma.workrp_bus;
+	ATA_OUTL(ch->r_mem, EDMA_RESQBAH, work >> 32);
+	ATA_OUTL(ch->r_mem, EDMA_RESQIP, work & 0xffffffff);
+	ATA_OUTL(ch->r_mem, EDMA_RESQOP, work & 0xffffffff);
+	bus_dmamap_sync(ch->dma.workrp_tag, ch->dma.workrp_map, BUS_DMASYNC_PREREAD);
+	ch->out_idx = 0;
+	ch->in_idx = 0;
+}
+
+static void
+mvs_set_edma_mode(device_t dev, enum mvs_edma_mode mode)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int timeout;
+	uint32_t ecfg, fcfg, hc, ltm, unkn;
+
+	if (mode == ch->curr_mode)
+		return;
+	/* If we are running, we should stop first. */
+	if (ch->curr_mode != MVS_EDMA_OFF) {
+		ATA_OUTL(ch->r_mem, EDMA_CMD, EDMA_CMD_EDSEDMA);
+		timeout = 0;
+		while (ATA_INL(ch->r_mem, EDMA_CMD) & EDMA_CMD_EENEDMA) {
+			DELAY(1000);
+			if (timeout++ > 1000) {
+				device_printf(dev, "stopping EDMA engine failed\n");
+				break;
+			}
+		};
+	}
+	ch->curr_mode = mode;
+	ch->fbs_enabled = 0;
+	ch->fake_busy = 0;
+	/* Report mode to controller. Needed for correct CCC operation. */
+	MVS_EDMA(device_get_parent(dev), dev, mode);
+	/* Configure new mode. */
+	ecfg = EDMA_CFG_RESERVED | EDMA_CFG_RESERVED2 | EDMA_CFG_EHOSTQUEUECACHEEN;
+	if (ch->pm_present) {
+		ecfg |= EDMA_CFG_EMASKRXPM;
+		if (ch->quirks & MVS_Q_GENIIE) {
+			ecfg |= EDMA_CFG_EEDMAFBS;
+			ch->fbs_enabled = 1;
+		}
+	}
+	if (ch->quirks & MVS_Q_GENI)
+		ecfg |= EDMA_CFG_ERDBSZ;
+	else if (ch->quirks & MVS_Q_GENII)
+		ecfg |= EDMA_CFG_ERDBSZEXT | EDMA_CFG_EWRBUFFERLEN;
+	if (ch->quirks & MVS_Q_CT)
+		ecfg |= EDMA_CFG_ECUTTHROUGHEN;
+	if (mode != MVS_EDMA_OFF)
+		ecfg |= EDMA_CFG_EEARLYCOMPLETIONEN;
+	if (mode == MVS_EDMA_QUEUED)
+		ecfg |= EDMA_CFG_EQUE;
+	else if (mode == MVS_EDMA_NCQ)
+		ecfg |= EDMA_CFG_ESATANATVCMDQUE;
+	ATA_OUTL(ch->r_mem, EDMA_CFG, ecfg);
+	mvs_setup_edma_queues(dev);
+	if (ch->quirks & MVS_Q_GENIIE) {
+		/* Configure FBS-related registers */
+		fcfg = ATA_INL(ch->r_mem, SATA_FISC);
+		ltm = ATA_INL(ch->r_mem, SATA_LTM);
+		hc = ATA_INL(ch->r_mem, EDMA_HC);
+		if (ch->fbs_enabled) {
+			fcfg |= SATA_FISC_FISDMAACTIVATESYNCRESP;
+			if (mode == MVS_EDMA_NCQ) {
+				fcfg &= ~SATA_FISC_FISWAIT4HOSTRDYEN_B0;
+				hc &= ~EDMA_IE_EDEVERR;
+			} else {
+				fcfg |= SATA_FISC_FISWAIT4HOSTRDYEN_B0;
+				hc |= EDMA_IE_EDEVERR;
+			}
+			ltm |= (1 << 8);
+		} else {
+			fcfg &= ~SATA_FISC_FISDMAACTIVATESYNCRESP;
+			fcfg &= ~SATA_FISC_FISWAIT4HOSTRDYEN_B0;
+			hc |= EDMA_IE_EDEVERR;
+			ltm &= ~(1 << 8);
+		}
+		ATA_OUTL(ch->r_mem, SATA_FISC, fcfg);
+		ATA_OUTL(ch->r_mem, SATA_LTM, ltm);
+		ATA_OUTL(ch->r_mem, EDMA_HC, hc);
+		/* This is some magic, required to handle several DRQs
+		 * with basic DMA. */
+		unkn = ATA_INL(ch->r_mem, EDMA_UNKN_RESD);
+		if (mode == MVS_EDMA_OFF)
+			unkn |= 1;
+		else
+			unkn &= ~1;
+		ATA_OUTL(ch->r_mem, EDMA_UNKN_RESD, unkn);
+	}
+	/* Run EDMA. */
+	if (mode != MVS_EDMA_OFF)
+		ATA_OUTL(ch->r_mem, EDMA_CMD, EDMA_CMD_EENEDMA);
+}
+
+devclass_t mvs_devclass;
+devclass_t mvsch_devclass;
+static device_method_t mvsch_methods[] = {
+	DEVMETHOD(device_probe,     mvs_ch_probe),
+	DEVMETHOD(device_attach,    mvs_ch_attach),
+	DEVMETHOD(device_detach,    mvs_ch_detach),
+	DEVMETHOD(device_suspend,   mvs_ch_suspend),
+	DEVMETHOD(device_resume,    mvs_ch_resume),
+	{ 0, 0 }
+};
+static driver_t mvsch_driver = {
+        "mvsch",
+        mvsch_methods,
+        sizeof(struct mvs_channel)
+};
+DRIVER_MODULE(mvsch, mvs, mvsch_driver, mvsch_devclass, 0, 0);
+DRIVER_MODULE(mvsch, sata, mvsch_driver, mvsch_devclass, 0, 0);
+
+static void
+mvs_phy_check_events(device_t dev, u_int32_t serr)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	if (ch->pm_level == 0) {
+		u_int32_t status = ATA_INL(ch->r_mem, SATA_SS);
+		union ccb *ccb;
+
+		if (bootverbose) {
+			if (((status & SATA_SS_DET_MASK) == SATA_SS_DET_PHY_ONLINE) &&
+			    ((status & SATA_SS_SPD_MASK) != SATA_SS_SPD_NO_SPEED) &&
+			    ((status & SATA_SS_IPM_MASK) == SATA_SS_IPM_ACTIVE)) {
+				device_printf(dev, "CONNECT requested\n");
+			} else
+				device_printf(dev, "DISCONNECT requested\n");
+		}
+		mvs_reset(dev);
+		if ((ccb = xpt_alloc_ccb_nowait()) == NULL)
+			return;
+		if (xpt_create_path(&ccb->ccb_h.path, NULL,
+		    cam_sim_path(ch->sim),
+		    CAM_TARGET_WILDCARD, CAM_LUN_WILDCARD) != CAM_REQ_CMP) {
+			xpt_free_ccb(ccb);
+			return;
+		}
+		xpt_rescan(ccb);
+	}
+}
+
+static void
+mvs_notify_events(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct cam_path *dpath;
+	uint32_t fis;
+	int d;
+
+	/* Try to read PMP field from SDB FIS. Present only for Gen-IIe. */
+	fis = ATA_INL(ch->r_mem, SATA_FISDW0);
+	if ((fis & 0x80ff) == 0x80a1)
+		d = (fis & 0x0f00) >> 8;
+	else
+		d = ch->pm_present ? 15 : 0;
+	if (bootverbose)
+		device_printf(dev, "SNTF %d\n", d);
+	if (xpt_create_path(&dpath, NULL,
+	    xpt_path_path_id(ch->path), d, 0) == CAM_REQ_CMP) {
+		xpt_async(AC_SCSI_AEN, dpath, NULL);
+		xpt_free_path(dpath);
+	}
+}
+
+static void
+mvs_ch_intr_locked(void *data)
+{
+	struct mvs_intr_arg *arg = (struct mvs_intr_arg *)data;
+	device_t dev = (device_t)arg->arg;
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	mtx_lock(&ch->mtx);
+	mvs_ch_intr(data);
+	mtx_unlock(&ch->mtx);
+}
+
+static void
+mvs_ch_pm(void *arg)
+{
+	device_t dev = (device_t)arg;
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint32_t work;
+
+	if (ch->numrslots != 0)
+		return;
+	/* If we are idle - request power state transition. */
+	work = ATA_INL(ch->r_mem, SATA_SC);
+	work &= ~SATA_SC_SPM_MASK;
+	if (ch->pm_level == 4)
+		work |= SATA_SC_SPM_PARTIAL;
+	else
+		work |= SATA_SC_SPM_SLUMBER;
+	ATA_OUTL(ch->r_mem, SATA_SC, work);
+}
+
+static void
+mvs_ch_pm_wake(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint32_t work;
+	int timeout = 0;
+
+	work = ATA_INL(ch->r_mem, SATA_SS);
+	if (work & SATA_SS_IPM_ACTIVE)
+		return;
+	/* If we are not in active state - request power state transition. */
+	work = ATA_INL(ch->r_mem, SATA_SC);
+	work &= ~SATA_SC_SPM_MASK;
+	work |= SATA_SC_SPM_ACTIVE;
+	ATA_OUTL(ch->r_mem, SATA_SC, work);
+	/* Wait for transition to happen. */
+	while ((ATA_INL(ch->r_mem, SATA_SS) & SATA_SS_IPM_ACTIVE) == 0 &&
+	    timeout++ < 100) {
+		DELAY(100);
+	}
+}
+
+static void
+mvs_ch_intr(void *data)
+{
+	struct mvs_intr_arg *arg = (struct mvs_intr_arg *)data;
+	device_t dev = (device_t)arg->arg;
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint32_t iec, serr = 0, fisic = 0;
+	enum mvs_err_type et;
+	int i, ccs, port = -1, selfdis = 0;
+	int edma = (ch->numtslots != 0 || ch->numdslots != 0);
+
+//device_printf(dev, "irq cause %02x EDMA %d IEC %08x\n",
+//    arg->cause, edma, ATA_INL(ch->r_mem, EDMA_IEC));
+	/* New item in response queue. */
+	if ((arg->cause & 2) && edma)
+		mvs_crbq_intr(dev);
+	/* Some error or special event. */
+	if (arg->cause & 1) {
+		iec = ATA_INL(ch->r_mem, EDMA_IEC);
+//device_printf(dev, "irq cause %02x EDMA %d IEC %08x\n",
+//    arg->cause, edma, iec);
+		if (iec & EDMA_IE_SERRINT) {
+			serr = ATA_INL(ch->r_mem, SATA_SE);
+			ATA_OUTL(ch->r_mem, SATA_SE, serr);
+//device_printf(dev, "SERR %08x\n", serr);
+		}
+		/* EDMA self-disabled due to error. */
+		if (iec & EDMA_IE_ESELFDIS)
+			selfdis = 1;
+		/* Transport interrupt. */
+		if (iec & EDMA_IE_ETRANSINT) {
+			/* For Gen-I this bit means self-disable. */
+			if (ch->quirks & MVS_Q_GENI)
+				selfdis = 1;
+			/* For Gen-II this bit means SDB-N. */
+			else if (ch->quirks & MVS_Q_GENII)
+				fisic = SATA_FISC_FISWAIT4HOSTRDYEN_B1;
+			else	/* For Gen-IIe - read FIS interrupt cause. */
+				fisic = ATA_INL(ch->r_mem, SATA_FISIC);
+//device_printf(dev, "FISIC %08x\n", fisic);
+		}
+		if (selfdis)
+			ch->curr_mode = MVS_EDMA_UNKNOWN;
+		ATA_OUTL(ch->r_mem, EDMA_IEC, ~iec);
+		/* Interface errors or Device error. */
+		if (iec & (0xfc1e9000 | EDMA_IE_EDEVERR)) {
+			port = -1;
+			if (ch->numpslots != 0) {
+				ccs = 0;
+			} else {
+				if (ch->quirks & MVS_Q_GENIIE)
+					ccs = EDMA_S_EIOID(ATA_INL(ch->r_mem, EDMA_S));
+				else
+					ccs = EDMA_S_EDEVQUETAG(ATA_INL(ch->r_mem, EDMA_S));
+				/* Check if error is one-PMP-port-specific, */
+				if (ch->fbs_enabled) {
+					/* Which ports were active. */
+					for (i = 0; i < 16; i++) {
+						if (ch->numrslotspd[i] == 0)
+							continue;
+						if (port == -1)
+							port = i;
+						else if (port != i) {
+							port = -2;
+							break;
+						}
+					}
+					/* If several ports were active and EDMA still enabled - 
+					 * other ports are probably unaffected and may continue.
+					 */
+					if (port == -2 && !selfdis) {
+						uint16_t p = ATA_INL(ch->r_mem, SATA_SATAITC) >> 16;
+						port = ffs(p) - 1;
+						if (port != (fls(p) - 1))
+							port = -2;
+					}
+				}
+			}
+//device_printf(dev, "err slot %d port %d\n", ccs, port);
+			mvs_requeue_frozen(dev);
+			for (i = 0; i < MVS_MAX_SLOTS; i++) {
+				/* XXX: reqests in loading state. */
+				if (((ch->rslots >> i) & 1) == 0)
+					continue;
+				if (port >= 0 &&
+				    ch->slot[i].ccb->ccb_h.target_id != port)
+					continue;
+				if (iec & EDMA_IE_EDEVERR) { /* Device error. */
+				    if (port != -2) {
+					if (ch->numtslots == 0) {
+						/* Untagged operation. */
+						if (i == ccs)
+							et = MVS_ERR_TFE;
+						else
+							et = MVS_ERR_INNOCENT;
+					} else {
+						/* Tagged operation. */
+						et = MVS_ERR_NCQ;
+					}
+				    } else {
+					et = MVS_ERR_TFE;
+					ch->fatalerr = 1;
+				    }
+				} else if (iec & 0xfc1e9000) {
+					if (ch->numtslots == 0 && i != ccs && port != -2)
+						et = MVS_ERR_INNOCENT;
+					else
+						et = MVS_ERR_SATA;
+				} else
+					et = MVS_ERR_INVALID;
+				mvs_end_transaction(&ch->slot[i], et);
+			}
+		}
+		/* Process SDB-N. */
+		if (fisic & SATA_FISC_FISWAIT4HOSTRDYEN_B1)
+			mvs_notify_events(dev);
+		if (fisic)
+			ATA_OUTL(ch->r_mem, SATA_FISIC, ~fisic);
+		/* Process hot-plug. */
+		if ((iec & (EDMA_IE_EDEVDIS | EDMA_IE_EDEVCON)) ||
+		    (serr & SATA_SE_PHY_CHANGED))
+			mvs_phy_check_events(dev, serr);
+	}
+	/* Legacy mode device interrupt. */
+	if ((arg->cause & 2) && !edma)
+		mvs_legacy_intr(dev);
+}
+
+static uint8_t
+mvs_getstatus(device_t dev, int clear)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint8_t status = ATA_INB(ch->r_mem, clear ? ATA_STATUS : ATA_ALTSTAT);
+
+	if (ch->fake_busy) {
+		if (status & (ATA_S_BUSY | ATA_S_DRQ | ATA_S_ERROR))
+			ch->fake_busy = 0;
+		else
+			status |= ATA_S_BUSY;
+	}
+	return (status);
+}
+
+static void
+mvs_legacy_intr(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct mvs_slot *slot = &ch->slot[0]; /* PIO is always in slot 0. */
+	union ccb *ccb = slot->ccb;
+	enum mvs_err_type et = MVS_ERR_NONE;
+	int port;
+	u_int length;
+	uint8_t status, ireason;
+
+	/* Clear interrupt and get status. */
+	status = mvs_getstatus(dev, 1);
+//	device_printf(dev, "Legacy intr status %02x\n",
+//	    status);
+	if (slot->state < MVS_SLOT_RUNNING)
+	    return;
+	port = ccb->ccb_h.target_id & 0x0f;
+	/* Wait a bit for late !BUSY status update. */
+	if (status & ATA_S_BUSY) {
+		DELAY(100);
+		if ((status = mvs_getstatus(dev, 1)) & ATA_S_BUSY) {
+			DELAY(1000);
+			if ((status = mvs_getstatus(dev, 1)) & ATA_S_BUSY)
+				return;
+		}
+	}
+	/* If we got an error, we are done. */
+	if (status & ATA_S_ERROR) {
+		et = MVS_ERR_TFE;
+		goto end_finished;
+	}
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) { /* ATA PIO */
+		ccb->ataio.res.status = status;
+		/* Are we moving data? */
+		if ((ccb->ccb_h.flags & CAM_DIR_MASK) != CAM_DIR_NONE) {
+		    /* If data read command - get them. */
+		    if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_IN) {
+			if (mvs_wait(dev, ATA_S_DRQ, ATA_S_BUSY, 1000) < 0) {
+			    device_printf(dev, "timeout waiting for read DRQ\n");
+			    et = MVS_ERR_TIMEOUT;
+			    goto end_finished;
+			}
+			ATA_INSW_STRM(ch->r_mem, ATA_DATA,
+			   (uint16_t *)(ccb->ataio.data_ptr + ch->donecount),
+			   ch->transfersize / 2);
+		    }
+		    /* Update how far we've gotten. */
+		    ch->donecount += ch->transfersize;
+		    /* Do we need more? */
+		    if (ccb->ataio.dxfer_len > ch->donecount) {
+			/* Set this transfer size according to HW capabilities */
+			ch->transfersize = min(ccb->ataio.dxfer_len - ch->donecount,
+			    ch->curr[ccb->ccb_h.target_id].bytecount);
+			/* If data write command - put them */
+			if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_OUT) {
+				if (mvs_wait(dev, ATA_S_DRQ, ATA_S_BUSY, 1000) < 0) {
+				    device_printf(dev, "timeout waiting for write DRQ\n");
+				    et = MVS_ERR_TIMEOUT;
+				    goto end_finished;
+				}
+				ATA_OUTSW_STRM(ch->r_mem, ATA_DATA,
+				   (uint16_t *)(ccb->ataio.data_ptr + ch->donecount),
+				   ch->transfersize / 2);
+				return;
+			}
+			/* If data read command, return & wait for interrupt */
+			if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_IN)
+				return;
+		    }
+		}
+	} else if (ch->basic_dma) {	/* ATAPI DMA */
+		if (status & ATA_S_DWF)
+			et = MVS_ERR_TFE;
+		else if (ATA_INL(ch->r_mem, DMA_S) & DMA_S_ERR)
+			et = MVS_ERR_TFE;
+		/* Stop basic DMA. */
+		ATA_OUTL(ch->r_mem, DMA_C, 0);
+		goto end_finished;
+	} else {			/* ATAPI PIO */
+		length = ATA_INB(ch->r_mem,ATA_CYL_LSB) | (ATA_INB(ch->r_mem,ATA_CYL_MSB) << 8);
+		ireason = ATA_INB(ch->r_mem,ATA_IREASON);
+//device_printf(dev, "status %02x, ireason %02x, length %d\n", status, ireason, length);
+		switch ((ireason & (ATA_I_CMD | ATA_I_IN)) |
+			(status & ATA_S_DRQ)) {
+
+		case ATAPI_P_CMDOUT:
+device_printf(dev, "ATAPI CMDOUT\n");
+		    /* Return wait for interrupt */
+		    return;
+
+		case ATAPI_P_WRITE:
+//device_printf(dev, "ATAPI WRITE\n");
+		    if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_IN) {
+			device_printf(dev, "trying to write on read buffer\n");
+			et = MVS_ERR_TFE;
+			goto end_finished;
+			break;
+		    }
+		    ATA_OUTSW_STRM(ch->r_mem, ATA_DATA,
+			(uint16_t *)(ccb->csio.data_ptr + ch->donecount),
+			length / 2);
+		    ch->donecount += length;
+		    /* Set next transfer size according to HW capabilities */
+		    ch->transfersize = min(ccb->csio.dxfer_len - ch->donecount,
+			    ch->curr[ccb->ccb_h.target_id].bytecount);
+		    /* Return wait for interrupt */
+		    return;
+
+		case ATAPI_P_READ:
+//device_printf(dev, "ATAPI READ\n");
+		    if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_OUT) {
+			device_printf(dev, "trying to read on write buffer\n");
+			et = MVS_ERR_TFE;
+			goto end_finished;
+		    }
+		    ATA_INSW_STRM(ch->r_mem, ATA_DATA,
+			(uint16_t *)(ccb->csio.data_ptr + ch->donecount),
+			length / 2);
+		    ch->donecount += length;
+		    /* Set next transfer size according to HW capabilities */
+		    ch->transfersize = min(ccb->csio.dxfer_len - ch->donecount,
+			    ch->curr[ccb->ccb_h.target_id].bytecount);
+		    /* Return wait for interrupt */
+		    return;
+
+		case ATAPI_P_DONEDRQ:
+device_printf(dev, "ATAPI DONEDRQ\n");
+		    device_printf(dev,
+			  "WARNING - DONEDRQ non conformant device\n");
+		    if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_IN) {
+			ATA_INSW_STRM(ch->r_mem, ATA_DATA,
+			    (uint16_t *)(ccb->csio.data_ptr + ch->donecount),
+			    length / 2);
+			ch->donecount += length;
+		    }
+		    else if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_OUT) {
+			ATA_OUTSW_STRM(ch->r_mem, ATA_DATA,
+			    (uint16_t *)(ccb->csio.data_ptr + ch->donecount),
+			    length / 2);
+			ch->donecount += length;
+		    }
+		    else
+			et = MVS_ERR_TFE;
+		    /* FALLTHROUGH */
+
+		case ATAPI_P_ABORT:
+		case ATAPI_P_DONE:
+//device_printf(dev, "ATAPI ABORT/DONE\n");
+		    if (status & (ATA_S_ERROR | ATA_S_DWF))
+			et = MVS_ERR_TFE;
+		    goto end_finished;
+
+		default:
+		    device_printf(dev, "unknown transfer phase (status %02x, ireason %02x)\n",
+			status, ireason);
+		    et = MVS_ERR_TFE;
+		}
+	}
+
+end_finished:
+	mvs_end_transaction(slot, et);
+}
+
+static void
+mvs_crbq_intr(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct mvs_crpb *crpb;
+	union ccb *ccb;
+	int in_idx, cin_idx, slot;
+	uint16_t flags;
+
+	in_idx = (ATA_INL(ch->r_mem, EDMA_RESQIP) & EDMA_RESQP_ERPQP_MASK) >>
+	    EDMA_RESQP_ERPQP_SHIFT;
+	bus_dmamap_sync(ch->dma.workrp_tag, ch->dma.workrp_map,
+	    BUS_DMASYNC_POSTREAD);
+	cin_idx = ch->in_idx;
+	ch->in_idx = in_idx;
+	while (in_idx != cin_idx) {
+		crpb = (struct mvs_crpb *)
+		    (ch->dma.workrp + MVS_CRPB_OFFSET + (MVS_CRPB_SIZE * cin_idx));
+		slot = le16toh(crpb->id) & MVS_CRPB_TAG_MASK;
+		flags = le16toh(crpb->rspflg);
+//device_printf(dev, "CRPB %d %d %04x\n", cin_idx, slot, flags);
+		/*
+		 * Handle only successfull completions here.
+		 * Errors will be handled by main intr handler.
+		 */
+		if (ch->numtslots != 0 || (flags & EDMA_IE_EDEVERR) == 0) {
+if ((flags >> 8) & ATA_S_ERROR)
+device_printf(dev, "ERROR STATUS CRPB %d %d %04x\n", cin_idx, slot, flags);
+			if (ch->slot[slot].state >= MVS_SLOT_RUNNING) {
+				ccb = ch->slot[slot].ccb;
+				ccb->ataio.res.status = (flags & MVS_CRPB_ATASTS_MASK) >>
+				    MVS_CRPB_ATASTS_SHIFT;
+				mvs_end_transaction(&ch->slot[slot], MVS_ERR_NONE);
+			} else 
+device_printf(dev, "EMPTY CRPB %d (->%d) %d %04x\n", cin_idx, in_idx, slot, flags);
+		} else
+device_printf(dev, "ERROR FLAGS CRPB %d %d %04x\n", cin_idx, slot, flags);
+
+		cin_idx = (cin_idx + 1) & (MVS_MAX_SLOTS - 1);
+	}
+	bus_dmamap_sync(ch->dma.workrp_tag, ch->dma.workrp_map,
+	    BUS_DMASYNC_PREREAD);
+	if (cin_idx == ch->in_idx) {
+		ATA_OUTL(ch->r_mem, EDMA_RESQOP,
+		    ch->dma.workrp_bus | (cin_idx << EDMA_RESQP_ERPQP_SHIFT));
+	}
+}
+
+/* Must be called with channel locked. */
+static int
+mvs_check_collision(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+		/* NCQ DMA */
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			/* Can't mix NCQ and non-NCQ DMA commands. */
+			if (ch->numdslots != 0)
+				return (1);
+			/* Can't mix NCQ and PIO commands. */
+			if (ch->numpslots != 0)
+				return (1);
+			/* If we have no FBS */
+			if (!ch->fbs_enabled) {
+				/* Tagged command while tagged to other target is active. */
+				if (ch->numtslots != 0 &&
+				    ch->taggedtarget != ccb->ccb_h.target_id)
+					return (1);
+			}
+		/* Non-NCQ DMA */
+		} else if (ccb->ataio.cmd.flags & CAM_ATAIO_DMA) {
+			/* Can't mix non-NCQ DMA and NCQ commands. */
+			if (ch->numtslots != 0)
+				return (1);
+			/* Can't mix non-NCQ DMA and PIO commands. */
+			if (ch->numpslots != 0)
+				return (1);
+		/* PIO */
+		} else {
+			/* Can't mix PIO with anything. */
+			if (ch->numrslots != 0)
+				return (1);
+		}
+		if (ccb->ataio.cmd.flags & (CAM_ATAIO_CONTROL | CAM_ATAIO_NEEDRESULT)) {
+			/* Atomic command while anything active. */
+			if (ch->numrslots != 0)
+				return (1);
+		}
+	} else { /* ATAPI */
+		/* ATAPI goes without EDMA, so can't mix it with anything. */
+		if (ch->numrslots != 0)
+			return (1);
+	}
+	/* We have some atomic command running. */
+	if (ch->aslots != 0)
+		return (1);
+	return (0);
+}
+
+static void
+mvs_tfd_read(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct ata_res *res = &ccb->ataio.res;
+
+	res->status = ATA_INB(ch->r_mem, ATA_ALTSTAT);
+	res->error =  ATA_INB(ch->r_mem, ATA_ERROR);
+	res->device = ATA_INB(ch->r_mem, ATA_DRIVE);
+	ATA_OUTB(ch->r_mem, ATA_CONTROL, ATA_A_HOB);
+	res->sector_count_exp = ATA_INB(ch->r_mem, ATA_COUNT);
+	res->lba_low_exp = ATA_INB(ch->r_mem, ATA_SECTOR);
+	res->lba_mid_exp = ATA_INB(ch->r_mem, ATA_CYL_LSB);
+	res->lba_high_exp = ATA_INB(ch->r_mem, ATA_CYL_MSB);
+	ATA_OUTB(ch->r_mem, ATA_CONTROL, 0);
+	res->sector_count = ATA_INB(ch->r_mem, ATA_COUNT);
+	res->lba_low = ATA_INB(ch->r_mem, ATA_SECTOR);
+	res->lba_mid = ATA_INB(ch->r_mem, ATA_CYL_LSB);
+	res->lba_high = ATA_INB(ch->r_mem, ATA_CYL_MSB);
+}
+
+static void
+mvs_tfd_write(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct ata_cmd *cmd = &ccb->ataio.cmd;
+
+	ATA_OUTB(ch->r_mem, ATA_DRIVE, cmd->device);
+	ATA_OUTB(ch->r_mem, ATA_CONTROL, cmd->control);
+	ATA_OUTB(ch->r_mem, ATA_FEATURE, cmd->features_exp);
+	ATA_OUTB(ch->r_mem, ATA_FEATURE, cmd->features);
+	ATA_OUTB(ch->r_mem, ATA_COUNT, cmd->sector_count_exp);
+	ATA_OUTB(ch->r_mem, ATA_COUNT, cmd->sector_count);
+	ATA_OUTB(ch->r_mem, ATA_SECTOR, cmd->lba_low_exp);
+	ATA_OUTB(ch->r_mem, ATA_SECTOR, cmd->lba_low);
+	ATA_OUTB(ch->r_mem, ATA_CYL_LSB, cmd->lba_mid_exp);
+	ATA_OUTB(ch->r_mem, ATA_CYL_LSB, cmd->lba_mid);
+	ATA_OUTB(ch->r_mem, ATA_CYL_MSB, cmd->lba_high_exp);
+	ATA_OUTB(ch->r_mem, ATA_CYL_MSB, cmd->lba_high);
+	ATA_OUTB(ch->r_mem, ATA_COMMAND, cmd->command);
+}
+
+
+/* Must be called with channel locked. */
+static void
+mvs_begin_transaction(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	struct mvs_slot *slot;
+	int slotn, tag;
+
+	if (ch->pm_level > 0)
+		mvs_ch_pm_wake(dev);
+	/* Softreset is a special case. */
+	if (ccb->ccb_h.func_code == XPT_ATA_IO &&
+	    (ccb->ataio.cmd.flags & CAM_ATAIO_CONTROL)) {
+		mvs_softreset(dev, ccb);
+		return;
+	}
+	/* Choose empty slot. */
+	slotn = ffs(~ch->oslots) - 1;
+	if ((ccb->ccb_h.func_code == XPT_ATA_IO) &&
+	    (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA)) {
+		if (ch->quirks & MVS_Q_GENIIE)
+			tag = ffs(~ch->otagspd[ccb->ccb_h.target_id]) - 1;
+		else
+			tag = slotn;
+	} else
+		tag = 0;
+	/* Occupy chosen slot. */
+	slot = &ch->slot[slotn];
+	slot->ccb = ccb;
+	slot->tag = tag;
+	/* Stop PM timer. */
+	if (ch->numrslots == 0 && ch->pm_level > 3)
+		callout_stop(&ch->pm_timer);
+	/* Update channel stats. */
+	ch->oslots |= (1 << slot->slot);
+	ch->numrslots++;
+	ch->numrslotspd[ccb->ccb_h.target_id]++;
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			ch->otagspd[ccb->ccb_h.target_id] |= (1 << slot->tag);
+			ch->numtslots++;
+			ch->numtslotspd[ccb->ccb_h.target_id]++;
+			ch->taggedtarget = ccb->ccb_h.target_id;
+			mvs_set_edma_mode(dev, MVS_EDMA_NCQ);
+		} else if (ccb->ataio.cmd.flags & CAM_ATAIO_DMA) {
+			ch->numdslots++;
+			mvs_set_edma_mode(dev, MVS_EDMA_ON);
+		} else {
+			ch->numpslots++;
+			mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+		}
+		if (ccb->ataio.cmd.flags &
+		    (CAM_ATAIO_CONTROL | CAM_ATAIO_NEEDRESULT)) {
+			ch->aslots |= (1 << slot->slot);
+		}
+	} else {
+		uint8_t *cdb = (ccb->ccb_h.flags & CAM_CDB_POINTER) ?
+		    ccb->csio.cdb_io.cdb_ptr : ccb->csio.cdb_io.cdb_bytes;
+		ch->numpslots++;
+		/* Use ATAPI DMA only for commands without under-/overruns. */
+		if ((ccb->ccb_h.flags & CAM_DIR_MASK) != CAM_DIR_NONE &&
+		    ch->curr[ccb->ccb_h.target_id].mode >= ATA_DMA &&
+		    (ch->quirks & MVS_Q_SOC) == 0 &&
+		    (cdb[0] == 0x08 ||
+		     cdb[0] == 0x0a ||
+		     cdb[0] == 0x28 ||
+		     cdb[0] == 0x2a ||
+		     cdb[0] == 0x88 ||
+		     cdb[0] == 0x8a ||
+		     cdb[0] == 0xa8 ||
+		     cdb[0] == 0xaa ||
+		     cdb[0] == 0xbe)) {
+			ch->basic_dma = 1;
+		}
+		mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+	}
+	if (ch->numpslots == 0 || ch->basic_dma) {
+		void *buf;
+		bus_size_t size;
+
+		slot->state = MVS_SLOT_LOADING;
+		if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+			buf = ccb->ataio.data_ptr;
+			size = ccb->ataio.dxfer_len;
+		} else {
+			buf = ccb->csio.data_ptr;
+			size = ccb->csio.dxfer_len;
+		}
+		bus_dmamap_load(ch->dma.data_tag, slot->dma.data_map,
+		    buf, size, mvs_dmasetprd, slot, 0);
+	} else
+		mvs_legacy_execute_transaction(slot);
+}
+
+/* Locked by busdma engine. */
+static void
+mvs_dmasetprd(void *arg, bus_dma_segment_t *segs, int nsegs, int error)
+{    
+	struct mvs_slot *slot = arg;
+	struct mvs_channel *ch = device_get_softc(slot->dev);
+	struct mvs_eprd *eprd;
+	int i;
+
+	if (error) {
+		device_printf(slot->dev, "DMA load error\n");
+		mvs_end_transaction(slot, MVS_ERR_INVALID);
+		return;
+	}
+	KASSERT(nsegs <= MVS_SG_ENTRIES, ("too many DMA segment entries\n"));
+	/* If there is only one segment - no need to use S/G table on Gen-IIe. */
+	if (nsegs == 1 && ch->basic_dma == 0 && (ch->quirks & MVS_Q_GENIIE)) {
+		slot->dma.addr = segs[0].ds_addr;
+		slot->dma.len = segs[0].ds_len;
+	} else {
+		slot->dma.addr = 0;
+		/* Get a piece of the workspace for this EPRD */
+		eprd = (struct mvs_eprd *)
+		    (ch->dma.workrq + MVS_EPRD_OFFSET + (MVS_EPRD_SIZE * slot->slot));
+		/* Fill S/G table */
+		for (i = 0; i < nsegs; i++) {
+			eprd[i].prdbal = htole32(segs[i].ds_addr);
+			eprd[i].bytecount = htole32(segs[i].ds_len & MVS_EPRD_MASK);
+			eprd[i].prdbah = htole32((segs[i].ds_addr >> 16) >> 16);
+		}
+		eprd[i - 1].bytecount |= htole32(MVS_EPRD_EOF);
+	}
+	bus_dmamap_sync(ch->dma.data_tag, slot->dma.data_map,
+	    ((slot->ccb->ccb_h.flags & CAM_DIR_IN) ?
+	    BUS_DMASYNC_PREREAD : BUS_DMASYNC_PREWRITE));
+	if (ch->basic_dma)
+		mvs_legacy_execute_transaction(slot);
+	else
+		mvs_execute_transaction(slot);
+}
+
+static void
+mvs_legacy_execute_transaction(struct mvs_slot *slot)
+{
+	device_t dev = slot->dev;
+	struct mvs_channel *ch = device_get_softc(dev);
+	bus_addr_t eprd;
+	union ccb *ccb = slot->ccb;
+	int port = ccb->ccb_h.target_id & 0x0f;
+	int timeout;
+
+	slot->state = MVS_SLOT_RUNNING;
+	ch->rslots |= (1 << slot->slot);
+	ATA_OUTB(ch->r_mem, SATA_SATAICTL, port << SATA_SATAICTL_PMPTX_SHIFT);
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+//		device_printf(dev, "%d Legacy command %02x size %d\n",
+//		    port, ccb->ataio.cmd.command, ccb->ataio.dxfer_len);
+		mvs_tfd_write(dev, ccb);
+		/* Device reset doesn't interrupt. */
+		if (ccb->ataio.cmd.command == ATA_DEVICE_RESET) {
+			int timeout = 1000000;
+			do {
+			    DELAY(10);
+			    ccb->ataio.res.status = ATA_INB(ch->r_mem, ATA_STATUS);
+			} while (ccb->ataio.res.status & ATA_S_BUSY && timeout--);
+			mvs_legacy_intr(dev);
+			return;
+		}
+		ch->donecount = 0;
+		ch->transfersize = min(ccb->ataio.dxfer_len,
+		    ch->curr[port].bytecount);
+		if ((ccb->ccb_h.flags & CAM_DIR_MASK) != CAM_DIR_NONE)
+			ch->fake_busy = 1;
+		/* If data write command - output the data */
+		if ((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_OUT) {
+			if (mvs_wait(dev, ATA_S_DRQ, ATA_S_BUSY, 1000) < 0) {
+				device_printf(dev, "timeout waiting for write DRQ\n");
+				mvs_end_transaction(slot, MVS_ERR_TIMEOUT);
+				return;
+			}
+			ATA_OUTSW_STRM(ch->r_mem, ATA_DATA,
+			   (uint16_t *)(ccb->ataio.data_ptr + ch->donecount),
+			   ch->transfersize / 2);
+		}
+	} else {
+//		device_printf(dev, "%d ATAPI command %02x size %d dma %d\n",
+//		    port, ccb->csio.cdb_io.cdb_bytes[0], ccb->csio.dxfer_len,
+//		    ch->basic_dma);
+		ch->donecount = 0;
+		ch->transfersize = min(ccb->csio.dxfer_len,
+		    ch->curr[port].bytecount);
+		/* Write ATA PACKET command. */
+		if (ch->basic_dma) {
+			ATA_OUTB(ch->r_mem, ATA_FEATURE, ATA_F_DMA);
+			ATA_OUTB(ch->r_mem, ATA_CYL_LSB, 0);
+		    	ATA_OUTB(ch->r_mem, ATA_CYL_MSB, 0);
+		} else {
+			ATA_OUTB(ch->r_mem, ATA_FEATURE, 0);
+			ATA_OUTB(ch->r_mem, ATA_CYL_LSB, ch->transfersize);
+		    	ATA_OUTB(ch->r_mem, ATA_CYL_MSB, ch->transfersize >> 8);
+		}
+		ATA_OUTB(ch->r_mem, ATA_COMMAND, ATA_PACKET_CMD);
+		ch->fake_busy = 1;
+		/* Wait for ready to write ATAPI command block */
+		if (mvs_wait(dev, 0, ATA_S_BUSY, 1000) < 0) {
+			device_printf(dev, "timeout waiting for ATAPI !BUSY\n");
+			mvs_end_transaction(slot, MVS_ERR_TIMEOUT);
+			return;
+		}
+		timeout = 5000;
+		while (timeout--) {
+		    int reason = ATA_INB(ch->r_mem, ATA_IREASON);
+		    int status = ATA_INB(ch->r_mem, ATA_STATUS);
+
+		    if (((reason & (ATA_I_CMD | ATA_I_IN)) |
+			 (status & (ATA_S_DRQ | ATA_S_BUSY))) == ATAPI_P_CMDOUT)
+			break;
+		    DELAY(20);
+		}
+		if (timeout <= 0) {
+			device_printf(dev, "timeout waiting for ATAPI command ready\n");
+			mvs_end_transaction(slot, MVS_ERR_TIMEOUT);
+			return;
+		}
+		/* Write ATAPI command. */
+		ATA_OUTSW_STRM(ch->r_mem, ATA_DATA,
+		   (uint16_t *)((ccb->ccb_h.flags & CAM_CDB_POINTER) ?
+		    ccb->csio.cdb_io.cdb_ptr : ccb->csio.cdb_io.cdb_bytes),
+		   ch->curr[port].atapi / 2);
+		DELAY(10);
+		if (ch->basic_dma) {
+			/* Start basic DMA. */
+			eprd = ch->dma.workrq_bus + MVS_EPRD_OFFSET +
+			    (MVS_EPRD_SIZE * slot->slot);
+			ATA_OUTL(ch->r_mem, DMA_DTLBA, eprd);
+			ATA_OUTL(ch->r_mem, DMA_DTHBA, (eprd >> 16) >> 16);
+			ATA_OUTL(ch->r_mem, DMA_C, DMA_C_START |
+			    (((ccb->ccb_h.flags & CAM_DIR_MASK) == CAM_DIR_IN) ?
+			    DMA_C_READ : 0));
+		} else if ((ccb->ccb_h.flags & CAM_DIR_MASK) != CAM_DIR_NONE)
+			ch->fake_busy = 1;
+	}
+	/* Start command execution timeout */
+	callout_reset(&slot->timeout, (int)ccb->ccb_h.timeout * hz / 1000,
+	    (timeout_t*)mvs_timeout, slot);
+}
+
+/* Must be called with channel locked. */
+static void
+mvs_execute_transaction(struct mvs_slot *slot)
+{
+	device_t dev = slot->dev;
+	struct mvs_channel *ch = device_get_softc(dev);
+	bus_addr_t eprd;
+	struct mvs_crqb *crqb;
+	struct mvs_crqb_gen2e *crqb2e;
+	union ccb *ccb = slot->ccb;
+	int port = ccb->ccb_h.target_id & 0x0f;
+	int i;
+
+//	device_printf(dev, "%d EDMA command %02x size %d slot %d tag %d\n",
+//	    port, ccb->ataio.cmd.command, ccb->ataio.dxfer_len, slot->slot, slot->tag);
+	/* Get address of the prepared EPRD */
+	eprd = ch->dma.workrq_bus + MVS_EPRD_OFFSET + (MVS_EPRD_SIZE * slot->slot);
+	/* Prepare CRQB. Gen IIe uses different CRQB format. */
+	if (ch->quirks & MVS_Q_GENIIE) {
+		crqb2e = (struct mvs_crqb_gen2e *)
+		    (ch->dma.workrq + MVS_CRQB_OFFSET + (MVS_CRQB_SIZE * ch->out_idx));
+		crqb2e->ctrlflg = htole32(
+		    ((ccb->ccb_h.flags & CAM_DIR_IN) ? MVS_CRQB2E_READ : 0) |
+		    (slot->tag << MVS_CRQB2E_DTAG_SHIFT) |
+		    (port << MVS_CRQB2E_PMP_SHIFT) |
+		    (slot->slot << MVS_CRQB2E_HTAG_SHIFT));
+		/* If there is only one segment - no need to use S/G table. */
+		if (slot->dma.addr != 0) {
+			eprd = slot->dma.addr;
+			crqb2e->ctrlflg |= htole32(MVS_CRQB2E_CPRD);
+			crqb2e->drbc = slot->dma.len;
+		}
+		crqb2e->cprdbl = htole32(eprd);
+		crqb2e->cprdbh = htole32((eprd >> 16) >> 16);
+		crqb2e->cmd[0] = 0;
+		crqb2e->cmd[1] = 0;
+		crqb2e->cmd[2] = ccb->ataio.cmd.command;
+		crqb2e->cmd[3] = ccb->ataio.cmd.features;
+		crqb2e->cmd[4] = ccb->ataio.cmd.lba_low;
+		crqb2e->cmd[5] = ccb->ataio.cmd.lba_mid;
+		crqb2e->cmd[6] = ccb->ataio.cmd.lba_high;
+		crqb2e->cmd[7] = ccb->ataio.cmd.device;
+		crqb2e->cmd[8] = ccb->ataio.cmd.lba_low_exp;
+		crqb2e->cmd[9] = ccb->ataio.cmd.lba_mid_exp;
+		crqb2e->cmd[10] = ccb->ataio.cmd.lba_high_exp;
+		crqb2e->cmd[11] = ccb->ataio.cmd.features_exp;
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			crqb2e->cmd[12] = slot->tag << 3;
+			crqb2e->cmd[13] = 0;
+		} else {
+			crqb2e->cmd[12] = ccb->ataio.cmd.sector_count;
+			crqb2e->cmd[13] = ccb->ataio.cmd.sector_count_exp;
+		}
+		crqb2e->cmd[14] = 0;
+		crqb2e->cmd[15] = 0;
+	} else {
+		crqb = (struct mvs_crqb *)
+		    (ch->dma.workrq + MVS_CRQB_OFFSET + (MVS_CRQB_SIZE * ch->out_idx));
+		crqb->cprdbl = htole32(eprd);
+		crqb->cprdbh = htole32((eprd >> 16) >> 16);
+		crqb->ctrlflg = htole16(
+		    ((ccb->ccb_h.flags & CAM_DIR_IN) ? MVS_CRQB_READ : 0) |
+		    (slot->slot << MVS_CRQB_TAG_SHIFT) |
+		    (port << MVS_CRQB_PMP_SHIFT));
+		i = 0;
+		/*
+		 * Controller can handle only 11 of 12 ATA registers,
+		 * so we have to choose which one to skip.
+		 */
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			crqb->cmd[i++] = ccb->ataio.cmd.features_exp;
+			crqb->cmd[i++] = 0x11;
+		}
+		crqb->cmd[i++] = ccb->ataio.cmd.features;
+		crqb->cmd[i++] = 0x11;
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			crqb->cmd[i++] = slot->tag << 3;
+			crqb->cmd[i++] = 0x12;
+		} else {
+			crqb->cmd[i++] = ccb->ataio.cmd.sector_count_exp;
+			crqb->cmd[i++] = 0x12;
+			crqb->cmd[i++] = ccb->ataio.cmd.sector_count;
+			crqb->cmd[i++] = 0x12;
+		}
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_low_exp;
+		crqb->cmd[i++] = 0x13;
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_low;
+		crqb->cmd[i++] = 0x13;
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_mid_exp;
+		crqb->cmd[i++] = 0x14;
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_mid;
+		crqb->cmd[i++] = 0x14;
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_high_exp;
+		crqb->cmd[i++] = 0x15;
+		crqb->cmd[i++] = ccb->ataio.cmd.lba_high;
+		crqb->cmd[i++] = 0x15;
+		crqb->cmd[i++] = ccb->ataio.cmd.device;
+		crqb->cmd[i++] = 0x16;
+		crqb->cmd[i++] = ccb->ataio.cmd.command;
+		crqb->cmd[i++] = 0x97;
+	}
+	bus_dmamap_sync(ch->dma.workrq_tag, ch->dma.workrq_map,
+	    BUS_DMASYNC_PREWRITE);
+	bus_dmamap_sync(ch->dma.workrp_tag, ch->dma.workrp_map,
+	    BUS_DMASYNC_PREREAD);
+	slot->state = MVS_SLOT_RUNNING;
+	ch->rslots |= (1 << slot->slot);
+	/* Issue command to the controller. */
+	ch->out_idx = (ch->out_idx + 1) & (MVS_MAX_SLOTS - 1);
+	ATA_OUTL(ch->r_mem, EDMA_REQQIP,
+	    ch->dma.workrq_bus + MVS_CRQB_OFFSET + (MVS_CRQB_SIZE * ch->out_idx));
+	/* Start command execution timeout */
+	callout_reset(&slot->timeout, (int)ccb->ccb_h.timeout * hz / 1000,
+	    (timeout_t*)mvs_timeout, slot);
+	return;
+}
+
+/* Must be called with channel locked. */
+static void
+mvs_process_timeout(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int i;
+
+	mtx_assert(&ch->mtx, MA_OWNED);
+	/* Handle the rest of commands. */
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		/* Do we have a running request on slot? */
+		if (ch->slot[i].state < MVS_SLOT_RUNNING)
+			continue;
+		mvs_end_transaction(&ch->slot[i], MVS_ERR_TIMEOUT);
+	}
+}
+
+/* Must be called with channel locked. */
+static void
+mvs_rearm_timeout(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int i;
+
+	mtx_assert(&ch->mtx, MA_OWNED);
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		struct mvs_slot *slot = &ch->slot[i];
+
+		/* Do we have a running request on slot? */
+		if (slot->state < MVS_SLOT_RUNNING)
+			continue;
+		if ((ch->toslots & (1 << i)) == 0)
+			continue;
+		callout_reset(&slot->timeout,
+		    (int)slot->ccb->ccb_h.timeout * hz / 2000,
+		    (timeout_t*)mvs_timeout, slot);
+	}
+}
+
+/* Locked by callout mechanism. */
+static void
+mvs_timeout(struct mvs_slot *slot)
+{
+	device_t dev = slot->dev;
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	/* Check for stale timeout. */
+	if (slot->state < MVS_SLOT_RUNNING)
+		return;
+	device_printf(dev, "Timeout on slot %d\n", slot->slot);
+	device_printf(dev, "iec %08x sstat %08x serr %08x edma_s %08x "
+	    "dma_c %08x dma_s %08x rs %08x status %02x\n",
+	    ATA_INL(ch->r_mem, EDMA_IEC),
+	    ATA_INL(ch->r_mem, SATA_SS), ATA_INL(ch->r_mem, SATA_SE),
+	    ATA_INL(ch->r_mem, EDMA_S), ATA_INL(ch->r_mem, DMA_C),
+	    ATA_INL(ch->r_mem, DMA_S), ch->rslots,
+	    ATA_INB(ch->r_mem, ATA_ALTSTAT));
+	/* Handle frozen command. */
+	mvs_requeue_frozen(dev);
+	/* We wait for other commands timeout and pray. */
+	if (ch->toslots == 0)
+		xpt_freeze_simq(ch->sim, 1);
+	ch->toslots |= (1 << slot->slot);
+	if ((ch->rslots & ~ch->toslots) == 0)
+		mvs_process_timeout(dev);
+	else
+		device_printf(dev, " ... waiting for slots %08x\n",
+		    ch->rslots & ~ch->toslots);
+}
+
+/* Must be called with channel locked. */
+static void
+mvs_end_transaction(struct mvs_slot *slot, enum mvs_err_type et)
+{
+	device_t dev = slot->dev;
+	struct mvs_channel *ch = device_get_softc(dev);
+	union ccb *ccb = slot->ccb;
+
+//device_printf(dev, "cmd done status %d\n", et);
+	bus_dmamap_sync(ch->dma.workrq_tag, ch->dma.workrq_map,
+	    BUS_DMASYNC_POSTWRITE);
+	/* Read result registers to the result struct
+	 * May be incorrect if several commands finished same time,
+	 * so read only when sure or have to.
+	 */
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+		struct ata_res *res = &ccb->ataio.res;
+
+		if ((et == MVS_ERR_TFE) ||
+		    (ccb->ataio.cmd.flags & CAM_ATAIO_NEEDRESULT)) {
+			mvs_tfd_read(dev, ccb);
+		} else
+			bzero(res, sizeof(*res));
+	}
+	if (ch->numpslots == 0 || ch->basic_dma) {
+		if ((ccb->ccb_h.flags & CAM_DIR_MASK) != CAM_DIR_NONE) {
+			bus_dmamap_sync(ch->dma.data_tag, slot->dma.data_map,
+			    (ccb->ccb_h.flags & CAM_DIR_IN) ?
+			    BUS_DMASYNC_POSTREAD : BUS_DMASYNC_POSTWRITE);
+			bus_dmamap_unload(ch->dma.data_tag, slot->dma.data_map);
+		}
+	}
+	if (et != MVS_ERR_NONE)
+		ch->eslots |= (1 << slot->slot);
+	/* In case of error, freeze device for proper recovery. */
+	if ((et != MVS_ERR_NONE) && (!ch->readlog) &&
+	    !(ccb->ccb_h.status & CAM_DEV_QFRZN)) {
+		xpt_freeze_devq(ccb->ccb_h.path, 1);
+		ccb->ccb_h.status |= CAM_DEV_QFRZN;
+	}
+	/* Set proper result status. */
+	ccb->ccb_h.status &= ~CAM_STATUS_MASK;
+	switch (et) {
+	case MVS_ERR_NONE:
+		ccb->ccb_h.status |= CAM_REQ_CMP;
+		if (ccb->ccb_h.func_code == XPT_SCSI_IO)
+			ccb->csio.scsi_status = SCSI_STATUS_OK;
+		break;
+	case MVS_ERR_INVALID:
+		ch->fatalerr = 1;
+		ccb->ccb_h.status |= CAM_REQ_INVALID;
+		break;
+	case MVS_ERR_INNOCENT:
+		ccb->ccb_h.status |= CAM_REQUEUE_REQ;
+		break;
+	case MVS_ERR_TFE:
+	case MVS_ERR_NCQ:
+		if (ccb->ccb_h.func_code == XPT_SCSI_IO) {
+			ccb->ccb_h.status |= CAM_SCSI_STATUS_ERROR;
+			ccb->csio.scsi_status = SCSI_STATUS_CHECK_COND;
+		} else {
+			ccb->ccb_h.status |= CAM_ATA_STATUS_ERROR;
+		}
+		break;
+	case MVS_ERR_SATA:
+		ch->fatalerr = 1;
+		if (!ch->readlog) {
+			xpt_freeze_simq(ch->sim, 1);
+			ccb->ccb_h.status &= ~CAM_STATUS_MASK;
+			ccb->ccb_h.status |= CAM_RELEASE_SIMQ;
+		}
+		ccb->ccb_h.status |= CAM_UNCOR_PARITY;
+		break;
+	case MVS_ERR_TIMEOUT:
+		if (!ch->readlog) {
+			xpt_freeze_simq(ch->sim, 1);
+			ccb->ccb_h.status &= ~CAM_STATUS_MASK;
+			ccb->ccb_h.status |= CAM_RELEASE_SIMQ;
+		}
+		ccb->ccb_h.status |= CAM_CMD_TIMEOUT;
+		break;
+	default:
+		ch->fatalerr = 1;
+		ccb->ccb_h.status |= CAM_REQ_CMP_ERR;
+	}
+	/* Free slot. */
+	ch->oslots &= ~(1 << slot->slot);
+	ch->rslots &= ~(1 << slot->slot);
+	ch->aslots &= ~(1 << slot->slot);
+	if (et != MVS_ERR_TIMEOUT) {
+		if (ch->toslots == (1 << slot->slot))
+			xpt_release_simq(ch->sim, TRUE);
+		ch->toslots &= ~(1 << slot->slot);
+	}
+	slot->state = MVS_SLOT_EMPTY;
+	slot->ccb = NULL;
+	/* Update channel stats. */
+	ch->numrslots--;
+	ch->numrslotspd[ccb->ccb_h.target_id]--;
+	if (ccb->ccb_h.func_code == XPT_ATA_IO) {
+		if (ccb->ataio.cmd.flags & CAM_ATAIO_FPDMA) {
+			ch->otagspd[ccb->ccb_h.target_id] &= ~(1 << slot->tag);
+			ch->numtslots--;
+			ch->numtslotspd[ccb->ccb_h.target_id]--;
+		} else if (ccb->ataio.cmd.flags & CAM_ATAIO_DMA) {
+			ch->numdslots--;
+		} else {
+			ch->numpslots--;
+		}
+	} else {
+		ch->numpslots--;
+		ch->basic_dma = 0;
+	}
+	/* If it was our READ LOG command - process it. */
+	if (ch->readlog) {
+		mvs_process_read_log(dev, ccb);
+	/* If it was NCQ command error, put result on hold. */
+	} else if (et == MVS_ERR_NCQ) {
+		ch->hold[slot->slot] = ccb;
+		ch->holdtag[slot->slot] = slot->tag;
+		ch->numhslots++;
+	} else
+		xpt_done(ccb);
+	/* Unfreeze frozen command. */
+	if (ch->frozen && !mvs_check_collision(dev, ch->frozen)) {
+		union ccb *fccb = ch->frozen;
+		ch->frozen = NULL;
+		mvs_begin_transaction(dev, fccb);
+		xpt_release_simq(ch->sim, TRUE);
+	}
+	/* If we have no other active commands, ... */
+	if (ch->rslots == 0) {
+		/* if there was fatal error - reset port. */
+		if (ch->toslots != 0 || ch->fatalerr) {
+			mvs_reset(dev);
+		} else {
+			/* if we have slots in error, we can reinit port. */
+			if (ch->eslots != 0) {
+				mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+				ch->eslots = 0;
+			}
+			/* if there commands on hold, we can do READ LOG. */
+			if (!ch->readlog && ch->numhslots)
+				mvs_issue_read_log(dev);
+		}
+	/* If all the rest of commands are in timeout - give them chance. */
+	} else if ((ch->rslots & ~ch->toslots) == 0 &&
+	    et != MVS_ERR_TIMEOUT)
+		mvs_rearm_timeout(dev);
+	/* Start PM timer. */
+	if (ch->numrslots == 0 && ch->pm_level > 3 &&
+	    (ch->curr[ch->pm_present ? 15 : 0].caps & CTS_SATA_CAPS_D_PMREQ)) {
+		callout_schedule(&ch->pm_timer,
+		    (ch->pm_level == 4) ? hz / 1000 : hz / 8);
+	}
+}
+
+static void
+mvs_issue_read_log(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	union ccb *ccb;
+	struct ccb_ataio *ataio;
+	int i;
+
+	ch->readlog = 1;
+	/* Find some holden command. */
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		if (ch->hold[i])
+			break;
+	}
+	ccb = xpt_alloc_ccb_nowait();
+	if (ccb == NULL) {
+		device_printf(dev, "Unable allocate READ LOG command");
+		return; /* XXX */
+	}
+	ccb->ccb_h = ch->hold[i]->ccb_h;	/* Reuse old header. */
+	ccb->ccb_h.func_code = XPT_ATA_IO;
+	ccb->ccb_h.flags = CAM_DIR_IN;
+	ccb->ccb_h.timeout = 1000;	/* 1s should be enough. */
+	ataio = &ccb->ataio;
+	ataio->data_ptr = malloc(512, M_MVS, M_NOWAIT);
+	if (ataio->data_ptr == NULL) {
+		device_printf(dev, "Unable allocate memory for READ LOG command");
+		return; /* XXX */
+	}
+	ataio->dxfer_len = 512;
+	bzero(&ataio->cmd, sizeof(ataio->cmd));
+	ataio->cmd.flags = CAM_ATAIO_48BIT;
+	ataio->cmd.command = 0x2F;	/* READ LOG EXT */
+	ataio->cmd.sector_count = 1;
+	ataio->cmd.sector_count_exp = 0;
+	ataio->cmd.lba_low = 0x10;
+	ataio->cmd.lba_mid = 0;
+	ataio->cmd.lba_mid_exp = 0;
+	/* Freeze SIM while doing READ LOG EXT. */
+	xpt_freeze_simq(ch->sim, 1);
+	mvs_begin_transaction(dev, ccb);
+}
+
+static void
+mvs_process_read_log(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	uint8_t *data;
+	struct ata_res *res;
+	int i;
+
+	ch->readlog = 0;
+
+	data = ccb->ataio.data_ptr;
+	if ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_CMP &&
+	    (data[0] & 0x80) == 0) {
+		for (i = 0; i < MVS_MAX_SLOTS; i++) {
+			if (!ch->hold[i])
+				continue;
+			if (ch->hold[i]->ccb_h.target_id != ccb->ccb_h.target_id)
+				continue;
+			if ((data[0] & 0x1F) == ch->holdtag[i]) {
+				res = &ch->hold[i]->ataio.res;
+				res->status = data[2];
+				res->error = data[3];
+				res->lba_low = data[4];
+				res->lba_mid = data[5];
+				res->lba_high = data[6];
+				res->device = data[7];
+				res->lba_low_exp = data[8];
+				res->lba_mid_exp = data[9];
+				res->lba_high_exp = data[10];
+				res->sector_count = data[12];
+				res->sector_count_exp = data[13];
+			} else {
+				ch->hold[i]->ccb_h.status &= ~CAM_STATUS_MASK;
+				ch->hold[i]->ccb_h.status |= CAM_REQUEUE_REQ;
+			}
+			xpt_done(ch->hold[i]);
+			ch->hold[i] = NULL;
+			ch->numhslots--;
+		}
+	} else {
+		if ((ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_CMP)
+			device_printf(dev, "Error while READ LOG EXT\n");
+		else if ((data[0] & 0x80) == 0) {
+			device_printf(dev, "Non-queued command error in READ LOG EXT\n");
+		}
+		for (i = 0; i < MVS_MAX_SLOTS; i++) {
+			if (!ch->hold[i])
+				continue;
+			if (ch->hold[i]->ccb_h.target_id != ccb->ccb_h.target_id)
+				continue;
+			xpt_done(ch->hold[i]);
+			ch->hold[i] = NULL;
+			ch->numhslots--;
+		}
+	}
+	free(ccb->ataio.data_ptr, M_MVS);
+	xpt_free_ccb(ccb);
+	xpt_release_simq(ch->sim, TRUE);
+}
+
+static int
+mvs_wait(device_t dev, u_int s, u_int c, int t)
+{
+	int timeout = 0;
+	uint8_t st;
+
+	while (((st =  mvs_getstatus(dev, 0)) & (s | c)) != s) {
+		DELAY(1000);
+		if (timeout++ > t) {
+			device_printf(dev, "Wait status %02x\n", st);
+			return (-1);
+		}
+	} 
+	return (timeout);
+}
+
+static void
+mvs_requeue_frozen(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	union ccb *fccb = ch->frozen;
+
+	if (fccb) {
+		ch->frozen = NULL;
+		fccb->ccb_h.status = CAM_REQUEUE_REQ | CAM_RELEASE_SIMQ;
+		if (!(fccb->ccb_h.status & CAM_DEV_QFRZN)) {
+			xpt_freeze_devq(fccb->ccb_h.path, 1);
+			fccb->ccb_h.status |= CAM_DEV_QFRZN;
+		}
+		xpt_done(fccb);
+	}
+}
+
+static void
+mvs_reset(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int i;
+
+	xpt_freeze_simq(ch->sim, 1);
+	if (bootverbose)
+		device_printf(dev, "MVS reset...\n");
+	/* Requeue freezed command. */
+	mvs_requeue_frozen(dev);
+	/* Kill the engine and requeue all running commands. */
+	mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+	ATA_OUTL(ch->r_mem, DMA_C, 0);
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		/* Do we have a running request on slot? */
+		if (ch->slot[i].state < MVS_SLOT_RUNNING)
+			continue;
+		/* XXX; Commands in loading state. */
+		mvs_end_transaction(&ch->slot[i], MVS_ERR_INNOCENT);
+	}
+	for (i = 0; i < MVS_MAX_SLOTS; i++) {
+		if (!ch->hold[i])
+			continue;
+		xpt_done(ch->hold[i]);
+		ch->hold[i] = NULL;
+		ch->numhslots--;
+	}
+	if (ch->toslots != 0)
+		xpt_release_simq(ch->sim, TRUE);
+	ch->eslots = 0;
+	ch->toslots = 0;
+	ch->fatalerr = 0;
+	/* Tell the XPT about the event */
+	xpt_async(AC_BUS_RESET, ch->path, NULL);
+	ATA_OUTL(ch->r_mem, EDMA_IEM, 0);
+	ATA_OUTL(ch->r_mem, EDMA_CMD, EDMA_CMD_EATARST);
+	DELAY(25);
+	ATA_OUTL(ch->r_mem, EDMA_CMD, 0);
+	/* Reset and reconnect PHY, */
+	if (!mvs_sata_phy_reset(dev)) {
+		if (bootverbose)
+			device_printf(dev,
+			    "MVS reset done: phy reset found no device\n");
+		ch->devices = 0;
+		ATA_OUTL(ch->r_mem, SATA_SE, 0xffffffff);
+		ATA_OUTL(ch->r_mem, EDMA_IEC, 0);
+		ATA_OUTL(ch->r_mem, EDMA_IEM, ~EDMA_IE_TRANSIENT);
+		xpt_release_simq(ch->sim, TRUE);
+		return;
+	}
+	/* Wait for clearing busy status. */
+	if ((i = mvs_wait(dev, 0, ATA_S_BUSY | ATA_S_DRQ, 15000)) < 0)
+		device_printf(dev, "device is not ready\n");
+	else if (bootverbose)                                                        
+		device_printf(dev, "ready wait time=%dms\n", i);
+	ch->devices = 1;
+	ATA_OUTL(ch->r_mem, SATA_SE, 0xffffffff);
+	ATA_OUTL(ch->r_mem, EDMA_IEC, 0);
+	ATA_OUTL(ch->r_mem, EDMA_IEM, ~EDMA_IE_TRANSIENT);
+	if (bootverbose)
+		device_printf(dev, "MVS reset done: device found\n");
+	xpt_release_simq(ch->sim, TRUE);
+}
+
+static void
+mvs_softreset(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int port = ccb->ccb_h.target_id & 0x0f;
+	int i;
+
+	mvs_set_edma_mode(dev, MVS_EDMA_OFF);
+	ATA_OUTB(ch->r_mem, SATA_SATAICTL, port << SATA_SATAICTL_PMPTX_SHIFT);
+	ATA_OUTB(ch->r_mem, ATA_CONTROL, ATA_A_RESET);
+	DELAY(10000);
+	ATA_OUTB(ch->r_mem, ATA_CONTROL, 0);
+	ccb->ccb_h.status &= ~CAM_STATUS_MASK;
+	/* Wait for clearing busy status. */
+	if ((i = mvs_wait(dev, 0, ATA_S_BUSY | ATA_S_DRQ, ccb->ccb_h.timeout)) < 0) {
+		ccb->ccb_h.status |= CAM_CMD_TIMEOUT;
+	} else {
+		ccb->ccb_h.status |= CAM_REQ_CMP;
+	}
+	mvs_tfd_read(dev, ccb);
+	xpt_done(ccb);
+}
+
+static int
+mvs_sata_connect(struct mvs_channel *ch)
+{
+	u_int32_t status;
+	int timeout;
+
+	/* Wait up to 100ms for "connect well" */
+	for (timeout = 0; timeout < 100 ; timeout++) {
+		status = ATA_INL(ch->r_mem, SATA_SS);
+		if (((status & SATA_SS_DET_MASK) == SATA_SS_DET_PHY_ONLINE) &&
+		    ((status & SATA_SS_SPD_MASK) != SATA_SS_SPD_NO_SPEED) &&
+		    ((status & SATA_SS_IPM_MASK) == SATA_SS_IPM_ACTIVE))
+			break;
+		if ((status & SATA_SS_DET_MASK) == SATA_SS_DET_PHY_OFFLINE) {
+			if (bootverbose) {
+				device_printf(ch->dev, "SATA offline status=%08x\n",
+				    status);
+			}
+			return (0);
+		}
+		DELAY(1000);
+	}
+	if (timeout >= 100) {
+		if (bootverbose) {
+			device_printf(ch->dev, "SATA connect timeout status=%08x\n",
+			    status);
+		}
+		return (0);
+	}
+	if (bootverbose) {
+		device_printf(ch->dev, "SATA connect time=%dms status=%08x\n",
+		    timeout, status);
+	}
+	/* Clear SATA error register */
+	ATA_OUTL(ch->r_mem, SATA_SE, 0xffffffff);
+	return (1);
+}
+
+static int
+mvs_sata_phy_reset(device_t dev)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+	int sata_rev;
+	uint32_t val;
+
+	sata_rev = ch->user[ch->pm_present ? 15 : 0].revision;
+	if (sata_rev == 1)
+		val = SATA_SC_SPD_SPEED_GEN1;
+	else if (sata_rev == 2)
+		val = SATA_SC_SPD_SPEED_GEN2;
+	else if (sata_rev == 3)
+		val = SATA_SC_SPD_SPEED_GEN3;
+	else
+		val = 0;
+	ATA_OUTL(ch->r_mem, SATA_SC,
+	    SATA_SC_DET_RESET | val |
+	    SATA_SC_IPM_DIS_PARTIAL | SATA_SC_IPM_DIS_SLUMBER);
+	DELAY(5000);
+	ATA_OUTL(ch->r_mem, SATA_SC,
+	    SATA_SC_DET_IDLE | val | ((ch->pm_level > 0) ? 0 :
+	    (SATA_SC_IPM_DIS_PARTIAL | SATA_SC_IPM_DIS_SLUMBER)));
+	DELAY(5000);
+	if (!mvs_sata_connect(ch)) {
+		if (ch->pm_level > 0)
+			ATA_OUTL(ch->r_mem, SATA_SC, SATA_SC_DET_DISABLE);
+		return (0);
+	}
+	return (1);
+}
+
+static int
+mvs_check_ids(device_t dev, union ccb *ccb)
+{
+	struct mvs_channel *ch = device_get_softc(dev);
+
+	if (ccb->ccb_h.target_id > ((ch->quirks & MVS_Q_GENI) ? 0 : 15)) {
+		ccb->ccb_h.status = CAM_TID_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	if (ccb->ccb_h.target_lun != 0) {
+		ccb->ccb_h.status = CAM_LUN_INVALID;
+		xpt_done(ccb);
+		return (-1);
+	}
+	return (0);
+}
+
+static void
+mvsaction(struct cam_sim *sim, union ccb *ccb)
+{
+	device_t dev;
+	struct mvs_channel *ch;
+
+	CAM_DEBUG(ccb->ccb_h.path, CAM_DEBUG_TRACE, ("mvsaction func_code=%x\n",
+	    ccb->ccb_h.func_code));
+
+	ch = (struct mvs_channel *)cam_sim_softc(sim);
+	dev = ch->dev;
+	switch (ccb->ccb_h.func_code) {
+	/* Common cases first */
+	case XPT_ATA_IO:	/* Execute the requested I/O operation */
+	case XPT_SCSI_IO:
+		if (mvs_check_ids(dev, ccb))
+			return;
+		if (ch->devices == 0 ||
+		    (ch->pm_present == 0 &&
+		     ccb->ccb_h.target_id > 0 && ccb->ccb_h.target_id < 15)) {
+			ccb->ccb_h.status = CAM_SEL_TIMEOUT;
+			break;
+		}
+		/* Check for command collision. */
+		if (mvs_check_collision(dev, ccb)) {
+			/* Freeze command. */
+			ch->frozen = ccb;
+			/* We have only one frozen slot, so freeze simq also. */
+			xpt_freeze_simq(ch->sim, 1);
+			return;
+		}
+		mvs_begin_transaction(dev, ccb);
+		return;
+	case XPT_EN_LUN:		/* Enable LUN as a target */
+	case XPT_TARGET_IO:		/* Execute target I/O request */
+	case XPT_ACCEPT_TARGET_IO:	/* Accept Host Target Mode CDB */
+	case XPT_CONT_TARGET_IO:	/* Continue Host Target I/O Connection*/
+	case XPT_ABORT:			/* Abort the specified CCB */
+		/* XXX Implement */
+		ccb->ccb_h.status = CAM_REQ_INVALID;
+		break;
+	case XPT_SET_TRAN_SETTINGS:
+	{
+		struct	ccb_trans_settings *cts = &ccb->cts;
+		struct	mvs_device *d; 
+
+		if (mvs_check_ids(dev, ccb))
+			return;
+		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
+			d = &ch->curr[ccb->ccb_h.target_id];
+		else
+			d = &ch->user[ccb->ccb_h.target_id];
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_REVISION)
+			d->revision = cts->xport_specific.sata.revision;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_MODE)
+			d->mode = cts->xport_specific.sata.mode;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_BYTECOUNT) {
+			d->bytecount = min((ch->quirks & MVS_Q_GENIIE) ? 8192 : 2048,
+			    cts->xport_specific.sata.bytecount);
+		}
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_TAGS)
+			d->tags = min(MVS_MAX_SLOTS, cts->xport_specific.sata.tags);
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_PM)
+			ch->pm_present = cts->xport_specific.sata.pm_present;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_ATAPI)
+			d->atapi = cts->xport_specific.sata.atapi;
+		if (cts->xport_specific.sata.valid & CTS_SATA_VALID_CAPS)
+			d->caps = cts->xport_specific.sata.caps;
+		ccb->ccb_h.status = CAM_REQ_CMP;
+		break;
+	}
+	case XPT_GET_TRAN_SETTINGS:
+	/* Get default/user set transfer settings for the target */
+	{
+		struct	ccb_trans_settings *cts = &ccb->cts;
+		struct  mvs_device *d;
+		uint32_t status;
+
+		if (mvs_check_ids(dev, ccb))
+			return;
+		if (cts->type == CTS_TYPE_CURRENT_SETTINGS)
+			d = &ch->curr[ccb->ccb_h.target_id];
+		else
+			d = &ch->user[ccb->ccb_h.target_id];
+		cts->protocol = PROTO_ATA;
+		cts->protocol_version = PROTO_VERSION_UNSPECIFIED;
+		cts->transport = XPORT_SATA;
+		cts->transport_version = XPORT_VERSION_UNSPECIFIED;
+		cts->proto_specific.valid = 0;
+		cts->xport_specific.sata.valid = 0;
+		if (cts->type == CTS_TYPE_CURRENT_SETTINGS &&
+		    (ccb->ccb_h.target_id == 15 ||
+		    (ccb->ccb_h.target_id == 0 && !ch->pm_present))) {
+			status = ATA_INL(ch->r_mem, SATA_SS) & SATA_SS_SPD_MASK;
+			if (status & 0x0f0) {
+				cts->xport_specific.sata.revision =
+				    (status & 0x0f0) >> 4;
+				cts->xport_specific.sata.valid |=
+				    CTS_SATA_VALID_REVISION;
+			}
+			cts->xport_specific.sata.caps = d->caps & CTS_SATA_CAPS_D;
+//			if (ch->pm_level)
+//				cts->xport_specific.sata.caps |= CTS_SATA_CAPS_H_PMREQ;
+			cts->xport_specific.sata.caps &=
+			    ch->user[ccb->ccb_h.target_id].caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
+		} else {
+			cts->xport_specific.sata.revision = d->revision;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_REVISION;
+			cts->xport_specific.sata.caps = d->caps;
+			cts->xport_specific.sata.valid |= CTS_SATA_VALID_CAPS;
+		}
+		cts->xport_specific.sata.mode = d->mode;
+		cts->xport_specific.sata.valid |= CTS_SATA_VALID_MODE;
+		cts->xport_specific.sata.bytecount = d->bytecount;
+		cts->xport_specific.sata.valid |= CTS_SATA_VALID_BYTECOUNT;
+		cts->xport_specific.sata.pm_present = ch->pm_present;
+		cts->xport_specific.sata.valid |= CTS_SATA_VALID_PM;
+		cts->xport_specific.sata.tags = d->tags;
+		cts->xport_specific.sata.valid |= CTS_SATA_VALID_TAGS;
+		cts->xport_specific.sata.atapi = d->atapi;
+		cts->xport_specific.sata.valid |= CTS_SATA_VALID_ATAPI;
+		ccb->ccb_h.status = CAM_REQ_CMP;
+		break;
+	}
+	case XPT_RESET_BUS:		/* Reset the specified SCSI bus */
+	case XPT_RESET_DEV:	/* Bus Device Reset the specified SCSI device */
+		mvs_reset(dev);
+		ccb->ccb_h.status = CAM_REQ_CMP;
+		break;
+	case XPT_TERM_IO:		/* Terminate the I/O process */
+		/* XXX Implement */
+		ccb->ccb_h.status = CAM_REQ_INVALID;
+		break;
+	case XPT_PATH_INQ:		/* Path routing inquiry */
+	{
+		struct ccb_pathinq *cpi = &ccb->cpi;
+
+		cpi->version_num = 1; /* XXX??? */
+		cpi->hba_inquiry = PI_SDTR_ABLE;
+		if (!(ch->quirks & MVS_Q_GENI)) {
+			cpi->hba_inquiry |= PI_SATAPM;
+			/* Gen-II is extremely slow with NCQ on PMP. */
+			if ((ch->quirks & MVS_Q_GENIIE) || ch->pm_present == 0)
+				cpi->hba_inquiry |= PI_TAG_ABLE;
+		}
+		cpi->target_sprt = 0;
+		cpi->hba_misc = PIM_SEQSCAN;
+		cpi->hba_eng_cnt = 0;
+		if (!(ch->quirks & MVS_Q_GENI))
+			cpi->max_target = 15;
+		else
+			cpi->max_target = 0;
+		cpi->max_lun = 0;
+		cpi->initiator_id = 0;
+		cpi->bus_id = cam_sim_bus(sim);
+		cpi->base_transfer_speed = 150000;
+		strncpy(cpi->sim_vid, "FreeBSD", SIM_IDLEN);
+		strncpy(cpi->hba_vid, "Marvell", HBA_IDLEN);
+		strncpy(cpi->dev_name, cam_sim_name(sim), DEV_IDLEN);
+		cpi->unit_number = cam_sim_unit(sim);
+		cpi->transport = XPORT_SATA;
+		cpi->transport_version = XPORT_VERSION_UNSPECIFIED;
+		cpi->protocol = PROTO_ATA;
+		cpi->protocol_version = PROTO_VERSION_UNSPECIFIED;
+		cpi->maxio = MAXPHYS;
+		cpi->ccb_h.status = CAM_REQ_CMP;
+		break;
+	}
+	default:
+		ccb->ccb_h.status = CAM_REQ_INVALID;
+		break;
+	}
+	xpt_done(ccb);
+}
+
+static void
+mvspoll(struct cam_sim *sim)
+{
+	struct mvs_channel *ch = (struct mvs_channel *)cam_sim_softc(sim);
+	struct mvs_intr_arg arg;
+
+	arg.arg = ch->dev;
+	arg.cause = 2; /* XXX */
+	mvs_ch_intr(arg.arg);
+}
+
diff --git a/sys/dev/mvs/mvs.h b/sys/dev/mvs/mvs.h
new file mode 100644
index 00000000000..9ec4e3d265a
--- /dev/null
+++ b/sys/dev/mvs/mvs.h
@@ -0,0 +1,650 @@
+/*-
+ * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer,
+ *    without modification, immediately at the beginning of the file.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#include "mvs_if.h"
+
+/* Chip registers */
+#define CHIP_PCIEIC		0x1900	/* PCIe Interrupt Cause */
+#define CHIP_PCIEIM		0x1910	/* PCIe Interrupt Mask */
+#define CHIP_PCIIC		0x1d58	/* PCI Interrupt Cause */
+#define CHIP_PCIIM		0x1d5c	/* PCI Interrupt Mask */
+#define CHIP_MIC		0x1d60	/* Main Interrupt Cause */
+#define CHIP_MIM		0x1d64	/* Main Interrupt Mask */
+#define CHIP_SOC_MIC		0x20	/* SoC Main Interrupt Cause */
+#define CHIP_SOC_MIM		0x24	/* SoC Main Interrupt Mask */
+#define IC_ERR_IRQ		 (1 << 0)	/* shift by (2 * port #) */
+#define IC_DONE_IRQ		 (1 << 1)	/* shift by (2 * port #) */
+#define IC_HC0			 0x000001ff	/* bits 0-8 = HC0 */
+#define IC_HC_SHIFT		 9		/* HC1 shift */
+#define IC_HC1			 (IC_HC0 << IC_HC_SHIFT) /* 9-17 = HC1 */
+#define IC_ERR_HC0		 0x00000055	/* HC0 ERR_IRQ */
+#define IC_DONE_HC0		 0x000000aa	/* HC0 DONE_IRQ */
+#define IC_ERR_HC1		 (IC_ERR_HC0 << IC_HC_SHIFT) /* HC1 ERR_IRQ */
+#define IC_DONE_HC1		 (IC_DONE_HC0 << IC_HC_SHIFT) /* HC1 DONE_IRQ */
+#define IC_HC0_COAL_DONE	 (1 << 8)	/* HC0 IRQ coalescing */
+#define IC_HC1_COAL_DONE	 (1 << 17)	/* HC1 IRQ coalescing */
+#define IC_PCI_ERR		 (1 << 18)
+#define IC_TRAN_COAL_LO_DONE	 (1 << 19)	/* transaction coalescing */
+#define IC_TRAN_COAL_HI_DONE	 (1 << 20)	/* transaction coalescing */
+#define IC_ALL_PORTS_COAL_DONE	 (1 << 21)	/* GEN_II(E) IRQ coalescing */
+#define IC_GPIO_INT		 (1 << 22)
+#define IC_SELF_INT		 (1 << 23)
+#define IC_TWSI_INT		 (1 << 24)
+#define IC_MAIN_RSVD		 (0xfe000000)	/* bits 31-25 */
+#define IC_MAIN_RSVD_5		 (0xfff10000)	/* bits 31-19 */
+#define IC_MAIN_RSVD_SOC	 (0xfffffec0)	/* bits 31-9, 7-6 */
+
+#define CHIP_SOC_LED		0x2C	/* SoC LED Configuration */
+
+/* Chip CCC registers */
+#define CHIP_ICC		0x18008
+#define CHIP_ICC_ALL_PORTS	 (1 << 4)	/* all ports irq event */
+#define CHIP_ICT 		0x180cc
+#define CHIP_ITT		0x180d0
+#define CHIP_TRAN_COAL_CAUSE_LO	0x18088
+#define CHIP_TRAN_COAL_CAUSE_HI	0x1808c
+
+/* Host Controller registers */
+#define HC_SIZE			0x10000
+#define HC_OFFSET		0x20000
+#define HC_BASE(hc)		((hc) * HC_SIZE + HC_OFFSET)
+
+#define HC_CFG			0x0	/* Configuration */
+#define HC_CFG_TIMEOUT_MASK	 (0xff << 0)
+#define HC_CFG_NODMABS		 (1 << 8)
+#define HC_CFG_NOEDMABS		 (1 << 9)
+#define HC_CFG_NOPRDBS		 (1 << 10)
+#define HC_CFG_TIMEOUTEN	 (1 << 16)	/* Timer Enable */
+#define HC_CFG_COALDIS(p)	 (1 << ((p) + 24))/* Coalescing Disable*/
+#define HC_RQOP			0x4	/* Request Queue Out-Pointer */
+#define HC_RQIP			0x8	/* Response Queue In-Pointer */
+#define HC_ICT			0xc	/* Interrupt Coalescing Threshold */
+#define HC_ICT_SAICOALT_MASK	 0x000000ff
+#define HC_ITT			0x10	/* Interrupt Time Threshold */
+#define HC_ITT_SAITMTH_MASK	 0x00ffffff
+#define HC_IC			0x14	/* Interrupt Cause */
+#define HC_IC_DONE(p)		 (1 << (p))	/* SaCrpb/DMA Done */
+#define HC_IC_COAL		 (1 << 4)	/* Intr Coalescing */
+#define HC_IC_DEV(p)		 (1 << ((p) + 8)) /* Device Intr */
+
+/* Port registers */
+#define PORT_SIZE		0x2000
+#define PORT_OFFSET		0x2000
+#define PORT_BASE(hc)		((hc) * PORT_SIZE + PORT_OFFSET)
+
+#define EDMA_CFG		0x0	/* Configuration */
+#define EDMA_CFG_RESERVED	 (0x1f << 0)	/* Queue len ? */
+#define EDMA_CFG_ESATANATVCMDQUE (1 << 5)
+#define EDMA_CFG_ERDBSZ		 (1 << 8)
+#define EDMA_CFG_EQUE		 (1 << 9)
+#define EDMA_CFG_ERDBSZEXT	 (1 << 11)
+#define EDMA_CFG_RESERVED2	 (1 << 12)
+#define EDMA_CFG_EWRBUFFERLEN	 (1 << 13)
+#define EDMA_CFG_EDEVERR	 (1 << 14)
+#define EDMA_CFG_EEDMAFBS	 (1 << 16)
+#define EDMA_CFG_ECUTTHROUGHEN	 (1 << 17)
+#define EDMA_CFG_EEARLYCOMPLETIONEN (1 << 18)
+#define EDMA_CFG_EEDMAQUELEN	 (1 << 19)
+#define EDMA_CFG_EHOSTQUEUECACHEEN (1 << 22)
+#define EDMA_CFG_EMASKRXPM	 (1 << 23)
+#define EDMA_CFG_RESUMEDIS	 (1 << 24)
+#define EDMA_CFG_EDMAFBS	 (1 << 26)
+#define EDMA_T			0x4	/* Timer */
+#define EDMA_IEC		0x8	/* Interrupt Error Cause */
+#define EDMA_IEM		0xc	/* Interrupt Error Mask */
+#define EDMA_IE_EDEVERR		 (1 << 2)	/* EDMA Device Error */
+#define EDMA_IE_EDEVDIS		 (1 << 3)	/* EDMA Dev Disconn */
+#define EDMA_IE_EDEVCON		 (1 << 4)	/* EDMA Dev Conn */
+#define EDMA_IE_SERRINT		 (1 << 5)
+#define EDMA_IE_ESELFDIS	 (1 << 7)	/* EDMA Self Disable */
+#define EDMA_IE_ETRANSINT	 (1 << 8)	/* Transport Layer */
+#define EDMA_IE_EIORDYERR	 (1 << 12)	/* EDMA IORdy Error */
+#define EDMA_IE_LINKXERR_SATACRC (1 << 0)	/* SATA CRC error */
+#define EDMA_IE_LINKXERR_INTERNALFIFO	(1 << 1)	/* internal FIFO err */
+#define EDMA_IE_LINKXERR_LINKLAYERRESET	(1 << 2)
+	/* Link Layer is reset by the reception of SYNC primitive from device */
+#define EDMA_IE_LINKXERR_OTHERERRORS	(1 << 3)
+	/*
+	 * Link state errors, coding errors, or running disparity errors occur
+	 * during FIS reception.
+	 */
+#define EDMA_IE_LINKTXERR_FISTXABORTED   (1 << 4)	/* FIS Tx is aborted */
+#define EDMA_IE_LINKCTLRXERR(x)		((x) << 13)	/* Link Ctrl Recv Err */
+#define EDMA_IE_LINKDATARXERR(x)	((x) << 17)	/* Link Data Recv Err */
+#define EDMA_IE_LINKCTLTXERR(x)		((x) << 21)	/* Link Ctrl Tx Error */
+#define EDMA_IE_LINKDATATXERR(x)	((x) << 26)	/* Link Data Tx Error */
+#define EDMA_IE_TRANSPROTERR		(1 << 31)	/* Transport Proto E */
+#define EDMA_IE_TRANSIENT		(EDMA_IE_LINKCTLRXERR(0x0b) | \
+					 EDMA_IE_LINKCTLTXERR(0x1f))
+							/* Non-fatal Errors */
+#define EDMA_REQQBAH		0x10	/* Request Queue Base Address High */
+#define EDMA_REQQIP		0x14	/* Request Queue In-Pointer */
+#define EDMA_REQQOP		0x18	/* Request Queue Out-Pointer */
+#define EDMA_REQQP_ERQQP_SHIFT	 5
+#define EDMA_REQQP_ERQQP_MASK	 0x000003e0
+#define EDMA_REQQP_ERQQBAP_MASK	 0x00000c00
+#define EDMA_REQQP_ERQQBA_MASK	 0xfffff000
+#define EDMA_RESQBAH		0x1c	/* Response Queue Base Address High */
+#define EDMA_RESQIP		0x20	/* Response Queue In-Pointer */
+#define EDMA_RESQOP		0x24	/* Response Queue Out-Pointer */
+#define EDMA_RESQP_ERPQP_SHIFT	 3
+#define EDMA_RESQP_ERPQP_MASK	 0x000000f8
+#define EDMA_RESQP_ERPQBAP_MASK	 0x00000300
+#define EDMA_RESQP_ERPQBA_MASK	 0xfffffc00
+#define EDMA_CMD		0x28	/* Command */
+#define EDMA_CMD_EENEDMA	 (1 << 0)	/* Enable EDMA */
+#define EDMA_CMD_EDSEDMA	 (1 << 1)	/* Disable EDMA */
+#define EDMA_CMD_EATARST	 (1 << 2)	/* ATA Device Reset */
+#define EDMA_CMD_EEDMAFRZ	 (1 << 4)	/* EDMA Freeze */
+#define EDMA_TC			0x2c	/* Test Control */
+#define EDMA_S			0x30	/* Status */
+#define EDMA_S_EDEVQUETAG(s)	 ((s) & 0x0000001f)
+#define EDMA_S_EDEVDIR_WRITE	 (0 << 5)
+#define EDMA_S_EDEVDIR_READ	 (1 << 5)
+#define EDMA_S_ECACHEEMPTY	 (1 << 6)
+#define EDMA_S_EDMAIDLE		 (1 << 7)
+#define EDMA_S_ESTATE(s)	 (((s) & 0x0000ff00) >> 8)
+#define EDMA_S_EIOID(s)		 (((s) & 0x003f0000) >> 16)
+#define EDMA_IORT		0x34	/* IORdy Timeout */
+#define EDMA_CDT		0x40	/* Command Delay Threshold */
+#define EDMA_HC			0x60	/* Halt Condition */
+#define EDMA_UNKN_RESD		0x6C	/* Unknown register */
+#define EDMA_CQDCQOS(x)		(0x90 + ((x) << 2)
+					/* NCQ Done/TCQ Outstanding Status */
+
+/* ATA register defines */
+#define ATA_DATA                        0x100	/* (RW) data */
+#define ATA_FEATURE                     0x104	/* (W) feature */
+#define         ATA_F_DMA                0x01	/* enable DMA */
+#define         ATA_F_OVL                0x02	/* enable overlap */
+#define ATA_ERROR                       0x104	/* (R) error */
+#define         ATA_E_ILI                0x01	/* illegal length */
+#define         ATA_E_NM                 0x02	/* no media */
+#define         ATA_E_ABORT              0x04	/* command aborted */
+#define         ATA_E_MCR                0x08	/* media change request */
+#define         ATA_E_IDNF               0x10	/* ID not found */
+#define         ATA_E_MC                 0x20	/* media changed */
+#define         ATA_E_UNC                0x40	/* uncorrectable data */
+#define         ATA_E_ICRC               0x80	/* UDMA crc error */
+#define		ATA_E_ATAPI_SENSE_MASK	 0xf0	/* ATAPI sense key mask */
+#define ATA_COUNT                       0x108	/* (W) sector count */
+#define ATA_IREASON                     0x108   /* (R) interrupt reason */
+#define         ATA_I_CMD                0x01	/* cmd (1) | data (0) */
+#define         ATA_I_IN                 0x02	/* read (1) | write (0) */
+#define         ATA_I_RELEASE            0x04	/* released bus (1) */
+#define         ATA_I_TAGMASK            0xf8	/* tag mask */
+#define ATA_SECTOR                      0x10c	/* (RW) sector # */
+#define ATA_CYL_LSB                     0x110	/* (RW) cylinder# LSB */
+#define ATA_CYL_MSB                     0x114	/* (RW) cylinder# MSB */
+#define ATA_DRIVE                       0x118	/* (W) Sector/Drive/Head */
+#define         ATA_D_LBA                0x40	/* use LBA addressing */
+#define         ATA_D_IBM                0xa0	/* 512 byte sectors, ECC */
+#define ATA_COMMAND                     0x11c	/* (W) command */
+#define ATA_STATUS                      0x11c	/* (R) status */
+#define         ATA_S_ERROR              0x01	/* error */
+#define         ATA_S_INDEX              0x02	/* index */
+#define         ATA_S_CORR               0x04	/* data corrected */
+#define         ATA_S_DRQ                0x08	/* data request */
+#define         ATA_S_DSC                0x10	/* drive seek completed */
+#define         ATA_S_SERVICE            0x10	/* drive needs service */
+#define         ATA_S_DWF                0x20	/* drive write fault */
+#define         ATA_S_DMA                0x20	/* DMA ready */
+#define         ATA_S_READY              0x40	/* drive ready */
+#define         ATA_S_BUSY               0x80	/* busy */
+#define ATA_CONTROL                     0x120	/* (W) control */
+#define         ATA_A_IDS                0x02	/* disable interrupts */
+#define         ATA_A_RESET              0x04	/* RESET controller */
+#define         ATA_A_4BIT               0x08	/* 4 head bits */
+#define         ATA_A_HOB                0x80	/* High Order Byte enable */
+#define ATA_ALTSTAT                     0x120	/* (R) alternate status */
+#define ATAPI_P_READ                    (ATA_S_DRQ | ATA_I_IN)
+#define ATAPI_P_WRITE                   (ATA_S_DRQ)
+#define ATAPI_P_CMDOUT                  (ATA_S_DRQ | ATA_I_CMD)
+#define ATAPI_P_DONEDRQ                 (ATA_S_DRQ | ATA_I_CMD | ATA_I_IN)
+#define ATAPI_P_DONE                    (ATA_I_CMD | ATA_I_IN)
+#define ATAPI_P_ABORT                   0
+
+/* Basic DMA Registers */
+#define DMA_C				0x224	/* Basic DMA Command */
+#define DMA_C_START			 (1 << 0)
+#define DMA_C_READ			 (1 << 3)
+#define DMA_C_DREGIONVALID		 (1 << 8)
+#define DMA_C_DREGIONLAST		 (1 << 9)
+#define DMA_C_CONTFROMPREV		 (1 << 10)
+#define DMA_C_DRBC(n)			 (((n) & 0xffff) << 16)
+#define DMA_S				0x228	/* Basic DMA Status */
+#define DMA_S_ACT			 (1 << 0) /* Active */
+#define DMA_S_ERR			 (1 << 1) /* Error */
+#define DMA_S_PAUSED			 (1 << 2) /* Paused */
+#define DMA_S_LAST			 (1 << 3) /* Last */
+#define DMA_DTLBA			0x22c	/* Descriptor Table Low Base Address */
+#define DMA_DTLBA_MASK			 0xfffffff0
+#define DMA_DTHBA			0x230	/* Descriptor Table High Base Address */
+#define DMA_DRLA			0x234	/* Data Region Low Address */
+#define DMA_DRHA			0x238	/* Data Region High Address */
+
+/* Serial-ATA Registers */
+#define SATA_SS				0x300	/* SStatus */
+#define        SATA_SS_DET_MASK         0x0000000f
+#define        SATA_SS_DET_NO_DEVICE    0x00000000
+#define        SATA_SS_DET_DEV_PRESENT  0x00000001
+#define        SATA_SS_DET_PHY_ONLINE   0x00000003
+#define        SATA_SS_DET_PHY_OFFLINE  0x00000004
+
+#define        SATA_SS_SPD_MASK         0x000000f0
+#define        SATA_SS_SPD_NO_SPEED     0x00000000
+#define        SATA_SS_SPD_GEN1         0x00000010
+#define        SATA_SS_SPD_GEN2         0x00000020
+#define        SATA_SS_SPD_GEN3         0x00000040
+
+#define        SATA_SS_IPM_MASK         0x00000f00
+#define        SATA_SS_IPM_NO_DEVICE    0x00000000
+#define        SATA_SS_IPM_ACTIVE       0x00000100
+#define        SATA_SS_IPM_PARTIAL      0x00000200
+#define        SATA_SS_IPM_SLUMBER      0x00000600
+#define SATA_SE				0x304	/* SError */
+#define SATA_SEIM			0x340	/* SError Interrupt Mask */
+#define        SATA_SE_DATA_CORRECTED   0x00000001
+#define        SATA_SE_COMM_CORRECTED   0x00000002
+#define        SATA_SE_DATA_ERR         0x00000100
+#define        SATA_SE_COMM_ERR         0x00000200
+#define        SATA_SE_PROT_ERR         0x00000400
+#define        SATA_SE_HOST_ERR         0x00000800
+#define        SATA_SE_PHY_CHANGED      0x00010000
+#define        SATA_SE_PHY_IERROR       0x00020000
+#define        SATA_SE_COMM_WAKE        0x00040000
+#define        SATA_SE_DECODE_ERR       0x00080000
+#define        SATA_SE_PARITY_ERR       0x00100000
+#define        SATA_SE_CRC_ERR          0x00200000
+#define        SATA_SE_HANDSHAKE_ERR    0x00400000
+#define        SATA_SE_LINKSEQ_ERR      0x00800000
+#define        SATA_SE_TRANSPORT_ERR    0x01000000
+#define        SATA_SE_UNKNOWN_FIS      0x02000000
+#define SATA_SC				0x308	/* SControl */
+#define        SATA_SC_DET_MASK         0x0000000f
+#define        SATA_SC_DET_IDLE         0x00000000
+#define        SATA_SC_DET_RESET        0x00000001
+#define        SATA_SC_DET_DISABLE      0x00000004
+
+#define        SATA_SC_SPD_MASK         0x000000f0
+#define        SATA_SC_SPD_NO_SPEED     0x00000000
+#define        SATA_SC_SPD_SPEED_GEN1   0x00000010
+#define        SATA_SC_SPD_SPEED_GEN2   0x00000020
+#define        SATA_SC_SPD_SPEED_GEN3   0x00000040
+
+#define        SATA_SC_IPM_MASK         0x00000f00
+#define        SATA_SC_IPM_NONE         0x00000000
+#define        SATA_SC_IPM_DIS_PARTIAL  0x00000100
+#define        SATA_SC_IPM_DIS_SLUMBER  0x00000200
+
+#define        SATA_SC_SPM_MASK		0x0000f000
+#define        SATA_SC_SPM_NONE		0x00000000
+#define        SATA_SC_SPM_PARTIAL	0x00001000
+#define        SATA_SC_SPM_SLUMBER	0x00002000
+#define        SATA_SC_SPM_ACTIVE	0x00004000
+#define SATA_LTM			0x30c	/* LTMode */
+#define SATA_PHYM3			0x310	/* PHY Mode 3 */
+#define SATA_PHYM4			0x314	/* PHY Mode 4 */
+#define SATA_PHYM1			0x32c	/* PHY Mode 1 */
+#define SATA_PHYM2			0x330	/* PHY Mode 2 */
+#define SATA_BISTC			0x334	/* BIST Control */
+#define SATA_BISTDW1			0x338	/* BIST DW1 */
+#define SATA_BISTDW2			0x33c	/* BIST DW2 */
+#define SATA_SATAICFG			0x050	/* Serial-ATA Interface Configuration */
+#define SATA_SATAICFG_REFCLKCNF_20MHZ	 (0 << 0)
+#define SATA_SATAICFG_REFCLKCNF_25MHZ	 (1 << 0)
+#define SATA_SATAICFG_REFCLKCNF_30MHZ	 (2 << 0)
+#define SATA_SATAICFG_REFCLKCNF_40MHZ	 (3 << 0)
+#define SATA_SATAICFG_REFCLKCNF_MASK	 (3 << 0)
+#define SATA_SATAICFG_REFCLKDIV_1	 (0 << 2)
+#define SATA_SATAICFG_REFCLKDIV_2	 (1 << 2)	/* Used 20 or 25MHz */
+#define SATA_SATAICFG_REFCLKDIV_4	 (2 << 2)	/* Used 40MHz */
+#define SATA_SATAICFG_REFCLKDIV_3	 (3 << 2)	/* Used 30MHz */
+#define SATA_SATAICFG_REFCLKDIV_MASK	 (3 << 2)
+#define SATA_SATAICFG_REFCLKFEEDDIV_50	 (0 << 4)	/* or 100, when Gen2En is 1 */
+#define SATA_SATAICFG_REFCLKFEEDDIV_60	 (1 << 4)	/* or 120. Used 25MHz */
+#define SATA_SATAICFG_REFCLKFEEDDIV_75	 (2 << 4)	/* or 150. Used 20MHz */
+#define SATA_SATAICFG_REFCLKFEEDDIV_90	 (3 << 4)	/* or 180 */
+#define SATA_SATAICFG_REFCLKFEEDDIV_MASK (3 << 4)
+#define SATA_SATAICFG_PHYSSCEN		 (1 << 6)
+#define SATA_SATAICFG_GEN2EN		 (1 << 7)
+#define SATA_SATAICFG_COMMEN		 (1 << 8)
+#define SATA_SATAICFG_PHYSHUTDOWN	 (1 << 9)
+#define SATA_SATAICFG_TARGETMODE	 (1 << 10)	/* 1 = Initiator */
+#define SATA_SATAICFG_COMCHANNEL	 (1 << 11)
+#define SATA_SATAICFG_IGNOREBSY		 (1 << 24)
+#define SATA_SATAICFG_LINKRSTEN		 (1 << 25)
+#define SATA_SATAICFG_CMDRETXDS		 (1 << 26)
+#define SATA_SATAICTL			0x344	/* Serial-ATA Interface Control */
+#define SATA_SATAICTL_PMPTX_MASK	 0x0000000f
+#define SATA_SATAICTL_PMPTX_SHIFT	 0
+#define SATA_SATAICTL_VUM		 (1 << 8)
+#define SATA_SATAICTL_VUS		 (1 << 9)
+#define SATA_SATAICTL_EDMAACT		 (1 << 16)
+#define SATA_SATAICTL_CLEARSTAT		 (1 << 24)
+#define SATA_SATAICTL_SRST		 (1 << 25)
+#define SATA_SATAITC			0x348	/* Serial-ATA Interface Test Control */
+#define SATA_SATAIS			0x34c	/* Serial-ATA Interface Status */
+#define SATA_VU				0x35c	/* Vendor Unique */
+#define SATA_FISC			0x360	/* FIS Configuration */
+#define SATA_FISC_FISWAIT4RDYEN_B0	 (1 << 0) /* Device to Host FIS */
+#define SATA_FISC_FISWAIT4RDYEN_B1	 (1 << 1) /* SDB FIS rcv with <N>bit 0 */
+#define SATA_FISC_FISWAIT4RDYEN_B2	 (1 << 2) /* DMA Activate FIS */
+#define SATA_FISC_FISWAIT4RDYEN_B3	 (1 << 3) /* DMA Setup FIS */
+#define SATA_FISC_FISWAIT4RDYEN_B4	 (1 << 4) /* Data FIS first DW */
+#define SATA_FISC_FISWAIT4RDYEN_B5	 (1 << 5) /* Data FIS entire FIS */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B0	 (1 << 8)
+				/* Device to Host FIS with <ERR> or <DF> */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B1	 (1 << 9) /* SDB FIS rcv with <N>bit */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B2	 (1 << 10) /* SDB FIS rcv with <ERR> */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B3	 (1 << 11) /* BIST Acivate FIS */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B4	 (1 << 12) /* PIO Setup FIS */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B5	 (1 << 13) /* Data FIS with Link error */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B6	 (1 << 14) /* Unrecognized FIS type */
+#define SATA_FISC_FISWAIT4HOSTRDYEN_B7	 (1 << 15) /* Any FIS */
+#define SATA_FISC_FISDMAACTIVATESYNCRESP (1 << 16)
+#define SATA_FISC_FISUNRECTYPECONT	 (1 << 17)
+#define SATA_FISIC			0x364	/* FIS Interrupt Cause */
+#define SATA_FISIM			0x368	/* FIS Interrupt Mask */
+#define SATA_FISDW0			0x370	/* FIS DW0 */
+#define SATA_FISDW1			0x374	/* FIS DW1 */
+#define SATA_FISDW2			0x378	/* FIS DW2 */
+#define SATA_FISDW3			0x37c	/* FIS DW3 */
+#define SATA_FISDW4			0x380	/* FIS DW4 */
+#define SATA_FISDW5			0x384	/* FIS DW5 */
+#define SATA_FISDW6			0x388	/* FIS DW6 */
+
+#define MVS_MAX_PORTS			8
+#define MVS_MAX_SLOTS			32
+
+/* Pessimistic prognosis on number of required S/G entries */
+#define MVS_SG_ENTRIES		(btoc(MAXPHYS) + 1)
+
+/* EDMA Command Request Block (CRQB) Data */
+struct mvs_crqb {
+	uint32_t cprdbl;	/* cPRD Desriptor Table Base Low Address */
+	uint32_t cprdbh;	/* cPRD Desriptor Table Base High Address */
+	uint16_t ctrlflg;	/* Control Flags */
+#define MVS_CRQB_READ		0x0001
+#define MVS_CRQB_TAG_MASK	0x003e
+#define MVS_CRQB_TAG_SHIFT	1
+#define MVS_CRQB_PMP_MASK	0xf000
+#define MVS_CRQB_PMP_SHIFT	12
+	uint8_t cmd[22];
+} __packed;
+
+struct mvs_crqb_gen2e {
+	uint32_t cprdbl;	/* cPRD Desriptor Table Base Low Address */
+	uint32_t cprdbh;	/* cPRD Desriptor Table Base High Address */
+	uint32_t ctrlflg;	/* Control Flags */
+#define MVS_CRQB2E_READ		0x00000001
+#define MVS_CRQB2E_DTAG_MASK	0x0000003e
+#define MVS_CRQB2E_DTAG_SHIFT	1
+#define MVS_CRQB2E_PMP_MASK	0x0000f000
+#define MVS_CRQB2E_PMP_SHIFT	12
+#define MVS_CRQB2E_CPRD		0x00010000
+#define MVS_CRQB2E_HTAG_MASK	0x003e0000
+#define MVS_CRQB2E_HTAG_SHIFT	17
+	uint32_t drbc;		/* Data Region Byte Count */
+	uint8_t cmd[16];
+} __packed;
+
+/* EDMA Phisical Region Descriptors (ePRD) Table Data Structure */
+struct mvs_eprd {
+	uint32_t prdbal;	/* Address bits[31:1] */
+	uint32_t bytecount;	/* Byte Count */
+#define MVS_EPRD_MASK		0x0000ffff      /* max 64KB */
+#define MVS_EPRD_MAX		(MVS_EPRD_MASK + 1)
+#define MVS_EPRD_EOF		0x80000000
+	uint32_t prdbah;	/* Address bits[63:32] */
+	uint32_t resv;
+} __packed;
+
+/* Command request blocks. 32 commands. First 1Kbyte aligned. */
+#define MVS_CRQB_OFFSET		0
+#define MVS_CRQB_SIZE		32	/* sizeof(struct mvs_crqb) */
+#define MVS_CRQB_MASK		0x000003e0
+#define MVS_CRQB_SHIFT		5
+#define MVS_CRQB_TO_ADDR(slot)	((slot) << MVS_CRQB_SHIFT)
+#define MVS_ADDR_TO_CRQB(addr)	(((addr) & MVS_CRQB_MASK) >> MVS_CRQB_SHIFT)
+/* ePRD blocks. Up to 32 commands, Each 16byte aligned. */
+#define MVS_EPRD_OFFSET		(MVS_CRQB_OFFSET + MVS_CRQB_SIZE * MVS_MAX_SLOTS)
+#define MVS_EPRD_SIZE		(MVS_SG_ENTRIES * 16) /* sizeof(struct mvs_eprd) */
+/* Request work area. */
+#define MVS_WORKRQ_SIZE		(MVS_EPRD_OFFSET + MVS_EPRD_SIZE * MVS_MAX_SLOTS)
+
+/* EDMA Command Response Block (CRPB) Data */
+struct mvs_crpb {
+	uint16_t id;		/* CRPB ID */
+#define MVS_CRPB_TAG_MASK	0x001F
+#define MVS_CRPB_TAG_SHIFT	0
+	uint16_t rspflg;	/* CPRB Response Flags */
+#define MVS_CRPB_EDMASTS_MASK	0x007F
+#define MVS_CRPB_EDMASTS_SHIFT	0
+#define MVS_CRPB_ATASTS_MASK	0xFF00
+#define MVS_CRPB_ATASTS_SHIFT	8
+	uint32_t ts;		/* CPRB Time Stamp */
+} __packed;
+
+/* Command response blocks. 32 commands. First 256byte aligned. */
+#define MVS_CRPB_OFFSET		0
+#define MVS_CRPB_SIZE		sizeof(struct mvs_crpb)
+#define MVS_CRPB_MASK		0x000000f8
+#define MVS_CRPB_SHIFT		3
+#define MVS_CRPB_TO_ADDR(slot)	((slot) << MVS_CRPB_SHIFT)
+#define MVS_ADDR_TO_CRPB(addr)	(((addr) & MVS_CRPB_MASK) >> MVS_CRPB_SHIFT)
+/* Request work area. */
+#define MVS_WORKRP_SIZE		(MVS_CRPB_OFFSET + MVS_CRPB_SIZE * MVS_MAX_SLOTS)
+
+/* misc defines */
+#define ATA_IRQ_RID		0
+#define ATA_INTR_FLAGS		(INTR_MPSAFE|INTR_TYPE_BIO|INTR_ENTROPY)
+
+struct ata_dmaslot {
+    bus_dmamap_t                data_map;       /* Data DMA map */
+    bus_addr_t			addr;		/* Data address */
+    uint16_t			len;		/* Data size */
+};
+
+/* structure holding DMA related information */
+struct mvs_dma {
+    bus_dma_tag_t               workrq_tag;	/* Request workspace DMA tag */
+    bus_dmamap_t                workrq_map;	/* Request workspace DMA map */
+    uint8_t                     *workrq;	/* Request workspace */
+    bus_addr_t                  workrq_bus;	/* Request bus address */
+    bus_dma_tag_t               workrp_tag;	/* Reply workspace DMA tag */
+    bus_dmamap_t                workrp_map;	/* Reply workspace DMA map */
+    uint8_t                     *workrp;	/* Reply workspace */
+    bus_addr_t                  workrp_bus;	/* Reply bus address */
+    bus_dma_tag_t               data_tag;	/* Data DMA tag */
+};
+
+enum mvs_slot_states {
+	MVS_SLOT_EMPTY,
+	MVS_SLOT_LOADING,
+	MVS_SLOT_RUNNING,
+	MVS_SLOT_EXECUTING
+};
+
+struct mvs_slot {
+    device_t                    dev;            /* Device handle */
+    int				slot;           /* Number of this slot */
+    int				tag;            /* Used command tag */
+    enum mvs_slot_states	state;          /* Slot state */
+    union ccb			*ccb;		/* CCB occupying slot */
+    struct ata_dmaslot          dma;            /* DMA data of this slot */
+    struct callout              timeout;        /* Execution timeout */
+};
+
+struct mvs_device {
+	int			revision;
+	int			mode;
+	u_int			bytecount;
+	u_int			atapi;
+	u_int			tags;
+	u_int			caps;
+};
+
+enum mvs_edma_mode {
+	MVS_EDMA_UNKNOWN,
+	MVS_EDMA_OFF,
+	MVS_EDMA_ON,
+	MVS_EDMA_QUEUED,
+	MVS_EDMA_NCQ,
+};
+
+/* structure describing an ATA channel */
+struct mvs_channel {
+	device_t		dev;            /* Device handle */
+	int			unit;           /* Physical channel */
+	struct resource		*r_mem;		/* Memory of this channel */
+	struct resource		*r_irq;         /* Interrupt of this channel */
+	void			*ih;            /* Interrupt handle */
+	struct mvs_dma		dma;            /* DMA data */
+	struct cam_sim		*sim;
+	struct cam_path		*path;
+	int			quirks;
+#define MVS_Q_GENI	1
+#define MVS_Q_GENII	2
+#define MVS_Q_GENIIE	4
+#define MVS_Q_SOC	8
+#define MVS_Q_CT	16
+	int			pm_level;	/* power management level */
+
+	struct mvs_slot		slot[MVS_MAX_SLOTS];
+	union ccb		*hold[MVS_MAX_SLOTS];
+	int			holdtag[MVS_MAX_SLOTS]; /* Tags used for holden commands. */
+	struct mtx		mtx;		/* state lock */
+	int			devices;        /* What is present */
+	int			pm_present;	/* PM presence reported */
+	enum mvs_edma_mode	curr_mode;	/* Current EDMA mode */
+	int			fbs_enabled;	/* FIS-based switching enabled */
+	uint32_t		oslots;		/* Occupied slots */
+	uint32_t		otagspd[16];	/* Occupied device tags */
+	uint32_t		rslots;		/* Running slots */
+	uint32_t		aslots;		/* Slots with atomic commands  */
+	uint32_t		eslots;		/* Slots in error */
+	uint32_t		toslots;	/* Slots in timeout */
+	int			numrslots;	/* Number of running slots */
+	int			numrslotspd[16];/* Number of running slots per dev */
+	int			numpslots;	/* Number of PIO slots */
+	int			numdslots;	/* Number of DMA slots */
+	int			numtslots;	/* Number of NCQ slots */
+	int			numtslotspd[16];/* Number of NCQ slots per dev */
+	int			numhslots;	/* Number of holden slots */
+	int			readlog;	/* Our READ LOG active */
+	int			fatalerr;	/* Fatal error happend */
+	int			lastslot;	/* Last used slot */
+	int			taggedtarget;	/* Last tagged target */
+	int			out_idx;	/* Next written CRQB */
+	int			in_idx;		/* Next read CRPB */
+	u_int			transfersize;	/* PIO transfer size */
+	u_int			donecount;	/* PIO bytes sent/received */
+	u_int			basic_dma;	/* Basic DMA used for ATAPI */
+	u_int			fake_busy;	/* Fake busy bit after command submission */
+	union ccb		*frozen;	/* Frozen command */
+	struct callout		pm_timer;	/* Power management events */
+
+	struct mvs_device	user[16];	/* User-specified settings */
+	struct mvs_device	curr[16];	/* Current settings */
+};
+
+/* structure describing a MVS controller */
+struct mvs_controller {
+	device_t		dev;
+	int			r_rid;
+	struct resource		*r_mem;
+	struct rman		sc_iomem;
+	struct mvs_controller_irq {
+		struct resource		*r_irq;
+		void			*handle;
+		int			r_irq_rid;
+	} irq;
+	int			quirks;
+	int			channels;
+	int			ccc;		/* CCC timeout */
+	int			cccc;		/* CCC commands */
+	struct mtx		mtx;		/* MIM access lock */
+	int			gmim;		/* Globally wanted MIM bits */
+	int			pmim;		/* Port wanted MIM bits */
+	int			mim;		/* Current MIM bits */
+	int			msi;		/* MSI enabled */
+	int			msia;		/* MSI active */
+	struct {
+		void			(*function)(void *);
+		void			*argument;
+	} interrupt[MVS_MAX_PORTS];
+};
+
+enum mvs_err_type {
+	MVS_ERR_NONE,		/* No error */
+	MVS_ERR_INVALID,	/* Error detected by us before submitting. */
+	MVS_ERR_INNOCENT,	/* Innocent victim. */
+	MVS_ERR_TFE,		/* Task File Error. */
+	MVS_ERR_SATA,		/* SATA error. */
+	MVS_ERR_TIMEOUT,	/* Command execution timeout. */
+	MVS_ERR_NCQ,		/* NCQ command error. CCB should be put on hold
+				 * until READ LOG executed to reveal error. */
+};
+
+struct mvs_intr_arg {
+	void *arg;
+	u_int cause;
+};
+
+extern devclass_t mvs_devclass;
+
+/* macros to hide busspace uglyness */
+#define ATA_INB(res, offset) \
+	bus_read_1((res), (offset))
+#define ATA_INW(res, offset) \
+	bus_read_2((res), (offset))
+#define ATA_INL(res, offset) \
+	bus_read_4((res), (offset))
+#define ATA_INSW(res, offset, addr, count) \
+	bus_read_multi_2((res), (offset), (addr), (count))
+#define ATA_INSW_STRM(res, offset, addr, count) \
+	bus_read_multi_stream_2((res), (offset), (addr), (count))
+#define ATA_INSL(res, offset, addr, count) \
+	bus_read_multi_4((res), (offset), (addr), (count))
+#define ATA_INSL_STRM(res, offset, addr, count) \
+	bus_read_multi_stream_4((res), (offset), (addr), (count))
+#define ATA_OUTB(res, offset, value) \
+	bus_write_1((res), (offset), (value))
+#define ATA_OUTW(res, offset, value) \
+	bus_write_2((res), (offset), (value))
+#define ATA_OUTL(res, offset, value) \
+	bus_write_4((res), (offset), (value));
+#define ATA_OUTSW(res, offset, addr, count) \
+	bus_write_multi_2((res), (offset), (addr), (count))
+#define ATA_OUTSW_STRM(res, offset, addr, count) \
+	bus_write_multi_stream_2((res), (offset), (addr), (count))
+#define ATA_OUTSL(res, offset, addr, count) \
+	bus_write_multi_4((res), (offset), (addr), (count))
+#define ATA_OUTSL_STRM(res, offset, addr, count) \
+	bus_write_multi_stream_4((res), (offset), (addr), (count))
diff --git a/sys/dev/mvs/mvs_if.m b/sys/dev/mvs/mvs_if.m
new file mode 100644
index 00000000000..e7442192065
--- /dev/null
+++ b/sys/dev/mvs/mvs_if.m
@@ -0,0 +1,34 @@
+# Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
+# All rights reserved.
+#
+# Redistribution and use in source and binary forms, with or without
+# modification, are permitted provided that the following conditions
+# are met:
+# 1. Redistributions of source code must retain the above copyright
+#    notice, this list of conditions and the following disclaimer,
+#    without modification, immediately at the beginning of the file.
+# 2. Redistributions in binary form must reproduce the above copyright
+#    notice, this list of conditions and the following disclaimer in the
+#    documentation and/or other materials provided with the distribution.
+#
+# THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+# IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+# OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+# IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+# INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+# NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+# THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+#
+# $FreeBSD$
+
+INTERFACE mvs;
+
+METHOD void edma {
+    device_t    dev;
+    device_t    child;
+    int         mode;
+};
+
diff --git a/sys/dev/mvs/mvs_pci.c b/sys/dev/mvs/mvs_pci.c
new file mode 100644
index 00000000000..4fae627c6aa
--- /dev/null
+++ b/sys/dev/mvs/mvs_pci.c
@@ -0,0 +1,507 @@
+/*-
+ * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer,
+ *    without modification, immediately at the beginning of the file.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/module.h>
+#include <sys/systm.h>
+#include <sys/kernel.h>
+#include <sys/bus.h>
+#include <sys/endian.h>
+#include <sys/malloc.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <vm/uma.h>
+#include <machine/stdarg.h>
+#include <machine/resource.h>
+#include <machine/bus.h>
+#include <sys/rman.h>
+#include <dev/pci/pcivar.h>
+#include <dev/pci/pcireg.h>
+#include "mvs.h"
+
+/* local prototypes */
+static int mvs_setup_interrupt(device_t dev);
+static void mvs_intr(void *data);
+static int mvs_suspend(device_t dev);
+static int mvs_resume(device_t dev);
+static int mvs_ctlr_setup(device_t dev);
+
+static struct {
+	uint32_t	id;
+	uint8_t		rev;
+	const char	*name;
+	int		ports;
+	int		quirks;
+} mvs_ids[] = {
+	{0x504011ab, 0x00, "Marvell 88SX5040",	4,	MVS_Q_GENI},
+	{0x504111ab, 0x00, "Marvell 88SX5041",	4,	MVS_Q_GENI},
+	{0x508011ab, 0x00, "Marvell 88SX5080",	8,	MVS_Q_GENI},
+	{0x508111ab, 0x00, "Marvell 88SX5081",	8,	MVS_Q_GENI},
+	{0x604011ab, 0x00, "Marvell 88SX6040",	4,	MVS_Q_GENII},
+	{0x604111ab, 0x00, "Marvell 88SX6041",	4,	MVS_Q_GENII},
+	{0x604211ab, 0x00, "Marvell 88SX6042",	4,	MVS_Q_GENIIE},
+	{0x608011ab, 0x00, "Marvell 88SX6080",	8,	MVS_Q_GENII},
+	{0x608111ab, 0x00, "Marvell 88SX6081",	8,	MVS_Q_GENII},
+	{0x704211ab, 0x00, "Marvell 88SX7042",	4,	MVS_Q_GENIIE|MVS_Q_CT},
+	{0x02419005, 0x00, "Adaptec 1420SA",	4,	MVS_Q_GENII},
+	{0x02439005, 0x00, "Adaptec 1430SA",	4,	MVS_Q_GENIIE|MVS_Q_CT},
+	{0x00000000, 0x00, NULL,	0,	0}
+};
+
+static int
+mvs_probe(device_t dev)
+{
+	char buf[64];
+	int i;
+	uint32_t devid = pci_get_devid(dev);
+	uint8_t revid = pci_get_revid(dev);
+
+	for (i = 0; mvs_ids[i].id != 0; i++) {
+		if (mvs_ids[i].id == devid &&
+		    mvs_ids[i].rev <= revid) {
+			snprintf(buf, sizeof(buf), "%s SATA controller",
+			    mvs_ids[i].name);
+			device_set_desc_copy(dev, buf);
+			return (BUS_PROBE_VENDOR);
+		}
+	}
+	return (ENXIO);
+}
+
+static int
+mvs_attach(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	device_t child;
+	int	error, unit, i;
+	uint32_t devid = pci_get_devid(dev);
+	uint8_t revid = pci_get_revid(dev);
+
+	ctlr->dev = dev;
+	i = 0;
+	while (mvs_ids[i].id != 0 &&
+	    (mvs_ids[i].id != devid ||
+	     mvs_ids[i].rev > revid))
+		i++;
+	ctlr->channels = mvs_ids[i].ports;
+	ctlr->quirks = mvs_ids[i].quirks;
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "ccc", &ctlr->ccc);
+	ctlr->cccc = 8;
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "cccc", &ctlr->cccc);
+	if (ctlr->ccc == 0 || ctlr->cccc == 0) {
+		ctlr->ccc = 0;
+		ctlr->cccc = 0;
+	}
+	if (ctlr->ccc > 100000)
+		ctlr->ccc = 100000;
+	device_printf(dev,
+	    "Gen-%s, %d %sGbps ports, Port Multiplier %s%s\n",
+	    ((ctlr->quirks & MVS_Q_GENI) ? "I" :
+	     ((ctlr->quirks & MVS_Q_GENII) ? "II" : "IIe")),
+	    ctlr->channels,
+	    ((ctlr->quirks & MVS_Q_GENI) ? "1.5" : "3"),
+	    ((ctlr->quirks & MVS_Q_GENI) ?
+	    "not supported" : "supported"),
+	    ((ctlr->quirks & MVS_Q_GENIIE) ?
+	    " with FBS" : ""));
+	mtx_init(&ctlr->mtx, "MVS controller lock", NULL, MTX_DEF);
+	/* We should have a memory BAR(0). */
+	ctlr->r_rid = PCIR_BAR(0);
+	if (!(ctlr->r_mem = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
+	    &ctlr->r_rid, RF_ACTIVE)))
+		return ENXIO;
+	/* Setup our own memory management for channels. */
+	ctlr->sc_iomem.rm_type = RMAN_ARRAY;
+	ctlr->sc_iomem.rm_descr = "I/O memory addresses";
+	if ((error = rman_init(&ctlr->sc_iomem)) != 0) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		return (error);
+	}
+	if ((error = rman_manage_region(&ctlr->sc_iomem,
+	    rman_get_start(ctlr->r_mem), rman_get_end(ctlr->r_mem))) != 0) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		rman_fini(&ctlr->sc_iomem);
+		return (error);
+	}
+	pci_enable_busmaster(dev);
+	mvs_ctlr_setup(dev);
+	/* Setup interrupts. */
+	if (mvs_setup_interrupt(dev)) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		rman_fini(&ctlr->sc_iomem);
+		return ENXIO;
+	}
+	/* Attach all channels on this controller */
+	for (unit = 0; unit < ctlr->channels; unit++) {
+		child = device_add_child(dev, "mvsch", -1);
+		if (child == NULL)
+			device_printf(dev, "failed to add channel device\n");
+		else
+			device_set_ivars(child, (void *)(intptr_t)unit);
+	}
+	bus_generic_attach(dev);
+	return 0;
+}
+
+static int
+mvs_detach(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	device_t *children;
+	int nchildren, i;
+
+	/* Detach & delete all children */
+	if (!device_get_children(dev, &children, &nchildren)) {
+		for (i = 0; i < nchildren; i++)
+			device_delete_child(dev, children[i]);
+		free(children, M_TEMP);
+	}
+	/* Free interrupt. */
+	if (ctlr->irq.r_irq) {
+		bus_teardown_intr(dev, ctlr->irq.r_irq,
+		    ctlr->irq.handle);
+		bus_release_resource(dev, SYS_RES_IRQ,
+		    ctlr->irq.r_irq_rid, ctlr->irq.r_irq);
+	}
+	pci_release_msi(dev);
+	/* Free memory. */
+	rman_fini(&ctlr->sc_iomem);
+	if (ctlr->r_mem)
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+	mtx_destroy(&ctlr->mtx);
+	return (0);
+}
+
+static int
+mvs_ctlr_setup(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int i, ccc = ctlr->ccc, cccc = ctlr->cccc, ccim = 0;
+
+	/* Mask chip interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_MIM, 0x00000000);
+	/* Mask PCI interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_PCIIM, 0x00000000);
+	/* Clear PCI interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_PCIIC, 0x00000000);
+	if (ccc && bootverbose) {
+		device_printf(dev,
+		    "CCC with %dus/%dcmd enabled\n",
+		    ctlr->ccc, ctlr->cccc);
+	}
+	ccc *= 150;
+	/* Configure chip-global CCC */
+	if (ctlr->channels > 4 && (ctlr->quirks & MVS_Q_GENI) == 0) {
+		ATA_OUTL(ctlr->r_mem, CHIP_ICT, cccc);
+		ATA_OUTL(ctlr->r_mem, CHIP_ITT, ccc);
+		ATA_OUTL(ctlr->r_mem, CHIP_ICC, ~CHIP_ICC_ALL_PORTS);
+		if (ccc)
+			ccim |= IC_ALL_PORTS_COAL_DONE;
+		ccc = 0;
+		cccc = 0;
+	}
+	for (i = 0; i < ctlr->channels / 4; i++) {
+		/* Configure per-HC CCC */
+		ATA_OUTL(ctlr->r_mem, HC_BASE(i) + HC_ICT, cccc);
+		ATA_OUTL(ctlr->r_mem, HC_BASE(i) + HC_ITT, ccc);
+		if (ccc)
+			ccim |= (IC_HC0_COAL_DONE << (i * IC_HC_SHIFT));
+		/* Clear HC interrupts */
+		ATA_OUTL(ctlr->r_mem, HC_BASE(i) + HC_IC, 0x00000000);
+	}
+	/* Enable chip interrupts */
+	ctlr->gmim = (ccim ? ccim : (IC_DONE_HC0 | IC_DONE_HC1)) |
+	     IC_ERR_HC0 | IC_ERR_HC1;
+	ctlr->mim = ctlr->gmim | ctlr->pmim;
+	ATA_OUTL(ctlr->r_mem, CHIP_MIM, ctlr->mim);
+	/* Enable PCI interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_PCIIM, 0x007fffff);
+	return (0);
+}
+
+static void
+mvs_edma(device_t dev, device_t child, int mode)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = ((struct mvs_channel *)device_get_softc(child))->unit;
+	int bit = IC_DONE_IRQ << (unit * 2 + unit / 4) ;
+
+	if (ctlr->ccc == 0)
+		return;
+	/* CCC is not working for non-EDMA mode. Unmask device interrupts. */
+	mtx_lock(&ctlr->mtx);
+	if (mode == MVS_EDMA_OFF)
+		ctlr->pmim |= bit;
+	else
+		ctlr->pmim &= ~bit;
+	ctlr->mim = ctlr->gmim | ctlr->pmim;
+	if (!ctlr->msia)
+		ATA_OUTL(ctlr->r_mem, CHIP_MIM, ctlr->mim);
+	mtx_unlock(&ctlr->mtx);
+}
+
+static int
+mvs_suspend(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+
+	bus_generic_suspend(dev);
+	/* Mask chip interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_MIM, 0x00000000);
+	/* Mask PCI interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_PCIIM, 0x00000000);
+	return 0;
+}
+
+static int
+mvs_resume(device_t dev)
+{
+
+	mvs_ctlr_setup(dev);
+	return (bus_generic_resume(dev));
+}
+
+static int
+mvs_setup_interrupt(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int msi = 0;
+
+	/* Process hints. */
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "msi", &msi);
+	if (msi < 0)
+		msi = 0;
+	else if (msi > 0)
+		msi = min(1, pci_msi_count(dev));
+	/* Allocate MSI if needed/present. */
+	if (msi && pci_alloc_msi(dev, &msi) != 0)
+		msi = 0;
+	ctlr->msi = msi;
+	/* Allocate all IRQs. */
+	ctlr->irq.r_irq_rid = msi ? 1 : 0;
+	if (!(ctlr->irq.r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ,
+	    &ctlr->irq.r_irq_rid, RF_SHAREABLE | RF_ACTIVE))) {
+		device_printf(dev, "unable to map interrupt\n");
+		return (ENXIO);
+	}
+	if ((bus_setup_intr(dev, ctlr->irq.r_irq, ATA_INTR_FLAGS, NULL,
+	    mvs_intr, ctlr, &ctlr->irq.handle))) {
+		device_printf(dev, "unable to setup interrupt\n");
+		bus_release_resource(dev, SYS_RES_IRQ,
+		    ctlr->irq.r_irq_rid, ctlr->irq.r_irq);
+		ctlr->irq.r_irq = 0;
+		return (ENXIO);
+	}
+	return (0);
+}
+
+/*
+ * Common case interrupt handler.
+ */
+static void
+mvs_intr(void *data)
+{
+	struct mvs_controller *ctlr = data;
+	struct mvs_intr_arg arg;
+	void (*function)(void *);
+	int p;
+	u_int32_t ic, aic;
+
+	ic = ATA_INL(ctlr->r_mem, CHIP_MIC);
+//device_printf(ctlr->dev, "irq MIC:%08x\n", ic);
+	if (ctlr->msi) {
+		/* We have to to mask MSI during processing. */
+		mtx_lock(&ctlr->mtx);
+		ATA_OUTL(ctlr->r_mem, CHIP_MIM, 0);
+		ctlr->msia = 1; /* Deny MIM update during processing. */
+		mtx_unlock(&ctlr->mtx);
+	} else if (ic == 0)
+		return;
+	/* Acknowledge all-ports CCC interrupt. */
+	if (ic & IC_ALL_PORTS_COAL_DONE)
+		ATA_OUTL(ctlr->r_mem, CHIP_ICC, ~CHIP_ICC_ALL_PORTS);
+	for (p = 0; p < ctlr->channels; p++) {
+		if ((p & 3) == 0) {
+			if (p != 0)
+				ic >>= 1;
+			if ((ic & IC_HC0) == 0) {
+				p += 3;
+				ic >>= 8;
+				continue;
+			}
+			/* Acknowledge interrupts of this HC. */
+			aic = 0;
+			if (ic & (IC_DONE_IRQ << 0))
+				aic |= HC_IC_DONE(0) | HC_IC_DEV(0);
+			if (ic & (IC_DONE_IRQ << 2))
+				aic |= HC_IC_DONE(1) | HC_IC_DEV(1);
+			if (ic & (IC_DONE_IRQ << 4))
+				aic |= HC_IC_DONE(2) | HC_IC_DEV(2);
+			if (ic & (IC_DONE_IRQ << 6))
+				aic |= HC_IC_DONE(3) | HC_IC_DEV(3);
+			if (ic & IC_HC0_COAL_DONE)
+				aic |= HC_IC_COAL;
+			ATA_OUTL(ctlr->r_mem, HC_BASE(p == 4) + HC_IC, ~aic);
+		}
+		/* Call per-port interrupt handler. */
+		arg.cause = ic & (IC_ERR_IRQ|IC_DONE_IRQ);
+		if ((arg.cause != 0) &&
+		    (function = ctlr->interrupt[p].function)) {
+			arg.arg = ctlr->interrupt[p].argument;
+			function(&arg);
+		}
+		ic >>= 2;
+	}
+	if (ctlr->msi) {
+		/* Unmasking MSI triggers next interrupt, if needed. */
+		mtx_lock(&ctlr->mtx);
+		ctlr->msia = 0;	/* Allow MIM update. */
+		ATA_OUTL(ctlr->r_mem, CHIP_MIM, ctlr->mim);
+		mtx_unlock(&ctlr->mtx);
+	}
+}
+
+static struct resource *
+mvs_alloc_resource(device_t dev, device_t child, int type, int *rid,
+		       u_long start, u_long end, u_long count, u_int flags)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = ((struct mvs_channel *)device_get_softc(child))->unit;
+	struct resource *res = NULL;
+	int offset = HC_BASE(unit >> 2) + PORT_BASE(unit & 0x03);
+	long st;
+
+	switch (type) {
+	case SYS_RES_MEMORY:
+		st = rman_get_start(ctlr->r_mem);
+		res = rman_reserve_resource(&ctlr->sc_iomem, st + offset,
+		    st + offset + PORT_SIZE - 1, PORT_SIZE, RF_ACTIVE, child);
+		if (res) {
+			bus_space_handle_t bsh;
+			bus_space_tag_t bst;
+			bsh = rman_get_bushandle(ctlr->r_mem);
+			bst = rman_get_bustag(ctlr->r_mem);
+			bus_space_subregion(bst, bsh, offset, PORT_SIZE, &bsh);
+			rman_set_bushandle(res, bsh);
+			rman_set_bustag(res, bst);
+		}
+		break;
+	case SYS_RES_IRQ:
+		if (*rid == ATA_IRQ_RID)
+			res = ctlr->irq.r_irq;
+		break;
+	}
+	return (res);
+}
+
+static int
+mvs_release_resource(device_t dev, device_t child, int type, int rid,
+			 struct resource *r)
+{
+
+	switch (type) {
+	case SYS_RES_MEMORY:
+		rman_release_resource(r);
+		return (0);
+	case SYS_RES_IRQ:
+		if (rid != ATA_IRQ_RID)
+			return ENOENT;
+		return (0);
+	}
+	return (EINVAL);
+}
+
+static int
+mvs_setup_intr(device_t dev, device_t child, struct resource *irq, 
+		   int flags, driver_filter_t *filter, driver_intr_t *function, 
+		   void *argument, void **cookiep)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = (intptr_t)device_get_ivars(child);
+
+	if (filter != NULL) {
+		printf("mvs.c: we cannot use a filter here\n");
+		return (EINVAL);
+	}
+	ctlr->interrupt[unit].function = function;
+	ctlr->interrupt[unit].argument = argument;
+	return (0);
+}
+
+static int
+mvs_teardown_intr(device_t dev, device_t child, struct resource *irq,
+		      void *cookie)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = (intptr_t)device_get_ivars(child);
+
+	ctlr->interrupt[unit].function = NULL;
+	ctlr->interrupt[unit].argument = NULL;
+	return (0);
+}
+
+static int
+mvs_print_child(device_t dev, device_t child)
+{
+	int retval;
+
+	retval = bus_print_child_header(dev, child);
+	retval += printf(" at channel %d",
+	    (int)(intptr_t)device_get_ivars(child));
+	retval += bus_print_child_footer(dev, child);
+
+	return (retval);
+}
+
+static device_method_t mvs_methods[] = {
+	DEVMETHOD(device_probe,     mvs_probe),
+	DEVMETHOD(device_attach,    mvs_attach),
+	DEVMETHOD(device_detach,    mvs_detach),
+	DEVMETHOD(device_suspend,   mvs_suspend),
+	DEVMETHOD(device_resume,    mvs_resume),
+	DEVMETHOD(bus_print_child,  mvs_print_child),
+	DEVMETHOD(bus_alloc_resource,       mvs_alloc_resource),
+	DEVMETHOD(bus_release_resource,     mvs_release_resource),
+	DEVMETHOD(bus_setup_intr,   mvs_setup_intr),
+	DEVMETHOD(bus_teardown_intr,mvs_teardown_intr),
+	DEVMETHOD(mvs_edma,         mvs_edma),
+	{ 0, 0 }
+};
+static driver_t mvs_driver = {
+        "mvs",
+        mvs_methods,
+        sizeof(struct mvs_controller)
+};
+DRIVER_MODULE(mvs, pci, mvs_driver, mvs_devclass, 0, 0);
+MODULE_VERSION(mvs, 1);
+MODULE_DEPEND(mvs, cam, 1, 1, 1);
+
diff --git a/sys/dev/mvs/mvs_soc.c b/sys/dev/mvs/mvs_soc.c
new file mode 100644
index 00000000000..298a87336a7
--- /dev/null
+++ b/sys/dev/mvs/mvs_soc.c
@@ -0,0 +1,437 @@
+/*-
+ * Copyright (c) 2010 Alexander Motin <mav@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer,
+ *    without modification, immediately at the beginning of the file.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/module.h>
+#include <sys/systm.h>
+#include <sys/kernel.h>
+#include <sys/bus.h>
+#include <sys/endian.h>
+#include <sys/malloc.h>
+#include <sys/lock.h>
+#include <sys/mutex.h>
+#include <vm/uma.h>
+#include <machine/stdarg.h>
+#include <machine/resource.h>
+#include <machine/bus.h>
+#include <sys/rman.h>
+#include <arm/mv/mvreg.h>
+#include <arm/mv/mvvar.h>
+#include "mvs.h"
+
+/* local prototypes */
+static int mvs_setup_interrupt(device_t dev);
+static void mvs_intr(void *data);
+static int mvs_suspend(device_t dev);
+static int mvs_resume(device_t dev);
+static int mvs_ctlr_setup(device_t dev);
+
+static struct {
+	uint32_t	id;
+	uint8_t		rev;
+	const char	*name;
+	int		ports;
+	int		quirks;
+} mvs_ids[] = {
+	{MV_DEV_88F5182, 0x00,   "Marvell 88F5182",	2, MVS_Q_GENIIE|MVS_Q_SOC},
+	{MV_DEV_88F6281, 0x00,   "Marvell 88F6281",	2, MVS_Q_GENIIE|MVS_Q_SOC},
+	{MV_DEV_MV78100, 0x00,   "Marvell MV78100",	2, MVS_Q_GENIIE|MVS_Q_SOC},
+	{MV_DEV_MV78100_Z0, 0x00,"Marvell MV78100",	2, MVS_Q_GENIIE|MVS_Q_SOC},
+	{0,              0x00,   NULL,			0, 0}
+};
+
+static int
+mvs_probe(device_t dev)
+{
+	char buf[64];
+	int i;
+	uint32_t devid, revid;
+
+	soc_id(&devid, &revid);
+	for (i = 0; mvs_ids[i].id != 0; i++) {
+		if (mvs_ids[i].id == devid &&
+		    mvs_ids[i].rev <= revid) {
+			snprintf(buf, sizeof(buf), "%s SATA controller",
+			    mvs_ids[i].name);
+			device_set_desc_copy(dev, buf);
+			return (BUS_PROBE_VENDOR);
+		}
+	}
+	return (ENXIO);
+}
+
+static int
+mvs_attach(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	device_t child;
+	int	error, unit, i;
+	uint32_t devid, revid;
+
+	soc_id(&devid, &revid);
+	ctlr->dev = dev;
+	i = 0;
+	while (mvs_ids[i].id != 0 &&
+	    (mvs_ids[i].id != devid ||
+	     mvs_ids[i].rev > revid))
+		i++;
+	ctlr->channels = mvs_ids[i].ports;
+	ctlr->quirks = mvs_ids[i].quirks;
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "ccc", &ctlr->ccc);
+	ctlr->cccc = 8;
+	resource_int_value(device_get_name(dev),
+	    device_get_unit(dev), "cccc", &ctlr->cccc);
+	if (ctlr->ccc == 0 || ctlr->cccc == 0) {
+		ctlr->ccc = 0;
+		ctlr->cccc = 0;
+	}
+	if (ctlr->ccc > 100000)
+		ctlr->ccc = 100000;
+	device_printf(dev,
+	    "Gen-%s, %d %sGbps ports, Port Multiplier %s%s\n",
+	    ((ctlr->quirks & MVS_Q_GENI) ? "I" :
+	     ((ctlr->quirks & MVS_Q_GENII) ? "II" : "IIe")),
+	    ctlr->channels,
+	    ((ctlr->quirks & MVS_Q_GENI) ? "1.5" : "3"),
+	    ((ctlr->quirks & MVS_Q_GENI) ?
+	    "not supported" : "supported"),
+	    ((ctlr->quirks & MVS_Q_GENIIE) ?
+	    " with FBS" : ""));
+	mtx_init(&ctlr->mtx, "MVS controller lock", NULL, MTX_DEF);
+	/* We should have a memory BAR(0). */
+	ctlr->r_rid = 0;
+	if (!(ctlr->r_mem = bus_alloc_resource_any(dev, SYS_RES_MEMORY,
+	    &ctlr->r_rid, RF_ACTIVE)))
+		return ENXIO;
+	/* Setup our own memory management for channels. */
+	ctlr->sc_iomem.rm_type = RMAN_ARRAY;
+	ctlr->sc_iomem.rm_descr = "I/O memory addresses";
+	if ((error = rman_init(&ctlr->sc_iomem)) != 0) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		return (error);
+	}
+	if ((error = rman_manage_region(&ctlr->sc_iomem,
+	    rman_get_start(ctlr->r_mem), rman_get_end(ctlr->r_mem))) != 0) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		rman_fini(&ctlr->sc_iomem);
+		return (error);
+	}
+	mvs_ctlr_setup(dev);
+	/* Setup interrupts. */
+	if (mvs_setup_interrupt(dev)) {
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+		rman_fini(&ctlr->sc_iomem);
+		return ENXIO;
+	}
+	/* Attach all channels on this controller */
+	for (unit = 0; unit < ctlr->channels; unit++) {
+		child = device_add_child(dev, "mvsch", -1);
+		if (child == NULL)
+			device_printf(dev, "failed to add channel device\n");
+		else
+			device_set_ivars(child, (void *)(intptr_t)unit);
+	}
+	bus_generic_attach(dev);
+	return 0;
+}
+
+static int
+mvs_detach(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	device_t *children;
+	int nchildren, i;
+
+	/* Detach & delete all children */
+	if (!device_get_children(dev, &children, &nchildren)) {
+		for (i = 0; i < nchildren; i++)
+			device_delete_child(dev, children[i]);
+		free(children, M_TEMP);
+	}
+	/* Free interrupt. */
+	if (ctlr->irq.r_irq) {
+		bus_teardown_intr(dev, ctlr->irq.r_irq,
+		    ctlr->irq.handle);
+		bus_release_resource(dev, SYS_RES_IRQ,
+		    ctlr->irq.r_irq_rid, ctlr->irq.r_irq);
+	}
+	/* Free memory. */
+	rman_fini(&ctlr->sc_iomem);
+	if (ctlr->r_mem)
+		bus_release_resource(dev, SYS_RES_MEMORY, ctlr->r_rid, ctlr->r_mem);
+	mtx_destroy(&ctlr->mtx);
+	return (0);
+}
+
+static int
+mvs_ctlr_setup(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int ccc = ctlr->ccc, cccc = ctlr->cccc, ccim = 0;
+
+	/* Mask chip interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_SOC_MIM, 0x00000000);
+	/* Clear HC interrupts */
+	ATA_OUTL(ctlr->r_mem, HC_IC, 0x00000000);
+	/* Clear chip interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_SOC_MIC, 0);
+	/* Configure per-HC CCC */
+	if (ccc && bootverbose) {
+		device_printf(dev,
+		    "CCC with %dus/%dcmd enabled\n",
+		    ctlr->ccc, ctlr->cccc);
+	}
+	ccc *= 150;
+	ATA_OUTL(ctlr->r_mem, HC_ICT, cccc);
+	ATA_OUTL(ctlr->r_mem, HC_ITT, ccc);
+	if (ccc)
+		ccim |= IC_HC0_COAL_DONE;
+	/* Enable chip interrupts */
+	ctlr->gmim = (ccc ? IC_HC0_COAL_DONE : IC_DONE_HC0) | IC_ERR_HC0;
+	ATA_OUTL(ctlr->r_mem, CHIP_SOC_MIM, ctlr->gmim | ctlr->pmim);
+	return (0);
+}
+
+static void
+mvs_edma(device_t dev, device_t child, int mode)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = ((struct mvs_channel *)device_get_softc(child))->unit;
+	int bit = IC_DONE_IRQ << (unit * 2);
+
+	if (ctlr->ccc == 0)
+		return;
+	/* CCC is not working for non-EDMA mode. Unmask device interrupts. */
+	mtx_lock(&ctlr->mtx);
+	if (mode == MVS_EDMA_OFF)
+		ctlr->pmim |= bit;
+	else
+		ctlr->pmim &= ~bit;
+	ATA_OUTL(ctlr->r_mem, CHIP_SOC_MIM, ctlr->gmim | ctlr->pmim);
+	mtx_unlock(&ctlr->mtx);
+}
+
+static int
+mvs_suspend(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+
+	bus_generic_suspend(dev);
+	/* Mask chip interrupts */
+	ATA_OUTL(ctlr->r_mem, CHIP_SOC_MIM, 0x00000000);
+	return 0;
+}
+
+static int
+mvs_resume(device_t dev)
+{
+
+	mvs_ctlr_setup(dev);
+	return (bus_generic_resume(dev));
+}
+
+static int
+mvs_setup_interrupt(device_t dev)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+
+	/* Allocate all IRQs. */
+	ctlr->irq.r_irq_rid = 0;
+	if (!(ctlr->irq.r_irq = bus_alloc_resource_any(dev, SYS_RES_IRQ,
+	    &ctlr->irq.r_irq_rid, RF_SHAREABLE | RF_ACTIVE))) {
+		device_printf(dev, "unable to map interrupt\n");
+		return (ENXIO);
+	}
+	if ((bus_setup_intr(dev, ctlr->irq.r_irq, ATA_INTR_FLAGS, NULL,
+	    mvs_intr, ctlr, &ctlr->irq.handle))) {
+		device_printf(dev, "unable to setup interrupt\n");
+		bus_release_resource(dev, SYS_RES_IRQ,
+		    ctlr->irq.r_irq_rid, ctlr->irq.r_irq);
+		ctlr->irq.r_irq = 0;
+		return (ENXIO);
+	}
+	return (0);
+}
+
+/*
+ * Common case interrupt handler.
+ */
+static void
+mvs_intr(void *data)
+{
+	struct mvs_controller *ctlr = data;
+	struct mvs_intr_arg arg;
+	void (*function)(void *);
+	int p;
+	u_int32_t ic, aic;
+
+	ic = ATA_INL(ctlr->r_mem, CHIP_SOC_MIC);
+//device_printf(ctlr->dev, "irq MIC:%08x\n", ic);
+	if ((ic & IC_HC0) == 0)
+		return;
+	/* Acknowledge interrupts of this HC. */
+	aic = 0;
+	if (ic & (IC_DONE_IRQ << 0))
+		aic |= HC_IC_DONE(0) | HC_IC_DEV(0);
+	if (ic & (IC_DONE_IRQ << 2))
+		aic |= HC_IC_DONE(1) | HC_IC_DEV(1);
+	if (ic & (IC_DONE_IRQ << 4))
+		aic |= HC_IC_DONE(2) | HC_IC_DEV(2);
+	if (ic & (IC_DONE_IRQ << 6))
+		aic |= HC_IC_DONE(3) | HC_IC_DEV(3);
+	if (ic & IC_HC0_COAL_DONE)
+		aic |= HC_IC_COAL;
+	ATA_OUTL(ctlr->r_mem, HC_IC, ~aic);
+	/* Call per-port interrupt handler. */
+	for (p = 0; p < ctlr->channels; p++) {
+		arg.cause = ic & (IC_ERR_IRQ|IC_DONE_IRQ);
+		if ((arg.cause != 0) &&
+		    (function = ctlr->interrupt[p].function)) {
+			arg.arg = ctlr->interrupt[p].argument;
+			function(&arg);
+		}
+		ic >>= 2;
+	}
+}
+
+static struct resource *
+mvs_alloc_resource(device_t dev, device_t child, int type, int *rid,
+		       u_long start, u_long end, u_long count, u_int flags)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = ((struct mvs_channel *)device_get_softc(child))->unit;
+	struct resource *res = NULL;
+	int offset = PORT_BASE(unit & 0x03);
+	long st;
+
+	switch (type) {
+	case SYS_RES_MEMORY:
+		st = rman_get_start(ctlr->r_mem);
+		res = rman_reserve_resource(&ctlr->sc_iomem, st + offset,
+		    st + offset + PORT_SIZE - 1, PORT_SIZE, RF_ACTIVE, child);
+		if (res) {
+			bus_space_handle_t bsh;
+			bus_space_tag_t bst;
+			bsh = rman_get_bushandle(ctlr->r_mem);
+			bst = rman_get_bustag(ctlr->r_mem);
+			bus_space_subregion(bst, bsh, offset, PORT_SIZE, &bsh);
+			rman_set_bushandle(res, bsh);
+			rman_set_bustag(res, bst);
+		}
+		break;
+	case SYS_RES_IRQ:
+		if (*rid == ATA_IRQ_RID)
+			res = ctlr->irq.r_irq;
+		break;
+	}
+	return (res);
+}
+
+static int
+mvs_release_resource(device_t dev, device_t child, int type, int rid,
+			 struct resource *r)
+{
+
+	switch (type) {
+	case SYS_RES_MEMORY:
+		rman_release_resource(r);
+		return (0);
+	case SYS_RES_IRQ:
+		if (rid != ATA_IRQ_RID)
+			return ENOENT;
+		return (0);
+	}
+	return (EINVAL);
+}
+
+static int
+mvs_setup_intr(device_t dev, device_t child, struct resource *irq, 
+		   int flags, driver_filter_t *filter, driver_intr_t *function, 
+		   void *argument, void **cookiep)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = (intptr_t)device_get_ivars(child);
+
+	if (filter != NULL) {
+		printf("mvs.c: we cannot use a filter here\n");
+		return (EINVAL);
+	}
+	ctlr->interrupt[unit].function = function;
+	ctlr->interrupt[unit].argument = argument;
+	return (0);
+}
+
+static int
+mvs_teardown_intr(device_t dev, device_t child, struct resource *irq,
+		      void *cookie)
+{
+	struct mvs_controller *ctlr = device_get_softc(dev);
+	int unit = (intptr_t)device_get_ivars(child);
+
+	ctlr->interrupt[unit].function = NULL;
+	ctlr->interrupt[unit].argument = NULL;
+	return (0);
+}
+
+static int
+mvs_print_child(device_t dev, device_t child)
+{
+	int retval;
+
+	retval = bus_print_child_header(dev, child);
+	retval += printf(" at channel %d",
+	    (int)(intptr_t)device_get_ivars(child));
+	retval += bus_print_child_footer(dev, child);
+
+	return (retval);
+}
+
+static device_method_t mvs_methods[] = {
+	DEVMETHOD(device_probe,     mvs_probe),
+	DEVMETHOD(device_attach,    mvs_attach),
+	DEVMETHOD(device_detach,    mvs_detach),
+	DEVMETHOD(device_suspend,   mvs_suspend),
+	DEVMETHOD(device_resume,    mvs_resume),
+	DEVMETHOD(bus_print_child,  mvs_print_child),
+	DEVMETHOD(bus_alloc_resource,       mvs_alloc_resource),
+	DEVMETHOD(bus_release_resource,     mvs_release_resource),
+	DEVMETHOD(bus_setup_intr,   mvs_setup_intr),
+	DEVMETHOD(bus_teardown_intr,mvs_teardown_intr),
+	DEVMETHOD(mvs_edma,         mvs_edma),
+	{ 0, 0 }
+};
+static driver_t mvs_driver = {
+        "sata",
+        mvs_methods,
+        sizeof(struct mvs_controller)
+};
+DRIVER_MODULE(sata, mbus, mvs_driver, mvs_devclass, 0, 0);
+MODULE_VERSION(sata, 1);
+
diff --git a/sys/modules/Makefile b/sys/modules/Makefile
index 83c7757c7e2..601b73be9bb 100644
--- a/sys/modules/Makefile
+++ b/sys/modules/Makefile
@@ -188,6 +188,7 @@ SUBDIR=	${_3dfx} \
 	msdosfs_iconv \
 	${_mse} \
 	msk \
+	mvs \
 	mwl \
 	mxge \
 	my \
diff --git a/sys/modules/mvs/Makefile b/sys/modules/mvs/Makefile
new file mode 100644
index 00000000000..f33a0927858
--- /dev/null
+++ b/sys/modules/mvs/Makefile
@@ -0,0 +1,10 @@
+# $FreeBSD$
+
+.PATH: ${.CURDIR}/../../dev/mvs
+
+KMOD=	mvs
+SRCS=	mvs.c mvs_pci.c mvs.h mvs_if.h device_if.h bus_if.h pci_if.h opt_cam.h
+
+MFILES=	kern/bus_if.m kern/device_if.m dev/pci/pci_if.m dev/mvs/mvs_if.m
+
+.include <bsd.kmod.mk>

From 5a8336816e21dbfaae43f22578b4f676a3c0358e Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Sun, 2 May 2010 19:38:17 +0000
Subject: [PATCH 405/532] Add support for SPARC64 V (and where it already makes
 sense for other HAL/Fujitsu) CPUs. For the most part this consists of
 fleshing out the MMU and cache handling, it doesn't add pmap optimizations
 possible with these CPU, yet, though. With these changes FreeBSD runs stable
 on Fujitsu Siemens PRIMEPOWER 250 and likely also other models based on
 SPARC64 V like 450, 650 and 850. Thanks go to Michael Moll for providing
 access to a PRIMEPOWER 250.

---
 sys/boot/sparc64/loader/main.c   |  8 +++-
 sys/conf/files.sparc64           |  1 +
 sys/sparc64/include/asi.h        |  7 +++-
 sys/sparc64/include/cache.h      |  4 ++
 sys/sparc64/include/mcntl.h      | 62 ++++++++++++++++++++++++++++++
 sys/sparc64/sparc64/cache.c      | 20 ++++++++--
 sys/sparc64/sparc64/cheetah.c    | 15 +++++++-
 sys/sparc64/sparc64/identcpu.c   | 24 +++++++++++-
 sys/sparc64/sparc64/machdep.c    |  6 ++-
 sys/sparc64/sparc64/mp_locore.S  | 22 +++++++----
 sys/sparc64/sparc64/mp_machdep.c |  9 +++--
 sys/sparc64/sparc64/pmap.c       |  9 +++--
 sys/sparc64/sparc64/tick.c       |  9 +++--
 sys/sparc64/sparc64/zeus.c       | 65 ++++++++++++++++++++++++++++++++
 14 files changed, 232 insertions(+), 29 deletions(-)
 create mode 100644 sys/sparc64/include/mcntl.h
 create mode 100644 sys/sparc64/sparc64/zeus.c

diff --git a/sys/boot/sparc64/loader/main.c b/sys/boot/sparc64/loader/main.c
index 2afbfeeb75e..b3e79084ad3 100644
--- a/sys/boot/sparc64/loader/main.c
+++ b/sys/boot/sparc64/loader/main.c
@@ -451,7 +451,8 @@ dtlb_va_to_pa_sun4u(vm_offset_t va)
 		reg = dtlb_get_data_sun4u(i);
 		wrpr(pstate, pstate, 0);
 		reg >>= TD_PA_SHIFT;
-		if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+		if (cpu_impl == CPU_IMPL_SPARC64V ||
+		    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 			return (reg & TD_PA_CH_MASK);
 		return (reg & TD_PA_SF_MASK);
 	}
@@ -474,7 +475,8 @@ itlb_va_to_pa_sun4u(vm_offset_t va)
 		reg = itlb_get_data_sun4u(i);
 		wrpr(pstate, pstate, 0);
 		reg >>= TD_PA_SHIFT;
-		if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+		if (cpu_impl == CPU_IMPL_SPARC64V ||
+		    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 			return (reg & TD_PA_CH_MASK);
 		return (reg & TD_PA_SF_MASK);
 	}
@@ -696,6 +698,7 @@ cpu_cpuid_prop_sun4u(void)
 
 	switch (cpu_impl) {
 	case CPU_IMPL_SPARC64:
+	case CPU_IMPL_SPARC64V:
 	case CPU_IMPL_ULTRASPARCI:
 	case CPU_IMPL_ULTRASPARCII:
 	case CPU_IMPL_ULTRASPARCIIi:
@@ -720,6 +723,7 @@ cpu_get_mid_sun4u(void)
 
 	switch (cpu_impl) {
 	case CPU_IMPL_SPARC64:
+	case CPU_IMPL_SPARC64V:
 	case CPU_IMPL_ULTRASPARCI:
 	case CPU_IMPL_ULTRASPARCII:
 	case CPU_IMPL_ULTRASPARCIIi:
diff --git a/sys/conf/files.sparc64 b/sys/conf/files.sparc64
index 63ee2c5be7f..3a0620a2d09 100644
--- a/sys/conf/files.sparc64
+++ b/sys/conf/files.sparc64
@@ -138,3 +138,4 @@ sparc64/sparc64/tsb.c		standard
 sparc64/sparc64/uio_machdep.c	standard
 sparc64/sparc64/upa.c		optional	creator
 sparc64/sparc64/vm_machdep.c	standard
+sparc64/sparc64/zeus.c		standard
diff --git a/sys/sparc64/include/asi.h b/sys/sparc64/include/asi.h
index 90dc3f394e1..e5769c125bd 100644
--- a/sys/sparc64/include/asi.h
+++ b/sys/sparc64/include/asi.h
@@ -82,7 +82,10 @@
 #define	ASI_DCACHE_SNOOP_TAG			0x44	/* US-III Cu */
 
 /* Named ASI_DCUCR on US-III, but is mostly identical except for added bits. */
-#define	ASI_LSU_CTL_REG				0x45
+#define	ASI_LSU_CTL_REG				0x45	/* US only */
+
+#define	ASI_MCNTL				0x45	/* SPARC64 only */
+#define		AA_MCNTL			0x08
 
 #define	ASI_DCACHE_DATA				0x46
 #define	ASI_DCACHE_TAG				0x47
@@ -167,6 +170,8 @@
 #define	ASI_ICACHE_PRE_DECODE			0x6e	/* US-I, II */
 #define	ASI_ICACHE_PRE_NEXT_FIELD		0x6f	/* US-I, II */
 
+#define	ASI_FLUSH_L1I				0x67	/* SPARC64 only */
+
 #define	ASI_BLK_AUIP				0x70
 #define	ASI_BLK_AIUS				0x71
 
diff --git a/sys/sparc64/include/cache.h b/sys/sparc64/include/cache.h
index 9c4804b9776..29bff69d937 100644
--- a/sys/sparc64/include/cache.h
+++ b/sys/sparc64/include/cache.h
@@ -113,6 +113,10 @@ extern cache_flush_t *cache_flush;
 extern dcache_page_inval_t *dcache_page_inval;
 extern icache_page_inval_t *icache_page_inval;
 
+cache_flush_t zeus_cache_flush;
+dcache_page_inval_t zeus_dcache_page_inval;
+icache_page_inval_t zeus_icache_page_inval;
+
 #endif /* KERNEL */
 
 #endif /* !LOCORE */
diff --git a/sys/sparc64/include/mcntl.h b/sys/sparc64/include/mcntl.h
new file mode 100644
index 00000000000..e9483c05477
--- /dev/null
+++ b/sys/sparc64/include/mcntl.h
@@ -0,0 +1,62 @@
+/*-
+ * Copyright (c) 2010 Marius Strobl <marius@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef	_MACHINE_MCNTL_H
+#define	_MACHINE_MCNTL_H
+
+/*
+ * Definitions for the SPARC64 V, VI, VII and VIIIfx Memory Control Register
+ */
+#define	MCNTL_JPS1_TSBP		(1UL << 8)
+
+#define	MCNTL_RMD_SHIFT		12
+#define	MCNTL_RMD_BITS		2
+#define	MCNTL_RMD_MASK							\
+	(((1UL << MCNTL_RMD_BITS) - 1) << MCNTL_RMD_SHIFT)
+#define	MCNTL_RMD_FULL		(0UL << MCNTL_RMD_SHIFT)
+#define	MCNTL_RMD_1024		(2UL << MCNTL_RMD_SHIFT)
+#define	MCNTL_RMD_512		(3UL << MCNTL_RMD_SHIFT)
+
+#define	MCNTL_FW_FDTLB		(1UL << 14)
+#define	MCNTL_FW_FITLB		(1UL << 15)
+#define	MCNTL_NC_CACHE		(1UL << 16)
+
+/* The following bits are valid for the SPARC64 VI, VII and VIIIfx only. */
+#define	MCNTL_MPG_SDTLB		(1UL << 6)
+#define	MCNTL_MPG_SITLB		(1UL << 7)
+
+/* The following bits are valid for the SPARC64 VIIIfx only. */
+#define	MCNTL_HPF_SHIFT		18
+#define	MCNTL_HPF_BITS		2
+#define	MCNTL_HPF_MASK							\
+	(((1UL << MCNTL_HPF_BITS) - 1) << MCNTL_HPF_SHIFT)
+#define	MCNTL_HPF_STRONG	(0UL << MCNTL_HPF_SHIFT)
+#define	MCNTL_HPF_NOT		(1UL << MCNTL_HPF_SHIFT)
+#define	MCNTL_HPF_WEAK		(2UL << MCNTL_HPF_SHIFT)
+
+#endif	/* _MACHINE_MCNTL_H */
diff --git a/sys/sparc64/sparc64/cache.c b/sys/sparc64/sparc64/cache.c
index 8cb8dee8a20..a3c6119ab97 100644
--- a/sys/sparc64/sparc64/cache.c
+++ b/sys/sparc64/sparc64/cache.c
@@ -141,7 +141,12 @@ cache_init(struct pcpu *pcpu)
 	if ((pcpu->pc_cache.dc_size &
 	    ~(1UL << (ffs(pcpu->pc_cache.dc_size) - 1))) != 0)
 		panic("cache_init: D$ size not a power of 2");
-	if (((pcpu->pc_cache.dc_size / pcpu->pc_cache.dc_assoc) /
+	/*
+	 * For CPUs which don't support unaliasing in hardware ensure that
+	 * the data cache doesn't have too many virtual colors.
+	 */
+	if (pcpu->pc_impl != CPU_IMPL_SPARC64V &&
+	    ((pcpu->pc_cache.dc_size / pcpu->pc_cache.dc_assoc) /
 	    PAGE_SIZE) != DCACHE_COLORS)
 		panic("cache_init: too many D$ colors");
 	set = pcpu->pc_cache.ec_size / pcpu->pc_cache.ec_assoc;
@@ -155,12 +160,21 @@ cache_init(struct pcpu *pcpu)
 		icache_page_inval = cheetah_icache_page_inval;
 		tlb_flush_nonlocked = cheetah_tlb_flush_nonlocked;
 		tlb_flush_user = cheetah_tlb_flush_user;
-	} else {
+	} else if (pcpu->pc_impl == CPU_IMPL_SPARC64V) {
+		cache_enable = cheetah_cache_enable;
+		cache_flush = zeus_cache_flush;
+		dcache_page_inval = zeus_dcache_page_inval;
+		icache_page_inval = zeus_icache_page_inval;
+		tlb_flush_nonlocked = cheetah_tlb_flush_nonlocked;
+		tlb_flush_user = cheetah_tlb_flush_user;
+	} else if (pcpu->pc_impl >= CPU_IMPL_ULTRASPARCI &&
+	    pcpu->pc_impl < CPU_IMPL_ULTRASPARCIII) {
 		cache_enable = spitfire_cache_enable;
 		cache_flush = spitfire_cache_flush;
 		dcache_page_inval = spitfire_dcache_page_inval;
 		icache_page_inval = spitfire_icache_page_inval;
 		tlb_flush_nonlocked = spitfire_tlb_flush_nonlocked;
 		tlb_flush_user = spitfire_tlb_flush_user;
-	}
+	} else
+		panic("cache_init: unknown CPU");
 }
diff --git a/sys/sparc64/sparc64/cheetah.c b/sys/sparc64/sparc64/cheetah.c
index 836618737d1..99d38c966ff 100644
--- a/sys/sparc64/sparc64/cheetah.c
+++ b/sys/sparc64/sparc64/cheetah.c
@@ -39,11 +39,13 @@ __FBSDID("$FreeBSD$");
 #include <vm/vm.h>
 #include <vm/pmap.h>
 
+#include <machine/asi.h>
 #include <machine/cache.h>
 #include <machine/cpu.h>
 #include <machine/cpufunc.h>
 #include <machine/dcr.h>
 #include <machine/lsu.h>
+#include <machine/mcntl.h>
 #include <machine/smp.h>
 #include <machine/tlb.h>
 #include <machine/ver.h>
@@ -52,7 +54,8 @@ __FBSDID("$FreeBSD$");
 #define	CHEETAH_ICACHE_TAG_LOWER	0x30
 
 /*
- * CPU-specific initialization
+ * CPU-specific initialization - this is used for both the Sun Cheetah and
+ * later as well as the Fujitsu Zeus and later CPUs.
  */
 void
 cheetah_init(u_int cpu_impl)
@@ -75,6 +78,14 @@ cheetah_init(u_int cpu_impl)
 	stxa(AA_IMMU_TSB_NEXT_REG, ASI_IMMU, 0);
 	membar(Sync);
 
+	if (cpu_impl == CPU_IMPL_SPARC64V) {
+		/* Ensure MCNTL_JPS1_TSBP is 0. */
+		val = ldxa(AA_MCNTL, ASI_MCNTL);
+		val &= ~MCNTL_JPS1_TSBP;
+		stxa(AA_MCNTL, ASI_MCNTL, val);
+		return;
+	}
+
 	/*
 	 * Configure the first large dTLB to hold 4MB pages (e.g. for direct
 	 * mappings) for all three contexts and ensure the second one is set
@@ -207,7 +218,7 @@ cheetah_dcache_page_inval(vm_paddr_t spa)
  * consistency is maintained by hardware.
  */
 void
-cheetah_icache_page_inval(vm_paddr_t pa)
+cheetah_icache_page_inval(vm_paddr_t pa __unused)
 {
 
 }
diff --git a/sys/sparc64/sparc64/identcpu.c b/sys/sparc64/sparc64/identcpu.c
index 0b731515f16..455aa6719bc 100644
--- a/sys/sparc64/sparc64/identcpu.c
+++ b/sys/sparc64/sparc64/identcpu.c
@@ -41,7 +41,7 @@ cpu_identify(u_long vers, u_int freq, u_int id)
 
 	switch (VER_MANUF(vers)) {
 	case 0x04:
-		manus = "HAL";
+		manus = "HAL/Fujitsu";
 		break;
 	case 0x13:
 	case 0x17:
@@ -57,6 +57,27 @@ cpu_identify(u_long vers, u_int freq, u_int id)
 	case CPU_IMPL_SPARC64:
 		impls = "SPARC64";
 		break;
+	case CPU_IMPL_SPARC64II:
+		impls = "SPARC64-II";
+		break;
+	case CPU_IMPL_SPARC64III:
+		impls = "SPARC64-III";
+		break;
+	case CPU_IMPL_SPARC64IV:
+		impls = "SPARC64-IV";
+		break;
+	case CPU_IMPL_SPARC64V:
+		impls = "SPARC64-V";
+		break;
+	case CPU_IMPL_SPARC64VI:
+		impls = "SPARC64-VI";
+		break;
+	case CPU_IMPL_SPARC64VII:
+		impls = "SPARC64-VII";
+		break;
+	case CPU_IMPL_SPARC64VIIIfx:
+		impls = "SPARC64-VIIIfx";
+		break;
 	case CPU_IMPL_ULTRASPARCI:
 		impls = "UltraSparc-I";
 		break;
@@ -67,7 +88,6 @@ cpu_identify(u_long vers, u_int freq, u_int id)
 		impls = "UltraSparc-IIi";
 		break;
 	case CPU_IMPL_ULTRASPARCIIe:
-		/* V9 Manual says `UltraSparc-e'.  I assume this is wrong. */
 		impls = "UltraSparc-IIe";
 		break;
 	case CPU_IMPL_ULTRASPARCIII:
diff --git a/sys/sparc64/sparc64/machdep.c b/sys/sparc64/sparc64/machdep.c
index c9040601642..eda2e01670a 100644
--- a/sys/sparc64/sparc64/machdep.c
+++ b/sys/sparc64/sparc64/machdep.c
@@ -276,6 +276,7 @@ cpu_cpuid_prop(u_int cpu_impl)
 
 	switch (cpu_impl) {
 	case CPU_IMPL_SPARC64:
+	case CPU_IMPL_SPARC64V:
 	case CPU_IMPL_ULTRASPARCI:
 	case CPU_IMPL_ULTRASPARCII:
 	case CPU_IMPL_ULTRASPARCIIi:
@@ -300,6 +301,7 @@ cpu_get_mid(u_int cpu_impl)
 
 	switch (cpu_impl) {
 	case CPU_IMPL_SPARC64:
+	case CPU_IMPL_SPARC64V:
 	case CPU_IMPL_ULTRASPARCI:
 	case CPU_IMPL_ULTRASPARCII:
 	case CPU_IMPL_ULTRASPARCIIi:
@@ -343,7 +345,8 @@ sparc64_init(caddr_t mdp, u_long o1, u_long o2, u_long o3, ofw_vec_t *vec)
 	/*
 	 * Do CPU-specific initialization.
 	 */
-	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+	if (cpu_impl == CPU_IMPL_SPARC64V ||
+	    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 		cheetah_init(cpu_impl);
 
 	/*
@@ -491,6 +494,7 @@ sparc64_init(caddr_t mdp, u_long o1, u_long o2, u_long o3, ofw_vec_t *vec)
 	if (cpu_use_vis) {
 		switch (cpu_impl) {
 		case CPU_IMPL_SPARC64:
+		case CPU_IMPL_SPARC64V:
 		case CPU_IMPL_ULTRASPARCI:
 		case CPU_IMPL_ULTRASPARCII:
 		case CPU_IMPL_ULTRASPARCIIi:
diff --git a/sys/sparc64/sparc64/mp_locore.S b/sys/sparc64/sparc64/mp_locore.S
index bd4fd13320d..e78b4b9afb3 100644
--- a/sys/sparc64/sparc64/mp_locore.S
+++ b/sys/sparc64/sparc64/mp_locore.S
@@ -200,19 +200,25 @@ ENTRY(mp_startup)
 	srlx	%l1, VER_IMPL_SHIFT, %l1
 	sll	%l1, VER_IMPL_SIZE, %l1
 	srl	%l1, VER_IMPL_SIZE, %l1
+	cmp	%l1, CPU_IMPL_SPARC64V
+	bl	%icc, 4f
+	 nop
+	cmp	%l1, CPU_IMPL_ULTRASPARCI
+	bl	%icc, 2f
+	 nop
 	cmp	%l1, CPU_IMPL_ULTRASPARCIII
 	bl	%icc, 3f
 	 nop
-	mov	CPU_STICKSYNC, %l2
+2:	mov	CPU_STICKSYNC, %l2
 	membar	#StoreLoad
 	stw	%l2, [%l0 + CSA_STATE]
 
-2:	ldx	[%l0 + CSA_STICK], %l2
-	brz	%l2, 2b
+3:	ldx	[%l0 + CSA_STICK], %l2
+	brz	%l2, 3b
 	 nop
 	wr	%l2, 0, %asr24
 
-3:	call	cpu_get_mid
+4:	call	cpu_get_mid
 	 mov	%l1, %o0
 
 	/*
@@ -225,9 +231,9 @@ ENTRY(mp_startup)
 	/*
 	 * Wait till its our turn to bootstrap.
 	 */
-4:	lduw	[%l0 + CSA_MID], %l1
+5:	lduw	[%l0 + CSA_MID], %l1
 	cmp	%l1, %o0
-	bne	%xcc, 4b
+	bne	%xcc, 5b
 	 nop
 
 	add	%l0, CSA_TTES, %l1
@@ -236,7 +242,7 @@ ENTRY(mp_startup)
 	/*
 	 * Map the per-CPU pages.
 	 */
-5:	sllx	%l2, TTE_SHIFT, %l3
+6:	sllx	%l2, TTE_SHIFT, %l3
 	add	%l1, %l3, %l3
 
 	ldx	[%l3 + TTE_VPN], %l4
@@ -251,7 +257,7 @@ ENTRY(mp_startup)
 
 	add	%l2, 1, %l2
 	cmp	%l2, PCPU_PAGES
-	bne	%xcc, 5b
+	bne	%xcc, 6b
 	 nop
 
 	/*
diff --git a/sys/sparc64/sparc64/mp_machdep.c b/sys/sparc64/sparc64/mp_machdep.c
index ce66a4b0ee2..428120437bc 100644
--- a/sys/sparc64/sparc64/mp_machdep.c
+++ b/sys/sparc64/sparc64/mp_machdep.c
@@ -164,7 +164,8 @@ mp_init(u_int cpu_impl)
 	if (cpu_impl == CPU_IMPL_ULTRASPARCIIIi ||
 	    cpu_impl == CPU_IMPL_ULTRASPARCIIIip)
 		isjbus = 1;
-	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+	if (cpu_impl == CPU_IMPL_SPARC64V ||
+	    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 		cpu_ipi_selected = cheetah_ipi_selected;
 	else
 		cpu_ipi_selected = spitfire_ipi_selected;
@@ -315,7 +316,8 @@ ap_start(phandle_t node, u_int mid, u_int cpu_impl)
 		;
 	membar(StoreLoad);
 	csa->csa_tick = rd(tick);
-	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII) {
+	if (cpu_impl == CPU_IMPL_SPARC64V ||
+	    cpu_impl >= CPU_IMPL_ULTRASPARCIII) {
 		while (csa->csa_state != CPU_STICKSYNC)
 			;
 		membar(StoreLoad);
@@ -411,7 +413,8 @@ cpu_mp_bootstrap(struct pcpu *pc)
 	csa = &cpu_start_args;
 
 	/* Do CPU-specific initialization. */
-	if (pc->pc_impl >= CPU_IMPL_ULTRASPARCIII)
+	if (pc->pc_impl == CPU_IMPL_SPARC64V ||
+	    pc->pc_impl >= CPU_IMPL_ULTRASPARCIII)
 		cheetah_init(pc->pc_impl);
 	/*
 	 * Enable the caches.  Note that his may include applying workarounds.
diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index 5d69c13e758..4a2853ca7b0 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -528,7 +528,8 @@ pmap_bootstrap(u_int cpu_impl)
 			tp->tte_data =
 			    ((translations[i].om_tte &
 			    ~((TD_SOFT2_MASK << TD_SOFT2_SHIFT) |
-			    (cpu_impl < CPU_IMPL_ULTRASPARCIII ?
+			    (cpu_impl >= CPU_IMPL_ULTRASPARCI &&
+			    cpu_impl < CPU_IMPL_ULTRASPARCIII ?
 			    (TD_DIAG_SF_MASK << TD_DIAG_SF_SHIFT) :
 			    (TD_RSVD_CH_MASK << TD_RSVD_CH_SHIFT)) |
 			    (TD_SOFT_MASK << TD_SOFT_SHIFT))) | TD_EXEC) +
@@ -704,16 +705,16 @@ retry:
 		if (va >= VM_MIN_DIRECT_ADDRESS) {
 			tp = NULL;
 			m = PHYS_TO_VM_PAGE(TLB_DIRECT_TO_PHYS(va));
-			(void)vm_page_pa_tryrelock(pm, TLB_DIRECT_TO_PHYS(va), &pa);
+			(void)vm_page_pa_tryrelock(pm, TLB_DIRECT_TO_PHYS(va),
+			    &pa);
 			vm_page_hold(m);
 		} else {
 			tp = tsb_kvtotte(va);
 			if ((tp->tte_data & TD_V) == 0)
 				tp = NULL;
 		}
-	} else {
+	} else
 		tp = tsb_tte_lookup(pm, va);
-	}
 	if (tp != NULL && ((tp->tte_data & TD_SW) ||
 	    (prot & VM_PROT_WRITE) == 0)) {
 		if (vm_page_pa_tryrelock(pm, TTE_GET_PA(tp), &pa))
diff --git a/sys/sparc64/sparc64/tick.c b/sys/sparc64/sparc64/tick.c
index 1a38013e884..041faa0ab04 100644
--- a/sys/sparc64/sparc64/tick.c
+++ b/sys/sparc64/sparc64/tick.c
@@ -120,7 +120,8 @@ cpu_initclocks(void)
 		 */
 	} else {
 		clock = PCPU_GET(clock);
-		intr_setup(PIL_TICK, PCPU_GET(impl) < CPU_IMPL_ULTRASPARCIII ?
+		intr_setup(PIL_TICK, PCPU_GET(impl) >= CPU_IMPL_ULTRASPARCI &&
+		    PCPU_GET(impl) < CPU_IMPL_ULTRASPARCIII ?
 		    tick_hardclock_bbwar : tick_hardclock, -1, NULL, NULL);
 		set_cputicker(tick_cputicks, clock, 0);
 	}
@@ -325,7 +326,8 @@ void
 tick_clear(u_int cpu_impl)
 {
 
-	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+	if (cpu_impl == CPU_IMPL_SPARC64V ||
+	    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 		wrstick(0, 0);
 	wrpr(tick, 0, 0);
 }
@@ -334,7 +336,8 @@ void
 tick_stop(u_int cpu_impl)
 {
 
-	if (cpu_impl >= CPU_IMPL_ULTRASPARCIII)
+	if (cpu_impl == CPU_IMPL_SPARC64V ||
+	    cpu_impl >= CPU_IMPL_ULTRASPARCIII)
 		wrstickcmpr(1L << 63, 0);
 	wrtickcmpr(1L << 63, 0);
 }
diff --git a/sys/sparc64/sparc64/zeus.c b/sys/sparc64/sparc64/zeus.c
new file mode 100644
index 00000000000..a3dc4e56e92
--- /dev/null
+++ b/sys/sparc64/sparc64/zeus.c
@@ -0,0 +1,65 @@
+/*-
+ * Copyright (c) 2010 Marius Strobl <marius@FreeBSD.org>
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/param.h>
+#include <sys/systm.h>
+
+#include <machine/asi.h>
+#include <machine/cache.h>
+#include <machine/cpufunc.h>
+
+/*
+ * Flush all lines from the level 1 caches.
+ */
+void
+zeus_cache_flush(void)
+{
+
+	stxa_sync(0, ASI_FLUSH_L1I, 0);
+}
+
+/*
+ * Flush a physical page from the data cache.  Data cache consistency is
+ * maintained by hardware.
+ */
+void
+zeus_dcache_page_inval(vm_paddr_t spa __unused)
+{
+
+}
+
+/*
+ * Flush a physical page from the intsruction cache.  Instruction cache
+ * consistency is maintained by hardware.
+ */
+void
+zeus_icache_page_inval(vm_paddr_t pa __unused)
+{
+
+}

From f623e552699270a192be864bbe84cd09af1b37c2 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 20:24:25 +0000
Subject: [PATCH 406/532] Simplify vm_fault().  The introduction of the new
 page lock renders a bit of cleverness by vm_fault() to avoid repeatedly
 releasing and reacquiring the page queues lock pointless.

Reviewed by:	kib, kmacy
---
 sys/vm/vm_fault.c | 18 +++++-------------
 1 file changed, 5 insertions(+), 13 deletions(-)

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 5ee4ab439a0..f9255944a01 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -215,7 +215,7 @@ vm_fault(vm_map_t map, vm_offset_t vaddr, vm_prot_t fault_type,
 {
 	vm_prot_t prot;
 	int is_first_object_locked, result;
-	boolean_t are_queues_locked, growstack, wired;
+	boolean_t growstack, wired;
 	int map_generation;
 	vm_object_t next_object;
 	vm_page_t marray[VM_FAULT_READ];
@@ -478,7 +478,6 @@ readrest:
 				else
 					firstpindex = fs.first_pindex - 2 * VM_FAULT_READ;
 
-				are_queues_locked = FALSE;
 				/*
 				 * note: partially valid pages cannot be 
 				 * included in the lookahead - NFS piecemeal
@@ -495,17 +494,11 @@ readrest:
 					if (mt->busy ||
 					    (mt->oflags & VPO_BUSY))
 						continue;
-					if (!are_queues_locked) {
-						are_queues_locked = TRUE;
-						vm_page_lock(mt);
-						vm_page_lock_queues();
-					} else {
-						vm_page_unlock_queues();
-						vm_page_lock(mt);
-						vm_page_lock_queues();
-					}
+					vm_page_lock(mt);
+					vm_page_lock_queues();
 					if (mt->hold_count ||
 					    mt->wire_count) {
+						vm_page_unlock_queues();
 						vm_page_unlock(mt);
 						continue;
 					}
@@ -515,10 +508,9 @@ readrest:
 					} else {
 						vm_page_cache(mt);
 					}
+					vm_page_unlock_queues();
 					vm_page_unlock(mt);
 				}
-				if (are_queues_locked)
-					vm_page_unlock_queues();
 				ahead += behind;
 				behind = 0;
 			}

From 447fe2a4c6fa56ce15c48080b0a73c5f8e7bed70 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 20:46:17 +0000
Subject: [PATCH 407/532] Defer the acquisition of the page and page queues
 locks in vm_pageout_object_deactivate_pages().

---
 sys/vm/vm_pageout.c | 16 ++++++++--------
 1 file changed, 8 insertions(+), 8 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 39a06f84939..3ce8a69e891 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -540,6 +540,7 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 	for (object = first_object;; object = backing_object) {
 		if (pmap_resident_count(pmap) <= desired)
 			goto unlock_return;
+		VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 		if (object->type == OBJT_PHYS || object->paging_in_progress)
 			goto unlock_return;
 
@@ -551,19 +552,18 @@ vm_pageout_object_deactivate_pages(pmap, first_object, desired)
 		 */
 		p = TAILQ_FIRST(&object->memq);
 		while (p != NULL) {
+			if (pmap_resident_count(pmap) <= desired)
+				goto unlock_return;
+			next = TAILQ_NEXT(p, listq);
+			if ((p->oflags & VPO_BUSY) != 0 || p->busy != 0) {
+				p = next;
+				continue;
+			}
 			vm_page_lock(p);
 			vm_page_lock_queues();
-			if (pmap_resident_count(pmap) <= desired) {
-				vm_page_unlock_queues();
-				vm_page_unlock(p);
-				goto unlock_return;
-			}
-			next = TAILQ_NEXT(p, listq);
 			cnt.v_pdpages++;
 			if (p->wire_count != 0 ||
 			    p->hold_count != 0 ||
-			    p->busy != 0 ||
-			    (p->oflags & VPO_BUSY) ||
 			    !pmap_page_exists_quick(pmap, p)) {
 				vm_page_unlock_queues();
 				vm_page_unlock(p);

From 6c56db5c9ec13c083b101b17aaa66d8d1dfc8e19 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 21:04:59 +0000
Subject: [PATCH 408/532] Eliminate an assignment that was made redundant by
 r207410.

---
 sys/vm/vm_pageout.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 3ce8a69e891..9921a84b6bb 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -1370,8 +1370,6 @@ vm_pageout_page_stats()
 		    ("vm_pageout_page_stats: page %p isn't active", m));
 
 		next = TAILQ_NEXT(m, pageq);
-		object = m->object;
-
 		if ((m->flags & PG_MARKER) != 0) {
 			m = next;
 			continue;

From 5eb6cd026301e64e3a97c1c63459f51010678196 Mon Sep 17 00:00:00 2001
From: Matt Jacob <mjacob@FreeBSD.org>
Date: Sun, 2 May 2010 22:48:27 +0000
Subject: [PATCH 409/532] Print IR_RESYNC updates informatively.

Obtained from:	pluknet
MFC after:	1 week
---
 sys/dev/mpt/mpt_cam.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/sys/dev/mpt/mpt_cam.c b/sys/dev/mpt/mpt_cam.c
index 56ea3e0f00e..39b57de20ac 100644
--- a/sys/dev/mpt/mpt_cam.c
+++ b/sys/dev/mpt/mpt_cam.c
@@ -2575,6 +2575,10 @@ mpt_cam_event(struct mpt_softc *mpt, request_t *req,
 		CAMLOCK_2_MPTLOCK(mpt);
 		break;
 	}
+	case MPI_EVENT_IR_RESYNC_UPDATE:
+		mpt_prt(mpt, "IR resync update %d completed\n",
+		    (data0 >> 16) & 0xff);
+		break;
 	case MPI_EVENT_EVENT_CHANGE:
 	case MPI_EVENT_INTEGRATED_RAID:
 	case MPI_EVENT_SAS_DEVICE_STATUS_CHANGE:

From b8d36afcfed7922db31d877a8123e1dfc6a238b0 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Sun, 2 May 2010 23:33:10 +0000
Subject: [PATCH 410/532] Add page lock assertions where we access the page's
 hold_count.

---
 sys/vm/vm_page.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index e73586a6007..59f82266413 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1657,6 +1657,7 @@ vm_page_try_to_cache(vm_page_t m)
 {
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
 	if (m->dirty || m->hold_count || m->busy || m->wire_count ||
 	    (m->oflags & VPO_BUSY) || (m->flags & PG_UNMANAGED)) {
@@ -1680,6 +1681,7 @@ vm_page_try_to_free(vm_page_t m)
 {
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	if (m->object != NULL)
 		VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
 	if (m->dirty || m->hold_count || m->busy || m->wire_count ||
@@ -1707,6 +1709,7 @@ vm_page_cache(vm_page_t m)
 	vm_page_t root;
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	object = m->object;
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	if ((m->flags & PG_UNMANAGED) || (m->oflags & VPO_BUSY) || m->busy ||

From 464aa6d5fb8f405c8f5358d66a55f1e934239786 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Mon, 3 May 2010 00:56:26 +0000
Subject: [PATCH 411/532] Fix wrong dma tag usage. Previously it used TX
 descriptor ring dma tag which should be TX mbuf dma tag.

Reported by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c | 7 ++++---
 1 file changed, 4 insertions(+), 3 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 37476a287ae..43b7022b983 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1421,7 +1421,7 @@ sge_encap(struct sge_softc *sc, struct mbuf **m_head)
 		}
 		*m_head = m;
 	}
-	error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_tx_tag, map,
+	error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_txmbuf_tag, map,
 	    *m_head, txsegs, &nsegs, 0);
 	if (error != 0) {
 		m_freem(*m_head);
@@ -1430,10 +1430,11 @@ sge_encap(struct sge_softc *sc, struct mbuf **m_head)
 	}
 	/* Check descriptor overrun. */
 	if (sc->sge_cdata.sge_tx_cnt + nsegs >= SGE_TX_RING_CNT) {
-		bus_dmamap_unload(sc->sge_cdata.sge_tx_tag, map);
+		bus_dmamap_unload(sc->sge_cdata.sge_txmbuf_tag, map);
 		return (ENOBUFS);
 	}
-	bus_dmamap_sync(sc->sge_cdata.sge_tx_tag, map, BUS_DMASYNC_PREWRITE);
+	bus_dmamap_sync(sc->sge_cdata.sge_txmbuf_tag, map,
+	    BUS_DMASYNC_PREWRITE);
 
 	cflags = 0;
 	if ((*m_head)->m_pkthdr.csum_flags & CSUM_IP)

From 913814935a618cfb68301dc38aa47042fd1130c3 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 05:41:50 +0000
Subject: [PATCH 412/532] This is the first step in transitioning
 responsibility for synchronizing access to the page's wire_count from the
 page queues lock to the page lock.

Submitted by:	kmacy
---
 sys/kern/uipc_cow.c      | 4 ++++
 sys/kern/uipc_syscalls.c | 4 ++++
 sys/net/bpf_zerocopy.c   | 2 ++
 3 files changed, 10 insertions(+)

diff --git a/sys/kern/uipc_cow.c b/sys/kern/uipc_cow.c
index 5c0dcd230ee..88a5625fb5b 100644
--- a/sys/kern/uipc_cow.c
+++ b/sys/kern/uipc_cow.c
@@ -80,6 +80,7 @@ socow_iodone(void *addr, void *args)
 	pp = sf_buf_page(sf);
 	sf_buf_free(sf);
 	/* remove COW mapping  */
+	vm_page_lock(pp);
 	vm_page_lock_queues();
 	vm_page_cowclear(pp);
 	vm_page_unwire(pp, 0);
@@ -91,6 +92,7 @@ socow_iodone(void *addr, void *args)
 	if (pp->wire_count == 0 && pp->object == NULL)
 		vm_page_free(pp);
 	vm_page_unlock_queues();
+	vm_page_unlock(pp);
 	socow_stats.iodone++;
 }
 
@@ -149,6 +151,7 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	 */
 	sf = sf_buf_alloc(pp, SFB_CATCH);
 	if (!sf) {
+		vm_page_lock(pp);
 		vm_page_lock_queues();
 		vm_page_cowclear(pp);
 		vm_page_unwire(pp, 0);
@@ -160,6 +163,7 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 		if (pp->wire_count == 0 && pp->object == NULL)
 			vm_page_free(pp);
 		vm_page_unlock_queues();
+		vm_page_unlock(pp);
 		socow_stats.fail_sf_buf++;
 		return(0);
 	}
diff --git a/sys/kern/uipc_syscalls.c b/sys/kern/uipc_syscalls.c
index 3271235a6f9..a1378758836 100644
--- a/sys/kern/uipc_syscalls.c
+++ b/sys/kern/uipc_syscalls.c
@@ -1715,6 +1715,7 @@ sf_buf_mext(void *addr, void *args)
 
 	m = sf_buf_page(args);
 	sf_buf_free(args);
+	vm_page_lock(m);
 	vm_page_lock_queues();
 	vm_page_unwire(m, 0);
 	/*
@@ -1725,6 +1726,7 @@ sf_buf_mext(void *addr, void *args)
 	if (m->wire_count == 0 && m->object == NULL)
 		vm_page_free(m);
 	vm_page_unlock_queues();
+	vm_page_unlock(m);
 	if (addr == NULL)
 		return;
 	sfs = addr;
@@ -2136,6 +2138,7 @@ retry_space:
 			if ((sf = sf_buf_alloc(pg,
 			    (mnw ? SFB_NOWAIT : SFB_CATCH))) == NULL) {
 				mbstat.sf_allocfail++;
+				vm_page_lock(pg);
 				vm_page_lock_queues();
 				vm_page_unwire(pg, 0);
 				/*
@@ -2144,6 +2147,7 @@ retry_space:
 				if (pg->wire_count == 0 && pg->object == NULL)
 					vm_page_free(pg);
 				vm_page_unlock_queues();
+				vm_page_unlock(pg);
 				error = (mnw ? EAGAIN : EINTR);
 				break;
 			}
diff --git a/sys/net/bpf_zerocopy.c b/sys/net/bpf_zerocopy.c
index 3cf4b8d9a51..5ca69782ee3 100644
--- a/sys/net/bpf_zerocopy.c
+++ b/sys/net/bpf_zerocopy.c
@@ -112,11 +112,13 @@ static void
 zbuf_page_free(vm_page_t pp)
 {
 
+	vm_page_lock(pp);
 	vm_page_lock_queues();
 	vm_page_unwire(pp, 0);
 	if (pp->wire_count == 0 && pp->object == NULL)
 		vm_page_free(pp);
 	vm_page_unlock_queues();
+	vm_page_unlock(pp);
 }
 
 /*

From 9f2512bab5d3a80df59c8793a38c943474010fd9 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 07:00:50 +0000
Subject: [PATCH 413/532] Assert that the page queues lock is held in
 vm_page_remove() and vm_page_unwire() only if the page is managed, i.e.,
 pageable.

---
 sys/vm/vm_page.c | 6 ++++--
 1 file changed, 4 insertions(+), 2 deletions(-)

diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index 59f82266413..acec3b7263b 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -790,6 +790,8 @@ vm_page_remove(vm_page_t m)
 	vm_object_t object;
 	vm_page_t root;
 
+	if ((m->flags & PG_UNMANAGED) == 0)
+		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	if ((object = m->object) == NULL)
 		return;
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
@@ -797,7 +799,6 @@ vm_page_remove(vm_page_t m)
 		m->oflags &= ~VPO_BUSY;
 		vm_page_flash(m);
 	}
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 
 	/*
 	 * Now remove from the object's list of backed pages.
@@ -1586,7 +1587,8 @@ void
 vm_page_unwire(vm_page_t m, int activate)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	if ((m->flags & PG_UNMANAGED) == 0)
+		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	if (m->flags & PG_FICTITIOUS)
 		return;
 	if (m->wire_count > 0) {

From 746c2ddee82e7839310c53a5601f8b1567026b1b Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 07:08:16 +0000
Subject: [PATCH 414/532] The pages allocated by kmem_alloc_attr() and
 kmem_malloc() are unmanaged. Consequently, neither the page lock nor the page
 queues lock is needed to unwire and free them.

---
 sys/vm/vm_contig.c | 4 ----
 sys/vm/vm_kern.c   | 4 ----
 2 files changed, 8 deletions(-)

diff --git a/sys/vm/vm_contig.c b/sys/vm/vm_contig.c
index e6a9b78c4a0..c1c861807bc 100644
--- a/sys/vm/vm_contig.c
+++ b/sys/vm/vm_contig.c
@@ -268,11 +268,7 @@ retry:
 				i -= PAGE_SIZE;
 				m = vm_page_lookup(object, OFF_TO_IDX(offset +
 				    i));
-				vm_page_lock(m);
-				vm_page_lock_queues();
 				vm_page_free(m);
-				vm_page_unlock_queues();
-				vm_page_unlock(m);
 			}
 			VM_OBJECT_UNLOCK(object);
 			vm_map_delete(map, addr, addr + size);
diff --git a/sys/vm/vm_kern.c b/sys/vm/vm_kern.c
index ead6d0d71a6..95a4e9d6302 100644
--- a/sys/vm/vm_kern.c
+++ b/sys/vm/vm_kern.c
@@ -380,12 +380,8 @@ retry:
 				i -= PAGE_SIZE;
 				m = vm_page_lookup(kmem_object,
 						   OFF_TO_IDX(offset + i));
-				vm_page_lock(m);
-				vm_page_lock_queues();
 				vm_page_unwire(m, 0);
 				vm_page_free(m);
-				vm_page_unlock_queues();
-				vm_page_unlock(m);
 			}
 			VM_OBJECT_UNLOCK(kmem_object);
 			vm_map_delete(map, addr, addr + size);

From 551e75c7af421b58b9ba54d639af1e5f7a79ca23 Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Mon, 3 May 2010 07:32:24 +0000
Subject: [PATCH 415/532] Implement the no_user_check option to pam_krb5.

This option is available in the Linux implementation of pam_krb5
and allows to authorize a user not known to the local system.

Ccache is not used as we don't have a secure uid/gid for the cache file.

Usable for authentication of external kerberos users (e.g Active Directory)
via PAM from applications like Cyrus saslauthd, PHP or perl.

PR:		bin/146186
Submitted by:	myself
Approved by:	deplhij (mentor)
MFC after:	2 weeks
---
 lib/libpam/modules/pam_krb5/pam_krb5.8 | 4 ++++
 lib/libpam/modules/pam_krb5/pam_krb5.c | 9 ++++++++-
 2 files changed, 12 insertions(+), 1 deletion(-)

diff --git a/lib/libpam/modules/pam_krb5/pam_krb5.8 b/lib/libpam/modules/pam_krb5/pam_krb5.8
index 3e0db91e009..a182e62e70d 100644
--- a/lib/libpam/modules/pam_krb5/pam_krb5.8
+++ b/lib/libpam/modules/pam_krb5/pam_krb5.8
@@ -108,6 +108,10 @@ and
 .Ql %p ,
 to designate the current process ID; can be used in
 .Ar name .
+.It Cm no_user_check
+Do not verify if a user exists on the local system. This option implies the
+.Cm no_ccache
+option because there is no secure local uid/gid for the cache file.
 .El
 .Ss Kerberos 5 Account Management Module
 The Kerberos 5 account management component
diff --git a/lib/libpam/modules/pam_krb5/pam_krb5.c b/lib/libpam/modules/pam_krb5/pam_krb5.c
index b56e0a3c17e..9623a17191f 100644
--- a/lib/libpam/modules/pam_krb5/pam_krb5.c
+++ b/lib/libpam/modules/pam_krb5/pam_krb5.c
@@ -89,6 +89,7 @@ static void	compat_free_data_contents(krb5_context, krb5_data *);
 #define PAM_OPT_DEBUG		"debug"
 #define PAM_OPT_FORWARDABLE	"forwardable"
 #define PAM_OPT_NO_CCACHE	"no_ccache"
+#define PAM_OPT_NO_USER_CHECK	"no_user_check"
 #define PAM_OPT_REUSE_CCACHE	"reuse_ccache"
 
 /*
@@ -194,6 +195,10 @@ pam_sm_authenticate(pam_handle_t *pamh, int flags __unused,
 
 	PAM_LOG("Got password");
 
+	if (openpam_get_option(pamh, PAM_OPT_NO_USER_CHECK))
+		PAM_LOG("Skipping local user check");
+	else {
+
 	/* Verify the local user exists (AFTER getting the password) */
 	if (strchr(user, '@')) {
 		/* get a local account name for this principal */
@@ -221,6 +226,7 @@ pam_sm_authenticate(pam_handle_t *pamh, int flags __unused,
 	}
 
 	PAM_LOG("Done getpwnam()");
+	}
 
 	/* Get a TGT */
 	memset(&creds, 0, sizeof(krb5_creds));
@@ -366,7 +372,8 @@ pam_sm_setcred(pam_handle_t *pamh, int flags,
 		return (PAM_SERVICE_ERR);
 
 	/* If a persistent cache isn't desired, stop now. */
-	if (openpam_get_option(pamh, PAM_OPT_NO_CCACHE))
+	if (openpam_get_option(pamh, PAM_OPT_NO_CCACHE) ||
+		openpam_get_option(pamh, PAM_OPT_NO_USER_CHECK))
 		return (PAM_SUCCESS);
 
 	PAM_LOG("Establishing credentials");

From e50d35e6c6c68e83fa329aff043d53ab84982fd3 Mon Sep 17 00:00:00 2001
From: Maxim Sobolev <sobomax@FreeBSD.org>
Date: Mon, 3 May 2010 07:32:50 +0000
Subject: [PATCH 416/532] Add new tunable 'net.link.ifqmaxlen' to set default
 send interface queue length. The default value for this parameter is 50,
 which is quite low for many of today's uses and the only way to modify this
 parameter right now is to edit if_var.h file. Also add read-only sysctl with
 the same name, so that it's possible to retrieve the current value.

MFC after:	1 month
---
 sys/arm/at91/if_ate.c          | 4 ++--
 sys/arm/xscale/ixp425/if_npe.c | 2 +-
 sys/dev/ae/if_ae.c             | 2 +-
 sys/dev/an/if_an.c             | 4 ++--
 sys/dev/ath/if_ath.c           | 4 ++--
 sys/dev/bwi/if_bwi.c           | 4 ++--
 sys/dev/bwn/if_bwn.c           | 4 ++--
 sys/dev/ce/if_ce.c             | 8 ++++----
 sys/dev/cm/smc90cx6.c          | 2 +-
 sys/dev/cp/if_cp.c             | 4 ++--
 sys/dev/cs/if_cs.c             | 2 +-
 sys/dev/ctau/if_ct.c           | 4 ++--
 sys/dev/cx/if_cx.c             | 4 ++--
 sys/dev/cxgb/cxgb_main.c       | 2 +-
 sys/dev/ed/if_ed.c             | 4 ++--
 sys/dev/ep/if_ep.c             | 4 ++--
 sys/dev/ex/if_ex.c             | 2 +-
 sys/dev/fe/if_fe.c             | 2 +-
 sys/dev/ie/if_ie.c             | 2 +-
 sys/dev/iicbus/if_ic.c         | 2 +-
 sys/dev/ipw/if_ipw.c           | 4 ++--
 sys/dev/iwi/if_iwi.c           | 4 ++--
 sys/dev/iwn/if_iwn.c           | 4 ++--
 sys/dev/le/lance.c             | 4 ++--
 sys/dev/malo/if_malo.c         | 4 ++--
 sys/dev/mwl/if_mwl.c           | 4 ++--
 sys/dev/my/if_my.c             | 4 ++--
 sys/dev/nxge/if_nxge.c         | 2 +-
 sys/dev/pdq/pdq_ifsubr.c       | 2 +-
 sys/dev/ppbus/if_plip.c        | 2 +-
 sys/dev/ral/rt2560.c           | 4 ++--
 sys/dev/ral/rt2661.c           | 4 ++--
 sys/dev/sbni/if_sbni.c         | 2 +-
 sys/dev/smc/if_smc.c           | 2 +-
 sys/dev/sn/if_sn.c             | 4 ++--
 sys/dev/snc/dp83932.c          | 2 +-
 sys/dev/usb/net/uhso.c         | 4 ++--
 sys/dev/usb/net/usb_ethernet.c | 4 ++--
 sys/dev/usb/wlan/if_rum.c      | 4 ++--
 sys/dev/usb/wlan/if_run.c      | 4 ++--
 sys/dev/usb/wlan/if_uath.c     | 4 ++--
 sys/dev/usb/wlan/if_upgt.c     | 2 +-
 sys/dev/usb/wlan/if_ural.c     | 4 ++--
 sys/dev/usb/wlan/if_urtw.c     | 4 ++--
 sys/dev/usb/wlan/if_zyd.c      | 2 +-
 sys/dev/vx/if_vx.c             | 2 +-
 sys/dev/wi/if_wi.c             | 4 ++--
 sys/dev/wl/if_wl.c             | 2 +-
 sys/dev/wpi/if_wpi.c           | 4 ++--
 sys/dev/xe/if_xe.c             | 2 +-
 sys/mips/adm5120/if_admsw.c    | 4 ++--
 sys/mips/atheros/if_arge.c     | 4 ++--
 sys/net/if.c                   | 4 ++++
 sys/net/if_ef.c                | 2 +-
 sys/net/if_gif.c               | 2 +-
 sys/net/if_gre.c               | 2 +-
 sys/net/if_stf.c               | 2 +-
 sys/net80211/ieee80211.c       | 4 ++--
 sys/netgraph/ng_eiface.c       | 2 +-
 sys/netgraph/ng_fec.c          | 2 +-
 sys/netgraph/ng_iface.c        | 4 ++--
 sys/netgraph/ng_tty.c          | 2 +-
 sys/pci/if_rl.c                | 4 ++--
 63 files changed, 103 insertions(+), 99 deletions(-)

diff --git a/sys/arm/at91/if_ate.c b/sys/arm/at91/if_ate.c
index 2f6b741cc10..b0654823470 100644
--- a/sys/arm/at91/if_ate.c
+++ b/sys/arm/at91/if_ate.c
@@ -272,8 +272,8 @@ ate_attach(device_t dev)
 	ifp->if_ioctl = ateioctl;
 	ifp->if_init = ateinit;
 	ifp->if_baudrate = 10000000;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 	ifp->if_linkmib = &sc->mibdata;
 	ifp->if_linkmiblen = sizeof(sc->mibdata);
diff --git a/sys/arm/xscale/ixp425/if_npe.c b/sys/arm/xscale/ixp425/if_npe.c
index 5b605aeb6d2..3d8d670eec9 100644
--- a/sys/arm/xscale/ixp425/if_npe.c
+++ b/sys/arm/xscale/ixp425/if_npe.c
@@ -360,7 +360,7 @@ npe_attach(device_t dev)
 	ifp->if_ioctl = npeioctl;
 	ifp->if_init = npeinit;
 	IFQ_SET_MAXLEN(&ifp->if_snd, sc->txdma.nbuf - 1);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 	ifp->if_linkmib = &sc->mibdata;
 	ifp->if_linkmiblen = sizeof(sc->mibdata);
diff --git a/sys/dev/ae/if_ae.c b/sys/dev/ae/if_ae.c
index 68181f4b6d4..d2812a4990b 100644
--- a/sys/dev/ae/if_ae.c
+++ b/sys/dev/ae/if_ae.c
@@ -378,7 +378,7 @@ ae_attach(device_t dev)
 	ifp->if_init = ae_init;
 	ifp->if_capabilities = IFCAP_VLAN_MTU | IFCAP_VLAN_HWTAGGING;
 	ifp->if_hwassist = 0;
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_MAXLEN(&ifp->if_snd, ifp->if_snd.ifq_drv_maxlen);
 	IFQ_SET_READY(&ifp->if_snd);
 	if (pci_find_extcap(dev, PCIY_PMG, &pmc) == 0) {
diff --git a/sys/dev/an/if_an.c b/sys/dev/an/if_an.c
index 6d9ef96831b..645fe419ebe 100644
--- a/sys/dev/an/if_an.c
+++ b/sys/dev/an/if_an.c
@@ -767,8 +767,8 @@ an_attach(struct an_softc *sc, int flags)
 	ifp->if_start = an_start;
 	ifp->if_init = an_init;
 	ifp->if_baudrate = 10000000;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	bzero(sc->an_config.an_nodename, sizeof(sc->an_config.an_nodename));
diff --git a/sys/dev/ath/if_ath.c b/sys/dev/ath/if_ath.c
index 58474890e72..15765f2b5b0 100644
--- a/sys/dev/ath/if_ath.c
+++ b/sys/dev/ath/if_ath.c
@@ -562,8 +562,8 @@ ath_attach(u_int16_t devid, struct ath_softc *sc)
 	ifp->if_start = ath_start;
 	ifp->if_ioctl = ath_ioctl;
 	ifp->if_init = ath_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/bwi/if_bwi.c b/sys/dev/bwi/if_bwi.c
index 96c5cc6d3e5..724778d192c 100644
--- a/sys/dev/bwi/if_bwi.c
+++ b/sys/dev/bwi/if_bwi.c
@@ -461,8 +461,8 @@ bwi_attach(struct bwi_softc *sc)
 	ifp->if_init = bwi_init;
 	ifp->if_ioctl = bwi_ioctl;
 	ifp->if_start = bwi_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 	callout_init_mtx(&sc->sc_watchdog_timer, &sc->sc_mtx, 0);
 
diff --git a/sys/dev/bwn/if_bwn.c b/sys/dev/bwn/if_bwn.c
index 5ca5098115c..fd364035d3c 100644
--- a/sys/dev/bwn/if_bwn.c
+++ b/sys/dev/bwn/if_bwn.c
@@ -1193,8 +1193,8 @@ bwn_attach_pre(struct bwn_softc *sc)
 	ifp->if_init = bwn_init;
 	ifp->if_ioctl = bwn_ioctl;
 	ifp->if_start = bwn_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	return (0);
diff --git a/sys/dev/ce/if_ce.c b/sys/dev/ce/if_ce.c
index f04fd0916e6..d151b4243a8 100644
--- a/sys/dev/ce/if_ce.c
+++ b/sys/dev/ce/if_ce.c
@@ -701,9 +701,9 @@ static int ce_attach (device_t dev)
 #endif
 			continue;
 		}
-		d->queue.ifq_maxlen	= IFQ_MAXLEN;
-		d->hi_queue.ifq_maxlen	= IFQ_MAXLEN;
-		d->rqueue.ifq_maxlen	= IFQ_MAXLEN;
+		d->queue.ifq_maxlen	= ifqmaxlen;
+		d->hi_queue.ifq_maxlen	= ifqmaxlen;
+		d->rqueue.ifq_maxlen	= ifqmaxlen;
 #if __FreeBSD_version >= 500000
 		mtx_init (&d->queue.ifq_mtx, "ce_queue", NULL, MTX_DEF);
 		mtx_init (&d->hi_queue.ifq_mtx, "ce_queue_hi", NULL, MTX_DEF);
@@ -732,7 +732,7 @@ static int ce_attach (device_t dev)
 		d->ifp->if_ioctl	= ce_sioctl;
 		d->ifp->if_start	= ce_ifstart;
 		d->ifp->if_init		= ce_initialize;
-		d->rqueue.ifq_maxlen	= IFQ_MAXLEN;
+		d->rqueue.ifq_maxlen	= ifqmaxlen;
 #if __FreeBSD_version >= 500000
 		mtx_init (&d->rqueue.ifq_mtx, "ce_rqueue", NULL, MTX_DEF);
 #endif		
diff --git a/sys/dev/cm/smc90cx6.c b/sys/dev/cm/smc90cx6.c
index 0e9015ea17c..719e46868a1 100644
--- a/sys/dev/cm/smc90cx6.c
+++ b/sys/dev/cm/smc90cx6.c
@@ -189,7 +189,7 @@ cm_attach(dev)
 	ifp->if_ioctl = cm_ioctl;
 	ifp->if_init = cm_init;
 	/* XXX IFQ_SET_READY(&ifp->if_snd); */
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX;
 
 	arc_ifattach(ifp, linkaddress);
diff --git a/sys/dev/cp/if_cp.c b/sys/dev/cp/if_cp.c
index f6eeed5058e..e6402309846 100644
--- a/sys/dev/cp/if_cp.c
+++ b/sys/dev/cp/if_cp.c
@@ -505,8 +505,8 @@ static int cp_attach (device_t dev)
 			NG_NODE_UNREF (d->node);
 			continue;
 		}
-		d->queue.ifq_maxlen = IFQ_MAXLEN;
-		d->hi_queue.ifq_maxlen = IFQ_MAXLEN;
+		d->queue.ifq_maxlen = ifqmaxlen;
+		d->hi_queue.ifq_maxlen = ifqmaxlen;
 		mtx_init (&d->queue.ifq_mtx, "cp_queue", NULL, MTX_DEF);
 		mtx_init (&d->hi_queue.ifq_mtx, "cp_queue_hi", NULL, MTX_DEF);
 #else /*NETGRAPH*/
diff --git a/sys/dev/cs/if_cs.c b/sys/dev/cs/if_cs.c
index b32a03d57cf..476b9ddbb96 100644
--- a/sys/dev/cs/if_cs.c
+++ b/sys/dev/cs/if_cs.c
@@ -500,7 +500,7 @@ cs_attach(device_t dev)
 	ifp->if_start=cs_start;
 	ifp->if_ioctl=cs_ioctl;
 	ifp->if_init=cs_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 	ifp->if_flags=(IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST);
 
diff --git a/sys/dev/ctau/if_ct.c b/sys/dev/ctau/if_ct.c
index 125dfa5ed80..dbdca97411d 100644
--- a/sys/dev/ctau/if_ct.c
+++ b/sys/dev/ctau/if_ct.c
@@ -722,8 +722,8 @@ static int ct_attach (device_t dev)
 			ct_bus_dma_mem_free (&d->dmamem);
 			continue;
 		}
-		d->queue.ifq_maxlen = IFQ_MAXLEN;
-		d->hi_queue.ifq_maxlen = IFQ_MAXLEN;
+		d->queue.ifq_maxlen = ifqmaxlen;
+		d->hi_queue.ifq_maxlen = ifqmaxlen;
 		mtx_init (&d->queue.ifq_mtx, "ct_queue", NULL, MTX_DEF);
 		mtx_init (&d->hi_queue.ifq_mtx, "ct_queue_hi", NULL, MTX_DEF);
 #else /*NETGRAPH*/
diff --git a/sys/dev/cx/if_cx.c b/sys/dev/cx/if_cx.c
index 4e9750e1483..c8f53a00bec 100644
--- a/sys/dev/cx/if_cx.c
+++ b/sys/dev/cx/if_cx.c
@@ -831,8 +831,8 @@ static int cx_attach (device_t dev)
 			cx_bus_dma_mem_free (&d->dmamem);
 			continue;
 		}
-		d->lo_queue.ifq_maxlen = IFQ_MAXLEN;
-		d->hi_queue.ifq_maxlen = IFQ_MAXLEN;
+		d->lo_queue.ifq_maxlen = ifqmaxlen;
+		d->hi_queue.ifq_maxlen = ifqmaxlen;
 		mtx_init (&d->lo_queue.ifq_mtx, "cx_queue_lo", NULL, MTX_DEF);
 		mtx_init (&d->hi_queue.ifq_mtx, "cx_queue_hi", NULL, MTX_DEF);
 #else /*NETGRAPH*/
diff --git a/sys/dev/cxgb/cxgb_main.c b/sys/dev/cxgb/cxgb_main.c
index a47284fe9ae..a805b5d9826 100644
--- a/sys/dev/cxgb/cxgb_main.c
+++ b/sys/dev/cxgb/cxgb_main.c
@@ -1012,7 +1012,7 @@ cxgb_port_attach(device_t dev)
 	ifp->if_ioctl = cxgb_ioctl;
 	ifp->if_start = cxgb_start;
 
-	ifp->if_snd.ifq_drv_maxlen = cxgb_snd_queue_len;
+	ifp->if_snd.ifq_drv_maxlen = max(cxgb_snd_queue_len, ifqmaxlen);
 	IFQ_SET_MAXLEN(&ifp->if_snd, ifp->if_snd.ifq_drv_maxlen);
 	IFQ_SET_READY(&ifp->if_snd);
 
diff --git a/sys/dev/ed/if_ed.c b/sys/dev/ed/if_ed.c
index e065a670d5c..839e2a52cc9 100644
--- a/sys/dev/ed/if_ed.c
+++ b/sys/dev/ed/if_ed.c
@@ -283,8 +283,8 @@ ed_attach(device_t dev)
 	ifp->if_start = ed_start;
 	ifp->if_ioctl = ed_ioctl;
 	ifp->if_init = ed_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 	ifp->if_linkmib = &sc->mibdata;
 	ifp->if_linkmiblen = sizeof sc->mibdata;
diff --git a/sys/dev/ep/if_ep.c b/sys/dev/ep/if_ep.c
index 59dd1c4cc99..24362558ff3 100644
--- a/sys/dev/ep/if_ep.c
+++ b/sys/dev/ep/if_ep.c
@@ -306,8 +306,8 @@ ep_attach(struct ep_softc *sc)
 	ifp->if_start = epstart;
 	ifp->if_ioctl = epioctl;
 	ifp->if_init = epinit;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	callout_init_mtx(&sc->watchdog_timer, &sc->sc_mtx, 0);
diff --git a/sys/dev/ex/if_ex.c b/sys/dev/ex/if_ex.c
index d9e0c98f9fa..efd54bca0e5 100644
--- a/sys/dev/ex/if_ex.c
+++ b/sys/dev/ex/if_ex.c
@@ -237,7 +237,7 @@ ex_attach(device_t dev)
 	ifp->if_start = ex_start;
 	ifp->if_ioctl = ex_ioctl;
 	ifp->if_init = ex_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 	ifmedia_init(&sc->ifmedia, 0, ex_ifmedia_upd, ex_ifmedia_sts);
 	mtx_init(&sc->lock, device_get_nameunit(dev), MTX_NETWORK_LOCK,
diff --git a/sys/dev/fe/if_fe.c b/sys/dev/fe/if_fe.c
index eff92b77366..9fe1f2c2a09 100644
--- a/sys/dev/fe/if_fe.c
+++ b/sys/dev/fe/if_fe.c
@@ -766,7 +766,7 @@ fe_attach (device_t dev)
 	 * Set fixed interface flags.
 	 */
  	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 #if FE_SINGLE_TRANSMISSION
 	/* Override txb config to allocate minimum.  */
diff --git a/sys/dev/ie/if_ie.c b/sys/dev/ie/if_ie.c
index 34c658eef3d..172bf9e5bdf 100644
--- a/sys/dev/ie/if_ie.c
+++ b/sys/dev/ie/if_ie.c
@@ -318,7 +318,7 @@ ie_attach(device_t dev)
 	ifp->if_start = iestart;
 	ifp->if_ioctl = ieioctl;
 	ifp->if_init = ieinit;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 	ether_ifattach(ifp, sc->enaddr);
 
diff --git a/sys/dev/iicbus/if_ic.c b/sys/dev/iicbus/if_ic.c
index bb5186cc401..4a05b163f6b 100644
--- a/sys/dev/iicbus/if_ic.c
+++ b/sys/dev/iicbus/if_ic.c
@@ -181,7 +181,7 @@ icattach(device_t dev)
 	ifp->if_output = icoutput;
 	ifp->if_hdrlen = 0;
 	ifp->if_addrlen = 0;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 
 	ic_alloc_buffers(sc, ICMTU);
 
diff --git a/sys/dev/ipw/if_ipw.c b/sys/dev/ipw/if_ipw.c
index 3315d654b1c..2329c3377e0 100644
--- a/sys/dev/ipw/if_ipw.c
+++ b/sys/dev/ipw/if_ipw.c
@@ -289,8 +289,8 @@ ipw_attach(device_t dev)
 	ifp->if_init = ipw_init;
 	ifp->if_ioctl = ipw_ioctl;
 	ifp->if_start = ipw_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/iwi/if_iwi.c b/sys/dev/iwi/if_iwi.c
index 1105425da63..4b765ef16e9 100644
--- a/sys/dev/iwi/if_iwi.c
+++ b/sys/dev/iwi/if_iwi.c
@@ -363,8 +363,8 @@ iwi_attach(device_t dev)
 	ifp->if_init = iwi_init;
 	ifp->if_ioctl = iwi_ioctl;
 	ifp->if_start = iwi_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/iwn/if_iwn.c b/sys/dev/iwn/if_iwn.c
index 09c1efdbe71..5a17d1e3a7a 100644
--- a/sys/dev/iwn/if_iwn.c
+++ b/sys/dev/iwn/if_iwn.c
@@ -627,8 +627,8 @@ iwn_attach(device_t dev)
 	ifp->if_init = iwn_init;
 	ifp->if_ioctl = iwn_ioctl;
 	ifp->if_start = iwn_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ieee80211_ifattach(ic, macaddr);
diff --git a/sys/dev/le/lance.c b/sys/dev/le/lance.c
index db200995fa7..8700aa2cd43 100644
--- a/sys/dev/le/lance.c
+++ b/sys/dev/le/lance.c
@@ -133,8 +133,8 @@ lance_config(struct lance_softc *sc, const char* name, int unit)
 	ifp->if_flags &= ~IFF_MULTICAST;
 #endif
 	ifp->if_baudrate = IF_Mbps(10);
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	/* Initialize ifmedia structures. */
diff --git a/sys/dev/malo/if_malo.c b/sys/dev/malo/if_malo.c
index 57e0d3d8fa3..ae6ef882652 100644
--- a/sys/dev/malo/if_malo.c
+++ b/sys/dev/malo/if_malo.c
@@ -275,8 +275,8 @@ malo_attach(uint16_t devid, struct malo_softc *sc)
 	ifp->if_start = malo_start;
 	ifp->if_ioctl = malo_ioctl;
 	ifp->if_init = malo_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/mwl/if_mwl.c b/sys/dev/mwl/if_mwl.c
index e4e469b5bcf..662f201c195 100644
--- a/sys/dev/mwl/if_mwl.c
+++ b/sys/dev/mwl/if_mwl.c
@@ -404,8 +404,8 @@ mwl_attach(uint16_t devid, struct mwl_softc *sc)
 	ifp->if_start = mwl_start;
 	ifp->if_ioctl = mwl_ioctl;
 	ifp->if_init = mwl_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/my/if_my.c b/sys/dev/my/if_my.c
index 99f6071ba20..951473a3ba7 100644
--- a/sys/dev/my/if_my.c
+++ b/sys/dev/my/if_my.c
@@ -902,8 +902,8 @@ my_attach(device_t dev)
 	ifp->if_start = my_start;
 	ifp->if_init = my_init;
 	ifp->if_baudrate = 10000000;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	if (sc->my_info->my_did == MTD803ID)
diff --git a/sys/dev/nxge/if_nxge.c b/sys/dev/nxge/if_nxge.c
index c86d7b61bad..0e67413747c 100644
--- a/sys/dev/nxge/if_nxge.c
+++ b/sys/dev/nxge/if_nxge.c
@@ -1190,7 +1190,7 @@ xge_interface_setup(device_t dev)
 	ifnetp->if_start    = xge_send;
 
 	/* TODO: Check and assign optimal value */
-	ifnetp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifnetp->if_snd.ifq_maxlen = ifqmaxlen;
 
 	ifnetp->if_capabilities = IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU |
 	    IFCAP_HWCSUM;
diff --git a/sys/dev/pdq/pdq_ifsubr.c b/sys/dev/pdq/pdq_ifsubr.c
index ec141502a9e..4df00825fe4 100644
--- a/sys/dev/pdq/pdq_ifsubr.c
+++ b/sys/dev/pdq/pdq_ifsubr.c
@@ -471,7 +471,7 @@ pdq_ifattach(pdq_softc_t *sc, const pdq_uint8_t *llc, pdq_type_t type)
     if_initname(ifp, device_get_name(sc->dev), device_get_unit(sc->dev));
     ifp->if_softc = sc;
     ifp->if_init = pdq_ifinit;
-    ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+    ifp->if_snd.ifq_maxlen = ifqmaxlen;
     ifp->if_flags = IFF_BROADCAST|IFF_SIMPLEX|IFF_NOTRAILERS|IFF_MULTICAST;
 
     ifp->if_ioctl = pdq_ifioctl;
diff --git a/sys/dev/ppbus/if_plip.c b/sys/dev/ppbus/if_plip.c
index 3e4c1ca3b2b..fc81aec97b8 100644
--- a/sys/dev/ppbus/if_plip.c
+++ b/sys/dev/ppbus/if_plip.c
@@ -262,7 +262,7 @@ lp_attach(device_t dev)
 	ifp->if_output = lpoutput;
 	ifp->if_hdrlen = 0;
 	ifp->if_addrlen = 0;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	if_attach(ifp);
 
 	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
diff --git a/sys/dev/ral/rt2560.c b/sys/dev/ral/rt2560.c
index 9549107005f..8589768d422 100644
--- a/sys/dev/ral/rt2560.c
+++ b/sys/dev/ral/rt2560.c
@@ -267,8 +267,8 @@ rt2560_attach(device_t dev, int id)
 	ifp->if_init = rt2560_init;
 	ifp->if_ioctl = rt2560_ioctl;
 	ifp->if_start = rt2560_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/ral/rt2661.c b/sys/dev/ral/rt2661.c
index 9bc88385768..9b77c23dc0e 100644
--- a/sys/dev/ral/rt2661.c
+++ b/sys/dev/ral/rt2661.c
@@ -269,8 +269,8 @@ rt2661_attach(device_t dev, int id)
 	ifp->if_init = rt2661_init;
 	ifp->if_ioctl = rt2661_ioctl;
 	ifp->if_start = rt2661_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/sbni/if_sbni.c b/sys/dev/sbni/if_sbni.c
index 8c8687c1295..7ca46d31115 100644
--- a/sys/dev/sbni/if_sbni.c
+++ b/sys/dev/sbni/if_sbni.c
@@ -235,7 +235,7 @@ sbni_attach(struct sbni_softc *sc, int unit, struct sbni_flags flags)
 	ifp->if_init	= sbni_init;
 	ifp->if_start	= sbni_start;
 	ifp->if_ioctl	= sbni_ioctl;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 	/* report real baud rate */
 	csr0 = sbni_inb(sc, CSR0);
diff --git a/sys/dev/smc/if_smc.c b/sys/dev/smc/if_smc.c
index f7f54076a96..6d9ba592679 100644
--- a/sys/dev/smc/if_smc.c
+++ b/sys/dev/smc/if_smc.c
@@ -347,7 +347,7 @@ smc_attach(device_t dev)
 	ifp->if_init = smc_init;
 	ifp->if_ioctl = smc_ioctl;
 	ifp->if_start = smc_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ifp->if_capabilities = ifp->if_capenable = 0;
diff --git a/sys/dev/sn/if_sn.c b/sys/dev/sn/if_sn.c
index 217d811cb0a..0f86f92e76d 100644
--- a/sys/dev/sn/if_sn.c
+++ b/sys/dev/sn/if_sn.c
@@ -207,8 +207,8 @@ sn_attach(device_t dev)
 	ifp->if_ioctl = snioctl;
 	ifp->if_init = sninit;
 	ifp->if_baudrate = 10000000;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ether_ifattach(ifp, eaddr);
diff --git a/sys/dev/snc/dp83932.c b/sys/dev/snc/dp83932.c
index c659c82fd2c..0b62053e2fe 100644
--- a/sys/dev/snc/dp83932.c
+++ b/sys/dev/snc/dp83932.c
@@ -177,7 +177,7 @@ sncconfig(sc, media, nmedia, defmedia, myea)
 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
         ifp->if_init = sncinit;
         ifp->if_mtu = ETHERMTU;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 
 	/* Initialize media goo. */
 	ifmedia_init(&sc->sc_media, 0, snc_mediachange,
diff --git a/sys/dev/usb/net/uhso.c b/sys/dev/usb/net/uhso.c
index a810ec48c5d..21a23c9f6fc 100644
--- a/sys/dev/usb/net/uhso.c
+++ b/sys/dev/usb/net/uhso.c
@@ -1471,8 +1471,8 @@ static int uhso_attach_ifnet(struct uhso_softc *sc, struct usb_interface *iface,
 	ifp->if_output = uhso_if_output;
 	ifp->if_flags = 0;
 	ifp->if_softc = sc;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	if_attach(ifp);
diff --git a/sys/dev/usb/net/usb_ethernet.c b/sys/dev/usb/net/usb_ethernet.c
index 6cf446016e7..bd75cef9657 100644
--- a/sys/dev/usb/net/usb_ethernet.c
+++ b/sys/dev/usb/net/usb_ethernet.c
@@ -214,8 +214,8 @@ ue_attach_post_task(struct usb_proc_msg *_task)
 		ifp->if_ioctl = uether_ioctl;
 	ifp->if_start = ue_start;
 	ifp->if_init = ue_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 	ue->ue_ifp = ifp;
 
diff --git a/sys/dev/usb/wlan/if_rum.c b/sys/dev/usb/wlan/if_rum.c
index 4a5fbae2779..155b76787f0 100644
--- a/sys/dev/usb/wlan/if_rum.c
+++ b/sys/dev/usb/wlan/if_rum.c
@@ -478,8 +478,8 @@ rum_attach(device_t self)
 	ifp->if_init = rum_init;
 	ifp->if_ioctl = rum_ioctl;
 	ifp->if_start = rum_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/usb/wlan/if_run.c b/sys/dev/usb/wlan/if_run.c
index 3ab607fa342..c798c32b823 100644
--- a/sys/dev/usb/wlan/if_run.c
+++ b/sys/dev/usb/wlan/if_run.c
@@ -609,8 +609,8 @@ run_attach(device_t self)
 	ifp->if_init = run_init;
 	ifp->if_ioctl = run_ioctl;
 	ifp->if_start = run_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/usb/wlan/if_uath.c b/sys/dev/usb/wlan/if_uath.c
index 7838d5ad353..40b7c83da10 100644
--- a/sys/dev/usb/wlan/if_uath.c
+++ b/sys/dev/usb/wlan/if_uath.c
@@ -438,8 +438,8 @@ uath_attach(device_t dev)
 	ifp->if_ioctl = uath_ioctl;
 	ifp->if_start = uath_start;
 	/* XXX UATH_TX_DATA_LIST_COUNT */
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic = ifp->if_l2com;
diff --git a/sys/dev/usb/wlan/if_upgt.c b/sys/dev/usb/wlan/if_upgt.c
index 20f06c6ddea..091b9dfe45b 100644
--- a/sys/dev/usb/wlan/if_upgt.c
+++ b/sys/dev/usb/wlan/if_upgt.c
@@ -331,7 +331,7 @@ upgt_attach(device_t dev)
 	ifp->if_init = upgt_init;
 	ifp->if_ioctl = upgt_ioctl;
 	ifp->if_start = upgt_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic = ifp->if_l2com;
diff --git a/sys/dev/usb/wlan/if_ural.c b/sys/dev/usb/wlan/if_ural.c
index 4a4378b90ef..6839bcff1b1 100644
--- a/sys/dev/usb/wlan/if_ural.c
+++ b/sys/dev/usb/wlan/if_ural.c
@@ -467,8 +467,8 @@ ural_attach(device_t self)
 	ifp->if_init = ural_init;
 	ifp->if_ioctl = ural_ioctl;
 	ifp->if_start = ural_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/usb/wlan/if_urtw.c b/sys/dev/usb/wlan/if_urtw.c
index 6604268e009..9ff16f3b137 100644
--- a/sys/dev/usb/wlan/if_urtw.c
+++ b/sys/dev/usb/wlan/if_urtw.c
@@ -867,8 +867,8 @@ urtw_attach(device_t dev)
 	ifp->if_ioctl = urtw_ioctl;
 	ifp->if_start = urtw_start;
 	/* XXX URTW_TX_DATA_LIST_COUNT */
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic = ifp->if_l2com;
diff --git a/sys/dev/usb/wlan/if_zyd.c b/sys/dev/usb/wlan/if_zyd.c
index a0b8179a4a3..1a298f8ff58 100644
--- a/sys/dev/usb/wlan/if_zyd.c
+++ b/sys/dev/usb/wlan/if_zyd.c
@@ -381,7 +381,7 @@ zyd_attach(device_t dev)
 	ifp->if_init = zyd_init;
 	ifp->if_ioctl = zyd_ioctl;
 	ifp->if_start = zyd_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic = ifp->if_l2com;
diff --git a/sys/dev/vx/if_vx.c b/sys/dev/vx/if_vx.c
index 0dea9bf9093..1ea0faf2b11 100644
--- a/sys/dev/vx/if_vx.c
+++ b/sys/dev/vx/if_vx.c
@@ -189,7 +189,7 @@ vx_attach(device_t dev)
 	}
 
 	ifp->if_mtu = ETHERMTU;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	ifp->if_flags = IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST;
 	ifp->if_start = vx_start;
 	ifp->if_ioctl = vx_ioctl;
diff --git a/sys/dev/wi/if_wi.c b/sys/dev/wi/if_wi.c
index 616c5d7b330..fe0d903d173 100644
--- a/sys/dev/wi/if_wi.c
+++ b/sys/dev/wi/if_wi.c
@@ -333,8 +333,8 @@ wi_attach(device_t dev)
 	ifp->if_ioctl = wi_ioctl;
 	ifp->if_start = wi_start;
 	ifp->if_init = wi_init;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ic->ic_ifp = ifp;
diff --git a/sys/dev/wl/if_wl.c b/sys/dev/wl/if_wl.c
index ed291ed8dbd..a95997e4225 100644
--- a/sys/dev/wl/if_wl.c
+++ b/sys/dev/wl/if_wl.c
@@ -559,7 +559,7 @@ wlattach(device_t device)
     ifp->if_init = wlinit;
     ifp->if_start = wlstart;
     ifp->if_ioctl = wlioctl;
-    ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+    ifp->if_snd.ifq_maxlen = ifqmaxlen;
     /* no entries
        ifp->if_done
        ifp->if_reset
diff --git a/sys/dev/wpi/if_wpi.c b/sys/dev/wpi/if_wpi.c
index 9bf9342140a..f6edc916569 100644
--- a/sys/dev/wpi/if_wpi.c
+++ b/sys/dev/wpi/if_wpi.c
@@ -661,8 +661,8 @@ wpi_attach(device_t dev)
 	ifp->if_init = wpi_init;
 	ifp->if_ioctl = wpi_ioctl;
 	ifp->if_start = wpi_start;
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ieee80211_ifattach(ic, macaddr);
diff --git a/sys/dev/xe/if_xe.c b/sys/dev/xe/if_xe.c
index f18fd9842c9..77b953ac886 100644
--- a/sys/dev/xe/if_xe.c
+++ b/sys/dev/xe/if_xe.c
@@ -254,7 +254,7 @@ xe_attach(device_t dev)
 	scp->ifp->if_ioctl = xe_ioctl;
 	scp->ifp->if_init = xe_init;
 	scp->ifp->if_baudrate = 100000000;
-	IFQ_SET_MAXLEN(&scp->ifp->if_snd, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&scp->ifp->if_snd, ifqmaxlen);
 
 	/* Initialise the ifmedia structure */
 	ifmedia_init(scp->ifm, 0, xe_media_change, xe_media_status);
diff --git a/sys/mips/adm5120/if_admsw.c b/sys/mips/adm5120/if_admsw.c
index 164e84008a2..472a80414a7 100644
--- a/sys/mips/adm5120/if_admsw.c
+++ b/sys/mips/adm5120/if_admsw.c
@@ -540,8 +540,8 @@ admsw_attach(device_t dev)
 		ifp->if_init = admsw_init;
 		ifp->if_mtu = ETHERMTU;
 		ifp->if_baudrate = IF_Mbps(100);
-		IFQ_SET_MAXLEN(&ifp->if_snd, max(ADMSW_NTXLDESC, IFQ_MAXLEN));
-		ifp->if_snd.ifq_drv_maxlen = max(ADMSW_NTXLDESC, IFQ_MAXLEN);
+		IFQ_SET_MAXLEN(&ifp->if_snd, max(ADMSW_NTXLDESC, ifqmaxlen));
+		ifp->if_snd.ifq_drv_maxlen = max(ADMSW_NTXLDESC, ifqmaxlen);
 		IFQ_SET_READY(&ifp->if_snd);
 		ifp->if_capabilities |= IFCAP_VLAN_MTU;
 
diff --git a/sys/mips/atheros/if_arge.c b/sys/mips/atheros/if_arge.c
index 4367bfb001d..eef1dcf037b 100644
--- a/sys/mips/atheros/if_arge.c
+++ b/sys/mips/atheros/if_arge.c
@@ -312,8 +312,8 @@ arge_attach(device_t dev)
 	sc->arge_if_flags = ifp->if_flags;
 
 	/* XXX: add real size */
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	ifp->if_capenable = ifp->if_capabilities;
diff --git a/sys/net/if.c b/sys/net/if.c
index 1acfb24dc51..780f2c291aa 100644
--- a/sys/net/if.c
+++ b/sys/net/if.c
@@ -104,6 +104,10 @@ struct ifindex_entry {
 SYSCTL_NODE(_net, PF_LINK, link, CTLFLAG_RW, 0, "Link layers");
 SYSCTL_NODE(_net_link, 0, generic, CTLFLAG_RW, 0, "Generic link-management");
 
+TUNABLE_INT("net.link.ifqmaxlen", &ifqmaxlen);
+SYSCTL_UINT(_net_link, OID_AUTO, ifqmaxlen, CTLFLAG_RDTUN,
+    &ifqmaxlen, 0, "max send queue size");
+
 /* Log link state change events */
 static int log_link_state_change = 1;
 
diff --git a/sys/net/if_ef.c b/sys/net/if_ef.c
index 94fb03d13a4..61f7cb7933f 100644
--- a/sys/net/if_ef.c
+++ b/sys/net/if_ef.c
@@ -129,7 +129,7 @@ ef_attach(struct efnet *sc)
 
 	ifp->if_start = ef_start;
 	ifp->if_init = ef_init;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	ifp->if_flags = (IFF_BROADCAST | IFF_SIMPLEX | IFF_MULTICAST);
 	/*
 	 * Attach the interface
diff --git a/sys/net/if_gif.c b/sys/net/if_gif.c
index ae0ff4e3369..76838396bc4 100644
--- a/sys/net/if_gif.c
+++ b/sys/net/if_gif.c
@@ -185,7 +185,7 @@ gif_clone_create(ifc, unit, params)
 	GIF2IFP(sc)->if_ioctl  = gif_ioctl;
 	GIF2IFP(sc)->if_start  = gif_start;
 	GIF2IFP(sc)->if_output = gif_output;
-	GIF2IFP(sc)->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	GIF2IFP(sc)->if_snd.ifq_maxlen = ifqmaxlen;
 	if_attach(GIF2IFP(sc));
 	bpfattach(GIF2IFP(sc), DLT_NULL, sizeof(u_int32_t));
 	if (ng_gif_attach_p != NULL)
diff --git a/sys/net/if_gre.c b/sys/net/if_gre.c
index bac3d7954fa..1f1e1082d9b 100644
--- a/sys/net/if_gre.c
+++ b/sys/net/if_gre.c
@@ -184,7 +184,7 @@ gre_clone_create(ifc, unit, params)
 	GRE2IFP(sc)->if_softc = sc;
 	if_initname(GRE2IFP(sc), ifc->ifc_name, unit);
 
-	GRE2IFP(sc)->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	GRE2IFP(sc)->if_snd.ifq_maxlen = ifqmaxlen;
 	GRE2IFP(sc)->if_addrlen = 0;
 	GRE2IFP(sc)->if_hdrlen = 24; /* IP + GRE */
 	GRE2IFP(sc)->if_mtu = GREMTU;
diff --git a/sys/net/if_stf.c b/sys/net/if_stf.c
index 5c082c4926a..1ef5581ae62 100644
--- a/sys/net/if_stf.c
+++ b/sys/net/if_stf.c
@@ -252,7 +252,7 @@ stf_clone_create(struct if_clone *ifc, char *name, size_t len, caddr_t params)
 	ifp->if_mtu    = IPV6_MMTU;
 	ifp->if_ioctl  = stf_ioctl;
 	ifp->if_output = stf_output;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	if_attach(ifp);
 	bpfattach(ifp, DLT_NULL, sizeof(u_int32_t));
 	return (0);
diff --git a/sys/net80211/ieee80211.c b/sys/net80211/ieee80211.c
index 3d5669caec5..97ff7c21ad4 100644
--- a/sys/net80211/ieee80211.c
+++ b/sys/net80211/ieee80211.c
@@ -394,8 +394,8 @@ ieee80211_vap_setup(struct ieee80211com *ic, struct ieee80211vap *vap,
 	ifp->if_ioctl = ieee80211_ioctl;
 	ifp->if_init = ieee80211_init;
 	/* NB: input+output filled in by ether_ifattach */
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	vap->iv_ifp = ifp;
diff --git a/sys/netgraph/ng_eiface.c b/sys/netgraph/ng_eiface.c
index 72cc55caae2..4239328967e 100644
--- a/sys/netgraph/ng_eiface.c
+++ b/sys/netgraph/ng_eiface.c
@@ -369,7 +369,7 @@ ng_eiface_constructor(node_p node)
 	ifp->if_output = ether_output;
 	ifp->if_start = ng_eiface_start;
 	ifp->if_ioctl = ng_eiface_ioctl;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	ifp->if_flags = (IFF_SIMPLEX | IFF_BROADCAST | IFF_MULTICAST);
 
 	/* Give this node the same name as the interface (if possible) */
diff --git a/sys/netgraph/ng_fec.c b/sys/netgraph/ng_fec.c
index 919947acab4..ce5a2d58f81 100644
--- a/sys/netgraph/ng_fec.c
+++ b/sys/netgraph/ng_fec.c
@@ -1227,7 +1227,7 @@ ng_fec_constructor(node_p node)
 	ifp->if_start = ng_fec_start;
 	ifp->if_ioctl = ng_fec_ioctl;
 	ifp->if_init = ng_fec_init;
-	ifp->if_snd.ifq_maxlen = IFQ_MAXLEN;
+	ifp->if_snd.ifq_maxlen = ifqmaxlen;
 	ifp->if_mtu = NG_FEC_MTU_DEFAULT;
 	ifp->if_flags = (IFF_SIMPLEX|IFF_BROADCAST|IFF_MULTICAST);
 	ifp->if_addrlen = 0;			/* XXX */
diff --git a/sys/netgraph/ng_iface.c b/sys/netgraph/ng_iface.c
index d53bf768b10..f94fd14930f 100644
--- a/sys/netgraph/ng_iface.c
+++ b/sys/netgraph/ng_iface.c
@@ -564,8 +564,8 @@ ng_iface_constructor(node_p node)
 	ifp->if_addrlen = 0;			/* XXX */
 	ifp->if_hdrlen = 0;			/* XXX */
 	ifp->if_baudrate = 64000;		/* XXX */
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	/* Give this node the same name as the interface (if possible) */
diff --git a/sys/netgraph/ng_tty.c b/sys/netgraph/ng_tty.c
index 2fa33fc8005..84083170d8a 100644
--- a/sys/netgraph/ng_tty.c
+++ b/sys/netgraph/ng_tty.c
@@ -164,7 +164,7 @@ ngt_constructor(node_p node)
 	sc->node = node;
 
 	mtx_init(&sc->outq.ifq_mtx, "ng_tty node+queue", NULL, MTX_DEF);
-	IFQ_SET_MAXLEN(&sc->outq, IFQ_MAXLEN);
+	IFQ_SET_MAXLEN(&sc->outq, ifqmaxlen);
 
 	return (0);
 }
diff --git a/sys/pci/if_rl.c b/sys/pci/if_rl.c
index 8b225169a9c..cd7609db697 100644
--- a/sys/pci/if_rl.c
+++ b/sys/pci/if_rl.c
@@ -942,8 +942,8 @@ rl_attach(device_t dev)
 #ifdef DEVICE_POLLING
 	ifp->if_capabilities |= IFCAP_POLLING;
 #endif
-	IFQ_SET_MAXLEN(&ifp->if_snd, IFQ_MAXLEN);
-	ifp->if_snd.ifq_drv_maxlen = IFQ_MAXLEN;
+	IFQ_SET_MAXLEN(&ifp->if_snd, ifqmaxlen);
+	ifp->if_snd.ifq_drv_maxlen = ifqmaxlen;
 	IFQ_SET_READY(&ifp->if_snd);
 
 	/*

From 54c7282725e84db4062a1aa3756a47b5574e1968 Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Mon, 3 May 2010 07:39:51 +0000
Subject: [PATCH 417/532] Code indent according to style(9).

PR:		bin/146186
Submitted by:	myself
Approved by:	delphij (mentor)
MFC after:	2 weeks
---
 lib/libpam/modules/pam_krb5/pam_krb5.c | 46 +++++++++++++-------------
 1 file changed, 23 insertions(+), 23 deletions(-)

diff --git a/lib/libpam/modules/pam_krb5/pam_krb5.c b/lib/libpam/modules/pam_krb5/pam_krb5.c
index 9623a17191f..439fcf9f1a2 100644
--- a/lib/libpam/modules/pam_krb5/pam_krb5.c
+++ b/lib/libpam/modules/pam_krb5/pam_krb5.c
@@ -199,33 +199,33 @@ pam_sm_authenticate(pam_handle_t *pamh, int flags __unused,
 		PAM_LOG("Skipping local user check");
 	else {
 
-	/* Verify the local user exists (AFTER getting the password) */
-	if (strchr(user, '@')) {
-		/* get a local account name for this principal */
-		krbret = krb5_aname_to_localname(pam_context, princ,
-		    sizeof(luser), luser);
-		if (krbret != 0) {
-			PAM_VERBOSE_ERROR("Kerberos 5 error");
-			PAM_LOG("Error krb5_aname_to_localname(): %s",
-			    krb5_get_err_text(pam_context, krbret));
+		/* Verify the local user exists (AFTER getting the password) */
+		if (strchr(user, '@')) {
+			/* get a local account name for this principal */
+			krbret = krb5_aname_to_localname(pam_context, princ,
+			    sizeof(luser), luser);
+			if (krbret != 0) {
+				PAM_VERBOSE_ERROR("Kerberos 5 error");
+				PAM_LOG("Error krb5_aname_to_localname(): %s",
+				    krb5_get_err_text(pam_context, krbret));
+				retval = PAM_USER_UNKNOWN;
+				goto cleanup2;
+			}
+
+			retval = pam_set_item(pamh, PAM_USER, luser);
+			if (retval != PAM_SUCCESS)
+				goto cleanup2;
+
+			PAM_LOG("PAM_USER Redone");
+		}
+
+		pwd = getpwnam(user);
+		if (pwd == NULL) {
 			retval = PAM_USER_UNKNOWN;
 			goto cleanup2;
 		}
 
-		retval = pam_set_item(pamh, PAM_USER, luser);
-		if (retval != PAM_SUCCESS)
-			goto cleanup2;
-
-		PAM_LOG("PAM_USER Redone");
-	}
-
-	pwd = getpwnam(user);
-	if (pwd == NULL) {
-		retval = PAM_USER_UNKNOWN;
-		goto cleanup2;
-	}
-
-	PAM_LOG("Done getpwnam()");
+		PAM_LOG("Done getpwnam()");
 	}
 
 	/* Get a TGT */

From f93beda98af381f5ce34d8bf3322669895857d9f Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Mon, 3 May 2010 09:49:42 +0000
Subject: [PATCH 418/532] Bump .Dd date.

Forgotten by:	delphij
---
 lib/libpam/modules/pam_krb5/pam_krb5.8 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/lib/libpam/modules/pam_krb5/pam_krb5.8 b/lib/libpam/modules/pam_krb5/pam_krb5.8
index a182e62e70d..bd7ac5b9ca0 100644
--- a/lib/libpam/modules/pam_krb5/pam_krb5.8
+++ b/lib/libpam/modules/pam_krb5/pam_krb5.8
@@ -1,7 +1,7 @@
 .\"
 .\" $Id: pam_krb5.5,v 1.5 2000/01/05 00:59:56 fcusack Exp $
 .\" $FreeBSD$
-.Dd January 15, 1999
+.Dd May 3, 2010
 .Dt PAM_KRB5 8
 .Os
 .Sh NAME

From eddc400373205638b2812283866f2cf8ddcd50c0 Mon Sep 17 00:00:00 2001
From: Alexander Leidinger <netchild@FreeBSD.org>
Date: Mon, 3 May 2010 14:19:58 +0000
Subject: [PATCH 419/532] - #ifdef out the cliplist part, skype seems like
 using an uninitialized   variable and can cause problems, without the
 cliplist handling it works   without problems - improve the cliplist error
 handling - fix VIDIOCGTUNER and VIDIOCSMICROCODE (still no hardware available
 to test)

Submitted by:	J.R. Oldroyd <jr@opal.com>
X-MFC after:	soon (together with all the v4l stuff)
---
 sys/compat/linux/linux_ioctl.c | 31 ++++++++++++++++++++++++-------
 1 file changed, 24 insertions(+), 7 deletions(-)

diff --git a/sys/compat/linux/linux_ioctl.c b/sys/compat/linux/linux_ioctl.c
index 66009763404..261126bf706 100644
--- a/sys/compat/linux/linux_ioctl.c
+++ b/sys/compat/linux/linux_ioctl.c
@@ -2628,6 +2628,7 @@ bsd_to_linux_v4l_tuner(struct video_tuner *vt, struct l_video_tuner *lvt)
 	return (0);
 }
 
+#ifdef COMPAT_LINUX_V4L_CLIPLIST
 static int
 linux_to_bsd_v4l_clip(struct l_video_clip *lvc, struct video_clip *vc)
 {
@@ -2638,6 +2639,7 @@ linux_to_bsd_v4l_clip(struct l_video_clip *lvc, struct video_clip *vc)
 	vc->next = PTRIN(lvc->next);	/* possible pointer size conversion */
 	return (0);
 }
+#endif
 
 static int
 linux_to_bsd_v4l_window(struct l_video_window *lvw, struct video_window *vw)
@@ -2698,6 +2700,7 @@ linux_to_bsd_v4l_code(struct l_video_code *lvc, struct video_code *vc)
 	return (0);
 }
 
+#ifdef COMPAT_LINUX_V4L_CLIPLIST
 static int
 linux_v4l_clip_copy(void *lvc, struct video_clip **ppvc)
 {
@@ -2772,15 +2775,18 @@ linux_v4l_cliplist_copy(struct l_video_window *lvw, struct video_window *vw)
 		 *	example of cliplist use.
 		 */
 		plvc = PTRIN(lvw->clips);
+		vw->clips = NULL;
 		ppvc = &(vw->clips);
 		while (clipcount-- > 0) {
-			if (plvc == 0)
+			if (plvc == 0) {
 				error = EFAULT;
-			if (!error)
-				error = linux_v4l_clip_copy(plvc, ppvc);
-			if (error) {
-				linux_v4l_cliplist_free(vw);
 				break;
+			} else {
+				error = linux_v4l_clip_copy(plvc, ppvc);
+				if (error) {
+					linux_v4l_cliplist_free(vw);
+					break;
+				}
 			}
 			ppvc = &((*ppvc)->next);
 		        plvc = PTRIN(((struct l_video_clip *) plvc)->next);
@@ -2795,6 +2801,7 @@ linux_v4l_cliplist_copy(struct l_video_window *lvw, struct video_window *vw)
 	}
 	return (error);
 }
+#endif
 
 static int
 linux_ioctl_v4l(struct thread *td, struct linux_ioctl_args *args)
@@ -2818,6 +2825,12 @@ linux_ioctl_v4l(struct thread *td, struct linux_ioctl_args *args)
 	case LINUX_VIDIOCGTUNER:
 		if ((error = fget(td, args->fd, &fp)) != 0)
 			return (error);
+		error = copyin((void *) args->arg, &l_vtun, sizeof(l_vtun));
+		if (error) {
+			fdrop(fp, td);
+			return (error);
+		}
+		linux_to_bsd_v4l_tuner(&l_vtun, &vtun);
 		error = fo_ioctl(fp, VIDIOCGTUNER, &vtun, td->td_ucred, td);
 		if (!error) {
 			bsd_to_linux_v4l_tuner(&vtun, &l_vtun);
@@ -2836,7 +2849,7 @@ linux_ioctl_v4l(struct thread *td, struct linux_ioctl_args *args)
 			return (error);
 		}
 		linux_to_bsd_v4l_tuner(&l_vtun, &vtun);
-		error = fo_ioctl(fp, VIDIOCSMICROCODE, &vtun, td->td_ucred, td);
+		error = fo_ioctl(fp, VIDIOCSTUNER, &vtun, td->td_ucred, td);
 		fdrop(fp, td);
 		return (error);
 
@@ -2865,14 +2878,18 @@ linux_ioctl_v4l(struct thread *td, struct linux_ioctl_args *args)
 			return (error);
 		}
 		linux_to_bsd_v4l_window(&l_vwin, &vwin);
+#ifdef COMPAT_LINUX_V4L_CLIPLIST
 		error = linux_v4l_cliplist_copy(&l_vwin, &vwin);
 		if (error) {
 			fdrop(fp, td);
 			return (error);
 		}
+#endif
 		error = fo_ioctl(fp, VIDIOCSWIN, &vwin, td->td_ucred, td);
 		fdrop(fp, td);
+#ifdef COMPAT_LINUX_V4L_CLIPLIST
 		linux_v4l_cliplist_free(&vwin);
+#endif
 		return (error);
 
 	case LINUX_VIDIOCGFBUF:
@@ -2924,7 +2941,7 @@ linux_ioctl_v4l(struct thread *td, struct linux_ioctl_args *args)
 			return (error);
 		}
 		linux_to_bsd_v4l_code(&l_vcode, &vcode);
-		error = fo_ioctl(fp, VIDIOCSTUNER, &vcode, td->td_ucred, td);
+		error = fo_ioctl(fp, VIDIOCSMICROCODE, &vcode, td->td_ucred, td);
 		fdrop(fp, td);
 		return (error);
 

From bf6f1b56c56b12212ddf22dabe23fee448174bf4 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Mon, 3 May 2010 14:30:49 +0000
Subject: [PATCH 420/532] Style and comment adjustements.

Suggested and reviewed by:	bde
MFC after:	3 days
---
 sys/amd64/amd64/exception.S | 79 +++++++++++++++++++++----------------
 1 file changed, 45 insertions(+), 34 deletions(-)

diff --git a/sys/amd64/amd64/exception.S b/sys/amd64/amd64/exception.S
index 0d5a50ddc0c..7b1fed87a76 100644
--- a/sys/amd64/amd64/exception.S
+++ b/sys/amd64/amd64/exception.S
@@ -50,14 +50,14 @@
 	.bss
 	.globl	dtrace_invop_jump_addr
 	.align	8
-	.type	dtrace_invop_jump_addr, @object
-        .size	dtrace_invop_jump_addr, 8
+	.type	dtrace_invop_jump_addr,@object
+	.size	dtrace_invop_jump_addr,8
 dtrace_invop_jump_addr:
 	.zero	8
 	.globl	dtrace_invop_calltrap_addr
 	.align	8
-	.type	dtrace_invop_calltrap_addr, @object
-        .size	dtrace_invop_calltrap_addr, 8
+	.type	dtrace_invop_calltrap_addr,@object
+	.size	dtrace_invop_calltrap_addr,8
 dtrace_invop_calltrap_addr:
 	.zero	8
 #endif
@@ -157,7 +157,6 @@ IDTVEC(align)
 	 * kernel from userland.  Reenable interrupts if they were enabled
 	 * before the trap.  This approximates SDT_SYS386TGT on the i386 port.
 	 */
-
 	SUPERALIGN_TEXT
 	.globl	alltraps
 	.type	alltraps,@function
@@ -211,16 +210,16 @@ alltraps_pushregs_no_rdi:
 	 * Set our jump address for the jump back in the event that
 	 * the breakpoint wasn't caused by DTrace at all.
 	 */
-	movq	$calltrap, dtrace_invop_calltrap_addr(%rip)
+	movq	$calltrap,dtrace_invop_calltrap_addr(%rip)
 
 	/* Jump to the code hooked in by DTrace. */
-	movq	dtrace_invop_jump_addr, %rax
+	movq	dtrace_invop_jump_addr,%rax
 	jmpq	*dtrace_invop_jump_addr
 #endif
 	.globl	calltrap
 	.type	calltrap,@function
 calltrap:
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	MEXITCOUNT
 	jmp	doreti			/* Handle any pending ASTs */
@@ -274,9 +273,11 @@ IDTVEC(dblfault)
 	testb	$SEL_RPL_MASK,TF_CS(%rsp) /* Did we come from kernel? */
 	jz	1f			/* already running with kernel GS.base */
 	swapgs
-1:	movq	%rsp, %rdi
+1:
+	movq	%rsp,%rdi
 	call	dblfault_handler
-2:	hlt
+2:
+	hlt
 	jmp	2b
 
 IDTVEC(page)
@@ -369,7 +370,7 @@ IDTVEC(fast_syscall)
 	movq	%r15,TF_R15(%rsp)	/* C preserved */
 	movl	$TF_HASSEGS,TF_FLAGS(%rsp)
 	FAKE_MCOUNT(TF_RIP(%rsp))
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	syscall
 	movq	PCPU(CURPCB),%rax
 	andq	$~PCB_FULLCTX,PCB_FLAGS(%rax)
@@ -456,7 +457,7 @@ nmi_fromuserspace:
 /* Note: this label is also used by ddb and gdb: */
 nmi_calltrap:
 	FAKE_MCOUNT(TF_RIP(%rsp))
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	MEXITCOUNT
 #ifdef HWPMC_HOOKS
@@ -555,9 +556,9 @@ nmi_restoreregs:
 	iretq
 
 ENTRY(fork_trampoline)
-	movq	%r12, %rdi		/* function */
-	movq	%rbx, %rsi		/* arg1 */
-	movq	%rsp, %rdx		/* trapframe pointer */
+	movq	%r12,%rdi		/* function */
+	movq	%rbx,%rsi		/* arg1 */
+	movq	%rsp,%rdx		/* trapframe pointer */
 	call	fork_exit
 	MEXITCOUNT
 	jmp	doreti			/* Handle any ASTs */
@@ -628,7 +629,7 @@ doreti_ast:
 	testl	$TDF_ASTPENDING | TDF_NEEDRESCHED,TD_FLAGS(%rax)
 	je	doreti_exit
 	sti
-	movq	%rsp, %rdi			/* pass a pointer to the trapframe */
+	movq	%rsp,%rdi	/* pass a pointer to the trapframe */
 	call	ast
 	jmp	doreti_ast
 
@@ -648,8 +649,8 @@ doreti_exit:
 	 * Do not reload segment registers for kernel.
 	 * Since we do not reload segments registers with sane
 	 * values on kernel entry, descriptors referenced by
-	 * segments registers may be not valid. This is fatal
-	 * for the usermode, but is innocent for the kernel.
+	 * segments registers might be not valid.  This is fatal
+	 * for user mode, but is not a problem for the kernel.
 	 */
 	testb	$SEL_RPL_MASK,TF_CS(%rsp)
 	jz	ld_regs
@@ -662,14 +663,16 @@ do_segs:
 	/* Restore %fs and fsbase */
 	movw	TF_FS(%rsp),%ax
 	.globl	ld_fs
-ld_fs:	movw	%ax,%fs
+ld_fs:
+	movw	%ax,%fs
 	cmpw	$KUF32SEL,%ax
 	jne	1f
 	movl	$MSR_FSBASE,%ecx
 	movl	PCB_FSBASE(%r8),%eax
 	movl	PCB_FSBASE+4(%r8),%edx
 	.globl	ld_fsbase
-ld_fsbase: wrmsr
+ld_fsbase:
+	wrmsr
 1:
 	/* Restore %gs and gsbase */
 	movw	TF_GS(%rsp),%si
@@ -678,7 +681,8 @@ ld_fsbase: wrmsr
 	movl	$MSR_GSBASE,%ecx
 	rdmsr
 	.globl	ld_gs
-ld_gs:	movw	%si,%gs
+ld_gs:
+	movw	%si,%gs
 	wrmsr
 	popfq
 	cmpw	$KUG32SEL,%si
@@ -687,12 +691,17 @@ ld_gs:	movw	%si,%gs
 	movl	PCB_GSBASE(%r8),%eax
 	movl	PCB_GSBASE+4(%r8),%edx
 	.globl	ld_gsbase
-ld_gsbase: wrmsr
-1:	.globl	ld_es
-ld_es:	movw	TF_ES(%rsp),%es
+ld_gsbase:
+	wrmsr
+1:
+	.globl	ld_es
+ld_es:
+	movw	TF_ES(%rsp),%es
 	.globl	ld_ds
-ld_ds:	movw	TF_DS(%rsp),%ds
-ld_regs:movq	TF_RDI(%rsp),%rdi
+ld_ds:
+	movw	TF_DS(%rsp),%ds
+ld_regs:
+	movq	TF_RDI(%rsp),%rdi
 	movq	TF_RSI(%rsp),%rsi
 	movq	TF_RDX(%rsp),%rdx
 	movq	TF_RCX(%rsp),%rcx
@@ -711,7 +720,8 @@ ld_regs:movq	TF_RDI(%rsp),%rdi
 	jz	1f			/* keep running with kernel GS.base */
 	cli
 	swapgs
-1:	addq	$TF_RIP,%rsp		/* skip over tf_err, tf_trapno */
+1:
+	addq	$TF_RIP,%rsp		/* skip over tf_err, tf_trapno */
 	.globl	doreti_iret
 doreti_iret:
 	iretq
@@ -738,7 +748,8 @@ doreti_iret_fault:
 	testl	$PSL_I,TF_RFLAGS(%rsp)
 	jz	1f
 	sti
-1:	movw	%fs,TF_FS(%rsp)
+1:
+	movw	%fs,TF_FS(%rsp)
 	movw	%gs,TF_GS(%rsp)
 	movw	%es,TF_ES(%rsp)
 	movw	%ds,TF_DS(%rsp)
@@ -768,7 +779,7 @@ doreti_iret_fault:
 	.globl	ds_load_fault
 ds_load_fault:
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movw	$KUDSEL,TF_DS(%rsp)
 	jmp	doreti
@@ -777,7 +788,7 @@ ds_load_fault:
 	.globl	es_load_fault
 es_load_fault:
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movw	$KUDSEL,TF_ES(%rsp)
 	jmp	doreti
@@ -786,7 +797,7 @@ es_load_fault:
 	.globl	fs_load_fault
 fs_load_fault:
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movw	$KUF32SEL,TF_FS(%rsp)
 	jmp	doreti
@@ -796,7 +807,7 @@ fs_load_fault:
 gs_load_fault:
 	popfq
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movw	$KUG32SEL,TF_GS(%rsp)
 	jmp	doreti
@@ -805,7 +816,7 @@ gs_load_fault:
 	.globl	fsbase_load_fault
 fsbase_load_fault:
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movq	PCPU(CURTHREAD),%r8
 	movq	TD_PCB(%r8),%r8
@@ -816,7 +827,7 @@ fsbase_load_fault:
 	.globl	gsbase_load_fault
 gsbase_load_fault:
 	movl	$T_PROTFLT,TF_TRAPNO(%rsp)
-	movq	%rsp, %rdi
+	movq	%rsp,%rdi
 	call	trap
 	movq	PCPU(CURTHREAD),%r8
 	movq	TD_PCB(%r8),%r8

From 492d01762f9243f7a112f8ab3935f1e0b8d43e49 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 15:51:59 +0000
Subject: [PATCH 421/532] Neither the page lock nor the page queues lock is
 required to unwire and free a VM_ALLOC_NOOBJ page.  (Such pages are
 unmanaged.)

---
 sys/dev/ti/if_ti.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/sys/dev/ti/if_ti.c b/sys/dev/ti/if_ti.c
index 20130bcd15f..7eeee36fe27 100644
--- a/sys/dev/ti/if_ti.c
+++ b/sys/dev/ti/if_ti.c
@@ -1488,10 +1488,8 @@ ti_newbuf_jumbo(sc, idx, m_old)
 			}
 			sf[i] = sf_buf_alloc(frame, SFB_NOWAIT);
 			if (sf[i] == NULL) {
-				vm_page_lock_queues();
 				vm_page_unwire(frame, 0);
 				vm_page_free(frame);
-				vm_page_unlock_queues();
 				device_printf(sc->ti_dev, "buffer allocation "
 				    "failed -- packet dropped!\n");
 				printf("      index %d page %d\n", idx, i);

From a3f93b7269debfea7808b22094ab217561b26d8b Mon Sep 17 00:00:00 2001
From: Marko Zec <zec@FreeBSD.org>
Date: Mon, 3 May 2010 16:08:24 +0000
Subject: [PATCH 422/532] When destroying a vnet, shut down all netgraph nodes
 tied to that vnet before proceeding with dismantling other protocol domains.

This change only affects options VIMAGE builds.

Reviewed by:	julian, bz
MFC after:	3 days
---
 sys/netgraph/ng_base.c | 52 +++++++++++++++++++++++++++---------------
 1 file changed, 33 insertions(+), 19 deletions(-)

diff --git a/sys/netgraph/ng_base.c b/sys/netgraph/ng_base.c
index fdfe878dbc9..2a93e588820 100644
--- a/sys/netgraph/ng_base.c
+++ b/sys/netgraph/ng_base.c
@@ -3067,28 +3067,42 @@ ng_mod_event(module_t mod, int event, void *data)
 static void
 vnet_netgraph_uninit(const void *unused __unused)
 {
-#if 0
-	node_p node, last_killed = NULL;
+	node_p node = NULL, last_killed = NULL;
+	int i;
 
-	/* XXXRW: utterly bogus. */
-	while ((node = LIST_FIRST(&V_ng_allnodes)) != NULL) {
-		if (node == last_killed) {
-			/* This should never happen */
-			node->nd_flags |= NGF_REALLY_DIE;
-			printf("netgraph node %s needs NGF_REALLY_DIE\n",
-			    node->nd_name);
-			ng_rmnode(node, NULL, NULL, 0);
-			/* This must never happen */
-			if (node == LIST_FIRST(&V_ng_allnodes))
-				panic("netgraph node %s won't die",
-				    node->nd_name);
+	do {
+		/* Find a node to kill */
+		mtx_lock(&ng_namehash_mtx);
+		for (i = 0; i < NG_NAME_HASH_SIZE; i++) {
+			LIST_FOREACH(node, &V_ng_name_hash[i], nd_nodes) {
+				if (node != &ng_deadnode) {
+					NG_NODE_REF(node);
+					break;
+				}
+			}
+			if (node != NULL)
+				break;
 		}
-		ng_rmnode(node, NULL, NULL, 0);
-		last_killed = node;
-	}
-#endif
+		mtx_unlock(&ng_namehash_mtx);
+
+		/* Attempt to kill it only if it is a regular node */
+		if (node != NULL) {
+			if (node == last_killed) {
+				/* This should never happen */
+				printf("ng node %s needs"
+				    "NGF_REALLY_DIE\n", node->nd_name);
+				if (node->nd_flags & NGF_REALLY_DIE)
+					panic("ng node %s won't die",
+					    node->nd_name);
+				node->nd_flags |= NGF_REALLY_DIE;
+			}
+			ng_rmnode(node, NULL, NULL, 0);
+			NG_NODE_UNREF(node);
+			last_killed = node;
+		}
+	} while (node != NULL);
 }
-VNET_SYSUNINIT(vnet_netgraph_uninit, SI_SUB_NETGRAPH, SI_ORDER_ANY,
+VNET_SYSUNINIT(vnet_netgraph_uninit, SI_SUB_PROTO_IFATTACHDOMAIN, SI_ORDER_ANY,
     vnet_netgraph_uninit, NULL);
 #endif /* VIMAGE */
 

From c5a648516efef5bd582a816cb7c36ce070e4583c Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 16:41:11 +0000
Subject: [PATCH 423/532] Acquire the page lock around vm_page_unwire() and
 vm_page_wire().

Reviewed by:	kib
---
 sys/fs/tmpfs/tmpfs_vnops.c |  4 ++++
 sys/kern/vfs_bio.c         | 13 +++++++++----
 2 files changed, 13 insertions(+), 4 deletions(-)

diff --git a/sys/fs/tmpfs/tmpfs_vnops.c b/sys/fs/tmpfs/tmpfs_vnops.c
index 3d3755d0528..2d6005a5958 100644
--- a/sys/fs/tmpfs/tmpfs_vnops.c
+++ b/sys/fs/tmpfs/tmpfs_vnops.c
@@ -460,9 +460,11 @@ tmpfs_nocacheread(vm_object_t tobj, vm_pindex_t idx,
 	error = uiomove_fromphys(&m, offset, tlen, uio);
 	VM_OBJECT_LOCK(tobj);
 out:
+	vm_page_lock(m);
 	vm_page_lock_queues();
 	vm_page_unwire(m, TRUE);
 	vm_page_unlock_queues();
+	vm_page_unlock(m);
 	vm_page_wakeup(m);
 	vm_object_pip_subtract(tobj, 1);
 	VM_OBJECT_UNLOCK(tobj);
@@ -691,6 +693,7 @@ nocache:
 out:
 	if (vobj != NULL)
 		VM_OBJECT_LOCK(vobj);
+	vm_page_lock(tpg);
 	vm_page_lock_queues();
 	if (error == 0) {
 		KASSERT(tpg->valid == VM_PAGE_BITS_ALL,
@@ -699,6 +702,7 @@ out:
 	}
 	vm_page_unwire(tpg, TRUE);
 	vm_page_unlock_queues();
+	vm_page_unlock(tpg);
 	vm_page_wakeup(tpg);
 	if (vpg != NULL)
 		vm_page_wakeup(vpg);
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 10489985388..5db2d9e9773 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -2942,7 +2942,6 @@ allocbuf(struct buf *bp, int size)
 				vm_page_t m;
 
 				VM_OBJECT_LOCK(bp->b_bufobj->bo_object);
-				vm_page_lock_queues();
 				for (i = desiredpages; i < bp->b_npages; i++) {
 					/*
 					 * the page is not freed here -- it
@@ -2952,13 +2951,17 @@ allocbuf(struct buf *bp, int size)
 					m = bp->b_pages[i];
 					KASSERT(m != bogus_page,
 					    ("allocbuf: bogus page found"));
-					while (vm_page_sleep_if_busy(m, TRUE, "biodep"))
-						vm_page_lock_queues();
+					while (vm_page_sleep_if_busy(m, TRUE,
+					    "biodep"))
+						continue;
 
 					bp->b_pages[i] = NULL;
+					vm_page_lock(m);
+					vm_page_lock_queues();
 					vm_page_unwire(m, 0);
+					vm_page_unlock_queues();
+					vm_page_unlock(m);
 				}
-				vm_page_unlock_queues();
 				VM_OBJECT_UNLOCK(bp->b_bufobj->bo_object);
 				pmap_qremove((vm_offset_t) trunc_page((vm_offset_t)bp->b_data) +
 				    (desiredpages << PAGE_SHIFT), (bp->b_npages - desiredpages));
@@ -3039,9 +3042,11 @@ allocbuf(struct buf *bp, int size)
 				/*
 				 * We have a good page.
 				 */
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				vm_page_wire(m);
 				vm_page_unlock_queues();
+				vm_page_unlock(m);
 				bp->b_pages[bp->b_npages] = m;
 				++bp->b_npages;
 			}

From 5fdd0a335f567c95f479d2739e1df0f664626f21 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 16:55:50 +0000
Subject: [PATCH 424/532] Acquire the page lock around vm_page_unwire().  For
 consistency, extend the scope of the object lock in agp_i810.c.  (In this
 specific case, the scope of the object lock shouldn't matter, but I don't
 want to create a bad example that might be copied to a case where it did
 matter.)

Reviewed by:	kib
---
 sys/dev/agp/agp.c         | 4 ++++
 sys/dev/agp/agp_i810.c    | 4 +++-
 sys/dev/drm/via_dmablit.c | 2 ++
 3 files changed, 9 insertions(+), 1 deletion(-)

diff --git a/sys/dev/agp/agp.c b/sys/dev/agp/agp.c
index c992c931846..89b68b000e1 100644
--- a/sys/dev/agp/agp.c
+++ b/sys/dev/agp/agp.c
@@ -623,9 +623,11 @@ bad:
 		m = vm_page_lookup(mem->am_obj, OFF_TO_IDX(k));
 		if (k >= i)
 			vm_page_wakeup(m);
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(mem->am_obj);
 
@@ -657,9 +659,11 @@ agp_generic_unbind_memory(device_t dev, struct agp_memory *mem)
 	VM_OBJECT_LOCK(mem->am_obj);
 	for (i = 0; i < mem->am_size; i += PAGE_SIZE) {
 		m = vm_page_lookup(mem->am_obj, atop(i));
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(mem->am_obj);
 		
diff --git a/sys/dev/agp/agp_i810.c b/sys/dev/agp/agp_i810.c
index 95a9cc40c9f..44eb50da27f 100644
--- a/sys/dev/agp/agp_i810.c
+++ b/sys/dev/agp/agp_i810.c
@@ -1010,10 +1010,12 @@ agp_i810_free_memory(device_t dev, struct agp_memory *mem)
 	
 			VM_OBJECT_LOCK(mem->am_obj);
 			m = vm_page_lookup(mem->am_obj, 0);
-			VM_OBJECT_UNLOCK(mem->am_obj);
+			vm_page_lock(m);
 			vm_page_lock_queues();
 			vm_page_unwire(m, 0);
 			vm_page_unlock_queues();
+			vm_page_unlock(m);
+			VM_OBJECT_UNLOCK(mem->am_obj);
 		} else {
 			contigfree(sc->argb_cursor, mem->am_size, M_AGP);
 			sc->argb_cursor = NULL;
diff --git a/sys/dev/drm/via_dmablit.c b/sys/dev/drm/via_dmablit.c
index 3a66b261574..72d914eacd8 100644
--- a/sys/dev/drm/via_dmablit.c
+++ b/sys/dev/drm/via_dmablit.c
@@ -178,9 +178,11 @@ via_free_sg_info(drm_via_sg_info_t *vsg)
 	case dr_via_pages_locked:
 		for (i=0; i < vsg->num_pages; ++i) {
 			if ( NULL != (page = vsg->pages[i])) {
+				vm_page_lock(page);
 				vm_page_lock_queues();
 				vm_page_unwire(page, 0);
 				vm_page_unlock_queues();
+				vm_page_unlock(page);
 			}
 		}
 	case dr_via_pages_alloc:

From 451033a48ac5930e486f46db2235bd01f9eca9c3 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 17:35:31 +0000
Subject: [PATCH 425/532] It makes more sense for the object-based backend
 allocator to use OBJT_PHYS objects instead of OBJT_DEFAULT objects because we
 never reclaim or pageout the allocated pages.  Moreover, they are mapped with
 pmap_qenter(), which creates unmanaged mappings.

Reviewed by:	kib
---
 sys/vm/uma_core.c | 14 ++++----------
 1 file changed, 4 insertions(+), 10 deletions(-)

diff --git a/sys/vm/uma_core.c b/sys/vm/uma_core.c
index 00a8276d36f..e1b9a08109d 100644
--- a/sys/vm/uma_core.c
+++ b/sys/vm/uma_core.c
@@ -1022,12 +1022,8 @@ obj_alloc(uma_zone_t zone, int bytes, u_int8_t *flags, int wait)
 			while (pages != startpages) {
 				pages--;
 				p = TAILQ_LAST(&object->memq, pglist);
-				vm_page_lock(p);
-				vm_page_lock_queues();
 				vm_page_unwire(p, 0);
 				vm_page_free(p);
-				vm_page_unlock_queues();
-				vm_page_unlock(p);
 			}
 			retkva = 0;
 			goto done;
@@ -2893,13 +2889,11 @@ uma_zone_set_obj(uma_zone_t zone, struct vm_object *obj, int count)
 
 	if (kva == 0)
 		return (0);
-	if (obj == NULL) {
-		obj = vm_object_allocate(OBJT_DEFAULT,
-		    pages);
-	} else {
+	if (obj == NULL)
+		obj = vm_object_allocate(OBJT_PHYS, pages);
+	else {
 		VM_OBJECT_LOCK_INIT(obj, "uma object");
-		_vm_object_allocate(OBJT_DEFAULT,
-		    pages, obj);
+		_vm_object_allocate(OBJT_PHYS, pages, obj);
 	}
 	ZONE_LOCK(zone);
 	keg->uk_kva = kva;

From 2d5d7f7f61f2217db46e1f9673aecc111d651df9 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Mon, 3 May 2010 17:55:32 +0000
Subject: [PATCH 426/532] Acquire the page lock around vm_page_wire() in
 vm_page_grab().

Assert that the page lock is held in vm_page_wire().
---
 sys/vm/vm_page.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index acec3b7263b..e18e7ba1b3b 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1544,6 +1544,7 @@ vm_page_wire(vm_page_t m)
 	 * it is already off the queues).
 	 */
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	if (m->flags & PG_FICTITIOUS)
 		return;
 	if (m->wire_count == 0) {
@@ -1914,9 +1915,11 @@ retrylookup:
 			goto retrylookup;
 		} else {
 			if ((allocflags & VM_ALLOC_WIRED) != 0) {
+				vm_page_lock(m);
 				vm_page_lock_queues();
 				vm_page_wire(m);
 				vm_page_unlock_queues();
+				vm_page_unlock(m);
 			}
 			if ((allocflags & VM_ALLOC_NOBUSY) == 0)
 				vm_page_busy(m);

From 59879e52c1d56b4c7e7ff1cee46770722d254fe8 Mon Sep 17 00:00:00 2001
From: Alexander Kabaev <kan@FreeBSD.org>
Date: Mon, 3 May 2010 18:04:17 +0000
Subject: [PATCH 427/532] Do not encode more than CTF_MAX_VLEN(1023) enum
 members.

CTF can not represent enums with more than CTF_MAX_VLEN members, but
ctfconvert will happily ignore that limitation and create CTF section no
other tool can interpret.

This change is different from similar change from upstream, which just
returns an error if big enum is encountered.  Doing that means that
every FreeBSD kernel with compiled in hwpmc will have no useable CTF
information due to pmc_event enum having 1236+ members.
---
 cddl/contrib/opensolaris/tools/ctf/cvt/ctf.c | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/cddl/contrib/opensolaris/tools/ctf/cvt/ctf.c b/cddl/contrib/opensolaris/tools/ctf/cvt/ctf.c
index 6a9d67aa300..d95dd1db165 100644
--- a/cddl/contrib/opensolaris/tools/ctf/cvt/ctf.c
+++ b/cddl/contrib/opensolaris/tools/ctf/cvt/ctf.c
@@ -355,14 +355,21 @@ write_type(void *arg1, void *arg2)
 		for (i = 0, ep = tp->t_emem; ep != NULL; ep = ep->el_next)
 			i++; /* count up enum members */
 
+		if (i > CTF_MAX_VLEN) {
+			warning("enum %s has too many values: %d > %d\n",
+			    tdesc_name(tp), i, CTF_MAX_VLEN);
+			i = CTF_MAX_VLEN;
+		}
+
 		ctt.ctt_info = CTF_TYPE_INFO(CTF_K_ENUM, isroot, i);
 		write_sized_type_rec(b, &ctt, tp->t_size);
 
-		for (ep = tp->t_emem; ep != NULL; ep = ep->el_next) {
+		for (ep = tp->t_emem; ep != NULL && i > 0; ep = ep->el_next) {
 			offset = strtab_insert(&b->ctb_strtab, ep->el_name);
 			cte.cte_name = CTF_TYPE_NAME(CTF_STRTAB_0, offset);
 			cte.cte_value = ep->el_number;
 			ctf_buf_write(b, &cte, sizeof (cte));
+			i--;
 		}
 		break;
 

From 17bc427d0b1ebb2c0686f8133982541c034f0e10 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 3 May 2010 18:39:40 +0000
Subject: [PATCH 428/532] On sparc64 obtain the initiator ID from the Open
 Firmware device tree in order to match what the PROM built-in driver uses.

Approved by:	mjacob
---
 sys/dev/isp/isp_pci.c  | 9 +++++++++
 sys/dev/isp/isp_sbus.c | 8 +++-----
 2 files changed, 12 insertions(+), 5 deletions(-)

diff --git a/sys/dev/isp/isp_pci.c b/sys/dev/isp/isp_pci.c
index 8410ea64d72..e152e99ee67 100644
--- a/sys/dev/isp/isp_pci.c
+++ b/sys/dev/isp/isp_pci.c
@@ -46,6 +46,11 @@ __FBSDID("$FreeBSD$");
 #include <sys/malloc.h>
 #include <sys/uio.h>
 
+#ifdef __sparc64__
+#include <dev/ofw/openfirm.h>
+#include <machine/ofw_machdep.h>
+#endif
+
 #include <dev/isp/isp_freebsd.h>
 
 static uint32_t isp_pci_rd_reg(ispsoftc_t *, int);
@@ -517,7 +522,11 @@ isp_get_specific_options(device_t dev, int chan, ispsoftc_t *isp)
 		if (IS_FC(isp)) {
 			ISP_FC_PC(isp, chan)->default_id = 109 - chan;
 		} else {
+#ifdef __sparc64__
+			ISP_SPI_PC(isp, chan)->iid = OF_getscsinitid(dev);
+#else
 			ISP_SPI_PC(isp, chan)->iid = 7;
+#endif
 		}
 	} else {
 		if (IS_FC(isp)) {
diff --git a/sys/dev/isp/isp_sbus.c b/sys/dev/isp/isp_sbus.c
index 895645a673a..34dcc09f9e5 100644
--- a/sys/dev/isp/isp_sbus.c
+++ b/sys/dev/isp/isp_sbus.c
@@ -41,8 +41,10 @@ __FBSDID("$FreeBSD$");
 #include <sys/resource.h>
 
 #include <dev/ofw/ofw_bus.h>
+#include <dev/ofw/openfirm.h>
 
 #include <machine/bus.h>
+#include <machine/ofw_machdep.h>
 #include <machine/resource.h>
 #include <sys/rman.h>
 #include <sparc64/sbus/sbusvar.h>
@@ -264,11 +266,7 @@ isp_sbus_attach(device_t dev)
 		isp->isp_confopts |= ISP_CFG_OWNLOOPID;
 	}
 	if (default_id == -1) {
-		/*
-		 * XXX: should be a way to get properties w/o having
-		 * XXX: to call OF_xxx functions
-		 */
-		default_id = 7;
+		default_id = OF_getscsinitid(dev);
 	}
 	ISP_SPI_PC(isp, 0)->iid = default_id;
 

From 5637a5914382985bbd9e7d21e6966611eb3d8224 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Mon, 3 May 2010 19:19:58 +0000
Subject: [PATCH 429/532] Handle busy status of the page in a way expected for
 pager_getpage(). Flush requested page, unbusy other pages, do not clear
 m->busy.

Reviewed by:	alc
MFC after:	1 week
---
 sys/vm/phys_pager.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/sys/vm/phys_pager.c b/sys/vm/phys_pager.c
index 42cdab3ebec..97674e35cf7 100644
--- a/sys/vm/phys_pager.c
+++ b/sys/vm/phys_pager.c
@@ -152,10 +152,10 @@ phys_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 		KASSERT(m[i]->dirty == 0,
 		    ("phys_pager_getpages: dirty page %p", m[i]));
 		/* The requested page must remain busy, the others not. */
-		if (reqpage != i) {
-			m[i]->oflags &= ~VPO_BUSY;
-			m[i]->busy = 0;
-		}
+		if (i == reqpage)
+			vm_page_flash(m[i]);
+		else
+			vm_page_wakeup(m[i]);
 	}
 	return (VM_PAGER_OK);
 }

From fc0c3802f0ab94acd342a9be61d417169d628ce6 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Mon, 3 May 2010 20:31:13 +0000
Subject: [PATCH 430/532] Lock the page around vm_page_activate() and
 vm_page_deactivate() calls where it was missed. The wrapped fragments now
 protect wire_count with page lock.

Reviewed by:	alc
---
 sys/dev/md/md.c              | 2 ++
 sys/fs/nfsclient/nfs_clbio.c | 8 ++++++--
 sys/fs/nwfs/nwfs_io.c        | 8 ++++++--
 sys/fs/smbfs/smbfs_io.c      | 8 ++++++--
 sys/nfsclient/nfs_bio.c      | 8 ++++++--
 5 files changed, 26 insertions(+), 8 deletions(-)

diff --git a/sys/dev/md/md.c b/sys/dev/md/md.c
index 78f2af3673c..edd687fca9e 100644
--- a/sys/dev/md/md.c
+++ b/sys/dev/md/md.c
@@ -665,11 +665,13 @@ mdstart_swap(struct md_s *sc, struct bio *bp)
 		sf_buf_free(sf);
 		sched_unpin();
 		vm_page_wakeup(m);
+		vm_page_lock(m);
 		vm_page_lock_queues();
 		vm_page_activate(m);
 		if (bp->bio_cmd == BIO_WRITE)
 			vm_page_dirty(m);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 
 		/* Actions on further pages start at offset 0 */
 		p += PAGE_SIZE - offs;
diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index 2401c887be7..e23da725d3f 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -192,12 +192,14 @@ ncl_getpages(struct vop_getpages_args *ap)
 
 	size = count - uio.uio_resid;
 	VM_OBJECT_LOCK(object);
-	vm_page_lock_queues();
 	for (i = 0, toff = 0; i < npages; i++, toff = nextoff) {
 		vm_page_t m;
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
+		vm_page_lock(m);
+		vm_page_lock_queues();
+
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -244,8 +246,10 @@ ncl_getpages(struct vop_getpages_args *ap)
 				vm_page_free(m);
 			}
 		}
+
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(object);
 	return (0);
 }
diff --git a/sys/fs/nwfs/nwfs_io.c b/sys/fs/nwfs/nwfs_io.c
index 75b1c188a18..ef1c800a525 100644
--- a/sys/fs/nwfs/nwfs_io.c
+++ b/sys/fs/nwfs/nwfs_io.c
@@ -449,12 +449,14 @@ nwfs_getpages(ap)
 
 	size = count - uio.uio_resid;
 
-	vm_page_lock_queues();
 	for (i = 0, toff = 0; i < npages; i++, toff = nextoff) {
 		vm_page_t m;
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
+		vm_page_lock(m);
+		vm_page_lock_queues();
+
 		if (nextoff <= size) {
 			m->valid = VM_PAGE_BITS_ALL;
 			KASSERT(m->dirty == 0,
@@ -489,8 +491,10 @@ nwfs_getpages(ap)
 				vm_page_free(m);
 			}
 		}
+
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(object);
 	return 0;
 #endif /* NWFS_RWCACHE */
diff --git a/sys/fs/smbfs/smbfs_io.c b/sys/fs/smbfs/smbfs_io.c
index df779a6151b..6537ed4b29f 100644
--- a/sys/fs/smbfs/smbfs_io.c
+++ b/sys/fs/smbfs/smbfs_io.c
@@ -500,12 +500,14 @@ smbfs_getpages(ap)
 
 	size = count - uio.uio_resid;
 
-	vm_page_lock_queues();
 	for (i = 0, toff = 0; i < npages; i++, toff = nextoff) {
 		vm_page_t m;
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
+		vm_page_lock(m);
+		vm_page_lock_queues();
+
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -553,8 +555,10 @@ smbfs_getpages(ap)
 				vm_page_free(m);
 			}
 		}
+
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(object);
 	return 0;
 #endif /* SMBFS_RWGENERIC */
diff --git a/sys/nfsclient/nfs_bio.c b/sys/nfsclient/nfs_bio.c
index cec0220b7f1..3a0705ef88d 100644
--- a/sys/nfsclient/nfs_bio.c
+++ b/sys/nfsclient/nfs_bio.c
@@ -189,12 +189,14 @@ nfs_getpages(struct vop_getpages_args *ap)
 
 	size = count - uio.uio_resid;
 	VM_OBJECT_LOCK(object);
-	vm_page_lock_queues();
 	for (i = 0, toff = 0; i < npages; i++, toff = nextoff) {
 		vm_page_t m;
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
+		vm_page_lock(m);
+		vm_page_lock_queues();
+
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -241,8 +243,10 @@ nfs_getpages(struct vop_getpages_args *ap)
 				vm_page_free(m);
 			}
 		}
+
+		vm_page_unlock_queues();
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
 	VM_OBJECT_UNLOCK(object);
 	return (0);
 }

From 917825acb9bbc575102001b79aa3fc25b47fe035 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 3 May 2010 20:57:16 +0000
Subject: [PATCH 431/532] - Don't set CAS_PCS_DATAPATH to anything except
 CAS_PCS_DATAPATH_SERDES   on Cassini using the external PCS SERDES otherwise
 unaligned access   traps and other strange effects happen with some machines.
 Don't touch   the MIF which is unused in that case either. These changes
 require the   PHY type to use to be determined via the OFW device tree or
 from the   VPD in machines without the former. - Disable the SERDES pins of
 Saturn when not used in order to save power   and ensure they are enabled
 otherwise. - In cas_attach() use the correct register offset for
 CAS_PCS_CONF_EN. - Add some bus space barriers missing in the PCS code path.

These changes make the Sun GigaSwift Ethernet 1.0 MMF cards as well as
the on-board interfaces found in Sun Fire B100s Blade Server work.

PR:	144867
---
 sys/dev/cas/if_cas.c    | 232 ++++++++++++++++++++++++++++------------
 sys/dev/cas/if_casreg.h |  12 +++
 2 files changed, 174 insertions(+), 70 deletions(-)

diff --git a/sys/dev/cas/if_cas.c b/sys/dev/cas/if_cas.c
index edcfec47cdd..3522f7ca3b7 100644
--- a/sys/dev/cas/if_cas.c
+++ b/sys/dev/cas/if_cas.c
@@ -76,6 +76,7 @@ __FBSDID("$FreeBSD$");
 
 #include <machine/bus.h>
 #if defined(__powerpc__) || defined(__sparc64__)
+#include <dev/ofw/ofw_bus.h>
 #include <dev/ofw/openfirm.h>
 #include <machine/ofw_machdep.h>
 #endif
@@ -321,55 +322,82 @@ cas_attach(struct cas_softc *sc)
 		}
 	}
 
-	CAS_WRITE_4(sc, CAS_PCS_DATAPATH, CAS_PCS_DATAPATH_MII);
-
-	cas_mifinit(sc);
-
-	/*
-	 * Look for an external PHY.
-	 */
-	error = ENXIO;
-	v = CAS_READ_4(sc, CAS_MIF_CONF);
-	if ((v & CAS_MIF_CONF_MDI1) != 0) {
-		v |= CAS_MIF_CONF_PHY_SELECT;
-		CAS_WRITE_4(sc, CAS_MIF_CONF, v);
-		switch (sc->sc_variant) {
-		default:
-			sc->sc_phyad = -1;
-			break;
+	if ((sc->sc_flags & CAS_SERDES) == 0) {
+		CAS_WRITE_4(sc, CAS_PCS_DATAPATH, CAS_PCS_DATAPATH_MII);
+		CAS_BARRIER(sc, CAS_PCS_DATAPATH, 4,
+		    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
+		cas_mifinit(sc);
+		/*
+		 * Look for an external PHY.
+		 */
+		error = ENXIO;
+		v = CAS_READ_4(sc, CAS_MIF_CONF);
+		if ((v & CAS_MIF_CONF_MDI1) != 0) {
+			v |= CAS_MIF_CONF_PHY_SELECT;
+			CAS_WRITE_4(sc, CAS_MIF_CONF, v);
+			CAS_BARRIER(sc, CAS_MIF_CONF, 4,
+			    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
+			/* Enable/unfreeze the GMII pins of Saturn. */
+			if (sc->sc_variant == CAS_SATURN) {
+				CAS_WRITE_4(sc, CAS_SATURN_PCFG, 0);
+				CAS_BARRIER(sc, CAS_SATURN_PCFG, 4,
+				    BUS_SPACE_BARRIER_READ |
+				    BUS_SPACE_BARRIER_WRITE);
+			}
+			switch (sc->sc_variant) {
+			default:
+				sc->sc_phyad = -1;
+				break;
+			}
+			error = mii_phy_probe(sc->sc_dev, &sc->sc_miibus,
+			    cas_mediachange, cas_mediastatus);
 		}
-		error = mii_phy_probe(sc->sc_dev, &sc->sc_miibus,
-		    cas_mediachange, cas_mediastatus);
-	}
-
-	/*
-	 * Fall back on an internal PHY if no external PHY was found.
-	 */
-	if (error != 0 && (v & CAS_MIF_CONF_MDI0) != 0) {
-		v &= ~CAS_MIF_CONF_PHY_SELECT;
-		CAS_WRITE_4(sc, CAS_MIF_CONF, v);
-		switch (sc->sc_variant) {
-		default:
-			sc->sc_phyad = -1;
-			break;
+		/*
+		 * Fall back on an internal PHY if no external PHY was found.
+		 */
+		if (error != 0 && (v & CAS_MIF_CONF_MDI0) != 0) {
+			v &= ~CAS_MIF_CONF_PHY_SELECT;
+			CAS_WRITE_4(sc, CAS_MIF_CONF, v);
+			CAS_BARRIER(sc, CAS_MIF_CONF, 4,
+			    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
+			/* Freeze the GMII pins of Saturn for saving power. */
+			if (sc->sc_variant == CAS_SATURN) {
+				CAS_WRITE_4(sc, CAS_SATURN_PCFG,
+				    CAS_SATURN_PCFG_FSI);
+				CAS_BARRIER(sc, CAS_SATURN_PCFG, 4,
+				    BUS_SPACE_BARRIER_READ |
+				    BUS_SPACE_BARRIER_WRITE);
+			}
+			switch (sc->sc_variant) {
+			default:
+				sc->sc_phyad = -1;
+				break;
+			}
+			error = mii_phy_probe(sc->sc_dev, &sc->sc_miibus,
+			    cas_mediachange, cas_mediastatus);
 		}
-		error = mii_phy_probe(sc->sc_dev, &sc->sc_miibus,
-		    cas_mediachange, cas_mediastatus);
-	}
-
-	/*
-	 * Try the external PCS SERDES if we didn't find any PHYs.
-	 */
-	if (error != 0) {
+	} else {
+		/*
+		 * Use the external PCS SERDES.
+		 */
 		CAS_WRITE_4(sc, CAS_PCS_DATAPATH, CAS_PCS_DATAPATH_SERDES);
+		CAS_BARRIER(sc, CAS_PCS_DATAPATH, 4, BUS_SPACE_BARRIER_WRITE);
+		/* Enable/unfreeze the SERDES pins of Saturn. */
+		if (sc->sc_variant == CAS_SATURN) {
+			CAS_WRITE_4(sc, CAS_SATURN_PCFG, 0);
+			CAS_BARRIER(sc, CAS_SATURN_PCFG, 4,
+			    BUS_SPACE_BARRIER_WRITE);
+		}
 		CAS_WRITE_4(sc, CAS_PCS_SERDES_CTRL, CAS_PCS_SERDES_CTRL_ESD);
-		CAS_WRITE_4(sc, CAS_PCS_CONF_EN, CAS_PCS_CONF_EN);
-		sc->sc_flags |= CAS_SERDES;
+		CAS_BARRIER(sc, CAS_PCS_SERDES_CTRL, 4,
+		    BUS_SPACE_BARRIER_WRITE);
+		CAS_WRITE_4(sc, CAS_PCS_CONF, CAS_PCS_CONF_EN);
+		CAS_BARRIER(sc, CAS_PCS_CONF, 4,
+		    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
 		sc->sc_phyad = CAS_PHYAD_EXTERNAL;
 		error = mii_phy_probe(sc->sc_dev, &sc->sc_miibus,
 		    cas_mediachange, cas_mediastatus);
 	}
-
 	if (error != 0) {
 		device_printf(sc->sc_dev, "PHY probe failed: %d\n", error);
 		goto fail_rxmap;
@@ -956,8 +984,9 @@ cas_init_locked(struct cas_softc *sc)
 	    __func__);
 #endif
 
-	/* Re-initialize the MIF. */
-	cas_mifinit(sc);
+	if ((sc->sc_flags & CAS_SERDES) == 0)
+		/* Re-initialize the MIF. */
+		cas_mifinit(sc);
 
 	/* step 3.  Setup data structures in host memory. */
 	cas_meminit(sc);
@@ -2105,6 +2134,8 @@ cas_mifinit(struct cas_softc *sc)
 	/* Configure the MIF in frame mode. */
 	CAS_WRITE_4(sc, CAS_MIF_CONF,
 	    CAS_READ_4(sc, CAS_MIF_CONF) & ~CAS_MIF_CONF_BB_MODE);
+	CAS_BARRIER(sc, CAS_MIF_CONF, 4,
+	    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
 }
 
 /*
@@ -2219,10 +2250,16 @@ cas_mii_writereg(device_t dev, int phy, int reg, int val)
 			CAS_BARRIER(sc, CAS_PCS_CONF, 4,
 			    BUS_SPACE_BARRIER_WRITE);
 			CAS_WRITE_4(sc, CAS_PCS_ANAR, val);
+			CAS_BARRIER(sc, CAS_PCS_ANAR, 4,
+			    BUS_SPACE_BARRIER_WRITE);
 			CAS_WRITE_4(sc, CAS_PCS_SERDES_CTRL,
 			    CAS_PCS_SERDES_CTRL_ESD);
+			CAS_BARRIER(sc, CAS_PCS_CONF, 4,
+			    BUS_SPACE_BARRIER_WRITE);
 			CAS_WRITE_4(sc, CAS_PCS_CONF,
 			    CAS_PCS_CONF_EN);
+			CAS_BARRIER(sc, CAS_PCS_CONF, 4,
+			    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
 			return (0);
 		case MII_ANLPAR:
 			reg = CAS_PCS_ANLPAR;
@@ -2233,6 +2270,8 @@ cas_mii_writereg(device_t dev, int phy, int reg, int val)
 			return (0);
 		}
 		CAS_WRITE_4(sc, reg, val);
+		CAS_BARRIER(sc, reg, 4,
+		    BUS_SPACE_BARRIER_READ | BUS_SPACE_BARRIER_WRITE);
 		return (0);
 	}
 
@@ -2630,15 +2669,20 @@ static struct resource_spec cas_pci_res_spec[] = {
 	{ -1, 0 }
 };
 
+#define	CAS_LOCAL_MAC_ADDRESS	"local-mac-address"
+#define	CAS_PHY_INTERFACE	"phy-interface"
+#define	CAS_PHY_TYPE		"phy-type"
+#define	CAS_PHY_TYPE_PCS	"pcs"
+
 static int
 cas_pci_attach(device_t dev)
 {
+	char buf[sizeof(CAS_LOCAL_MAC_ADDRESS)];
 	struct cas_softc *sc;
 	int i;
 #if !(defined(__powerpc__) || defined(__sparc64__))
 	u_char enaddr[4][ETHER_ADDR_LEN];
-	char lma[sizeof("local-mac-address")];
-	int found, j;
+	u_int j, k, lma, pcs[4], phy;
 #endif
 
 	sc = device_get_softc(dev);
@@ -2679,13 +2723,20 @@ cas_pci_attach(device_t dev)
 
 #if defined(__powerpc__) || defined(__sparc64__)
 	OF_getetheraddr(dev, sc->sc_enaddr);
+	if (OF_getprop(ofw_bus_get_node(dev), CAS_PHY_INTERFACE, buf,
+	    sizeof(buf)) > 0 || OF_getprop(ofw_bus_get_node(dev),
+	    CAS_PHY_TYPE, buf, sizeof(buf)) > 0) {
+		buf[sizeof(buf) - 1] = '\0';
+		if (strcmp(buf, CAS_PHY_TYPE_PCS) == 0)
+			sc->sc_flags |= CAS_SERDES;
+	}
 #else
 	/*
-	 * Dig out VPD (vital product data) and read the MAX address.
-	 * The VPD resides in the PCI Expansion ROM (PCI FCode) and
-	 * can't be accessed via the PCI capability pointer.
-	 * SUNW,pci-ce and SUNW,pci-qge use the Enhanced VPD format
-	 * described in US Patent 7149820.
+	 * Dig out VPD (vital product data) and read the MAC address as well
+	 * as the PHY type.  The VPD resides in the PCI Expansion ROM (PCI
+	 * FCode) and can't be accessed via the PCI capability pointer.
+	 * SUNW,pci-ce and SUNW,pci-qge use the Enhanced VPD format described
+	 * in the free US Patent 7149820.
 	 */
 
 #define	PCI_ROMHDR_SIZE			0x1c
@@ -2719,7 +2770,10 @@ cas_pci_attach(device_t dev)
 #define	CAS_ROM_READ_4(sc, offs)					\
 	CAS_READ_4((sc), CAS_PCI_ROM_OFFSET + (offs))
 
-	found = 0;
+	lma = phy = 0;
+	memset(enaddr, 0, sizeof(enaddr));
+	memset(pcs, 0, sizeof(pcs));
+
 	/* Enable PCI Expansion ROM access. */
 	CAS_WRITE_4(sc, CAS_BIM_LDEV_OEN,
 	    CAS_BIM_LDEV_OEN_PAD | CAS_BIM_LDEV_OEN_PROM);
@@ -2768,23 +2822,51 @@ cas_pci_attach(device_t dev)
 			if (CAS_ROM_READ_1(sc, j + PCI_VPD_SIZE) != 'I')
 				/* no instance property */
 				continue;
-			if (CAS_ROM_READ_1(sc, j + PCI_VPD_SIZE + 3) != 'B')
-				/* no byte array */
-				continue;
-			if (CAS_ROM_READ_1(sc, j + PCI_VPD_SIZE + 4) !=
-			    ETHER_ADDR_LEN)
-				continue;
-			bus_read_region_1(sc->sc_res[CAS_RES_MEM],
-			    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE + 5,
-			    lma, sizeof(lma));
-			if (strcmp(lma, "local-mac-address") != 0)
-				continue;
-			bus_read_region_1(sc->sc_res[CAS_RES_MEM],
-			    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE + 5 +
-			    sizeof(lma), enaddr[found],
-			    sizeof(enaddr[found]));
-			if (found++ == 4)
-				break;
+			if (CAS_ROM_READ_1(sc, j + PCI_VPD_SIZE + 3) == 'B') {
+				/* byte array */
+				if (CAS_ROM_READ_1(sc,
+				    j + PCI_VPD_SIZE + 4) != ETHER_ADDR_LEN)
+					continue;
+				bus_read_region_1(sc->sc_res[CAS_RES_MEM],
+				    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE + 5,
+				    buf, sizeof(buf));
+				buf[sizeof(buf) - 1] = '\0';
+				if (strcmp(buf, CAS_LOCAL_MAC_ADDRESS) != 0)
+					continue;
+				bus_read_region_1(sc->sc_res[CAS_RES_MEM],
+				    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE +
+				    5 + sizeof(CAS_LOCAL_MAC_ADDRESS),
+				    enaddr[lma], sizeof(enaddr[lma]));
+				lma++;
+				if (lma == 4 && phy == 4)
+					break;
+			} else if (CAS_ROM_READ_1(sc, j + PCI_VPD_SIZE + 3) ==
+			   'S') {
+				/* string */
+				if (CAS_ROM_READ_1(sc,
+				    j + PCI_VPD_SIZE + 4) !=
+				    sizeof(CAS_PHY_TYPE_PCS))
+					continue;
+				bus_read_region_1(sc->sc_res[CAS_RES_MEM],
+				    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE + 5,
+				    buf, sizeof(buf));
+				buf[sizeof(buf) - 1] = '\0';
+				if (strcmp(buf, CAS_PHY_INTERFACE) == 0)
+					k = sizeof(CAS_PHY_INTERFACE);
+				else if (strcmp(buf, CAS_PHY_TYPE) == 0)
+					k = sizeof(CAS_PHY_TYPE);
+				else
+					continue;
+				bus_read_region_1(sc->sc_res[CAS_RES_MEM],
+				    CAS_PCI_ROM_OFFSET + j + PCI_VPD_SIZE +
+				    5 + k, buf, sizeof(buf));
+				buf[sizeof(buf) - 1] = '\0';
+				if (strcmp(buf, CAS_PHY_TYPE_PCS) == 0)
+					pcs[phy] = 1;
+				phy++;
+				if (lma == 4 && phy == 4)
+					break;
+			}
 		}
 		break;
 	default:
@@ -2795,14 +2877,24 @@ cas_pci_attach(device_t dev)
  fail_prom:
 	CAS_WRITE_4(sc, CAS_BIM_LDEV_OEN, 0);
 
-	if (found == 0) {
+	if (lma == 0) {
 		device_printf(dev, "could not determine Ethernet address\n");
 		goto fail;
 	}
 	i = 0;
-	if (found > 1 && pci_get_slot(dev) < sizeof(enaddr) / sizeof(*enaddr))
+	if (lma > 1 && pci_get_slot(dev) < sizeof(enaddr) / sizeof(*enaddr))
 		i = pci_get_slot(dev);
 	memcpy(sc->sc_enaddr, enaddr[i], ETHER_ADDR_LEN);
+
+	if (phy == 0) {
+		device_printf(dev, "could not determine PHY type\n");
+		goto fail;
+	}
+	i = 0;
+	if (phy > 1 && pci_get_slot(dev) < sizeof(pcs) / sizeof(*pcs))
+		i = pci_get_slot(dev);
+	if (pcs[i] != 0)
+		sc->sc_flags |= CAS_SERDES;
 #endif
 
 	if (cas_attach(sc) != 0) {
diff --git a/sys/dev/cas/if_casreg.h b/sys/dev/cas/if_casreg.h
index 97250e02791..0d1c1d822b9 100644
--- a/sys/dev/cas/if_casreg.h
+++ b/sys/dev/cas/if_casreg.h
@@ -68,6 +68,7 @@
 #define	CAS_STATUS4		0x105c	/* interrupt status 4 for INTD */
 #define	CAS_CLEAR_ALIAS4	0x1060	/* clear mask alias 4 for INTD */
 #define	CAS_STATUS_ALIAS4	0x1064	/* interrupt status alias 4 for INTD */
+#define	CAS_SATURN_PCFG		0x106c	/* internal MACPHY pin configuration */
 
 #define	CAS_CAW_RX_WGHT_MASK	0x00000003	/* RX DMA factor for... */
 #define	CAS_CAW_RX_WGHT_SHFT	0		/* ...weighted round robin */
@@ -171,6 +172,17 @@
 /* INTn enable bit for CAS_INTMASK[2-4] */
 #define	CAS_INTMASKN_EN		0x00000080	/* INT[B-D] enable */
 
+#define	CAS_SATURN_PCFG_TLA	0x00000001	/* PHY activity LED */
+#define	CAS_SATURN_PCFG_FLA	0x00000002	/* PHY 10MBit/sec LED */
+#define	CAS_SATURN_PCFG_CLA	0x00000004	/* PHY 100MBit/sec LED */
+#define	CAS_SATURN_PCFG_LLA	0x00000008	/* PHY 1000MBit/sec LED */
+#define	CAS_SATURN_PCFG_RLA	0x00000010	/* PHY full-duplex LED */
+#define	CAS_SATURN_PCFG_PDS	0x00000020	/* PHY debug mode */
+#define	CAS_SATURN_PCFG_MTP	0x00000080	/* test point select */
+#define	CAS_SATURN_PCFG_GMO	0x00000100	/* GMII observe */
+#define	CAS_SATURN_PCFG_FSI	0x00000200	/* freeze GMII/SERDES */
+#define	CAS_SATURN_PCFG_LAD	0x00000800	/* MAC LED control active low */
+
 /* TX DMA registers */
 #define	CAS_TX_CONF		0x2004	/* TX configuration */
 #define	CAS_TX_FIFO_WR		0x2014	/* FIFO write pointer */

From c41e7be93cba7d78b1ec125961f21189a5bccc34 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Mon, 3 May 2010 20:59:27 +0000
Subject: [PATCH 432/532] Sun GigaSwift Ethernet 1.0 MMF work with r207585 in
 place.

---
 share/man/man4/cas.4 | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/share/man/man4/cas.4 b/share/man/man4/cas.4
index 7851a994662..cc050c9a42b 100644
--- a/share/man/man4/cas.4
+++ b/share/man/man4/cas.4
@@ -25,7 +25,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd June 15, 2009
+.Dd May 3, 2010
 .Dt CAS 4
 .Os
 .Sh NAME
@@ -91,6 +91,9 @@ driver at this time:
 .Pp
 .Bl -bullet -compact
 .It
+Sun GigaSwift Ethernet 1.0 MMF (Cassini Kuheen)
+(part no.\& 501-5524)
+.It
 Sun GigaSwift Ethernet 1.0 UTP (Cassini)
 (part no.\& 501-5902)
 .It

From 7775701f3a6c82671c29240ebba0a23980b37082 Mon Sep 17 00:00:00 2001
From: Ed Maste <emaste@FreeBSD.org>
Date: Mon, 3 May 2010 22:32:26 +0000
Subject: [PATCH 433/532] Update GRANDPARENTED text to match the contents of
 tzdata/factory r19879. This eliminates "warning: time zone abbreviation
 differs from POSIX" on installworld to an empty filesystem.

Reviewed by:	edwin
MFC after:	1 week
---
 contrib/tzcode/zic/private.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/contrib/tzcode/zic/private.h b/contrib/tzcode/zic/private.h
index ecbf6124859..d00bf9bae14 100644
--- a/contrib/tzcode/zic/private.h
+++ b/contrib/tzcode/zic/private.h
@@ -34,7 +34,7 @@ static const char	privatehid[] = "@(#)private.h	8.6";
 #endif /* !defined NOID */
 #endif /* !defined lint */
 
-#define GRANDPARENTED	"Local time zone must be set--see zic manual page"
+#define GRANDPARENTED	"Local time zone must be set--use tzsetup"
 
 /*
 ** Defaults for preprocessor symbols.

From fd73ccb1432c7350485b957f82446cc93c821ea5 Mon Sep 17 00:00:00 2001
From: Doug Barton <dougb@FreeBSD.org>
Date: Tue, 4 May 2010 01:46:58 +0000
Subject: [PATCH 434/532] Make address assignment via ipv6_prefix_IF work again

---
 etc/network.subr | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/etc/network.subr b/etc/network.subr
index 52c99db431c..d0b309abdb6 100644
--- a/etc/network.subr
+++ b/etc/network.subr
@@ -399,6 +399,8 @@ ipv6if()
 	$_if|"$_if "*|*" $_if"|*" $_if "*|[Aa][Uu][Tt][Oo])
 		# True if $ifconfig_IF_ipv6 is defined.
 		_tmpargs=`_ifconfig_getargs $_if ipv6`
+		# Also true if ipv6_prefix_IF is defined
+		[ -n "$_tmpargs" ] || _tmpargs=`get_if_var $_if ipv6_prefix_IF`
 		;;
 	esac
 

From d54a8529a7b4999c58f2d3f6318c4732715aa749 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 05:47:17 +0000
Subject: [PATCH 435/532] Move definition of struct rusage_ext before struct
 thread.

MFC after:	1 week
---
 sys/sys/proc.h | 41 +++++++++++++++++++++--------------------
 1 file changed, 21 insertions(+), 20 deletions(-)

diff --git a/sys/sys/proc.h b/sys/sys/proc.h
index fb31cfc1318..24c918e1121 100644
--- a/sys/sys/proc.h
+++ b/sys/sys/proc.h
@@ -172,6 +172,27 @@ struct kdtrace_proc;
 struct kdtrace_thread;
 struct cpuset;
 
+/*
+ * XXX: Does this belong in resource.h or resourcevar.h instead?
+ * Resource usage extension.  The times in rusage structs in the kernel are
+ * never up to date.  The actual times are kept as runtimes and tick counts
+ * (with control info in the "previous" times), and are converted when
+ * userland asks for rusage info.  Backwards compatibility prevents putting
+ * this directly in the user-visible rusage struct.
+ *
+ * Locking for p_rux: (cj) means (j) for p_rux and (c) for p_crux.
+ * Locking for td_rux: (t) for all fields.
+ */
+struct rusage_ext {
+	u_int64_t	rux_runtime;    /* (cj) Real time. */
+	u_int64_t	rux_uticks;     /* (cj) Statclock hits in user mode. */
+	u_int64_t	rux_sticks;     /* (cj) Statclock hits in sys mode. */
+	u_int64_t	rux_iticks;     /* (cj) Statclock hits in intr mode. */
+	u_int64_t	rux_uu;         /* (c) Previous user time in usec. */
+	u_int64_t	rux_su;         /* (c) Previous sys time in usec. */
+	u_int64_t	rux_tu;         /* (c) Previous total time in usec. */
+};
+
 /*
  * Kernel runnable context (thread).
  * This is what is put to sleep and reactivated.
@@ -425,26 +446,6 @@ do {									\
 #define	TD_SET_RUNQ(td)		(td)->td_state = TDS_RUNQ
 #define	TD_SET_CAN_RUN(td)	(td)->td_state = TDS_CAN_RUN
 
-/*
- * XXX: Does this belong in resource.h or resourcevar.h instead?
- * Resource usage extension.  The times in rusage structs in the kernel are
- * never up to date.  The actual times are kept as runtimes and tick counts
- * (with control info in the "previous" times), and are converted when
- * userland asks for rusage info.  Backwards compatibility prevents putting
- * this directly in the user-visible rusage struct.
- *
- * Locking: (cj) means (j) for p_rux and (c) for p_crux.
- */
-struct rusage_ext {
-	u_int64_t	rux_runtime;    /* (cj) Real time. */
-	u_int64_t	rux_uticks;     /* (cj) Statclock hits in user mode. */
-	u_int64_t	rux_sticks;     /* (cj) Statclock hits in sys mode. */
-	u_int64_t	rux_iticks;     /* (cj) Statclock hits in intr mode. */
-	u_int64_t	rux_uu;         /* (c) Previous user time in usec. */
-	u_int64_t	rux_su;         /* (c) Previous sys time in usec. */
-	u_int64_t	rux_tu;         /* (c) Previous total time in usec. */
-};
-
 /*
  * Process structure.
  */

From 0c41a69e71a87add59a92e43b9a7f4e5749a9167 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Tue, 4 May 2010 05:55:19 +0000
Subject: [PATCH 436/532] Add lock assertions.

---
 sys/vm/vm_page.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index e18e7ba1b3b..d9288d98f4d 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1385,6 +1385,7 @@ vm_page_activate(vm_page_t m)
 {
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	if (VM_PAGE_GETKNOWNQUEUE2(m) != PQ_ACTIVE) {
 		vm_pageq_remove(m);
 		if (m->wire_count == 0 && (m->flags & PG_UNMANAGED) == 0) {
@@ -1588,8 +1589,10 @@ void
 vm_page_unwire(vm_page_t m, int activate)
 {
 
-	if ((m->flags & PG_UNMANAGED) == 0)
+	if ((m->flags & PG_UNMANAGED) == 0) {
 		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+		vm_page_lock_assert(m, MA_OWNED);
+	}
 	if (m->flags & PG_FICTITIOUS)
 		return;
 	if (m->wire_count > 0) {
@@ -1626,6 +1629,7 @@ _vm_page_deactivate(vm_page_t m, int athead)
 {
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 
 	/*
 	 * Ignore if already inactive.
@@ -1845,6 +1849,8 @@ vm_page_dontneed(vm_page_t m)
 	int head;
 
 	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
+	VM_OBJECT_LOCK_ASSERT(m->object, MA_OWNED);
 	dnw = ++dnweight;
 
 	/*

From bed4c5241663420ce37d648d07a2c564ee96e35a Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 05:55:37 +0000
Subject: [PATCH 437/532] Implement RUSAGE_THREAD. Add td_rux to keep extended
 runtime and ticks information for thread to allow calcru1() (re)use.

Rename ruxagg()->ruxagg_locked(), ruxagg_tlock()->ruxagg() [1].
The ruxagg_locked() function no longer clears thread ticks nor
td_incruntime.

Requested by:	attilio [1]
Discussed with:	attilio, bde
Reviewed by:	bde
Based on submission by:	Alexander Krizhanovsky <ak natsys-lab com>
MFC after:	1 week
X-MFC-Note:	td_rux shall be moved to the end of struct thread
---
 sys/kern/kern_resource.c | 37 ++++++++++++++++++++++++-------------
 sys/kern/kern_thread.c   |  2 +-
 sys/sys/proc.h           |  3 ++-
 sys/sys/resource.h       |  1 +
 sys/sys/resourcevar.h    |  2 +-
 5 files changed, 29 insertions(+), 16 deletions(-)

diff --git a/sys/kern/kern_resource.c b/sys/kern/kern_resource.c
index a3ed75d3583..0bc78d0ae2e 100644
--- a/sys/kern/kern_resource.c
+++ b/sys/kern/kern_resource.c
@@ -76,7 +76,7 @@ static void	calcru1(struct proc *p, struct rusage_ext *ruxp,
 		    struct timeval *up, struct timeval *sp);
 static int	donice(struct thread *td, struct proc *chgp, int n);
 static struct uidinfo *uilookup(uid_t uid);
-static void	ruxagg_tlock(struct proc *p, struct thread *td);
+static void	ruxagg(struct proc *p, struct thread *td);
 
 /*
  * Resource controls and accounting.
@@ -630,7 +630,7 @@ lim_cb(void *arg)
 		return;
 	PROC_SLOCK(p);
 	FOREACH_THREAD_IN_PROC(p, td) {
-		ruxagg_tlock(p, td);
+		ruxagg(p, td);
 	}
 	PROC_SUNLOCK(p);
 	if (p->p_rux.rux_runtime > p->p_cpulimit * cpu_tickrate()) {
@@ -841,7 +841,7 @@ calcru(struct proc *p, struct timeval *up, struct timeval *sp)
 	FOREACH_THREAD_IN_PROC(p, td) {
 		if (td->td_incruntime == 0)
 			continue;
-		ruxagg_tlock(p, td);
+		ruxagg(p, td);
 	}
 	calcru1(p, &p->p_rux, up, sp);
 }
@@ -961,6 +961,16 @@ kern_getrusage(struct thread *td, int who, struct rusage *rup)
 		calccru(p, &rup->ru_utime, &rup->ru_stime);
 		break;
 
+	case RUSAGE_THREAD:
+		PROC_SLOCK(p);
+		ruxagg(p, td);
+		PROC_SUNLOCK(p);
+		thread_lock(td);
+		*rup = td->td_ru;
+		calcru1(p, &td->td_rux, &rup->ru_utime, &rup->ru_stime);
+		thread_unlock(td);
+		break;
+
 	default:
 		error = EINVAL;
 	}
@@ -1001,7 +1011,7 @@ ruadd(struct rusage *ru, struct rusage_ext *rux, struct rusage *ru2,
  * Aggregate tick counts into the proc's rusage_ext.
  */
 void
-ruxagg(struct rusage_ext *rux, struct thread *td)
+ruxagg_locked(struct rusage_ext *rux, struct thread *td)
 {
 
 	THREAD_LOCK_ASSERT(td, MA_OWNED);
@@ -1010,18 +1020,19 @@ ruxagg(struct rusage_ext *rux, struct thread *td)
 	rux->rux_uticks += td->td_uticks;
 	rux->rux_sticks += td->td_sticks;
 	rux->rux_iticks += td->td_iticks;
+}
+
+static void
+ruxagg(struct proc *p, struct thread *td)
+{
+
+	thread_lock(td);
+	ruxagg_locked(&p->p_rux, td);
+	ruxagg_locked(&td->td_rux, td);
 	td->td_incruntime = 0;
 	td->td_uticks = 0;
 	td->td_iticks = 0;
 	td->td_sticks = 0;
-}
-
-static void
-ruxagg_tlock(struct proc *p, struct thread *td)
-{
-
-	thread_lock(td);
-	ruxagg(&p->p_rux, td);
 	thread_unlock(td);
 }
 
@@ -1039,7 +1050,7 @@ rufetch(struct proc *p, struct rusage *ru)
 	*ru = p->p_ru;
 	if (p->p_numthreads > 0)  {
 		FOREACH_THREAD_IN_PROC(p, td) {
-			ruxagg_tlock(p, td);
+			ruxagg(p, td);
 			rucollect(ru, &td->td_ru);
 		}
 	}
diff --git a/sys/kern/kern_thread.c b/sys/kern/kern_thread.c
index 9be4c2f3eb3..b32c5842198 100644
--- a/sys/kern/kern_thread.c
+++ b/sys/kern/kern_thread.c
@@ -432,7 +432,7 @@ thread_exit(void)
 	PROC_UNLOCK(p);
 	thread_lock(td);
 	/* Save our tick information with both the thread and proc locked */
-	ruxagg(&p->p_rux, td);
+	ruxagg_locked(&p->p_rux, td);
 	PROC_SUNLOCK(p);
 	td->td_state = TDS_INACTIVE;
 #ifdef WITNESS
diff --git a/sys/sys/proc.h b/sys/sys/proc.h
index 24c918e1121..e32e494f00f 100644
--- a/sys/sys/proc.h
+++ b/sys/sys/proc.h
@@ -240,7 +240,8 @@ struct thread {
 	u_int		td_estcpu;	/* (t) estimated cpu utilization */
 	int		td_slptick;	/* (t) Time at sleep. */
 	int		td_blktick;	/* (t) Time spent blocked. */
-	struct rusage	td_ru;		/* (t) rusage information */
+	struct rusage	td_ru;		/* (t) rusage information. */
+	struct rusage_ext td_rux;	/* (t) Internal rusage information. */
 	uint64_t	td_incruntime;	/* (t) Cpu ticks to transfer to proc. */
 	uint64_t	td_runtime;	/* (t) How many cpu ticks we've run. */
 	u_int 		td_pticks;	/* (t) Statclock hits for profiling */
diff --git a/sys/sys/resource.h b/sys/sys/resource.h
index 9af96afe0e0..e7037443246 100644
--- a/sys/sys/resource.h
+++ b/sys/sys/resource.h
@@ -56,6 +56,7 @@
 
 #define	RUSAGE_SELF	0
 #define	RUSAGE_CHILDREN	-1
+#define	RUSAGE_THREAD	1
 
 struct rusage {
 	struct timeval ru_utime;	/* user time used */
diff --git a/sys/sys/resourcevar.h b/sys/sys/resourcevar.h
index 21728aa67e4..95a9b49e1a6 100644
--- a/sys/sys/resourcevar.h
+++ b/sys/sys/resourcevar.h
@@ -131,7 +131,7 @@ void	 rucollect(struct rusage *ru, struct rusage *ru2);
 void	 rufetch(struct proc *p, struct rusage *ru);
 void	 rufetchcalc(struct proc *p, struct rusage *ru, struct timeval *up,
 	    struct timeval *sp);
-void	 ruxagg(struct rusage_ext *rux, struct thread *td);
+void	 ruxagg_locked(struct rusage_ext *rux, struct thread *td);
 int	 suswintr(void *base, int word);
 struct uidinfo
 	*uifind(uid_t uid);

From 03d13670c2e7d8ee66e6528ba49f33b9b6587723 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 06:00:39 +0000
Subject: [PATCH 438/532] Use td_rux.rux_runtime for ki_runtime instead of
 redoing calculation.

Submitted by:	bde
MFC after:	1 week
---
 sys/kern/kern_proc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index 49a30971502..6c50f1f3618 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -901,7 +901,7 @@ fill_kinfo_thread(struct thread *td, struct kinfo_proc *kp, int preferthread)
 	kp->ki_pri.pri_user = td->td_user_pri;
 
 	if (preferthread) {
-		kp->ki_runtime = cputick2usec(td->td_runtime);
+		kp->ki_runtime = td->td_rux.rux_runtime;
 		kp->ki_pctcpu = sched_pctcpu(td);
 		kp->ki_estcpu = td->td_estcpu;
 	}

From f742a317e67eb71d666cee5b6bbf5fa5842d1ed9 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 06:01:25 +0000
Subject: [PATCH 439/532] Document RUSAGE_THREAD.

Reviewed by:	bde
MFC after:	1 week
---
 lib/libc/sys/getrusage.2 | 12 +++++++++---
 1 file changed, 9 insertions(+), 3 deletions(-)

diff --git a/lib/libc/sys/getrusage.2 b/lib/libc/sys/getrusage.2
index bdf5d4591cb..423503f8282 100644
--- a/lib/libc/sys/getrusage.2
+++ b/lib/libc/sys/getrusage.2
@@ -28,7 +28,7 @@
 .\"     @(#)getrusage.2	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd June 4, 1993
+.Dd May 1, 2010
 .Dt GETRUSAGE 2
 .Os
 .Sh NAME
@@ -42,6 +42,7 @@
 .In sys/resource.h
 .Fd "#define	RUSAGE_SELF	 0"
 .Fd "#define	RUSAGE_CHILDREN	-1"
+.Fd "#define	RUSAGE_THREAD	1"
 .Ft int
 .Fn getrusage "int who" "struct rusage *rusage"
 .Sh DESCRIPTION
@@ -49,11 +50,12 @@ The
 .Fn getrusage
 system call
 returns information describing the resources utilized by the current
-process, or all its terminated child processes.
+thread, the current process, or all its terminated child processes.
 The
 .Fa who
 argument is either
-.Dv RUSAGE_SELF
+.Dv RUSAGE_THREAD ,
+.Dv RUSAGE_SELF ,
 or
 .Dv RUSAGE_CHILDREN .
 The buffer to which
@@ -175,6 +177,10 @@ The
 .Fn getrusage
 system call appeared in
 .Bx 4.2 .
+The
+.Dv RUSAGE_THREAD
+facility first appeared in
+.Fx 8.1 .
 .Sh BUGS
 There is no way to obtain information about a child process
 that has not yet terminated.

From 603a4d7f411b69db731e4c5d6e51a8e81a6cd510 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 06:04:33 +0000
Subject: [PATCH 440/532] Remove a comment that merely repeats code.

Submitted by:	bde
MFC after:	1 week
---
 sys/kern/kern_thread.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/kern/kern_thread.c b/sys/kern/kern_thread.c
index b32c5842198..dd5cb62273d 100644
--- a/sys/kern/kern_thread.c
+++ b/sys/kern/kern_thread.c
@@ -431,7 +431,6 @@ thread_exit(void)
 #endif
 	PROC_UNLOCK(p);
 	thread_lock(td);
-	/* Save our tick information with both the thread and proc locked */
 	ruxagg_locked(&p->p_rux, td);
 	PROC_SUNLOCK(p);
 	td->td_state = TDS_INACTIVE;

From 9182554ae9bdc596ed5aa8bcd7bab95c6c809151 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Tue, 4 May 2010 06:06:01 +0000
Subject: [PATCH 441/532] Fix typo in comment.

MFC after:	3 days
---
 sys/kern/kern_thread.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/kern_thread.c b/sys/kern/kern_thread.c
index dd5cb62273d..d7a9199c3e5 100644
--- a/sys/kern/kern_thread.c
+++ b/sys/kern/kern_thread.c
@@ -398,7 +398,7 @@ thread_exit(void)
 
 			/*
 			 * The test below is NOT true if we are the
-			 * sole exiting thread. P_STOPPED_SNGL is unset
+			 * sole exiting thread. P_STOPPED_SINGLE is unset
 			 * in exit1() after it is the only survivor.
 			 */
 			if (P_SHOULDSTOP(p) == P_STOPPED_SINGLE) {

From 752fa694029c20397f7928fa5f6b80f787677be2 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 4 May 2010 06:13:17 +0000
Subject: [PATCH 442/532] Go ahead and merge the work edwin@ on tftpd into the
 tree.  It is a lot better than what's in the tree now.  Edwin tested it at a
 prior employer, but can't test it today.  I've found that it works a lot
 better with the various uboot versions that I've used in my embedded work. 
 Here's the pkg-descr from the port that describes the changes:

It all started when we got some new routers, which told me the
following when trying to upload configuration or download images
from it: The TFTP server doesn't support the blocksize option.

My curiousity was triggered, it took me some reading of RFCs and
other documentation to find out what was possible and what could
be done. Was plain TFTP very simple in its handshake, TFTP with
options was kind of messy because of its backwards capability: The
first packet returned could either be an acknowledgement of options,
or the first data packet.

Going through the source code of src/libexec/tftpd and going through
the code of src/usr.bin/tftp showed that there was a lot of duplicate
code, and the addition of options would only increase the amount
of duplicate code. After all, both the client and the server can
act as a sender and receiver.

At the end, it ended up with a nearly complete rewrite of the tftp
client and server. It has been tested against the following TFTP
clients and servers:

- Itself (yay!)
- The standard FreeBSD tftp client and server
- The Fedora Core 6 tftp client and server
- Cisco router tftp client
- Extreme Networks tftp client

It supports the following RFCs:

RFC1350 - THE TFTP PROTOCOL (REVISION 2)
RFC2347 - TFTP Option Extension
RFC2348 - TFTP Blocksize Option
RFC2349 - TFTP Timeout Interval and Transfer Size Options
RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability
          Statement for the Trivial File Transfer Protocol (TFTP)

It supports the following unofficial TFTP Options as described at
http://www.compuphase.com/tftp.htm:

blksize2 - Block size restricted to powers of 2, excluding protocol headers
rollover - Block counter roll-over (roll back to zero or to one)

From the tftp program point of view the following things are changed:

- New commands: "blocksize", "blocksize2", "rollover" and "options"
- Development features: "debug" and "packetdrop"

If you try this tftp/tftpd implementation, please let me know if
it works (or doesn't work) and against which implementaion so I can
get a list of confirmed working systems.

Author: Edwin Groothuis <edwin@FreeBSD.org>
---
 usr.bin/tftp/Makefile |   6 +-
 usr.bin/tftp/main.c   | 654 +++++++++++++++++++++++++++++++-----------
 usr.bin/tftp/tftp.1   |  60 ++--
 usr.bin/tftp/tftp.c   | 586 ++++++++++++-------------------------
 4 files changed, 700 insertions(+), 606 deletions(-)

diff --git a/usr.bin/tftp/Makefile b/usr.bin/tftp/Makefile
index a51afed9974..9be599a607e 100644
--- a/usr.bin/tftp/Makefile
+++ b/usr.bin/tftp/Makefile
@@ -1,9 +1,13 @@
 #	@(#)Makefile	8.1 (Berkeley) 6/6/93
 # $FreeBSD$
 
+CFLAGS=-g -Wall
+WARNS=	3
 PROG=	tftp
-SRCS=	main.c tftp.c tftpsubs.c
+SRCS=	main.c tftp.c tftp-utils.c tftp-io.c tftp-file.c tftp-transfer.c tftp-options.c
 DPADD=	${LIBEDIT} ${LIBTERMCAP}
 LDADD=	-ledit -ltermcap
+CFLAGS+=-I${.CURDIR}/../../libexec/tftpd -I${.CURDIR}/../../usr.bin/tftp
+.PATH:  ${.CURDIR}/../../libexec/tftpd
 
 .include <bsd.prog.mk>
diff --git a/usr.bin/tftp/main.c b/usr.bin/tftp/main.c
index f01d3d8b8ee..26ca024454d 100644
--- a/usr.bin/tftp/main.c
+++ b/usr.bin/tftp/main.c
@@ -54,12 +54,14 @@ __FBSDID("$FreeBSD$");
 #include <sys/param.h>
 #include <sys/types.h>
 #include <sys/socket.h>
+#include <sys/sysctl.h>
 #include <sys/file.h>
 #include <sys/param.h>
+#include <sys/stat.h>
 
 #include <netinet/in.h>
-
 #include <arpa/inet.h>
+#include <arpa/tftp.h>
 
 #include <ctype.h>
 #include <err.h>
@@ -72,119 +74,212 @@ __FBSDID("$FreeBSD$");
 #include <string.h>
 #include <unistd.h>
 
-#include "extern.h"
+#include "tftp-utils.h"
+#include "tftp-io.h"
+#include "tftp-options.h"
+#include "tftp.h"
 
 #define	MAXLINE		200
 #define	TIMEOUT		5		/* secs between rexmt's */
 
-struct	sockaddr_storage peeraddr;
-int	f;
-int	trace;
-int	verbose;
-int	connected;
-char	mode[32];
-char	line[MAXLINE];
-int	margc;
-#define	MAX_MARGV	20
-char	*margv[MAX_MARGV];
-jmp_buf	toplevel;
-volatile int txrx_error;
+static struct	sockaddr_storage peeraddr;
+static int	connected;
+static char	mode[32];
+jmp_buf		toplevel;
+volatile int	txrx_error;
+static int	peer;
 
-void	get(int, char **);
-void	help(int, char **);
-void	intr(int);
-void	modecmd(int, char **);
-void	put(int, char **);
-void	quit(int, char **);
-void	setascii(int, char **);
-void	setbinary(int, char **);
-void	setpeer0(char *, const char *);
-void	setpeer(int, char **);
-void	setrexmt(int, char **);
-void	settimeout(int, char **);
-void	settrace(int, char **);
-void	setverbose(int, char **);
-void	status(int, char **);
+#define	MAX_MARGV	20
+static int	margc;
+static char	*margv[MAX_MARGV];
+
+int		verbose;
+char		*port = NULL;
+
+static void	get(int, char **);
+static void	help(int, char **);
+static void	intr(int);
+static void	modecmd(int, char **);
+static void	put(int, char **);
+static void	quit(int, char **);
+static void	setascii(int, char **);
+static void	setbinary(int, char **);
+static void	setpeer0(char *, const char *);
+static void	setpeer(int, char **);
+static void	settimeoutpacket(int, char **);
+static void	settimeoutnetwork(int, char **);
+static void	setdebug(int, char **);
+static void	setverbose(int, char **);
+static void	showstatus(int, char **);
+static void	setblocksize(int, char **);
+static void	setblocksize2(int, char **);
+static void	setoptions(int, char **);
+static void	setrollover(int, char **);
+static void	setpacketdrop(int, char **);
 
 static void command(void) __dead2;
 static const char *command_prompt(void);
 
-static void getusage(const char *);
-static void makeargv(void);
-static void putusage(const char *);
+static void urihandling(char *URI);
+static void getusage(char *);
+static void makeargv(char *line);
+static void putusage(char *);
 static void settftpmode(const char *);
 
-char	*tail(char *);
-struct	cmd *getcmd(char *);
+static char	*tail(char *);
+static struct	cmd *getcmd(char *);
 
 #define HELPINDENT (sizeof("connect"))
 
 struct cmd {
 	const char	*name;
-	char	*help;
 	void	(*handler)(int, char **);
+	const char	*help;
 };
 
-char	vhelp[] = "toggle verbose mode";
-char	thelp[] = "toggle packet tracing";
-char	chelp[] = "connect to remote tftp";
-char	qhelp[] = "exit tftp";
-char	hhelp[] = "print help information";
-char	shelp[] = "send file";
-char	rhelp[] = "receive file";
-char	mhelp[] = "set file transfer mode";
-char	sthelp[] = "show current status";
-char	xhelp[] = "set per-packet retransmission timeout";
-char	ihelp[] = "set total retransmission timeout";
-char    ashelp[] = "set mode to netascii";
-char    bnhelp[] = "set mode to octet";
+static struct cmd cmdtab[] = {
+	{ "connect",	setpeer,	"connect to remote tftp"	},
+	{ "mode",	modecmd,	"set file transfer mode"	},
+	{ "put",	put,		"send file"			},
+	{ "get",	get,		"receive file"			},
+	{ "quit",	quit,		"exit tftp"			},
+	{ "verbose",	setverbose,	"toggle verbose mode"		},
+	{ "status",	showstatus,	"show current status"		},
+	{ "binary",     setbinary,	"set mode to octet"		},
+	{ "ascii",      setascii,	"set mode to netascii"		},
+	{ "rexmt",	settimeoutpacket,
+	  "set per-packet retransmission timeout[-]" },
+	{ "timeout",	settimeoutnetwork,
+	  "set total retransmission timeout" },
+	{ "trace",	setdebug,	"enable 'debug packet'[-]"	},
+	{ "debug",	setdebug,	"enable verbose output"		},
+	{ "blocksize",	setblocksize,	"set blocksize[*]"		},
+	{ "blocksize2",	setblocksize2,	"set blocksize as a power of 2[**]" },
+	{ "rollover",	setrollover,	"rollover after 64K packets[**]" },
+	{ "options",	setoptions,
+	  "enable or disable RFC2347 style options" },
+	{ "help",	help,		"print help information"	},
+	{ "packetdrop",	setpacketdrop,	"artifical packetloss feature"	},
+	{ "?",		help,		"print help information"	},
+	{ NULL,		NULL,		NULL				}
+};
 
-struct cmd cmdtab[] = {
-	{ "connect",	chelp,		setpeer },
-	{ "mode",       mhelp,          modecmd },
-	{ "put",	shelp,		put },
-	{ "get",	rhelp,		get },
-	{ "quit",	qhelp,		quit },
-	{ "verbose",	vhelp,		setverbose },
-	{ "trace",	thelp,		settrace },
-	{ "status",	sthelp,		status },
-	{ "binary",     bnhelp,         setbinary },
-	{ "ascii",      ashelp,         setascii },
-	{ "rexmt",	xhelp,		setrexmt },
-	{ "timeout",	ihelp,		settimeout },
-	{ "?",		hhelp,		help },
-	{ NULL, NULL, NULL }
+static struct	modes {
+	const char *m_name;
+	const char *m_mode;
+} modes[] = {
+	{ "ascii",	"netascii" },
+	{ "netascii",	"netascii" },
+	{ "binary",	"octet" },
+	{ "image",	"octet" },
+	{ "octet",	"octet" },
+	{ NULL,		NULL }
 };
 
 int
 main(int argc, char *argv[])
 {
-	f = -1;
+
+	acting_as_client = 1;
+	peer = -1;
 	strcpy(mode, "netascii");
 	signal(SIGINT, intr);
 	if (argc > 1) {
 		if (setjmp(toplevel) != 0)
 			exit(txrx_error);
+
+		if (strncmp(argv[1], "tftp://", 7) == 0) {
+			urihandling(argv[1]);
+			exit(txrx_error);
+		}
+
 		setpeer(argc, argv);
 	}
 	if (setjmp(toplevel) != 0)
 		(void)putchar('\n');
+
+	init_options();
 	command();
 }
 
-char    hostname[MAXHOSTNAMELEN];
+/*
+ * RFC3617 handling of TFTP URIs:
+ *
+ *    tftpURI         = "tftp://" host "/" file [ mode ]
+ *    mode            = ";"  "mode=" ( "netascii" / "octet" )
+ *    file            = *( unreserved / escaped )
+ *    host            = <as specified by RFC 2732>
+ *    unreserved      = <as specified in RFC 2396>
+ *    escaped         = <as specified in RFC 2396>
+ *
+ * We are cheating a little bit by allowing any mode as specified in the
+ * modes table defined earlier on in this file and mapping it on the real
+ * mode.
+ */
+static void
+urihandling(char *URI)
+{
+	char	uri[ARG_MAX];
+	char	*host = NULL;
+	char	*path = NULL;
+	char	*options = NULL;
+	char	*mode = "octet";
+	char	*s;
+	char	line[MAXLINE];
+	int	i;
 
-void
-setpeer0(char *host, const char *port)
+	strncpy(uri, URI, ARG_MAX);
+	host = uri + 7;
+
+	if ((s = strchr(host, '/')) == NULL) {
+		fprintf(stderr,
+		    "Invalid URI: Couldn't find / after hostname\n");
+		exit(1);
+	}
+	*s = '\0';
+	path = s + 1;
+
+	if ((s = strchr(path, ';')) != NULL) {
+		*s = '\0';
+		options = s + 1;
+
+		if (strncmp(options, "mode=", 5) == 0) {
+			mode = options;
+			mode += 5;
+
+			for (i = 0; modes[i].m_name != NULL; i++) {
+				if (strcmp(modes[i].m_name, mode) == 0)
+					break;
+			}
+			if (modes[i].m_name == NULL) {
+				fprintf(stderr, "Invalid mode: '%s'\n", mode);
+				exit(1);
+			}
+			settftpmode(modes[i].m_mode);
+		}
+	} else {
+		settftpmode("octet");
+	}
+
+	setpeer0(host, NULL);
+
+	sprintf(line, "get %s", path);
+	makeargv(line);
+	get(margc, margv);
+}
+
+static char    hostname[MAXHOSTNAMELEN];
+
+static void
+setpeer0(char *host, const char *lport)
 {
 	struct addrinfo hints, *res0, *res;
 	int error;
-	struct sockaddr_storage ss;
 	const char *cause = "unknown";
 
 	if (connected) {
-		close(f);
-		f = -1;
+		close(peer);
+		peer = -1;
 	}
 	connected = 0;
 
@@ -193,9 +288,9 @@ setpeer0(char *host, const char *port)
 	hints.ai_socktype = SOCK_DGRAM;
 	hints.ai_protocol = IPPROTO_UDP;
 	hints.ai_flags = AI_CANONNAME;
-	if (!port)
-		port = "tftp";
-	error = getaddrinfo(host, port, &hints, &res0);
+	if (!lport)
+		lport = "tftp";
+	error = getaddrinfo(host, lport, &hints, &res0);
 	if (error) {
 		warnx("%s", gai_strerror(error));
 		return;
@@ -204,50 +299,53 @@ setpeer0(char *host, const char *port)
 	for (res = res0; res; res = res->ai_next) {
 		if (res->ai_addrlen > sizeof(peeraddr))
 			continue;
-		f = socket(res->ai_family, res->ai_socktype, res->ai_protocol);
-		if (f < 0) {
+		peer = socket(res->ai_family, res->ai_socktype,
+			res->ai_protocol);
+		if (peer < 0) {
 			cause = "socket";
 			continue;
 		}
 
-		memset(&ss, 0, sizeof(ss));
-		ss.ss_family = res->ai_family;
-		ss.ss_len = res->ai_addrlen;
-		if (bind(f, (struct sockaddr *)&ss, ss.ss_len) < 0) {
+		memset(&peer_sock, 0, sizeof(peer_sock));
+		peer_sock.ss_family = res->ai_family;
+		peer_sock.ss_len = res->ai_addrlen;
+		if (bind(peer, (struct sockaddr *)&peer_sock, peer_sock.ss_len) < 0) {
 			cause = "bind";
-			close(f);
-			f = -1;
+			close(peer);
+			peer = -1;
 			continue;
 		}
 
 		break;
 	}
 
-	if (f < 0)
+	if (peer < 0)
 		warn("%s", cause);
 	else {
 		/* res->ai_addr <= sizeof(peeraddr) is guaranteed */
-		memcpy(&peeraddr, res->ai_addr, res->ai_addrlen);
+		memcpy(&peer_sock, res->ai_addr, res->ai_addrlen);
 		if (res->ai_canonname) {
-			(void) strlcpy(hostname, res->ai_canonname,
+			(void) strncpy(hostname, res->ai_canonname,
 				sizeof(hostname));
 		} else
-			(void) strlcpy(hostname, host, sizeof(hostname));
+			(void) strncpy(hostname, host, sizeof(hostname));
+		hostname[sizeof(hostname)-1] = 0;
 		connected = 1;
 	}
 
 	freeaddrinfo(res0);
 }
 
-void
+static void
 setpeer(int argc, char *argv[])
 {
+	char	line[MAXLINE];
 
 	if (argc < 2) {
 		strcpy(line, "Connect ");
 		printf("(to) ");
 		fgets(&line[strlen(line)], sizeof line - strlen(line), stdin);
-		makeargv();
+		makeargv(line);
 		argc = margc;
 		argv = margv;
 	}
@@ -255,26 +353,14 @@ setpeer(int argc, char *argv[])
 		printf("usage: %s [host [port]]\n", argv[0]);
 		return;
 	}
-	if (argc == 3)
+	if (argc == 3) {
+		port = argv[2];
 		setpeer0(argv[1], argv[2]);
-	else
+	} else
 		setpeer0(argv[1], NULL);
 }
 
-struct	modes {
-	const char *m_name;
-	const char *m_mode;
-} modes[] = {
-	{ "ascii",	"netascii" },
-	{ "netascii",   "netascii" },
-	{ "binary",     "octet" },
-	{ "image",      "octet" },
-	{ "octet",     "octet" },
-/*      { "mail",       "mail" },       */
-	{ 0,		0 }
-};
-
-void
+static void
 modecmd(int argc, char *argv[])
 {
 	struct modes *p;
@@ -298,7 +384,7 @@ modecmd(int argc, char *argv[])
 
 	printf("usage: %s [", argv[0]);
 	sep = " ";
-	for (p = modes; p->m_name; p++) {
+	for (p = modes; p->m_name != NULL; p++) {
 		printf("%s%s", sep, p->m_name);
 		if (*sep == ' ')
 			sep = " | ";
@@ -307,14 +393,14 @@ modecmd(int argc, char *argv[])
 	return;
 }
 
-void
+static void
 setbinary(int argc __unused, char *argv[] __unused)
 {
 
 	settftpmode("octet");
 }
 
-void
+static void
 setascii(int argc __unused, char *argv[] __unused)
 {
 
@@ -324,6 +410,7 @@ setascii(int argc __unused, char *argv[] __unused)
 static void
 settftpmode(const char *newmode)
 {
+
 	strcpy(mode, newmode);
 	if (verbose)
 		printf("mode set to %s\n", mode);
@@ -333,18 +420,20 @@ settftpmode(const char *newmode)
 /*
  * Send file(s).
  */
-void
+static void
 put(int argc, char *argv[])
 {
-	int fd;
-	int n;
-	char *cp, *targ;
+	int	fd;
+	int	n;
+	char	*cp, *targ;
+	char	line[MAXLINE];
+	struct stat sb;
 
 	if (argc < 2) {
 		strcpy(line, "send ");
 		printf("(file) ");
 		fgets(&line[strlen(line)], sizeof line - strlen(line), stdin);
-		makeargv();
+		makeargv(line);
 		argc = margc;
 		argv = margv;
 	}
@@ -381,10 +470,14 @@ put(int argc, char *argv[])
 			warn("%s", cp);
 			return;
 		}
+
+		stat(cp, &sb);
+		asprintf(&options[OPT_TSIZE].o_request, "%ju", sb.st_size);
+
 		if (verbose)
 			printf("putting %s to %s:%s [%s]\n",
-				cp, hostname, targ, mode);
-		xmitfile(fd, targ, mode);
+			    cp, hostname, targ, mode);
+		xmitfile(peer, port, fd, targ, mode);
 		return;
 	}
 				/* this assumes the target is a directory */
@@ -398,36 +491,43 @@ put(int argc, char *argv[])
 			warn("%s", argv[n]);
 			continue;
 		}
+
+		stat(cp, &sb);
+		asprintf(&options[OPT_TSIZE].o_request, "%ju", sb.st_size);
+
 		if (verbose)
 			printf("putting %s to %s:%s [%s]\n",
-				argv[n], hostname, targ, mode);
-		xmitfile(fd, targ, mode);
+			    argv[n], hostname, targ, mode);
+		xmitfile(peer, port, fd, targ, mode);
 	}
 }
 
 static void
-putusage(const char *s)
+putusage(char *s)
 {
-	printf("usage: %s file [[host:]remotename]\n", s);
+
+	printf("usage: %s file [remotename]\n", s);
+	printf("       %s file host:remotename\n", s);
 	printf("       %s file1 file2 ... fileN [[host:]remote-directory]\n", s);
 }
 
 /*
  * Receive file(s).
  */
-void
+static void
 get(int argc, char *argv[])
 {
 	int fd;
 	int n;
 	char *cp;
 	char *src;
+	char	line[MAXLINE];
 
 	if (argc < 2) {
 		strcpy(line, "get ");
 		printf("(files) ");
 		fgets(&line[strlen(line)], sizeof line - strlen(line), stdin);
-		makeargv();
+		makeargv(line);
 		argc = margc;
 		argv = margv;
 	}
@@ -438,6 +538,8 @@ get(int argc, char *argv[])
 	if (!connected) {
 		for (n = 1; n < argc ; n++)
 			if (rindex(argv[n], ':') == 0) {
+				printf("No remote host specified and "
+				    "no host given for file '%s'\n", argv[n]);
 				getusage(argv[0]);
 				return;
 			}
@@ -468,8 +570,8 @@ get(int argc, char *argv[])
 			}
 			if (verbose)
 				printf("getting from %s:%s to %s [%s]\n",
-					hostname, src, cp, mode);
-			recvfile(fd, src, mode);
+				    hostname, src, cp, mode);
+			recvfile(peer, port, fd, src, mode);
 			break;
 		}
 		cp = tail(src);         /* new .. jdg */
@@ -480,30 +582,31 @@ get(int argc, char *argv[])
 		}
 		if (verbose)
 			printf("getting from %s:%s to %s [%s]\n",
-				hostname, src, cp, mode);
-		recvfile(fd, src, mode);
+			    hostname, src, cp, mode);
+		recvfile(peer, port, fd, src, mode);
 	}
 }
 
 static void
-getusage(const char *s)
+getusage(char *s)
 {
-	printf("usage: %s [host:]file [localname]\n", s);
+
+	printf("usage: %s file [localname]\n", s);
+	printf("       %s [host:]file [localname]\n", s);
 	printf("       %s [host1:]file1 [host2:]file2 ... [hostN:]fileN\n", s);
 }
 
-int	rexmtval = TIMEOUT;
-
-void
-setrexmt(int argc, char *argv[])
+static void
+settimeoutpacket(int argc, char *argv[])
 {
 	int t;
+	char	line[MAXLINE];
 
 	if (argc < 2) {
-		strcpy(line, "Rexmt-timeout ");
+		strcpy(line, "Packet timeout ");
 		printf("(value) ");
 		fgets(&line[strlen(line)], sizeof line - strlen(line), stdin);
-		makeargv();
+		makeargv(line);
 		argc = margc;
 		argv = margv;
 	}
@@ -512,24 +615,25 @@ setrexmt(int argc, char *argv[])
 		return;
 	}
 	t = atoi(argv[1]);
-	if (t < 0)
+	if (t < 0) {
 		printf("%s: bad value\n", argv[1]);
-	else
-		rexmtval = t;
+		return;
+	}
+
+	settimeouts(t, timeoutnetwork, maxtimeouts);
 }
 
-int	maxtimeout = 5 * TIMEOUT;
-
-void
-settimeout(int argc, char *argv[])
+static void
+settimeoutnetwork(int argc, char *argv[])
 {
 	int t;
+	char	line[MAXLINE];
 
 	if (argc < 2) {
-		strcpy(line, "Maximum-timeout ");
+		strcpy(line, "Network timeout ");
 		printf("(value) ");
 		fgets(&line[strlen(line)], sizeof line - strlen(line), stdin);
-		makeargv();
+		makeargv(line);
 		argc = margc;
 		argv = margv;
 	}
@@ -538,26 +642,36 @@ settimeout(int argc, char *argv[])
 		return;
 	}
 	t = atoi(argv[1]);
-	if (t < 0)
+	if (t < 0) {
 		printf("%s: bad value\n", argv[1]);
-	else
-		maxtimeout = t;
+		return;
+	}
+
+	settimeouts(timeoutpacket, t, maxtimeouts);
 }
 
-void
-status(int argc __unused, char *argv[] __unused)
+static void
+showstatus(int argc __unused, char *argv[] __unused)
 {
-	if (connected)
-		printf("Connected to %s.\n", hostname);
-	else
-		printf("Not connected.\n");
-	printf("Mode: %s Verbose: %s Tracing: %s\n", mode,
-		verbose ? "on" : "off", trace ? "on" : "off");
-	printf("Rexmt-interval: %d seconds, Max-timeout: %d seconds\n",
-		rexmtval, maxtimeout);
+
+	printf("Remote host: %s\n",
+	    connected ? hostname : "none specified yet");
+	printf("RFC2347 Options support: %s\n",
+	    options_rfc_enabled ? "enabled" : "disabled");
+	printf("Non-RFC defined options support: %s\n",
+	    options_extra_enabled ? "enabled" : "disabled");
+	printf("Mode: %s\n", mode);
+	printf("Verbose: %s\n", verbose ? "on" : "off");
+	printf("Debug: %s\n", debug_show(debug));
+	printf("Artificial packetloss: %d in 100 packets\n",
+	    packetdroppercentage);
+	printf("Segment size: %d bytes\n", segsize);
+	printf("Network timeout: %d seconds\n", timeoutpacket);
+	printf("Maximum network timeout: %d seconds\n", timeoutnetwork);
+	printf("Maximum timeouts: %d \n", maxtimeouts);
 }
 
-void
+static void
 intr(int dummy __unused)
 {
 
@@ -566,7 +680,7 @@ intr(int dummy __unused)
 	longjmp(toplevel, -1);
 }
 
-char *
+static char *
 tail(char *filename)
 {
 	char *s;
@@ -583,7 +697,7 @@ tail(char *filename)
 }
 
 static const char *
-command_prompt(void)
+command_prompt()
 {
 
 	return ("tftp> ");
@@ -602,6 +716,7 @@ command(void)
 	const char *bp;
 	char *cp;
 	int len, num, vrbose;
+	char	line[MAXLINE];
 
 	vrbose = isatty(0);
 	if (vrbose) {
@@ -623,6 +738,7 @@ command(void)
                         line[len] = '\0';
                         history(hist, &he, H_ENTER, bp);
 		} else {
+			line[0] = 0;
 			if (fgets(line, sizeof line , stdin) == 0) {
 				if (feof(stdin)) {
 					exit(txrx_error);
@@ -635,7 +751,7 @@ command(void)
 			*cp = '\0';
 		if (line[0] == 0)
 			continue;
-		makeargv();
+		makeargv(line);
 		if (margc == 0)
 			continue;
 		c = getcmd(margv[0]);
@@ -651,7 +767,7 @@ command(void)
 	}
 }
 
-struct cmd *
+static struct cmd *
 getcmd(char *name)
 {
 	const char *p, *q;
@@ -683,15 +799,15 @@ getcmd(char *name)
  * Slice a string up into argc/argv.
  */
 static void
-makeargv(void)
+makeargv(char *line)
 {
 	char *cp;
 	char **argp = margv;
 
 	margc = 0;
-	if ((cp = strchr(line, '\n')))
+	if ((cp = strchr(line, '\n')) != NULL)
 		*cp = '\0';
-	for (cp = line; margc < MAX_MARGV - 1 && *cp;) {
+	for (cp = line; margc < MAX_MARGV - 1 && *cp != '\0';) {
 		while (isspace(*cp))
 			cp++;
 		if (*cp == '\0')
@@ -707,16 +823,17 @@ makeargv(void)
 	*argp++ = 0;
 }
 
-void
+static void
 quit(int argc __unused, char *argv[] __unused)
 {
+
 	exit(txrx_error);
 }
 
 /*
  * Help command.
  */
-void
+static void
 help(int argc, char *argv[])
 {
 	struct cmd *c;
@@ -725,6 +842,10 @@ help(int argc, char *argv[])
 		printf("Commands may be abbreviated.  Commands are:\n\n");
 		for (c = cmdtab; c->name; c++)
 			printf("%-*s\t%s\n", (int)HELPINDENT, c->name, c->help);
+
+		printf("\n[-] : You shouldn't use these ones anymore.\n");
+		printf("[*] : RFC2834 options support required.\n");
+		printf("[**] : Non-standard RFC2834 option.\n");
 		return;
 	}
 	while (--argc > 0) {
@@ -732,24 +853,211 @@ help(int argc, char *argv[])
 		arg = *++argv;
 		c = getcmd(arg);
 		if (c == (struct cmd *)-1)
-			printf("?Ambiguous help command %s\n", arg);
+			printf("?Ambiguous help command: %s\n", arg);
 		else if (c == (struct cmd *)0)
-			printf("?Invalid help command %s\n", arg);
+			printf("?Invalid help command: %s\n", arg);
 		else
 			printf("%s\n", c->help);
 	}
 }
 
-void
-settrace(int argc __unused, char **argv __unused)
+static void
+setverbose(int argc __unused, char *argv[] __unused)
 {
-	trace = !trace;
-	printf("Packet tracing %s.\n", trace ? "on" : "off");
-}
 
-void
-setverbose(int argc __unused, char **argv __unused)
-{
 	verbose = !verbose;
 	printf("Verbose mode %s.\n", verbose ? "on" : "off");
 }
+
+static void
+setoptions(int argc, char *argv[])
+{
+
+	if (argc == 2) {
+		if (strcasecmp(argv[1], "enable") == 0 ||
+		    strcasecmp(argv[1], "on") == 0) {
+			options_extra_enabled = 1;
+			options_rfc_enabled = 1;
+		}
+		if (strcasecmp(argv[1], "disable") == 0 ||
+		    strcasecmp(argv[1], "off") == 0) {
+			options_extra_enabled = 0;
+			options_rfc_enabled = 0;
+		}
+		if (strcasecmp(argv[1], "extra") == 0)
+			options_extra_enabled = !options_extra_enabled;
+	}
+	printf("Support for RFC2347 style options are now %s.\n",
+	    options_rfc_enabled ? "enabled" : "disabled");
+	printf("Support for non-RFC defined options are now %s.\n",
+	    options_extra_enabled ? "enabled" : "disabled");
+
+	printf("\nThe following options are available:\n"
+	    "\toptions on	: enable support for RFC2347 style options\n"
+	    "\toptions off	: disable support for RFC2347 style options\n"
+	    "\toptions extra	: toggle support for non-RFC defined options\n"
+	);
+}
+
+static void
+setrollover(int argc, char *argv[])
+{
+
+	if (argc == 2) {
+		if (strcasecmp(argv[1], "never") == 0 ||
+		    strcasecmp(argv[1], "none") == 0) {
+			free(options[OPT_ROLLOVER].o_request);
+			options[OPT_ROLLOVER].o_request = NULL;
+		}
+		if (strcasecmp(argv[1], "1") == 0) {
+			free(options[OPT_ROLLOVER].o_request);
+			options[OPT_ROLLOVER].o_request = strdup("1");
+		}
+		if (strcasecmp(argv[1], "0") == 0) {
+			free(options[OPT_ROLLOVER].o_request);
+			options[OPT_ROLLOVER].o_request = strdup("0");
+		}
+	}
+	printf("Support for the rollover options is %s.\n",
+	    options[OPT_ROLLOVER].o_request != NULL ? "enabled" : "disabled");
+	if (options[OPT_ROLLOVER].o_request != NULL)
+		printf("Block rollover will be to block %s.\n",
+		    options[OPT_ROLLOVER].o_request);
+
+
+	printf("\nThe following rollover options are available:\n"
+	    "\trollover 0	: rollover to block zero (default)\n"
+	    "\trollover 1	: rollover to block one\n"
+	    "\trollover never	: do not support the rollover option\n"
+	    "\trollover none	: do not support the rollover option\n"
+	);
+}
+
+static void
+setdebug(int argc, char *argv[])
+{
+	int i;
+
+	if (argc != 1) {
+		i = 1;
+		while (i < argc)
+			debug ^= debug_find(argv[i++]);
+	}
+	printf("The following debugging is enabled: %s\n", debug_show(debug));
+
+	printf("\nThe following debugs are available:\n");
+	i = 0;
+	while (debugs[i].name != NULL) {
+		printf("\t%s\t%s\n", debugs[i].name, debugs[i].desc);
+		i++;
+	}
+}
+
+static void
+setblocksize(int argc, char *argv[])
+{
+
+	if (!options_rfc_enabled)
+		printf("RFC2347 style options are not enabled "
+		    "(but proceding anyway)\n");
+
+	if (argc != 1) {
+		int size = atoi(argv[1]);
+		size_t max;
+		char maxbuffer[100];
+		int *maxdgram;
+
+		max = sizeof(maxbuffer);
+		if (sysctlbyname("net.inet.udp.maxdgram",
+			maxbuffer, &max, NULL, 0) < 0) {
+			perror("sysctl: net.inet.udp.maxdgram");
+			return;
+		}
+		maxdgram = (int *)maxbuffer;
+
+		if (size < BLKSIZE_MIN || size > BLKSIZE_MAX) {
+			printf("Blocksize should be between %d and %d bytes.\n",
+				BLKSIZE_MIN, BLKSIZE_MAX);
+			return;
+		} else if (size > *maxdgram - 4) {
+			printf("Blocksize can't be bigger than %d bytes due "
+			    "to the net.inet.udp.maxdgram sysctl limitation.\n",
+			    *maxdgram - 4);
+			asprintf(&options[OPT_BLKSIZE].o_request,
+			    "%d", *maxdgram - 4);
+		} else {
+			asprintf(&options[OPT_BLKSIZE].o_request, "%d", size);
+		}
+	}
+	printf("Blocksize is now %s bytes.\n", options[OPT_BLKSIZE].o_request);
+}
+
+static void
+setblocksize2(int argc, char *argv[])
+{
+
+	if (!options_rfc_enabled || !options_extra_enabled)
+		printf(
+		    "RFC2347 style or non-RFC defined options are not enabled "
+		    "(but proceding anyway)\n");
+
+	if (argc != 1) {
+		int size = atoi(argv[1]);
+		int i;
+		size_t max;
+		char maxbuffer[100];
+		int *maxdgram;
+
+		int sizes[] = {
+			8, 16, 32, 64, 128, 256, 512, 1024,
+			2048, 4096, 8192, 16384, 32768, 0
+		};
+
+		max = sizeof(maxbuffer);
+		if (sysctlbyname("net.inet.udp.maxdgram",
+			maxbuffer, &max, NULL, 0) < 0) {
+			perror("sysctl: net.inet.udp.maxdgram");
+			return;
+		}
+		maxdgram = (int *)maxbuffer;
+
+		for (i = 0; sizes[i] != 0; i++) {
+			if (sizes[i] == size) break;
+		}
+		if (sizes[i] == 0) {
+			printf("Blocksize2 should be a power of two between "
+			    "8 and 32768.\n");
+			return;
+		}
+
+		if (size < BLKSIZE_MIN || size > BLKSIZE_MAX) {
+			printf("Blocksize2 should be between "
+			    "%d and %d bytes.\n", BLKSIZE_MIN, BLKSIZE_MAX);
+			return;
+		} else if (size > *maxdgram - 4) {
+			printf("Blocksize2 can't be bigger than %d bytes due "
+			    "to the net.inet.udp.maxdgram sysctl limitation.\n",
+			    *maxdgram - 4);
+			for (i = 0; sizes[i+1] != 0; i++) {
+				if (*maxdgram < sizes[i+1]) break;
+			}
+			asprintf(&options[OPT_BLKSIZE2].o_request,
+			    "%d", sizes[i]);
+		} else {
+			asprintf(&options[OPT_BLKSIZE2].o_request, "%d", size);
+		}
+	}
+	printf("Blocksize2 is now %s bytes.\n",
+	    options[OPT_BLKSIZE2].o_request);
+}
+
+static void
+setpacketdrop(int argc, char *argv[])
+{
+
+	if (argc != 1)
+		packetdroppercentage = atoi(argv[1]);
+
+	printf("Randomly %d in 100 packets will be dropped\n",
+	    packetdroppercentage);
+}
diff --git a/usr.bin/tftp/tftp.1 b/usr.bin/tftp/tftp.1
index 8924eb5e561..04daa87aace 100644
--- a/usr.bin/tftp/tftp.1
+++ b/usr.bin/tftp/tftp.1
@@ -37,7 +37,7 @@
 .Os
 .Sh NAME
 .Nm tftp
-.Nd "trivial file transfer program"
+.Nd trivial file transfer program
 .Sh SYNOPSIS
 .Nm
 .Op Ar host Op Ar port
@@ -55,28 +55,26 @@ may be specified on the command line, in which case
 uses
 .Ar host
 as the default host for future transfers (see the
-.Ic connect
+.Cm connect
 command below).
 .Sh COMMANDS
 Once
 .Nm
 is running, it issues the prompt
-.Dq Li "tftp> "
+.Dq Li tftp>
 and recognizes the following commands:
 .Pp
-.Bl -tag -width ".Ic verbose" -compact
-.It Ic \&? Ar command-name ...
+.Bl -tag -width verbose -compact
+.It Cm \&? Ar command-name ...
 Print help information.
 .Pp
-.It Ic ascii
-Shorthand for
-.Ic mode Cm ascii .
+.It Cm ascii
+Shorthand for "mode ascii"
 .Pp
-.It Ic binary
-Shorthand for
-.Ic mode Cm binary .
+.It Cm binary
+Shorthand for "mode binary"
 .Pp
-.It Ic connect Ar host Op Ar port
+.It Cm connect Ar host Op Ar port
 Set the
 .Ar host
 (and optionally
@@ -88,19 +86,19 @@ protocol, unlike the
 .Tn FTP
 protocol,
 does not maintain connections between transfers; thus, the
-.Ic connect
+.Cm connect
 command does not actually create a connection,
 but merely remembers what host is to be used for transfers.
 You do not have to use the
-.Ic connect
+.Cm connect
 command; the remote host can be specified as part of the
-.Ic get
+.Cm get
 or
-.Ic put
+.Cm put
 commands.
 .Pp
-.It Ic get Oo Ar host : Oc Ns Ar file Op Ar localname
-.It Ic get Xo
+.It Cm get Oo Ar host : Oc Ns Ar file Op Ar localname
+.It Cm get Xo
 .Oo Ar host1 : Oc Ns Ar file1
 .Oo Ar host2 : Oc Ns Ar file2 ...
 .Oo Ar hostN : Oc Ns Ar fileN
@@ -126,18 +124,18 @@ to disambiguate the
 colons used in the IPv6 address from the colon separating the host and
 the filename.
 .Pp
-.It Ic mode Ar transfer-mode
+.It Cm mode Ar transfer-mode
 Set the mode for transfers;
 .Ar transfer-mode
 may be one of
-.Cm ascii
+.Em ascii
 or
-.Cm binary .
+.Em binary .
 The default is
-.Cm ascii .
+.Em ascii .
 .Pp
-.It Ic put Ar file Op Oo Ar host : Oc Ns Ar remotename
-.It Ic put Ar file1 file2 ... fileN Op Oo Ar host : Oc Ns Ar remote-directory
+.It Cm put Ar file Op Oo Ar host : Oc Ns Ar remotename
+.It Cm put Ar file1 file2 ... fileN Op Oo Ar host : Oc Ns Ar remote-directory
 Put a file or set of files to the remote host.
 When
 .Ar remotename
@@ -152,27 +150,27 @@ machine.
 To specify an IPv6 numeric address for a
 .Ar host ,
 see the example under the
-.Ic get
+.Cm get
 command.
 .Pp
-.It Ic quit
+.It Cm quit
 Exit
 .Nm .
 An end of file also exits.
 .Pp
-.It Ic rexmt Ar retransmission-timeout
+.It Cm rexmt Ar retransmission-timeout
 Set the per-packet retransmission timeout, in seconds.
 .Pp
-.It Ic status
+.It Cm status
 Show current status.
 .Pp
-.It Ic timeout Ar total-transmission-timeout
+.It Cm timeout Ar total-transmission-timeout
 Set the total transmission timeout, in seconds.
 .Pp
-.It Ic trace
+.It Cm trace
 Toggle packet tracing.
 .Pp
-.It Ic verbose
+.It Cm verbose
 Toggle verbose mode.
 .El
 .Sh HISTORY
diff --git a/usr.bin/tftp/tftp.c b/usr.bin/tftp/tftp.c
index 43009029386..29ba2b26c74 100644
--- a/usr.bin/tftp/tftp.c
+++ b/usr.bin/tftp/tftp.c
@@ -45,446 +45,230 @@ __FBSDID("$FreeBSD$");
 /*
  * TFTP User Program -- Protocol Machines
  */
-#include <sys/types.h>
 #include <sys/socket.h>
-#include <sys/time.h>
+#include <sys/stat.h>
 
 #include <netinet/in.h>
 
-#include <arpa/inet.h>
 #include <arpa/tftp.h>
 
 #include <err.h>
-#include <errno.h>
-#include <setjmp.h>
-#include <signal.h>
-#include <stdio.h>
-#include <string.h>
-#include <unistd.h>
 #include <netdb.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
 
-#include "extern.h"
-#include "tftpsubs.h"
-
-extern  struct sockaddr_storage peeraddr; /* filled in by main */
-extern  int     f;			/* the opened socket */
-extern  int     trace;
-extern  int     verbose;
-extern  int     rexmtval;
-extern  int     maxtimeout;
-extern  volatile int txrx_error;
-
-#define PKTSIZE    SEGSIZE+4
-char    ackbuf[PKTSIZE];
-int	timeout;
-jmp_buf	toplevel;
-jmp_buf	timeoutbuf;
-
-static void nak(int, const struct sockaddr *);
-static int makerequest(int, const char *, struct tftphdr *, const char *);
-static void printstats(const char *, unsigned long);
-static void startclock(void);
-static void stopclock(void);
-static void timer(int);
-static void tpacket(const char *, struct tftphdr *, int);
-static int cmpport(const struct sockaddr *, const struct sockaddr *);
+#include "tftp.h"
+#include "tftp-file.h"
+#include "tftp-utils.h"
+#include "tftp-io.h"
+#include "tftp-transfer.h"
+#include "tftp-options.h"
 
 /*
  * Send the requested file.
  */
 void
-xmitfile(int fd, const char *name, const char *mode)
+xmitfile(int peer, char *port, int fd, char *name, char *mode)
 {
-	struct tftphdr *ap;	   /* data and ack packets */
-	struct tftphdr *dp;
-	int n;
-	volatile unsigned short block;
-	volatile int size, convert;
-	volatile unsigned long amount;
-	struct sockaddr_storage from;
-	socklen_t fromlen;
-	FILE *file;
-	struct sockaddr_storage peer;
+	struct tftphdr *rp;
+	int n, i;
+	uint16_t block;
+	uint32_t amount;
 	struct sockaddr_storage serv;	/* valid server port number */
+	char recvbuffer[MAXPKTSIZE];
+	struct tftp_stats tftp_stats;
+
+	stats_init(&tftp_stats);
 
-	startclock();		/* start stat's clock */
-	dp = r_init();		/* reset fillbuf/read-ahead code */
-	ap = (struct tftphdr *)ackbuf;
-	file = fdopen(fd, "r");
-	convert = !strcmp(mode, "netascii");
-	block = 0;
-	amount = 0;
-	memcpy(&peer, &peeraddr, peeraddr.ss_len);
 	memset(&serv, 0, sizeof(serv));
+	rp = (struct tftphdr *)recvbuffer;
 
-	signal(SIGALRM, timer);
-	do {
-		if (block == 0)
-			size = makerequest(WRQ, name, dp, mode) - 4;
-		else {
-		/*	size = read(fd, dp->th_data, SEGSIZE);	 */
-			size = readit(file, &dp, convert);
-			if (size < 0) {
-				nak(errno + 100, (struct sockaddr *)&peer);
-				break;
-			}
-			dp->th_opcode = htons((u_short)DATA);
-			dp->th_block = htons((u_short)block);
-		}
-		timeout = 0;
-		(void) setjmp(timeoutbuf);
-send_data:
-		if (trace)
-			tpacket("sent", dp, size + 4);
-		n = sendto(f, dp, size + 4, 0,
-		    (struct sockaddr *)&peer, peer.ss_len);
-		if (n != size + 4) {
-			warn("sendto");
-			txrx_error = 1;
-			goto abort;
-		}
-		read_ahead(file, convert);
-		for ( ; ; ) {
-			alarm(rexmtval);
-			do {
-				fromlen = sizeof(from);
-				n = recvfrom(f, ackbuf, sizeof(ackbuf), 0,
-				    (struct sockaddr *)&from, &fromlen);
-			} while (n <= 0);
-			alarm(0);
-			if (n < 0) {
-				warn("recvfrom");
-				txrx_error = 1;
-				goto abort;
-			}
-			if (!serv.ss_family)
-				serv = from;
-			else if (!cmpport((struct sockaddr *)&serv,
-			    (struct sockaddr *)&from)) {
-				warn("server port mismatch");
-				txrx_error = 1;
-				goto abort;
-			}
-			peer = from;
-			if (trace)
-				tpacket("received", ap, n);
-			/* should verify packet came from server */
-			ap->th_opcode = ntohs(ap->th_opcode);
-			ap->th_block = ntohs(ap->th_block);
-			if (ap->th_opcode == ERROR) {
-				printf("Error code %d: %s\n", ap->th_code,
-					ap->th_msg);
-				txrx_error = 1;
-				goto abort;
-			}
-			if (ap->th_opcode == ACK) {
-				int j;
+	if (port == NULL) {
+		struct servent *se;
+		se = getservbyname("tftp", "udp");
+		((struct sockaddr_in *)&peer_sock)->sin_port = se->s_port;
+	} else
+		((struct sockaddr_in *)&peer_sock)->sin_port =
+		    htons(atoi(port));
 
-				if (ap->th_block == block) {
-					break;
-				}
-				/* On an error, try to synchronize
-				 * both sides.
-				 */
-				j = synchnet(f);
-				if (j && trace) {
-					printf("discarded %d packets\n",
-							j);
-				}
-				if (ap->th_block == (block-1)) {
-					goto send_data;
-				}
-			}
+	for (i = 0; i < 12; i++) {
+		struct sockaddr_storage from;
+
+		/* Tell the other side what we want to do */
+		if (debug&DEBUG_SIMPLE)
+			printf("Sending %s\n", name);
+
+		n = send_wrq(peer, name, mode);
+		if (n > 0) {
+			printf("Cannot send WRQ packet\n");
+			return;
 		}
-		if (block > 0)
-			amount += size;
-		block++;
-	} while (size == SEGSIZE || block == 1);
-abort:
-	fclose(file);
-	stopclock();
+
+		/*
+		 * The first packet we receive has the new destination port
+		 * we have to send the next packets to.
+		 */
+		n = receive_packet(peer, recvbuffer,
+		    MAXPKTSIZE, &from, timeoutpacket);
+
+		/* We got some data! */
+		if (n >= 0) {
+			((struct sockaddr_in *)&peer_sock)->sin_port =
+			    ((struct sockaddr_in *)&from)->sin_port;
+			break;
+		}
+
+		/* This should be retried */
+		if (n == RP_TIMEOUT) {
+			printf("Try %d, didn't receive answer from remote.\n",
+			    i + 1);
+			continue;
+		}
+
+		/* Everything else is fatal */
+		break;
+	}
+	if (i == 12) {
+		printf("Transfer timed out.\n");
+		return;
+	}
+	if (rp->th_opcode == ERROR) {
+		printf("Got ERROR, aborted\n");
+		return;
+	}
+
+	/*
+	 * If the first packet is an OACK instead of an ACK packet,
+	 * handle it different.
+	 */
+	if (rp->th_opcode == OACK) {
+		if (!options_rfc_enabled) {
+			printf("Got OACK while options are not enabled!\n");
+			send_error(peer, EBADOP);
+			return;
+		}
+
+		parse_options(peer, rp->th_stuff, n + 2);
+	}
+
+	if (read_init(fd, NULL, mode) < 0) {
+		warn("read_init()");
+		return;
+	}
+
+	block = 1;
+	tftp_send(peer, &block, &tftp_stats);
+
+	read_close();
 	if (amount > 0)
-		printstats("Sent", amount);
+		printstats("Sent", verbose, &tftp_stats);
+
+	txrx_error = 1;
 }
 
 /*
  * Receive a file.
  */
 void
-recvfile(int fd, const char *name, const char *mode)
+recvfile(int peer, char *port, int fd, char *name, char *mode)
 {
-	struct tftphdr *ap;
-	struct tftphdr *dp;
-	int n;
-	volatile unsigned short block;
-	volatile int size, firsttrip;
-	volatile unsigned long amount;
-	struct sockaddr_storage from;
-	socklen_t fromlen;
-	FILE *file;
-	volatile int convert;		/* true if converting crlf -> lf */
-	struct sockaddr_storage peer;
-	struct sockaddr_storage serv;	/* valid server port number */
+	struct tftphdr *rp;
+	uint16_t block;
+	char recvbuffer[MAXPKTSIZE];
+	int n, i;
+	struct tftp_stats tftp_stats;
 
-	startclock();
-	dp = w_init();
-	ap = (struct tftphdr *)ackbuf;
-	file = fdopen(fd, "w");
-	convert = !strcmp(mode, "netascii");
-	block = 1;
-	firsttrip = 1;
-	amount = 0;
-	memcpy(&peer, &peeraddr, peeraddr.ss_len);
-	memset(&serv, 0, sizeof(serv));
+	stats_init(&tftp_stats);
 
-	signal(SIGALRM, timer);
-	do {
-		if (firsttrip) {
-			size = makerequest(RRQ, name, ap, mode);
-			firsttrip = 0;
-		} else {
-			ap->th_opcode = htons((u_short)ACK);
-			ap->th_block = htons((u_short)(block));
-			size = 4;
-			block++;
-		}
-		timeout = 0;
-		(void) setjmp(timeoutbuf);
-send_ack:
-		if (trace)
-			tpacket("sent", ap, size);
-		if (sendto(f, ackbuf, size, 0, (struct sockaddr *)&peer,
-		    peer.ss_len) != size) {
-			alarm(0);
-			warn("sendto");
-			txrx_error = 1;
-			goto abort;
-		}
-		write_behind(file, convert);
-		for ( ; ; ) {
-			alarm(rexmtval);
-			do  {
-				fromlen = sizeof(from);
-				n = recvfrom(f, dp, PKTSIZE, 0,
-				    (struct sockaddr *)&from, &fromlen);
-			} while (n <= 0);
-			alarm(0);
-			if (n < 0) {
-				warn("recvfrom");
-				txrx_error = 1;
-				goto abort;
-			}
-			if (!serv.ss_family)
-				serv = from;
-			else if (!cmpport((struct sockaddr *)&serv,
-			    (struct sockaddr *)&from)) {
-				warn("server port mismatch");
-				txrx_error = 1;
-				goto abort;
-			}
-			peer = from;
-			if (trace)
-				tpacket("received", dp, n);
-			/* should verify client address */
-			dp->th_opcode = ntohs(dp->th_opcode);
-			dp->th_block = ntohs(dp->th_block);
-			if (dp->th_opcode == ERROR) {
-				printf("Error code %d: %s\n", dp->th_code,
-					dp->th_msg);
-				txrx_error = 1;
-				goto abort;
-			}
-			if (dp->th_opcode == DATA) {
-				int j;
+	rp = (struct tftphdr *)recvbuffer;
 
-				if (dp->th_block == block) {
-					break;		/* have next packet */
-				}
-				/* On an error, try to synchronize
-				 * both sides.
-				 */
-				j = synchnet(f);
-				if (j && trace) {
-					printf("discarded %d packets\n", j);
-				}
-				if (dp->th_block == (block-1)) {
-					goto send_ack;	/* resend ack */
-				}
-			}
+	if (port == NULL) {
+		struct servent *se;
+		se = getservbyname("tftp", "udp");
+		((struct sockaddr_in *)&peer_sock)->sin_port = se->s_port;
+	} else
+		((struct sockaddr_in *)&peer_sock)->sin_port =
+		    htons(atoi(port));
+
+	for (i = 0; i < 12; i++) {
+		struct sockaddr_storage from;
+
+		/* Tell the other side what we want to do */
+		if (debug&DEBUG_SIMPLE)
+			printf("Requesting %s\n", name);
+
+		n = send_rrq(peer, name, mode);
+		if (n > 0) {
+			printf("Cannot send RRQ packet\n");
+			return;
 		}
-	/*	size = write(fd, dp->th_data, n - 4); */
-		size = writeit(file, &dp, n - 4, convert);
-		if (size < 0) {
-			nak(errno + 100, (struct sockaddr *)&peer);
+
+		/*
+		 * The first packet we receive has the new destination port
+		 * we have to send the next packets to.
+		 */
+		n = receive_packet(peer, recvbuffer,
+		    MAXPKTSIZE, &from, timeoutpacket);
+
+		/* We got something useful! */
+		if (n >= 0) {
+			((struct sockaddr_in *)&peer_sock)->sin_port =
+			    ((struct sockaddr_in *)&from)->sin_port;
 			break;
 		}
-		amount += size;
-	} while (size == SEGSIZE);
-abort:						/* ok to ack, since user */
-	ap->th_opcode = htons((u_short)ACK);	/* has seen err msg */
-	ap->th_block = htons((u_short)block);
-	(void) sendto(f, ackbuf, 4, 0, (struct sockaddr *)&peer,
-	    peer.ss_len);
-	write_behind(file, convert);		/* flush last buffer */
-	fclose(file);
-	stopclock();
-	if (amount > 0)
-		printstats("Received", amount);
-}
 
-static int
-makerequest(int request, const char *name, struct tftphdr *tp, const char *mode)
-{
-	char *cp;
+		/* We should retry if this happens */
+		if (n == RP_TIMEOUT) {
+			printf("Try %d, didn't receive answer from remote.\n",
+			    i + 1);
+			continue;
+		}
 
-	tp->th_opcode = htons((u_short)request);
-	cp = tp->th_stuff;
-	strcpy(cp, name);
-	cp += strlen(name);
-	*cp++ = '\0';
-	strcpy(cp, mode);
-	cp += strlen(mode);
-	*cp++ = '\0';
-	return (cp - (char *)tp);
-}
-
-struct errmsg {
-	int	e_code;
-	const char	*e_msg;
-} errmsgs[] = {
-	{ EUNDEF,	"Undefined error code" },
-	{ ENOTFOUND,	"File not found" },
-	{ EACCESS,	"Access violation" },
-	{ ENOSPACE,	"Disk full or allocation exceeded" },
-	{ EBADOP,	"Illegal TFTP operation" },
-	{ EBADID,	"Unknown transfer ID" },
-	{ EEXISTS,	"File already exists" },
-	{ ENOUSER,	"No such user" },
-	{ -1,		0 }
-};
-
-/*
- * Send a nak packet (error message).
- * Error code passed in is one of the
- * standard TFTP codes, or a UNIX errno
- * offset by 100.
- */
-static void
-nak(int error, const struct sockaddr *peer)
-{
-	struct errmsg *pe;
-	struct tftphdr *tp;
-	int length;
-
-	tp = (struct tftphdr *)ackbuf;
-	tp->th_opcode = htons((u_short)ERROR);
-	tp->th_code = htons((u_short)error);
-	for (pe = errmsgs; pe->e_code >= 0; pe++)
-		if (pe->e_code == error)
-			break;
-	if (pe->e_code < 0) {
-		pe->e_msg = strerror(error - 100);
-		tp->th_code = EUNDEF;
-	}
-	strcpy(tp->th_msg, pe->e_msg);
-	length = strlen(pe->e_msg) + 4;
-	if (trace)
-		tpacket("sent", tp, length);
-	if (sendto(f, ackbuf, length, 0, peer, peer->sa_len) != length)
-		warn("nak");
-}
-
-static void
-tpacket(const char *s, struct tftphdr *tp, int n)
-{
-	static const char *opcodes[] =
-	   { "#0", "RRQ", "WRQ", "DATA", "ACK", "ERROR" };
-	char *cp, *file;
-	u_short op = ntohs(tp->th_opcode);
-
-	if (op < RRQ || op > ERROR)
-		printf("%s opcode=%x ", s, op);
-	else
-		printf("%s %s ", s, opcodes[op]);
-	switch (op) {
-
-	case RRQ:
-	case WRQ:
-		n -= 2;
-		file = cp = tp->th_stuff;
-		cp = index(cp, '\0');
-		printf("<file=%s, mode=%s>\n", file, cp + 1);
-		break;
-
-	case DATA:
-		printf("<block=%d, %d bytes>\n", ntohs(tp->th_block), n - 4);
-		break;
-
-	case ACK:
-		printf("<block=%d>\n", ntohs(tp->th_block));
-		break;
-
-	case ERROR:
-		printf("<code=%d, msg=%s>\n", ntohs(tp->th_code), tp->th_msg);
+		/* Otherwise it is a fatal error */
 		break;
 	}
-}
 
-struct timeval tstart;
-struct timeval tstop;
-
-static void
-startclock(void)
-{
-
-	(void)gettimeofday(&tstart, NULL);
-}
-
-static void
-stopclock(void)
-{
-
-	(void)gettimeofday(&tstop, NULL);
-}
-
-static void
-printstats(const char *direction, unsigned long amount)
-{
-	double delta;
-			/* compute delta in 1/10's second units */
-	delta = ((tstop.tv_sec*10.)+(tstop.tv_usec/100000)) -
-		((tstart.tv_sec*10.)+(tstart.tv_usec/100000));
-	delta = delta/10.;      /* back to seconds */
-	printf("%s %ld bytes in %.1f seconds", direction, amount, delta);
-	if (verbose)
-		printf(" [%.0f bits/sec]", (amount*8.)/delta);
-	putchar('\n');
-}
-
-static void
-timer(int sig __unused)
-{
-
-	timeout += rexmtval;
-	if (timeout >= maxtimeout) {
-		printf("Transfer timed out.\n");
-		longjmp(toplevel, -1);
+	if (rp->th_opcode == ERROR) {
+		tftp_log(LOG_ERR, "Error code %d: %s", rp->th_code, rp->th_msg);
+		return;
 	}
-        txrx_error = 1;
-	longjmp(timeoutbuf, 1);
-}
-
-static int
-cmpport(const struct sockaddr *sa, const struct sockaddr *sb)
-{
-	char a[NI_MAXSERV], b[NI_MAXSERV];
-
-	if (getnameinfo(sa, sa->sa_len, NULL, 0, a, sizeof(a), NI_NUMERICSERV))
-		return 0;
-	if (getnameinfo(sb, sb->sa_len, NULL, 0, b, sizeof(b), NI_NUMERICSERV))
-		return 0;
-	if (strcmp(a, b) != 0)
-		return 0;
-
-	return 1;
+
+	if (write_init(fd, NULL, mode) < 0) {
+		warn("write_init");
+		return;
+	}
+
+	stats_init(&tftp_stats);
+
+	/*
+	 * If the first packet is an OACK packet instead of an DATA packet,
+	 * handle it different.
+	 */
+	if (rp->th_opcode == OACK) {
+		if (!options_rfc_enabled) {
+			printf("Got OACK while options are not enabled!\n");
+			send_error(peer, EBADOP);
+			return;
+		}
+
+		parse_options(peer, rp->th_stuff, n + 2);
+
+		n = send_ack(peer, 0);
+		if (n > 0) {
+			printf("Cannot send ACK on OACK.\n");
+			return;
+		}
+		block = 0;
+		tftp_receive(peer, &block, &tftp_stats, NULL, 0);
+	} else {
+		block = 1;
+		tftp_receive(peer, &block, &tftp_stats, rp, n);
+	}
+
+	write_close();
+	if (tftp_stats.amount > 0)
+		printstats("Received", verbose, &tftp_stats);
+	return;
 }

From 5276e6398137bcf9ea40715400a287fd62c23c3e Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 4 May 2010 06:19:19 +0000
Subject: [PATCH 443/532] Go ahead and merge the work edwin@ on tftpd into the
 tree.  It is a lot better than what's in the tree now.  Edwin tested it at a
 prior employer, but can't test it today.  I've found that it works a lot
 better with the various uboot versions that I've used in my embedded work. 
 Here's the pkg-descr from the port that describes the changes:

It all started when we got some new routers, which told me the
following when trying to upload configuration or download images
from it: The TFTP server doesn't support the blocksize option.

My curiousity was triggered, it took me some reading of RFCs and
other documentation to find out what was possible and what could
be done. Was plain TFTP very simple in its handshake, TFTP with
options was kind of messy because of its backwards capability: The
first packet returned could either be an acknowledgement of options,
or the first data packet.

Going through the source code of src/libexec/tftpd and going through
the code of src/usr.bin/tftp showed that there was a lot of duplicate
code, and the addition of options would only increase the amount
of duplicate code. After all, both the client and the server can
act as a sender and receiver.

At the end, it ended up with a nearly complete rewrite of the tftp
client and server. It has been tested against the following TFTP
clients and servers:

- Itself (yay!)
- The standard FreeBSD tftp client and server
- The Fedora Core 6 tftp client and server
- Cisco router tftp client
- Extreme Networks tftp client

It supports the following RFCs:

RFC1350 - THE TFTP PROTOCOL (REVISION 2)
RFC2347 - TFTP Option Extension
RFC2348 - TFTP Blocksize Option
RFC2349 - TFTP Timeout Interval and Transfer Size Options
RFC3617 - Uniform Resource Identifier (URI) Scheme and Applicability
          Statement for the Trivial File Transfer Protocol (TFTP)

It supports the following unofficial TFTP Options as described at
http://www.compuphase.com/tftp.htm:

blksize2 - Block size restricted to powers of 2, excluding protocol headers
rollover - Block counter roll-over (roll back to zero or to one)

From the tftp program point of view the following things are changed:

- New commands: "blocksize", "blocksize2", "rollover" and "options"
- Development features: "debug" and "packetdrop"

If you try this tftp/tftpd implementation, please let me know if
it works (or doesn't work) and against which implementaion so I can
get a list of confirmed working systems.

Author: Edwin Groothuis <edwin@FreeBSD.org>
---
 libexec/tftpd/Makefile |  13 +-
 libexec/tftpd/tftpd.8  |  17 +-
 libexec/tftpd/tftpd.c  | 834 +++++++++++++++++------------------------
 3 files changed, 358 insertions(+), 506 deletions(-)

diff --git a/libexec/tftpd/Makefile b/libexec/tftpd/Makefile
index 6c29243a666..b9db2a9e0ae 100644
--- a/libexec/tftpd/Makefile
+++ b/libexec/tftpd/Makefile
@@ -2,15 +2,14 @@
 # $FreeBSD$
 
 PROG=	tftpd
-SRCS=	tftpd.c tftpsubs.c
-DPADD=	${LIBUTIL}
-LDADD=	-lutil
-
-WARNS?=	1
+SRCS=	tftpd.c tftp-io.c tftp-utils.c tftp-file.c tftp-transfer.c tftp-options.c
+WARNS=	3
 WFORMAT=0
-
 MAN=	tftpd.8
-CFLAGS+=-I${.CURDIR}/../../usr.bin/tftp
+CFLAGS=-g -Wall
+CFLAGS+=-I${.CURDIR}/../../usr.bin/tftp -I${.CURDIR}/../../libexec/tftpd
 .PATH:	${.CURDIR}/../../usr.bin/tftp
+COPTFLAGS = -O
+LDFLAGS= -lwrap
 
 .include <bsd.prog.mk>
diff --git a/libexec/tftpd/tftpd.8 b/libexec/tftpd/tftpd.8
index f9d7aecc064..08f40f35e11 100644
--- a/libexec/tftpd/tftpd.8
+++ b/libexec/tftpd/tftpd.8
@@ -40,7 +40,7 @@
 .Nd Internet Trivial File Transfer Protocol server
 .Sh SYNOPSIS
 .Nm tftpd
-.Op Fl cClnwW
+.Op Fl cdClnow
 .Op Fl F Ar strftime-format
 .Op Fl s Ar directory
 .Op Fl u Ar user
@@ -150,6 +150,9 @@ compatible format string for the creation of the suffix if
 .Fl W
 is specified.
 By default the string "%Y%m%d" is used.
+.It Fl d
+Enables debug output.
+If specified twice, it will log DATA and ACK packets too.
 .It Fl l
 Log all requests using
 .Xr syslog 3
@@ -164,6 +167,8 @@ must also be enabled in the syslog configuration file,
 .It Fl n
 Suppress negative acknowledgement of requests for nonexistent
 relative filenames.
+.It Fl o
+Disable support for RFC2347 style TFTP Options.
 .It Fl s Ar directory
 Cause
 .Nm
@@ -240,10 +245,16 @@ and the
 and
 .Fl W
 options were introduced in
-.Fx 8.0 .
+.Fx 7 .
 .Pp
+Support for Timeout Interval and Transfer Size Options (RFC2349)
+was introduced in
+.Fx 5.0 ,
+support for the TFTP Blocksize Option (RFC2348) and the blksize2 option
+was introduced in
+.Fx 7 .
 .Sh BUGS
 Files larger than 33488896 octets (65535 blocks) cannot be transferred
-without client and server supporting blocksize negotiation (RFC1783).
+without client and server supporting blocksize negotiation (RFC2348).
 .Pp
 Many tftp clients will not transfer files over 16744448 octets (32767 blocks).
diff --git a/libexec/tftpd/tftpd.c b/libexec/tftpd/tftpd.c
index 027e4bc3b20..93943f5b43d 100644
--- a/libexec/tftpd/tftpd.c
+++ b/libexec/tftpd/tftpd.c
@@ -41,9 +41,9 @@ static const char copyright[] =
 #if 0
 static char sccsid[] = "@(#)tftpd.c	8.1 (Berkeley) 6/4/93";
 #endif
-static const char rcsid[] =
-  "$FreeBSD$";
 #endif /* not lint */
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
 
 /*
  * Trivial file transfer protocol server.
@@ -56,43 +56,30 @@ static const char rcsid[] =
 #include <sys/ioctl.h>
 #include <sys/stat.h>
 #include <sys/socket.h>
-#include <sys/types.h>
-#include <sys/time.h>
 
 #include <netinet/in.h>
 #include <arpa/tftp.h>
-#include <arpa/inet.h>
 
 #include <ctype.h>
 #include <errno.h>
 #include <fcntl.h>
-#include <libutil.h>
 #include <netdb.h>
 #include <pwd.h>
-#include <setjmp.h>
-#include <signal.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include <syslog.h>
+#include <tcpd.h>
 #include <unistd.h>
 
-#include "tftpsubs.h"
+#include "tftp-file.h"
+#include "tftp-io.h"
+#include "tftp-utils.h"
+#include "tftp-transfer.h"
+#include "tftp-options.h"
 
-#define	TIMEOUT		5
-#define	MAX_TIMEOUTS	5
-
-int	peer;
-int	rexmtval = TIMEOUT;
-int	max_rexmtval = 2*TIMEOUT;
-
-#define	PKTSIZE	SEGSIZE+4
-char	buf[PKTSIZE];
-char	ackbuf[PKTSIZE];
-struct	sockaddr_storage from;
-
-void	tftp(struct tftphdr *, int);
-static void unmappedaddr(struct sockaddr_in6 *);
+static void	tftp_wrq(int peer, char *, ssize_t);
+static void	tftp_rrq(int peer, char *, ssize_t);
 
 /*
  * Null-terminated directory prefix list for absolute pathname requests and
@@ -112,31 +99,44 @@ static int	ipchroot;
 static int	create_new = 0;
 static char	*newfile_format = "%Y%m%d";
 static int	increase_name = 0;
-static mode_t	mask = S_IWGRP|S_IWOTH;
+static mode_t	mask = S_IWGRP | S_IWOTH;
 
-static const char *errtomsg(int);
-static void  nak(int);
-static void  oack(void);
+struct formats;
+static void	tftp_recvfile(int peer, const char *mode);
+static void	tftp_xmitfile(int peer, const char *mode);
+static int	validate_access(int peer, char **, int);
+static char	peername[NI_MAXHOST];
 
-static void  timer(int);
-static void  justquit(int);
+FILE *file;
+
+struct formats {
+	const char	*f_mode;
+	int	f_convert;
+} formats[] = {
+	{ "netascii",	1 },
+	{ "octet",	0 },
+	{ NULL,		0 }
+};
 
 int
 main(int argc, char *argv[])
 {
 	struct tftphdr *tp;
-	socklen_t fromlen, len;
-	int n;
-	int ch, on;
-	struct sockaddr_storage me;
-	char *chroot_dir = NULL;
-	struct passwd *nobody;
-	const char *chuser = "nobody";
+	int		peer;
+	socklen_t	peerlen, len;
+	ssize_t		n;
+	int		ch;
+	char		*chroot_dir = NULL;
+	struct passwd	*nobody;
+	const char	*chuser = "nobody";
+	char		recvbuffer[MAXPKTSIZE];
+	int		allow_ro = 1, allow_wo = 1;
 
 	tzset();			/* syslog in localtime */
+	acting_as_client = 0;
 
-	openlog("tftpd", LOG_PID | LOG_NDELAY, LOG_FTP);
-	while ((ch = getopt(argc, argv, "cCF:lns:u:U:wW")) != -1) {
+	tftp_openlog("tftpd", LOG_PID | LOG_NDELAY, LOG_FTP);
+	while ((ch = getopt(argc, argv, "cCd:F:lnoOp:s:u:U:wW")) != -1) {
 		switch (ch) {
 		case 'c':
 			ipchroot = 1;
@@ -144,6 +144,12 @@ main(int argc, char *argv[])
 		case 'C':
 			ipchroot = 2;
 			break;
+		case 'd':
+			if (atoi(optarg) != 0)
+				debug += atoi(optarg);
+			else
+				debug |= debug_finds(optarg);
+			break;
 		case 'F':
 			newfile_format = optarg;
 			break;
@@ -153,6 +159,18 @@ main(int argc, char *argv[])
 		case 'n':
 			suppress_naks = 1;
 			break;
+		case 'o':
+			options_rfc_enabled = 0;
+			break;
+		case 'O':
+			options_extra_enabled = 0;
+			break;
+		case 'p':
+			packetdroppercentage = atoi(optarg);
+			tftp_log(LOG_INFO,
+			    "Randomly dropping %d out of 100 packets",
+			    packetdroppercentage);
+			break;
 		case 's':
 			chroot_dir = optarg;
 			break;
@@ -170,7 +188,8 @@ main(int argc, char *argv[])
 			increase_name = 1;
 			break;
 		default:
-			syslog(LOG_WARNING, "ignoring unknown option -%c", ch);
+			tftp_log(LOG_WARNING,
+				"ignoring unknown option -%c", ch);
 		}
 	}
 	if (optind < argc) {
@@ -191,24 +210,31 @@ main(int argc, char *argv[])
 		dirs->len = 1;
 	}
 	if (ipchroot > 0 && chroot_dir == NULL) {
-		syslog(LOG_ERR, "-c requires -s");
+		tftp_log(LOG_ERR, "-c requires -s");
 		exit(1);
 	}
 
 	umask(mask);
 
-	on = 1;
-	if (ioctl(0, FIONBIO, &on) < 0) {
-		syslog(LOG_ERR, "ioctl(FIONBIO): %m");
-		exit(1);
+	{
+		int on = 1;
+		if (ioctl(0, FIONBIO, &on) < 0) {
+			tftp_log(LOG_ERR, "ioctl(FIONBIO): %s", strerror(errno));
+			exit(1);
+		}
 	}
-	fromlen = sizeof (from);
-	n = recvfrom(0, buf, sizeof (buf), 0,
-	    (struct sockaddr *)&from, &fromlen);
+
+	/* Find out who we are talking to and what we are going to do */
+	peerlen = sizeof(peer_sock);
+	n = recvfrom(0, recvbuffer, MAXPKTSIZE, 0,
+	    (struct sockaddr *)&peer_sock, &peerlen);
 	if (n < 0) {
-		syslog(LOG_ERR, "recvfrom: %m");
+		tftp_log(LOG_ERR, "recvfrom: %s", strerror(errno));
 		exit(1);
 	}
+	getnameinfo((struct sockaddr *)&peer_sock, peer_sock.ss_len,
+	    peername, sizeof(peername), NULL, 0, NI_NUMERICHOST);
+
 	/*
 	 * Now that we have read the message out of the UDP
 	 * socket, we fork and exit.  Thus, inetd will go back
@@ -240,9 +266,9 @@ main(int argc, char *argv[])
 				 * than one tftpd being started up to service
 				 * a single request from a single client.
 				 */
-				fromlen = sizeof from;
-				i = recvfrom(0, buf, sizeof (buf), 0,
-				    (struct sockaddr *)&from, &fromlen);
+				peerlen = sizeof peer_sock;
+				i = recvfrom(0, recvbuffer, MAXPKTSIZE, 0,
+				    (struct sockaddr *)&peer_sock, &peerlen);
 				if (i > 0) {
 					n = i;
 				}
@@ -251,13 +277,62 @@ main(int argc, char *argv[])
 		    }
 		}
 		if (pid < 0) {
-			syslog(LOG_ERR, "fork: %m");
+			tftp_log(LOG_ERR, "fork: %s", strerror(errno));
 			exit(1);
 		} else if (pid != 0) {
 			exit(0);
 		}
 	}
 
+	/*
+	 * See if the client is allowed to talk to me.
+	 * (This needs to be done before the chroot())
+	 */
+	{
+		struct request_info req;
+
+		request_init(&req, RQ_CLIENT_ADDR, peername, 0);
+		request_set(&req, RQ_DAEMON, "tftpd", 0);
+
+		if (hosts_access(&req) == 0) {
+			if (debug&DEBUG_ACCESS)
+				tftp_log(LOG_WARNING,
+				    "Access denied by 'tftpd' entry "
+				    "in /etc/hosts.allow");
+
+			/*
+			 * Full access might be disabled, but maybe the
+			 * client is allowed to do read-only access.
+			 */
+			request_set(&req, RQ_DAEMON, "tftpd-ro", 0);
+			allow_ro = hosts_access(&req);
+
+			request_set(&req, RQ_DAEMON, "tftpd-wo", 0);
+			allow_wo = hosts_access(&req);
+
+			if (allow_ro == 0 && allow_wo == 0) {
+				tftp_log(LOG_WARNING,
+				    "Unauthorized access from %s", peername);
+				exit(1);
+			}
+
+			if (debug&DEBUG_ACCESS) {
+				if (allow_ro)
+					tftp_log(LOG_WARNING,
+					    "But allowed readonly access "
+					    "via 'tftpd-ro' entry");
+				if (allow_wo)
+					tftp_log(LOG_WARNING,
+					    "But allowed writeonly access "
+					    "via 'tftpd-wo' entry");
+			}
+		} else
+			if (debug&DEBUG_ACCESS)
+				tftp_log(LOG_WARNING,
+				    "Full access allowed"
+				    "in /etc/hosts.allow");
+	}
+
 	/*
 	 * Since we exit here, we should do that only after the above
 	 * recvfrom to keep inetd from constantly forking should there
@@ -271,7 +346,8 @@ main(int argc, char *argv[])
 			struct sockaddr_storage ss;
 			char hbuf[NI_MAXHOST];
 
-			memcpy(&ss, &from, from.ss_len);
+			statret = -1;
+			memcpy(&ss, &peer_sock, peer_sock.ss_len);
 			unmappedaddr((struct sockaddr_in6 *)&ss);
 			getnameinfo((struct sockaddr *)&ss, ss.ss_len,
 				    hbuf, sizeof(hbuf), NULL, 0,
@@ -285,11 +361,12 @@ main(int argc, char *argv[])
 		}
 		/* Must get this before chroot because /etc might go away */
 		if ((nobody = getpwnam(chuser)) == NULL) {
-			syslog(LOG_ERR, "%s: no such user", chuser);
+			tftp_log(LOG_ERR, "%s: no such user", chuser);
 			exit(1);
 		}
 		if (chroot(chroot_dir)) {
-			syslog(LOG_ERR, "chroot: %s: %m", chroot_dir);
+			tftp_log(LOG_ERR, "chroot: %s: %s",
+			    chroot_dir, strerror(errno));
 			exit(1);
 		}
 		chdir("/");
@@ -297,44 +374,56 @@ main(int argc, char *argv[])
 		setuid(nobody->pw_uid);
 	}
 
-	len = sizeof(me);
-	if (getsockname(0, (struct sockaddr *)&me, &len) == 0) {
-		switch (me.ss_family) {
+	len = sizeof(me_sock);
+	if (getsockname(0, (struct sockaddr *)&me_sock, &len) == 0) {
+		switch (me_sock.ss_family) {
 		case AF_INET:
-			((struct sockaddr_in *)&me)->sin_port = 0;
+			((struct sockaddr_in *)&me_sock)->sin_port = 0;
 			break;
 		case AF_INET6:
-			((struct sockaddr_in6 *)&me)->sin6_port = 0;
+			((struct sockaddr_in6 *)&me_sock)->sin6_port = 0;
 			break;
 		default:
 			/* unsupported */
 			break;
 		}
 	} else {
-		memset(&me, 0, sizeof(me));
-		me.ss_family = from.ss_family;
-		me.ss_len = from.ss_len;
+		memset(&me_sock, 0, sizeof(me_sock));
+		me_sock.ss_family = peer_sock.ss_family;
+		me_sock.ss_len = peer_sock.ss_len;
 	}
-	alarm(0);
 	close(0);
 	close(1);
-	peer = socket(from.ss_family, SOCK_DGRAM, 0);
+	peer = socket(peer_sock.ss_family, SOCK_DGRAM, 0);
 	if (peer < 0) {
-		syslog(LOG_ERR, "socket: %m");
+		tftp_log(LOG_ERR, "socket: %s", strerror(errno));
 		exit(1);
 	}
-	if (bind(peer, (struct sockaddr *)&me, me.ss_len) < 0) {
-		syslog(LOG_ERR, "bind: %m");
+	if (bind(peer, (struct sockaddr *)&me_sock, me_sock.ss_len) < 0) {
+		tftp_log(LOG_ERR, "bind: %s", strerror(errno));
 		exit(1);
 	}
-	if (connect(peer, (struct sockaddr *)&from, from.ss_len) < 0) {
-		syslog(LOG_ERR, "connect: %m");
-		exit(1);
-	}
-	tp = (struct tftphdr *)buf;
+
+	tp = (struct tftphdr *)recvbuffer;
 	tp->th_opcode = ntohs(tp->th_opcode);
-	if (tp->th_opcode == RRQ || tp->th_opcode == WRQ)
-		tftp(tp, n);
+	if (tp->th_opcode == RRQ) {
+		if (allow_ro)
+			tftp_rrq(peer, tp->th_stuff, n - 1);
+		else {
+			tftp_log(LOG_WARNING,
+			    "%s read access denied", peername);
+			exit(1);
+		}
+	}
+	if (tp->th_opcode == WRQ) {
+		if (allow_wo)
+			tftp_wrq(peer, tp->th_stuff, n - 1);
+		else {
+			tftp_log(LOG_WARNING,
+			    "%s write access denied", peername);
+			exit(1);
+		}
+	}
 	exit(1);
 }
 
@@ -369,138 +458,145 @@ reduce_path(char *fn)
 	}
 }
 
-struct formats;
-int	validate_access(char **, int);
-void	xmitfile(struct formats *);
-void	recvfile(struct formats *);
-
-struct formats {
-	const char	*f_mode;
-	int	(*f_validate)(char **, int);
-	void	(*f_send)(struct formats *);
-	void	(*f_recv)(struct formats *);
-	int	f_convert;
-} formats[] = {
-	{ "netascii",	validate_access,	xmitfile,	recvfile, 1 },
-	{ "octet",	validate_access,	xmitfile,	recvfile, 0 },
-#ifdef notdef
-	{ "mail",	validate_user,		sendmail,	recvmail, 1 },
-#endif
-	{ 0,		NULL,			NULL,		NULL,	  0 }
-};
-
-struct options {
-	const char	*o_type;
-	char	*o_request;
-	int	o_reply;	/* turn into union if need be */
-} options[] = {
-	{ "tsize",	NULL, 0 },		/* OPT_TSIZE */
-	{ "timeout",	NULL, 0 },		/* OPT_TIMEOUT */
-	{ NULL,		NULL, 0 }
-};
-
-enum opt_enum {
-	OPT_TSIZE = 0,
-	OPT_TIMEOUT,
-};
-
-/*
- * Handle initial connection protocol.
- */
-void
-tftp(struct tftphdr *tp, int size)
+static char *
+parse_header(int peer, char *recvbuffer, ssize_t size,
+	char **filename, char **mode)
 {
-	char *cp;
-	int i, first = 1, has_options = 0, ecode;
+	char	*cp;
+	int	i;
 	struct formats *pf;
-	char *filename, *mode, *option, *ccp;
-	char fnbuf[PATH_MAX];
 
-	cp = tp->th_stuff;
-again:
-	while (cp < buf + size) {
-		if (*cp == '\0')
-			break;
-		cp++;
-	}
-	if (*cp != '\0') {
-		nak(EBADOP);
+	*mode = NULL;
+	cp = recvbuffer;
+
+	i = get_field(peer, recvbuffer, size);
+	if (i >= PATH_MAX) {
+		tftp_log(LOG_ERR, "Bad option - filename too long");
+		send_error(peer, EBADOP);
 		exit(1);
 	}
-	i = cp - tp->th_stuff;
-	if (i >= sizeof(fnbuf)) {
-		nak(EBADOP);
-		exit(1);
-	}
-	memcpy(fnbuf, tp->th_stuff, i);
-	fnbuf[i] = '\0';
-	reduce_path(fnbuf);
-	filename = fnbuf;
-	if (first) {
-		mode = ++cp;
-		first = 0;
-		goto again;
-	}
-	for (cp = mode; *cp; cp++)
+	*filename = recvbuffer;
+	tftp_log(LOG_INFO, "Filename: '%s'", *filename);
+	cp += i;
+
+	i = get_field(peer, cp, size);
+	*mode = cp;
+	cp += i;
+
+	/* Find the file transfer mode */
+	for (cp = *mode; *cp; cp++)
 		if (isupper(*cp))
 			*cp = tolower(*cp);
 	for (pf = formats; pf->f_mode; pf++)
-		if (strcmp(pf->f_mode, mode) == 0)
+		if (strcmp(pf->f_mode, *mode) == 0)
 			break;
-	if (pf->f_mode == 0) {
-		nak(EBADOP);
+	if (pf->f_mode == NULL) {
+		tftp_log(LOG_ERR,
+		    "Bad option - Unknown transfer mode (%s)", *mode);
+		send_error(peer, EBADOP);
 		exit(1);
 	}
-	while (++cp < buf + size) {
-		for (i = 2, ccp = cp; i > 0; ccp++) {
-			if (ccp >= buf + size) {
-				/*
-				 * Don't reject the request, just stop trying
-				 * to parse the option and get on with it.
-				 * Some Apple Open Firmware versions have
-				 * trailing garbage on the end of otherwise
-				 * valid requests.
-				 */
-				goto option_fail;
-			} else if (*ccp == '\0')
-				i--;
-		}
-		for (option = cp; *cp; cp++)
-			if (isupper(*cp))
-				*cp = tolower(*cp);
-		for (i = 0; options[i].o_type != NULL; i++)
-			if (strcmp(option, options[i].o_type) == 0) {
-				options[i].o_request = ++cp;
-				has_options = 1;
-			}
-		cp = ccp-1;
-	}
+	tftp_log(LOG_INFO, "Mode: '%s'", *mode);
 
-option_fail:
-	if (options[OPT_TIMEOUT].o_request) {
-		int to = atoi(options[OPT_TIMEOUT].o_request);
-		if (to < 1 || to > 255) {
-			nak(EBADOP);
-			exit(1);
-		}
-		else if (to <= max_rexmtval)
-			options[OPT_TIMEOUT].o_reply = rexmtval = to;
+	return (cp + 1);
+}
+
+/*
+ * WRQ - receive a file from the client
+ */
+void
+tftp_wrq(int peer, char *recvbuffer, ssize_t size)
+{
+	char *cp;
+	int has_options = 0, ecode;
+	char *filename, *mode;
+	char fnbuf[PATH_MAX];
+
+	cp = parse_header(peer, recvbuffer, size, &filename, &mode);
+	size -= (cp - recvbuffer) + 1;
+
+	strcpy(fnbuf, filename);
+	reduce_path(fnbuf);
+	filename = fnbuf;
+
+	if (size > 0) {
+		if (options_rfc_enabled)
+			has_options = !parse_options(peer, cp, size);
 		else
-			options[OPT_TIMEOUT].o_request = NULL;
+			tftp_log(LOG_INFO, "Options found but not enabled");
 	}
 
-	ecode = (*pf->f_validate)(&filename, tp->th_opcode);
-	if (has_options && ecode == 0)
-		oack();
+	ecode = validate_access(peer, &filename, WRQ);
+	if (ecode == 0) {
+		if (has_options)
+			send_oack(peer);
+		else
+			send_ack(peer, 0);
+	}
 	if (logging) {
-		char hbuf[NI_MAXHOST];
-
-		getnameinfo((struct sockaddr *)&from, from.ss_len,
-			    hbuf, sizeof(hbuf), NULL, 0, 0);
-		syslog(LOG_INFO, "%s: %s request for %s: %s", hbuf,
-			tp->th_opcode == WRQ ? "write" : "read",
-			filename, errtomsg(ecode));
+		tftp_log(LOG_INFO, "%s: write request for %s: %s", peername,
+			    filename, errtomsg(ecode));
 	}
+
+	tftp_recvfile(peer, mode);
+	exit(0);
+}
+
+/*
+ * RRQ - send a file to the client
+ */
+void
+tftp_rrq(int peer, char *recvbuffer, ssize_t size)
+{
+	char *cp;
+	int has_options = 0, ecode;
+	char *filename, *mode;
+	char	fnbuf[PATH_MAX];
+
+	cp = parse_header(peer, recvbuffer, size, &filename, &mode);
+	size -= (cp - recvbuffer) + 1;
+
+	strcpy(fnbuf, filename);
+	reduce_path(fnbuf);
+	filename = fnbuf;
+
+	if (size > 0) {
+		if (options_rfc_enabled)
+			has_options = !parse_options(peer, cp, size);
+		else
+			tftp_log(LOG_INFO, "Options found but not enabled");
+	}
+
+	ecode = validate_access(peer, &filename, RRQ);
+	if (ecode == 0) {
+		if (has_options) {
+			int n;
+			char lrecvbuffer[MAXPKTSIZE];
+			struct tftphdr *rp = (struct tftphdr *)lrecvbuffer;
+
+			send_oack(peer);
+			n = receive_packet(peer, lrecvbuffer, MAXPKTSIZE,
+				NULL, timeoutpacket);
+			if (n < 0) {
+				if (debug&DEBUG_SIMPLE)
+					tftp_log(LOG_DEBUG, "Aborting: %s",
+					    rp_strerror(n));
+				return;
+			}
+			if (rp->th_opcode != ACK) {
+				if (debug&DEBUG_SIMPLE)
+					tftp_log(LOG_DEBUG,
+					    "Expected ACK, got %s on OACK",
+					    packettype(rp->th_opcode));
+				return;
+			}
+		}
+	}
+
+	if (logging)
+		tftp_log(LOG_INFO, "%s: read request for %s: %s", peername,
+			    filename, errtomsg(ecode));
+
 	if (ecode) {
 		/*
 		 * Avoid storms of naks to a RRQ broadcast for a relative
@@ -508,19 +604,13 @@ option_fail:
 		 */
 		if (suppress_naks && *filename != '/' && ecode == ENOTFOUND)
 			exit(0);
-		nak(ecode);
+		tftp_log(LOG_ERR, "Prevent NAK storm");
+		send_error(peer, ecode);
 		exit(1);
 	}
-	if (tp->th_opcode == WRQ)
-		(*pf->f_recv)(pf);
-	else
-		(*pf->f_send)(pf);
-	exit(0);
+	tftp_xmitfile(peer, mode);
 }
 
-
-FILE *file;
-
 /*
  * Find the next value for YYYYMMDD.nn when the file to be written should
  * be unique. Due to the limitations of nn, we will fail if nn reaches 100.
@@ -536,8 +626,6 @@ find_next_name(char *filename, int *fd)
 	struct tm lt;
 	char yyyymmdd[MAXPATHLEN];
 	char newname[MAXPATHLEN];
-	struct stat sb;
-	int ret;
 
 	/* Create the YYYYMMDD part of the filename */
 	time(&tval);
@@ -553,7 +641,7 @@ find_next_name(char *filename, int *fd)
 	/* Make sure the new filename is not too long */
 	if (strlen(filename) > MAXPATHLEN - len - 5) {
 		syslog(LOG_WARNING,
-			"Filename too long (%d characters, %d maximum)",
+			"Filename too long (%zd characters, %zd maximum)",
 			strlen(filename), MAXPATHLEN - len - 5);
 		return (EACCESS);
 	}
@@ -584,7 +672,7 @@ find_next_name(char *filename, int *fd)
  * given as we have no login directory.
  */
 int
-validate_access(char **filep, int mode)
+validate_access(int peer, char **filep, int mode)
 {
 	struct stat stbuf;
 	int	fd;
@@ -660,14 +748,13 @@ validate_access(char **filep, int mode)
 		else if (mode == RRQ)
 			return (err);
 	}
-	if (options[OPT_TSIZE].o_request) {
-		if (mode == RRQ) 
-			options[OPT_TSIZE].o_reply = stbuf.st_size;
-		else
-			/* XXX Allows writes of all sizes. */
-			options[OPT_TSIZE].o_reply =
-				atoi(options[OPT_TSIZE].o_request);
-	}
+
+	/*
+	 * This option is handled here because it (might) require(s) the
+	 * size of the file.
+	 */
+	option_tsize(peer, NULL, mode, &stbuf);
+
 	if (mode == RRQ)
 		fd = open(filename, O_RDONLY);
 	else {
@@ -694,305 +781,60 @@ validate_access(char **filep, int mode)
 	return (0);
 }
 
-int	timeouts;
-jmp_buf	timeoutbuf;
-
-void
-timer(int sig __unused)
+static void
+tftp_xmitfile(int peer, const char *mode)
 {
-	if (++timeouts > MAX_TIMEOUTS)
-		exit(1);
-	longjmp(timeoutbuf, 1);
-}
+	uint16_t block;
+	uint32_t amount;
+	time_t now;
+	struct tftp_stats ts;
 
-/*
- * Send the requested file.
- */
-void
-xmitfile(struct formats *pf)
-{
-	struct tftphdr *dp;
-	struct tftphdr *ap;    /* ack packet */
-	int size, n;
-	volatile unsigned short block;
+	now = time(NULL);
+	if (debug&DEBUG_SIMPLE)
+		tftp_log(LOG_DEBUG, "Transmitting file");
 
-	signal(SIGALRM, timer);
-	dp = r_init();
-	ap = (struct tftphdr *)ackbuf;
+	read_init(0, file, mode);
 	block = 1;
-	do {
-		size = readit(file, &dp, pf->f_convert);
-		if (size < 0) {
-			nak(errno + 100);
-			goto abort;
-		}
-		dp->th_opcode = htons((u_short)DATA);
-		dp->th_block = htons((u_short)block);
-		timeouts = 0;
-		(void)setjmp(timeoutbuf);
-
-send_data:
-		{
-			int i, t = 1;
-			for (i = 0; ; i++){
-				if (send(peer, dp, size + 4, 0) != size + 4) {
-					sleep(t);
-					t = (t < 32) ? t<< 1 : t;
-					if (i >= 12) {
-						syslog(LOG_ERR, "write: %m");
-						goto abort;
-					}
-				}
-				break;
-			}
-		}
-		read_ahead(file, pf->f_convert);
-		for ( ; ; ) {
-			alarm(rexmtval);        /* read the ack */
-			n = recv(peer, ackbuf, sizeof (ackbuf), 0);
-			alarm(0);
-			if (n < 0) {
-				syslog(LOG_ERR, "read: %m");
-				goto abort;
-			}
-			ap->th_opcode = ntohs((u_short)ap->th_opcode);
-			ap->th_block = ntohs((u_short)ap->th_block);
-
-			if (ap->th_opcode == ERROR)
-				goto abort;
-
-			if (ap->th_opcode == ACK) {
-				if (ap->th_block == block)
-					break;
-				/* Re-synchronize with the other side */
-				(void) synchnet(peer);
-				if (ap->th_block == (block -1))
-					goto send_data;
-			}
-
-		}
-		block++;
-	} while (size == SEGSIZE);
-abort:
-	(void) fclose(file);
+	tftp_send(peer, &block, &ts);
+	read_close();
+	if (debug&DEBUG_SIMPLE)
+		tftp_log(LOG_INFO, "Sent %d bytes in %d seconds",
+		    amount, time(NULL) - now);
 }
 
-void
-justquit(int sig __unused)
+static void
+tftp_recvfile(int peer, const char *mode)
 {
-	exit(0);
-}
+	uint32_t filesize; 
+	uint16_t block;
+	struct timeval now1, now2;
+	struct tftp_stats ts;
 
+	gettimeofday(&now1, NULL);
+	if (debug&DEBUG_SIMPLE)
+		tftp_log(LOG_DEBUG, "Receiving file");
 
-/*
- * Receive a file.
- */
-void
-recvfile(struct formats *pf)
-{
-	struct tftphdr *dp;
-	struct tftphdr *ap;    /* ack buffer */
-	int n, size;
-	volatile unsigned short block;
+	write_init(0, file, mode);
 
-	signal(SIGALRM, timer);
-	dp = w_init();
-	ap = (struct tftphdr *)ackbuf;
 	block = 0;
-	do {
-		timeouts = 0;
-		ap->th_opcode = htons((u_short)ACK);
-		ap->th_block = htons((u_short)block);
-		block++;
-		(void) setjmp(timeoutbuf);
-send_ack:
-		if (send(peer, ackbuf, 4, 0) != 4) {
-			syslog(LOG_ERR, "write: %m");
-			goto abort;
-		}
-		write_behind(file, pf->f_convert);
-		for ( ; ; ) {
-			alarm(rexmtval);
-			n = recv(peer, dp, PKTSIZE, 0);
-			alarm(0);
-			if (n < 0) {            /* really? */
-				syslog(LOG_ERR, "read: %m");
-				goto abort;
-			}
-			dp->th_opcode = ntohs((u_short)dp->th_opcode);
-			dp->th_block = ntohs((u_short)dp->th_block);
-			if (dp->th_opcode == ERROR)
-				goto abort;
-			if (dp->th_opcode == DATA) {
-				if (dp->th_block == block) {
-					break;   /* normal */
-				}
-				/* Re-synchronize with the other side */
-				(void) synchnet(peer);
-				if (dp->th_block == (block-1))
-					goto send_ack;          /* rexmit */
-			}
-		}
-		/*  size = write(file, dp->th_data, n - 4); */
-		size = writeit(file, &dp, n - 4, pf->f_convert);
-		if (size != (n-4)) {                    /* ahem */
-			if (size < 0) nak(errno + 100);
-			else nak(ENOSPACE);
-			goto abort;
-		}
-	} while (size == SEGSIZE);
-	write_behind(file, pf->f_convert);
-	(void) fclose(file);            /* close data file */
+	tftp_receive(peer, &block, &ts, NULL, 0);
 
-	ap->th_opcode = htons((u_short)ACK);    /* send the "final" ack */
-	ap->th_block = htons((u_short)(block));
-	(void) send(peer, ackbuf, 4, 0);
+	write_close();
 
-	signal(SIGALRM, justquit);      /* just quit on timeout */
-	alarm(rexmtval);
-	n = recv(peer, buf, sizeof (buf), 0); /* normally times out and quits */
-	alarm(0);
-	if (n >= 4 &&                   /* if read some data */
-	    dp->th_opcode == DATA &&    /* and got a data block */
-	    block == dp->th_block) {	/* then my last ack was lost */
-		(void) send(peer, ackbuf, 4, 0);     /* resend final ack */
+	if (debug&DEBUG_SIMPLE) {
+		double f;
+		if (now1.tv_usec > now2.tv_usec) {
+			now2.tv_usec += 1000000;
+			now2.tv_sec--;
+		}
+
+		f = now2.tv_sec - now1.tv_sec +
+		    (now2.tv_usec - now1.tv_usec) / 100000.0;
+		tftp_log(LOG_INFO,
+		    "Download of %d bytes in %d blocks completed after %0.1f seconds\n",
+		    filesize, block, f);
 	}
-abort:
+
 	return;
 }
 
-struct errmsg {
-	int	e_code;
-	const char	*e_msg;
-} errmsgs[] = {
-	{ EUNDEF,	"Undefined error code" },
-	{ ENOTFOUND,	"File not found" },
-	{ EACCESS,	"Access violation" },
-	{ ENOSPACE,	"Disk full or allocation exceeded" },
-	{ EBADOP,	"Illegal TFTP operation" },
-	{ EBADID,	"Unknown transfer ID" },
-	{ EEXISTS,	"File already exists" },
-	{ ENOUSER,	"No such user" },
-	{ EOPTNEG,	"Option negotiation" },
-	{ -1,		0 }
-};
-
-static const char *
-errtomsg(int error)
-{
-	static char ebuf[20];
-	struct errmsg *pe;
-	if (error == 0)
-		return "success";
-	for (pe = errmsgs; pe->e_code >= 0; pe++)
-		if (pe->e_code == error)
-			return pe->e_msg;
-	snprintf(ebuf, sizeof(buf), "error %d", error);
-	return ebuf;
-}
-
-/*
- * Send a nak packet (error message).
- * Error code passed in is one of the
- * standard TFTP codes, or a UNIX errno
- * offset by 100.
- */
-static void
-nak(int error)
-{
-	struct tftphdr *tp;
-	int length;
-	struct errmsg *pe;
-
-	tp = (struct tftphdr *)buf;
-	tp->th_opcode = htons((u_short)ERROR);
-	tp->th_code = htons((u_short)error);
-	for (pe = errmsgs; pe->e_code >= 0; pe++)
-		if (pe->e_code == error)
-			break;
-	if (pe->e_code < 0) {
-		pe->e_msg = strerror(error - 100);
-		tp->th_code = EUNDEF;   /* set 'undef' errorcode */
-	}
-	strcpy(tp->th_msg, pe->e_msg);
-	length = strlen(pe->e_msg);
-	tp->th_msg[length] = '\0';
-	length += 5;
-	if (send(peer, buf, length, 0) != length)
-		syslog(LOG_ERR, "nak: %m");
-}
-
-/* translate IPv4 mapped IPv6 address to IPv4 address */
-static void
-unmappedaddr(struct sockaddr_in6 *sin6)
-{
-	struct sockaddr_in *sin4;
-	u_int32_t addr;
-	int port;
-
-	if (sin6->sin6_family != AF_INET6 ||
-	    !IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr))
-		return;
-	sin4 = (struct sockaddr_in *)sin6;
-	addr = *(u_int32_t *)&sin6->sin6_addr.s6_addr[12];
-	port = sin6->sin6_port;
-	memset(sin4, 0, sizeof(struct sockaddr_in));
-	sin4->sin_addr.s_addr = addr;
-	sin4->sin_port = port;
-	sin4->sin_family = AF_INET;
-	sin4->sin_len = sizeof(struct sockaddr_in);
-}
-
-/*
- * Send an oack packet (option acknowledgement).
- */
-static void
-oack(void)
-{
-	struct tftphdr *tp, *ap;
-	int size, i, n;
-	char *bp;
-
-	tp = (struct tftphdr *)buf;
-	bp = buf + 2;
-	size = sizeof(buf) - 2;
-	tp->th_opcode = htons((u_short)OACK);
-	for (i = 0; options[i].o_type != NULL; i++) {
-		if (options[i].o_request) {
-			n = snprintf(bp, size, "%s%c%d", options[i].o_type,
-				     0, options[i].o_reply);
-			bp += n+1;
-			size -= n+1;
-			if (size < 0) {
-				syslog(LOG_ERR, "oack: buffer overflow");
-				exit(1);
-			}
-		}
-	}
-	size = bp - buf;
-	ap = (struct tftphdr *)ackbuf;
-	signal(SIGALRM, timer);
-	timeouts = 0;
-
-	(void)setjmp(timeoutbuf);
-	if (send(peer, buf, size, 0) != size) {
-		syslog(LOG_INFO, "oack: %m");
-		exit(1);
-	}
-
-	for (;;) {
-		alarm(rexmtval);
-		n = recv(peer, ackbuf, sizeof (ackbuf), 0);
-		alarm(0);
-		if (n < 0) {
-			syslog(LOG_ERR, "recv: %m");
-			exit(1);
-		}
-		ap->th_opcode = ntohs((u_short)ap->th_opcode);
-		ap->th_block = ntohs((u_short)ap->th_block);
-		if (ap->th_opcode == ERROR)
-			exit(1);
-		if (ap->th_opcode == ACK && ap->th_block == 0)
-			break;
-	}
-}

From 64c68f1c50be89a011a05886728960002b94f69b Mon Sep 17 00:00:00 2001
From: Kevin Lo <kevlo@FreeBSD.org>
Date: Tue, 4 May 2010 10:14:05 +0000
Subject: [PATCH 444/532] Add support for FA626TE. Tested on GM8181 development
 board.

---
 sys/arm/arm/cpufunc.c           | 99 ++++++++++++++++-----------------
 sys/arm/arm/cpufunc_asm_fa526.S |  4 ++
 sys/arm/arm/elf_trampoline.c    |  2 +-
 sys/arm/arm/identcpu.c          |  4 +-
 sys/arm/include/cpuconf.h       |  6 +-
 sys/arm/include/cpufunc.h       | 12 ++--
 sys/conf/options.arm            |  1 +
 7 files changed, 67 insertions(+), 61 deletions(-)

diff --git a/sys/arm/arm/cpufunc.c b/sys/arm/arm/cpufunc.c
index 2f4c3a1f58a..09c20264819 100644
--- a/sys/arm/arm/cpufunc.c
+++ b/sys/arm/arm/cpufunc.c
@@ -783,69 +783,66 @@ struct cpu_functions xscalec3_cpufuncs = {
 #endif /* CPU_XSCALE_81342 */
 
 
-#if defined(CPU_FA526)
+#if defined(CPU_FA526) || defined(CPU_FA626TE)
 struct cpu_functions fa526_cpufuncs = {
 	/* CPU functions */
 
-	.cf_id			= cpufunc_id,
-	.cf_cpwait		= cpufunc_nullop,
+	cpufunc_id,			/* id			*/
+	cpufunc_nullop,			/* cpwait		*/
 
 	/* MMU functions */
 
-	.cf_control		= cpufunc_control,
-	.cf_domains		= cpufunc_domains,
-	.cf_setttb		= fa526_setttb,
-	.cf_faultstatus		= cpufunc_faultstatus,
-	.cf_faultaddress	= cpufunc_faultaddress,
+	cpufunc_control,		/* control		*/
+	cpufunc_domains,		/* domain		*/
+	fa526_setttb,			/* setttb		*/
+	cpufunc_faultstatus,		/* faultstatus		*/
+	cpufunc_faultaddress,		/* faultaddress		*/
 
 	/* TLB functions */
 
-	.cf_tlb_flushID		= armv4_tlb_flushID,
-	.cf_tlb_flushID_SE	= fa526_tlb_flushID_SE,
-	.cf_tlb_flushI		= armv4_tlb_flushI,
-	.cf_tlb_flushI_SE	= fa526_tlb_flushI_SE,
-	.cf_tlb_flushD		= armv4_tlb_flushD,
-	.cf_tlb_flushD_SE	= armv4_tlb_flushD_SE,
+	armv4_tlb_flushID,		/* tlb_flushID		*/
+	fa526_tlb_flushID_SE,		/* tlb_flushID_SE	*/
+	armv4_tlb_flushI,		/* tlb_flushI		*/
+	fa526_tlb_flushI_SE,		/* tlb_flushI_SE	*/
+	armv4_tlb_flushD,		/* tlb_flushD		*/
+	armv4_tlb_flushD_SE,		/* tlb_flushD_SE	*/
 
 	/* Cache operations */
 
-	.cf_icache_sync_all	= fa526_icache_sync_all,
-	.cf_icache_sync_range	= fa526_icache_sync_range,
+	fa526_icache_sync_all,		/* icache_sync_all	*/
+	fa526_icache_sync_range,	/* icache_sync_range	*/
 
-	.cf_dcache_wbinv_all	= fa526_dcache_wbinv_all,
-	.cf_dcache_wbinv_range	= fa526_dcache_wbinv_range,
-	.cf_dcache_inv_range	= fa526_dcache_inv_range,
-	.cf_dcache_wb_range	= fa526_dcache_wb_range,
-
-	.cf_idcache_wbinv_all	= fa526_idcache_wbinv_all,
-	.cf_idcache_wbinv_range	= fa526_idcache_wbinv_range,
-
-
-	.cf_l2cache_wbinv_all = cpufunc_nullop,
-	.cf_l2cache_wbinv_range = (void *)cpufunc_nullop,
-	.cf_l2cache_inv_range = (void *)cpufunc_nullop,
-	.cf_l2cache_wb_range = (void *)cpufunc_nullop,
+	fa526_dcache_wbinv_all,		/* dcache_wbinv_all	*/
+	fa526_dcache_wbinv_range,	/* dcache_wbinv_range	*/
+	fa526_dcache_inv_range,		/* dcache_inv_range	*/
+	fa526_dcache_wb_range,		/* dcache_wb_range	*/
 
+	fa526_idcache_wbinv_all,	/* idcache_wbinv_all	*/
+	fa526_idcache_wbinv_range,	/* idcache_wbinv_range	*/
+	cpufunc_nullop,			/* l2cache_wbinv_all	*/
+	(void *)cpufunc_nullop,		/* l2cache_wbinv_range	*/
+	(void *)cpufunc_nullop,		/* l2cache_inv_range	*/
+	(void *)cpufunc_nullop,		/* l2cache_wb_range	*/
 
 	/* Other functions */
 
-	.cf_flush_prefetchbuf	= fa526_flush_prefetchbuf,
-	.cf_drain_writebuf	= armv4_drain_writebuf,
-	.cf_flush_brnchtgt_C	= cpufunc_nullop,
-	.cf_flush_brnchtgt_E	= fa526_flush_brnchtgt_E,
+	fa526_flush_prefetchbuf,	/* flush_prefetchbuf	*/
+	armv4_drain_writebuf,		/* drain_writebuf	*/
+	cpufunc_nullop,			/* flush_brnchtgt_C	*/
+	fa526_flush_brnchtgt_E,		/* flush_brnchtgt_E	*/
 
-	.cf_sleep		= fa526_cpu_sleep,
+	fa526_cpu_sleep,		/* sleep		*/
 
 	/* Soft functions */
 
-	.cf_dataabt_fixup	= cpufunc_null_fixup,
-	.cf_prefetchabt_fixup	= cpufunc_null_fixup,
+	cpufunc_null_fixup,		/* dataabt_fixup	*/
+	cpufunc_null_fixup,		/* prefetchabt_fixup	*/
 
-	.cf_context_switch	= fa526_context_switch,
+	fa526_context_switch,		/* context_switch	*/
 
-	.cf_setup		= fa526_setup
-};
-#endif	/* CPU_FA526 */
+	fa526_setup			/* cpu setup 		*/
+};          
+#endif	/* CPU_FA526 || CPU_FA626TE */
 
 
 /*
@@ -856,11 +853,11 @@ struct cpu_functions cpufuncs;
 u_int cputype;
 u_int cpu_reset_needs_v4_MMU_disable;	/* flag used in locore.s */
 
-#if defined(CPU_ARM7TDMI) || defined(CPU_ARM8) || defined(CPU_ARM9) || \
-  defined (CPU_ARM9E) || defined (CPU_ARM10) ||			       \
-  defined(CPU_XSCALE_80200) || defined(CPU_XSCALE_80321) ||	       \
-  defined(CPU_XSCALE_PXA2X0) || defined(CPU_XSCALE_IXP425) ||	       \
-  defined(CPU_FA526) ||					       \
+#if defined(CPU_ARM7TDMI) || defined(CPU_ARM8) || defined(CPU_ARM9) ||	\
+  defined (CPU_ARM9E) || defined (CPU_ARM10) ||				\
+  defined(CPU_XSCALE_80200) || defined(CPU_XSCALE_80321) ||		\
+  defined(CPU_XSCALE_PXA2X0) || defined(CPU_XSCALE_IXP425) ||		\
+  defined(CPU_FA526) || defined(CPU_FA626TE) ||				\
   defined(CPU_XSCALE_80219) || defined(CPU_XSCALE_81342)
 
 static void get_cachetype_cp15(void);
@@ -1141,8 +1138,8 @@ set_cpufuncs()
 		goto out;
 	}
 #endif	/* CPU_SA1110 */
-#ifdef CPU_FA526
-	if (cputype == CPU_ID_FA526) {
+#if defined(CPU_FA526) || defined(CPU_FA626TE)
+	if (cputype == CPU_ID_FA526 || cputype == CPU_ID_FA626TE) {
 		cpufuncs = fa526_cpufuncs;
 		cpu_reset_needs_v4_MMU_disable = 1;	/* SA needs it	*/
 		get_cachetype_cp15();
@@ -1153,7 +1150,7 @@ set_cpufuncs()
 
 		goto out;
 	}
-#endif	/* CPU_FA526 */
+#endif	/* CPU_FA526 || CPU_FA626TE */
 #ifdef CPU_IXP12X0
         if (cputype == CPU_ID_IXP1200) {
                 cpufuncs = ixp12x0_cpufuncs;
@@ -1629,7 +1626,7 @@ late_abort_fixup(arg)
   defined(CPU_XSCALE_PXA2X0) || defined(CPU_XSCALE_IXP425) ||		\
   defined(CPU_XSCALE_80219) || defined(CPU_XSCALE_81342) || \
   defined(CPU_ARM10) ||  defined(CPU_ARM11) || \
-  defined(CPU_FA526)
+  defined(CPU_FA526) || defined(CPU_FA626TE)
 
 #define IGN	0
 #define OR	1
@@ -2095,7 +2092,7 @@ sa11x0_setup(args)
 }
 #endif	/* CPU_SA1100 || CPU_SA1110 */
 
-#if defined(CPU_FA526)
+#if defined(CPU_FA526) || defined(CPU_FA626TE)
 struct cpu_option fa526_options[] = {
 #ifdef COMPAT_12
 	{ "nocache",		IGN, BIC, (CPU_CONTROL_IC_ENABLE |
@@ -2149,7 +2146,7 @@ fa526_setup(char *args)
 	ctrl = cpuctrl;
 	cpu_control(0xffffffff, cpuctrl);
 }
-#endif	/* CPU_FA526 */
+#endif	/* CPU_FA526 || CPU_FA626TE */
 
 
 #if defined(CPU_IXP12X0)
diff --git a/sys/arm/arm/cpufunc_asm_fa526.S b/sys/arm/arm/cpufunc_asm_fa526.S
index d68d7a6f26b..d53d29a1cfb 100644
--- a/sys/arm/arm/cpufunc_asm_fa526.S
+++ b/sys/arm/arm/cpufunc_asm_fa526.S
@@ -32,7 +32,11 @@
 #include <machine/asm.h>
 __FBSDID("$FreeBSD$");
 
+#ifdef CPU_FA526
 #define	CACHELINE_SIZE	16
+#else
+#define	CACHELINE_SIZE	32
+#endif
 
 ENTRY(fa526_setttb)
 	mov	r1, #0
diff --git a/sys/arm/arm/elf_trampoline.c b/sys/arm/arm/elf_trampoline.c
index 6addbc387d5..0f725c8d8b6 100644
--- a/sys/arm/arm/elf_trampoline.c
+++ b/sys/arm/arm/elf_trampoline.c
@@ -57,7 +57,7 @@ void __startC(void);
 #define cpu_idcache_wbinv_all	arm8_cache_purgeID
 #elif defined(CPU_ARM9)
 #define cpu_idcache_wbinv_all	arm9_idcache_wbinv_all
-#elif defined(CPU_FA526)
+#elif defined(CPU_FA526) || defined(CPU_FA626TE)
 #define cpu_idcache_wbinv_all	fa526_idcache_wbinv_all
 #elif defined(CPU_ARM9E)
 #define cpu_idcache_wbinv_all	armv5_ec_idcache_wbinv_all
diff --git a/sys/arm/arm/identcpu.c b/sys/arm/arm/identcpu.c
index 6d2a0f01c44..1a9c996bab0 100644
--- a/sys/arm/arm/identcpu.c
+++ b/sys/arm/arm/identcpu.c
@@ -220,7 +220,9 @@ const struct cpuidtab cpuids[] = {
 	  generic_steppings },
 	{ CPU_ID_ARM966ESR1,	CPU_CLASS_ARM9ES,	"ARM966E-S",
 	  generic_steppings },
-	{ CPU_ID_FA526,	CPU_CLASS_ARM9,	"FA526",
+	{ CPU_ID_FA526,		CPU_CLASS_ARM9,		"FA526",
+	  generic_steppings },
+	{ CPU_ID_FA626TE,	CPU_CLASS_ARM9ES,	"FA626TE",
 	  generic_steppings },
 
 	{ CPU_ID_TI925T,	CPU_CLASS_ARM9TDMI,	"TI ARM925T",
diff --git a/sys/arm/include/cpuconf.h b/sys/arm/include/cpuconf.h
index 942b5916933..c8a9b6e191e 100644
--- a/sys/arm/include/cpuconf.h
+++ b/sys/arm/include/cpuconf.h
@@ -62,6 +62,7 @@
 			 defined(CPU_XSCALE_80321) +			\
 			 defined(CPU_XSCALE_PXA2X0) +			\
 			 defined(CPU_FA526) +				\
+			 defined(CPU_FA626TE) +				\
 			 defined(CPU_XSCALE_IXP425))
 
 /*
@@ -78,7 +79,7 @@
 #if (defined(CPU_ARM9E) || defined(CPU_ARM10) ||			\
      defined(CPU_XSCALE_80200) || defined(CPU_XSCALE_80321) ||		\
      defined(CPU_XSCALE_80219) || defined(CPU_XSCALE_81342) ||		\
-     defined(CPU_XSCALE_PXA2X0))
+     defined(CPU_XSCALE_PXA2X0) || defined(CPU_FA626TE))
 #define	ARM_ARCH_5	1
 #else
 #define	ARM_ARCH_5	0
@@ -126,7 +127,8 @@
 
 #if (defined(CPU_ARM6) || defined(CPU_ARM7) || defined(CPU_ARM7TDMI) ||	\
      defined(CPU_ARM8) || defined(CPU_ARM9) || defined(CPU_ARM9E) ||	\
-     defined(CPU_ARM10) || defined(CPU_ARM11) || defined(CPU_FA526))
+     defined(CPU_ARM10) || defined(CPU_ARM11) || defined(CPU_FA526) ||	\
+     defined(CPU_FA626TE))
 #define	ARM_MMU_GENERIC		1
 #else
 #define	ARM_MMU_GENERIC		0
diff --git a/sys/arm/include/cpufunc.h b/sys/arm/include/cpufunc.h
index 84949668535..4ab63fbe3fa 100644
--- a/sys/arm/include/cpufunc.h
+++ b/sys/arm/include/cpufunc.h
@@ -284,7 +284,7 @@ u_int	arm8_clock_config	(u_int, u_int);
 #endif
 
 
-#ifdef CPU_FA526
+#if defined(CPU_FA526) || defined(CPU_FA626TE)
 void	fa526_setup		(char *arg);
 void	fa526_setttb		(u_int ttb);
 void	fa526_context_switch	(void);
@@ -464,11 +464,11 @@ extern unsigned armv5_dcache_index_max;
 extern unsigned armv5_dcache_index_inc;
 #endif
 
-#if defined(CPU_ARM9) || defined(CPU_ARM9E) || defined(CPU_ARM10) || \
-  defined(CPU_SA110) || defined(CPU_SA1100) || defined(CPU_SA1110) || \
-  defined(CPU_XSCALE_80200) || defined(CPU_XSCALE_80321) ||	     \
-    defined(CPU_FA526) || \
-  defined(CPU_XSCALE_PXA2X0) || defined(CPU_XSCALE_IXP425) ||	     \
+#if defined(CPU_ARM9) || defined(CPU_ARM9E) || defined(CPU_ARM10) ||	\
+  defined(CPU_SA110) || defined(CPU_SA1100) || defined(CPU_SA1110) ||	\
+  defined(CPU_XSCALE_80200) || defined(CPU_XSCALE_80321) ||		\
+  defined(CPU_FA526) || defined(CPU_FA626TE) ||				\
+  defined(CPU_XSCALE_PXA2X0) || defined(CPU_XSCALE_IXP425) ||		\
   defined(CPU_XSCALE_80219) || defined(CPU_XSCALE_81342)
   
 void	armv4_tlb_flushID	(void);
diff --git a/sys/conf/options.arm b/sys/conf/options.arm
index 922d38fda0c..6b1928195c7 100644
--- a/sys/conf/options.arm
+++ b/sys/conf/options.arm
@@ -37,3 +37,4 @@ AT91_BWCT		opt_at91.h
 AT91_TSC		opt_at91.h
 AT91_KWIKBYTE		opt_at91.h
 CPU_FA526		opt_global.h
+CPU_FA626TE		opt_global.h

From 2db9a1171cb6b8a6f009261d10354315c0ea2f13 Mon Sep 17 00:00:00 2001
From: Norikatsu Shigemura <nork@FreeBSD.org>
Date: Tue, 4 May 2010 11:25:04 +0000
Subject: [PATCH 445/532] Add support run services_mkdb(8).

Approved by:	dougb, imp (mentor)
Reviewed by:	ume
MFC after:	2 weeks
---
 usr.sbin/mergemaster/mergemaster.sh | 14 ++++++++++++++
 1 file changed, 14 insertions(+)

diff --git a/usr.sbin/mergemaster/mergemaster.sh b/usr.sbin/mergemaster/mergemaster.sh
index 7f0696846e1..84ee455d5fb 100755
--- a/usr.sbin/mergemaster/mergemaster.sh
+++ b/usr.sbin/mergemaster/mergemaster.sh
@@ -849,6 +849,9 @@ mm_install () {
     /etc/login.conf)
       NEED_CAP_MKDB=yes
       ;;
+    /etc/services)
+      NEED_SERVICES_MKDB=yes
+      ;;
     /etc/master.passwd)
       do_install_and_rm 600 "${1}" "${DESTDIR}${INSTALL_DIR}"
       NEED_PWD_MKDB=yes
@@ -1278,6 +1281,17 @@ case "${NEED_CAP_MKDB}" in
   ;;
 esac
 
+case "${NEED_SERVICES_MKDB}" in
+'') ;;
+*)
+  echo ''
+  echo "*** You installed a services file, so make sure that you run"
+  echo "    '/usr/sbin/services_mkdb -q -o ${DESTDIR}/var/db/services.db ${DESTDIR}/etc/services'"
+  echo "     to rebuild your services database"
+  run_it_now "/usr/sbin/services_mkdb -q -o ${DESTDIR}/var/db/services.db ${DESTDIR}/etc/services"
+  ;;
+esac
+
 case "${NEED_PWD_MKDB}" in
 '') ;;
 *)

From 11666ce50ae9eab40af280a181a4c8b0355feff7 Mon Sep 17 00:00:00 2001
From: Ed Maste <emaste@FreeBSD.org>
Date: Tue, 4 May 2010 11:34:13 +0000
Subject: [PATCH 446/532] Restore historical behaviour of only executing
 chflags on files that exist. This eliminates cosmetic errors of the form
 "chflags: ...: No such file or directory" during an installworld to an empty
 destination.

---
 usr.bin/chpass/Makefile | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/usr.bin/chpass/Makefile b/usr.bin/chpass/Makefile
index a5571d72ffb..566173e14d8 100644
--- a/usr.bin/chpass/Makefile
+++ b/usr.bin/chpass/Makefile
@@ -38,7 +38,9 @@ MLINKS+= chpass.1 ypchpass.1 chpass.1 ypchfn.1 chpass.1 ypchsh.1
 
 beforeinstall:
 .for i in chpass chfn chsh ypchpass ypchfn ypchsh
+.if exists(${DESTDIR}${BINDIR}/$i)
 	-chflags noschg ${DESTDIR}${BINDIR}/$i
+.endif
 .endfor
 
 .if !defined(NO_FSCHG)

From e7ff54750b0edb2e2f72bfa7398e6c24a251d770 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 4 May 2010 13:07:40 +0000
Subject: [PATCH 447/532] Bring in new files from edwin's tftp

---
 libexec/tftpd/tftp-file.c     | 257 ++++++++++++++++++
 libexec/tftpd/tftp-file.h     |  37 +++
 libexec/tftpd/tftp-io.c       | 478 ++++++++++++++++++++++++++++++++++
 libexec/tftpd/tftp-io.h       |  47 ++++
 libexec/tftpd/tftp-options.c  | 390 +++++++++++++++++++++++++++
 libexec/tftpd/tftp-options.h  |  62 +++++
 libexec/tftpd/tftp-transfer.c | 318 ++++++++++++++++++++++
 libexec/tftpd/tftp-transfer.h |  31 +++
 libexec/tftpd/tftp-utils.c    | 320 +++++++++++++++++++++++
 libexec/tftpd/tftp-utils.h    | 124 +++++++++
 10 files changed, 2064 insertions(+)
 create mode 100644 libexec/tftpd/tftp-file.c
 create mode 100644 libexec/tftpd/tftp-file.h
 create mode 100644 libexec/tftpd/tftp-io.c
 create mode 100644 libexec/tftpd/tftp-io.h
 create mode 100644 libexec/tftpd/tftp-options.c
 create mode 100644 libexec/tftpd/tftp-options.h
 create mode 100644 libexec/tftpd/tftp-transfer.c
 create mode 100644 libexec/tftpd/tftp-transfer.h
 create mode 100644 libexec/tftpd/tftp-utils.c
 create mode 100644 libexec/tftpd/tftp-utils.h

diff --git a/libexec/tftpd/tftp-file.c b/libexec/tftpd/tftp-file.c
new file mode 100644
index 00000000000..e0f8e784a4b
--- /dev/null
+++ b/libexec/tftpd/tftp-file.c
@@ -0,0 +1,257 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/types.h>
+#include <sys/stat.h>
+
+#include <netinet/in.h>
+#include <arpa/tftp.h>
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
+#include <unistd.h>
+
+#include "tftp-file.h"
+#include "tftp-utils.h"
+
+static FILE	*file;
+static int	convert;
+
+static char	convbuffer[66000];
+static int	gotcr = 0;
+
+static size_t
+convert_from_net(char *buffer, size_t count)
+{
+	size_t i, n;
+
+	/*
+	 * Convert all CR/LF to LF and all CR,NUL to CR
+	 */
+
+	n = 0;
+	for (i = 0; i < count; i++) {
+
+		if (gotcr == 0) {
+			convbuffer[n++] = buffer[i];
+			gotcr = (buffer[i] == '\r');
+			continue;
+		}
+
+		/* CR, NULL -> CR */
+		if (buffer[i] == '\0') {
+			gotcr = 0;
+			continue;
+		}
+
+		/* CR, LF -> LF */
+		if (buffer[i] == '\n') {
+			if (n == 0) {
+				if (ftell(file) != 0) {
+					fseek(file, -1, SEEK_END);
+					convbuffer[n++] = '\n';
+				} else {
+					/* This shouldn't happen */
+					tftp_log(LOG_ERR,
+					    "Received LF as first character");
+					abort();
+				}
+			} else
+				convbuffer[n-1] = '\n';
+			gotcr = 0;
+			continue;
+		}
+
+		/* Everything else just accept as is */
+		convbuffer[n++] = buffer[i];
+		gotcr = (buffer[i] == '\r');
+		continue;
+	}
+
+	return fwrite(convbuffer, 1, n, file);
+}
+
+static size_t
+convert_to_net(char *buffer, size_t count, int init)
+{
+	size_t i;
+	static size_t n = 0, read = 0;
+	static int newline = 0;
+
+	if (init) {
+		newline = 0;
+		n = 0;
+		read = 0;
+		return 0 ;
+	}
+
+	/*
+	 * Convert all LF to CR,LF and all CR to CR,NUL
+	 */
+	i = 0;
+
+	if (newline) {
+		buffer[i++] = newline;
+		newline = 0;
+	}
+
+	while (i < count) {
+		if (n == read) {
+			/* When done we're done */
+			if (feof(file)) break;
+
+			/* Otherwise read another bunch */
+			read = fread(convbuffer, 1, count, file);
+			if (read == 0) break;
+			n = 0;
+		}
+
+		/* CR -> CR,NULL */
+		if (convbuffer[n] == '\r') {
+			buffer[i++] = '\r';
+			buffer[i++] = '\0';
+			n++;
+			continue;
+		}
+
+		/* LF -> CR,LF */
+		if (convbuffer[n] == '\n') {
+			buffer[i++] = '\r';
+			buffer[i++] = '\n';
+			n++;
+			continue;
+		}
+
+		buffer[i++] = convbuffer[n++];
+	}
+
+	if (i > count) {
+		/*
+		 * Whoops... that isn't alllowed (but it will happen
+		 * when there is a CR or LF at the end of the buffer)
+		 */
+		newline = buffer[i-1];
+	}
+
+	if (i < count) {
+		/* We are done! */
+		return i;
+	} else
+		return count;
+
+}
+
+int
+write_init(int fd, FILE *f, const char *mode)
+{
+
+	if (f == NULL) {
+		file = fdopen(fd, "w");
+		if (file == NULL) {
+			int en = errno;
+			tftp_log(LOG_ERR, "fdopen() failed: %s",
+			    strerror(errno));
+			return en;
+		}
+	} else
+		file = f;
+	convert = !strcmp(mode, "netascii");
+	return 0;
+}
+
+size_t
+write_file(char *buffer, int count)
+{
+
+	if (convert == 0)
+		return fwrite(buffer, 1, count, file);
+
+	return convert_from_net(buffer, count);
+}
+
+int
+write_close(void)
+{
+
+	if (fclose(file) != 0) {
+		tftp_log(LOG_ERR, "fclose() failed: %s", strerror(errno));
+		return 1;
+	}
+	return 0;
+}
+
+int
+read_init(int fd, FILE *f, const char *mode)
+{
+
+	convert_to_net(NULL, 0, 1);
+	if (f == NULL) {
+		file = fdopen(fd, "r");
+		if (file == NULL) {
+			int en = errno;
+			tftp_log(LOG_ERR, "fdopen() failed: %s",
+			    strerror(errno));
+			return en;
+		}
+	} else
+		file = f;
+	convert = !strcmp(mode, "netascii");
+	return 0;
+}
+
+size_t
+read_file(char *buffer, int count)
+{
+
+	if (convert == 0)
+		return fread(buffer, 1, count, file);
+
+	return convert_to_net(buffer, count, 0);
+}
+
+int
+read_close(void)
+{
+
+	if (fclose(file) != 0) {
+		tftp_log(LOG_ERR, "fclose() failed: %s", strerror(errno));
+		return 1;
+	}
+	return 0;
+}
+
+
+int
+synchnet(int peer)
+{
+
+	return 0;
+}
diff --git a/libexec/tftpd/tftp-file.h b/libexec/tftpd/tftp-file.h
new file mode 100644
index 00000000000..fcc4d0d0321
--- /dev/null
+++ b/libexec/tftpd/tftp-file.h
@@ -0,0 +1,37 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+int	write_init(int fd, FILE *f, const char *mode);
+size_t	write_file(char *buffer, int count);
+int	write_close(void);
+
+int	read_init(int fd, FILE *f, const char *mode);
+size_t	read_file(char *buffer, int count);
+int	read_close(void);
+
+int	synchnet(int peer);
diff --git a/libexec/tftpd/tftp-io.c b/libexec/tftpd/tftp-io.c
new file mode 100644
index 00000000000..28628bab459
--- /dev/null
+++ b/libexec/tftpd/tftp-io.c
@@ -0,0 +1,478 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <sys/socket.h>
+
+#include <netinet/in.h>
+#include <arpa/tftp.h>
+#include <arpa/inet.h>
+
+#include <errno.h>
+#include <setjmp.h>
+#include <signal.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
+#include <unistd.h>
+
+#include "tftp-file.h"
+#include "tftp-io.h"
+#include "tftp-utils.h"
+#include "tftp-options.h"
+
+struct sockaddr_storage peer_sock;
+struct sockaddr_storage me_sock;
+
+static int send_packet(int peer, uint16_t block, char *pkt, int size);
+
+struct errmsg {
+	int	e_code;
+	const char	*e_msg;
+} errmsgs[] = {
+	{ EUNDEF,	"Undefined error code" },
+	{ ENOTFOUND,	"File not found" },
+	{ EACCESS,	"Access violation" },
+	{ ENOSPACE,	"Disk full or allocation exceeded" },
+	{ EBADOP,	"Illegal TFTP operation" },
+	{ EBADID,	"Unknown transfer ID" },
+	{ EEXISTS,	"File already exists" },
+	{ ENOUSER,	"No such user" },
+	{ EOPTNEG,	"Option negotiation" },
+	{ -1,		NULL }
+};
+
+#define DROPPACKET(s)							\
+	if (packetdroppercentage != 0 &&				\
+	    random()%100 < packetdroppercentage) {			\
+		tftp_log(LOG_DEBUG, "Artifical packet drop in %s", s);	\
+		return;							\
+	}
+#define DROPPACKETn(s,n)						\
+	if (packetdroppercentage != 0 &&				\
+	    random()%100 < packetdroppercentage) {			\
+		tftp_log(LOG_DEBUG, "Artifical packet drop in %s", s);	\
+		return (n);						\
+	}
+
+const char *
+errtomsg(int error)
+{
+	static char ebuf[40];
+	struct errmsg *pe;
+	char buf[MAXPKTSIZE];
+
+	if (error == 0)
+		return ("success");
+	for (pe = errmsgs; pe->e_code >= 0; pe++)
+		if (pe->e_code == error)
+			return (pe->e_msg);
+	snprintf(ebuf, sizeof(buf), "error %d", error);
+	return (ebuf);
+}
+
+static int
+send_packet(int peer, uint16_t block, char *pkt, int size)
+{
+	int i;
+	int t = 1;
+
+	for (i = 0; i < 12 ; i++) {
+		DROPPACKETn("send_packet", 0);
+
+		if (sendto(peer, pkt, size, 0,
+			(struct sockaddr *)&peer_sock, peer_sock.ss_len)
+			== size) {
+			if (i)
+				tftp_log(LOG_ERR,
+				    "%s block %d, attempt %d successful",
+				    block, i);
+			return (0);
+		}
+		tftp_log(LOG_ERR,
+		    "%s block %d, attempt %d failed (Error %d: %s)", 
+		    packettype(ntohs(((struct tftphdr *)(pkt))->th_opcode)),
+		    block, i, errno, strerror(errno));
+		sleep(t);
+		if (t < 32)
+			t <<= 1;
+	}
+	tftp_log(LOG_ERR, "send_packet: %s", strerror(errno));
+	return (1);
+}
+
+/*
+ * Send an ERROR packet (error message).
+ * Error code passed in is one of the
+ * standard TFTP codes, or a UNIX errno
+ * offset by 100.
+ */
+void
+send_error(int peer, int error)
+{
+	struct tftphdr *tp;
+	int length;
+	struct errmsg *pe;
+	char buf[MAXPKTSIZE];
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending ERROR %d: %s", error);
+
+	DROPPACKET("send_error");
+
+	tp = (struct tftphdr *)buf;
+	tp->th_opcode = htons((u_short)ERROR);
+	tp->th_code = htons((u_short)error);
+	for (pe = errmsgs; pe->e_code >= 0; pe++)
+		if (pe->e_code == error)
+			break;
+	if (pe->e_code < 0) {
+		pe->e_msg = strerror(error - 100);
+		tp->th_code = EUNDEF;   /* set 'undef' errorcode */
+	}
+	strcpy(tp->th_msg, pe->e_msg);
+	length = strlen(pe->e_msg);
+	tp->th_msg[length] = '\0';
+	length += 5;
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending ERROR %d: %s", error, tp->th_msg);
+
+	if (sendto(peer, buf, length, 0,
+		(struct sockaddr *)&peer_sock, peer_sock.ss_len) != length)
+		tftp_log(LOG_ERR, "send_error: %s", strerror(errno));
+}
+
+/*
+ * Send an WRQ packet (write request).
+ */
+int
+send_wrq(int peer, char *filename, char *mode)
+{
+	int n;
+	struct tftphdr *tp;
+	char *bp;
+	char buf[MAXPKTSIZE];
+	int size;
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending WRQ: filename: '%s', mode '%s'",
+			filename, mode
+		);
+
+	DROPPACKETn("send_wrq", 1);
+
+	tp = (struct tftphdr *)buf;
+	tp->th_opcode = htons((u_short)WRQ);
+	size = 2;
+
+	bp = tp->th_stuff;
+	strcpy(bp, filename);
+	bp += strlen(filename);
+	*bp = 0;
+	bp++;
+	size += strlen(filename) + 1;
+
+	strcpy(bp, mode);
+	bp += strlen(mode);
+	*bp = 0;
+	bp++;
+	size += strlen(mode) + 1;
+
+	if (options_rfc_enabled)
+		size += make_options(peer, bp, sizeof(buf) - size);
+
+	n = sendto(peer, buf, size, 0,
+	    (struct sockaddr *)&peer_sock, peer_sock.ss_len);
+	if (n != size) {
+		tftp_log(LOG_ERR, "send_wrq: %s", strerror(errno));
+		return (1);
+	}
+	return (0);
+}
+
+/*
+ * Send an RRQ packet (write request).
+ */
+int
+send_rrq(int peer, char *filename, char *mode)
+{
+	int n;
+	struct tftphdr *tp;
+	char *bp;
+	char buf[MAXPKTSIZE];
+	int size;
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending RRQ: filename: '%s', mode '%s'",
+			filename, mode
+		);
+
+	DROPPACKETn("send_rrq", 1);
+
+	tp = (struct tftphdr *)buf;
+	tp->th_opcode = htons((u_short)RRQ);
+	size = 2;
+
+	bp = tp->th_stuff;
+	strcpy(bp, filename);
+	bp += strlen(filename);
+	*bp = 0;
+	bp++;
+	size += strlen(filename) + 1;
+
+	strcpy(bp, mode);
+	bp += strlen(mode);
+	*bp = 0;
+	bp++;
+	size += strlen(mode) + 1;
+
+	if (options_rfc_enabled) {
+		options[OPT_TSIZE].o_request = strdup("0");
+		size += make_options(peer, bp, sizeof(buf) - size);
+	}
+
+	n = sendto(peer, buf, size, 0,
+	    (struct sockaddr *)&peer_sock, peer_sock.ss_len);
+	if (n != size) {
+		tftp_log(LOG_ERR, "send_rrq: %s", n, strerror(errno));
+		return (1);
+	}
+	return (0);
+}
+
+/*
+ * Send an OACK packet (option acknowledgement).
+ */
+int
+send_oack(int peer)
+{
+	struct tftphdr *tp;
+	int size, i, n;
+	char *bp;
+	char buf[MAXPKTSIZE];
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending OACK");
+
+	DROPPACKETn("send_oack", 0);
+
+	/*
+	 * Send back an options acknowledgement (only the ones with
+	 * a reply for)
+	 */
+	tp = (struct tftphdr *)buf;
+	bp = buf + 2;
+	size = sizeof(buf) - 2;
+	tp->th_opcode = htons((u_short)OACK);
+	for (i = 0; options[i].o_type != NULL; i++) {
+		if (options[i].o_reply != NULL) {
+			n = snprintf(bp, size, "%s%c%s", options[i].o_type,
+				     0, options[i].o_reply);
+			bp += n+1;
+			size -= n+1;
+			if (size < 0) {
+				tftp_log(LOG_ERR, "oack: buffer overflow");
+				exit(1);
+			}
+		}
+	}
+	size = bp - buf;
+
+	if (sendto(peer, buf, size, 0,
+		(struct sockaddr *)&peer_sock, peer_sock.ss_len) != size) {
+		tftp_log(LOG_INFO, "send_oack: %s", strerror(errno));
+		return (1);
+	}
+
+	return (0);
+}
+
+/*
+ * Send an ACK packet (acknowledgement).
+ */
+int
+send_ack(int fp, uint16_t block)
+{
+	struct tftphdr *tp;
+	int size;
+	char *bp;
+	char buf[MAXPKTSIZE];
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending ACK for block %d", block);
+
+	DROPPACKETn("send_ack", 0);
+
+	tp = (struct tftphdr *)buf;
+	bp = buf + 2;
+	size = sizeof(buf) - 2;
+	tp->th_opcode = htons((u_short)ACK);
+	tp->th_block = htons((u_short)block);
+	size = 4;
+
+	if (sendto(fp, buf, size, 0,
+	    (struct sockaddr *)&peer_sock, peer_sock.ss_len) != size) {
+		tftp_log(LOG_INFO, "send_ack: %s", strerror(errno));
+		return (1);
+	}
+
+	return (0);
+}
+
+/*
+ * Send a DATA packet
+ */
+int
+send_data(int peer, uint16_t block, char *data, int size)
+{
+	char buf[MAXPKTSIZE];
+	struct tftphdr *pkt;
+	int n;
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Sending DATA packet %d of %d bytes",
+			block, size);
+
+	DROPPACKETn("send_data", 0);
+
+	pkt = (struct tftphdr *)buf;
+
+	pkt->th_opcode = htons((u_short)DATA);
+	pkt->th_block = htons((u_short)block);
+	memcpy(pkt->th_data, data, size);
+
+	n = send_packet(peer, block, (char *)pkt, size + 4);
+	return (n);
+}
+
+
+/*
+ * Receive a packet
+ */
+jmp_buf	timeoutbuf;
+
+static void
+timeout(int sig __unused)
+{
+
+	/* tftp_log(LOG_DEBUG, "Timeout\n");	Inside a signal handler... */
+	longjmp(timeoutbuf, 1);
+}
+
+int
+receive_packet(int peer, char *data, int size, struct sockaddr_storage *from,
+    int thistimeout)
+{
+	struct tftphdr *pkt;
+	struct sockaddr_storage from_local;
+	struct sockaddr_storage *pfrom;
+	socklen_t fromlen;
+	int n;
+	static int waiting;
+
+	pfrom = (from == NULL) ? &from_local : from;
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG,
+		    "Waiting %d seconds for packet", timeoutpacket);
+
+	pkt = (struct tftphdr *)data;
+
+	waiting = 0;
+	signal(SIGALRM, timeout);
+	setjmp(timeoutbuf);
+	alarm(thistimeout);
+
+	if (waiting > 0) {
+		alarm(0);
+		return (RP_TIMEOUT);
+	}
+
+	if (waiting > 0) {
+		tftp_log(LOG_ERR, "receive_packet: timeout");
+		alarm(0);
+		return (RP_TIMEOUT);
+	}
+
+	waiting++;
+	fromlen = sizeof(*pfrom);
+	n = recvfrom(peer, data, size, 0, (struct sockaddr *)pfrom, &fromlen);
+
+	alarm(0);
+
+	DROPPACKETn("receive_packet", RP_TIMEOUT);
+
+	if (n < 0) {
+		tftp_log(LOG_ERR, "receive_packet: timeout");
+		return (RP_TIMEOUT);
+	}
+
+	alarm(0);
+
+	if (n < 0) {
+		/* No idea what could have happened if it isn't a timeout */
+		tftp_log(LOG_ERR, "receive_packet: %s", strerror(errno));
+		return (RP_RECVFROM);
+	}
+	if (n < 4) {
+		tftp_log(LOG_ERR,
+		    "receive_packet: packet too small (%d bytes)", n);
+		return (RP_TOOSMALL);
+	}
+
+	pkt->th_opcode = ntohs((u_short)pkt->th_opcode);
+	if (pkt->th_opcode == DATA ||
+	    pkt->th_opcode == ACK)
+		pkt->th_block = ntohs((u_short)pkt->th_block);
+
+	if (pkt->th_opcode == DATA && n > pktsize) {
+		tftp_log(LOG_ERR, "receive_packet: packet too big");
+		return (RP_TOOBIG);
+	}
+
+	if (((struct sockaddr_in *)(pfrom))->sin_addr.s_addr !=
+	    ((struct sockaddr_in *)(&peer_sock))->sin_addr.s_addr) {
+		tftp_log(LOG_ERR,
+			"receive_packet: received packet from wrong source");
+		return (RP_WRONGSOURCE);
+	}
+
+	if (pkt->th_opcode == ERROR) {
+		tftp_log(LOG_ERR, "Got ERROR packet: %s", pkt->th_msg);
+		return (RP_ERROR);
+	}
+
+	if (debug&DEBUG_PACKETS)
+		tftp_log(LOG_DEBUG, "Received %d bytes in a %s packet",
+			n, packettype(pkt->th_opcode));
+
+	return n - 4;
+}
diff --git a/libexec/tftpd/tftp-io.h b/libexec/tftpd/tftp-io.h
new file mode 100644
index 00000000000..70558bc634e
--- /dev/null
+++ b/libexec/tftpd/tftp-io.h
@@ -0,0 +1,47 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#define RP_NONE		0
+#define	RP_RECVFROM	-1
+#define	RP_TOOSMALL	-2
+#define RP_ERROR	-3
+#define RP_WRONGSOURCE	-4
+#define	RP_TIMEOUT	-5
+#define	RP_TOOBIG	-6
+
+const char *errtomsg(int);
+void	send_error(int peer, int);
+int	send_wrq(int peer, char *, char *);
+int	send_rrq(int peer, char *, char *);
+int	send_oack(int peer);
+int	send_ack(int peer, unsigned short);
+int	send_data(int peer, uint16_t, char *, int);
+int	receive_packet(int peer, char *, int, struct sockaddr_storage *, int);
+
+extern struct sockaddr_storage peer_sock;
+extern struct sockaddr_storage me_sock;
diff --git a/libexec/tftpd/tftp-options.c b/libexec/tftpd/tftp-options.c
new file mode 100644
index 00000000000..0b97aaf18f6
--- /dev/null
+++ b/libexec/tftpd/tftp-options.c
@@ -0,0 +1,390 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/socket.h>
+#include <sys/types.h>
+#include <sys/sysctl.h>
+#include <sys/stat.h>
+
+#include <netinet/in.h>
+#include <arpa/tftp.h>
+
+#include <ctype.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
+
+#include "tftp-utils.h"
+#include "tftp-io.h"
+#include "tftp-options.h"
+
+/*
+ * Option handlers
+ */
+
+struct options options[] = {
+	{ "tsize",	NULL, NULL, NULL /* option_tsize */, 1 },
+	{ "timeout",	NULL, NULL, option_timeout, 1 },
+	{ "blksize",	NULL, NULL, option_blksize, 1 },
+	{ "blksize2",	NULL, NULL, option_blksize2, 0 },
+	{ "rollover",	NULL, NULL, option_rollover, 0 },
+	{ NULL,		NULL, NULL, NULL, 0 }
+};
+
+/* By default allow them */
+int options_rfc_enabled = 1;
+int options_extra_enabled = 1;
+
+/*
+ * Rules for the option handlers:
+ * - If there is no o_request, there will be no processing.
+ *
+ * For servers
+ * - Logging is done as warnings.
+ * - The handler exit()s if there is a serious problem with the
+ *   values submitted in the option.
+ *
+ * For clients
+ * - Logging is done as errors. After all, the server shouldn't
+ *   return rubbish.
+ * - The handler returns if there is a serious problem with the
+ *   values submitted in the option.
+ * - Sending the EBADOP packets is done by the handler.
+ */
+
+int
+option_tsize(int peer, struct tftphdr *tp, int mode, struct stat *stbuf)
+{
+
+	if (options[OPT_TSIZE].o_request == NULL)
+		return (0);
+
+	if (mode == RRQ) 
+		asprintf(&options[OPT_TSIZE].o_reply,
+			"%ju", stbuf->st_size);
+	else
+		/* XXX Allows writes of all sizes. */
+		options[OPT_TSIZE].o_reply =
+			strdup(options[OPT_TSIZE].o_request);
+	return (0);
+}
+
+int
+option_timeout(int peer)
+{
+
+	if (options[OPT_TIMEOUT].o_request == NULL)
+		return (0);
+
+	int to = atoi(options[OPT_TIMEOUT].o_request);
+	if (to < TIMEOUT_MIN || to > TIMEOUT_MAX) {
+		tftp_log(acting_as_client ? LOG_ERR : LOG_WARNING,
+		    "Received bad value for timeout. "
+		    "Should be between %d and %d, received %s",
+		    TIMEOUT_MIN, TIMEOUT_MAX);
+		send_error(peer, EBADOP);
+		if (acting_as_client)
+			return (1);
+		exit(1);
+	} else {
+		timeoutpacket = to;
+		options[OPT_TIMEOUT].o_reply =
+			strdup(options[OPT_TIMEOUT].o_request);
+	}
+	settimeouts(timeoutpacket, timeoutnetwork, maxtimeouts);
+
+	if (debug&DEBUG_OPTIONS)
+		tftp_log(LOG_DEBUG, "Setting timeout to '%s'",
+			options[OPT_TIMEOUT].o_reply);
+
+	return (0);
+}
+
+int
+option_rollover(int peer)
+{
+
+	if (options[OPT_ROLLOVER].o_request == NULL)
+		return (0);
+
+	if (strcmp(options[OPT_ROLLOVER].o_request, "0") != 0
+	 && strcmp(options[OPT_ROLLOVER].o_request, "1") != 0) {
+		tftp_log(acting_as_client ? LOG_ERR : LOG_WARNING,
+		    "Bad value for rollover, "
+		    "should be either 0 or 1, received '%s', "
+		    "ignoring request",
+		    options[OPT_ROLLOVER].o_request);
+		if (acting_as_client) {
+			send_error(peer, EBADOP);
+			return (1);
+		}
+		return (0);
+	}
+	options[OPT_ROLLOVER].o_reply =
+		strdup(options[OPT_ROLLOVER].o_request);
+
+	if (debug&DEBUG_OPTIONS)
+		tftp_log(LOG_DEBUG, "Setting rollover to '%s'",
+			options[OPT_ROLLOVER].o_reply);
+
+	return (0);
+}
+
+int
+option_blksize(int peer)
+{
+	int *maxdgram;
+	char maxbuffer[100];
+	size_t len;
+
+	if (options[OPT_BLKSIZE].o_request == NULL)
+		return (0);
+
+	/* maximum size of an UDP packet according to the system */
+	len = sizeof(maxbuffer);
+	if (sysctlbyname("net.inet.udp.maxdgram",
+	    maxbuffer, &len, NULL, 0) < 0) {
+		tftp_log(LOG_ERR, "sysctl: net.inet.udp.maxdgram");
+		return (acting_as_client ? 1 : 0);
+	}
+	maxdgram = (int *)maxbuffer;
+
+	int size = atoi(options[OPT_BLKSIZE].o_request);
+	if (size < BLKSIZE_MIN || size > BLKSIZE_MAX) {
+		if (acting_as_client) {
+			tftp_log(LOG_ERR,
+			    "Invalid blocksize (%d bytes), aborting",
+			    size);
+			send_error(peer, EBADOP);
+			return (1);
+		} else {
+			tftp_log(LOG_WARNING,
+			    "Invalid blocksize (%d bytes), ignoring request",
+			    size);
+			return (0);
+		}
+	}
+
+	if (size > *maxdgram) {
+		if (acting_as_client) {
+			tftp_log(LOG_ERR,
+			    "Invalid blocksize (%d bytes), "
+			    "net.inet.udp.maxdgram sysctl limits it to "
+			    "%d bytes.\n", size, *maxdgram);
+			send_error(peer, EBADOP);
+			return (1);
+		} else {
+			tftp_log(LOG_WARNING,
+			    "Invalid blocksize (%d bytes), "
+			    "net.inet.udp.maxdgram sysctl limits it to "
+			    "%d bytes.\n", size, *maxdgram);
+			size = *maxdgram;
+			/* No reason to return */
+		}
+	}
+
+	asprintf(&options[OPT_BLKSIZE].o_reply, "%d", size);
+	segsize = size;
+	pktsize = size + 4;
+	if (debug&DEBUG_OPTIONS)
+		tftp_log(LOG_DEBUG, "Setting blksize to '%s'",
+		    options[OPT_BLKSIZE].o_reply);
+
+	return (0);
+}
+
+int
+option_blksize2(int peer)
+{
+	int	*maxdgram;
+	char	maxbuffer[100];
+	int	size, i;
+	size_t	len;
+
+	int sizes[] = {
+		8, 16, 32, 64, 128, 256, 512, 1024,
+		2048, 4096, 8192, 16384, 32768, 0
+	};
+
+	if (options[OPT_BLKSIZE2].o_request == NULL)
+		return (0);
+
+	/* maximum size of an UDP packet according to the system */
+	len = sizeof(maxbuffer);
+	if (sysctlbyname("net.inet.udp.maxdgram",
+	    maxbuffer, &len, NULL, 0) < 0) {
+		tftp_log(LOG_ERR, "sysctl: net.inet.udp.maxdgram");
+		return (acting_as_client ? 1 : 0);
+	}
+	maxdgram = (int *)maxbuffer;
+
+	size = atoi(options[OPT_BLKSIZE2].o_request);
+	for (i = 0; sizes[i] != 0; i++) {
+		if (size == sizes[i]) break;
+	}
+	if (sizes[i] == 0) {
+		tftp_log(LOG_INFO,
+		    "Invalid blocksize2 (%d bytes), ignoring request", size);
+		return (acting_as_client ? 1 : 0);
+	}
+
+	if (size > *maxdgram) {
+		for (i = 0; sizes[i+1] != 0; i++) {
+			if (*maxdgram < sizes[i+1]) break;
+		}
+		tftp_log(LOG_INFO,
+		    "Invalid blocksize2 (%d bytes), net.inet.udp.maxdgram "
+		    "sysctl limits it to %d bytes.\n", size, *maxdgram);
+		size = sizes[i];
+		/* No need to return */
+	}
+
+	asprintf(&options[OPT_BLKSIZE2].o_reply, "%d", size);
+	segsize = size;
+	pktsize = size + 4;
+	if (debug&DEBUG_OPTIONS)
+		tftp_log(LOG_DEBUG, "Setting blksize2 to '%s'",
+		    options[OPT_BLKSIZE2].o_reply);
+
+	return (0);
+}
+
+/*
+ * Append the available options to the header
+ */
+uint16_t
+make_options(int peer, char *buffer, uint16_t size) {
+	int	i;
+	char	*value;
+	const char *option;
+	uint16_t length;
+	uint16_t returnsize = 0;
+
+	if (!options_rfc_enabled) return (0);
+
+	for (i = 0; options[i].o_type != NULL; i++) {
+		if (options[i].rfc == 0 && !options_extra_enabled)
+			continue;
+
+		option = options[i].o_type;
+		if (acting_as_client)
+			value = options[i].o_request;
+		else
+			value = options[i].o_reply;
+		if (value == NULL)
+			continue;
+
+		length = strlen(value) + strlen(option) + 2;
+		if (size <= length) {
+			tftp_log(LOG_ERR,
+			    "Running out of option space for "
+			    "option '%s' with value '%s': "
+			    "needed %d bytes, got %d bytes",
+			    option, value, size, length);
+			continue;
+		}
+
+		sprintf(buffer, "%s%c%s%c", option, '\000', value, '\000');
+		size -= length;
+		buffer += length;
+		returnsize += length;
+	}
+
+	return (returnsize);
+}
+
+/*
+ * Parse the received options in the header
+ */
+int
+parse_options(int peer, char *buffer, uint16_t size)
+{
+	int	i, options_failed;
+	char	*c, *cp, *option, *value;
+
+	if (!options_rfc_enabled) return (0);
+
+	/* Parse the options */
+	cp = buffer;
+	options_failed = 0;	
+	while (size > 0) {
+		option = cp;
+		i = get_field(peer, cp, size);
+		cp += i;
+
+		value = cp;
+		i = get_field(peer, cp, size);
+		cp += i;
+
+		/* We are at the end */
+		if (*option == '\0') break;
+
+		if (debug&DEBUG_OPTIONS)
+			tftp_log(LOG_DEBUG,
+			    "option: '%s' value: '%s'", option, value);
+
+		for (c = option; *c; c++)
+			if (isupper(*c))
+				*c = tolower(*c);
+		for (i = 0; options[i].o_type != NULL; i++) {
+			if (strcmp(option, options[i].o_type) == 0) {
+				if (!acting_as_client)
+					options[i].o_request = value;
+				if (!options_extra_enabled && !options[i].rfc) {
+					tftp_log(LOG_INFO,
+					    "Option '%s' with value '%s' found "
+					    "but it is not an RFC option",
+					    option, value);
+					continue;
+				}
+				if (options[i].o_handler)
+					options_failed +=
+					    (options[i].o_handler)(peer);
+				break;
+			}
+		}
+		if (options[i].o_type == NULL)
+			tftp_log(LOG_WARNING,
+			    "Unknown option: '%s'", option);
+
+		size -= strlen(option) + strlen(value) + 2;
+	}
+
+	return (options_failed);
+}
+
+/*
+ * Set some default values in the options
+ */
+void
+init_options(void)
+{
+
+	options[OPT_ROLLOVER].o_request = strdup("0");
+}
diff --git a/libexec/tftpd/tftp-options.h b/libexec/tftpd/tftp-options.h
new file mode 100644
index 00000000000..d8bd2fc6daa
--- /dev/null
+++ b/libexec/tftpd/tftp-options.h
@@ -0,0 +1,62 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+/*
+ * Options
+ */
+
+void		init_options(void);
+uint16_t	make_options(int peer, char *buffer, uint16_t size);
+int		parse_options(int peer, char *buffer, uint16_t size);
+
+/* Call back functions */
+int	option_tsize(int peer, struct tftphdr *, int, struct stat *);
+int	option_timeout(int peer);
+int	option_blksize(int peer);
+int	option_blksize2(int peer);
+int	option_rollover(int peer);
+
+extern int options_extra_enabled;
+extern int options_rfc_enabled;
+
+struct options {
+	const char	*o_type;
+	char		*o_request;
+	char		*o_reply;
+	int		(*o_handler)(int peer);
+	int		rfc;
+};
+
+extern struct options	options[];
+enum opt_enum {
+	OPT_TSIZE = 0,
+	OPT_TIMEOUT,
+	OPT_BLKSIZE,
+	OPT_BLKSIZE2,
+	OPT_ROLLOVER,
+};
diff --git a/libexec/tftpd/tftp-transfer.c b/libexec/tftpd/tftp-transfer.c
new file mode 100644
index 00000000000..fd0e00d9409
--- /dev/null
+++ b/libexec/tftpd/tftp-transfer.c
@@ -0,0 +1,318 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/types.h>
+#include <sys/param.h>
+#include <sys/ioctl.h>
+#include <sys/stat.h>
+#include <sys/socket.h>
+
+#include <netinet/in.h>
+#include <arpa/tftp.h>
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <syslog.h>
+
+#include "tftp-file.h"
+#include "tftp-io.h"
+#include "tftp-utils.h"
+#include "tftp-options.h"
+#include "tftp-transfer.h"
+
+/*
+ * Send a file via the TFTP data session.
+ */
+void
+tftp_send(int peer, uint16_t *block, struct tftp_stats *ts)
+{
+	struct tftphdr *rp;
+	int size, n_data, n_ack, try;
+	uint16_t oldblock;
+	char sendbuffer[MAXPKTSIZE];
+	char recvbuffer[MAXPKTSIZE];
+
+	rp = (struct tftphdr *)recvbuffer;
+	*block = 1;
+	ts->amount = 0;
+	do {
+		if (debug&DEBUG_SIMPLE)
+			tftp_log(LOG_DEBUG, "Sending block %d", *block);
+
+		size = read_file(sendbuffer, segsize);
+		if (size < 0) {
+			tftp_log(LOG_ERR, "read_file returned %d", size);
+			send_error(peer, errno + 100);
+			goto abort;
+		}
+
+		for (try = 0; ; try++) {
+			n_data = send_data(peer, *block, sendbuffer, size);
+			if (n_data > 0) {
+				if (try == maxtimeouts) {
+					tftp_log(LOG_ERR,
+					    "Cannot send DATA packet #%d, "
+					    "giving up", *block);
+					return;
+				}
+				tftp_log(LOG_ERR,
+				    "Cannot send DATA packet #%d, trying again",
+				    *block);
+				continue;
+			}
+
+			n_ack = receive_packet(peer, recvbuffer,
+			    MAXPKTSIZE, NULL, timeoutpacket);
+			if (n_ack < 0) {
+				if (n_ack == RP_TIMEOUT) {
+					if (try == maxtimeouts) {
+						tftp_log(LOG_ERR,
+						    "Timeout #%d send ACK %d "
+						    "giving up", try, *block);
+						return;
+					}
+					tftp_log(LOG_WARNING,
+					    "Timeout #%d on ACK %d",
+					    try, *block);
+					continue;
+				}
+
+				/* Either read failure or ERROR packet */
+				if (debug&DEBUG_SIMPLE)
+					tftp_log(LOG_ERR, "Aborting: %s",
+					    rp_strerror(n_ack));
+				goto abort;
+			}
+			if (rp->th_opcode == ACK) {
+				ts->blocks++;
+				if (rp->th_block == *block) {
+					ts->amount += size;
+					break;
+				}
+
+				/* Re-synchronize with the other side */
+				(void) synchnet(peer);
+				if (rp->th_block == (*block - 1)) {
+					ts->retries++;
+					continue;
+				}
+			}
+
+		}
+		oldblock = *block;
+		(*block)++;
+		if (oldblock > *block) {
+			if (options[OPT_ROLLOVER].o_request == NULL) {
+				tftp_log(LOG_ERR,
+				    "Block rollover but not allowed.");
+				send_error(peer, EBADOP);
+				gettimeofday(&(ts->tstop), NULL);
+				return;
+			}
+
+			*block = atoi(options[OPT_ROLLOVER].o_request);
+			ts->rollovers++;
+		}
+		gettimeofday(&(ts->tstop), NULL);
+	} while (size == segsize);
+abort:
+	return;
+}
+
+/*
+ * Receive a file via the TFTP data session.
+ *
+ * - It could be that the first block has already arrived while
+ *   trying to figure out if we were receiving options or not. In
+ *   that case it is passed to this function.
+ */
+void
+tftp_receive(int peer, uint16_t *block, struct tftp_stats *ts,
+    struct tftphdr *firstblock, size_t fb_size)
+{
+	struct tftphdr *rp;
+	uint16_t oldblock;
+	int n_data, n_ack, writesize, i, retry;
+	char recvbuffer[MAXPKTSIZE];
+
+	ts->amount = 0;
+
+	if (firstblock != NULL) {
+		writesize = write_file(firstblock->th_data, fb_size);
+		ts->amount += writesize;
+		for (i = 0; ; i++) {
+			n_ack = send_ack(peer, *block);
+			if (n_ack > 0) {
+				if (i == maxtimeouts) {
+					tftp_log(LOG_ERR,
+					    "Cannot send ACK packet #%d, "
+					    "giving up", *block);
+					return;
+				}
+				tftp_log(LOG_ERR,
+				    "Cannot send ACK packet #%d, trying again",
+				    *block);
+				continue;
+			}
+
+			break;
+		}
+
+		if (fb_size != segsize) {
+			gettimeofday(&(ts->tstop), NULL);
+			return;
+		}
+	}
+
+	rp = (struct tftphdr *)recvbuffer;
+	do {
+		oldblock = *block;
+		(*block)++;
+		if (oldblock > *block) {
+			if (options[OPT_ROLLOVER].o_request == NULL) {
+				tftp_log(LOG_ERR,
+				    "Block rollover but not allowed.");
+				send_error(peer, EBADOP);
+				gettimeofday(&(ts->tstop), NULL);
+				return;
+			}
+
+			*block = atoi(options[OPT_ROLLOVER].o_request);
+			ts->rollovers++;
+		}
+
+		for (retry = 0; ; retry++) {
+			if (debug&DEBUG_SIMPLE)
+				tftp_log(LOG_DEBUG,
+				    "Receiving DATA block %d", *block);
+
+			n_data = receive_packet(peer, recvbuffer,
+			    MAXPKTSIZE, NULL, timeoutpacket);
+			if (n_data < 0) {
+				if (retry == maxtimeouts) {
+					tftp_log(LOG_ERR,
+					    "Timeout #%d on DATA block %d, "
+					    "giving up", retry, *block);
+					return;
+				}
+				if (n_data == RP_TIMEOUT) {
+					tftp_log(LOG_WARNING,
+					    "Timeout #%d on DATA block %d",
+					    retry, *block);
+					send_ack(peer, oldblock);
+					continue;
+				}
+
+				/* Either read failure or ERROR packet */
+				if (debug&DEBUG_SIMPLE)
+					tftp_log(LOG_DEBUG, "Aborting: %s",
+					    rp_strerror(n_data));
+				goto abort;
+			}
+			if (rp->th_opcode == DATA) {
+				ts->blocks++;
+
+				if (rp->th_block == *block)
+					break;
+
+				tftp_log(LOG_WARNING,
+				    "Expected DATA block %d, got block %d",
+				    *block, rp->th_block);
+
+				/* Re-synchronize with the other side */
+				(void) synchnet(peer);
+				if (rp->th_block == (*block-1)) {
+					tftp_log(LOG_INFO, "Trying to sync");
+					*block = oldblock;
+					ts->retries++;
+					goto send_ack;	/* rexmit */
+				}
+
+			} else {
+				tftp_log(LOG_WARNING,
+				    "Expected DATA block, got %s block",
+				    packettype(rp->th_opcode));
+			}
+		}
+
+		if (n_data > 0) {
+			writesize = write_file(rp->th_data, n_data);
+			ts->amount += writesize;
+			if (writesize <= 0) {
+				tftp_log(LOG_ERR,
+				    "write_file returned %d", writesize);
+				if (writesize < 0)
+					send_error(peer, errno + 100);
+				else
+					send_error(peer, ENOSPACE);
+				goto abort;
+			}
+		}
+
+send_ack:
+		for (i = 0; ; i++) {
+			n_ack = send_ack(peer, *block);
+			if (n_ack > 0) {
+
+				if (i == maxtimeouts) {
+					tftp_log(LOG_ERR,
+					    "Cannot send ACK packet #%d, "
+					    "giving up", *block);
+					return;
+				}
+
+				tftp_log(LOG_ERR,
+				    "Cannot send ACK packet #%d, trying again",
+				    *block);
+				continue;
+			}
+
+			break;
+		}
+		gettimeofday(&(ts->tstop), NULL);
+	} while (n_data == segsize);
+
+	/* Don't do late packet management for the client implementation */
+	if (acting_as_client)
+		return;
+
+	for (i = 0; ; i++) {
+		n_data = receive_packet(peer, (char *)rp, pktsize,
+		    NULL, timeoutpacket);
+		if (n_data <= 0)
+			break;
+		if (n_data > 0 &&
+		    rp->th_opcode == DATA &&	/* and got a data block */
+		    *block == rp->th_block)	/* then my last ack was lost */
+			send_ack(peer, *block);	/* resend final ack */
+	}
+
+abort:
+	return;
+}
diff --git a/libexec/tftpd/tftp-transfer.h b/libexec/tftpd/tftp-transfer.h
new file mode 100644
index 00000000000..2cfa2df8ee8
--- /dev/null
+++ b/libexec/tftpd/tftp-transfer.h
@@ -0,0 +1,31 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+void	tftp_send(int peer, uint16_t *block, struct tftp_stats *tp);
+void	tftp_receive(int peer, uint16_t *block, struct tftp_stats *tp,
+	    struct tftphdr *firstblock, size_t fb_size);
diff --git a/libexec/tftpd/tftp-utils.c b/libexec/tftpd/tftp-utils.c
new file mode 100644
index 00000000000..da58064dc06
--- /dev/null
+++ b/libexec/tftpd/tftp-utils.c
@@ -0,0 +1,320 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+#include <sys/socket.h>
+#include <sys/stat.h>
+
+#include <netinet/in.h>
+#include <arpa/tftp.h>
+
+#include <stdarg.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <syslog.h>
+
+#include "tftp-utils.h"
+#include "tftp-io.h"
+
+/*
+ * Default values, can be changed later via the TFTP Options
+ */
+int		timeoutpacket = TIMEOUT;
+int		timeoutnetwork = MAX_TIMEOUTS * TIMEOUT;
+int		maxtimeouts = MAX_TIMEOUTS;
+uint16_t	segsize = SEGSIZE;
+uint16_t	pktsize = SEGSIZE + 4;
+
+int	acting_as_client;
+
+
+/*
+ * Set timeout values for packet reception. The idea is that you
+ * get 'maxtimeouts' of 5 seconds between 'timeoutpacket' (i.e. the
+ * first timeout) to 'timeoutnetwork' (i.e. the last timeout)
+ */
+int
+settimeouts(int _timeoutpacket, int _timeoutnetwork, int _maxtimeouts)
+{
+	int i;
+	
+	/* We cannot do impossible things */
+	if (_timeoutpacket >= _timeoutnetwork)
+		return (0);
+	
+	maxtimeouts = 0;
+	i = _timeoutpacket;
+	while (i < _timeoutnetwork || maxtimeouts < MIN_TIMEOUTS) {
+		maxtimeouts++;
+		i += 5;
+	}
+
+	timeoutpacket = _timeoutpacket;
+	timeoutnetwork = i;
+	return (1);
+}
+
+/* translate IPv4 mapped IPv6 address to IPv4 address */
+void
+unmappedaddr(struct sockaddr_in6 *sin6)
+{
+	struct sockaddr_in *sin4;
+	u_int32_t addr;
+	int port;
+
+	if (sin6->sin6_family != AF_INET6 ||
+	    !IN6_IS_ADDR_V4MAPPED(&sin6->sin6_addr))
+		return;
+	sin4 = (struct sockaddr_in *)sin6;
+	addr = *(u_int32_t *)&sin6->sin6_addr.s6_addr[12];
+	port = sin6->sin6_port;
+	memset(sin4, 0, sizeof(struct sockaddr_in));
+	sin4->sin_addr.s_addr = addr;
+	sin4->sin_port = port;
+	sin4->sin_family = AF_INET;
+	sin4->sin_len = sizeof(struct sockaddr_in);
+}
+
+/* Get a field from a \0 seperated string */
+ssize_t
+get_field(int peer, char *buffer, ssize_t size)
+{
+	char *cp = buffer;
+
+	while (cp < buffer + size) {
+		if (*cp == '\0') break;
+		cp++;
+	}
+	if (*cp != '\0') {
+		tftp_log(LOG_ERR, "Bad option - no trailing \\0 found");
+		send_error(peer, EBADOP);
+		exit(1);
+	}
+	return (cp - buffer + 1);
+}
+
+/*
+ * Logging functions
+ */
+int	_tftp_logtostdout = 1;
+
+void
+tftp_openlog(const char *ident, int logopt, int facility)
+{
+
+	_tftp_logtostdout = (ident == NULL);
+	if (_tftp_logtostdout == 0)
+		openlog(ident, logopt, facility);
+}
+
+void
+tftp_closelog(void)
+{
+
+	if (_tftp_logtostdout == 0)
+		closelog();
+}
+
+void
+tftp_log(int priority, const char *message, ...)
+{
+	va_list ap;
+	char *s;
+
+	va_start(ap, message);
+	if (_tftp_logtostdout == 0) {
+		vasprintf(&s, message, ap);
+		syslog(priority, "%s", s);
+	} else {
+		vprintf(message, ap);
+		printf("\n");
+	}
+	va_end(ap);
+}
+
+/*
+ * Packet types
+ */
+struct packettypes packettypes[] = {
+	{ RRQ,		"RRQ"	},
+	{ WRQ,		"WRQ"	},
+	{ DATA,		"DATA"	},
+	{ ACK,		"ACK"	},
+	{ ERROR,	"ERROR"	},
+	{ OACK,		"OACK"	},
+	{ 0,		NULL	},
+};
+
+char *
+packettype(int type)
+{
+	static char failed[100];
+	int i = 0;
+
+	while (packettypes[i].name != NULL) {
+		if (packettypes[i].value == type)
+			break;
+		i++;
+	}
+	if (packettypes[i].name != NULL)
+		return packettypes[i].name;
+	sprintf(failed, "unknown (type: %d)", type);
+	return (failed);
+}
+
+/*
+ * Debugs
+ */
+int	debug = DEBUG_NONE;
+struct debugs debugs[] = {
+	{ DEBUG_PACKETS,	"packet",	"Packet debugging"	},
+	{ DEBUG_SIMPLE,		"simple",	"Simple debugging"	},
+	{ DEBUG_OPTIONS,	"options",	"Options debugging"	},
+	{ DEBUG_ACCESS,		"access",	"TCPd access debugging"	},
+	{ DEBUG_NONE,		NULL,		"No debugging"		},
+};
+int	packetdroppercentage = 0;
+
+int
+debug_find(char *s)
+{
+	int i = 0;
+
+	while (debugs[i].name != NULL) {
+		if (strcasecmp(debugs[i].name, s) == 0)
+			break;
+		i++;
+	}
+	return (debugs[i].value);
+}
+
+int
+debug_finds(char *s)
+{
+	int i = 0;
+	char *ps = s;
+
+	while (s != NULL) {
+		ps = strchr(s, ' ');
+		if (ps != NULL)
+			*ps = '\0';
+		i += debug_find(s);
+		if (ps != NULL)
+			*ps = ' ';
+		s = ps;
+	}
+	return (i);
+}
+
+char *
+debug_show(int d)
+{
+	static char s[100];
+	int i = 0;
+
+	s[0] = '\0';
+	while (debugs[i].name != NULL) {
+		if (d&debugs[i].value) {
+			if (s[0] != '\0') 
+				strcat(s, " ");
+			strcat(s, debugs[i].name);
+		}
+		i++;
+	}
+	if (s[0] != '\0')
+		return (s);
+	return ("none");
+}
+
+/*
+ * RP_
+ */
+struct rp_errors rp_errors[] = {
+	{ RP_TIMEOUT,		"Network timeout" },
+	{ RP_TOOSMALL,		"Not enough data bytes" },
+	{ RP_WRONGSOURCE,	"Invalid IP address of UDP port" },
+	{ RP_ERROR,		"Error packet" },
+	{ RP_RECVFROM,		"recvfrom() complained" },
+	{ RP_TOOBIG,		"Too many data bytes" },
+	{ RP_NONE,		NULL }
+};
+
+char *
+rp_strerror(int error)
+{
+	static char s[100];
+	int i = 0;
+
+	while (rp_errors[i].desc != NULL) {
+		if (rp_errors[i].error == error) {
+			strcpy(s, rp_errors[i].desc);
+		}
+		i++;
+	}
+	if (s[0] == '\0')
+		sprintf(s, "unknown (error=%d)", error);
+	return (s);
+}
+
+/*
+ * Performance figures
+ */
+
+void
+stats_init(struct tftp_stats *ts)
+{
+
+	ts->amount = 0;
+	ts->rollovers = 0;
+	ts->retries = 0;
+	ts->blocks = 0;
+	ts->amount = 0;
+	gettimeofday(&(ts->tstart), NULL);
+}
+
+void
+printstats(const char *direction, int verbose, struct tftp_stats *ts)
+{
+	double delta;	/* compute delta in 1/10's second units */
+
+	delta = ((ts->tstop.tv_sec*10.)+(ts->tstop.tv_usec/100000)) -
+		((ts->tstart.tv_sec*10.)+(ts->tstart.tv_usec/100000));
+	delta = delta/10.;      /* back to seconds */
+
+	printf("%s %zu bytes during %.1f seconds in %u blocks",
+	    direction, ts->amount, delta, ts->blocks);
+
+	if (ts->rollovers != 0)
+		printf(" with %d rollover%s",
+		    ts->rollovers, ts->rollovers != 1 ? "s" : "");
+
+	if (verbose)
+		printf(" [%.0f bits/sec]", (ts->amount*8.)/delta);
+	putchar('\n');
+}
+
diff --git a/libexec/tftpd/tftp-utils.h b/libexec/tftpd/tftp-utils.h
new file mode 100644
index 00000000000..d07247908d9
--- /dev/null
+++ b/libexec/tftpd/tftp-utils.h
@@ -0,0 +1,124 @@
+/*
+ * Copyright (C) 2008 Edwin Groothuis. All rights reserved.
+ * 
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 
+ * THIS SOFTWARE IS PROVIDED BY AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
+#include <sys/cdefs.h>
+__FBSDID("$FreeBSD$");
+
+/*
+ */
+#define	TIMEOUT		5
+#define	MAX_TIMEOUTS	5
+
+/* Generic values */
+#define MAXSEGSIZE	65464		/* Maximum size of the data segment */
+#define	MAXPKTSIZE	(MAXSEGSIZE + 4) /* Maximum size of the packet */
+
+/* For the blksize option */
+#define BLKSIZE_MIN	8		/* Minumum size of the data segment */
+#define BLKSIZE_MAX	MAXSEGSIZE	/* Maximum size of the data segment */
+
+/* For the timeout option */
+#define TIMEOUT_MIN	0		/* Minumum timeout value */
+#define TIMEOUT_MAX	255		/* Maximum timeout value */
+#define MIN_TIMEOUTS	3
+
+extern int	timeoutpacket;
+extern int	timeoutnetwork;
+extern int	maxtimeouts;
+int	settimeouts(int timeoutpacket, int timeoutnetwork, int maxtimeouts);
+
+extern uint16_t	segsize;
+extern uint16_t	pktsize;
+
+extern int	acting_as_client;
+
+/*
+ */
+void	unmappedaddr(struct sockaddr_in6 *sin6);
+ssize_t	get_field(int peer, char *buffer, ssize_t size);
+
+/*
+ * Packet types
+ */
+struct packettypes {
+	int	value;
+	char	*name;
+};
+extern struct packettypes packettypes[];
+char *packettype(int);
+
+/*
+ * RP_
+ */
+struct rp_errors {
+	int	error;
+	char	*desc;
+};
+extern struct rp_errors rp_errors[];
+char	*rp_strerror(int error);
+
+/*
+ * Debug features
+ */
+#define	DEBUG_NONE	0x0000
+#define DEBUG_PACKETS	0x0001
+#define DEBUG_SIMPLE	0x0002
+#define DEBUG_OPTIONS	0x0004
+#define DEBUG_ACCESS	0x0008
+struct debugs {
+	int	value;
+	char	*name;
+	char	*desc;
+};
+extern int	debug;
+extern struct debugs debugs[];
+extern int	packetdroppercentage;
+int	debug_find(char *s);
+int	debug_finds(char *s);
+char	*debug_show(int d);
+
+/*
+ * Log routines
+ */
+#define DEBUG(s) tftp_log(LOG_DEBUG, "%s", s)
+extern int tftp_logtostdout;
+void	tftp_openlog(const char *ident, int logopt, int facility);
+void	tftp_closelog(void);
+void	tftp_log(int priority, const char *message, ...);
+
+/*
+ * Performance figures
+ */
+struct tftp_stats {
+	size_t		amount;
+	int		rollovers;
+	uint32_t	blocks;
+	int		retries;
+	struct timeval	tstart;
+	struct timeval	tstop;
+};
+
+void	stats_init(struct tftp_stats *ts);
+void	printstats(const char *direction, int verbose, struct tftp_stats *ts);

From 24ffe7244813469cd4819dd90132f9083b16f317 Mon Sep 17 00:00:00 2001
From: "Christian S.J. Peron" <csjp@FreeBSD.org>
Date: Tue, 4 May 2010 15:29:07 +0000
Subject: [PATCH 448/532] Add a case to make sure that internal audit records
 get converted to BSM format for lpathconf(2) events.

MFC after:	2 weeks
---
 sys/security/audit/audit_bsm.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/security/audit/audit_bsm.c b/sys/security/audit/audit_bsm.c
index faa45835318..b4713cc8333 100644
--- a/sys/security/audit/audit_bsm.c
+++ b/sys/security/audit/audit_bsm.c
@@ -740,6 +740,7 @@ kaudit_to_bsm(struct kaudit_record *kar, struct au_record **pau)
 	case AUE_LUTIMES:
 	case AUE_NFS_GETFH:
 	case AUE_LSTAT:
+	case AUE_LPATHCONF:
 	case AUE_PATHCONF:
 	case AUE_READLINK:
 	case AUE_REVOKE:

From a7283d32132a41a985e7c92afbc81fcee183a6e5 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Tue, 4 May 2010 15:55:41 +0000
Subject: [PATCH 449/532] Add page locking to the vm_page_cow* functions.

Push down the acquisition and release of the page queues lock into
vm_page_wire().

Reviewed by:	kib
---
 sys/dev/drm/via_dmablit.c |  2 --
 sys/kern/uipc_cow.c       |  3 ---
 sys/kern/vfs_bio.c        |  2 --
 sys/net/bpf_zerocopy.c    |  2 --
 sys/vm/vm_fault.c         |  6 ------
 sys/vm/vm_page.c          | 20 +++++++++++++-------
 6 files changed, 13 insertions(+), 22 deletions(-)

diff --git a/sys/dev/drm/via_dmablit.c b/sys/dev/drm/via_dmablit.c
index 72d914eacd8..81d438aea01 100644
--- a/sys/dev/drm/via_dmablit.c
+++ b/sys/dev/drm/via_dmablit.c
@@ -251,10 +251,8 @@ via_lock_all_dma_pages(drm_via_sg_info_t *vsg,  drm_via_dmablit_t *xfer)
 		if (m == NULL)
 			break;
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_wire(m);
 		vm_page_unhold(m);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 		vsg->pages[i] = m;
 	}
diff --git a/sys/kern/uipc_cow.c b/sys/kern/uipc_cow.c
index 88a5625fb5b..bc9a56a3dcf 100644
--- a/sys/kern/uipc_cow.c
+++ b/sys/kern/uipc_cow.c
@@ -131,10 +131,8 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	 * set up COW
 	 */
 	vm_page_lock(pp);
-	vm_page_lock_queues();
 	if (vm_page_cowsetup(pp) != 0) {
 		vm_page_unhold(pp);
-		vm_page_unlock_queues();
 		vm_page_unlock(pp);
 		return (0);
 	}
@@ -144,7 +142,6 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	 */
 	vm_page_wire(pp);
 	vm_page_unhold(pp);
-	vm_page_unlock_queues();
 	vm_page_unlock(pp);
 	/*
 	 * Allocate an sf buf
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 5db2d9e9773..589bdbd8093 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -3043,9 +3043,7 @@ allocbuf(struct buf *bp, int size)
 				 * We have a good page.
 				 */
 				vm_page_lock(m);
-				vm_page_lock_queues();
 				vm_page_wire(m);
-				vm_page_unlock_queues();
 				vm_page_unlock(m);
 				bp->b_pages[bp->b_npages] = m;
 				++bp->b_npages;
diff --git a/sys/net/bpf_zerocopy.c b/sys/net/bpf_zerocopy.c
index 5ca69782ee3..5629093fc62 100644
--- a/sys/net/bpf_zerocopy.c
+++ b/sys/net/bpf_zerocopy.c
@@ -171,10 +171,8 @@ zbuf_sfbuf_get(struct vm_map *map, vm_offset_t uaddr)
 	if (pp == NULL)
 		return (NULL);
 	vm_page_lock(pp);
-	vm_page_lock_queues();
 	vm_page_wire(pp);
 	vm_page_unhold(pp);
-	vm_page_unlock_queues();
 	vm_page_unlock(pp);
 	sf = sf_buf_alloc(pp, SFB_NOWAIT);
 	if (sf == NULL) {
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index f9255944a01..142a9ef5d5a 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -315,8 +315,6 @@ RetryFault:;
 			    (fault_type & VM_PROT_WRITE) &&
 			    (fs.object == fs.first_object)) {
 				vm_page_cowfault(fs.m);
-				vm_page_unlock_queues();
-				vm_page_unlock(fs.m);
 				unlock_and_deallocate(&fs);
 				goto RetryFault;
 			}
@@ -797,9 +795,7 @@ vnode_locked:
 				if (wired && (fault_flags &
 				    VM_FAULT_CHANGE_WIRING) == 0) {
 					vm_page_lock(fs.first_m);
-					vm_page_lock_queues();
 					vm_page_wire(fs.first_m);
-					vm_page_unlock_queues();
 					vm_page_unlock(fs.first_m);
 					
 					vm_page_lock(fs.m);
@@ -1285,9 +1281,7 @@ vm_fault_copy_entry(vm_map_t dst_map, vm_map_t src_map,
 			vm_page_unlock(src_m);
 
 			vm_page_lock(dst_m);
-			vm_page_lock_queues();
 			vm_page_wire(dst_m);
-			vm_page_unlock_queues();
 			vm_page_unlock(dst_m);
 		} else {
 			vm_page_lock(dst_m);
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index d9288d98f4d..957bf824679 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1544,13 +1544,15 @@ vm_page_wire(vm_page_t m)
 	 * and only unqueue the page if it is on some queue (if it is unmanaged
 	 * it is already off the queues).
 	 */
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	vm_page_lock_assert(m, MA_OWNED);
 	if (m->flags & PG_FICTITIOUS)
 		return;
 	if (m->wire_count == 0) {
-		if ((m->flags & PG_UNMANAGED) == 0)
+		if ((m->flags & PG_UNMANAGED) == 0) {
+			vm_page_lock_queues();
 			vm_pageq_remove(m);
+			vm_page_unlock_queues();
+		}
 		atomic_add_int(&cnt.v_wire_count, 1);
 	}
 	m->wire_count++;
@@ -1922,9 +1924,7 @@ retrylookup:
 		} else {
 			if ((allocflags & VM_ALLOC_WIRED) != 0) {
 				vm_page_lock(m);
-				vm_page_lock_queues();
 				vm_page_wire(m);
-				vm_page_unlock_queues();
 				vm_page_unlock(m);
 			}
 			if ((allocflags & VM_ALLOC_NOBUSY) == 0)
@@ -2224,6 +2224,7 @@ vm_page_cowfault(vm_page_t m)
 	vm_object_t object;
 	vm_pindex_t pindex;
 
+	vm_page_lock_assert(m, MA_OWNED);
 	object = m->object;
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
 	KASSERT(object->paging_in_progress != 0,
@@ -2238,17 +2239,18 @@ vm_page_cowfault(vm_page_t m)
 	if (mnew == NULL) {
 		vm_page_insert(m, object, pindex);
 		vm_page_unlock_queues();
+		vm_page_unlock(m);
 		VM_OBJECT_UNLOCK(object);
 		VM_WAIT;
 		VM_OBJECT_LOCK(object);
 		if (m == vm_page_lookup(object, pindex)) {
+			vm_page_lock(m);
 			vm_page_lock_queues();
 			goto retry_alloc;
 		} else {
 			/*
 			 * Page disappeared during the wait.
 			 */
-			vm_page_lock_queues();
 			return;
 		}
 	}
@@ -2269,13 +2271,15 @@ vm_page_cowfault(vm_page_t m)
 		mnew->wire_count = m->wire_count - m->cow;
 		m->wire_count = m->cow;
 	}
+	vm_page_unlock_queues();
+	vm_page_unlock(m);
 }
 
 void 
 vm_page_cowclear(vm_page_t m)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	if (m->cow) {
 		m->cow--;
 		/* 
@@ -2291,11 +2295,13 @@ int
 vm_page_cowsetup(vm_page_t m)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	vm_page_lock_assert(m, MA_OWNED);
 	if (m->cow == USHRT_MAX - 1)
 		return (EBUSY);
 	m->cow++;
+	vm_page_lock_queues();
 	pmap_remove_write(m);
+	vm_page_unlock_queues();
 	return (0);
 }
 

From c909aebcbb2125164da7ec014ca9ec245074486f Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Tue, 4 May 2010 16:56:59 +0000
Subject: [PATCH 450/532] - Remove more dead code[1].  Since r207330, we only
 need to check division by zero of the second argument 'from'. - Prefer
 u_int32_t over unsigned int to make its intention more clearer. - Move the
 function to a header file and make it a static inline function.

Pointed out by:	Andrew Reilly (areilly at bigpond dot net dot au)[1]
MFC after:	3 days
---
 sys/dev/sound/pcm/buffer.c | 13 -------------
 sys/dev/sound/pcm/buffer.h | 12 +++++++++++-
 2 files changed, 11 insertions(+), 14 deletions(-)

diff --git a/sys/dev/sound/pcm/buffer.c b/sys/dev/sound/pcm/buffer.c
index a9053d751d8..687542e4a84 100644
--- a/sys/dev/sound/pcm/buffer.c
+++ b/sys/dev/sound/pcm/buffer.c
@@ -565,19 +565,6 @@ sndbuf_updateprevtotal(struct snd_dbuf *b)
 	b->prev_total = b->total;
 }
 
-unsigned int
-snd_xbytes(unsigned int v, unsigned int from, unsigned int to)
-{
-
-	if (from == to)
-		return v;
-
-	if (from == 0 || to == 0 || v == 0)
-		return 0;
-
-	return (unsigned int)(((u_int64_t)v * to) / from);
-}
-
 unsigned int
 sndbuf_xbytes(unsigned int v, struct snd_dbuf *from, struct snd_dbuf *to)
 {
diff --git a/sys/dev/sound/pcm/buffer.h b/sys/dev/sound/pcm/buffer.h
index 91a63af0760..d079cdbad4f 100644
--- a/sys/dev/sound/pcm/buffer.h
+++ b/sys/dev/sound/pcm/buffer.h
@@ -111,7 +111,6 @@ u_int64_t sndbuf_getblocks(struct snd_dbuf *b);
 u_int64_t sndbuf_getprevblocks(struct snd_dbuf *b);
 u_int64_t sndbuf_gettotal(struct snd_dbuf *b);
 u_int64_t sndbuf_getprevtotal(struct snd_dbuf *b);
-unsigned int snd_xbytes(unsigned int v, unsigned int from, unsigned int to);
 unsigned int sndbuf_xbytes(unsigned int v, struct snd_dbuf *from, struct snd_dbuf *to);
 u_int8_t sndbuf_zerodata(u_int32_t fmt);
 void sndbuf_updateprevtotal(struct snd_dbuf *b);
@@ -132,3 +131,14 @@ void sndbuf_dmabounce(struct snd_dbuf *b);
 #ifdef OSSV4_EXPERIMENT
 void sndbuf_getpeaks(struct snd_dbuf *b, int *lp, int *rp);
 #endif
+
+static inline u_int32_t
+snd_xbytes(u_int32_t v, u_int32_t from, u_int32_t to)
+{
+
+	if (from == to)
+		return (v);
+	if (from == 0)
+		return (0);
+	return ((u_int64_t)v * to / from);
+}

From a3255f9b15cf339989ee69a0d00bd773c72a3333 Mon Sep 17 00:00:00 2001
From: Warner Losh <imp@FreeBSD.org>
Date: Tue, 4 May 2010 17:00:18 +0000
Subject: [PATCH 451/532] Doh!  Add another new file forgotten by the importer
 of edwin@'s tftp improvements.

MFC after:	1 week
Pointy Hat: imp-o-rama...
---
 usr.bin/tftp/tftp.h | 42 ++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 42 insertions(+)
 create mode 100644 usr.bin/tftp/tftp.h

diff --git a/usr.bin/tftp/tftp.h b/usr.bin/tftp/tftp.h
new file mode 100644
index 00000000000..5928dea4bcc
--- /dev/null
+++ b/usr.bin/tftp/tftp.h
@@ -0,0 +1,42 @@
+/*
+ * Copyright (c) 1993
+ *	The Regents of the University of California.  All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ * 3. All advertising materials mentioning features or use of this software
+ *    must display the following acknowledgement:
+ *	This product includes software developed by the University of
+ *	California, Berkeley and its contributors.
+ * 4. Neither the name of the University nor the names of its contributors
+ *    may be used to endorse or promote products derived from this software
+ *    without specific prior written permission.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ *	@(#)extern.h	8.1 (Berkeley) 6/6/93
+ * $FreeBSD$
+ */
+
+void	recvfile(int peer, char *port, int fd, char *name, char *mode);
+void	xmitfile(int peer, char *port, int fd, char *name, char *mode);
+
+extern int	verbose;
+extern int	maxtimeout;
+extern volatile int txrx_error;

From 3d763c313346277f8b43c6a659de5b68bd092f7c Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 17:02:34 +0000
Subject: [PATCH 452/532] Drop driver lock before exiting from interrupt
 handler.

Submitted by:	jhb
MFC after:	3 days
---
 sys/dev/msk/if_msk.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index e15ccfe738d..1c871a80d4d 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -3594,6 +3594,7 @@ msk_intr(void *xsc)
 	    (sc->msk_pflags & MSK_FLAG_SUSPEND) != 0 ||
 	    (status & sc->msk_intrmask) == 0) {
 		CSR_WRITE_4(sc, B0_Y2_SP_ICR, 2);
+		MSK_UNLOCK(sc);
 		return;
 	}
 

From 31fefd0d5d7f18d280a1ef35b233d64a055434f5 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 17:12:36 +0000
Subject: [PATCH 453/532] Make sure to check whether driver is running before
 processing received frames. Also check driver has valid ifp pointer before
 calling msk_stop() in device_shutdown handler. While I'm here remove
 unnecessary accesses to interrupt mask registers in device_shutdown handler
 because driver puts the controller into reset state. With these changes,
 msk(4) now survive from heavy RX traffic(1byte UDP frame) while reboot is in
 progress.

Reported by:	Mark Atkinson < atkin901 <> gmail dot com >
---
 sys/dev/msk/if_msk.c | 15 ++++++---------
 1 file changed, 6 insertions(+), 9 deletions(-)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index 1c871a80d4d..af977bfecf6 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -2904,20 +2904,15 @@ mskc_shutdown(device_t dev)
 	sc = device_get_softc(dev);
 	MSK_LOCK(sc);
 	for (i = 0; i < sc->msk_num_port; i++) {
-		if (sc->msk_if[i] != NULL)
+		if (sc->msk_if[i] != NULL && sc->msk_if[i]->msk_ifp != NULL &&
+		    ((sc->msk_if[i]->msk_ifp->if_drv_flags &
+		    IFF_DRV_RUNNING) != 0))
 			msk_stop(sc->msk_if[i]);
 	}
-
-	/* Disable all interrupts. */
-	CSR_WRITE_4(sc, B0_IMSK, 0);
-	CSR_READ_4(sc, B0_IMSK);
-	CSR_WRITE_4(sc, B0_HWE_IMSK, 0);
-	CSR_READ_4(sc, B0_HWE_IMSK);
+	MSK_UNLOCK(sc);
 
 	/* Put hardware reset. */
 	CSR_WRITE_2(sc, B0_CTST, CS_RST_SET);
-
-	MSK_UNLOCK(sc);
 	return (0);
 }
 
@@ -3525,6 +3520,8 @@ msk_handle_events(struct msk_softc *sc)
 			sc_if->msk_csum = status;
 			break;
 		case OP_RXSTAT:
+			if (!(sc_if->msk_ifp->if_drv_flags & IFF_DRV_RUNNING))
+				break;
 			if (sc_if->msk_framesize >
 			    (MCLBYTES - MSK_RX_BUF_ALIGN))
 				msk_jumbo_rxeof(sc_if, status, control, len);

From 77a7f647494ffa5e8524a16ea01a9143ba1090bd Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Tue, 4 May 2010 17:30:07 +0000
Subject: [PATCH 454/532] Fix deadlock during zfs receive.

OpenSolaris onnv revision:	9299:8809e849f63e

PR:		kern/146296
Submitted by:	myself
Approved by:	pjd, delphij (mentor)
Obtained from:	OpenSolaris (Bug ID 6783818, 6826836)
MFC after:	1 week
---
 .../opensolaris/uts/common/fs/zfs/dbuf.c      |  6 ++--
 .../uts/common/fs/zfs/dmu_object.c            | 28 ++++++++++---------
 2 files changed, 18 insertions(+), 16 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c
index 3bf0939a521..053c1e1baa2 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dbuf.c
@@ -464,15 +464,15 @@ dbuf_read_impl(dmu_buf_impl_t *db, zio_t *zio, uint32_t *flags)
 	ASSERT(db->db_buf == NULL);
 
 	if (db->db_blkid == DB_BONUS_BLKID) {
-		int bonuslen = dn->dn_bonuslen;
+		int bonuslen = MIN(dn->dn_bonuslen, dn->dn_phys->dn_bonuslen);
 
 		ASSERT3U(bonuslen, <=, db->db.db_size);
 		db->db.db_data = zio_buf_alloc(DN_MAX_BONUSLEN);
 		arc_space_consume(DN_MAX_BONUSLEN);
 		if (bonuslen < DN_MAX_BONUSLEN)
 			bzero(db->db.db_data, DN_MAX_BONUSLEN);
-		bcopy(DN_BONUS(dn->dn_phys), db->db.db_data,
-		    bonuslen);
+		if (bonuslen)
+			bcopy(DN_BONUS(dn->dn_phys), db->db.db_data, bonuslen);
 		dbuf_update_data(db);
 		db->db_state = DB_CACHED;
 		mutex_exit(&db->db_mtx);
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c
index 25dfafd4f2c..1f91fc1ad36 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c
@@ -128,15 +128,6 @@ dmu_object_reclaim(objset_t *os, uint64_t object, dmu_object_type_t ot,
 		return (0);
 	}
 
-	tx = dmu_tx_create(os);
-	dmu_tx_hold_bonus(tx, object);
-	err = dmu_tx_assign(tx, TXG_WAIT);
-	if (err) {
-		dmu_tx_abort(tx);
-		dnode_rele(dn, FTAG);
-		return (err);
-	}
-
 	nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT);
 
 	/*
@@ -144,16 +135,27 @@ dmu_object_reclaim(objset_t *os, uint64_t object, dmu_object_type_t ot,
 	 * be a new file instance.   We must clear out the previous file
 	 * contents before we can change this type of metadata in the dnode.
 	 */
-	if (dn->dn_nblkptr > nblkptr || dn->dn_datablksz != blocksize)
-		dmu_free_long_range(os, object, 0, DMU_OBJECT_END);
+	if (dn->dn_nblkptr > nblkptr || dn->dn_datablksz != blocksize) {
+		err = dmu_free_long_range(os, object, 0, DMU_OBJECT_END);
+		if (err)
+			goto out;
+	}
+
+	tx = dmu_tx_create(os);
+	dmu_tx_hold_bonus(tx, object);
+	err = dmu_tx_assign(tx, TXG_WAIT);
+	if (err) {
+		dmu_tx_abort(tx);
+		goto out;
+	}
 
 	dnode_reallocate(dn, ot, blocksize, bonustype, bonuslen, tx);
 
 	dmu_tx_commit(tx);
-
+out:
 	dnode_rele(dn, FTAG);
 
-	return (0);
+	return (err);
 }
 
 int

From 7135e11ef94e8130e7f4418a8a4c21cd8a8451de Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 17:34:00 +0000
Subject: [PATCH 455/532] Remove clearing RxHashTable2 register. The register
 is reprogrammed in sge_rxfilter().

---
 sys/dev/sge/if_sge.c | 1 -
 1 file changed, 1 deletion(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 43b7022b983..753c6de4ff0 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1577,7 +1577,6 @@ sge_init_locked(struct sge_softc *sc)
 	CSR_WRITE_4(sc, RX_DESC, SGE_ADDR_LO(sc->sge_ldata.sge_rx_paddr));
 
 	CSR_WRITE_4(sc, TxMacControl, 0x60);
-	CSR_WRITE_4(sc, 0x6c, 0);
 	CSR_WRITE_4(sc, RxWakeOnLan, 0);
 	CSR_WRITE_4(sc, RxWakeOnLanData, 0);
 	/* Allow receiving VLAN frames. */

From 7d4daf9a103b25106fcff80f1588384650c0e346 Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Tue, 4 May 2010 17:40:24 +0000
Subject: [PATCH 456/532] Speed up ZFS list operation with objset prefetching.

Partial import of OpenSolaris onnv revisions:
8415:8809e849f63e, 10474:0e96dd3b905a

PR:		kern/146297
Submitted by:	myself
Approved by:	pjd, delphij (mentor)
Obtained from:	OpenSolaris (Bug ID 6386929, 6755389, 6847118)
MFC after:	2 weeks
---
 .../uts/common/fs/zfs/dmu_objset.c            | 33 +++++++++++++++++++
 .../uts/common/fs/zfs/sys/dmu_objset.h        |  3 +-
 .../opensolaris/uts/common/fs/zfs/zfs_ioctl.c | 11 +++++++
 3 files changed, 45 insertions(+), 2 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c
index 7981e06825c..c9e00d51151 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_objset.c
@@ -1213,6 +1213,39 @@ dmu_objset_find_spa(spa_t *spa, const char *name,
 	return (err);
 }
 
+/* ARGSUSED */
+int
+dmu_objset_prefetch(char *name, void *arg)
+{
+	dsl_dataset_t *ds;
+
+	if (dsl_dataset_hold(name, FTAG, &ds))
+		return (0);
+
+	if (!BP_IS_HOLE(&ds->ds_phys->ds_bp)) {
+		mutex_enter(&ds->ds_opening_lock);
+		if (!dsl_dataset_get_user_ptr(ds)) {
+			uint32_t aflags = ARC_NOWAIT | ARC_PREFETCH;
+			zbookmark_t zb;
+
+			zb.zb_objset = ds->ds_object;
+			zb.zb_object = 0;
+			zb.zb_level = -1;
+			zb.zb_blkid = 0;
+
+			(void) arc_read_nolock(NULL, dsl_dataset_get_spa(ds),
+			    &ds->ds_phys->ds_bp, NULL, NULL,
+			    ZIO_PRIORITY_ASYNC_READ,
+			    ZIO_FLAG_CANFAIL | ZIO_FLAG_SPECULATIVE,
+			    &aflags, &zb);
+		}
+		mutex_exit(&ds->ds_opening_lock);
+	}
+
+	dsl_dataset_rele(ds, FTAG);
+	return (0);
+}
+
 void
 dmu_objset_set_user(objset_t *os, void *user_ptr)
 {
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_objset.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_objset.h
index 15df29a1779..1d65727808c 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_objset.h
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/dmu_objset.h
@@ -26,8 +26,6 @@
 #ifndef	_SYS_DMU_OBJSET_H
 #define	_SYS_DMU_OBJSET_H
 
-#pragma ident	"%Z%%M%	%I%	%E% SMI"
-
 #include <sys/spa.h>
 #include <sys/arc.h>
 #include <sys/txg.h>
@@ -118,6 +116,7 @@ int dmu_objset_find(char *name, int func(char *, void *), void *arg,
     int flags);
 int dmu_objset_find_spa(spa_t *spa, const char *name,
     int func(spa_t *, uint64_t, const char *, void *), void *arg, int flags);
+int dmu_objset_prefetch(char *name, void *arg);
 void dmu_objset_byteswap(void *buf, size_t size);
 int dmu_objset_evict_dbufs(objset_t *os);
 
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
index 00e446b82e9..1536e0b4da6 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
@@ -1349,6 +1349,14 @@ zfs_ioc_dataset_list_next(zfs_cmd_t *zc)
 		(void) strlcat(zc->zc_name, "/", sizeof (zc->zc_name));
 	p = zc->zc_name + strlen(zc->zc_name);
 
+	if (zc->zc_cookie == 0) {
+		uint64_t cookie = 0;
+		int len = sizeof (zc->zc_name) - (p - zc->zc_name);
+
+		while (dmu_dir_list_next(os, len, p, NULL, &cookie) == 0)
+			dmu_objset_prefetch(p, NULL);
+	}
+
 	do {
 		error = dmu_dir_list_next(os,
 		    sizeof (zc->zc_name) - (p - zc->zc_name), p,
@@ -1387,6 +1395,9 @@ zfs_ioc_snapshot_list_next(zfs_cmd_t *zc)
 	objset_t *os;
 	int error;
 
+	if (zc->zc_cookie == 0)
+		dmu_objset_find(zc->zc_name, dmu_objset_prefetch,
+		    NULL, DS_FIND_SNAPSHOTS);
 	error = dmu_objset_open(zc->zc_name,
 	    DMU_OST_ANY, DS_MODE_USER | DS_MODE_READONLY, &os);
 	if (error)

From e2725a2c5c63f771a8d033c309fe6159982cfe7b Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Tue, 4 May 2010 17:44:40 +0000
Subject: [PATCH 457/532] Enable "zfs list" to list explicitly requested
 snapshots.

Partial import of OpenSolaris onnv revision:
8415:8809e849f63e

PR:		kern/146297
Submitted by:	myself
Approved by:	pjd, delphij (mentor)
Obtained from:	OpenSolaris (Bug ID 6758338)
MFC after:	2 weeks
---
 cddl/contrib/opensolaris/cmd/zfs/zfs_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c b/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c
index 79428f7f541..a9d3c01bec2 100644
--- a/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c
+++ b/cddl/contrib/opensolaris/cmd/zfs/zfs_main.c
@@ -1790,7 +1790,7 @@ zfs_do_list(int argc, char **argv)
 	boolean_t scripted = B_FALSE;
 	static char default_fields[] =
 	    "name,used,available,referenced,mountpoint";
-	int types = ZFS_TYPE_FILESYSTEM | ZFS_TYPE_VOLUME;
+	int types = ZFS_TYPE_DATASET;
 	boolean_t types_specified = B_FALSE;
 	char *fields = NULL;
 	list_cbdata_t cb = { 0 };

From 55c978ba9dfb99e56033b789d3a364b0c377090d Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 19:04:51 +0000
Subject: [PATCH 458/532] Enable multi-descriptor transmisstion for fragmented
 mbufs. There is no more need to defragment mbufs. After transmitting the
 multi-fragmented frame, the controller updates only the first descriptor of
 multi-descriptor transmission so it's driver's responsibility to clear OWN
 bits of remaining descriptor of multi-descriptor transmission. It seems the
 controller behaves much like jme(4) controllers in descriptor handling.

Tested by:	xclin <xclin <> cs dot nctu dot edu dot tw >
---
 sys/dev/sge/if_sge.c    | 232 +++++++++++++++++++++++-----------------
 sys/dev/sge/if_sgereg.h |  19 +++-
 2 files changed, 148 insertions(+), 103 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index 753c6de4ff0..c3849832650 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -756,6 +756,8 @@ sge_dma_alloc(struct sge_softc *sc)
 {
 	struct sge_chain_data *cd;
 	struct sge_list_data *ld;
+	struct sge_rxdesc *rxd;
+	struct sge_txdesc *txd;
 	int error, i;
 
 	cd = &sc->sge_cdata;
@@ -869,8 +871,12 @@ sge_dma_alloc(struct sge_softc *sc)
 
 	/* Create DMA maps for Tx buffers. */
 	for (i = 0; i < SGE_TX_RING_CNT; i++) {
+		txd = &cd->sge_txdesc[i];
+		txd->tx_m = NULL;
+		txd->tx_dmamap = NULL;
+		txd->tx_ndesc = 0;
 		error = bus_dmamap_create(cd->sge_txmbuf_tag, 0,
-		    &cd->sge_tx_map[i]);
+		    &txd->tx_dmamap);
 		if (error != 0) {
 			device_printf(sc->sge_dev,
 			    "could not create Tx DMA map.\n");
@@ -886,8 +892,11 @@ sge_dma_alloc(struct sge_softc *sc)
 	}
 	/* Create DMA maps for Rx buffers. */
 	for (i = 0; i < SGE_RX_RING_CNT; i++) {
+		rxd = &cd->sge_rxdesc[i];
+		rxd->rx_m = NULL;
+		rxd->rx_dmamap = NULL;
 		error = bus_dmamap_create(cd->sge_rxmbuf_tag, 0,
-		    &cd->sge_rx_map[i]);
+		    &rxd->rx_dmamap);
 		if (error) {
 			device_printf(sc->sge_dev,
 			    "could not create Rx DMA map.\n");
@@ -903,6 +912,8 @@ sge_dma_free(struct sge_softc *sc)
 {
 	struct sge_chain_data *cd;
 	struct sge_list_data *ld;
+	struct sge_rxdesc *rxd;
+	struct sge_txdesc *txd;
 	int i;
 
 	cd = &sc->sge_cdata;
@@ -934,10 +945,11 @@ sge_dma_free(struct sge_softc *sc)
 	/* Rx buffers. */
 	if (cd->sge_rxmbuf_tag != NULL) {
 		for (i = 0; i < SGE_RX_RING_CNT; i++) {
-			if (cd->sge_rx_map[i] != NULL) {
+			rxd = &cd->sge_rxdesc[i];
+			if (rxd->rx_dmamap != NULL) {
 				bus_dmamap_destroy(cd->sge_rxmbuf_tag,
-				    cd->sge_rx_map[i]);
-				cd->sge_rx_map[i] = NULL;
+				    rxd->rx_dmamap);
+				rxd->rx_dmamap = NULL;
 			}
 		}
 		if (cd->sge_rx_spare_map != NULL) {
@@ -951,10 +963,11 @@ sge_dma_free(struct sge_softc *sc)
 	/* Tx buffers. */
 	if (cd->sge_txmbuf_tag != NULL) {
 		for (i = 0; i < SGE_TX_RING_CNT; i++) {
-			if (cd->sge_tx_map[i] != NULL) {
+			txd = &cd->sge_txdesc[i];
+			if (txd->tx_dmamap != NULL) {
 				bus_dmamap_destroy(cd->sge_txmbuf_tag,
-				    cd->sge_tx_map[i]);
-				cd->sge_tx_map[i] = NULL;
+				    txd->tx_dmamap);
+				txd->tx_dmamap = NULL;
 			}
 		}
 		bus_dma_tag_destroy(cd->sge_txmbuf_tag);
@@ -991,18 +1004,20 @@ static int
 sge_list_tx_free(struct sge_softc *sc)
 {
 	struct sge_chain_data *cd;
+	struct sge_txdesc *txd;
 	int i;
 
 	SGE_LOCK_ASSERT(sc);
 	cd = &sc->sge_cdata;
 	for (i = 0; i < SGE_TX_RING_CNT; i++) {
-		if (cd->sge_tx_mbuf[i] != NULL) {
-			bus_dmamap_sync(cd->sge_txmbuf_tag,
-			    cd->sge_tx_map[i], BUS_DMASYNC_POSTWRITE);
-			bus_dmamap_unload(cd->sge_txmbuf_tag,
-			    cd->sge_tx_map[i]);
-			m_free(cd->sge_tx_mbuf[i]);
-			cd->sge_tx_mbuf[i] = NULL;
+		txd = &cd->sge_txdesc[i];
+		if (txd->tx_m != NULL) {
+			bus_dmamap_sync(cd->sge_txmbuf_tag, txd->tx_dmamap,
+			    BUS_DMASYNC_POSTWRITE);
+			bus_dmamap_unload(cd->sge_txmbuf_tag, txd->tx_dmamap);
+			m_free(txd->tx_m);
+			txd->tx_m = NULL;
+			txd->tx_ndesc = 0;
 		}
 	}
 
@@ -1037,18 +1052,20 @@ static int
 sge_list_rx_free(struct sge_softc *sc)
 {
 	struct sge_chain_data *cd;
+	struct sge_rxdesc *rxd;
 	int i;
 
 	SGE_LOCK_ASSERT(sc);
 	cd = &sc->sge_cdata;
 	for (i = 0; i < SGE_RX_RING_CNT; i++) {
-		if (cd->sge_rx_mbuf[i] != NULL) {
-			bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[i],
+		rxd = &cd->sge_rxdesc[i];
+		if (rxd->rx_m != NULL) {
+			bus_dmamap_sync(cd->sge_rxmbuf_tag, rxd->rx_dmamap,
 			    BUS_DMASYNC_POSTREAD);
 			bus_dmamap_unload(cd->sge_rxmbuf_tag,
-			    cd->sge_rx_map[i]);
-			m_free(cd->sge_rx_mbuf[i]);
-			cd->sge_rx_mbuf[i] = NULL;
+			    rxd->rx_dmamap);
+			m_free(rxd->rx_m);
+			rxd->rx_m = NULL;
 		}
 	}
 	return (0);
@@ -1063,6 +1080,7 @@ sge_newbuf(struct sge_softc *sc, int prod)
 	struct mbuf *m;
 	struct sge_desc *desc;
 	struct sge_chain_data *cd;
+	struct sge_rxdesc *rxd;
 	bus_dma_segment_t segs[1];
 	bus_dmamap_t map;
 	int error, nsegs;
@@ -1082,17 +1100,18 @@ sge_newbuf(struct sge_softc *sc, int prod)
 		return (error);
 	}
 	KASSERT(nsegs == 1, ("%s: %d segments returned!", __func__, nsegs));
-	if (cd->sge_rx_mbuf[prod] != NULL) {
-		bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod],
+	rxd = &cd->sge_rxdesc[prod];
+	if (rxd->rx_m != NULL) {
+		bus_dmamap_sync(cd->sge_rxmbuf_tag, rxd->rx_dmamap,
 		    BUS_DMASYNC_POSTREAD);
-		bus_dmamap_unload(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod]);
+		bus_dmamap_unload(cd->sge_rxmbuf_tag, rxd->rx_dmamap);
 	}
-	map = cd->sge_rx_map[prod];
-	cd->sge_rx_map[prod] =  cd->sge_rx_spare_map;
+	map = rxd->rx_dmamap;
+	rxd->rx_dmamap = cd->sge_rx_spare_map;
 	cd->sge_rx_spare_map = map;
-	bus_dmamap_sync(cd->sge_rxmbuf_tag, cd->sge_rx_map[prod],
+	bus_dmamap_sync(cd->sge_rxmbuf_tag, rxd->rx_dmamap,
 	    BUS_DMASYNC_PREREAD);
-	cd->sge_rx_mbuf[prod] = m;
+	rxd->rx_m = m;
 
 	desc = &sc->sge_ldata.sge_rx_ring[prod];
 	desc->sge_sts_size = 0;
@@ -1178,7 +1197,7 @@ sge_rxeof(struct sge_softc *sc)
 			ifp->if_ierrors++;
 			continue;
 		}
-		m = cd->sge_rx_mbuf[cons];
+		m = cd->sge_rxdesc[cons].rx_m;
 		if (sge_newbuf(sc, cons) != 0) {
 			sge_discard_rxbuf(sc, cons);
 			ifp->if_iqdrops++;
@@ -1245,8 +1264,9 @@ sge_txeof(struct sge_softc *sc)
 	struct ifnet *ifp;
 	struct sge_list_data *ld;
 	struct sge_chain_data *cd;
+	struct sge_txdesc *txd;
 	uint32_t txstat;
-	int cons, prod;
+	int cons, nsegs, prod;
 
 	SGE_LOCK_ASSERT(sc);
 
@@ -1260,33 +1280,47 @@ sge_txeof(struct sge_softc *sc)
 	    BUS_DMASYNC_POSTREAD | BUS_DMASYNC_POSTWRITE);
 	cons = cd->sge_tx_cons;
 	prod = cd->sge_tx_prod;
-	for (; cons != prod; SGE_INC(cons, SGE_TX_RING_CNT)) {
+	for (; cons != prod;) {
 		txstat = le32toh(ld->sge_tx_ring[cons].sge_cmdsts);
 		if ((txstat & TDC_OWN) != 0)
 			break;
-		cd->sge_tx_cnt--;
-		ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
-		if (cd->sge_tx_mbuf[cons] != NULL) {
-			bus_dmamap_sync(cd->sge_txmbuf_tag,
-			    cd->sge_tx_map[cons], BUS_DMASYNC_POSTWRITE);
-			bus_dmamap_unload(cd->sge_txmbuf_tag,
-			    cd->sge_tx_map[cons]);
-			m_freem(cd->sge_tx_mbuf[cons]);
-			cd->sge_tx_mbuf[cons] = NULL;
-			if (SGE_TX_ERROR(txstat) != 0) {
+		/*
+		 * Only the first descriptor of multi-descriptor transmission
+		 * is updated by controller.  Driver should skip entire
+		 * chained buffers for the transmitted frame. In other words
+		 * TDC_OWN bit is valid only at the first descriptor of a
+		 * multi-descriptor transmission.
+		 */
+		if (SGE_TX_ERROR(txstat) != 0) {
 #ifdef SGE_SHOW_ERRORS
-				device_printf(sc->sge_dev, "Tx error : 0x%b\n",
-				    txstat, TX_ERR_BITS);
+			device_printf(sc->sge_dev, "Tx error : 0x%b\n",
+			    txstat, TX_ERR_BITS);
 #endif
-				ifp->if_oerrors++;
-			} else {
+			ifp->if_oerrors++;
+		} else {
 #ifdef notyet
-				ifp->if_collisions += (txstat & 0xFFFF) - 1;
+			ifp->if_collisions += (txstat & 0xFFFF) - 1;
 #endif
-				ifp->if_opackets++;
-			}
+			ifp->if_opackets++;
 		}
-
+		txd = &cd->sge_txdesc[cons];
+		for (nsegs = 0; nsegs < txd->tx_ndesc; nsegs++) {
+			ld->sge_tx_ring[cons].sge_cmdsts = 0;
+			SGE_INC(cons, SGE_TX_RING_CNT);
+		}
+		/* Reclaim transmitted mbuf. */
+		KASSERT(txd->tx_m != NULL,
+		    ("%s: freeing NULL mbuf\n", __func__));
+		bus_dmamap_sync(cd->sge_txmbuf_tag, txd->tx_dmamap,
+		    BUS_DMASYNC_POSTWRITE);
+		bus_dmamap_unload(cd->sge_txmbuf_tag, txd->tx_dmamap);
+		m_freem(txd->tx_m);
+		txd->tx_m = NULL;
+		cd->sge_tx_cnt -= txd->tx_ndesc;
+		KASSERT(cd->sge_tx_cnt >= 0,
+		    ("%s: Active Tx desc counter was garbled\n", __func__));
+		txd->tx_ndesc = 0;
+		ifp->if_drv_flags &= ~IFF_DRV_OACTIVE;
 	}
 	cd->sge_tx_cons = cons;
 	if (cd->sge_tx_cnt == 0)
@@ -1388,73 +1422,78 @@ sge_encap(struct sge_softc *sc, struct mbuf **m_head)
 {
 	struct mbuf *m;
 	struct sge_desc *desc;
+	struct sge_txdesc *txd;
 	bus_dma_segment_t txsegs[SGE_MAXTXSEGS];
-	bus_dmamap_t map;
 	uint32_t cflags;
-	int error, nsegs, prod;
+	int error, i, nsegs, prod, si;
 
 	SGE_LOCK_ASSERT(sc);
 
-	prod = sc->sge_cdata.sge_tx_prod;
-	map = sc->sge_cdata.sge_tx_map[prod];
-	/*
-	 * Reading Windows inf file indicates SiS controller supports
-	 * TSO, VLAN hardware tag insertion/stripping, interrupt
-	 * moderation and Tx/Rx checksum offloading.  Unfortunately
-	 * vendor didn't release these information so we're guessing
-	 * descriptor usage with trial and errors.
-	 *
-	 * Controller seems to support multi-fragmented buffers but
-	 * don't know how to enable that feature so limit number of
-	 * fragmented Tx buffers to single buffer until we understand
-	 * the controller internals.
-	 * I assume the controller can pad zero bytes if frame length
-	 * is less than 60 bytes and I also think the controller has
-	 * no Tx buffer alignment limitation. - Need testing!
-	 */
-	if ((*m_head)->m_next != NULL) {
-		m = m_defrag(*m_head, M_DONTWAIT);
+	si = prod = sc->sge_cdata.sge_tx_prod;
+	txd = &sc->sge_cdata.sge_txdesc[prod];
+	error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_txmbuf_tag,
+	    txd->tx_dmamap, *m_head, txsegs, &nsegs, 0);
+	if (error == EFBIG) {
+		m = m_collapse(*m_head, M_DONTWAIT, SGE_MAXTXSEGS);
 		if (m == NULL) {
 			m_freem(*m_head);
 			*m_head = NULL;
 			return (ENOBUFS);
 		}
 		*m_head = m;
-	}
-	error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_txmbuf_tag, map,
-	    *m_head, txsegs, &nsegs, 0);
-	if (error != 0) {
-		m_freem(*m_head);
-		*m_head = NULL;
+		error = bus_dmamap_load_mbuf_sg(sc->sge_cdata.sge_txmbuf_tag,
+		    txd->tx_dmamap, *m_head, txsegs, &nsegs, 0);
+		if (error != 0) {
+			m_freem(*m_head);
+			*m_head = NULL;
+			return (error);
+		}
+	} else if (error != 0)
 		return (error);
-	}
+
+	KASSERT(nsegs != 0, ("zero segment returned"));
 	/* Check descriptor overrun. */
 	if (sc->sge_cdata.sge_tx_cnt + nsegs >= SGE_TX_RING_CNT) {
-		bus_dmamap_unload(sc->sge_cdata.sge_txmbuf_tag, map);
+		bus_dmamap_unload(sc->sge_cdata.sge_txmbuf_tag, txd->tx_dmamap);
 		return (ENOBUFS);
 	}
-	bus_dmamap_sync(sc->sge_cdata.sge_txmbuf_tag, map,
+	bus_dmamap_sync(sc->sge_cdata.sge_txmbuf_tag, txd->tx_dmamap,
 	    BUS_DMASYNC_PREWRITE);
 
+	m = *m_head;
 	cflags = 0;
-	if ((*m_head)->m_pkthdr.csum_flags & CSUM_IP)
+	if (m->m_pkthdr.csum_flags & CSUM_IP)
 		cflags |= TDC_IP_CSUM;
-	if ((*m_head)->m_pkthdr.csum_flags & CSUM_TCP)
+	if (m->m_pkthdr.csum_flags & CSUM_TCP)
 		cflags |= TDC_TCP_CSUM;
-	if ((*m_head)->m_pkthdr.csum_flags & CSUM_UDP)
+	if (m->m_pkthdr.csum_flags & CSUM_UDP)
 		cflags |= TDC_UDP_CSUM;
-	desc = &sc->sge_ldata.sge_tx_ring[prod];
-	desc->sge_sts_size = htole32((*m_head)->m_pkthdr.len);
-	desc->sge_ptr = htole32(SGE_ADDR_LO(txsegs[0].ds_addr));
-	desc->sge_flags = htole32(txsegs[0].ds_len);
-	if (prod == SGE_TX_RING_CNT - 1)
-		desc->sge_flags |= htole32(RING_END);
+	for (i = 0; i < nsegs; i++) {
+		desc = &sc->sge_ldata.sge_tx_ring[prod];
+		if (i == 0) {
+			desc->sge_sts_size = htole32(m->m_pkthdr.len);
+			desc->sge_cmdsts = 0;
+		} else {
+			desc->sge_sts_size = 0;
+			desc->sge_cmdsts = htole32(TDC_OWN);
+		}
+		desc->sge_ptr = htole32(SGE_ADDR_LO(txsegs[i].ds_addr));
+		desc->sge_flags = htole32(txsegs[i].ds_len);
+		if (prod == SGE_TX_RING_CNT - 1)
+			desc->sge_flags |= htole32(RING_END);
+		sc->sge_cdata.sge_tx_cnt++;
+		SGE_INC(prod, SGE_TX_RING_CNT);
+	}
+	/* Update producer index. */
+	sc->sge_cdata.sge_tx_prod = prod;
+
+	desc = &sc->sge_ldata.sge_tx_ring[si];
 	/* Configure VLAN. */
-	if(((*m_head)->m_flags & M_VLANTAG) != 0) {
-		cflags |= (*m_head)->m_pkthdr.ether_vtag;
+	if((m->m_flags & M_VLANTAG) != 0) {
+		cflags |= m->m_pkthdr.ether_vtag;
 		desc->sge_sts_size |= htole32(TDS_INS_VLAN);
 	}
-	desc->sge_cmdsts = htole32(TDC_DEF | TDC_CRC | TDC_PAD | cflags);
+	desc->sge_cmdsts |= htole32(TDC_DEF | TDC_CRC | TDC_PAD | cflags);
 #if 1
 	if ((sc->sge_flags & SGE_FLAG_SPEED_1000) != 0)
 		desc->sge_cmdsts |= htole32(TDC_BST);
@@ -1466,13 +1505,9 @@ sge_encap(struct sge_softc *sc, struct mbuf **m_head)
 	}
 #endif
 	/* Request interrupt and give ownership to controller. */
-	if ((prod % SGE_TX_INTR_FRAMES) == 0)
-		desc->sge_cmdsts |= htole32(TDC_OWN | TDC_INTR);
-	else
-		desc->sge_cmdsts |= htole32(TDC_OWN);
-	sc->sge_cdata.sge_tx_mbuf[prod] = *m_head;
-	sc->sge_cdata.sge_tx_cnt++;
-	SGE_INC(sc->sge_cdata.sge_tx_prod, SGE_TX_RING_CNT);
+	desc->sge_cmdsts |= htole32(TDC_OWN | TDC_INTR);
+	txd->tx_m = m;
+	txd->tx_ndesc = nsegs;
 	return (0);
 }
 
@@ -1503,7 +1538,8 @@ sge_start_locked(struct ifnet *ifp)
 		return;
 
 	for (queued = 0; !IFQ_DRV_IS_EMPTY(&ifp->if_snd); ) {
-		if (sc->sge_cdata.sge_tx_cnt == SGE_TX_RING_CNT - 1) {
+		if (sc->sge_cdata.sge_tx_cnt > (SGE_TX_RING_CNT -
+		    SGE_MAXTXSEGS)) {
 			ifp->if_drv_flags |= IFF_DRV_OACTIVE;
 			break;
 		}
diff --git a/sys/dev/sge/if_sgereg.h b/sys/dev/sge/if_sgereg.h
index 722dd3c490d..c06072cdeb9 100644
--- a/sys/dev/sge/if_sgereg.h
+++ b/sys/dev/sge/if_sgereg.h
@@ -283,7 +283,7 @@ struct sge_desc {
 #define	SGE_RX_RING_CNT		256 /* [8, 1024] */
 #define	SGE_TX_RING_CNT		256 /* [8, 8192] */
 #define	SGE_DESC_ALIGN		16
-#define	SGE_MAXTXSEGS		1
+#define	SGE_MAXTXSEGS		16
 #define	SGE_RX_BUF_ALIGN	sizeof(uint64_t)
 
 #define	SGE_RX_RING_SZ		(SGE_RX_RING_CNT * sizeof(struct sge_desc))
@@ -298,6 +298,17 @@ struct sge_list_data {
 	bus_addr_t		sge_tx_paddr;
 };
 
+struct sge_txdesc {
+	struct mbuf		*tx_m;
+	bus_dmamap_t		tx_dmamap;
+	int			tx_ndesc;
+};
+
+struct sge_rxdesc {
+	struct mbuf		*rx_m;
+	bus_dmamap_t		rx_dmamap;
+};
+
 struct sge_chain_data {
 	bus_dma_tag_t		sge_tag;
 	bus_dma_tag_t		sge_rx_tag;
@@ -306,11 +317,9 @@ struct sge_chain_data {
 	bus_dmamap_t		sge_tx_dmamap;
 	bus_dma_tag_t		sge_txmbuf_tag;
 	bus_dma_tag_t		sge_rxmbuf_tag;
-	struct mbuf		*sge_rx_mbuf[SGE_RX_RING_CNT];
-	struct mbuf		*sge_tx_mbuf[SGE_TX_RING_CNT];
-	bus_dmamap_t		sge_rx_map[SGE_RX_RING_CNT];
+	struct sge_txdesc	sge_txdesc[SGE_TX_RING_CNT];
+	struct sge_rxdesc	sge_rxdesc[SGE_RX_RING_CNT];
 	bus_dmamap_t		sge_rx_spare_map;
-	bus_dmamap_t		sge_tx_map[SGE_TX_RING_CNT];
 	int			sge_rx_cons;
 	int			sge_tx_prod;
 	int			sge_tx_cons;

From 54934285c76c26be9eba25128ae8260ddd367147 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Tue, 4 May 2010 19:58:55 +0000
Subject: [PATCH 459/532] Remove if_ar, if_ray, if_sr, if_ppp, if_sl to reflect
 the current modules available, they were removed due to NEEDSGIANT.

While I'm there, add if_et which was missed quite a while ago.

MFC after:	2 weeks
---
 sys/boot/forth/loader.conf | 8 ++------
 1 file changed, 2 insertions(+), 6 deletions(-)

diff --git a/sys/boot/forth/loader.conf b/sys/boot/forth/loader.conf
index e0e4409f03b..d48d8c44529 100644
--- a/sys/boot/forth/loader.conf
+++ b/sys/boot/forth/loader.conf
@@ -197,8 +197,6 @@ if_epair_load="NO"		# Virtual b-t-b Ethernet-like interface pair
 if_faith_load="NO"		# IPv6-to-IPv4 TCP relay capturing interface
 if_gif_load="NO"		# generic tunnel interface
 if_gre_load="NO"		# encapsulating network device
-if_ppp_load="NO"		# Kernel ppp
-if_sl_load="NO"			# SLIP
 if_stf_load="NO"		# 6to4 tunnel interface
 if_tap_load="NO"		# Ethernet tunnel software network interface
 if_tun_load="NO"		# Tunnel driver (user process ppp)
@@ -217,7 +215,6 @@ if_age_load="NO"		# Attansic/Atheros L1 Gigabit Ethernet
 if_alc_load="NO"		# Atheros AR8131/AR8132 Ethernet
 if_ale_load="NO"		# Atheros AR8121/AR8113/AR8114 Ethernet
 if_an_load="NO"			# Aironet 4500/4800 802.11 wireless NICs
-if_ar_load="NO"			# Digi SYNC/570i
 if_ath_load="NO"		# Atheros IEEE 802.11 wireless NICs
 if_aue_load="NO"		# ADMtek AN986 Pegasus USB Ethernet
 if_awi_load="NO"		# AMD PCnetMobile IEEE 802.11 wireless NICs
@@ -237,6 +234,7 @@ if_ed_load="NO"			# National Semiconductor DS8390/WD83C690
 if_em_load="NO"			# Intel(R) PRO/1000 Gigabit Ethernet
 if_en_load="NO"			# Midway-based ATM interfaces
 if_ep_load="NO"			# 3Com Etherlink III (3c5x9)
+if_et_load="NO"			# Agere ET1310 10/100/Gigabit Ethernet
 if_ex_load="NO"			# Intel EtherExpress Pro/10 Ethernet
 if_fe_load="NO"			# Fujitsu MB86960A/MB86965A based Ethernet
 				# adapters
@@ -265,17 +263,15 @@ if_nve_load="NO"		# NVIDIA nForce MCP Networking Adapter
 if_nxge_load="NO"		# Neterion Xframe 10Gb Ethernet
 if_pcn_load="NO"		# AMD PCnet PCI
 if_ral_load="NO"		# Ralink Technology wireless
-if_ray_load="NO"		# Raytheon Raylink/Webgear Aviator PCCard
 if_re_load="NO"			# RealTek 8139C+/8169/8169S/8110S
 if_rl_load="NO"			# RealTek 8129/8139
 if_rue_load="NO"		# RealTek RTL8150 USB to Fast Ethernet
 if_sbni_load="NO"		# Granch SBNI12 leased line adapters
 if_sf_load="NO"			# Adaptec Duralink PCI (AIC-6915 "starfire")
-if_sge_load="NO"		# Silicon Integrated Systems SiS190/191
+if_sge_load="NO"		# Silicon Integrated Systems SiS 190/191
 if_sis_load="NO"		# Silicon Integrated Systems SiS 900/7016
 if_sk_load="NO"			# SysKonnect SK-984x series PCI Gigabit Ethernet
 if_sn_load="NO"			# SMC 91Cxx
-if_sr_load="NO"			# synchronous RISCom/N2 / WANic 400/405
 if_ste_load="NO"		# Sundance Technologies ST201 Fast Ethernet
 if_stge_load="NO"		# Sundance/Tamarack TC9021 Gigabit Ethernet
 if_ti_load="NO"			# Alteon Networks Tigon 1 and Tigon 2

From f648f6bb0c3ddce0e9e244d0dd61e71143aa63a1 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 21:23:59 +0000
Subject: [PATCH 460/532] Free entire mbuf chain instead of the first mbuf.

---
 sys/dev/sge/if_sge.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/sge/if_sge.c b/sys/dev/sge/if_sge.c
index c3849832650..ee88cb6fb1d 100644
--- a/sys/dev/sge/if_sge.c
+++ b/sys/dev/sge/if_sge.c
@@ -1015,7 +1015,7 @@ sge_list_tx_free(struct sge_softc *sc)
 			bus_dmamap_sync(cd->sge_txmbuf_tag, txd->tx_dmamap,
 			    BUS_DMASYNC_POSTWRITE);
 			bus_dmamap_unload(cd->sge_txmbuf_tag, txd->tx_dmamap);
-			m_free(txd->tx_m);
+			m_freem(txd->tx_m);
 			txd->tx_m = NULL;
 			txd->tx_ndesc = 0;
 		}
@@ -1064,7 +1064,7 @@ sge_list_rx_free(struct sge_softc *sc)
 			    BUS_DMASYNC_POSTREAD);
 			bus_dmamap_unload(cd->sge_rxmbuf_tag,
 			    rxd->rx_dmamap);
-			m_free(rxd->rx_m);
+			m_freem(rxd->rx_m);
 			rxd->rx_m = NULL;
 		}
 	}

From 3edfecaa37cdf0d4e11b819bc27056d6f37e9579 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Tue, 4 May 2010 22:24:19 +0000
Subject: [PATCH 461/532] When VLAN hardware tagging is disabled, make sure to
 disable VLAN checksum offloading as well as TSO over VLAN.

Reported by:	jhb
---
 sys/dev/msk/if_msk.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/sys/dev/msk/if_msk.c b/sys/dev/msk/if_msk.c
index af977bfecf6..3f97c3a9f5c 100644
--- a/sys/dev/msk/if_msk.c
+++ b/sys/dev/msk/if_msk.c
@@ -1101,7 +1101,8 @@ msk_ioctl(struct ifnet *ifp, u_long command, caddr_t data)
 		    (IFCAP_VLAN_HWTAGGING & ifp->if_capabilities) != 0) {
 			ifp->if_capenable ^= IFCAP_VLAN_HWTAGGING;
 			if ((IFCAP_VLAN_HWTAGGING & ifp->if_capenable) == 0)
-				ifp->if_capenable &= ~IFCAP_VLAN_HWTSO;
+				ifp->if_capenable &=
+				    ~(IFCAP_VLAN_HWTSO | IFCAP_VLAN_HWCSUM);
 			msk_setvlan(sc_if, ifp);
 		}
 		if (ifp->if_mtu > ETHERMTU &&

From 2caefebb079b6bd13ffd683214b3c10eddce830f Mon Sep 17 00:00:00 2001
From: Navdeep Parhar <np@FreeBSD.org>
Date: Tue, 4 May 2010 23:55:08 +0000
Subject: [PATCH 462/532] Add IFCAP_LINKSTATE to cxgb's capabilities.

MFC after:	3 days
---
 sys/dev/cxgb/cxgb_main.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/cxgb/cxgb_main.c b/sys/dev/cxgb/cxgb_main.c
index a805b5d9826..01095c4e0b7 100644
--- a/sys/dev/cxgb/cxgb_main.c
+++ b/sys/dev/cxgb/cxgb_main.c
@@ -981,7 +981,7 @@ cxgb_makedev(struct port_info *pi)
 
 #define CXGB_CAP (IFCAP_VLAN_HWTAGGING | IFCAP_VLAN_MTU | IFCAP_HWCSUM | \
     IFCAP_VLAN_HWCSUM | IFCAP_TSO | IFCAP_JUMBO_MTU | IFCAP_LRO | \
-    IFCAP_VLAN_HWTSO)
+    IFCAP_VLAN_HWTSO | IFCAP_LINKSTATE)
 #define CXGB_CAP_ENABLE (CXGB_CAP & ~IFCAP_TSO6)
 
 static int

From d6da8362017ea667d5d92d90ad3fe01b16c2cbe4 Mon Sep 17 00:00:00 2001
From: Navdeep Parhar <np@FreeBSD.org>
Date: Wed, 5 May 2010 00:41:40 +0000
Subject: [PATCH 463/532] Add support for hardware filters to cxgb(4).  The T3
 chip can inspect L2/3/4 headers and can drop or steer packets as instructed. 
 Filtering based on src ip, dst ip, src port, dst port, 802.1q, udp/tcp, and
 mac addr is possible.  Add support in cxgbtool to program these filters. Some
 simple examples:

Drop all tcp/80 traffic coming from the subnet specified.
# cxgbtool cxgb2 filter 0 sip 192.168.1.0/24 dport 80 type tcp action drop

Steer all incoming UDP traffic to qset 0.
# cxgbtool cxgb2 filter 1 type udp queue 0 action pass

Steer all tcp traffic from 192.168.1.1 to qset 1.
# cxgbtool cxgb2 filter 2 sip 192.168.1.1 type tcp queue 1 action pass

Drop fragments.
# cxgbtool cxgb2 filter 3 type frag action drop

List all filters.
# cxgbtool cxgb2 filter list
index         SIP                DIP     sport dport VLAN PRI P/MAC type Q
    0     192.168.1.0/24         0.0.0.0     *    80    0 0/1 */*    tcp -
    1         0.0.0.0/0          0.0.0.0     *     *    0 0/1 */*    udp 0
    2     192.168.1.1/32         0.0.0.0     *     *    0 0/1 */*    tcp 1
    3         0.0.0.0/0          0.0.0.0     *     *    0 0/1 */*   frag -
16367         0.0.0.0/0          0.0.0.0     *     *    0 0/1 */*      * *

MFC after:	2 weeks
---
 sys/dev/cxgb/cxgb_ioctl.h    |  29 ++++
 sys/dev/cxgb/cxgb_main.c     | 273 +++++++++++++++++++++++++++++++++++
 usr.sbin/cxgbtool/cxgbtool.c | 186 +++++++++++++++++++++++-
 3 files changed, 482 insertions(+), 6 deletions(-)

diff --git a/sys/dev/cxgb/cxgb_ioctl.h b/sys/dev/cxgb/cxgb_ioctl.h
index e4b487505a0..3f3ac5a1866 100644
--- a/sys/dev/cxgb/cxgb_ioctl.h
+++ b/sys/dev/cxgb/cxgb_ioctl.h
@@ -59,6 +59,9 @@ enum {
 	CH_CLEAR_STATS,
 	CH_GET_UP_LA,
 	CH_GET_UP_IOQS,
+	CH_SET_FILTER,
+	CH_DEL_FILTER,
+	CH_GET_FILTER,
 };
 
 /* statistics categories */
@@ -215,6 +218,29 @@ struct ch_up_ioqs {
 	struct t3_ioq_entry *data;
 };
 
+struct ch_filter_tuple {
+	uint32_t sip;
+	uint32_t dip;
+	uint16_t sport;
+	uint16_t dport;
+	uint16_t vlan:12;
+	uint16_t vlan_prio:3;
+};
+
+struct ch_filter {
+	uint32_t filter_id;
+	struct ch_filter_tuple val;
+	struct ch_filter_tuple mask;
+	uint16_t mac_addr_idx;
+	uint8_t mac_hit:1;
+	uint8_t proto:2;
+
+	uint8_t want_filter_id:1;
+	uint8_t pass:1;
+	uint8_t rss:1;
+	uint8_t qset;
+};
+
 #define CHELSIO_SETREG		_IOW('f', CH_SETREG, struct ch_reg)
 #define CHELSIO_GETREG		_IOWR('f', CH_GETREG, struct ch_reg)
 #define CHELSIO_GETMTUTAB	_IOR('f', CH_GETMTUTAB, struct ch_mtus)
@@ -239,4 +265,7 @@ struct ch_up_ioqs {
 #define CHELSIO_GET_EEPROM	_IOWR('f', CH_GET_EEPROM, struct ch_eeprom)
 #define CHELSIO_GET_UP_LA	_IOWR('f', CH_GET_UP_LA, struct ch_up_la)
 #define CHELSIO_GET_UP_IOQS	_IOWR('f', CH_GET_UP_IOQS, struct ch_up_ioqs)
+#define CHELSIO_SET_FILTER	_IOW('f', CH_SET_FILTER, struct ch_filter)
+#define CHELSIO_DEL_FILTER	_IOW('f', CH_DEL_FILTER, struct ch_filter)
+#define CHELSIO_GET_FILTER	_IOWR('f', CH_GET_FILTER, struct ch_filter)
 #endif
diff --git a/sys/dev/cxgb/cxgb_main.c b/sys/dev/cxgb/cxgb_main.c
index 01095c4e0b7..ad2caed218f 100644
--- a/sys/dev/cxgb/cxgb_main.c
+++ b/sys/dev/cxgb/cxgb_main.c
@@ -99,6 +99,13 @@ static void cxgb_ext_intr_handler(void *, int);
 static void cxgb_tick_handler(void *, int);
 static void cxgb_tick(void *);
 static void setup_rss(adapter_t *sc);
+static int alloc_filters(struct adapter *);
+static int setup_hw_filters(struct adapter *);
+static int set_filter(struct adapter *, int, const struct filter_info *);
+static inline void mk_set_tcb_field(struct cpl_set_tcb_field *, unsigned int,
+    unsigned int, u64, u64);
+static inline void set_tcb_field_ulp(struct cpl_set_tcb_field *, unsigned int,
+    unsigned int, u64, u64);
 
 /* Attachment glue for the PCI controller end of the device.  Each port of
  * the device is attached separately, as defined later.
@@ -1661,6 +1668,13 @@ cxgb_up(struct adapter *sc)
 			if ((err = update_tpsram(sc)))
 				goto out;
 
+		if (is_offload(sc)) {
+			sc->params.mc5.nservers = 0;
+			sc->params.mc5.nroutes = 0;
+			sc->params.mc5.nfilters = t3_mc5_size(&sc->mc5) -
+			    MC5_MIN_TIDS;
+		}
+
 		err = t3_init_hw(sc, 0);
 		if (err)
 			goto out;
@@ -1672,6 +1686,7 @@ cxgb_up(struct adapter *sc)
 		if (err)
 			goto out;
 
+		alloc_filters(sc);
 		setup_rss(sc);
 
 		t3_intr_clear(sc);
@@ -1698,6 +1713,7 @@ cxgb_up(struct adapter *sc)
 	
 	if (!(sc->flags & QUEUES_BOUND)) {
 		bind_qsets(sc);
+		setup_hw_filters(sc);
 		sc->flags |= QUEUES_BOUND;		
 	}
 
@@ -3076,6 +3092,139 @@ cxgb_extension_ioctl(struct cdev *dev, unsigned long cmd, caddr_t data,
 		free(buf, M_DEVBUF);
 		break;
 	}
+	case CHELSIO_SET_FILTER: {
+		struct ch_filter *f = (struct ch_filter *)data;;
+		struct filter_info *p;
+		unsigned int nfilters = sc->params.mc5.nfilters;
+
+		if (!is_offload(sc))
+			return (EOPNOTSUPP);	/* No TCAM */
+		if (!(sc->flags & FULL_INIT_DONE))
+			return (EAGAIN);	/* mc5 not setup yet */
+		if (nfilters == 0)
+			return (EBUSY);		/* TOE will use TCAM */
+
+		/* sanity checks */
+		if (f->filter_id >= nfilters ||
+		    (f->val.dip && f->mask.dip != 0xffffffff) ||
+		    (f->val.sport && f->mask.sport != 0xffff) ||
+		    (f->val.dport && f->mask.dport != 0xffff) ||
+		    (f->val.vlan && f->mask.vlan != 0xfff) ||
+		    (f->val.vlan_prio &&
+			f->mask.vlan_prio != FILTER_NO_VLAN_PRI) ||
+		    (f->mac_addr_idx != 0xffff && f->mac_addr_idx > 15) ||
+		    f->qset >= SGE_QSETS ||
+		    sc->rrss_map[f->qset] >= RSS_TABLE_SIZE)
+			return (EINVAL);
+
+		/* Was allocated with M_WAITOK */
+		KASSERT(sc->filters, ("filter table NULL\n"));
+
+		p = &sc->filters[f->filter_id];
+		if (p->locked)
+			return (EPERM);
+
+		bzero(p, sizeof(*p));
+		p->sip = f->val.sip;
+		p->sip_mask = f->mask.sip;
+		p->dip = f->val.dip;
+		p->sport = f->val.sport;
+		p->dport = f->val.dport;
+		p->vlan = f->mask.vlan ? f->val.vlan : 0xfff;
+		p->vlan_prio = f->mask.vlan_prio ? (f->val.vlan_prio & 6) :
+		    FILTER_NO_VLAN_PRI;
+		p->mac_hit = f->mac_hit;
+		p->mac_vld = f->mac_addr_idx != 0xffff;
+		p->mac_idx = f->mac_addr_idx;
+		p->pkt_type = f->proto;
+		p->report_filter_id = f->want_filter_id;
+		p->pass = f->pass;
+		p->rss = f->rss;
+		p->qset = f->qset;
+
+		error = set_filter(sc, f->filter_id, p);
+		if (error == 0)
+			p->valid = 1;
+		break;
+	}
+	case CHELSIO_DEL_FILTER: {
+		struct ch_filter *f = (struct ch_filter *)data;
+		struct filter_info *p;
+		unsigned int nfilters = sc->params.mc5.nfilters;
+
+		if (!is_offload(sc))
+			return (EOPNOTSUPP);
+		if (!(sc->flags & FULL_INIT_DONE))
+			return (EAGAIN);
+		if (nfilters == 0 || sc->filters == NULL)
+			return (EINVAL);
+		if (f->filter_id >= nfilters)
+		       return (EINVAL);
+
+		p = &sc->filters[f->filter_id];
+		if (p->locked)
+			return (EPERM);
+		if (!p->valid)
+			return (EFAULT); /* Read "Bad address" as "Bad index" */
+
+		bzero(p, sizeof(*p));
+		p->sip = p->sip_mask = 0xffffffff;
+		p->vlan = 0xfff;
+		p->vlan_prio = FILTER_NO_VLAN_PRI;
+		p->pkt_type = 1;
+		error = set_filter(sc, f->filter_id, p);
+		break;
+	}
+	case CHELSIO_GET_FILTER: {
+		struct ch_filter *f = (struct ch_filter *)data;
+		struct filter_info *p;
+		unsigned int i, nfilters = sc->params.mc5.nfilters;
+
+		if (!is_offload(sc))
+			return (EOPNOTSUPP);
+		if (!(sc->flags & FULL_INIT_DONE))
+			return (EAGAIN);
+		if (nfilters == 0 || sc->filters == NULL)
+			return (EINVAL);
+
+		i = f->filter_id == 0xffffffff ? 0 : f->filter_id + 1;
+		for (; i < nfilters; i++) {
+			p = &sc->filters[i];
+			if (!p->valid)
+				continue;
+
+			bzero(f, sizeof(*f));
+
+			f->filter_id = i;
+			f->val.sip = p->sip;
+			f->mask.sip = p->sip_mask;
+			f->val.dip = p->dip;
+			f->mask.dip = p->dip ? 0xffffffff : 0;
+			f->val.sport = p->sport;
+			f->mask.sport = p->sport ? 0xffff : 0;
+			f->val.dport = p->dport;
+			f->mask.dport = p->dport ? 0xffff : 0;
+			f->val.vlan = p->vlan == 0xfff ? 0 : p->vlan;
+			f->mask.vlan = p->vlan == 0xfff ? 0 : 0xfff;
+			f->val.vlan_prio = p->vlan_prio == FILTER_NO_VLAN_PRI ?
+			    0 : p->vlan_prio;
+			f->mask.vlan_prio = p->vlan_prio == FILTER_NO_VLAN_PRI ?
+			    0 : FILTER_NO_VLAN_PRI;
+			f->mac_hit = p->mac_hit;
+			f->mac_addr_idx = p->mac_vld ? p->mac_idx : 0xffff;
+			f->proto = p->pkt_type;
+			f->want_filter_id = p->report_filter_id;
+			f->pass = p->pass;
+			f->rss = p->rss;
+			f->qset = p->qset;
+
+			break;
+		}
+		
+		if (i == nfilters)
+			f->filter_id = 0xffffffff;
+		break;
+	}
 	default:
 		return (EOPNOTSUPP);
 		break;
@@ -3130,5 +3279,129 @@ cxgb_get_regs(adapter_t *sc, struct ch_ifconf_regs *regs, uint8_t *buf)
 		       XGM_REG(A_XGM_RX_SPI4_SOP_EOP_CNT, 1));
 }
 
+static int
+alloc_filters(struct adapter *sc)
+{
+	struct filter_info *p;
+	unsigned int nfilters = sc->params.mc5.nfilters;
+
+	if (nfilters == 0)
+		return (0);
+
+	p = malloc(sizeof(*p) * nfilters, M_DEVBUF, M_WAITOK | M_ZERO);
+	sc->filters = p;
+
+	p = &sc->filters[nfilters - 1];
+	p->vlan = 0xfff;
+	p->vlan_prio = FILTER_NO_VLAN_PRI;
+	p->pass = p->rss = p->valid = p->locked = 1;
+
+	return (0);
+}
+
+static int
+setup_hw_filters(struct adapter *sc)
+{
+	int i, rc;
+	unsigned int nfilters = sc->params.mc5.nfilters;
+
+	if (!sc->filters)
+		return (0);
+
+	t3_enable_filters(sc);
+
+	for (i = rc = 0; i < nfilters && !rc; i++) {
+		if (sc->filters[i].locked)
+			rc = set_filter(sc, i, &sc->filters[i]);
+	}
+
+	return (rc);
+}
+
+static int
+set_filter(struct adapter *sc, int id, const struct filter_info *f)
+{
+	int len;
+	struct mbuf *m;
+	struct ulp_txpkt *txpkt;
+	struct work_request_hdr *wr;
+	struct cpl_pass_open_req *oreq;
+	struct cpl_set_tcb_field *sreq;
+
+	len = sizeof(*wr) + sizeof(*oreq) + 2 * sizeof(*sreq);
+	KASSERT(len <= MHLEN, ("filter request too big for an mbuf"));
+
+	id += t3_mc5_size(&sc->mc5) - sc->params.mc5.nroutes -
+	      sc->params.mc5.nfilters;
+
+	m = m_gethdr(M_WAITOK, MT_DATA);
+	m->m_len = m->m_pkthdr.len = len;
+	bzero(mtod(m, char *), len);
+
+	wr = mtod(m, struct work_request_hdr *);
+	wr->wrh_hi = htonl(V_WR_OP(FW_WROPCODE_BYPASS) | F_WR_ATOMIC);
+
+	oreq = (struct cpl_pass_open_req *)(wr + 1);
+	txpkt = (struct ulp_txpkt *)oreq;
+	txpkt->cmd_dest = htonl(V_ULPTX_CMD(ULP_TXPKT));
+	txpkt->len = htonl(V_ULPTX_NFLITS(sizeof(*oreq) / 8));
+	OPCODE_TID(oreq) = htonl(MK_OPCODE_TID(CPL_PASS_OPEN_REQ, id));
+	oreq->local_port = htons(f->dport);
+	oreq->peer_port = htons(f->sport);
+	oreq->local_ip = htonl(f->dip);
+	oreq->peer_ip = htonl(f->sip);
+	oreq->peer_netmask = htonl(f->sip_mask);
+	oreq->opt0h = 0;
+	oreq->opt0l = htonl(F_NO_OFFLOAD);
+	oreq->opt1 = htonl(V_MAC_MATCH_VALID(f->mac_vld) |
+			 V_CONN_POLICY(CPL_CONN_POLICY_FILTER) |
+			 V_VLAN_PRI(f->vlan_prio >> 1) |
+			 V_VLAN_PRI_VALID(f->vlan_prio != FILTER_NO_VLAN_PRI) |
+			 V_PKT_TYPE(f->pkt_type) | V_OPT1_VLAN(f->vlan) |
+			 V_MAC_MATCH(f->mac_idx | (f->mac_hit << 4)));
+
+	sreq = (struct cpl_set_tcb_field *)(oreq + 1);
+	set_tcb_field_ulp(sreq, id, 1, 0x1800808000ULL,
+			  (f->report_filter_id << 15) | (1 << 23) |
+			  ((u64)f->pass << 35) | ((u64)!f->rss << 36));
+	set_tcb_field_ulp(sreq + 1, id, 0, 0xffffffff, (2 << 19) | 1);
+	t3_mgmt_tx(sc, m);
+
+	if (f->pass && !f->rss) {
+		len = sizeof(*sreq);
+		m = m_gethdr(M_WAITOK, MT_DATA);
+		m->m_len = m->m_pkthdr.len = len;
+		bzero(mtod(m, char *), len);
+		sreq = mtod(m, struct cpl_set_tcb_field *);
+		sreq->wr.wrh_hi = htonl(V_WR_OP(FW_WROPCODE_FORWARD));
+		mk_set_tcb_field(sreq, id, 25, 0x3f80000,
+				 (u64)sc->rrss_map[f->qset] << 19);
+		t3_mgmt_tx(sc, m);
+	}
+	return 0;
+}
+
+static inline void
+mk_set_tcb_field(struct cpl_set_tcb_field *req, unsigned int tid,
+    unsigned int word, u64 mask, u64 val)
+{
+	OPCODE_TID(req) = htonl(MK_OPCODE_TID(CPL_SET_TCB_FIELD, tid));
+	req->reply = V_NO_REPLY(1);
+	req->cpu_idx = 0;
+	req->word = htons(word);
+	req->mask = htobe64(mask);
+	req->val = htobe64(val);
+}
+
+static inline void
+set_tcb_field_ulp(struct cpl_set_tcb_field *req, unsigned int tid,
+    unsigned int word, u64 mask, u64 val)
+{
+	struct ulp_txpkt *txpkt = (struct ulp_txpkt *)req;
+
+	txpkt->cmd_dest = htonl(V_ULPTX_CMD(ULP_TXPKT));
+	txpkt->len = htonl(V_ULPTX_NFLITS(sizeof(*req) / 8));
+	mk_set_tcb_field(req, tid, word, mask, val);
+}
 
 MODULE_DEPEND(if_cxgb, cxgb_t3fw, 1, 1, 1);
diff --git a/usr.sbin/cxgbtool/cxgbtool.c b/usr.sbin/cxgbtool/cxgbtool.c
index 28d91f96dfe..b705ff88986 100644
--- a/usr.sbin/cxgbtool/cxgbtool.c
+++ b/usr.sbin/cxgbtool/cxgbtool.c
@@ -1,6 +1,6 @@
 /**************************************************************************
 
-Copyright (c) 2007-2009, Chelsio Inc.
+Copyright (c) 2007-2010, Chelsio Inc.
 All rights reserved.
 
 Redistribution and use in source and binary forms, with or without
@@ -93,6 +93,9 @@ usage(FILE *fp)
 	    	"\tclearstats                          clear MAC statistics\n"
 		"\tcontext <type> <id>                 show an SGE context\n"
 		"\tdesc <qset> <queue> <idx> [<cnt>]   dump SGE descriptors\n"
+		"\tfilter <idx> [<param> <val>] ...    set a filter\n"
+		"\tfilter <idx> delete|clear           delete a filter\n"
+		"\tfilter list                         list all filters\n"
 		"\tioqs                                dump uP IOQs\n"
 		"\tla                                  dump uP logic analyzer info\n"
 		"\tloadboot <boot image>               download boot image\n"
@@ -1177,24 +1180,25 @@ parse_ipaddr(const char *s, uint32_t *addr, uint32_t *mask)
  * Parse a string containing a value and an optional colon separated mask.
  */
 static int
-parse_val_mask_param(const char *s, uint32_t *val, uint32_t *mask)
+parse_val_mask_param(const char *s, uint32_t *val, uint32_t *mask,
+    uint32_t default_mask)
 {
 	char *p;
 
-	*mask = 0xffffffffU;
+	*mask = default_mask;
 	*val = strtoul(s, &p, 0);
-	if (p == s)
+	if (p == s || *val > default_mask)
 		return -1;
 	if (*p == ':' && p[1])
 		*mask = strtoul(p + 1, &p, 0);
-	return *p ? -1 : 0;
+	return *p || *mask > default_mask ? -1 : 0;
 }
 
 static int
 parse_trace_param(const char *s, uint32_t *val, uint32_t *mask)
 {
 	return strchr(s, '.') ? parse_ipaddr(s, val, mask) :
-				parse_val_mask_param(s, val, mask);
+				parse_val_mask_param(s, val, mask, 0xffffffffU);
 }
 
 static int
@@ -1273,6 +1277,174 @@ trace_config(int argc, char *argv[], int start_arg, const char *iff_name)
 	return 0;
 }
 
+static void
+show_filters(const char *iff_name)
+{
+	static const char *pkt_type[] = { "*", "tcp", "udp", "frag" };
+	struct ch_filter op;
+	union {
+		uint32_t nip;
+		uint8_t octet[4];
+	} nsip, ndip;
+	char sip[20], dip[20];
+	int header = 0;
+
+	bzero(&op, sizeof(op));
+	op.filter_id = 0xffffffff;
+
+	do {
+		if (doit(iff_name, CHELSIO_GET_FILTER, &op) < 0)
+			err(1, "list filters");
+
+		if (op.filter_id == 0xffffffff)
+			break;
+
+		if (!header) {
+			printf("index         SIP                DIP     sport "
+			    "dport VLAN PRI P/MAC type Q\n");
+			header = 1;
+		}
+
+		nsip.nip = htonl(op.val.sip);
+		ndip.nip = htonl(op.val.dip);
+
+		sprintf(sip, "%u.%u.%u.%u/%-2u", nsip.octet[0], nsip.octet[1],
+		    nsip.octet[2], nsip.octet[3],
+		    op.mask.sip ? 33 - ffs(op.mask.sip) : 0);
+		sprintf(dip, "%u.%u.%u.%u", ndip.octet[0], ndip.octet[1],
+		    ndip.octet[2], ndip.octet[3]);
+		printf("%5zu %18s %15s ", (size_t)op.filter_id, sip, dip);
+		printf(op.val.sport ? "%5u " : "    * ", op.val.sport);
+		printf(op.val.dport ? "%5u " : "    * ", op.val.dport);
+		printf(op.val.vlan != 0xfff ? "%4u " : "   * ", op.val.vlan);
+		printf(op.val.vlan_prio == 7 ?  "  * " :
+		    "%1u/%1u ", op.val.vlan_prio, op.val.vlan_prio | 1);
+		if (op.mac_addr_idx == 0xffff)
+			printf("*/*   ");
+		else if (op.mac_hit)
+			printf("%1u/%3u ", (op.mac_addr_idx >> 3) & 0x1,
+			    (op.mac_addr_idx) & 0x7);
+		else
+			printf("%1u/  * ", (op.mac_addr_idx >> 3) & 0x1);
+		printf("%4s ", pkt_type[op.proto]);
+		if (!op.pass)
+			printf("-\n");
+		else if (op.rss)
+			printf("*\n");
+		else
+			printf("%1u\n", op.qset);
+	} while (1);
+}
+
+static int
+filter_config(int argc, char *argv[], int start_arg, const char *iff_name)
+{
+	int ret = 0;
+	uint32_t val, mask;
+	struct ch_filter op;
+
+	if (argc < start_arg + 1)
+		return -1;
+
+	memset(&op, 0, sizeof(op));
+	op.mac_addr_idx = 0xffff;
+	op.rss = 1;
+
+	if (argc == start_arg + 1 && !strcmp(argv[start_arg], "list")) {
+		show_filters(iff_name);
+		return 0;
+	}
+
+	if (get_int_arg(argv[start_arg++], &op.filter_id))
+		return -1;
+	if (argc == start_arg + 1 && (!strcmp(argv[start_arg], "delete") ||
+				      !strcmp(argv[start_arg], "clear"))) {
+		if (doit(iff_name, CHELSIO_DEL_FILTER, &op) < 0) {
+			if (errno == EBUSY)
+				err(1, "no filter support when offload in use");
+			err(1, "delete filter");
+		}
+		return 0;
+	}
+
+	while (start_arg + 2 <= argc) {
+		if (!strcmp(argv[start_arg], "sip")) {
+			ret = parse_ipaddr(argv[start_arg + 1], &op.val.sip,
+					   &op.mask.sip);
+		} else if (!strcmp(argv[start_arg], "dip")) {
+			ret = parse_ipaddr(argv[start_arg + 1], &op.val.dip,
+					   &op.mask.dip);
+		} else if (!strcmp(argv[start_arg], "sport")) {
+			ret = parse_val_mask_param(argv[start_arg + 1],
+						   &val, &mask, 0xffff);
+			op.val.sport = val;
+			op.mask.sport = mask;
+		} else if (!strcmp(argv[start_arg], "dport")) {
+			ret = parse_val_mask_param(argv[start_arg + 1],
+						   &val, &mask, 0xffff);
+			op.val.dport = val;
+			op.mask.dport = mask;
+		} else if (!strcmp(argv[start_arg], "vlan")) {
+			ret = parse_val_mask_param(argv[start_arg + 1],
+						   &val, &mask, 0xfff);
+			op.val.vlan = val;
+			op.mask.vlan = mask;
+		} else if (!strcmp(argv[start_arg], "prio")) {
+			ret = parse_val_mask_param(argv[start_arg + 1],
+						   &val, &mask, 7);
+			op.val.vlan_prio = val;
+			op.mask.vlan_prio = mask;
+		} else if (!strcmp(argv[start_arg], "mac")) {
+			if (!strcmp(argv[start_arg + 1], "none"))
+				val = -1;
+			else
+				ret = get_int_arg(argv[start_arg + 1], &val);
+			op.mac_hit = val != (uint32_t)-1;
+			op.mac_addr_idx = op.mac_hit ? val : 0;
+		} else if (!strcmp(argv[start_arg], "type")) {
+			if (!strcmp(argv[start_arg + 1], "tcp"))
+				op.proto = 1;
+			else if (!strcmp(argv[start_arg + 1], "udp"))
+				op.proto = 2;
+			else if (!strcmp(argv[start_arg + 1], "frag"))
+				op.proto = 3;
+			else
+				errx(1, "unknown type \"%s\"; must be one of "
+				     "\"tcp\", \"udp\", or \"frag\"",
+				     argv[start_arg + 1]);
+		} else if (!strcmp(argv[start_arg], "queue")) {
+			ret = get_int_arg(argv[start_arg + 1], &val);
+			op.qset = val;
+			op.rss = 0;
+		} else if (!strcmp(argv[start_arg], "action")) {
+			if (!strcmp(argv[start_arg + 1], "pass"))
+				op.pass = 1;
+			else if (strcmp(argv[start_arg + 1], "drop"))
+				errx(1, "unknown action \"%s\"; must be one of "
+				     "\"pass\" or \"drop\"",
+				     argv[start_arg + 1]);
+		} else
+ 			errx(1, "unknown filter parameter \"%s\"\n"
+			     "known parameters are \"mac\", \"sip\", "
+			     "\"dip\", \"sport\", \"dport\", \"vlan\", "
+			     "\"prio\", \"type\", \"queue\", and \"action\"",
+			     argv[start_arg]);
+		if (ret < 0)
+			errx(1, "bad value \"%s\" for parameter \"%s\"",
+			     argv[start_arg + 1], argv[start_arg]);
+		start_arg += 2;
+	}
+	if (start_arg != argc)
+		errx(1, "no value for \"%s\"", argv[start_arg]);
+
+	if (doit(iff_name, CHELSIO_SET_FILTER, &op) < 0) {
+		if (errno == EBUSY)
+			err(1, "no filter support when offload in use");
+		err(1, "set filter");
+	}
+	
+	return 0;
+}
 static int
 get_sched_param(int argc, char *argv[], int pos, unsigned int *valp)
 {
@@ -1501,6 +1673,8 @@ run_cmd(int argc, char *argv[], const char *iff_name)
 		r = pktsched(argc, argv, 3, iff_name);
 	else if (!strcmp(argv[2], "tcb"))
 		r = get_tcb2(argc, argv, 3, iff_name);
+	else if (!strcmp(argv[2], "filter"))
+		r = filter_config(argc, argv, 3, iff_name);
 	else if (!strcmp(argv[2], "clearstats"))
 		r = clear_stats(argc, argv, 3, iff_name);
 	else if (!strcmp(argv[2], "la"))

From e3ef0d2fcf5ed3e334f33ab15d52b2fcf5f8223d Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 5 May 2010 03:45:46 +0000
Subject: [PATCH 464/532] Push down the acquisition of the page queues lock
 into vm_page_unwire().

Update the comment describing which lock should be held on entry to
vm_page_wire().

Reviewed by:	kib
---
 sys/dev/agp/agp.c          |  4 ----
 sys/dev/agp/agp_i810.c     |  2 --
 sys/fs/tmpfs/tmpfs_vnops.c |  2 --
 sys/kern/vfs_bio.c         |  6 ++----
 sys/vm/vm_fault.c          | 14 +++++---------
 sys/vm/vm_glue.c           |  2 --
 sys/vm/vm_page.c           | 16 ++++++++--------
 7 files changed, 15 insertions(+), 31 deletions(-)

diff --git a/sys/dev/agp/agp.c b/sys/dev/agp/agp.c
index 89b68b000e1..577a7e9ec95 100644
--- a/sys/dev/agp/agp.c
+++ b/sys/dev/agp/agp.c
@@ -624,9 +624,7 @@ bad:
 		if (k >= i)
 			vm_page_wakeup(m);
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(mem->am_obj);
@@ -660,9 +658,7 @@ agp_generic_unbind_memory(device_t dev, struct agp_memory *mem)
 	for (i = 0; i < mem->am_size; i += PAGE_SIZE) {
 		m = vm_page_lookup(mem->am_obj, atop(i));
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(mem->am_obj);
diff --git a/sys/dev/agp/agp_i810.c b/sys/dev/agp/agp_i810.c
index 44eb50da27f..ff3ad1c614a 100644
--- a/sys/dev/agp/agp_i810.c
+++ b/sys/dev/agp/agp_i810.c
@@ -1011,9 +1011,7 @@ agp_i810_free_memory(device_t dev, struct agp_memory *mem)
 			VM_OBJECT_LOCK(mem->am_obj);
 			m = vm_page_lookup(mem->am_obj, 0);
 			vm_page_lock(m);
-			vm_page_lock_queues();
 			vm_page_unwire(m, 0);
-			vm_page_unlock_queues();
 			vm_page_unlock(m);
 			VM_OBJECT_UNLOCK(mem->am_obj);
 		} else {
diff --git a/sys/fs/tmpfs/tmpfs_vnops.c b/sys/fs/tmpfs/tmpfs_vnops.c
index 2d6005a5958..85cd23db6d8 100644
--- a/sys/fs/tmpfs/tmpfs_vnops.c
+++ b/sys/fs/tmpfs/tmpfs_vnops.c
@@ -461,9 +461,7 @@ tmpfs_nocacheread(vm_object_t tobj, vm_pindex_t idx,
 	VM_OBJECT_LOCK(tobj);
 out:
 	vm_page_lock(m);
-	vm_page_lock_queues();
 	vm_page_unwire(m, TRUE);
-	vm_page_unlock_queues();
 	vm_page_unlock(m);
 	vm_page_wakeup(m);
 	vm_object_pip_subtract(tobj, 1);
diff --git a/sys/kern/vfs_bio.c b/sys/kern/vfs_bio.c
index 589bdbd8093..ec8ad67fab3 100644
--- a/sys/kern/vfs_bio.c
+++ b/sys/kern/vfs_bio.c
@@ -1571,7 +1571,6 @@ vfs_vmio_release(struct buf *bp)
 		 * everything on the inactive queue.
 		 */
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		/*
 		 * We don't mess with busy pages, it is
@@ -1580,6 +1579,7 @@ vfs_vmio_release(struct buf *bp)
 		 */
 		if ((m->oflags & VPO_BUSY) == 0 && m->busy == 0 &&
 		    m->wire_count == 0) {
+			vm_page_lock_queues();
 			/*
 			 * Might as well free the page if we can and it has
 			 * no valid data.  We also free the page if the
@@ -1593,8 +1593,8 @@ vfs_vmio_release(struct buf *bp)
 			} else if (buf_vm_page_count_severe()) {
 				vm_page_try_to_cache(m);
 			}
+			vm_page_unlock_queues();
 		}
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(bp->b_bufobj->bo_object);
@@ -2957,9 +2957,7 @@ allocbuf(struct buf *bp, int size)
 
 					bp->b_pages[i] = NULL;
 					vm_page_lock(m);
-					vm_page_lock_queues();
 					vm_page_unwire(m, 0);
-					vm_page_unlock_queues();
 					vm_page_unlock(m);
 				}
 				VM_OBJECT_UNLOCK(bp->b_bufobj->bo_object);
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 142a9ef5d5a..373433f1d4b 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -799,9 +799,7 @@ vnode_locked:
 					vm_page_unlock(fs.first_m);
 					
 					vm_page_lock(fs.m);
-					vm_page_lock_queues();
 					vm_page_unwire(fs.m, FALSE);
-					vm_page_unlock_queues();
 					vm_page_unlock(fs.m);
 				}
 				/*
@@ -1112,6 +1110,7 @@ vm_fault_unwire(vm_map_t map, vm_offset_t start, vm_offset_t end,
 {
 	vm_paddr_t pa;
 	vm_offset_t va;
+	vm_page_t m;
 	pmap_t pmap;
 
 	pmap = vm_map_pmap(map);
@@ -1125,11 +1124,10 @@ vm_fault_unwire(vm_map_t map, vm_offset_t start, vm_offset_t end,
 		if (pa != 0) {
 			pmap_change_wiring(pmap, va, FALSE);
 			if (!fictitious) {
-				vm_page_lock(PHYS_TO_VM_PAGE(pa));
-				vm_page_lock_queues();
-				vm_page_unwire(PHYS_TO_VM_PAGE(pa), 1);
-				vm_page_unlock_queues();
-				vm_page_unlock(PHYS_TO_VM_PAGE(pa));
+				m = PHYS_TO_VM_PAGE(pa);
+				vm_page_lock(m);
+				vm_page_unwire(m, TRUE);
+				vm_page_unlock(m);
 			}
 		}
 	}
@@ -1275,9 +1273,7 @@ vm_fault_copy_entry(vm_map_t dst_map, vm_map_t src_map,
 		
 		if (upgrade) {
 			vm_page_lock(src_m);
-			vm_page_lock_queues();
 			vm_page_unwire(src_m, 0);
-			vm_page_unlock_queues();
 			vm_page_unlock(src_m);
 
 			vm_page_lock(dst_m);
diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index 288c5d72adc..f86212f0291 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -529,9 +529,7 @@ vm_thread_swapout(struct thread *td)
 			panic("vm_thread_swapout: kstack already missing?");
 		vm_page_dirty(m);
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(ksobj);
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index 957bf824679..a77dc136002 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1532,7 +1532,7 @@ vm_page_free_toq(vm_page_t m)
  *	another map, removing it from paging queues
  *	as necessary.
  *
- *	The page queues must be locked.
+ *	The page must be locked.
  *	This routine may not block.
  */
 void
@@ -1584,31 +1584,31 @@ vm_page_wire(vm_page_t m)
  *	be placed in the cache - for example, just after dirtying a page.
  *	dirty pages in the cache are not allowed.
  *
- *	The page queues must be locked.
+ *	The page must be locked.
  *	This routine may not block.
  */
 void
 vm_page_unwire(vm_page_t m, int activate)
 {
 
-	if ((m->flags & PG_UNMANAGED) == 0) {
-		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+	if ((m->flags & PG_UNMANAGED) == 0)
 		vm_page_lock_assert(m, MA_OWNED);
-	}
 	if (m->flags & PG_FICTITIOUS)
 		return;
 	if (m->wire_count > 0) {
 		m->wire_count--;
 		if (m->wire_count == 0) {
 			atomic_subtract_int(&cnt.v_wire_count, 1);
-			if (m->flags & PG_UNMANAGED) {
-				;
-			} else if (activate)
+			if ((m->flags & PG_UNMANAGED) != 0)
+				return;
+			vm_page_lock_queues();
+			if (activate)
 				vm_page_enqueue(PQ_ACTIVE, m);
 			else {
 				vm_page_flag_clear(m, PG_WINATCFLS);
 				vm_page_enqueue(PQ_INACTIVE, m);
 			}
+			vm_page_unlock_queues();
 		}
 	} else {
 		panic("vm_page_unwire: invalid wire count: %d", m->wire_count);

From 7ee44a58d7a399a6da3695e41f9fdd78ec50c35f Mon Sep 17 00:00:00 2001
From: Neel Natu <neel@FreeBSD.org>
Date: Wed, 5 May 2010 04:37:45 +0000
Subject: [PATCH 465/532] Fix DDB backtrace involving kernel modules.

We can no longer assume that all valid program counter values reside
within the kernel object file.
---
 sys/mips/mips/db_trace.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/mips/mips/db_trace.c b/sys/mips/mips/db_trace.c
index 9417bcc56d5..116470c3329 100644
--- a/sys/mips/mips/db_trace.c
+++ b/sys/mips/mips/db_trace.c
@@ -181,7 +181,7 @@ loop:
 	}
 	/* check for bad PC */
 	/*XXX MIPS64 bad: These hard coded constants are lame */
-	if (pc & 3 || pc < (uintptr_t)0x80000000 || pc >= (uintptr_t)edata) {
+	if (pc & 3 || pc < (uintptr_t)0x80000000) {
 		(*printfn) ("PC 0x%x: not in kernel\n", pc);
 		ra = 0;
 		goto done;

From 28a50a192f615b677f963ec2e839383b77360173 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 5 May 2010 07:40:54 +0000
Subject: [PATCH 466/532] Push down the acquisition of the page queues lock
 into vm_page_unwire().

---
 sys/dev/drm/via_dmablit.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/sys/dev/drm/via_dmablit.c b/sys/dev/drm/via_dmablit.c
index 81d438aea01..9be42ffa7f1 100644
--- a/sys/dev/drm/via_dmablit.c
+++ b/sys/dev/drm/via_dmablit.c
@@ -179,9 +179,7 @@ via_free_sg_info(drm_via_sg_info_t *vsg)
 		for (i=0; i < vsg->num_pages; ++i) {
 			if ( NULL != (page = vsg->pages[i])) {
 				vm_page_lock(page);
-				vm_page_lock_queues();
 				vm_page_unwire(page, 0);
-				vm_page_unlock_queues();
 				vm_page_unlock(page);
 			}
 		}

From d43708c0495bb8626a317ca07eef6d833f24e6c8 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 5 May 2010 07:47:40 +0000
Subject: [PATCH 467/532] Use an OBJT_PHYS object and thus PG_UNMANAGED pages
 to implement the TSB. The TSB is not a pageable structure, so there is no
 point in using managed pages.

Reviewed by:	kib
---
 sys/sparc64/sparc64/pmap.c | 8 +-------
 1 file changed, 1 insertion(+), 7 deletions(-)

diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index 4a2853ca7b0..daba4b7fb32 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -1091,7 +1091,7 @@ pmap_pinit(pmap_t pm)
 	 * Allocate an object for it.
 	 */
 	if (pm->pm_tsb_obj == NULL)
-		pm->pm_tsb_obj = vm_object_allocate(OBJT_DEFAULT, TSB_PAGES);
+		pm->pm_tsb_obj = vm_object_allocate(OBJT_PHYS, TSB_PAGES);
 
 	VM_OBJECT_LOCK(pm->pm_tsb_obj);
 	for (i = 0; i < TSB_PAGES; i++) {
@@ -1152,16 +1152,10 @@ pmap_release(pmap_t pm)
 	KASSERT(obj->ref_count == 1, ("pmap_release: tsbobj ref count != 1"));
 	while (!TAILQ_EMPTY(&obj->memq)) {
 		m = TAILQ_FIRST(&obj->memq);
-		vm_page_lock_queues();
-		if (vm_page_sleep_if_busy(m, FALSE, "pmaprl"))
-			continue;
-		KASSERT(m->hold_count == 0,
-		    ("pmap_release: freeing held tsb page"));
 		m->md.pmap = NULL;
 		m->wire_count--;
 		atomic_subtract_int(&cnt.v_wire_count, 1);
 		vm_page_free_zero(m);
-		vm_page_unlock_queues();
 	}
 	VM_OBJECT_UNLOCK(obj);
 	pmap_qremove((vm_offset_t)pm->pm_tsb, TSB_PAGES);

From 35939b04d3299d8956dd7963ec2731bd2941e306 Mon Sep 17 00:00:00 2001
From: Daniel Gerzo <danger@FreeBSD.org>
Date: Wed, 5 May 2010 08:43:47 +0000
Subject: [PATCH 468/532] - fix typo

---
 usr.sbin/jail/jail.8 | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.sbin/jail/jail.8 b/usr.sbin/jail/jail.8
index 85fe988f1ea..8e49b07f781 100644
--- a/usr.sbin/jail/jail.8
+++ b/usr.sbin/jail/jail.8
@@ -239,7 +239,7 @@ The
 .Xr jexec 8 .
 .It Va ip4.addr
 A comma-separated list of IPv4 addresses assigned to the prison.
-If this is set, the jail is restricted to using only these address.
+If this is set, the jail is restricted to using only these addresses.
 Any attempts to use other addresses fail, and attempts to use wildcard
 addresses silently use the jailed address instead.
 For IPv4 the first address given will be kept used as the source address

From 2d2a2083f7b08af9aba55c6fa338714a535a8566 Mon Sep 17 00:00:00 2001
From: VANHULLEBUS Yvan <vanhu@FreeBSD.org>
Date: Wed, 5 May 2010 08:55:26 +0000
Subject: [PATCH 469/532] Update SA's NAT-T stuff before calling key_mature()
 in key_update(), as SA may be used as soon as key_mature() has been called.

Obtained from:	NETASQ
MFC after: 1 week
---
 sys/netipsec/key.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/sys/netipsec/key.c b/sys/netipsec/key.c
index 521976812eb..7fc199ab2f0 100644
--- a/sys/netipsec/key.c
+++ b/sys/netipsec/key.c
@@ -5156,12 +5156,6 @@ key_update(so, m, mhp)
 		return key_senderror(so, m, error);
 	}
 
-	/* check SA values to be mature. */
-	if ((mhp->msg->sadb_msg_errno = key_mature(sav)) != 0) {
-		KEY_FREESAV(&sav);
-		return key_senderror(so, m, 0);
-	}
-
 #ifdef IPSEC_NAT_T
 	/*
 	 * Handle more NAT-T info if present,
@@ -5188,6 +5182,12 @@ key_update(so, m, mhp)
 #endif
 #endif
 
+	/* check SA values to be mature. */
+	if ((mhp->msg->sadb_msg_errno = key_mature(sav)) != 0) {
+		KEY_FREESAV(&sav);
+		return key_senderror(so, m, 0);
+	}
+
     {
 	struct mbuf *n;
 

From 2e8d55c4e86314f1f7992efea6ee840c4b60c40f Mon Sep 17 00:00:00 2001
From: VANHULLEBUS Yvan <vanhu@FreeBSD.org>
Date: Wed, 5 May 2010 08:58:58 +0000
Subject: [PATCH 470/532] Set SA's natt_type before calling key_mature() in
 key_add(), as the SA may be used as soon as key_mature() has been done.

Obtained from:	NETASQ
MFC after:	1 week
---
 sys/netipsec/key.c | 12 ++++++------
 1 file changed, 6 insertions(+), 6 deletions(-)

diff --git a/sys/netipsec/key.c b/sys/netipsec/key.c
index 7fc199ab2f0..d00489db907 100644
--- a/sys/netipsec/key.c
+++ b/sys/netipsec/key.c
@@ -5422,12 +5422,6 @@ key_add(so, m, mhp)
 		return key_senderror(so, m, error);
 	}
 
-	/* check SA values to be mature. */
-	if ((error = key_mature(newsav)) != 0) {
-		KEY_FREESAV(&newsav);
-		return key_senderror(so, m, error);
-	}
-
 #ifdef IPSEC_NAT_T
 	/*
 	 * Handle more NAT-T info if present,
@@ -5447,6 +5441,12 @@ key_add(so, m, mhp)
 #endif
 #endif
 
+	/* check SA values to be mature. */
+	if ((error = key_mature(newsav)) != 0) {
+		KEY_FREESAV(&newsav);
+		return key_senderror(so, m, error);
+	}
+
 	/*
 	 * don't call key_freesav() here, as we would like to keep the SA
 	 * in the database on success.

From 213c077f62b2dc32de2ce12b26564dd325184926 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 5 May 2010 16:05:51 +0000
Subject: [PATCH 471/532] Fix a mistake in r207603. td_rux.rux_runtime still
 needs conversion.

Reported and tested by:	nwhitehorn
Pointy hat to:	kib
MFC after:	6 days
---
 sys/kern/kern_proc.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/kern_proc.c b/sys/kern/kern_proc.c
index 6c50f1f3618..055c844fcfe 100644
--- a/sys/kern/kern_proc.c
+++ b/sys/kern/kern_proc.c
@@ -901,7 +901,7 @@ fill_kinfo_thread(struct thread *td, struct kinfo_proc *kp, int preferthread)
 	kp->ki_pri.pri_user = td->td_user_pri;
 
 	if (preferthread) {
-		kp->ki_runtime = td->td_rux.rux_runtime;
+		kp->ki_runtime = cputick2usec(td->td_rux.rux_runtime);
 		kp->ki_pctcpu = sched_pctcpu(td);
 		kp->ki_estcpu = td->td_estcpu;
 	}

From b5f770bd8627d79822809f223a3ad5eab7bcadd6 Mon Sep 17 00:00:00 2001
From: Edward Tomasz Napierala <trasz@FreeBSD.org>
Date: Wed, 5 May 2010 16:44:25 +0000
Subject: [PATCH 472/532] Move checking against RLIMIT_FSIZE into one place,
 vn_rlimit_fsize().

Reviewed by:	kib
---
 sys/fs/ext2fs/ext2_readwrite.c     | 14 ++------------
 sys/fs/ext2fs/ext2_vnops.c         |  3 ---
 sys/fs/msdosfs/msdosfs_vnops.c     | 15 ++-------------
 sys/fs/nfsclient/nfs_clbio.c       | 16 ++--------------
 sys/fs/nwfs/nwfs_io.c              | 17 ++++-------------
 sys/fs/smbfs/smbfs_io.c            | 18 ++++--------------
 sys/fs/tmpfs/tmpfs_vnops.c         | 13 ++-----------
 sys/gnu/fs/xfs/FreeBSD/xfs_vnops.c | 12 ++----------
 sys/kern/vfs_vnops.c               | 19 +++++++++++++++++++
 sys/nfsclient/nfs_bio.c            | 15 ++-------------
 sys/sys/vnode.h                    |  1 +
 sys/ufs/ffs/ffs_vnops.c            | 17 ++---------------
 12 files changed, 42 insertions(+), 118 deletions(-)

diff --git a/sys/fs/ext2fs/ext2_readwrite.c b/sys/fs/ext2fs/ext2_readwrite.c
index 9c9749a21d7..1a713ca5949 100644
--- a/sys/fs/ext2fs/ext2_readwrite.c
+++ b/sys/fs/ext2fs/ext2_readwrite.c
@@ -168,7 +168,6 @@ WRITE(ap)
 	struct inode *ip;
 	FS *fs;
 	struct buf *bp;
-	struct thread *td;
 	daddr_t lbn;
 	off_t osize;
 	int blkoffset, error, flags, ioflag, resid, size, seqcount, xfersize;
@@ -213,17 +212,8 @@ WRITE(ap)
 	 * Maybe this should be above the vnode op call, but so long as
 	 * file servers have no limits, I don't think it matters.
 	 */
-	td = uio->uio_td;
-	if (vp->v_type == VREG && td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if (uio->uio_offset + uio->uio_resid >
-		    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+	if (vn_rlimit_fsize(vp, uio, uio->uio_td))
+		return (EFBIG);
 
 	resid = uio->uio_resid;
 	osize = ip->i_size;
diff --git a/sys/fs/ext2fs/ext2_vnops.c b/sys/fs/ext2fs/ext2_vnops.c
index 2b2f8c547c7..2d302c56e86 100644
--- a/sys/fs/ext2fs/ext2_vnops.c
+++ b/sys/fs/ext2fs/ext2_vnops.c
@@ -46,7 +46,6 @@
 
 #include <sys/param.h>
 #include <sys/systm.h>
-#include <sys/resourcevar.h>
 #include <sys/kernel.h>
 #include <sys/fcntl.h>
 #include <sys/stat.h>
@@ -54,7 +53,6 @@
 #include <sys/buf.h>
 #include <sys/endian.h>
 #include <sys/priv.h>
-#include <sys/proc.h>
 #include <sys/mount.h>
 #include <sys/unistd.h>
 #include <sys/time.h>
@@ -71,7 +69,6 @@
 
 #include <fs/fifofs/fifo.h>
 
-#include <sys/signalvar.h>
 #include <ufs/ufs/dir.h>
 
 #include <fs/ext2fs/inode.h>
diff --git a/sys/fs/msdosfs/msdosfs_vnops.c b/sys/fs/msdosfs/msdosfs_vnops.c
index e9c593bdc0c..f38c72b841b 100644
--- a/sys/fs/msdosfs/msdosfs_vnops.c
+++ b/sys/fs/msdosfs/msdosfs_vnops.c
@@ -61,9 +61,6 @@
 #include <sys/mutex.h>
 #include <sys/namei.h>
 #include <sys/priv.h>
-#include <sys/proc.h>
-#include <sys/resourcevar.h>
-#include <sys/signalvar.h>
 #include <sys/stat.h>
 #include <sys/unistd.h>
 #include <sys/vnode.h>
@@ -699,16 +696,8 @@ msdosfs_write(ap)
 	/*
 	 * If they've exceeded their filesize limit, tell them about it.
 	 */
-	if (td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if ((uoff_t)uio->uio_offset + uio->uio_resid >
-		    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+	if (vn_rlimit_fsize(vp, uio, td))
+		return (EFBIG);
 
 	/*
 	 * If the offset we are starting the write at is beyond the end of
diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index e23da725d3f..14138569c14 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -41,9 +41,6 @@ __FBSDID("$FreeBSD$");
 #include <sys/buf.h>
 #include <sys/kernel.h>
 #include <sys/mount.h>
-#include <sys/proc.h>
-#include <sys/resourcevar.h>
-#include <sys/signalvar.h>
 #include <sys/vmmeter.h>
 #include <sys/vnode.h>
 
@@ -880,7 +877,6 @@ ncl_write(struct vop_write_args *ap)
 	daddr_t lbn;
 	int bcount;
 	int n, on, error = 0;
-	struct proc *p = td?td->td_proc:NULL;
 
 #ifdef DIAGNOSTIC
 	if (uio->uio_rw != UIO_WRITE)
@@ -962,16 +958,8 @@ flush_and_restart:
 	 * Maybe this should be above the vnode op call, but so long as
 	 * file servers have no limits, i don't think it matters
 	 */
-	if (p != NULL) {
-		PROC_LOCK(p);
-		if (uio->uio_offset + uio->uio_resid >
-		    lim_cur(p, RLIMIT_FSIZE)) {
-			psignal(p, SIGXFSZ);
-			PROC_UNLOCK(p);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(p);
-	}
+	if (vn_rlimit_fsize(vp, uio, td))
+		return (EFBIG);
 
 	biosize = vp->v_mount->mnt_stat.f_iosize;
 	/*
diff --git a/sys/fs/nwfs/nwfs_io.c b/sys/fs/nwfs/nwfs_io.c
index ef1c800a525..83fbe7047dd 100644
--- a/sys/fs/nwfs/nwfs_io.c
+++ b/sys/fs/nwfs/nwfs_io.c
@@ -28,16 +28,13 @@
  */
 #include <sys/param.h>
 #include <sys/systm.h>
-#include <sys/resourcevar.h>	/* defines plimit structure in proc struct */
 #include <sys/kernel.h>
 #include <sys/bio.h>
 #include <sys/buf.h>
-#include <sys/proc.h>
 #include <sys/mount.h>
 #include <sys/namei.h>
 #include <sys/vnode.h>
 #include <sys/dirent.h>
-#include <sys/signalvar.h>
 #include <sys/sysctl.h>
 
 #include <vm/vm.h>
@@ -229,16 +226,10 @@ nwfs_writevnode(vp, uiop, cred, ioflag)
 		}
 	}
 	if (uiop->uio_resid == 0) return 0;
-	if (td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if  (uiop->uio_offset + uiop->uio_resid >
-		    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+
+	if (vn_rlimit_fsize(vp, uiop, td))
+		return (EFBIG);
+
 	error = ncp_write(NWFSTOCONN(nmp), &np->n_fh, uiop, cred);
 	NCPVNDEBUG("after: ofs=%d,resid=%d\n",(int)uiop->uio_offset, uiop->uio_resid);
 	if (!error) {
diff --git a/sys/fs/smbfs/smbfs_io.c b/sys/fs/smbfs/smbfs_io.c
index 6537ed4b29f..7e021c13418 100644
--- a/sys/fs/smbfs/smbfs_io.c
+++ b/sys/fs/smbfs/smbfs_io.c
@@ -28,9 +28,7 @@
  */
 #include <sys/param.h>
 #include <sys/systm.h>
-#include <sys/resourcevar.h>	/* defines plimit structure in proc struct */
 #include <sys/kernel.h>
-#include <sys/proc.h>
 #include <sys/fcntl.h>
 #include <sys/bio.h>
 #include <sys/buf.h>
@@ -235,7 +233,6 @@ smbfs_writevnode(struct vnode *vp, struct uio *uiop,
 	struct smbmount *smp = VTOSMBFS(vp);
 	struct smbnode *np = VTOSMB(vp);
 	struct smb_cred scred;
-	struct proc *p;
 	struct thread *td;
 	int error = 0;
 
@@ -249,7 +246,6 @@ smbfs_writevnode(struct vnode *vp, struct uio *uiop,
 /*	if (uiop->uio_offset + uiop->uio_resid > smp->nm_maxfilesize)
 		return (EFBIG);*/
 	td = uiop->uio_td;
-	p = td->td_proc;
 	if (ioflag & (IO_APPEND | IO_SYNC)) {
 		if (np->n_flag & NMODIFIED) {
 			smbfs_attr_cacheremove(vp);
@@ -271,16 +267,10 @@ smbfs_writevnode(struct vnode *vp, struct uio *uiop,
 	}
 	if (uiop->uio_resid == 0)
 		return 0;
-	if (p != NULL) {
-		PROC_LOCK(p);
-		if (uiop->uio_offset + uiop->uio_resid >
-		    lim_cur(p, RLIMIT_FSIZE)) {
-			psignal(p, SIGXFSZ);
-			PROC_UNLOCK(p);
-			return EFBIG;
-		}
-		PROC_UNLOCK(p);
-	}
+
+	if (vn_rlimit_fsize(vp, uiop, td))
+		return (EFBIG);
+
 	smb_makescred(&scred, td, cred);
 	error = smb_write(smp->sm_share, np->n_fid, uiop, &scred);
 	SMBVDEBUG("after: ofs=%d,resid=%d\n",(int)uiop->uio_offset, uiop->uio_resid);
diff --git a/sys/fs/tmpfs/tmpfs_vnops.c b/sys/fs/tmpfs/tmpfs_vnops.c
index 85cd23db6d8..3ffef6b6413 100644
--- a/sys/fs/tmpfs/tmpfs_vnops.c
+++ b/sys/fs/tmpfs/tmpfs_vnops.c
@@ -42,7 +42,6 @@ __FBSDID("$FreeBSD$");
 #include <sys/namei.h>
 #include <sys/priv.h>
 #include <sys/proc.h>
-#include <sys/resourcevar.h>
 #include <sys/sched.h>
 #include <sys/sf_buf.h>
 #include <sys/stat.h>
@@ -748,16 +747,8 @@ tmpfs_write(struct vop_write_args *v)
 	  VFS_TO_TMPFS(vp->v_mount)->tm_maxfilesize)
 		return (EFBIG);
 
-	if (vp->v_type == VREG && td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if (uio->uio_offset + uio->uio_resid >
-		  lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+	if (vn_rlimit_fsize(vp, uio, td))
+		return (EFBIG);
 
 	extended = uio->uio_offset + uio->uio_resid > node->tn_size;
 	if (extended) {
diff --git a/sys/gnu/fs/xfs/FreeBSD/xfs_vnops.c b/sys/gnu/fs/xfs/FreeBSD/xfs_vnops.c
index 20be557d771..b935b6468b6 100644
--- a/sys/gnu/fs/xfs/FreeBSD/xfs_vnops.c
+++ b/sys/gnu/fs/xfs/FreeBSD/xfs_vnops.c
@@ -598,16 +598,8 @@ xfs_write_file(xfs_inode_t *xip, struct uio *uio, int ioflag)
 	 */
 #if 0
 	td = uio->uio_td;
-	if (vp->v_type == VREG && td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if (uio->uio_offset + uio->uio_resid >
-		    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+	if (vn_rlimit_fsize(vp, uio, uio->uio_td))
+		return (EFBIG);
 #endif
 
 	resid = uio->uio_resid;
diff --git a/sys/kern/vfs_vnops.c b/sys/kern/vfs_vnops.c
index 838f8f7be31..edc7ea83d29 100644
--- a/sys/kern/vfs_vnops.c
+++ b/sys/kern/vfs_vnops.c
@@ -54,6 +54,7 @@ __FBSDID("$FreeBSD$");
 #include <sys/bio.h>
 #include <sys/buf.h>
 #include <sys/filio.h>
+#include <sys/resourcevar.h>
 #include <sys/sx.h>
 #include <sys/ttycom.h>
 #include <sys/conf.h>
@@ -1338,3 +1339,21 @@ vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags, struct vnode **rvp)
 	}
 	return (error);
 }
+
+int
+vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio, const struct thread *td)
+{
+	if (vp->v_type != VREG || td == NULL)
+		return (0);
+
+	PROC_LOCK(td->td_proc);
+	if (uio->uio_offset + uio->uio_resid >
+	    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
+		psignal(td->td_proc, SIGXFSZ);
+		PROC_UNLOCK(td->td_proc);
+		return (EFBIG);
+	}
+	PROC_UNLOCK(td->td_proc);
+
+	return (0);
+}
diff --git a/sys/nfsclient/nfs_bio.c b/sys/nfsclient/nfs_bio.c
index 3a0705ef88d..13effd0aa3a 100644
--- a/sys/nfsclient/nfs_bio.c
+++ b/sys/nfsclient/nfs_bio.c
@@ -45,8 +45,6 @@ __FBSDID("$FreeBSD$");
 #include <sys/mbuf.h>
 #include <sys/mount.h>
 #include <sys/proc.h>
-#include <sys/resourcevar.h>
-#include <sys/signalvar.h>
 #include <sys/vmmeter.h>
 #include <sys/vnode.h>
 
@@ -873,7 +871,6 @@ nfs_write(struct vop_write_args *ap)
 	daddr_t lbn;
 	int bcount;
 	int n, on, error = 0;
-	struct proc *p = td?td->td_proc:NULL;
 
 #ifdef DIAGNOSTIC
 	if (uio->uio_rw != UIO_WRITE)
@@ -954,16 +951,8 @@ flush_and_restart:
 	 * Maybe this should be above the vnode op call, but so long as
 	 * file servers have no limits, i don't think it matters
 	 */
-	if (p != NULL) {
-		PROC_LOCK(p);
-		if (uio->uio_offset + uio->uio_resid >
-		    lim_cur(p, RLIMIT_FSIZE)) {
-			psignal(p, SIGXFSZ);
-			PROC_UNLOCK(p);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(p);
-	}
+	if (vn_rlimit_fsize(vp, uio, td))
+		return (EFBIG);
 
 	biosize = vp->v_mount->mnt_stat.f_iosize;
 	/*
diff --git a/sys/sys/vnode.h b/sys/sys/vnode.h
index b5784e4a75a..7c4ae32df44 100644
--- a/sys/sys/vnode.h
+++ b/sys/sys/vnode.h
@@ -780,6 +780,7 @@ struct dirent;
 int vfs_read_dirent(struct vop_readdir_args *ap, struct dirent *dp, off_t off);
 
 int	vfs_unixify_accmode(accmode_t *accmode);
+int	vn_rlimit_fsize(const struct vnode *vn, const struct uio *uio, const struct thread *td);
 
 #endif /* _KERNEL */
 
diff --git a/sys/ufs/ffs/ffs_vnops.c b/sys/ufs/ffs/ffs_vnops.c
index e6617cbcdfa..85f8ffd4d8a 100644
--- a/sys/ufs/ffs/ffs_vnops.c
+++ b/sys/ufs/ffs/ffs_vnops.c
@@ -75,9 +75,6 @@ __FBSDID("$FreeBSD$");
 #include <sys/malloc.h>
 #include <sys/mount.h>
 #include <sys/priv.h>
-#include <sys/proc.h>
-#include <sys/resourcevar.h>
-#include <sys/signalvar.h>
 #include <sys/stat.h>
 #include <sys/vmmeter.h>
 #include <sys/vnode.h>
@@ -652,7 +649,6 @@ ffs_write(ap)
 	struct inode *ip;
 	struct fs *fs;
 	struct buf *bp;
-	struct thread *td;
 	ufs_lbn_t lbn;
 	off_t osize;
 	int seqcount;
@@ -704,17 +700,8 @@ ffs_write(ap)
 	 * Maybe this should be above the vnode op call, but so long as
 	 * file servers have no limits, I don't think it matters.
 	 */
-	td = uio->uio_td;
-	if (vp->v_type == VREG && td != NULL) {
-		PROC_LOCK(td->td_proc);
-		if (uio->uio_offset + uio->uio_resid >
-		    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
-			psignal(td->td_proc, SIGXFSZ);
-			PROC_UNLOCK(td->td_proc);
-			return (EFBIG);
-		}
-		PROC_UNLOCK(td->td_proc);
-	}
+	if (vn_rlimit_fsize(vp, uio, uio->uio_td))
+		return (EFBIG);
 
 	resid = uio->uio_resid;
 	osize = ip->i_size;

From 5ac59343be1105caac5ff4bbc4c20f94da925a7c Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Wed, 5 May 2010 18:16:06 +0000
Subject: [PATCH 473/532] Acquire the page lock around all remaining calls to
 vm_page_free() on managed pages that didn't already have that lock held. 
 (Freeing an unmanaged page, such as the various pmaps use, doesn't require
 the page lock.)

This allows a change in vm_page_remove()'s locking requirements.  It now
expects the page lock to be held instead of the page queues lock.
Consequently, the page queues lock is no longer required at all by callers
to vm_page_rename().

Discussed with: kib
---
 sys/fs/nfsclient/nfs_clbio.c | 18 ++++++++++++------
 sys/fs/nwfs/nwfs_io.c        |  9 ++++++---
 sys/fs/smbfs/smbfs_io.c      | 18 ++++++++++++------
 sys/kern/kern_exec.c         |  2 ++
 sys/kern/subr_uio.c          |  2 ++
 sys/nfsclient/nfs_bio.c      | 18 ++++++++++++------
 sys/ufs/ffs/ffs_vnops.c      |  6 ++++--
 sys/vm/vm_fault.c            |  2 --
 sys/vm/vm_object.c           |  4 ----
 sys/vm/vm_page.c             | 12 ++++++++----
 sys/vm/vm_page.h             |  2 +-
 11 files changed, 59 insertions(+), 34 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index 14138569c14..df027acd800 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -131,12 +131,15 @@ ncl_getpages(struct vop_getpages_args *ap)
 	 */
 	VM_OBJECT_LOCK(object);
 	if (pages[ap->a_reqpage]->valid != 0) {
-		vm_page_lock_queues();
 		for (i = 0; i < npages; ++i) {
-			if (i != ap->a_reqpage)
+			if (i != ap->a_reqpage) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return (0);
 	}
@@ -171,12 +174,15 @@ ncl_getpages(struct vop_getpages_args *ap)
 	if (error && (uio.uio_resid == count)) {
 		ncl_printf("nfs_getpages: error %d\n", error);
 		VM_OBJECT_LOCK(object);
-		vm_page_lock_queues();
 		for (i = 0; i < npages; ++i) {
-			if (i != ap->a_reqpage)
+			if (i != ap->a_reqpage) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return (VM_PAGER_ERROR);
 	}
diff --git a/sys/fs/nwfs/nwfs_io.c b/sys/fs/nwfs/nwfs_io.c
index 83fbe7047dd..0bb996fb3b3 100644
--- a/sys/fs/nwfs/nwfs_io.c
+++ b/sys/fs/nwfs/nwfs_io.c
@@ -428,12 +428,15 @@ nwfs_getpages(ap)
 	VM_OBJECT_LOCK(object);
 	if (error && (uio.uio_resid == count)) {
 		printf("nwfs_getpages: error %d\n",error);
-		vm_page_lock_queues();
 		for (i = 0; i < npages; i++) {
-			if (ap->a_reqpage != i)
+			if (ap->a_reqpage != i) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return VM_PAGER_ERROR;
 	}
diff --git a/sys/fs/smbfs/smbfs_io.c b/sys/fs/smbfs/smbfs_io.c
index 7e021c13418..c47fb1bcfe4 100644
--- a/sys/fs/smbfs/smbfs_io.c
+++ b/sys/fs/smbfs/smbfs_io.c
@@ -440,12 +440,15 @@ smbfs_getpages(ap)
 
 	VM_OBJECT_LOCK(object);
 	if (m->valid != 0) {
-		vm_page_lock_queues();
 		for (i = 0; i < npages; ++i) {
-			if (i != reqpage)
+			if (i != reqpage) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return 0;
 	}
@@ -478,12 +481,15 @@ smbfs_getpages(ap)
 	VM_OBJECT_LOCK(object);
 	if (error && (uio.uio_resid == count)) {
 		printf("smbfs_getpages: error %d\n",error);
-		vm_page_lock_queues();
 		for (i = 0; i < npages; i++) {
-			if (reqpage != i)
+			if (reqpage != i) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return VM_PAGER_ERROR;
 	}
diff --git a/sys/kern/kern_exec.c b/sys/kern/kern_exec.c
index bb92972fbe6..0ec9817140d 100644
--- a/sys/kern/kern_exec.c
+++ b/sys/kern/kern_exec.c
@@ -949,9 +949,11 @@ exec_map_first_page(imgp)
 		ma[0] = vm_page_lookup(object, 0);
 		if ((rv != VM_PAGER_OK) || (ma[0] == NULL)) {
 			if (ma[0]) {
+				vm_page_lock(ma[0]);
 				vm_page_lock_queues();
 				vm_page_free(ma[0]);
 				vm_page_unlock_queues();
+				vm_page_unlock(ma[0]);
 			}
 			VM_OBJECT_UNLOCK(object);
 			return (EIO);
diff --git a/sys/kern/subr_uio.c b/sys/kern/subr_uio.c
index 725b1a83f9b..783a8d625f3 100644
--- a/sys/kern/subr_uio.c
+++ b/sys/kern/subr_uio.c
@@ -104,9 +104,11 @@ retry:
 	if ((user_pg = vm_page_lookup(uobject, upindex)) != NULL) {
 		if (vm_page_sleep_if_busy(user_pg, TRUE, "vm_pgmoveco"))
 			goto retry;
+		vm_page_lock(user_pg);
 		vm_page_lock_queues();
 		pmap_remove_all(user_pg);
 		vm_page_free(user_pg);
+		vm_page_unlock(user_pg);
 	} else {
 		/*
 		 * Even if a physical page does not exist in the
diff --git a/sys/nfsclient/nfs_bio.c b/sys/nfsclient/nfs_bio.c
index 13effd0aa3a..c93d677eadd 100644
--- a/sys/nfsclient/nfs_bio.c
+++ b/sys/nfsclient/nfs_bio.c
@@ -129,12 +129,15 @@ nfs_getpages(struct vop_getpages_args *ap)
 	 */
 	VM_OBJECT_LOCK(object);
 	if (pages[ap->a_reqpage]->valid != 0) {
-		vm_page_lock_queues();
 		for (i = 0; i < npages; ++i) {
-			if (i != ap->a_reqpage)
+			if (i != ap->a_reqpage) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return (0);
 	}
@@ -169,12 +172,15 @@ nfs_getpages(struct vop_getpages_args *ap)
 	if (error && (uio.uio_resid == count)) {
 		nfs_printf("nfs_getpages: error %d\n", error);
 		VM_OBJECT_LOCK(object);
-		vm_page_lock_queues();
 		for (i = 0; i < npages; ++i) {
-			if (i != ap->a_reqpage)
+			if (i != ap->a_reqpage) {
+				vm_page_lock(pages[i]);
+				vm_page_lock_queues();
 				vm_page_free(pages[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(pages[i]);
+			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(object);
 		return (VM_PAGER_ERROR);
 	}
diff --git a/sys/ufs/ffs/ffs_vnops.c b/sys/ufs/ffs/ffs_vnops.c
index 85f8ffd4d8a..a2fee4ecb41 100644
--- a/sys/ufs/ffs/ffs_vnops.c
+++ b/sys/ufs/ffs/ffs_vnops.c
@@ -847,13 +847,15 @@ ffs_getpages(ap)
 	if (mreq->valid) {
 		if (mreq->valid != VM_PAGE_BITS_ALL)
 			vm_page_zero_invalid(mreq, TRUE);
-		vm_page_lock_queues();
 		for (i = 0; i < pcount; i++) {
 			if (i != ap->a_reqpage) {
+				vm_page_lock(ap->a_m[i]);
+				vm_page_lock_queues();
 				vm_page_free(ap->a_m[i]);
+				vm_page_unlock_queues();
+				vm_page_unlock(ap->a_m[i]);
 			}
 		}
-		vm_page_unlock_queues();
 		VM_OBJECT_UNLOCK(mreq->object);
 		return VM_PAGER_OK;
 	}
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 373433f1d4b..2978f7f3a82 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -778,9 +778,7 @@ vnode_locked:
 				 * automatically made dirty.
 				 */
 				vm_page_lock(fs.m);
-				vm_page_lock_queues();
 				vm_page_rename(fs.m, fs.first_object, fs.first_pindex);
-				vm_page_unlock_queues();
 				vm_page_unlock(fs.m);
 				vm_page_busy(fs.m);
 				fs.first_m = fs.m;
diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index b5f246c0bec..f8c50bc73a0 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -1461,9 +1461,7 @@ retry:
 			goto retry;
 		}
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_rename(m, new_object, idx);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 		/* page automatically made dirty by rename and cache handled */
 		vm_page_busy(m);
@@ -1691,9 +1689,7 @@ vm_object_backing_scan(vm_object_t object, int op)
 			 * mapped through the rename.
 			 */
 			vm_page_lock(p);
-			vm_page_lock_queues();
 			vm_page_rename(p, object, new_pindex);
-			vm_page_unlock_queues();
 			vm_page_unlock(p);
 			/* page automatically made dirty by rename */
 		}
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index a77dc136002..cc34c1b857e 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -791,7 +791,7 @@ vm_page_remove(vm_page_t m)
 	vm_page_t root;
 
 	if ((m->flags & PG_UNMANAGED) == 0)
-		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+		vm_page_lock_assert(m, MA_OWNED);
 	if ((object = m->object) == NULL)
 		return;
 	VM_OBJECT_LOCK_ASSERT(object, MA_OWNED);
@@ -2234,11 +2234,11 @@ vm_page_cowfault(vm_page_t m)
 
  retry_alloc:
 	pmap_remove_all(m);
+	vm_page_unlock_queues();
 	vm_page_remove(m);
 	mnew = vm_page_alloc(object, pindex, VM_ALLOC_NORMAL | VM_ALLOC_NOBUSY);
 	if (mnew == NULL) {
 		vm_page_insert(m, object, pindex);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 		VM_OBJECT_UNLOCK(object);
 		VM_WAIT;
@@ -2261,7 +2261,12 @@ vm_page_cowfault(vm_page_t m)
 		 * waiting to allocate a page.  If so, put things back 
 		 * the way they were 
 		 */
+		vm_page_unlock(m);
+		vm_page_lock(mnew);
+		vm_page_lock_queues();
 		vm_page_free(mnew);
+		vm_page_unlock_queues();
+		vm_page_unlock(mnew);
 		vm_page_insert(m, object, pindex);
 	} else { /* clear COW & copy page */
 		if (!so_zerocp_fullpage)
@@ -2270,9 +2275,8 @@ vm_page_cowfault(vm_page_t m)
 		vm_page_dirty(mnew);
 		mnew->wire_count = m->wire_count - m->cow;
 		m->wire_count = m->cow;
+		vm_page_unlock(m);
 	}
-	vm_page_unlock_queues();
-	vm_page_unlock(m);
 }
 
 void 
diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index 51a896f9ae3..da4d42a8a76 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -103,7 +103,7 @@ struct vm_page {
 	struct vm_page *left;		/* splay tree link (O)		*/
 	struct vm_page *right;		/* splay tree link (O)		*/
 
-	vm_object_t object;		/* which object am I in (O,Q)*/
+	vm_object_t object;		/* which object am I in (O,P)*/
 	vm_pindex_t pindex;		/* offset into object (O,Q) */
 	vm_paddr_t phys_addr;		/* physical address of page */
 	struct md_page md;		/* machine dependant stuff */

From d75554ec0451b31e7a923b25314b703cd66929ff Mon Sep 17 00:00:00 2001
From: Martin Matuska <mm@FreeBSD.org>
Date: Wed, 5 May 2010 18:22:29 +0000
Subject: [PATCH 474/532] Introduce hardforce export option (-F) for "zpool
 export". When exporting with this flag, zpool.cache remains untouched.

OpenSolaris onnv revision: 8211:32722be6ad3b

Approved by:	pjd, delphij (mentor)
Obtained from:	OpenSolaris (Bug ID: 6775357)
---
 .../opensolaris/cmd/zpool/zpool_main.c        | 12 +++++++++--
 cddl/contrib/opensolaris/cmd/ztest/ztest.c    |  2 +-
 .../opensolaris/lib/libzfs/common/libzfs.h    |  1 +
 .../lib/libzfs/common/libzfs_pool.c           | 15 ++++++++++++-
 .../opensolaris/uts/common/fs/zfs/spa.c       | 21 ++++++++++++-------
 .../opensolaris/uts/common/fs/zfs/sys/spa.h   |  3 ++-
 .../opensolaris/uts/common/fs/zfs/zfs_ioctl.c |  3 ++-
 7 files changed, 43 insertions(+), 14 deletions(-)

diff --git a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
index 297037178a9..eef60e6dedb 100644
--- a/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
+++ b/cddl/contrib/opensolaris/cmd/zpool/zpool_main.c
@@ -879,17 +879,21 @@ int
 zpool_do_export(int argc, char **argv)
 {
 	boolean_t force = B_FALSE;
+	boolean_t hardforce = B_FALSE;
 	int c;
 	zpool_handle_t *zhp;
 	int ret;
 	int i;
 
 	/* check options */
-	while ((c = getopt(argc, argv, "f")) != -1) {
+	while ((c = getopt(argc, argv, "fF")) != -1) {
 		switch (c) {
 		case 'f':
 			force = B_TRUE;
 			break;
+		case 'F':
+			hardforce = B_TRUE;
+			break;
 		case '?':
 			(void) fprintf(stderr, gettext("invalid option '%c'\n"),
 			    optopt);
@@ -919,8 +923,12 @@ zpool_do_export(int argc, char **argv)
 			continue;
 		}
 
-		if (zpool_export(zhp, force) != 0)
+		if (hardforce) {
+			if (zpool_export_force(zhp) != 0)
+				ret = 1;
+		} else if (zpool_export(zhp, force) != 0) {
 			ret = 1;
+		}
 
 		zpool_close(zhp);
 	}
diff --git a/cddl/contrib/opensolaris/cmd/ztest/ztest.c b/cddl/contrib/opensolaris/cmd/ztest/ztest.c
index b7ca302d3dd..9744a0a9729 100644
--- a/cddl/contrib/opensolaris/cmd/ztest/ztest.c
+++ b/cddl/contrib/opensolaris/cmd/ztest/ztest.c
@@ -3039,7 +3039,7 @@ ztest_spa_import_export(char *oldname, char *newname)
 	/*
 	 * Export it.
 	 */
-	error = spa_export(oldname, &config, B_FALSE);
+	error = spa_export(oldname, &config, B_FALSE, B_FALSE);
 	if (error)
 		fatal(0, "spa_export('%s') = %d", oldname, error);
 
diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h
index a959494cd52..a77317ef9fa 100644
--- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h
+++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs.h
@@ -289,6 +289,7 @@ extern int zpool_get_errlog(zpool_handle_t *, nvlist_t **);
  * Import and export functions
  */
 extern int zpool_export(zpool_handle_t *, boolean_t);
+extern int zpool_export_force(zpool_handle_t *);
 extern int zpool_import(libzfs_handle_t *, nvlist_t *, const char *,
     char *altroot);
 extern int zpool_import_props(libzfs_handle_t *, nvlist_t *, const char *,
diff --git a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c
index c3cb5470196..0369062bbc3 100644
--- a/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c
+++ b/cddl/contrib/opensolaris/lib/libzfs/common/libzfs_pool.c
@@ -1096,7 +1096,7 @@ zpool_add(zpool_handle_t *zhp, nvlist_t *nvroot)
  * mounted datasets in the pool.
  */
 int
-zpool_export(zpool_handle_t *zhp, boolean_t force)
+zpool_export_common(zpool_handle_t *zhp, boolean_t force, boolean_t hardforce)
 {
 	zfs_cmd_t zc = { 0 };
 	char msg[1024];
@@ -1109,6 +1109,7 @@ zpool_export(zpool_handle_t *zhp, boolean_t force)
 
 	(void) strlcpy(zc.zc_name, zhp->zpool_name, sizeof (zc.zc_name));
 	zc.zc_cookie = force;
+	zc.zc_guid = hardforce;
 
 	if (zfs_ioctl(zhp->zpool_hdl, ZFS_IOC_POOL_EXPORT, &zc) != 0) {
 		switch (errno) {
@@ -1129,6 +1130,18 @@ zpool_export(zpool_handle_t *zhp, boolean_t force)
 	return (0);
 }
 
+int
+zpool_export(zpool_handle_t *zhp, boolean_t force)
+{
+	return (zpool_export_common(zhp, force, B_FALSE));
+}
+
+int
+zpool_export_force(zpool_handle_t *zhp)
+{
+	return (zpool_export_common(zhp, B_TRUE, B_TRUE));
+}
+
 /*
  * zpool_import() is a contracted interface. Should be kept the same
  * if possible.
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c
index 90861bab4c7..888b8829dd0 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/spa.c
@@ -2564,11 +2564,12 @@ spa_tryimport(nvlist_t *tryconfig)
  * The act of destroying or exporting a pool is very simple.  We make sure there
  * is no more pending I/O and any references to the pool are gone.  Then, we
  * update the pool state and sync all the labels to disk, removing the
- * configuration from the cache afterwards.
+ * configuration from the cache afterwards. If the 'hardforce' flag is set, then
+ * we don't sync the labels or remove the configuration cache.
  */
 static int
 spa_export_common(char *pool, int new_state, nvlist_t **oldconfig,
-    boolean_t force)
+    boolean_t force, boolean_t hardforce)
 {
 	spa_t *spa;
 
@@ -2636,7 +2637,7 @@ spa_export_common(char *pool, int new_state, nvlist_t **oldconfig,
 		 * so mark them all dirty.  spa_unload() will do the
 		 * final sync that pushes these changes out.
 		 */
-		if (new_state != POOL_STATE_UNINITIALIZED) {
+		if (new_state != POOL_STATE_UNINITIALIZED && !hardforce) {
 			spa_config_enter(spa, SCL_ALL, FTAG, RW_WRITER);
 			spa->spa_state = new_state;
 			spa->spa_final_txg = spa_last_synced_txg(spa) + 1;
@@ -2656,7 +2657,8 @@ spa_export_common(char *pool, int new_state, nvlist_t **oldconfig,
 		VERIFY(nvlist_dup(spa->spa_config, oldconfig, 0) == 0);
 
 	if (new_state != POOL_STATE_UNINITIALIZED) {
-		spa_config_sync(spa, B_TRUE, B_TRUE);
+		if (!hardforce)
+			spa_config_sync(spa, B_TRUE, B_TRUE);
 		spa_remove(spa);
 	}
 	mutex_exit(&spa_namespace_lock);
@@ -2670,16 +2672,19 @@ spa_export_common(char *pool, int new_state, nvlist_t **oldconfig,
 int
 spa_destroy(char *pool)
 {
-	return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL, B_FALSE));
+	return (spa_export_common(pool, POOL_STATE_DESTROYED, NULL,
+	    B_FALSE, B_FALSE));
 }
 
 /*
  * Export a storage pool.
  */
 int
-spa_export(char *pool, nvlist_t **oldconfig, boolean_t force)
+spa_export(char *pool, nvlist_t **oldconfig, boolean_t force,
+    boolean_t hardforce)
 {
-	return (spa_export_common(pool, POOL_STATE_EXPORTED, oldconfig, force));
+	return (spa_export_common(pool, POOL_STATE_EXPORTED, oldconfig,
+	    force, hardforce));
 }
 
 /*
@@ -2690,7 +2695,7 @@ int
 spa_reset(char *pool)
 {
 	return (spa_export_common(pool, POOL_STATE_UNINITIALIZED, NULL,
-	    B_FALSE));
+	    B_FALSE, B_FALSE));
 }
 
 /*
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/spa.h b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/spa.h
index 99bcb915911..b0b758b3716 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/spa.h
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/sys/spa.h
@@ -333,7 +333,8 @@ extern int spa_import(const char *pool, nvlist_t *config, nvlist_t *props);
 extern int spa_import_faulted(const char *, nvlist_t *, nvlist_t *);
 extern nvlist_t *spa_tryimport(nvlist_t *tryconfig);
 extern int spa_destroy(char *pool);
-extern int spa_export(char *pool, nvlist_t **oldconfig, boolean_t force);
+extern int spa_export(char *pool, nvlist_t **oldconfig, boolean_t force,
+    boolean_t hardforce);
 extern int spa_reset(char *pool);
 extern void spa_async_request(spa_t *spa, int flag);
 extern void spa_async_unrequest(spa_t *spa, int flag);
diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
index 1536e0b4da6..c7bd8ad56c1 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_ioctl.c
@@ -882,9 +882,10 @@ zfs_ioc_pool_export(zfs_cmd_t *zc)
 {
 	int error;
 	boolean_t force = (boolean_t)zc->zc_cookie;
+	boolean_t hardforce = (boolean_t)zc->zc_guid;
 
 	zfs_log_history(zc);
-	error = spa_export(zc->zc_name, NULL, force);
+	error = spa_export(zc->zc_name, NULL, force, hardforce);
 	return (error);
 }
 

From 9061251f9ac7d2b9be5cceb89d0611a92b10d22e Mon Sep 17 00:00:00 2001
From: Jaakko Heinonen <jh@FreeBSD.org>
Date: Wed, 5 May 2010 18:53:24 +0000
Subject: [PATCH 475/532] Fix deadlock between GEOM class unloading and
 withering. Withering can't proceed while g_unload_class() blocks the event
 thread. Fix this by not running g_unload_class() as a GEOM event and dropping
 the topology lock when withering needs to proceed.

PR:		kern/139847
Silence on:	freebsd-geom
---
 sys/geom/geom.h      |   3 ++
 sys/geom/geom_subr.c | 100 ++++++++++++++++++++++---------------------
 2 files changed, 55 insertions(+), 48 deletions(-)

diff --git a/sys/geom/geom.h b/sys/geom/geom.h
index 82d04568aee..fe18a15e8c6 100644
--- a/sys/geom/geom.h
+++ b/sys/geom/geom.h
@@ -353,6 +353,9 @@ g_free(void *ptr)
 		sx_assert(&topology_lock, SX_UNLOCKED);		\
 	} while (0)
 
+#define g_topology_sleep(chan, timo)				\
+	sx_sleep(chan, &topology_lock, 0, "gtopol", timo)
+
 #define DECLARE_GEOM_CLASS(class, name) 			\
 	static moduledata_t name##_mod = {			\
 		#name, g_modevent, &class			\
diff --git a/sys/geom/geom_subr.c b/sys/geom/geom_subr.c
index 772277e02a7..b001c5fcaed 100644
--- a/sys/geom/geom_subr.c
+++ b/sys/geom/geom_subr.c
@@ -134,65 +134,73 @@ g_load_class(void *arg, int flag)
 	}
 }
 
-static void
-g_unload_class(void *arg, int flag)
+static int
+g_unload_class(struct g_class *mp)
 {
-	struct g_hh00 *hh;
-	struct g_class *mp;
 	struct g_geom *gp;
 	struct g_provider *pp;
 	struct g_consumer *cp;
 	int error;
 
-	g_topology_assert();
-	hh = arg;
-	mp = hh->mp;
-	G_VALID_CLASS(mp);
+	g_topology_lock();
 	g_trace(G_T_TOPOLOGY, "g_unload_class(%s)", mp->name);
+retry:
+	G_VALID_CLASS(mp);
+	LIST_FOREACH(gp, &mp->geom, geom) {
+		/* We refuse to unload if anything is open */
+		LIST_FOREACH(pp, &gp->provider, provider)
+			if (pp->acr || pp->acw || pp->ace) {
+				g_topology_unlock();
+				return (EBUSY);
+			}
+		LIST_FOREACH(cp, &gp->consumer, consumer)
+			if (cp->acr || cp->acw || cp->ace) {
+				g_topology_unlock();
+				return (EBUSY);
+			}
+		/* If the geom is withering, wait for it to finish. */
+		if (gp->flags & G_GEOM_WITHER) {
+			g_topology_sleep(mp, 1);
+			goto retry;
+		}
+	}
 
 	/*
 	 * We allow unloading if we have no geoms, or a class
 	 * method we can use to get rid of them.
 	 */
 	if (!LIST_EMPTY(&mp->geom) && mp->destroy_geom == NULL) {
-		hh->error = EOPNOTSUPP;
-		return;
-	}
-
-	/* We refuse to unload if anything is open */
-	LIST_FOREACH(gp, &mp->geom, geom) {
-		LIST_FOREACH(pp, &gp->provider, provider)
-			if (pp->acr || pp->acw || pp->ace) {
-				hh->error = EBUSY;
-				return;
-			}
-		LIST_FOREACH(cp, &gp->consumer, consumer)
-			if (cp->acr || cp->acw || cp->ace) {
-				hh->error = EBUSY;
-				return;
-			}
+		g_topology_unlock();
+		return (EOPNOTSUPP);
 	}
 
 	/* Bar new entries */
 	mp->taste = NULL;
 	mp->config = NULL;
 
-	error = 0;
+	LIST_FOREACH(gp, &mp->geom, geom) {
+		error = mp->destroy_geom(NULL, mp, gp);
+		if (error != 0) {
+			g_topology_unlock();
+			return (error);
+		}
+	}
+	/* Wait for withering to finish. */
 	for (;;) {
 		gp = LIST_FIRST(&mp->geom);
 		if (gp == NULL)
 			break;
-		error = mp->destroy_geom(NULL, mp, gp);
-		if (error != 0)
-			break;
+		KASSERT(gp->flags & G_GEOM_WITHER,
+		   ("Non-withering geom in class %s", mp->name));
+		g_topology_sleep(mp, 1);
 	}
-	if (error == 0) {
-		if (mp->fini != NULL)
-			mp->fini(mp);
-		LIST_REMOVE(mp, class);
-	}
-	hh->error = error;
-	return;
+	G_VALID_CLASS(mp);
+	if (mp->fini != NULL)
+		mp->fini(mp);
+	LIST_REMOVE(mp, class);
+	g_topology_unlock();
+
+	return (0);
 }
 
 int
@@ -213,12 +221,12 @@ g_modevent(module_t mod, int type, void *data)
 		g_ignition++;
 		g_init();
 	}
-	hh = g_malloc(sizeof *hh, M_WAITOK | M_ZERO);
-	hh->mp = data;
 	error = EOPNOTSUPP;
 	switch (type) {
 	case MOD_LOAD:
-		g_trace(G_T_TOPOLOGY, "g_modevent(%s, LOAD)", hh->mp->name);
+		g_trace(G_T_TOPOLOGY, "g_modevent(%s, LOAD)", mp->name);
+		hh = g_malloc(sizeof *hh, M_WAITOK | M_ZERO);
+		hh->mp = mp;
 		/*
 		 * Once the system is not cold, MOD_LOAD calls will be
 		 * from the userland and the g_event thread will be able
@@ -236,18 +244,14 @@ g_modevent(module_t mod, int type, void *data)
 		}
 		break;
 	case MOD_UNLOAD:
-		g_trace(G_T_TOPOLOGY, "g_modevent(%s, UNLOAD)", hh->mp->name);
-		error = g_waitfor_event(g_unload_class, hh, M_WAITOK, NULL);
-		if (error == 0)
-			error = hh->error;
+		g_trace(G_T_TOPOLOGY, "g_modevent(%s, UNLOAD)", mp->name);
+		DROP_GIANT();
+		error = g_unload_class(mp);
+		PICKUP_GIANT();
 		if (error == 0) {
-			KASSERT(LIST_EMPTY(&hh->mp->geom),
-			    ("Unloaded class (%s) still has geom", hh->mp->name));
+			KASSERT(LIST_EMPTY(&mp->geom),
+			    ("Unloaded class (%s) still has geom", mp->name));
 		}
-		g_free(hh);
-		break;
-	default:
-		g_free(hh);
 		break;
 	}
 	return (error);

From 8e0ad55abb607c30abd20e5a1c0c797e0e8772ca Mon Sep 17 00:00:00 2001
From: Joel Dahl <joel@FreeBSD.org>
Date: Wed, 5 May 2010 20:39:02 +0000
Subject: [PATCH 477/532] Switch to our preferred 2-clause BSD license.

Approved by:	kmacy
---
 sys/amd64/include/xen/xenfunc.h | 31 +++++++++++------------
 sys/amd64/include/xen/xenvar.h  | 26 +++++++++----------
 sys/dev/cxgb/sys/mvec.h         | 42 +++++++++++++++---------------
 sys/dev/cxgb/sys/uipc_mvec.c    | 44 +++++++++++++++-----------------
 sys/dev/xen/netfront/netfront.c | 32 ++++++++++++++---------
 sys/i386/include/xen/xenfunc.h  | 29 ++++++++++-----------
 sys/i386/include/xen/xenvar.h   | 27 ++++++++++----------
 sys/kern/subr_bufring.c         | 45 +++++++++++++++------------------
 sys/sys/buf_ring.h              | 42 +++++++++++++++---------------
 9 files changed, 154 insertions(+), 164 deletions(-)

diff --git a/sys/amd64/include/xen/xenfunc.h b/sys/amd64/include/xen/xenfunc.h
index b3a6672576c..d03d4f685e2 100644
--- a/sys/amd64/include/xen/xenfunc.h
+++ b/sys/amd64/include/xen/xenfunc.h
@@ -1,6 +1,5 @@
-/*
- *
- * Copyright (c) 2004,2005 Kip Macy
+/*-
+ * Copyright (c) 2004, 2005 Kip Macy
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -11,22 +10,22 @@
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
  *    documentation and/or other materials provided with the distribution.
- * 4. The name of the author may not be used to endorse or promote products
- *    derived from this software without specific prior written permission.
  *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ *
+ * $FreeBSD$
  */
 
-
 #ifndef _XEN_XENFUNC_H_
 #define _XEN_XENFUNC_H_
 
diff --git a/sys/amd64/include/xen/xenvar.h b/sys/amd64/include/xen/xenvar.h
index 1433b76871e..d9dbc5d9186 100644
--- a/sys/amd64/include/xen/xenvar.h
+++ b/sys/amd64/include/xen/xenvar.h
@@ -1,29 +1,27 @@
-/*
+/*-
  * Copyright (c) 2008 Kip Macy
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
- *
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
- *
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
  *    documentation and/or other materials provided with the distribution.
  *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
  * $FreeBSD$
  */
diff --git a/sys/dev/cxgb/sys/mvec.h b/sys/dev/cxgb/sys/mvec.h
index ed38dca3444..9db27ddf75d 100644
--- a/sys/dev/cxgb/sys/mvec.h
+++ b/sys/dev/cxgb/sys/mvec.h
@@ -1,33 +1,31 @@
-/**************************************************************************
- *
- * Copyright (c) 2007,2009 Kip Macy kmacy@freebsd.org
+/*-
+ * Copyright (c) 2007, 2009 Kip Macy <kmacy@freebsd.org>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. The name of Kip Macy nor the names of other
- *    contributors may be used to endorse or promote products derived from
- *    this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
  * $FreeBSD$
  *
- ***************************************************************************/
+ */
 
 #ifndef _MVEC_H_
 #define _MVEC_H_
diff --git a/sys/dev/cxgb/sys/uipc_mvec.c b/sys/dev/cxgb/sys/uipc_mvec.c
index 4e494ebfc15..f52daa681b1 100644
--- a/sys/dev/cxgb/sys/uipc_mvec.c
+++ b/sys/dev/cxgb/sys/uipc_mvec.c
@@ -1,32 +1,28 @@
-/**************************************************************************
- *
- * Copyright (c) 2007-2008, Kip Macy kmacy@freebsd.org
+/*-
+ * Copyright (c) 2007-2008 Kip Macy <kmacy@freebsd.org>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. The name of Kip Macy nor the names of other
- *    contributors may be used to endorse or promote products derived from
- *    this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
- *
- *
- ***************************************************************************/
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
 
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
diff --git a/sys/dev/xen/netfront/netfront.c b/sys/dev/xen/netfront/netfront.c
index d67d3544384..15cf455bff7 100644
--- a/sys/dev/xen/netfront/netfront.c
+++ b/sys/dev/xen/netfront/netfront.c
@@ -1,19 +1,27 @@
-/*
- *
+/*-
  * Copyright (c) 2004-2006 Kip Macy
  * All rights reserved.
  *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  */
 
 
diff --git a/sys/i386/include/xen/xenfunc.h b/sys/i386/include/xen/xenfunc.h
index 2851709cb95..47f04057aa6 100644
--- a/sys/i386/include/xen/xenfunc.h
+++ b/sys/i386/include/xen/xenfunc.h
@@ -1,6 +1,5 @@
-/*
- *
- * Copyright (c) 2004,2005 Kip Macy
+/*-
+ * Copyright (c) 2004, 2005 Kip Macy
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
@@ -11,24 +10,22 @@
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
  *    documentation and/or other materials provided with the distribution.
- * 4. The name of the author may not be used to endorse or promote products
- *    derived from this software without specific prior written permission.
  *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
  * $FreeBSD$
  */
 
-
 #ifndef _XEN_XENFUNC_H_
 #define _XEN_XENFUNC_H_
 
diff --git a/sys/i386/include/xen/xenvar.h b/sys/i386/include/xen/xenvar.h
index cefbb058c1c..1110f8506ae 100644
--- a/sys/i386/include/xen/xenvar.h
+++ b/sys/i386/include/xen/xenvar.h
@@ -1,32 +1,31 @@
-/*
+/*-
  * Copyright (c) 2008 Kip Macy
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
  * modification, are permitted provided that the following conditions
  * are met:
- *
  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer.
- *
  * 2. Redistributions in binary form must reproduce the above copyright
  *    notice, this list of conditions and the following disclaimer in the
  *    documentation and/or other materials provided with the distribution.
  *
- * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
- * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
- * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
- * IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
- * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
- * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
- * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
- * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
- * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
- * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
- *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
  * $FreeBSD$
  */
+
 #ifndef XENVAR_H_
 #define XENVAR_H_
 #define XBOOTUP 0x1
diff --git a/sys/kern/subr_bufring.c b/sys/kern/subr_bufring.c
index 63938eaca44..4cd39292622 100644
--- a/sys/kern/subr_bufring.c
+++ b/sys/kern/subr_bufring.c
@@ -1,32 +1,29 @@
-/**************************************************************************
- *
- * Copyright (c) 2007,2008 Kip Macy kmacy@freebsd.org
+/*-
+ * Copyright (c) 2007, 2008 Kip Macy <kmacy@freebsd.org>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. The name of Kip Macy nor the names of other
- *    contributors may be used to endorse or promote products derived from
- *    this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
- *
- *
- ***************************************************************************/
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
+ */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/sys/sys/buf_ring.h b/sys/sys/buf_ring.h
index efa667de189..57e42e5c189 100644
--- a/sys/sys/buf_ring.h
+++ b/sys/sys/buf_ring.h
@@ -1,33 +1,31 @@
-/**************************************************************************
- *
- * Copyright (c) 2007-2009 Kip Macy kmacy@freebsd.org
+/*-
+ * Copyright (c) 2007-2009 Kip Macy <kmacy@freebsd.org>
  * All rights reserved.
  *
  * Redistribution and use in source and binary forms, with or without
- * modification, are permitted provided that the following conditions are met:
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
  *
- * 1. Redistributions of source code must retain the above copyright notice,
- *    this list of conditions and the following disclaimer.
- *
- * 2. The name of Kip Macy nor the names of other
- *    contributors may be used to endorse or promote products derived from
- *    this software without specific prior written permission.
- *
- * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
- * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
+ * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
- * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
- * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
- * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
- * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
- * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
- * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
- * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
- * POSSIBILITY OF SUCH DAMAGE.
+ * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
+ * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
+ * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
+ * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
+ * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
+ * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
+ * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
+ * SUCH DAMAGE.
  *
  * $FreeBSD$
  *
- ***************************************************************************/
+ */
 
 #ifndef	_SYS_BUF_RING_H_
 #define	_SYS_BUF_RING_H_

From db8fd40e9fabb7207d19d8828480ca62c328d685 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Wed, 5 May 2010 21:07:47 +0000
Subject: [PATCH 478/532] Add definitions for Intel AESNI CPUID bits and print
 the capabilities on boot.

Hardware provided by:	Sentex Communications
MFC after:	1 week
---
 sys/amd64/amd64/identcpu.c     | 4 ++--
 sys/amd64/include/specialreg.h | 2 ++
 sys/i386/i386/identcpu.c       | 4 ++--
 sys/i386/include/specialreg.h  | 2 ++
 4 files changed, 8 insertions(+), 4 deletions(-)

diff --git a/sys/amd64/amd64/identcpu.c b/sys/amd64/amd64/identcpu.c
index c23102a5a2f..52e7568f7c0 100644
--- a/sys/amd64/amd64/identcpu.c
+++ b/sys/amd64/amd64/identcpu.c
@@ -240,7 +240,7 @@ printcpuinfo(void)
 				printf("\n  Features2=0x%b", cpu_feature2,
 				"\020"
 				"\001SSE3"	/* SSE3 */
-				"\002<b1>"
+				"\002PCLMULQDQ"	/* Carry-Less Mul Quadword */
 				"\003DTES64"	/* 64-bit Debug Trace */
 				"\004MON"	/* MONITOR/MWAIT Instructions */
 				"\005DS_CPL"	/* CPL Qualified Debug Store */
@@ -264,7 +264,7 @@ printcpuinfo(void)
 				"\027MOVBE"
 				"\030POPCNT"
 				"\031<b24>"
-				"\032<b25>"
+				"\032AESNI"	/* AES Crypto*/
 				"\033XSAVE"
 				"\034OSXSAVE"
 				"\035<b28>"
diff --git a/sys/amd64/include/specialreg.h b/sys/amd64/include/specialreg.h
index 86a08cec931..895619cf6f7 100644
--- a/sys/amd64/include/specialreg.h
+++ b/sys/amd64/include/specialreg.h
@@ -113,6 +113,7 @@
 #define	CPUID_PBE	0x80000000
 
 #define	CPUID2_SSE3	0x00000001
+#define	CPUID2_PCLMULQDQ 0x00000002
 #define	CPUID2_DTES64	0x00000004
 #define	CPUID2_MON	0x00000008
 #define	CPUID2_DS_CPL	0x00000010
@@ -131,6 +132,7 @@
 #define	CPUID2_X2APIC	0x00200000
 #define	CPUID2_MOVBE	0x00400000
 #define	CPUID2_POPCNT	0x00800000
+#define	CPUID2_AESNI	0x02000000
 
 /*
  * Important bits in the AMD extended cpuid flags
diff --git a/sys/i386/i386/identcpu.c b/sys/i386/i386/identcpu.c
index 931bfaf6777..6f2e4a40fd1 100644
--- a/sys/i386/i386/identcpu.c
+++ b/sys/i386/i386/identcpu.c
@@ -727,7 +727,7 @@ printcpuinfo(void)
 				printf("\n  Features2=0x%b", cpu_feature2,
 				"\020"
 				"\001SSE3"	/* SSE3 */
-				"\002<b1>"
+				"\002PCLMULQDQ"	/* Carry-Less Mul Quadword */
 				"\003DTES64"	/* 64-bit Debug Trace */
 				"\004MON"	/* MONITOR/MWAIT Instructions */
 				"\005DS_CPL"	/* CPL Qualified Debug Store */
@@ -751,7 +751,7 @@ printcpuinfo(void)
 				"\027MOVBE"
 				"\030POPCNT"
 				"\031<b24>"
-				"\032<b25>"
+				"\032AESNI"	/* AES Crypto*/
 				"\033XSAVE"
 				"\034OSXSAVE"
 				"\035<b28>"
diff --git a/sys/i386/include/specialreg.h b/sys/i386/include/specialreg.h
index d2494c72100..efcf9244869 100644
--- a/sys/i386/include/specialreg.h
+++ b/sys/i386/include/specialreg.h
@@ -110,6 +110,7 @@
 #define	CPUID_PBE	0x80000000
 
 #define	CPUID2_SSE3	0x00000001
+#define	CPUID2_PCLMULQDQ 0x00000002
 #define	CPUID2_DTES64	0x00000004
 #define	CPUID2_MON	0x00000008
 #define	CPUID2_DS_CPL	0x00000010
@@ -128,6 +129,7 @@
 #define	CPUID2_X2APIC	0x00200000
 #define	CPUID2_MOVBE	0x00400000
 #define	CPUID2_POPCNT	0x00800000
+#define	CPUID2_AESNI	0x02000000
 
 /*
  * Important bits in the AMD extended cpuid flags

From d18129ea1d0a63a8736020b2e6a3c241976511b5 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Wed, 5 May 2010 21:24:18 +0000
Subject: [PATCH 479/532] Move SCCS tags to comments as they were already #if
 0'ed.

---
 usr.bin/find/find.c     | 9 ++-------
 usr.bin/find/function.c | 8 ++------
 usr.bin/find/ls.c       | 8 ++------
 usr.bin/find/main.c     | 8 ++------
 usr.bin/find/misc.c     | 9 ++-------
 usr.bin/find/operator.c | 8 ++------
 usr.bin/find/option.c   | 8 ++------
 7 files changed, 14 insertions(+), 44 deletions(-)

diff --git a/usr.bin/find/find.c b/usr.bin/find/find.c
index cc2d797a541..35ef5b25f46 100644
--- a/usr.bin/find/find.c
+++ b/usr.bin/find/find.c
@@ -32,15 +32,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)find.c	8.5 (Berkeley) 8/5/94
  */
 
-#ifndef lint
-#if 0
-static char sccsid[] = "@(#)find.c	8.5 (Berkeley) 8/5/94";
-#else
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/function.c b/usr.bin/find/function.c
index 1714627d1fc..4329887b09e 100644
--- a/usr.bin/find/function.c
+++ b/usr.bin/find/function.c
@@ -32,14 +32,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)function.c	8.10 (Berkeley) 5/4/95
  */
 
-#ifndef lint
-#if 0
-static const char sccsid[] = "@(#)function.c	8.10 (Berkeley) 5/4/95";
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/ls.c b/usr.bin/find/ls.c
index 88e45938cc6..e96994b7184 100644
--- a/usr.bin/find/ls.c
+++ b/usr.bin/find/ls.c
@@ -29,14 +29,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)ls.c	8.1 (Berkeley) 6/6/93
  */
 
-#ifndef lint
-#if 0
-static char sccsid[] = "@(#)ls.c	8.1 (Berkeley) 6/6/93";
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/main.c b/usr.bin/find/main.c
index 8e2b42cd298..7d4c24ac41a 100644
--- a/usr.bin/find/main.c
+++ b/usr.bin/find/main.c
@@ -32,6 +32,8 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)main.c	8.4 (Berkeley) 5/4/95
  */
 
 #ifndef lint
@@ -40,12 +42,6 @@ char copyright[] =
 	The Regents of the University of California.  All rights reserved.\n";
 #endif /* not lint */
 
-#ifndef lint
-#if 0
-static char sccsid[] = "@(#)main.c	8.4 (Berkeley) 5/4/95";
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/misc.c b/usr.bin/find/misc.c
index 15329064cf4..3696c13305f 100644
--- a/usr.bin/find/misc.c
+++ b/usr.bin/find/misc.c
@@ -32,15 +32,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)misc.c	8.2 (Berkeley) 4/1/94
  */
 
-#ifndef lint
-#if 0
-static char sccsid[] = "@(#)misc.c	8.2 (Berkeley) 4/1/94";
-#else
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/operator.c b/usr.bin/find/operator.c
index c774efa0dc8..d1a7ea8e807 100644
--- a/usr.bin/find/operator.c
+++ b/usr.bin/find/operator.c
@@ -32,14 +32,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)operator.c	8.1 (Berkeley) 6/6/93
  */
 
-#ifndef lint
-#if 0
-static char sccsid[] = "@(#)operator.c	8.1 (Berkeley) 6/6/93";
-#endif
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/option.c b/usr.bin/find/option.c
index 7d06c910ae7..6ae3958315b 100644
--- a/usr.bin/find/option.c
+++ b/usr.bin/find/option.c
@@ -32,14 +32,10 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
+ *
+ *	@(#)option.c	8.2 (Berkeley) 4/16/94
  */
 
-#ifndef lint
-/*
-static char sccsid[] = "@(#)option.c	8.2 (Berkeley) 4/16/94";
-*/
-#endif /* not lint */
-
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 

From e4b50334ec633d9ca7dbb970b48c72753a0a9d66 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Wed, 5 May 2010 21:48:40 +0000
Subject: [PATCH 480/532] sh: Apply locale vars on builtins, recognize
 LC_MESSAGES as a locale var.

This allows doing things like LC_ALL=C some_builtin to run a builtin under a
different locale, just like is possible with external programs. The
immediate reason is that this allows making printf(1) a builtin without
breaking things like LC_NUMERIC=C printf '%f\n' 1.2

This change also affects special builtins, as even though the assignment is
persistent, the export is only to the builtin (unless the variable was
already exported).

Note: for this to work for builtins that also exist as external programs
such as /bin/test, the setlocale() call must be under #ifndef SHELL. The
shell will do the setlocale() calls which may not agree with the environment
variables.
---
 bin/sh/eval.c                              |   4 +
 bin/sh/var.c                               |  75 ++++++++++--
 bin/sh/var.h                               |   2 +
 tools/regression/bin/sh/builtins/locale1.0 | 133 +++++++++++++++++++++
 4 files changed, 207 insertions(+), 7 deletions(-)
 create mode 100644 tools/regression/bin/sh/builtins/locale1.0

diff --git a/bin/sh/eval.c b/bin/sh/eval.c
index e93fb147b57..c4495d04275 100644
--- a/bin/sh/eval.c
+++ b/bin/sh/eval.c
@@ -937,6 +937,8 @@ evalcommand(union node *cmd, int flags, struct backcmd *backcmd)
 			cmdentry.special = 1;
 		if (cmdentry.special)
 			listsetvar(cmdenviron);
+		if (argc > 0)
+			bltinsetlocale();
 		commandname = argv[0];
 		argptr = argv + 1;
 		nextopt_optptr = NULL;		/* initialize nextopt */
@@ -944,6 +946,8 @@ evalcommand(union node *cmd, int flags, struct backcmd *backcmd)
 		exitstatus = (*builtinfunc[cmdentry.u.index])(argc, argv);
 		flushall();
 cmddone:
+		if (argc > 0)
+			bltinunsetlocale();
 		cmdenviron = NULL;
 		out1 = &output;
 		out2 = &errout;
diff --git a/bin/sh/var.c b/bin/sh/var.c
index d46e2c3e464..75de239b775 100644
--- a/bin/sh/var.c
+++ b/bin/sh/var.c
@@ -122,6 +122,14 @@ STATIC const struct varinit varinit[] = {
 
 STATIC struct var *vartab[VTABSIZE];
 
+STATIC const char *const locale_names[7] = {
+	"LC_COLLATE", "LC_CTYPE", "LC_MONETARY",
+	"LC_NUMERIC", "LC_TIME", "LC_MESSAGES", NULL
+};
+STATIC const int locale_categories[7] = {
+	LC_COLLATE, LC_CTYPE, LC_MONETARY, LC_NUMERIC, LC_TIME, LC_MESSAGES, 0
+};
+
 STATIC struct var **hashvar(const char *);
 STATIC int varequal(const char *, const char *);
 STATIC int localevar(const char *);
@@ -258,11 +266,7 @@ setvar(const char *name, const char *val, int flags)
 STATIC int
 localevar(const char *s)
 {
-	static const char *lnames[7] = {
-		"ALL", "COLLATE", "CTYPE", "MONETARY",
-		"NUMERIC", "TIME", NULL
-	};
-	const char **ss;
+	const char *const *ss;
 
 	if (*s != 'L')
 		return 0;
@@ -270,8 +274,10 @@ localevar(const char *s)
 		return 1;
 	if (strncmp(s + 1, "C_", 2) != 0)
 		return 0;
-	for (ss = lnames; *ss ; ss++)
-		if (varequal(s + 3, *ss))
+	if (varequal(s + 3, "ALL"))
+		return 1;
+	for (ss = locale_names; *ss ; ss++)
+		if (varequal(s + 3, *ss + 3))
 			return 1;
 	return 0;
 }
@@ -437,6 +443,61 @@ bltinlookup(const char *name, int doall)
 }
 
 
+/*
+ * Set up locale for a builtin (LANG/LC_* assignments).
+ */
+void
+bltinsetlocale(void)
+{
+	struct strlist *lp;
+	int act = 0;
+	char *loc, *locdef;
+	int i;
+
+	for (lp = cmdenviron ; lp ; lp = lp->next) {
+		if (localevar(lp->text)) {
+			act = 1;
+			break;
+		}
+	}
+	if (!act)
+		return;
+	loc = bltinlookup("LC_ALL", 0);
+	INTOFF;
+	if (loc != NULL) {
+		setlocale(LC_ALL, loc);
+		INTON;
+		return;
+	}
+	locdef = bltinlookup("LANG", 0);
+	for (i = 0; locale_names[i] != NULL; i++) {
+		loc = bltinlookup(locale_names[i], 0);
+		if (loc == NULL)
+			loc = locdef;
+		if (loc != NULL)
+			setlocale(locale_categories[i], loc);
+	}
+	INTON;
+}
+
+/*
+ * Undo the effect of bltinlocaleset().
+ */
+void
+bltinunsetlocale(void)
+{
+	struct strlist *lp;
+
+	INTOFF;
+	for (lp = cmdenviron ; lp ; lp = lp->next) {
+		if (localevar(lp->text)) {
+			setlocale(LC_ALL, "");
+			return;
+		}
+	}
+	INTON;
+}
+
 
 /*
  * Generate a list of exported variables.  This routine is used to construct
diff --git a/bin/sh/var.h b/bin/sh/var.h
index 9c792c846ca..6a02630842d 100644
--- a/bin/sh/var.h
+++ b/bin/sh/var.h
@@ -107,6 +107,8 @@ struct strlist;
 void listsetvar(struct strlist *);
 char *lookupvar(const char *);
 char *bltinlookup(const char *, int);
+void bltinsetlocale(void);
+void bltinunsetlocale(void);
 char **environment(void);
 int showvarscmd(int, char **);
 int exportcmd(int, char **);
diff --git a/tools/regression/bin/sh/builtins/locale1.0 b/tools/regression/bin/sh/builtins/locale1.0
new file mode 100644
index 00000000000..93a37dfc876
--- /dev/null
+++ b/tools/regression/bin/sh/builtins/locale1.0
@@ -0,0 +1,133 @@
+# $FreeBSD$
+# Note: this test depends on strerror() using locale.
+
+failures=0
+
+check() {
+	if ! eval "[ $1 ]"; then
+		echo "Failed: $1 at $2"
+		: $((failures += 1))
+	fi
+}
+
+unset LANG LC_ALL LC_COLLATE LC_CTYPE LC_MONETARY LC_NUMERIC LC_TIME LC_MESSAGES
+
+msgeng="No such file or directory"
+msgdut="Bestand of map niet gevonden"
+
+# Verify C locale error message.
+case $(command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+# Various locale variables that should not affect the message.
+case $(LC_ALL=C command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_ALL=C LANG=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_ALL=C LC_MESSAGES=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_CTYPE=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+# Verify Dutch message.
+case $(export LANG=nl_NL.ISO8859-1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(export LC_MESSAGES=nl_NL.ISO8859-1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(export LC_ALL=nl_NL.ISO8859-1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LANG=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_MESSAGES=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_ALL=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+# Verify that command assignments do not set the locale persistently.
+case $(command . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LANG=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_MESSAGES=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(LC_ALL=nl_NL.ISO8859-1 command . /var/empty/foo 2>&1; command . /var/empty/foo 2>&1) in
+	*"$msgdut"*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+# Check special builtin; add colon invocation to avoid depending on certain fix.
+case $(LC_ALL=nl_NL.ISO8859-1 . /var/empty/foo 2>&1; :) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+# Assignments on special builtins are exported to that builtin; the export
+# is not persistent.
+case $(LC_ALL=nl_NL.ISO8859-1 . /dev/null; . /var/empty/foo 2>&1) in
+	*"$msgeng"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+case $(export LC_ALL; LC_ALL=nl_NL.ISO8859-1 . /dev/null; . /var/empty/foo 2>&1) in
+	*"$msgdut"*) ok=1 ;;
+	*) ok=0 ;;
+esac
+check '$ok -eq 1' $LINENO
+
+exit $((failures > 0))

From f8aab721b2d046452ffa9347cb20ab26f62d7b9a Mon Sep 17 00:00:00 2001
From: Marko Zec <zec@FreeBSD.org>
Date: Wed, 5 May 2010 22:06:05 +0000
Subject: [PATCH 481/532] Add an optional "persistent" flag to ng_hub and
 ng_bridge, which if set, disables automatic node shutdown when the last hook
 gets disconnected.

Reviewed by:	julian
---
 share/man/man4/ng_bridge.4 | 10 ++++--
 share/man/man4/ng_hub.4    | 15 ++++++--
 sys/netgraph/ng_bridge.c   | 18 ++++++++--
 sys/netgraph/ng_bridge.h   |  1 +
 sys/netgraph/ng_hub.c      | 74 +++++++++++++++++++++++++++++++++++++-
 sys/netgraph/ng_hub.h      |  5 +++
 6 files changed, 115 insertions(+), 8 deletions(-)

diff --git a/share/man/man4/ng_bridge.4 b/share/man/man4/ng_bridge.4
index 3b1dd71cb62..8884e339558 100644
--- a/share/man/man4/ng_bridge.4
+++ b/share/man/man4/ng_bridge.4
@@ -34,7 +34,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd August 31, 2000
+.Dd May 5, 2010
 .Dt NG_BRIDGE 4
 .Os
 .Sh NAME
@@ -181,11 +181,17 @@ but also atomically clears the statistics as well.
 .It Dv NGM_BRIDGE_GET_TABLE
 Returns the current host mapping table used to direct packets, in a
 .Dv "struct ng_bridge_host_ary" .
+.It Dv NGM_BRIDGE_SET_PERSISTENT
+This command sets the persistent flag on the node, and takes no arguments.
 .El
 .Sh SHUTDOWN
 This node shuts down upon receipt of a
 .Dv NGM_SHUTDOWN
-control message, or when all hooks have been disconnected.
+control message, or when all hooks have been disconnected.  Setting the
+persistent flag via a
+.Dv NGM_BRIDGE_SET_PERSISTENT
+control message disables automatic node shutdown when the last hook gets
+disconnected.
 .Sh FILES
 .Bl -tag -width XXXXXXXX -compact
 .It Pa /usr/share/examples/netgraph/ether.bridge
diff --git a/share/man/man4/ng_hub.4 b/share/man/man4/ng_hub.4
index a1ad0680171..c52aba04847 100644
--- a/share/man/man4/ng_hub.4
+++ b/share/man/man4/ng_hub.4
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd April 17, 2004
+.Dd May 5, 2010
 .Dt NG_HUB 4
 .Os
 .Sh NAME
@@ -45,11 +45,20 @@ A
 node accepts any request to connect, regardless of the hook name,
 as long as the name is unique.
 .Sh CONTROL MESSAGES
-This node type supports only the generic control messages.
+This node type supports the generic control messages, plus the
+following:
+.Bl -tag -width foo
+.It Dv NGM_HUB_SET_PERSISTENT
+This command sets the persistent flag on the node, and takes no arguments.
+.El
 .Sh SHUTDOWN
 This node shuts down upon receipt of a
 .Dv NGM_SHUTDOWN
-control message, or when all hooks have been disconnected.
+control message, or when all hooks have been disconnected.  Setting the
+persistent flag via a
+.Dv NGM_HUB_SET_PERSISTENT
+control message disables automatic node shutdown when the last hook gets
+disconnected.
 .Sh SEE ALSO
 .Xr netgraph 4 ,
 .Xr ng_bridge 4 ,
diff --git a/sys/netgraph/ng_bridge.c b/sys/netgraph/ng_bridge.c
index 35e13927812..0bf08662353 100644
--- a/sys/netgraph/ng_bridge.c
+++ b/sys/netgraph/ng_bridge.c
@@ -84,7 +84,7 @@
 #include <netgraph/ng_bridge.h>
 
 #ifdef NG_SEPARATE_MALLOC
-MALLOC_DEFINE(M_NETGRAPH_BRIDGE, "netgraph_bridge", "netgraph bridge node ");
+MALLOC_DEFINE(M_NETGRAPH_BRIDGE, "netgraph_bridge", "netgraph bridge node");
 #else
 #define M_NETGRAPH_BRIDGE M_NETGRAPH
 #endif
@@ -106,6 +106,7 @@ struct ng_bridge_private {
 	u_int			numBuckets;	/* num buckets in table */
 	u_int			hashMask;	/* numBuckets - 1 */
 	int			numLinks;	/* num connected links */
+	int			persistent;	/* can exist w/o hooks */
 	struct callout		timer;		/* one second periodic timer */
 };
 typedef struct ng_bridge_private *priv_p;
@@ -271,6 +272,13 @@ static const struct ng_cmdlist ng_bridge_cmdlist[] = {
 	  NULL,
 	  &ng_bridge_host_ary_type
 	},
+	{
+	  NGM_BRIDGE_COOKIE,
+	  NGM_BRIDGE_SET_PERSISTENT,
+	  "setpersistent",
+	  NULL,
+	  NULL
+	},
 	{ 0 }
 };
 
@@ -495,6 +503,11 @@ ng_bridge_rcvmsg(node_p node, item_p item, hook_p lasthook)
 			}
 			break;
 		    }
+		case NGM_BRIDGE_SET_PERSISTENT:
+		    {
+			priv->persistent = 1;
+			break;
+		    }
 		default:
 			error = EINVAL;
 			break;
@@ -800,7 +813,8 @@ ng_bridge_disconnect(hook_p hook)
 
 	/* If no more hooks, go away */
 	if ((NG_NODE_NUMHOOKS(NG_HOOK_NODE(hook)) == 0)
-	&& (NG_NODE_IS_VALID(NG_HOOK_NODE(hook)))) {
+	    && (NG_NODE_IS_VALID(NG_HOOK_NODE(hook)))
+	    && !priv->persistent) {
 		ng_rmnode_self(NG_HOOK_NODE(hook));
 	}
 	return (0);
diff --git a/sys/netgraph/ng_bridge.h b/sys/netgraph/ng_bridge.h
index 0524a96cb36..c34d27d9966 100644
--- a/sys/netgraph/ng_bridge.h
+++ b/sys/netgraph/ng_bridge.h
@@ -149,6 +149,7 @@ enum {
 	NGM_BRIDGE_CLR_STATS,		/* clear link stats */
 	NGM_BRIDGE_GETCLR_STATS,	/* atomically get & clear link stats */
 	NGM_BRIDGE_GET_TABLE,		/* get link table */
+	NGM_BRIDGE_SET_PERSISTENT,	/* set persistent mode */
 };
 
 #endif /* _NETGRAPH_NG_BRIDGE_H_ */
diff --git a/sys/netgraph/ng_hub.c b/sys/netgraph/ng_hub.c
index ba298e56b13..077074575b9 100644
--- a/sys/netgraph/ng_hub.c
+++ b/sys/netgraph/ng_hub.c
@@ -36,16 +36,46 @@
 #include <netgraph/ng_hub.h>
 #include <netgraph/netgraph.h>
 
+#ifdef NG_SEPARATE_MALLOC
+MALLOC_DEFINE(M_NETGRAPH_HUB, "netgraph_hub", "netgraph hub node");
+#else
+#define M_NETGRAPH_HUB M_NETGRAPH
+#endif
+
+/* Per-node private data */
+struct ng_hub_private {
+	int		persistent;	/* can exist w/o hooks */
+};
+typedef struct ng_hub_private *priv_p;
+
+/* Netgraph node methods */
 static ng_constructor_t	ng_hub_constructor;
+static ng_rcvmsg_t	ng_hub_rcvmsg;
+static ng_shutdown_t	ng_hub_shutdown;
 static ng_rcvdata_t	ng_hub_rcvdata;
 static ng_disconnect_t	ng_hub_disconnect;
 
+/* List of commands and how to convert arguments to/from ASCII */
+static const struct ng_cmdlist ng_hub_cmdlist[] = {
+	{
+		NGM_HUB_COOKIE,
+		NGM_HUB_SET_PERSISTENT,
+		"setpersistent",
+		NULL,
+		NULL
+	},
+	{ 0 }
+};
+
 static struct ng_type ng_hub_typestruct = {
 	.version =	NG_ABI_VERSION,
 	.name =		NG_HUB_NODE_TYPE,
 	.constructor =	ng_hub_constructor,
+	.rcvmsg =	ng_hub_rcvmsg,
+	.shutdown =	ng_hub_shutdown,
 	.rcvdata =	ng_hub_rcvdata,
 	.disconnect =	ng_hub_disconnect,
+	.cmdlist =	ng_hub_cmdlist,
 };
 NETGRAPH_INIT(hub, &ng_hub_typestruct);
 
@@ -53,10 +83,39 @@ NETGRAPH_INIT(hub, &ng_hub_typestruct);
 static int
 ng_hub_constructor(node_p node)
 {
+	priv_p priv;
 
+	/* Allocate and initialize private info */
+	priv = malloc(sizeof(*priv), M_NETGRAPH_HUB, M_NOWAIT | M_ZERO);
+	if (priv == NULL)
+		return (ENOMEM);
+
+	NG_NODE_SET_PRIVATE(node, priv);
 	return (0);
 }
 
+/*
+ * Receive a control message
+ */
+static int
+ng_hub_rcvmsg(node_p node, item_p item, hook_p lasthook)
+{
+	const priv_p priv = NG_NODE_PRIVATE(node);
+	int error = 0;
+	struct ng_mesg *msg;
+
+	NGI_GET_MSG(item, msg);
+	if (msg->header.typecookie == NGM_HUB_COOKIE &&
+	    msg->header.cmd == NGM_HUB_SET_PERSISTENT) {
+		priv->persistent = 1;
+	} else {
+		error = EINVAL;
+	}
+
+	NG_FREE_MSG(msg);
+	return (error);
+}
+
 static int
 ng_hub_rcvdata(hook_p hook, item_p item)
 {
@@ -89,12 +148,25 @@ ng_hub_rcvdata(hook_p hook, item_p item)
 	return (error);
 }
 
+/*
+ * Shutdown node
+ */
+static int
+ng_hub_shutdown(node_p node)
+{
+	const priv_p priv = NG_NODE_PRIVATE(node);
+
+	free(priv, M_NETGRAPH_HUB);
+	return (0);
+}
+
 static int
 ng_hub_disconnect(hook_p hook)
 {
+	const priv_p priv = NG_NODE_PRIVATE(NG_HOOK_NODE(hook));
 
 	if (NG_NODE_NUMHOOKS(NG_HOOK_NODE(hook)) == 0 &&
-	    NG_NODE_IS_VALID(NG_HOOK_NODE(hook)))
+	    NG_NODE_IS_VALID(NG_HOOK_NODE(hook)) && !priv->persistent)
 		ng_rmnode_self(NG_HOOK_NODE(hook));
 	return (0);
 }
diff --git a/sys/netgraph/ng_hub.h b/sys/netgraph/ng_hub.h
index a735d7b64df..d6b16e076c3 100644
--- a/sys/netgraph/ng_hub.h
+++ b/sys/netgraph/ng_hub.h
@@ -33,4 +33,9 @@
 #define	NG_HUB_NODE_TYPE	"hub"
 #define	NGM_HUB_COOKIE		1082189597
 
+/* Netgraph control messages */
+enum {
+	NGM_HUB_SET_PERSISTENT = 1,	/* set persistent mode */
+};
+
 #endif /* _NETGRAPH_NG_HUB_H_ */

From 626b7c61f8545503a69c594bea5d9a7c9f442278 Mon Sep 17 00:00:00 2001
From: Marius Strobl <marius@FreeBSD.org>
Date: Wed, 5 May 2010 22:15:20 +0000
Subject: [PATCH 482/532] - Fix broken symlinks on cross platform zfs
 send/recv. [1] - Enable zfs_ace_byteswap() on FreeBSD as it works just fine
 (tested between   amd64 and sparc64 in both directions by Michael Moll).

PR:		146272
Approved by:	mm, pjd
Obtained from:	OpenSolaris (onnv rev. 8283:1ca59f393041; Bug ID 6764193) [1]
MFC after:	3 days
---
 .../uts/common/fs/zfs/zfs_byteswap.c          | 37 ++++++++++++++-----
 1 file changed, 28 insertions(+), 9 deletions(-)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_byteswap.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_byteswap.c
index b6c43f4245f..cd36696f950 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_byteswap.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_byteswap.c
@@ -19,12 +19,10 @@
  * CDDL HEADER END
  */
 /*
- * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
+ * Copyright 2008 Sun Microsystems, Inc.  All rights reserved.
  * Use is subject to license terms.
  */
 
-#pragma ident	"%Z%%M%	%I%	%E% SMI"
-
 #include <sys/zfs_context.h>
 #include <sys/vfs.h>
 #include <sys/fs/zfs.h>
@@ -50,7 +48,6 @@ zfs_oldace_byteswap(ace_t *ace, int ace_cnt)
 void
 zfs_ace_byteswap(void *buf, size_t size, boolean_t zfs_layout)
 {
-#ifdef TODO
 	caddr_t end;
 	caddr_t ptr;
 	zfs_ace_t *zacep;
@@ -64,6 +61,20 @@ zfs_ace_byteswap(void *buf, size_t size, boolean_t zfs_layout)
 
 	while (ptr < end) {
 		if (zfs_layout) {
+			/*
+			 * Avoid overrun.  Embedded aces can have one
+			 * of several sizes.  We don't know exactly
+			 * how many our present, only the size of the
+			 * buffer containing them.  That size may be
+			 * larger than needed to hold the aces
+			 * present.  As long as we do not do any
+			 * swapping beyond the end of our block we are
+			 * okay.  It it safe to swap any non-ace data
+			 * within the block since it is just zeros.
+			 */
+			if (ptr + sizeof (zfs_ace_hdr_t) > end) {
+				break;
+			}
 			zacep = (zfs_ace_t *)ptr;
 			zacep->z_hdr.z_access_mask =
 			    BSWAP_32(zacep->z_hdr.z_access_mask);
@@ -72,6 +83,10 @@ zfs_ace_byteswap(void *buf, size_t size, boolean_t zfs_layout)
 			    BSWAP_16(zacep->z_hdr.z_type);
 			entry_type = zacep->z_hdr.z_flags & ACE_TYPE_FLAGS;
 		} else {
+			/* Overrun avoidance */
+			if (ptr + sizeof (ace_t) > end) {
+				break;
+			}
 			acep = (ace_t *)ptr;
 			acep->a_access_mask = BSWAP_32(acep->a_access_mask);
 			acep->a_flags = BSWAP_16(acep->a_flags);
@@ -88,8 +103,14 @@ zfs_ace_byteswap(void *buf, size_t size, boolean_t zfs_layout)
 			break;
 		case ACE_IDENTIFIER_GROUP:
 		default:
+			/* Overrun avoidance */
 			if (zfs_layout) {
-				zacep->z_fuid = BSWAP_64(zacep->z_fuid);
+				if (ptr + sizeof (zfs_ace_t) <= end) {
+					zacep->z_fuid = BSWAP_64(zacep->z_fuid);
+				} else {
+					entry_size = sizeof (zfs_ace_t);
+					break;
+				}
 			}
 			switch (ace_type) {
 			case ACE_ACCESS_ALLOWED_OBJECT_ACE_TYPE:
@@ -108,9 +129,6 @@ zfs_ace_byteswap(void *buf, size_t size, boolean_t zfs_layout)
 		}
 		ptr = ptr + entry_size;
 	}
-#else	/* TODO */
-	panic("%s:%u: TODO", __func__, __LINE__);
-#endif	/* TODO */
 }
 
 /* ARGSUSED */
@@ -173,7 +191,8 @@ zfs_znode_byteswap(void *buf, size_t size)
 	if (zp->zp_acl.z_acl_version == ZFS_ACL_VERSION) {
 		zfs_acl_byteswap((void *)&zp->zp_acl.z_ace_data[0],
 		    ZFS_ACE_SPACE);
-	} else
+	} else {
 		zfs_oldace_byteswap((ace_t *)&zp->zp_acl.z_ace_data[0],
 		    ACE_SLOT_CNT);
+	}
 }

From b85998cb48133c19c5504393840f47668cd094f7 Mon Sep 17 00:00:00 2001
From: Navdeep Parhar <np@FreeBSD.org>
Date: Wed, 5 May 2010 22:29:54 +0000
Subject: [PATCH 483/532] Do not hold the T3 firmware in memory all the time.
 firmware(9) can load/unload it as needed.

---
 sys/dev/cxgb/cxgb_main.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/sys/dev/cxgb/cxgb_main.c b/sys/dev/cxgb/cxgb_main.c
index ad2caed218f..13fed449fa7 100644
--- a/sys/dev/cxgb/cxgb_main.c
+++ b/sys/dev/cxgb/cxgb_main.c
@@ -3403,5 +3403,3 @@ set_tcb_field_ulp(struct cpl_set_tcb_field *req, unsigned int tid,
 	txpkt->len = htonl(V_ULPTX_NFLITS(sizeof(*req) / 8));
 	mk_set_tcb_field(req, tid, word, mask, val);
 }
-
-MODULE_DEPEND(if_cxgb, cxgb_t3fw, 1, 1, 1);

From 1d4942f42f0444835252be4ca5ccf9fd3e91a0fb Mon Sep 17 00:00:00 2001
From: Navdeep Parhar <np@FreeBSD.org>
Date: Wed, 5 May 2010 22:52:06 +0000
Subject: [PATCH 484/532] Don't ring the tx doorbell for every frame when we
 know more frames will follow.  Adjust the freelist and response queue
 doorbells too.

Discussed with:	kmacy
---
 sys/dev/cxgb/cxgb_adapter.h |  2 ++
 sys/dev/cxgb/cxgb_sge.c     | 48 +++++++++++++++++++++++--------------
 2 files changed, 32 insertions(+), 18 deletions(-)

diff --git a/sys/dev/cxgb/cxgb_adapter.h b/sys/dev/cxgb/cxgb_adapter.h
index d1f5ef64c50..1430ca15e81 100644
--- a/sys/dev/cxgb/cxgb_adapter.h
+++ b/sys/dev/cxgb/cxgb_adapter.h
@@ -204,6 +204,7 @@ struct sge_fl {
 	uint32_t	cidx;
 	uint32_t	pidx;
 	uint32_t	gen;
+	uint32_t	db_pending;
 	bus_addr_t	phys_addr;
 	uint32_t	cntxt_id;
 	uint32_t	empty;
@@ -232,6 +233,7 @@ struct sge_txq {
 	uint32_t	pidx;
 	uint32_t	gen;
 	uint32_t	unacked;
+	uint32_t	db_pending;
 	struct tx_desc	*desc;
 	struct tx_sw_desc *sdesc;
 	uint32_t	token;
diff --git a/sys/dev/cxgb/cxgb_sge.c b/sys/dev/cxgb/cxgb_sge.c
index 9bc36c94fa7..5fdae6be7e7 100644
--- a/sys/dev/cxgb/cxgb_sge.c
+++ b/sys/dev/cxgb/cxgb_sge.c
@@ -696,7 +696,7 @@ refill_fl(adapter_t *sc, struct sge_fl *q, int n)
 	struct refill_fl_cb_arg cb_arg;
 	struct mbuf *m;
 	caddr_t cl;
-	int err, count = 0;
+	int err;
 	
 	cb_arg.error = 0;
 	while (n--) {
@@ -754,12 +754,14 @@ refill_fl(adapter_t *sc, struct sge_fl *q, int n)
 			d = q->desc;
 		}
 		q->credits++;
-		count++;
+		q->db_pending++;
 	}
 
 done:
-	if (count)
+	if (q->db_pending >= 32) {
+		q->db_pending = 0;
 		t3_write_reg(sc, A_SG_KDOORBELL, V_EGRCNTX(q->cntxt_id));
+	}
 }
 
 
@@ -810,8 +812,10 @@ __refill_fl(adapter_t *adap, struct sge_fl *fl)
 static __inline void
 __refill_fl_lt(adapter_t *adap, struct sge_fl *fl, int max)
 {
-	if ((fl->size - fl->credits) < max)
-		refill_fl(adap, fl, min(max, fl->size - fl->credits));
+	uint32_t reclaimable = fl->size - fl->credits;
+
+	if (reclaimable > 0)
+		refill_fl(adap, fl, min(max, reclaimable));
 }
 
 /**
@@ -1261,7 +1265,7 @@ make_sgl(struct sg_ent *sgp, bus_dma_segment_t *segs, int nsegs)
  *	When GTS is disabled we unconditionally ring the doorbell.
  */
 static __inline void
-check_ring_tx_db(adapter_t *adap, struct sge_txq *q)
+check_ring_tx_db(adapter_t *adap, struct sge_txq *q, int mustring)
 {
 #if USE_GTS
 	clear_bit(TXQ_LAST_PKT_DB, &q->flags);
@@ -1275,9 +1279,12 @@ check_ring_tx_db(adapter_t *adap, struct sge_txq *q)
 			     F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id));
 	}
 #else
-	wmb();            /* write descriptors before telling HW */
-	t3_write_reg(adap, A_SG_KDOORBELL,
-		     F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id));
+	if (mustring || ++q->db_pending >= 32) {
+		wmb();            /* write descriptors before telling HW */
+		t3_write_reg(adap, A_SG_KDOORBELL,
+		    F_SELEGRCNTX | V_EGRCNTX(q->cntxt_id));
+		q->db_pending = 0;
+	}
 #endif
 }
 
@@ -1480,7 +1487,7 @@ t3_encap(struct sge_qset *qs, struct mbuf **m)
 		wmb();
 		ETHER_BPF_MTAP(pi->ifp, m0);
 		wr_gen2(txd, txqs.gen);
-		check_ring_tx_db(sc, txq);
+		check_ring_tx_db(sc, txq, 0);
 		return (0);		
 	} else if (tso_info) {
 		int eth_type;
@@ -1543,7 +1550,7 @@ t3_encap(struct sge_qset *qs, struct mbuf **m)
 			wmb();
 			ETHER_BPF_MTAP(pi->ifp, m0);
 			wr_gen2(txd, txqs.gen);
-			check_ring_tx_db(sc, txq);
+			check_ring_tx_db(sc, txq, 0);
 			m_freem(m0);
 			return (0);
 		}
@@ -1574,7 +1581,7 @@ t3_encap(struct sge_qset *qs, struct mbuf **m)
 			wmb();
 			ETHER_BPF_MTAP(pi->ifp, m0);
 			wr_gen2(txd, txqs.gen);
-			check_ring_tx_db(sc, txq);
+			check_ring_tx_db(sc, txq, 0);
 			m_freem(m0);
 			return (0);
 		}
@@ -1593,7 +1600,7 @@ t3_encap(struct sge_qset *qs, struct mbuf **m)
 	wr_lo = htonl(V_WR_TID(txq->token));
 	write_wr_hdr_sgl(ndesc, txd, &txqs, txq, sgl, flits,
 	    sgl_flits, wr_hi, wr_lo);
-	check_ring_tx_db(sc, txq);
+	check_ring_tx_db(sc, txq, 0);
 
 	return (0);
 }
@@ -1643,7 +1650,6 @@ cxgb_start_locked(struct sge_qset *qs)
 {
 	struct mbuf *m_head = NULL;
 	struct sge_txq *txq = &qs->txq[TXQ_ETH];
-	int in_use_init = txq->in_use;
 	struct port_info *pi = qs->port;
 	struct ifnet *ifp = pi->ifp;
 
@@ -1655,8 +1661,7 @@ cxgb_start_locked(struct sge_qset *qs)
 		return;
 	}
 	TXQ_LOCK_ASSERT(qs);
-	while ((txq->in_use - in_use_init < TX_START_MAX_DESC) &&
-	    !TXQ_RING_EMPTY(qs) && (ifp->if_drv_flags & IFF_DRV_RUNNING) &&
+	while (!TXQ_RING_EMPTY(qs) && (ifp->if_drv_flags & IFF_DRV_RUNNING) &&
 	    pi->link_config.link_ok) {
 		reclaim_completed_tx(qs, cxgb_tx_reclaim_threshold, TXQ_ETH);
 
@@ -1674,6 +1679,10 @@ cxgb_start_locked(struct sge_qset *qs)
 
 		m_head = NULL;
 	}
+
+	if (txq->db_pending)
+		check_ring_tx_db(pi->adapter, txq, 1);
+
 	if (!TXQ_RING_EMPTY(qs) && callout_pending(&txq->txq_timer) == 0 &&
 	    pi->link_config.link_ok)
 		callout_reset_on(&txq->txq_timer, 1, cxgb_tx_timeout,
@@ -1707,6 +1716,9 @@ cxgb_transmit_locked(struct ifnet *ifp, struct sge_qset *qs, struct mbuf *m)
 			    (error = drbr_enqueue(ifp, br, m)) != 0) 
 				return (error);
 		} else {
+			if (txq->db_pending)
+				check_ring_tx_db(pi->adapter, txq, 1);
+
 			/*
 			 * We've bypassed the buf ring so we need to update
 			 * the stats directly
@@ -2354,7 +2366,7 @@ again:	reclaim_completed_tx(qs, 16, TXQ_OFLD);
 	TXQ_UNLOCK(qs);
 
 	write_ofld_wr(adap, m, q, pidx, gen, ndesc, segs, nsegs);
-	check_ring_tx_db(adap, q);
+	check_ring_tx_db(adap, q, 1);
 	return (0);
 }
 
@@ -3033,7 +3045,7 @@ process_responses(adapter_t *adap, struct sge_qset *qs, int budget)
 			r = rspq->desc;
 		}
 
-		if (++rspq->credits >= (rspq->size / 4)) {
+		if (++rspq->credits >= 64) {
 			refill_rspq(adap, rspq, rspq->credits);
 			rspq->credits = 0;
 		}

From 56cfc5ed12258e5ca5cf968eb030d9300317295d Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Thu, 6 May 2010 00:24:08 +0000
Subject: [PATCH 485/532] If the "-alldirs" export option was used for the V4:
 line, mountd would crash in check_options() since dp == NULL for the V4:
 line. This patch moves the check for options allowed on the V4: line to ahead
 of where dp is used to avoid this crash.

Reported by:	mamalos AT eng.auth.gr
MFC after:	1 week
---
 usr.sbin/mountd/mountd.c | 8 ++++----
 1 file changed, 4 insertions(+), 4 deletions(-)

diff --git a/usr.sbin/mountd/mountd.c b/usr.sbin/mountd/mountd.c
index 340afd3c08f..41a845f5b03 100644
--- a/usr.sbin/mountd/mountd.c
+++ b/usr.sbin/mountd/mountd.c
@@ -2881,16 +2881,16 @@ check_options(dp)
 		syslog(LOG_ERR, "-mask and /masklen are mutually exclusive");
 		return (1);
 	}
-	if ((opt_flags & OP_ALLDIRS) && dp->dp_left) {
-	    syslog(LOG_ERR, "-alldirs has multiple directories");
-	    return (1);
-	}
 	if (v4root_phase > 0 &&
 	    (opt_flags &
 	     ~(OP_SEC | OP_MASK | OP_NET | OP_HAVEMASK | OP_MASKLEN)) != 0) {
 	    syslog(LOG_ERR,"only -sec,-net,-mask options allowed on V4:");
 	    return (1);
 	}
+	if ((opt_flags & OP_ALLDIRS) && dp->dp_left) {
+	    syslog(LOG_ERR, "-alldirs has multiple directories");
+	    return (1);
+	}
 	return (0);
 }
 

From d238560c7c0bf4d10d4c23d4a61c8476b881ee78 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 04:23:52 +0000
Subject: [PATCH 486/532] Eliminate dead code.

---
 sys/mips/include/pmap.h |  1 -
 sys/mips/mips/pmap.c    | 14 --------------
 2 files changed, 15 deletions(-)

diff --git a/sys/mips/include/pmap.h b/sys/mips/include/pmap.h
index f4df6cae01c..0091d587917 100644
--- a/sys/mips/include/pmap.h
+++ b/sys/mips/include/pmap.h
@@ -175,7 +175,6 @@ void pmap_unmapdev(vm_offset_t, vm_size_t);
 vm_offset_t pmap_steal_memory(vm_size_t size);
 void pmap_set_modified(vm_offset_t pa);
 int page_is_managed(vm_offset_t pa);
-void pmap_page_is_free(vm_page_t m);
 void pmap_kenter(vm_offset_t va, vm_paddr_t pa);
 void pmap_kremove(vm_offset_t va);
 void *pmap_kenter_temporary(vm_paddr_t pa, int i);
diff --git a/sys/mips/mips/pmap.c b/sys/mips/mips/pmap.c
index af9ef432037..dbce42a6d03 100644
--- a/sys/mips/mips/pmap.c
+++ b/sys/mips/mips/pmap.c
@@ -3029,20 +3029,6 @@ init_pte_prot(vm_offset_t va, vm_page_t m, vm_prot_t prot)
 	return rw;
 }
 
-/*
- *	pmap_page_is_free:
- *
- *	Called when a page is freed to allow pmap to clean up
- *	any extra state associated with the page.  In this case
- *	clear modified/referenced bits.
- */
-void
-pmap_page_is_free(vm_page_t m)
-{
-
-	m->md.pv_flags = 0;
-}
-
 /*
  *	pmap_set_modified:
  *

From 8c6162468b8575f42092c138f9885feea2097ce0 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 6 May 2010 04:57:33 +0000
Subject: [PATCH 487/532] Add a helper function vm_pageout_page_lock(), similar
 to tegge' vm_pageout_fallback_object_lock(), to obtain the page lock while
 having page queue lock locked, and still maintain the page position in a
 queue.

Use the helper to lock the page in the pageout daemon and contig launder
iterators instead of skipping the page if its lock is contested.
Skipping locked pages easily causes pagedaemon or launder to not make a
progress with page cleaning.

Proposed and reviewed by:	alc
---
 sys/vm/vm_contig.c  |  6 ++--
 sys/vm/vm_pageout.c | 72 +++++++++++++++++++++++++++++++++++++--------
 sys/vm/vm_pageout.h |  1 +
 3 files changed, 65 insertions(+), 14 deletions(-)

diff --git a/sys/vm/vm_contig.c b/sys/vm/vm_contig.c
index c1c861807bc..1286c170d1c 100644
--- a/sys/vm/vm_contig.c
+++ b/sys/vm/vm_contig.c
@@ -168,8 +168,10 @@ vm_contig_launder(int queue)
 		if ((m->flags & PG_MARKER) != 0)
 			continue;
 
-		if (!vm_page_trylock(m))
-			continue;
+		if (!vm_pageout_page_lock(m, &next)) {
+			vm_page_unlock(m);
+			return (FALSE);
+		}
 		KASSERT(VM_PAGE_INQUEUE2(m, queue),
 		    ("vm_contig_launder: page %p's queue is not %d", m, queue));
 		error = vm_contig_launder_page(m, &next);
diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 9921a84b6bb..a84b6d40965 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -215,6 +215,17 @@ static void vm_req_vmdaemon(int req);
 #endif
 static void vm_pageout_page_stats(void);
 
+static void
+vm_pageout_init_marker(vm_page_t marker, u_short queue)
+{
+
+	bzero(marker, sizeof(*marker));
+	marker->flags = PG_FICTITIOUS | PG_MARKER;
+	marker->oflags = VPO_BUSY;
+	marker->queue = queue;
+	marker->wire_count = 1;
+}
+
 /*
  * vm_pageout_fallback_object_lock:
  * 
@@ -237,16 +248,8 @@ vm_pageout_fallback_object_lock(vm_page_t m, vm_page_t *next)
 	u_short queue;
 	vm_object_t object;
 
-	/*
-	 * Initialize our marker
-	 */
-	bzero(&marker, sizeof(marker));
-	marker.flags = PG_FICTITIOUS | PG_MARKER;
-	marker.oflags = VPO_BUSY;
-	marker.queue = m->queue;
-	marker.wire_count = 1;
-
 	queue = m->queue;
+	vm_pageout_init_marker(&marker, queue);
 	object = m->object;
 	
 	TAILQ_INSERT_AFTER(&vm_page_queues[queue].pl,
@@ -267,6 +270,43 @@ vm_pageout_fallback_object_lock(vm_page_t m, vm_page_t *next)
 	return (unchanged);
 }
 
+/*
+ * Lock the page while holding the page queue lock.  Use marker page
+ * to detect page queue changes and maintain notion of next page on
+ * page queue.  Return TRUE if no changes were detected, FALSE
+ * otherwise.  The page is locked on return. The page queue lock might
+ * be dropped and reacquired.
+ *
+ * This function depends on normal struct vm_page being type stable.
+ */
+boolean_t
+vm_pageout_page_lock(vm_page_t m, vm_page_t *next)
+{
+	struct vm_page marker;
+	boolean_t unchanged;
+	u_short queue;
+
+	vm_page_lock_assert(m, MA_NOTOWNED);
+	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+
+	if (vm_page_trylock(m))
+		return (TRUE);
+
+	queue = m->queue;
+	vm_pageout_init_marker(&marker, queue);
+
+	TAILQ_INSERT_AFTER(&vm_page_queues[queue].pl, m, &marker, pageq);
+	vm_page_unlock_queues();
+	vm_page_lock(m);
+	vm_page_lock_queues();
+
+	/* Page queue might have changed. */
+	*next = TAILQ_NEXT(&marker, pageq);
+	unchanged = (m->queue == queue && &marker == TAILQ_NEXT(m, pageq));
+	TAILQ_REMOVE(&vm_page_queues[queue].pl, &marker, pageq);
+	return (unchanged);
+}
+
 /*
  * vm_pageout_clean:
  *
@@ -777,7 +817,11 @@ rescan0:
 		if (m->flags & PG_MARKER)
 			continue;
 
-		if (!vm_page_trylock(m)) {
+		/*
+		 * Lock the page.
+		 */
+		if (!vm_pageout_page_lock(m, &next)) {
+			vm_page_unlock(m);
 			addl_page_shortage++;
 			continue;
 		}
@@ -1112,7 +1156,9 @@ unlock_and_continue:
 			m = next;
 			continue;
 		}
-		if (!vm_page_trylock(m) || (object = m->object) == NULL) {
+		if (!vm_pageout_page_lock(m, &next) ||
+		    (object = m->object) == NULL) {
+			vm_page_unlock(m);
 			m = next;
 			continue;
 		}
@@ -1375,7 +1421,9 @@ vm_pageout_page_stats()
 			continue;
 		}
 		vm_page_lock_assert(m, MA_NOTOWNED);
-		if (vm_page_trylock(m) == 0 || (object = m->object) == NULL) {
+		if (!vm_pageout_page_lock(m, &next) ||
+		    (object = m->object) == NULL) {
+			vm_page_unlock(m);
 			m = next;
 			continue;
 		}
diff --git a/sys/vm/vm_pageout.h b/sys/vm/vm_pageout.h
index 15ca5708a0c..4759dd7c0ea 100644
--- a/sys/vm/vm_pageout.h
+++ b/sys/vm/vm_pageout.h
@@ -104,5 +104,6 @@ extern void vm_waitpfault(void);
 boolean_t vm_pageout_fallback_object_lock(vm_page_t, vm_page_t *);
 int vm_pageout_flush(vm_page_t *, int, int);
 void vm_pageout_oom(int shortage);
+boolean_t vm_pageout_page_lock(vm_page_t, vm_page_t *);
 #endif
 #endif	/* _VM_VM_PAGEOUT_H_ */

From 14496931b0ed02621085207869337c5d52bef821 Mon Sep 17 00:00:00 2001
From: Alexander Motin <mav@FreeBSD.org>
Date: Thu, 6 May 2010 06:44:26 +0000
Subject: [PATCH 488/532] Fix polled operation. Now it is possible to dump
 kernel via mvs(4).

---
 sys/dev/mvs/mvs.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/mvs/mvs.c b/sys/dev/mvs/mvs.c
index 038e9f21072..2694d9cc380 100644
--- a/sys/dev/mvs/mvs.c
+++ b/sys/dev/mvs/mvs.c
@@ -2168,6 +2168,6 @@ mvspoll(struct cam_sim *sim)
 
 	arg.arg = ch->dev;
 	arg.cause = 2; /* XXX */
-	mvs_ch_intr(arg.arg);
+	mvs_ch_intr(&arg);
 }
 

From 52683078a270e5edf8fa3fe78a63d55c72272e87 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 15:52:08 +0000
Subject: [PATCH 489/532] Eliminate a small bit of unneeded code from
 kern_sendfile(): While kern_sendfile() is running, the file's vm object can't
 be destroyed because kern_sendfile() increments the vm object's reference
 count.  (Once kern_sendfile() decrements the reference count and returns, the
 vm object can, however, be destroyed.  So, sf_buf_mext() must handle the case
 where the vm object is destroyed.)

Reviewed by:	kib
---
 sys/kern/uipc_syscalls.c | 9 ++-------
 1 file changed, 2 insertions(+), 7 deletions(-)

diff --git a/sys/kern/uipc_syscalls.c b/sys/kern/uipc_syscalls.c
index a1378758836..dd504f85e9f 100644
--- a/sys/kern/uipc_syscalls.c
+++ b/sys/kern/uipc_syscalls.c
@@ -2139,14 +2139,9 @@ retry_space:
 			    (mnw ? SFB_NOWAIT : SFB_CATCH))) == NULL) {
 				mbstat.sf_allocfail++;
 				vm_page_lock(pg);
-				vm_page_lock_queues();
 				vm_page_unwire(pg, 0);
-				/*
-				 * XXX: Not same check as above!?
-				 */
-				if (pg->wire_count == 0 && pg->object == NULL)
-					vm_page_free(pg);
-				vm_page_unlock_queues();
+				KASSERT(pg->object != NULL,
+				    ("kern_sendfile: object disappeared"));
 				vm_page_unlock(pg);
 				error = (mnw ? EAGAIN : EINTR);
 				break;

From b277fb56f9452b620776541c9a45bd6db2d3a7c8 Mon Sep 17 00:00:00 2001
From: "Andrey A. Chernov" <ache@FreeBSD.org>
Date: Thu, 6 May 2010 16:37:50 +0000
Subject: [PATCH 490/532] Fix recent space skipping: 1) add missing (unsigned
 char) cast to ctype() macro 2) fix off-by-one error causing last letter
 always doubled

---
 usr.bin/calendar/io.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/usr.bin/calendar/io.c b/usr.bin/calendar/io.c
index 42abca62b9b..d12783f474c 100644
--- a/usr.bin/calendar/io.c
+++ b/usr.bin/calendar/io.c
@@ -155,8 +155,8 @@ cal(void)
 		}
 
 		/* Get rid of leading spaces (non-standard) */
-		while (isspace(buf[0]))
-			memcpy(buf, buf + 1, strlen(buf) - 1);
+		while (isspace((unsigned char)buf[0]))
+			memcpy(buf, buf + 1, strlen(buf));
 
 		/* No tab in the line, then not a valid line */
 		if ((pp = strchr(buf, '\t')) == NULL)

From 7024db1d405527d9e5c78f5b075521e189bf04df Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 16:39:43 +0000
Subject: [PATCH 491/532] Push down the page queues lock inside of
 vm_page_free_toq() and pmap_page_is_mapped() in preparation for removing page
 queues locking around calls to vm_page_free().  Setting aside the assertion
 that calls pmap_page_is_mapped(), vm_page_free_toq() now acquires and holds
 the page queues lock just long enough to actually add or remove the page from
 the paging queues.

Update vm_page_unhold() to reflect the above change.
---
 sys/amd64/amd64/pmap.c     | 13 ++++++-------
 sys/i386/i386/pmap.c       | 13 ++++++-------
 sys/sparc64/sparc64/pmap.c | 15 ++++++++++-----
 sys/vm/vm_page.c           | 23 +++++++++++++----------
 sys/vm/vm_page.h           |  2 +-
 5 files changed, 36 insertions(+), 30 deletions(-)

diff --git a/sys/amd64/amd64/pmap.c b/sys/amd64/amd64/pmap.c
index 045b634a980..3cf15d1d4fe 100644
--- a/sys/amd64/amd64/pmap.c
+++ b/sys/amd64/amd64/pmap.c
@@ -3961,16 +3961,15 @@ pmap_pvh_wired_mappings(struct md_page *pvh, int count)
 boolean_t
 pmap_page_is_mapped(vm_page_t m)
 {
-	struct md_page *pvh;
+	boolean_t rv;
 
 	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
 		return (FALSE);
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
-	if (TAILQ_EMPTY(&m->md.pv_list)) {
-		pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
-		return (!TAILQ_EMPTY(&pvh->pv_list));
-	} else
-		return (TRUE);
+	vm_page_lock_queues();
+	rv = !TAILQ_EMPTY(&m->md.pv_list) ||
+	    !TAILQ_EMPTY(&pa_to_pvh(VM_PAGE_TO_PHYS(m))->pv_list);
+	vm_page_unlock_queues();
+	return (rv);
 }
 
 /*
diff --git a/sys/i386/i386/pmap.c b/sys/i386/i386/pmap.c
index 661fade94e3..b342a675ef5 100644
--- a/sys/i386/i386/pmap.c
+++ b/sys/i386/i386/pmap.c
@@ -4125,16 +4125,15 @@ pmap_pvh_wired_mappings(struct md_page *pvh, int count)
 boolean_t
 pmap_page_is_mapped(vm_page_t m)
 {
-	struct md_page *pvh;
+	boolean_t rv;
 
 	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
 		return (FALSE);
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
-	if (TAILQ_EMPTY(&m->md.pv_list)) {
-		pvh = pa_to_pvh(VM_PAGE_TO_PHYS(m));
-		return (!TAILQ_EMPTY(&pvh->pv_list));
-	} else
-		return (TRUE);
+	vm_page_lock_queues();
+	rv = !TAILQ_EMPTY(&m->md.pv_list) ||
+	    !TAILQ_EMPTY(&pa_to_pvh(VM_PAGE_TO_PHYS(m))->pv_list);
+	vm_page_unlock_queues();
+	return (rv);
 }
 
 /*
diff --git a/sys/sparc64/sparc64/pmap.c b/sys/sparc64/sparc64/pmap.c
index daba4b7fb32..0a90b4032ab 100644
--- a/sys/sparc64/sparc64/pmap.c
+++ b/sys/sparc64/sparc64/pmap.c
@@ -1834,14 +1834,19 @@ boolean_t
 pmap_page_is_mapped(vm_page_t m)
 {
 	struct tte *tp;
+	boolean_t rv;
 
+	rv = FALSE;
 	if ((m->flags & (PG_FICTITIOUS | PG_UNMANAGED)) != 0)
-		return (FALSE);
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
+		return (rv);
+	vm_page_lock_queues();
 	TAILQ_FOREACH(tp, &m->md.tte_list, tte_link)
-		if ((tp->tte_data & TD_PV) != 0)
-			return (TRUE);
-	return (FALSE);
+		if ((tp->tte_data & TD_PV) != 0) {
+			rv = TRUE;
+			break;
+		}
+	vm_page_unlock_queues();
+	return (rv);
 }
 
 /*
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index cc34c1b857e..9cbd2efa258 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -563,11 +563,8 @@ vm_page_unhold(vm_page_t mem)
 	vm_page_lock_assert(mem, MA_OWNED);
 	--mem->hold_count;
 	KASSERT(mem->hold_count >= 0, ("vm_page_unhold: hold count < 0!!!"));
-	if (mem->hold_count == 0 && VM_PAGE_INQUEUE2(mem, PQ_HOLD)) {
-		vm_page_lock_queues();
+	if (mem->hold_count == 0 && VM_PAGE_INQUEUE2(mem, PQ_HOLD))
 		vm_page_free_toq(mem);
-		vm_page_unlock_queues();
-	}
 }
 
 /*
@@ -1448,10 +1445,11 @@ void
 vm_page_free_toq(vm_page_t m)
 {
 
-	if (VM_PAGE_GETQUEUE(m) != PQ_NONE)
-		mtx_assert(&vm_page_queue_mtx, MA_OWNED);
-	KASSERT(!pmap_page_is_mapped(m),
-	    ("vm_page_free_toq: freeing mapped page %p", m));
+	if ((m->flags & PG_UNMANAGED) == 0) {
+		vm_page_lock_assert(m, MA_OWNED);
+		KASSERT(!pmap_page_is_mapped(m),
+		    ("vm_page_free_toq: freeing mapped page %p", m));
+	}
 	PCPU_INC(cnt.v_tfree);
 
 	if (m->busy || VM_PAGE_IS_FREE(m)) {
@@ -1471,7 +1469,11 @@ vm_page_free_toq(vm_page_t m)
 	 * callback routine until after we've put the page on the
 	 * appropriate free queue.
 	 */
-	vm_pageq_remove(m);
+	if (VM_PAGE_GETQUEUE(m) != PQ_NONE) {
+		vm_page_lock_queues();
+		vm_pageq_remove(m);
+		vm_page_unlock_queues();
+	}
 	vm_page_remove(m);
 
 	/*
@@ -1493,9 +1495,10 @@ vm_page_free_toq(vm_page_t m)
 		panic("vm_page_free: freeing wired page");
 	}
 	if (m->hold_count != 0) {
-		vm_page_lock_assert(m, MA_OWNED);
 		m->flags &= ~PG_ZERO;
+		vm_page_lock_queues();
 		vm_page_enqueue(PQ_HOLD, m);
+		vm_page_unlock_queues();
 	} else {
 		/*
 		 * Restore the default memory attribute to the page.
diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index da4d42a8a76..3e7a318e3b4 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -107,7 +107,7 @@ struct vm_page {
 	vm_pindex_t pindex;		/* offset into object (O,Q) */
 	vm_paddr_t phys_addr;		/* physical address of page */
 	struct md_page md;		/* machine dependant stuff */
-	uint8_t	queue;			/* page queue index */
+	uint8_t	queue;			/* page queue index (P,Q) */
 	int8_t segind;
 	u_short	flags;			/* see below */
 	uint8_t	order;			/* index of the buddy queue */

From e86a8937dc51680b8e944e8f07f60684cf87daad Mon Sep 17 00:00:00 2001
From: "Andrey A. Chernov" <ache@FreeBSD.org>
Date: Thu, 6 May 2010 16:54:46 +0000
Subject: [PATCH 492/532] While I am here, add more missing (unsigned char)
 casts to ctype() macros

---
 usr.bin/calendar/io.c        | 2 +-
 usr.bin/calendar/parsedata.c | 2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)

diff --git a/usr.bin/calendar/io.c b/usr.bin/calendar/io.c
index d12783f474c..f1ba1ae7b76 100644
--- a/usr.bin/calendar/io.c
+++ b/usr.bin/calendar/io.c
@@ -163,7 +163,7 @@ cal(void)
 			continue;
 
 		/* Trim spaces in front of the tab */
-		while (isspace(pp[-1]))
+		while (isspace((unsigned char)pp[-1]))
 			pp--;
 
 		p = *pp;
diff --git a/usr.bin/calendar/parsedata.c b/usr.bin/calendar/parsedata.c
index 4fab6407f2b..6482fd3e081 100644
--- a/usr.bin/calendar/parsedata.c
+++ b/usr.bin/calendar/parsedata.c
@@ -872,7 +872,7 @@ isonlydigits(char *s, int nostar)
 	for (i = 0; s[i] != '\0'; i++) {
 		if (nostar == 0 && s[i] == '*' && s[i + 1] == '\0')
 			return 1;
-		if (!isdigit(s[i]))
+		if (!isdigit((unsigned char)s[i]))
 			return (0);
 	}
 	return (1);

From 3549ef2f2b5be31492b26030855c8903fc4f959f Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 6 May 2010 17:06:36 +0000
Subject: [PATCH 493/532] Revert r207677 which is considered a violation of
 style(9).

Pointed out by:	bde
Pointy hat to:	delphij
---
 usr.bin/find/find.c     |  9 +++++++--
 usr.bin/find/function.c |  8 ++++++--
 usr.bin/find/ls.c       |  8 ++++++--
 usr.bin/find/main.c     |  8 ++++++--
 usr.bin/find/misc.c     |  9 +++++++--
 usr.bin/find/operator.c |  8 ++++++--
 usr.bin/find/option.c   | 12 ++++++++----
 7 files changed, 46 insertions(+), 16 deletions(-)

diff --git a/usr.bin/find/find.c b/usr.bin/find/find.c
index 35ef5b25f46..cc2d797a541 100644
--- a/usr.bin/find/find.c
+++ b/usr.bin/find/find.c
@@ -32,10 +32,15 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)find.c	8.5 (Berkeley) 8/5/94
  */
 
+#ifndef lint
+#if 0
+static char sccsid[] = "@(#)find.c	8.5 (Berkeley) 8/5/94";
+#else
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/function.c b/usr.bin/find/function.c
index 4329887b09e..1714627d1fc 100644
--- a/usr.bin/find/function.c
+++ b/usr.bin/find/function.c
@@ -32,10 +32,14 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)function.c	8.10 (Berkeley) 5/4/95
  */
 
+#ifndef lint
+#if 0
+static const char sccsid[] = "@(#)function.c	8.10 (Berkeley) 5/4/95";
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/ls.c b/usr.bin/find/ls.c
index e96994b7184..88e45938cc6 100644
--- a/usr.bin/find/ls.c
+++ b/usr.bin/find/ls.c
@@ -29,10 +29,14 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)ls.c	8.1 (Berkeley) 6/6/93
  */
 
+#ifndef lint
+#if 0
+static char sccsid[] = "@(#)ls.c	8.1 (Berkeley) 6/6/93";
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/main.c b/usr.bin/find/main.c
index 7d4c24ac41a..8e2b42cd298 100644
--- a/usr.bin/find/main.c
+++ b/usr.bin/find/main.c
@@ -32,8 +32,6 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)main.c	8.4 (Berkeley) 5/4/95
  */
 
 #ifndef lint
@@ -42,6 +40,12 @@ char copyright[] =
 	The Regents of the University of California.  All rights reserved.\n";
 #endif /* not lint */
 
+#ifndef lint
+#if 0
+static char sccsid[] = "@(#)main.c	8.4 (Berkeley) 5/4/95";
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/misc.c b/usr.bin/find/misc.c
index 3696c13305f..15329064cf4 100644
--- a/usr.bin/find/misc.c
+++ b/usr.bin/find/misc.c
@@ -32,10 +32,15 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)misc.c	8.2 (Berkeley) 4/1/94
  */
 
+#ifndef lint
+#if 0
+static char sccsid[] = "@(#)misc.c	8.2 (Berkeley) 4/1/94";
+#else
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/operator.c b/usr.bin/find/operator.c
index d1a7ea8e807..c774efa0dc8 100644
--- a/usr.bin/find/operator.c
+++ b/usr.bin/find/operator.c
@@ -32,10 +32,14 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)operator.c	8.1 (Berkeley) 6/6/93
  */
 
+#ifndef lint
+#if 0
+static char sccsid[] = "@(#)operator.c	8.1 (Berkeley) 6/6/93";
+#endif
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
diff --git a/usr.bin/find/option.c b/usr.bin/find/option.c
index 6ae3958315b..02287c9c17a 100644
--- a/usr.bin/find/option.c
+++ b/usr.bin/find/option.c
@@ -32,10 +32,14 @@
  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  * SUCH DAMAGE.
- *
- *	@(#)option.c	8.2 (Berkeley) 4/16/94
  */
 
+#ifndef lint
+/*
+static char sccsid[] = "@(#)option.c	8.2 (Berkeley) 4/16/94";
+*/
+#endif /* not lint */
+
 #include <sys/cdefs.h>
 __FBSDID("$FreeBSD$");
 
@@ -51,7 +55,7 @@ __FBSDID("$FreeBSD$");
 
 #include "find.h"
 
-int typecompare(const void *, const void *);
+static int typecompare(const void *, const void *);
 
 /* NB: the following table must be sorted lexically. */
 /* Options listed with C++ comments are in gnu find, but not our find */
@@ -190,7 +194,7 @@ lookup_option(const char *name)
 	    sizeof(options)/sizeof(OPTION), sizeof(OPTION), typecompare));
 }
 
-int
+static int
 typecompare(const void *a, const void *b)
 {
 	return (strcmp(((const OPTION *)a)->name, ((const OPTION *)b)->name));

From fd8c28bfdfb5e931098ba6e3fce5269e9b9fee9b Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 17:28:59 +0000
Subject: [PATCH 494/532] Update a comment to say that access to a page's wire
 count is now synchronized by the page lock.

---
 sys/vm/vm_page.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index 3e7a318e3b4..ef790ae959e 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -113,7 +113,7 @@ struct vm_page {
 	uint8_t	order;			/* index of the buddy queue */
 	uint8_t pool;
 	u_short cow;			/* page cow mapping count (Q) */
-	u_int wire_count;		/* wired down maps refs (Q) */
+	u_int wire_count;		/* wired down maps refs (P) */
 	short hold_count;		/* page hold count (P) */
 	u_short oflags;			/* page flags (O) */
 	u_char	act_count;		/* page usage count (Q) */

From f0c0d3998df3028b8057d1d322c9b0c1ce227958 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 17:43:41 +0000
Subject: [PATCH 495/532] Remove page queues locking from all
 sf_buf_mext()-like functions.  The page lock now suffices.

Fix a couple nearby style violations.
---
 sys/kern/uipc_cow.c      | 6 +-----
 sys/kern/uipc_syscalls.c | 7 +------
 sys/net/bpf_zerocopy.c   | 2 --
 3 files changed, 2 insertions(+), 13 deletions(-)

diff --git a/sys/kern/uipc_cow.c b/sys/kern/uipc_cow.c
index bc9a56a3dcf..853b70b3bf0 100644
--- a/sys/kern/uipc_cow.c
+++ b/sys/kern/uipc_cow.c
@@ -81,7 +81,6 @@ socow_iodone(void *addr, void *args)
 	sf_buf_free(sf);
 	/* remove COW mapping  */
 	vm_page_lock(pp);
-	vm_page_lock_queues();
 	vm_page_cowclear(pp);
 	vm_page_unwire(pp, 0);
 	/*
@@ -91,7 +90,6 @@ socow_iodone(void *addr, void *args)
 	 */
 	if (pp->wire_count == 0 && pp->object == NULL)
 		vm_page_free(pp);
-	vm_page_unlock_queues();
 	vm_page_unlock(pp);
 	socow_stats.iodone++;
 }
@@ -147,9 +145,8 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 	 * Allocate an sf buf
 	 */
 	sf = sf_buf_alloc(pp, SFB_CATCH);
-	if (!sf) {
+	if (sf == NULL) {
 		vm_page_lock(pp);
-		vm_page_lock_queues();
 		vm_page_cowclear(pp);
 		vm_page_unwire(pp, 0);
 		/*
@@ -159,7 +156,6 @@ socow_setup(struct mbuf *m0, struct uio *uio)
 		 */
 		if (pp->wire_count == 0 && pp->object == NULL)
 			vm_page_free(pp);
-		vm_page_unlock_queues();
 		vm_page_unlock(pp);
 		socow_stats.fail_sf_buf++;
 		return(0);
diff --git a/sys/kern/uipc_syscalls.c b/sys/kern/uipc_syscalls.c
index dd504f85e9f..adcb8528743 100644
--- a/sys/kern/uipc_syscalls.c
+++ b/sys/kern/uipc_syscalls.c
@@ -1716,7 +1716,6 @@ sf_buf_mext(void *addr, void *args)
 	m = sf_buf_page(args);
 	sf_buf_free(args);
 	vm_page_lock(m);
-	vm_page_lock_queues();
 	vm_page_unwire(m, 0);
 	/*
 	 * Check for the object going away on us. This can
@@ -1725,7 +1724,6 @@ sf_buf_mext(void *addr, void *args)
 	 */
 	if (m->wire_count == 0 && m->object == NULL)
 		vm_page_free(m);
-	vm_page_unlock_queues();
 	vm_page_unlock(m);
 	if (addr == NULL)
 		return;
@@ -2111,7 +2109,6 @@ retry_space:
 			}
 			if (error) {
 				vm_page_lock(pg);
-				vm_page_lock_queues();
 				vm_page_unwire(pg, 0);
 				/*
 				 * See if anyone else might know about
@@ -2120,10 +2117,8 @@ retry_space:
 				 */
 				if (pg->wire_count == 0 && pg->valid == 0 &&
 				    pg->busy == 0 && !(pg->oflags & VPO_BUSY) &&
-				    pg->hold_count == 0) {
+				    pg->hold_count == 0)
 					vm_page_free(pg);
-				}
-				vm_page_unlock_queues();
 				vm_page_unlock(pg);
 				VM_OBJECT_UNLOCK(obj);
 				if (error == EAGAIN)
diff --git a/sys/net/bpf_zerocopy.c b/sys/net/bpf_zerocopy.c
index 5629093fc62..91efd649db4 100644
--- a/sys/net/bpf_zerocopy.c
+++ b/sys/net/bpf_zerocopy.c
@@ -113,11 +113,9 @@ zbuf_page_free(vm_page_t pp)
 {
 
 	vm_page_lock(pp);
-	vm_page_lock_queues();
 	vm_page_unwire(pp, 0);
 	if (pp->wire_count == 0 && pp->object == NULL)
 		vm_page_free(pp);
-	vm_page_unlock_queues();
 	vm_page_unlock(pp);
 }
 

From 2e3a546047f21bce97d518742da3b68e7b364232 Mon Sep 17 00:00:00 2001
From: Bernhard Schmidt <bschmidt@FreeBSD.org>
Date: Thu, 6 May 2010 17:53:04 +0000
Subject: [PATCH 496/532] Add a workaround for a bug in the firmware regarding
 the transition from passive to active scans. Basicly disable it by increasing
 the amount packets to be received to an amount which can't be reached during
 dwell times.

Approved by:	rpaulo (mentor)
MFC after:	3 days
---
 sys/dev/iwn/if_iwn.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/sys/dev/iwn/if_iwn.c b/sys/dev/iwn/if_iwn.c
index 5a17d1e3a7a..39b26b5d55d 100644
--- a/sys/dev/iwn/if_iwn.c
+++ b/sys/dev/iwn/if_iwn.c
@@ -4730,7 +4730,7 @@ iwn_scan(struct iwn_softc *sc)
 			chan->passive = htole16(78);
 		else
 			chan->passive = htole16(110);
-		hdr->crc_threshold = htole16(1);
+		hdr->crc_threshold = 0xffff;
 	} else if (!(c->ic_flags & IEEE80211_CHAN_PASSIVE)) {
 		chan->rf_gain = 0x28;
 		chan->active  = htole16(36);
@@ -4743,7 +4743,7 @@ iwn_scan(struct iwn_softc *sc)
 			chan->passive = htole16(88);
 		else
 			chan->passive = htole16(120);
-		hdr->crc_threshold = htole16(1);
+		hdr->crc_threshold = 0xffff;
 	}
 
 	DPRINTF(sc, IWN_DEBUG_STATE,

From 307d88b787f223c5b487ea6ce7779e73dc5201d4 Mon Sep 17 00:00:00 2001
From: Edward Tomasz Napierala <trasz@FreeBSD.org>
Date: Thu, 6 May 2010 18:43:19 +0000
Subject: [PATCH 497/532] Style fixes and removal of unneeded variable.

Submitted by:	bde@
---
 sys/fs/msdosfs/msdosfs_vnops.c | 3 +--
 sys/fs/tmpfs/tmpfs_vnops.c     | 3 +--
 sys/kern/vfs_vnops.c           | 6 +++---
 sys/sys/vnode.h                | 4 +++-
 4 files changed, 8 insertions(+), 8 deletions(-)

diff --git a/sys/fs/msdosfs/msdosfs_vnops.c b/sys/fs/msdosfs/msdosfs_vnops.c
index f38c72b841b..7a194124d79 100644
--- a/sys/fs/msdosfs/msdosfs_vnops.c
+++ b/sys/fs/msdosfs/msdosfs_vnops.c
@@ -652,7 +652,6 @@ msdosfs_write(ap)
 	struct buf *bp;
 	int ioflag = ap->a_ioflag;
 	struct uio *uio = ap->a_uio;
-	struct thread *td = uio->uio_td;
 	struct vnode *vp = ap->a_vp;
 	struct vnode *thisvp;
 	struct denode *dep = VTODE(vp);
@@ -696,7 +695,7 @@ msdosfs_write(ap)
 	/*
 	 * If they've exceeded their filesize limit, tell them about it.
 	 */
-	if (vn_rlimit_fsize(vp, uio, td))
+	if (vn_rlimit_fsize(vp, uio, uio->uio_td))
 		return (EFBIG);
 
 	/*
diff --git a/sys/fs/tmpfs/tmpfs_vnops.c b/sys/fs/tmpfs/tmpfs_vnops.c
index 3ffef6b6413..330eea57538 100644
--- a/sys/fs/tmpfs/tmpfs_vnops.c
+++ b/sys/fs/tmpfs/tmpfs_vnops.c
@@ -717,7 +717,6 @@ tmpfs_write(struct vop_write_args *v)
 	struct vnode *vp = v->a_vp;
 	struct uio *uio = v->a_uio;
 	int ioflag = v->a_ioflag;
-	struct thread *td = uio->uio_td;
 
 	boolean_t extended;
 	int error = 0;
@@ -747,7 +746,7 @@ tmpfs_write(struct vop_write_args *v)
 	  VFS_TO_TMPFS(vp->v_mount)->tm_maxfilesize)
 		return (EFBIG);
 
-	if (vn_rlimit_fsize(vp, uio, td))
+	if (vn_rlimit_fsize(vp, uio, uio->uio_td))
 		return (EFBIG);
 
 	extended = uio->uio_offset + uio->uio_resid > node->tn_size;
diff --git a/sys/kern/vfs_vnops.c b/sys/kern/vfs_vnops.c
index edc7ea83d29..28a6e21d2de 100644
--- a/sys/kern/vfs_vnops.c
+++ b/sys/kern/vfs_vnops.c
@@ -1341,11 +1341,12 @@ vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags, struct vnode **rvp)
 }
 
 int
-vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio, const struct thread *td)
+vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio,
+    const struct thread *td)
 {
+
 	if (vp->v_type != VREG || td == NULL)
 		return (0);
-
 	PROC_LOCK(td->td_proc);
 	if (uio->uio_offset + uio->uio_resid >
 	    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
@@ -1354,6 +1355,5 @@ vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio, const struct thre
 		return (EFBIG);
 	}
 	PROC_UNLOCK(td->td_proc);
-
 	return (0);
 }
diff --git a/sys/sys/vnode.h b/sys/sys/vnode.h
index 7c4ae32df44..64b3ce4e89e 100644
--- a/sys/sys/vnode.h
+++ b/sys/sys/vnode.h
@@ -655,6 +655,8 @@ int	vn_rdwr_inchunks(enum uio_rw rw, struct vnode *vp, void *base,
 	    size_t len, off_t offset, enum uio_seg segflg, int ioflg,
 	    struct ucred *active_cred, struct ucred *file_cred, size_t *aresid,
 	    struct thread *td);
+int	vn_rlimit_fsize(const struct vnode *vn, const struct uio *uio,
+	    const struct thread *td);
 int	vn_stat(struct vnode *vp, struct stat *sb, struct ucred *active_cred,
 	    struct ucred *file_cred, struct thread *td);
 int	vn_start_write(struct vnode *vp, struct mount **mpp, int flags);
@@ -670,6 +672,7 @@ int	vn_extattr_rm(struct vnode *vp, int ioflg, int attrnamespace,
 int	vn_vget_ino(struct vnode *vp, ino_t ino, int lkflags,
 	    struct vnode **rvp);
 
+
 int	vfs_cache_lookup(struct vop_lookup_args *ap);
 void	vfs_timestamp(struct timespec *);
 void	vfs_write_resume(struct mount *mp);
@@ -780,7 +783,6 @@ struct dirent;
 int vfs_read_dirent(struct vop_readdir_args *ap, struct dirent *dp, off_t off);
 
 int	vfs_unixify_accmode(accmode_t *accmode);
-int	vn_rlimit_fsize(const struct vnode *vn, const struct uio *uio, const struct thread *td);
 
 #endif /* _KERNEL */
 

From 77dda2b96f116c81d24ffd6dcddb3d331f0e154f Mon Sep 17 00:00:00 2001
From: Edward Tomasz Napierala <trasz@FreeBSD.org>
Date: Thu, 6 May 2010 18:52:41 +0000
Subject: [PATCH 498/532] Avoid overflow.

Submitted by:	bde@
---
 sys/kern/vfs_vnops.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/kern/vfs_vnops.c b/sys/kern/vfs_vnops.c
index 28a6e21d2de..64f77dd46d0 100644
--- a/sys/kern/vfs_vnops.c
+++ b/sys/kern/vfs_vnops.c
@@ -1348,7 +1348,7 @@ vn_rlimit_fsize(const struct vnode *vp, const struct uio *uio,
 	if (vp->v_type != VREG || td == NULL)
 		return (0);
 	PROC_LOCK(td->td_proc);
-	if (uio->uio_offset + uio->uio_resid >
+	if ((uoff_t)uio->uio_offset + uio->uio_resid >
 	    lim_cur(td->td_proc, RLIMIT_FSIZE)) {
 		psignal(td->td_proc, SIGXFSZ);
 		PROC_UNLOCK(td->td_proc);

From eb00b276ab2a5a549620ec3fe92c22fd0ddf948e Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Thu, 6 May 2010 18:58:32 +0000
Subject: [PATCH 499/532] Eliminate page queues locking around most calls to
 vm_page_free().

---
 sys/fs/nfsclient/nfs_clbio.c |  4 ----
 sys/fs/nwfs/nwfs_io.c        |  2 --
 sys/fs/smbfs/smbfs_io.c      |  4 ----
 sys/kern/kern_exec.c         |  4 +---
 sys/nfsclient/nfs_bio.c      |  4 ----
 sys/ufs/ffs/ffs_vnops.c      |  2 --
 sys/vm/device_pager.c        |  4 ----
 sys/vm/sg_pager.c            |  2 --
 sys/vm/vm_fault.c            | 12 +-----------
 sys/vm/vm_glue.c             |  4 ----
 sys/vm/vm_object.c           |  2 --
 sys/vm/vnode_pager.c         | 18 ------------------
 12 files changed, 2 insertions(+), 60 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index df027acd800..0224e157b6e 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -134,9 +134,7 @@ ncl_getpages(struct vop_getpages_args *ap)
 		for (i = 0; i < npages; ++i) {
 			if (i != ap->a_reqpage) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
@@ -177,9 +175,7 @@ ncl_getpages(struct vop_getpages_args *ap)
 		for (i = 0; i < npages; ++i) {
 			if (i != ap->a_reqpage) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
diff --git a/sys/fs/nwfs/nwfs_io.c b/sys/fs/nwfs/nwfs_io.c
index 0bb996fb3b3..125201408ad 100644
--- a/sys/fs/nwfs/nwfs_io.c
+++ b/sys/fs/nwfs/nwfs_io.c
@@ -431,9 +431,7 @@ nwfs_getpages(ap)
 		for (i = 0; i < npages; i++) {
 			if (ap->a_reqpage != i) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
diff --git a/sys/fs/smbfs/smbfs_io.c b/sys/fs/smbfs/smbfs_io.c
index c47fb1bcfe4..1ffcde9fb80 100644
--- a/sys/fs/smbfs/smbfs_io.c
+++ b/sys/fs/smbfs/smbfs_io.c
@@ -443,9 +443,7 @@ smbfs_getpages(ap)
 		for (i = 0; i < npages; ++i) {
 			if (i != reqpage) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
@@ -484,9 +482,7 @@ smbfs_getpages(ap)
 		for (i = 0; i < npages; i++) {
 			if (reqpage != i) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
diff --git a/sys/kern/kern_exec.c b/sys/kern/kern_exec.c
index 0ec9817140d..fc87d639404 100644
--- a/sys/kern/kern_exec.c
+++ b/sys/kern/kern_exec.c
@@ -948,11 +948,9 @@ exec_map_first_page(imgp)
 		rv = vm_pager_get_pages(object, ma, initial_pagein, 0);
 		ma[0] = vm_page_lookup(object, 0);
 		if ((rv != VM_PAGER_OK) || (ma[0] == NULL)) {
-			if (ma[0]) {
+			if (ma[0] != NULL) {
 				vm_page_lock(ma[0]);
-				vm_page_lock_queues();
 				vm_page_free(ma[0]);
-				vm_page_unlock_queues();
 				vm_page_unlock(ma[0]);
 			}
 			VM_OBJECT_UNLOCK(object);
diff --git a/sys/nfsclient/nfs_bio.c b/sys/nfsclient/nfs_bio.c
index c93d677eadd..37da34b8cce 100644
--- a/sys/nfsclient/nfs_bio.c
+++ b/sys/nfsclient/nfs_bio.c
@@ -132,9 +132,7 @@ nfs_getpages(struct vop_getpages_args *ap)
 		for (i = 0; i < npages; ++i) {
 			if (i != ap->a_reqpage) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
@@ -175,9 +173,7 @@ nfs_getpages(struct vop_getpages_args *ap)
 		for (i = 0; i < npages; ++i) {
 			if (i != ap->a_reqpage) {
 				vm_page_lock(pages[i]);
-				vm_page_lock_queues();
 				vm_page_free(pages[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(pages[i]);
 			}
 		}
diff --git a/sys/ufs/ffs/ffs_vnops.c b/sys/ufs/ffs/ffs_vnops.c
index a2fee4ecb41..96d0c7a3e8d 100644
--- a/sys/ufs/ffs/ffs_vnops.c
+++ b/sys/ufs/ffs/ffs_vnops.c
@@ -850,9 +850,7 @@ ffs_getpages(ap)
 		for (i = 0; i < pcount; i++) {
 			if (i != ap->a_reqpage) {
 				vm_page_lock(ap->a_m[i]);
-				vm_page_lock_queues();
 				vm_page_free(ap->a_m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(ap->a_m[i]);
 			}
 		}
diff --git a/sys/vm/device_pager.c b/sys/vm/device_pager.c
index 996f740c0ee..7d8d3611b37 100644
--- a/sys/vm/device_pager.c
+++ b/sys/vm/device_pager.c
@@ -255,9 +255,7 @@ dev_pager_getpages(object, m, count, reqpage)
 			for (i = 0; i < count; i++) {
 				if (i != reqpage) {
 					vm_page_lock(m[i]);
-					vm_page_lock_queues();		
 					vm_page_free(m[i]);
-					vm_page_unlock_queues();
 					vm_page_unlock(m[i]);
 				}
 			}
@@ -272,9 +270,7 @@ dev_pager_getpages(object, m, count, reqpage)
 		TAILQ_INSERT_TAIL(&object->un_pager.devp.devp_pglist, page, pageq);
 		for (i = 0; i < count; i++) {
 			vm_page_lock(m[i]);
-			vm_page_lock_queues();
 			vm_page_free(m[i]);
-			vm_page_unlock_queues();
 			vm_page_unlock(m[i]);
 		}
 		vm_page_insert(page, object, offset);
diff --git a/sys/vm/sg_pager.c b/sys/vm/sg_pager.c
index 845401a4bbd..a7b373647a5 100644
--- a/sys/vm/sg_pager.c
+++ b/sys/vm/sg_pager.c
@@ -200,9 +200,7 @@ sg_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 	/* Free the original pages and insert this fake page into the object. */
 	for (i = 0; i < count; i++) {
 		vm_page_lock(m[i]);
-		vm_page_lock_queues();
 		vm_page_free(m[i]);
-		vm_page_unlock_queues();
 		vm_page_unlock(m[i]);
 	}
 	vm_page_insert(page, object, offset);
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 2978f7f3a82..ae1ee290707 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -164,9 +164,7 @@ unlock_and_deallocate(struct faultstate *fs)
 	if (fs->object != fs->first_object) {
 		VM_OBJECT_LOCK(fs->first_object);
 		vm_page_lock(fs->first_m);
-		vm_page_lock_queues();
 		vm_page_free(fs->first_m);
-		vm_page_unlock_queues();
 		vm_page_unlock(fs->first_m);
 		vm_object_pip_wakeup(fs->first_object);
 		VM_OBJECT_UNLOCK(fs->first_object);
@@ -348,9 +346,7 @@ RetryFault:;
 				if (fs.object != fs.first_object) {
 					VM_OBJECT_LOCK(fs.first_object);
 					vm_page_lock(fs.first_m);
-					vm_page_lock_queues();
 					vm_page_free(fs.first_m);
-					vm_page_unlock_queues();
 					vm_page_unlock(fs.first_m);
 					vm_object_pip_wakeup(fs.first_object);
 					VM_OBJECT_UNLOCK(fs.first_object);
@@ -638,9 +634,7 @@ vnode_locked:
 			if (((fs.map != kernel_map) && (rv == VM_PAGER_ERROR)) ||
 				(rv == VM_PAGER_BAD)) {
 				vm_page_lock(fs.m);
-				vm_page_lock_queues();
 				vm_page_free(fs.m);
-				vm_page_unlock_queues();
 				vm_page_unlock(fs.m);
 				fs.m = NULL;
 				unlock_and_deallocate(&fs);
@@ -648,9 +642,7 @@ vnode_locked:
 			}
 			if (fs.object != fs.first_object) {
 				vm_page_lock(fs.m);
-				vm_page_lock_queues();
 				vm_page_free(fs.m);
-				vm_page_unlock_queues();
 				vm_page_unlock(fs.m);
 				fs.m = NULL;
 				/*
@@ -764,13 +756,11 @@ vnode_locked:
 				 * We don't chase down the shadow chain
 				 */
 			    fs.object == fs.first_object->backing_object) {
-				vm_page_lock(fs.first_m);
-				vm_page_lock_queues();
 				/*
 				 * get rid of the unnecessary page
 				 */
+				vm_page_lock(fs.first_m);
 				vm_page_free(fs.first_m);
-				vm_page_unlock_queues();
 				vm_page_unlock(fs.first_m);
 				/*
 				 * grab the page and put it into the 
diff --git a/sys/vm/vm_glue.c b/sys/vm/vm_glue.c
index f86212f0291..0b9d92a1df1 100644
--- a/sys/vm/vm_glue.c
+++ b/sys/vm/vm_glue.c
@@ -258,9 +258,7 @@ vm_imgact_hold_page(vm_object_t object, vm_ooffset_t offset)
 			goto out;
 		if (rv != VM_PAGER_OK) {
 			vm_page_lock(m);
-			vm_page_lock_queues();
 			vm_page_free(m);
-			vm_page_unlock_queues();
 			vm_page_unlock(m);
 			m = NULL;
 			goto out;
@@ -437,10 +435,8 @@ vm_thread_stack_dispose(vm_object_t ksobj, vm_offset_t ks, int pages)
 		if (m == NULL)
 			panic("vm_thread_dispose: kstack already missing?");
 		vm_page_lock(m);
-		vm_page_lock_queues();
 		vm_page_unwire(m, 0);
 		vm_page_free(m);
-		vm_page_unlock_queues();
 		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(ksobj);
diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index f8c50bc73a0..3986d142d87 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -2046,9 +2046,7 @@ vm_object_populate(vm_object_t object, vm_pindex_t start, vm_pindex_t end)
 				break;
 			if (rv != VM_PAGER_OK) {
 				vm_page_lock(m);
-				vm_page_lock_queues();
 				vm_page_free(m);
-				vm_page_unlock_queues();
 				vm_page_unlock(m);
 				break;
 			}
diff --git a/sys/vm/vnode_pager.c b/sys/vm/vnode_pager.c
index eb21c60a7c3..69fd5d101c7 100644
--- a/sys/vm/vnode_pager.c
+++ b/sys/vm/vnode_pager.c
@@ -725,9 +725,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		for (i = 0; i < count; i++)
 			if (i != reqpage) {
 				vm_page_lock(m[i]);
-				vm_page_lock_queues();
 				vm_page_free(m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[i]);
 			}
 		PCPU_INC(cnt.v_vnodein);
@@ -740,9 +738,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		for (i = 0; i < count; i++)
 			if (i != reqpage) {
 				vm_page_lock(m[i]);
-				vm_page_lock_queues();
 				vm_page_free(m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[i]);
 			}
 		VM_OBJECT_UNLOCK(object);
@@ -759,9 +755,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		for (i = 0; i < count; i++)
 			if (i != reqpage) {
 				vm_page_lock(m[i]);
-				vm_page_lock_queues();
 				vm_page_free(m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[i]);
 			}
 		VM_OBJECT_UNLOCK(object);
@@ -780,9 +774,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		for (i = 0; i < count; i++)
 			if (i != reqpage) {
 				vm_page_lock(m[i]);
-				vm_page_lock_queues();
 				vm_page_free(m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[i]);
 			}
 		VM_OBJECT_UNLOCK(object);
@@ -795,9 +787,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		for (i = 0; i < count; i++)
 			if (i != reqpage) {
 				vm_page_lock(m[i]);
-				vm_page_lock_queues();
 				vm_page_free(m[i]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[i]);
 			}
 		VM_OBJECT_UNLOCK(object);
@@ -821,9 +811,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 			for (; i < count; i++)
 				if (i != reqpage) {
 					vm_page_lock(m[i]);
-					vm_page_lock_queues();
 					vm_page_free(m[i]);
-					vm_page_unlock_queues();
 					vm_page_unlock(m[i]);
 				}
 			VM_OBJECT_UNLOCK(object);
@@ -840,9 +828,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 				    (uintmax_t)object->un_pager.vnp.vnp_size);
 			}
 			vm_page_lock(m[i]);
-			vm_page_lock_queues();
 			vm_page_free(m[i]);
-			vm_page_unlock_queues();
 			vm_page_unlock(m[i]);
 			VM_OBJECT_UNLOCK(object);
 			runend = i + 1;
@@ -854,9 +840,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 			VM_OBJECT_LOCK(object);
 			for (j = i; j < runend; j++) {
 				vm_page_lock(m[j]);
-				vm_page_lock_queues();
 				vm_page_free(m[j]);
-				vm_page_unlock_queues();
 				vm_page_unlock(m[j]);
 			}
 			VM_OBJECT_UNLOCK(object);
@@ -865,9 +849,7 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 				VM_OBJECT_LOCK(object);
 				for (i = first + runpg; i < count; i++) {
 					vm_page_lock(m[i]);
-					vm_page_lock_queues();
 					vm_page_free(m[i]);
-					vm_page_unlock_queues();
 					vm_page_unlock(m[i]);
 				}
 				VM_OBJECT_UNLOCK(object);

From d2ba618a63b7e5efdd72a038e65c6d6d601799d1 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 6 May 2010 19:22:50 +0000
Subject: [PATCH 500/532] Add MAKEDEV_NOWAIT flag to make_dev_credf(9), to
 create a device node in a no-sleep context. If resource allocation cannot be
 done without sleep, make_dev_credf() fails and returns NULL.

Reviewed by:	jh
MFC after:	2 weeks
---
 sys/fs/devfs/devfs_devs.c |  9 ++++++--
 sys/fs/devfs/devfs_int.h  |  2 +-
 sys/kern/kern_conf.c      | 47 +++++++++++++++++++++++----------------
 sys/sys/conf.h            |  6 +++--
 4 files changed, 40 insertions(+), 24 deletions(-)

diff --git a/sys/fs/devfs/devfs_devs.c b/sys/fs/devfs/devfs_devs.c
index 81b2752583e..4bd672867cc 100644
--- a/sys/fs/devfs/devfs_devs.c
+++ b/sys/fs/devfs/devfs_devs.c
@@ -115,17 +115,21 @@ SYSCTL_INT(_debug_sizeof, OID_AUTO, cdev_priv, CTLFLAG_RD,
     0, sizeof(struct cdev_priv), "sizeof(struct cdev_priv)");
 
 struct cdev *
-devfs_alloc(void)
+devfs_alloc(int flags)
 {
 	struct cdev_priv *cdp;
 	struct cdev *cdev;
 	struct timespec ts;
 
-	cdp = malloc(sizeof *cdp, M_CDEVP, M_USE_RESERVE | M_ZERO | M_WAITOK);
+	cdp = malloc(sizeof *cdp, M_CDEVP, M_USE_RESERVE | M_ZERO |
+	    ((flags & MAKEDEV_NOWAIT) ? M_NOWAIT : M_WAITOK));
+	if (cdp == NULL)
+		return (NULL);
 
 	cdp->cdp_dirents = &cdp->cdp_dirent0;
 	cdp->cdp_dirent0 = NULL;
 	cdp->cdp_maxdirent = 0;
+	cdp->cdp_inode = 0;
 
 	cdev = &cdp->cdp_c;
 
@@ -133,6 +137,7 @@ devfs_alloc(void)
 	LIST_INIT(&cdev->si_children);
 	vfs_timestamp(&ts);
 	cdev->si_atime = cdev->si_mtime = cdev->si_ctime = ts;
+	cdev->si_cred = NULL;
 
 	return (cdev);
 }
diff --git a/sys/fs/devfs/devfs_int.h b/sys/fs/devfs/devfs_int.h
index 5a61dd4a79b..f5612e1580f 100644
--- a/sys/fs/devfs/devfs_int.h
+++ b/sys/fs/devfs/devfs_int.h
@@ -70,7 +70,7 @@ struct cdev_priv {
 
 #define	cdev2priv(c)	member2struct(cdev_priv, cdp_c, c)
 
-struct cdev *devfs_alloc(void);
+struct cdev *devfs_alloc(int);
 void devfs_free(struct cdev *);
 void devfs_create(struct cdev *dev);
 void devfs_destroy(struct cdev *dev);
diff --git a/sys/kern/kern_conf.c b/sys/kern/kern_conf.c
index a2e292e40f0..17a6827254e 100644
--- a/sys/kern/kern_conf.c
+++ b/sys/kern/kern_conf.c
@@ -505,7 +505,7 @@ giant_mmap_single(struct cdev *dev, vm_ooffset_t *offset, vm_size_t size,
 }
 
 static void
-notify(struct cdev *dev, const char *ev)
+notify(struct cdev *dev, const char *ev, int flags)
 {
 	static const char prefix[] = "cdev=";
 	char *data;
@@ -514,7 +514,8 @@ notify(struct cdev *dev, const char *ev)
 	if (cold)
 		return;
 	namelen = strlen(dev->si_name);
-	data = malloc(namelen + sizeof(prefix), M_TEMP, M_NOWAIT);
+	data = malloc(namelen + sizeof(prefix), M_TEMP,
+	     (flags & MAKEDEV_NOWAIT) ? M_NOWAIT : M_WAITOK);
 	if (data == NULL)
 		return;
 	memcpy(data, prefix, sizeof(prefix) - 1);
@@ -524,17 +525,17 @@ notify(struct cdev *dev, const char *ev)
 }
 
 static void
-notify_create(struct cdev *dev)
+notify_create(struct cdev *dev, int flags)
 {
 
-	notify(dev, "CREATE");
+	notify(dev, "CREATE", flags);
 }
 
 static void
 notify_destroy(struct cdev *dev)
 {
 
-	notify(dev, "DESTROY");
+	notify(dev, "DESTROY", MAKEDEV_WAITOK);
 }
 
 static struct cdev *
@@ -572,24 +573,27 @@ fini_cdevsw(struct cdevsw *devsw)
 	devsw->d_flags &= ~D_INIT;
 }
 
-static void
-prep_cdevsw(struct cdevsw *devsw)
+static int
+prep_cdevsw(struct cdevsw *devsw, int flags)
 {
 	struct cdevsw *dsw2;
 
 	mtx_assert(&devmtx, MA_OWNED);
 	if (devsw->d_flags & D_INIT)
-		return;
+		return (1);
 	if (devsw->d_flags & D_NEEDGIANT) {
 		dev_unlock();
-		dsw2 = malloc(sizeof *dsw2, M_DEVT, M_WAITOK);
+		dsw2 = malloc(sizeof *dsw2, M_DEVT,
+		     (flags & MAKEDEV_NOWAIT) ? M_NOWAIT : M_WAITOK);
 		dev_lock();
+		if (dsw2 == NULL && !(devsw->d_flags & D_INIT))
+			return (0);
 	} else
 		dsw2 = NULL;
 	if (devsw->d_flags & D_INIT) {
 		if (dsw2 != NULL)
 			cdevsw_free_devlocked(dsw2);
-		return;
+		return (1);
 	}
 
 	if (devsw->d_version != D_VERSION_03) {
@@ -647,6 +651,7 @@ prep_cdevsw(struct cdevsw *devsw)
 
 	if (dsw2 != NULL)
 		cdevsw_free_devlocked(dsw2);
+	return (1);
 }
 
 static struct cdev *
@@ -657,9 +662,15 @@ make_dev_credv(int flags, struct cdevsw *devsw, int unit,
 	struct cdev *dev;
 	int i;
 
-	dev = devfs_alloc();
+	dev = devfs_alloc(flags);
+	if (dev == NULL)
+		return (NULL);
 	dev_lock();
-	prep_cdevsw(devsw);
+	if (!prep_cdevsw(devsw, flags)) {
+		dev_unlock();
+		devfs_free(dev);
+		return (NULL);
+	}
 	dev = newdev(devsw, unit, dev);
 	if (flags & MAKEDEV_REF)
 		dev_refl(dev);
@@ -686,8 +697,6 @@ make_dev_credv(int flags, struct cdevsw *devsw, int unit,
 	dev->si_flags |= SI_NAMED;
 	if (cr != NULL)
 		dev->si_cred = crhold(cr);
-	else
-		dev->si_cred = NULL;
 	dev->si_uid = uid;
 	dev->si_gid = gid;
 	dev->si_mode = mode;
@@ -696,7 +705,7 @@ make_dev_credv(int flags, struct cdevsw *devsw, int unit,
 	clean_unrhdrl(devfs_inos);
 	dev_unlock_and_free();
 
-	notify_create(dev);
+	notify_create(dev, flags);
 
 	return (dev);
 }
@@ -771,7 +780,7 @@ make_dev_alias(struct cdev *pdev, const char *fmt, ...)
 	int i;
 
 	KASSERT(pdev != NULL, ("NULL pdev"));
-	dev = devfs_alloc();
+	dev = devfs_alloc(MAKEDEV_WAITOK);
 	dev_lock();
 	dev->si_flags |= SI_ALIAS;
 	dev->si_flags |= SI_NAMED;
@@ -788,7 +797,7 @@ make_dev_alias(struct cdev *pdev, const char *fmt, ...)
 	clean_unrhdrl(devfs_inos);
 	dev_unlock();
 
-	notify_create(dev);
+	notify_create(dev, MAKEDEV_WAITOK);
 
 	return (dev);
 }
@@ -973,9 +982,9 @@ clone_create(struct clonedevs **cdp, struct cdevsw *csw, int *up,
 	 *       the end of the list.
 	 */
 	unit = *up;
-	ndev = devfs_alloc();
+	ndev = devfs_alloc(MAKEDEV_WAITOK);
 	dev_lock();
-	prep_cdevsw(csw);
+	prep_cdevsw(csw, MAKEDEV_WAITOK);
 	low = extra;
 	de = dl = NULL;
 	cd = *cdp;
diff --git a/sys/sys/conf.h b/sys/sys/conf.h
index a18d6996a6c..937414d7b65 100644
--- a/sys/sys/conf.h
+++ b/sys/sys/conf.h
@@ -262,8 +262,10 @@ struct cdev *make_dev(struct cdevsw *_devsw, int _unit, uid_t _uid, gid_t _gid,
 struct cdev *make_dev_cred(struct cdevsw *_devsw, int _unit,
 		struct ucred *_cr, uid_t _uid, gid_t _gid, int _perms,
 		const char *_fmt, ...) __printflike(7, 8);
-#define MAKEDEV_REF     0x1
-#define MAKEDEV_WHTOUT	0x2
+#define	MAKEDEV_REF     0x1
+#define	MAKEDEV_WHTOUT	0x2
+#define	MAKEDEV_NOWAIT	0x4
+#define	MAKEDEV_WAITOK	0x8
 struct cdev *make_dev_credf(int _flags,
 		struct cdevsw *_devsw, int _unit,
 		struct ucred *_cr, uid_t _uid, gid_t _gid, int _mode,

From 2b848cb1f5b4045de8946e59374e0b9d34f1a558 Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Thu, 6 May 2010 19:23:49 +0000
Subject: [PATCH 501/532] Document MAKEDEV_NOWAIT flag for  make_dev_credf(9).

Reviewed by:	jh
MFC after:	2 weeks
---
 share/man/man9/make_dev.9 | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/share/man/man9/make_dev.9 b/share/man/man9/make_dev.9
index 50f7d46fc6c..00af8a89dc8 100644
--- a/share/man/man9/make_dev.9
+++ b/share/man/man9/make_dev.9
@@ -24,7 +24,7 @@
 .\"
 .\" $FreeBSD$
 .\"
-.Dd September 28, 2008
+.Dd May 6, 2010
 .Dt MAKE_DEV 9
 .Os
 .Sh NAME
@@ -133,9 +133,18 @@ The following values are currently accepted:
 .Pp
 .Bd -literal -offset indent -compact
 MAKEDEV_REF	reference the created device
+MAKEDEV_NOWAIT	do not sleep, may return NULL
+MAKEDEV_WAITOK	allow the function to sleep to satisfy malloc
 .Ed
 .Pp
 The
+.Dv MAKEDEV_WAITOK
+flag is assumed if none of
+.Dv MAKEDEV_WAITOK ,
+.Dv MAKEDEV_NOWAIT
+is specified.
+.Pp
+The
 .Xr dev_clone 9
 event handler shall specify
 .Dv MAKEDEV_REF

From 176755d33edcc9c92c2b58192772d2220e00c9b6 Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Thu, 6 May 2010 20:54:19 +0000
Subject: [PATCH 502/532] Exclude undefined symbol from ELF file when doing
 function resolve.

MFC after: 3 days
---
 usr.sbin/pmcstat/pmcstat_log.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/usr.sbin/pmcstat/pmcstat_log.c b/usr.sbin/pmcstat/pmcstat_log.c
index 6d8c57a760a..3458bd31699 100644
--- a/usr.sbin/pmcstat/pmcstat_log.c
+++ b/usr.sbin/pmcstat/pmcstat_log.c
@@ -539,6 +539,8 @@ pmcstat_image_add_symbols(struct pmcstat_image *image, Elf *e,
 			return;
 		if (GELF_ST_TYPE(sym.st_info) != STT_FUNC)
 			continue;
+		if (sym.st_shndx == STN_UNDEF)
+			continue;
 
 		if (!firsttime && pmcstat_symbol_search(image, sym.st_value))
 			continue; /* We've seen this symbol already. */

From f9e4dd7122c6e140e2660bcfbb62c99b550f5847 Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Thu, 6 May 2010 20:58:23 +0000
Subject: [PATCH 503/532] Fix an invalid parameter detected by INVARIANT and
 confirmed by r193272.

---
 sys/netgraph/ng_ksocket.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/netgraph/ng_ksocket.c b/sys/netgraph/ng_ksocket.c
index 260eba90d7e..f6cbccca97c 100644
--- a/sys/netgraph/ng_ksocket.c
+++ b/sys/netgraph/ng_ksocket.c
@@ -1272,7 +1272,7 @@ ng_ksocket_finish_accept(priv_p priv)
 	soupcall_set(so, SO_RCV, ng_ksocket_incoming, node);
 	SOCKBUF_UNLOCK(&so->so_rcv);
 	SOCKBUF_LOCK(&so->so_snd);
-	soupcall_set(so, SO_RCV, ng_ksocket_incoming, node);
+	soupcall_set(so, SO_SND, ng_ksocket_incoming, node);
 	SOCKBUF_UNLOCK(&so->so_snd);
 
 	/* Fill in the response data and send it or return it to the caller */

From 3f65dafda3d7ad24a8c9181558bc597285313a05 Mon Sep 17 00:00:00 2001
From: Xin LI <delphij@FreeBSD.org>
Date: Thu, 6 May 2010 21:57:38 +0000
Subject: [PATCH 504/532] Plug a memory leak.

MFC after:	2 weeks
---
 usr.bin/rpcgen/rpc_cout.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/usr.bin/rpcgen/rpc_cout.c b/usr.bin/rpcgen/rpc_cout.c
index 8bbd5d3c578..4c36ce6c583 100644
--- a/usr.bin/rpcgen/rpc_cout.c
+++ b/usr.bin/rpcgen/rpc_cout.c
@@ -489,6 +489,7 @@ inline_struct(definition *def, int flag)
 			}
 			size = 0;
 			i = 0;
+			free(sizestr);
 			sizestr = NULL;
 			print_stat(indent + 1, &dl->decl);
 		}

From 29b7bdea791d0a08f8c16eed487dc9a808a4cffd Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Thu, 6 May 2010 22:06:14 +0000
Subject: [PATCH 505/532] sigaltstack(2): document some modernizations: *
 un-document 'struct sigaltstack' tag for stack_t as this is BSD-specific;  
 this doesn't seem useful enough to document as such * alternate stacks are
 per thread, not per process * update error codes to what the kernel does and
 POSIX requires

MFC after:	1 week
---
 lib/libc/sys/sigaltstack.2 | 32 +++++++++++++++-----------------
 1 file changed, 15 insertions(+), 17 deletions(-)

diff --git a/lib/libc/sys/sigaltstack.2 b/lib/libc/sys/sigaltstack.2
index da6877e6486..6119c6a3abc 100644
--- a/lib/libc/sys/sigaltstack.2
+++ b/lib/libc/sys/sigaltstack.2
@@ -28,7 +28,7 @@
 .\"     @(#)sigaltstack.2	8.2 (Berkeley) 5/1/95
 .\" $FreeBSD$
 .\"
-.Dd May 1, 1995
+.Dd May 6, 2010
 .Dt SIGALTSTACK 2
 .Os
 .Sh NAME
@@ -39,7 +39,7 @@
 .Sh SYNOPSIS
 .In signal.h
 .Bd -literal
-typedef struct sigaltstack {
+typedef struct {
         char    *ss_sp;
         size_t  ss_size;
         int     ss_flags;
@@ -51,25 +51,25 @@ typedef struct sigaltstack {
 The
 .Fn sigaltstack
 system call
-allows users to define an alternate stack on which signals
-are to be processed.
+allows defining an alternate stack on which signals
+are to be processed for the current thread.
 If
 .Fa ss
 is non-zero,
 it specifies a pointer to and the size of a
 .Em "signal stack"
-on which to deliver signals,
-and tells the system if the process is currently executing
-on that stack.
+on which to deliver signals.
 When a signal's action indicates its handler
 should execute on the signal stack (specified with a
 .Xr sigaction 2
 system call), the system checks to see
-if the process is currently executing on that stack.
-If the process is not currently executing on the signal stack,
+if the thread is currently executing on that stack.
+If the thread is not currently executing on the signal stack,
 the system arranges a switch to the signal stack for the
 duration of the signal handler's execution.
 .Pp
+An active stack cannot be modified.
+.Pp
 If
 .Dv SS_DISABLE
 is set in
@@ -78,12 +78,6 @@ is set in
 and
 .Fa ss_size
 are ignored and the signal stack will be disabled.
-Trying to disable an active stack will cause
-.Fn sigaltstack
-to return -1 with
-.Va errno
-set to
-.Er EINVAL .
 A disabled stack will cause all signals to be
 taken on the regular user stack.
 If the stack is later re-enabled then all signals that were specified
@@ -96,7 +90,7 @@ The
 .Fa ss_flags
 field will contain the value
 .Dv SS_ONSTACK
-if the process is currently on a signal stack and
+if the thread is currently on a signal stack and
 .Dv SS_DISABLE
 if the signal stack is currently disabled.
 .Sh NOTES
@@ -146,8 +140,12 @@ or
 .Fa oss
 points to memory that is not a valid part of the process
 address space.
+.It Bq Er EPERM
+An attempt was made to modify an active stack.
 .It Bq Er EINVAL
-An attempt was made to disable an active stack.
+The
+.Fa ss_flags
+field was invalid.
 .It Bq Er ENOMEM
 Size of alternate stack area is less than or equal to
 .Dv MINSIGSTKSZ .

From c84cab5626f45a39acf9c322e9b45f4459d0300b Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Thu, 6 May 2010 22:49:54 +0000
Subject: [PATCH 506/532] Update xrefs from 4.3BSD to modern signal functions
 in various man pages.

sigvec(2) references have been updated to sigaction(2), sigsetmask(2) and
sigblock(2) to sigprocmask(2), sigpause(2) to sigsuspend(2).

Some legacy man pages still refer to them, that is OK.
---
 lib/libc/gen/alarm.3        | 3 +--
 lib/libc/gen/siginterrupt.3 | 5 ++---
 lib/libc/gen/ualarm.3       | 4 ++--
 lib/libc/sys/execve.2       | 4 ++--
 lib/libc/sys/fcntl.2        | 2 +-
 lib/libc/sys/getitimer.2    | 2 +-
 lib/libc/sys/getrlimit.2    | 2 +-
 lib/libc/sys/sigaction.2    | 4 ----
 lib/libc/sys/sigreturn.2    | 2 +-
 lib/libc/sys/vfork.2        | 2 +-
 usr.bin/lastcomm/lastcomm.1 | 2 +-
 11 files changed, 13 insertions(+), 19 deletions(-)

diff --git a/lib/libc/gen/alarm.3 b/lib/libc/gen/alarm.3
index 70e7c6e739a..92817b8fce2 100644
--- a/lib/libc/gen/alarm.3
+++ b/lib/libc/gen/alarm.3
@@ -76,8 +76,7 @@ If no alarm is currently set, the return value is 0.
 .Sh SEE ALSO
 .Xr setitimer 2 ,
 .Xr sigaction 2 ,
-.Xr sigpause 2 ,
-.Xr sigvec 2 ,
+.Xr sigsuspend 2 ,
 .Xr signal 3 ,
 .Xr sleep 3 ,
 .Xr ualarm 3 ,
diff --git a/lib/libc/gen/siginterrupt.3 b/lib/libc/gen/siginterrupt.3
index e5eae626014..e790132b370 100644
--- a/lib/libc/gen/siginterrupt.3
+++ b/lib/libc/gen/siginterrupt.3
@@ -109,9 +109,8 @@ is not a valid signal number.
 .El
 .Sh SEE ALSO
 .Xr sigaction 2 ,
-.Xr sigblock 2 ,
-.Xr sigpause 2 ,
-.Xr sigsetmask 2 ,
+.Xr sigprocmask 2 ,
+.Xr sigsuspend 2 ,
 .Xr signal 3
 .Sh HISTORY
 The
diff --git a/lib/libc/gen/ualarm.3 b/lib/libc/gen/ualarm.3
index 47575cf73f7..57198962800 100644
--- a/lib/libc/gen/ualarm.3
+++ b/lib/libc/gen/ualarm.3
@@ -84,8 +84,8 @@ A microsecond is 0.000001 seconds.
 .Sh SEE ALSO
 .Xr getitimer 2 ,
 .Xr setitimer 2 ,
-.Xr sigpause 2 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
+.Xr sigsuspend 2 ,
 .Xr alarm 3 ,
 .Xr signal 3 ,
 .Xr sleep 3 ,
diff --git a/lib/libc/sys/execve.2 b/lib/libc/sys/execve.2
index cdd4197a998..acc647113bc 100644
--- a/lib/libc/sys/execve.2
+++ b/lib/libc/sys/execve.2
@@ -189,8 +189,8 @@ the calling process:
 .It interval timers Ta see Xr getitimer 2
 .It resource limits Ta see Xr getrlimit 2
 .It file mode mask Ta see Xr umask 2
-.It signal mask Ta see Xr sigvec 2 ,
-.Xr sigsetmask 2
+.It signal mask Ta see Xr sigaction 2 ,
+.Xr sigprocmask 2
 .El
 .Pp
 When a program is executed as a result of an
diff --git a/lib/libc/sys/fcntl.2 b/lib/libc/sys/fcntl.2
index 250cef46d5d..a801b60dc1f 100644
--- a/lib/libc/sys/fcntl.2
+++ b/lib/libc/sys/fcntl.2
@@ -618,7 +618,7 @@ for the reasons as stated in
 .Xr flock 2 ,
 .Xr getdtablesize 2 ,
 .Xr open 2 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr lockf 3 ,
 .Xr tcgetpgrp 3 ,
 .Xr tcsetpgrp 3
diff --git a/lib/libc/sys/getitimer.2 b/lib/libc/sys/getitimer.2
index 260f8f0dcfb..e942cc80fc6 100644
--- a/lib/libc/sys/getitimer.2
+++ b/lib/libc/sys/getitimer.2
@@ -171,7 +171,7 @@ to be handled.
 .Sh SEE ALSO
 .Xr gettimeofday 2 ,
 .Xr select 2 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr clocks 7
 .Sh HISTORY
 The
diff --git a/lib/libc/sys/getrlimit.2 b/lib/libc/sys/getrlimit.2
index 48feea0e440..35198bce306 100644
--- a/lib/libc/sys/getrlimit.2
+++ b/lib/libc/sys/getrlimit.2
@@ -193,7 +193,7 @@ raised the maximum limit value, and the caller is not the super-user.
 .Xr quota 1 ,
 .Xr quotactl 2 ,
 .Xr sigaltstack 2 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr sysctl 3 ,
 .Xr ulimit 3
 .Sh HISTORY
diff --git a/lib/libc/sys/sigaction.2 b/lib/libc/sys/sigaction.2
index 3b4f2f0196d..47b3a72d8d6 100644
--- a/lib/libc/sys/sigaction.2
+++ b/lib/libc/sys/sigaction.2
@@ -609,13 +609,9 @@ or
 .Xr kill 2 ,
 .Xr ptrace 2 ,
 .Xr sigaltstack 2 ,
-.Xr sigblock 2 ,
-.Xr sigpause 2 ,
 .Xr sigpending 2 ,
 .Xr sigprocmask 2 ,
-.Xr sigsetmask 2 ,
 .Xr sigsuspend 2 ,
-.Xr sigvec 2 ,
 .Xr wait 2 ,
 .Xr fpsetmask 3 ,
 .Xr setjmp 3 ,
diff --git a/lib/libc/sys/sigreturn.2 b/lib/libc/sys/sigreturn.2
index a0224709b14..02d75a83b6e 100644
--- a/lib/libc/sys/sigreturn.2
+++ b/lib/libc/sys/sigreturn.2
@@ -83,7 +83,7 @@ The process status longword is invalid or would improperly
 raise the privilege level of the process.
 .El
 .Sh SEE ALSO
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr setjmp 3 ,
 .Xr ucontext 3
 .Sh HISTORY
diff --git a/lib/libc/sys/vfork.2 b/lib/libc/sys/vfork.2
index 3cd33686edc..1cfaa6130de 100644
--- a/lib/libc/sys/vfork.2
+++ b/lib/libc/sys/vfork.2
@@ -104,7 +104,7 @@ Same as for
 .Xr _exit 2 ,
 .Xr fork 2 ,
 .Xr rfork 2 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr wait 2 ,
 .Xr exit 3
 .Sh HISTORY
diff --git a/usr.bin/lastcomm/lastcomm.1 b/usr.bin/lastcomm/lastcomm.1
index 1d714c5514f..8cb2039ab87 100644
--- a/usr.bin/lastcomm/lastcomm.1
+++ b/usr.bin/lastcomm/lastcomm.1
@@ -168,7 +168,7 @@ while
 will print details of each terminating command.
 .Sh SEE ALSO
 .Xr last 1 ,
-.Xr sigvec 2 ,
+.Xr sigaction 2 ,
 .Xr acct 5 ,
 .Xr core 5
 .Sh HISTORY

From 9402dff3de6195cb2060113b695e49f850b4102f Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 7 May 2010 04:14:07 +0000
Subject: [PATCH 507/532] Push down the page queues lock into
 vm_page_deactivate().  Eliminate an incorrect comment.

---
 sys/vm/vm_fault.c |  2 --
 sys/vm/vm_page.c  | 15 ++++++++++-----
 2 files changed, 10 insertions(+), 7 deletions(-)

diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index ae1ee290707..0088248142d 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -138,9 +138,7 @@ release_page(struct faultstate *fs)
 
 	vm_page_wakeup(fs->m);
 	vm_page_lock(fs->m);
-	vm_page_lock_queues();
 	vm_page_deactivate(fs->m);
-	vm_page_unlock_queues();
 	vm_page_unlock(fs->m);
 	fs->m = NULL;
 }
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index 9cbd2efa258..afa5cfb7667 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1618,10 +1618,8 @@ vm_page_unwire(vm_page_t m, int activate)
 	}
 }
 
-
 /*
- * Move the specified page to the inactive queue.  If the page has
- * any associated swap, the swap is deallocated.
+ * Move the specified page to the inactive queue.
  *
  * Normally athead is 0 resulting in LRU operation.  athead is set
  * to 1 if we want this page to be 'as if it were placed in the cache',
@@ -1633,7 +1631,6 @@ static inline void
 _vm_page_deactivate(vm_page_t m, int athead)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	vm_page_lock_assert(m, MA_OWNED);
 
 	/*
@@ -1642,6 +1639,7 @@ _vm_page_deactivate(vm_page_t m, int athead)
 	if (VM_PAGE_INQUEUE2(m, PQ_INACTIVE))
 		return;
 	if (m->wire_count == 0 && (m->flags & PG_UNMANAGED) == 0) {
+		vm_page_lock_queues();
 		vm_page_flag_clear(m, PG_WINATCFLS);
 		vm_pageq_remove(m);
 		if (athead)
@@ -1650,13 +1648,20 @@ _vm_page_deactivate(vm_page_t m, int athead)
 			TAILQ_INSERT_TAIL(&vm_page_queues[PQ_INACTIVE].pl, m, pageq);
 		VM_PAGE_SETQUEUE2(m, PQ_INACTIVE);
 		cnt.v_inactive_count++;
+		vm_page_unlock_queues();
 	}
 }
 
+/*
+ * Move the specified page to the inactive queue.
+ *
+ * The page must be locked.
+ */
 void
 vm_page_deactivate(vm_page_t m)
 {
-    _vm_page_deactivate(m, 0);
+
+	_vm_page_deactivate(m, 0);
 }
 
 /*

From 7072188017d1e4bbb4de8cc5684548eb82c7e08a Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 7 May 2010 05:23:15 +0000
Subject: [PATCH 508/532] Eliminate acquisitions of the page queues lock that
 are no longer needed.

Switch to a per-processor counter for the number of pages freed during
process termination.
---
 sys/sys/vmmeter.h  |  2 +-
 sys/vm/vm_object.c | 11 ++---------
 2 files changed, 3 insertions(+), 10 deletions(-)

diff --git a/sys/sys/vmmeter.h b/sys/sys/vmmeter.h
index 6c866c7f1dc..9f5756d987b 100644
--- a/sys/sys/vmmeter.h
+++ b/sys/sys/vmmeter.h
@@ -74,7 +74,7 @@ struct vmmeter {
 
 	u_int v_tcached;	/* (q) total pages cached */
 	u_int v_dfree;		/* (q) pages freed by daemon */
-	u_int v_pfree;		/* (q) pages freed by exiting processes */
+	u_int v_pfree;		/* (p) pages freed by exiting processes */
 	u_int v_tfree;		/* (p) total pages freed */
 	/*
 	 * Distribution of page usages.
diff --git a/sys/vm/vm_object.c b/sys/vm/vm_object.c
index 3986d142d87..5428c98f544 100644
--- a/sys/vm/vm_object.c
+++ b/sys/vm/vm_object.c
@@ -722,14 +722,11 @@ vm_object_terminate(vm_object_t object)
 			("vm_object_terminate: freeing busy page %p "
 			"p->busy = %d, p->oflags %x\n", p, p->busy, p->oflags));
 		vm_page_lock(p);
-		vm_page_lock_queues();
 		if (p->wire_count == 0) {
 			vm_page_free(p);
-			cnt.v_pfree++;
-		} else {
+			PCPU_INC(cnt.v_pfree);
+		} else
 			vm_page_remove(p);
-		}
-		vm_page_unlock_queues();
 		vm_page_unlock(p);
 	}
 
@@ -1634,14 +1631,12 @@ vm_object_backing_scan(vm_object_t object, int op)
 				 * can simply destroy it. 
 				 */
 				vm_page_lock(p);
-				vm_page_lock_queues();
 				KASSERT(!pmap_page_is_mapped(p),
 				    ("freeing mapped page %p", p));
 				if (p->wire_count == 0)
 					vm_page_free(p);
 				else
 					vm_page_remove(p);
-				vm_page_unlock_queues();
 				vm_page_unlock(p);
 				p = next;
 				continue;
@@ -1660,14 +1655,12 @@ vm_object_backing_scan(vm_object_t object, int op)
 				 * Leave the parent's page alone
 				 */
 				vm_page_lock(p);
-				vm_page_lock_queues();
 				KASSERT(!pmap_page_is_mapped(p),
 				    ("freeing mapped page %p", p));
 				if (p->wire_count == 0)
 					vm_page_free(p);
 				else
 					vm_page_remove(p);
-				vm_page_unlock_queues();
 				vm_page_unlock(p);
 				p = next;
 				continue;

From c1f98960a37925b643ea044473b12a3cd6c00911 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 7 May 2010 06:58:53 +0000
Subject: [PATCH 509/532] Update the synchronization requirements for the page
 usage count.

---
 sys/vm/vm_page.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/vm/vm_page.h b/sys/vm/vm_page.h
index ef790ae959e..b2b92e92a2c 100644
--- a/sys/vm/vm_page.h
+++ b/sys/vm/vm_page.h
@@ -116,7 +116,7 @@ struct vm_page {
 	u_int wire_count;		/* wired down maps refs (P) */
 	short hold_count;		/* page hold count (P) */
 	u_short oflags;			/* page flags (O) */
-	u_char	act_count;		/* page usage count (Q) */
+	u_char	act_count;		/* page usage count (P) */
 	u_char	busy;			/* page busy count (O) */
 	/* NOTE that these must support one bit per DEV_BSIZE in a page!!! */
 	/* so, on normal X86 kernels, they must be at least 8 bits wide */

From 2c3ae115b6789421867309268dab0299ac41f130 Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Fri, 7 May 2010 08:20:56 +0000
Subject: [PATCH 510/532]  - Use the correct flag mask when determining whether
 an inode has    successfully made it to the free list yet or not.  This fixes
    a deadlock that can occur with unlinked but referenced files.    Journal
 space and inodedeps were not correctly reclaimed because    the inode block
 was not left dirty.

Tested/Reported by:	lwindschuh@googlemail.com
---
 sys/ufs/ffs/ffs_softdep.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index b05339b936c..90ed314d3b8 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -9301,7 +9301,7 @@ handle_written_inodeblock(inodedep, bp)
 			hadchanges = 1;
 	}
 	/* Leave this inodeblock dirty until it's in the list. */
-	if ((inodedep->id_state & (UNLINKED | DEPCOMPLETE)) == UNLINKED)
+	if ((inodedep->id_state & (UNLINKED | UNLINKONLIST)) == UNLINKED)
 		hadchanges = 1;
 	/*
 	 * If we had to rollback the inode allocation because of

From 8ef48de8882c9d0aaab7baf256dfb52c2d29e423 Mon Sep 17 00:00:00 2001
From: Jeff Roberson <jeff@FreeBSD.org>
Date: Fri, 7 May 2010 08:45:21 +0000
Subject: [PATCH 511/532]  - Call softdep_prealloc() before any of the balloc
 routines in the    snapshot code.  - Don't fsync() vnodes in prealloc if copy
 on write is in progress.  It    is not safe to recurse back into the write
 path here.

Reported by:	Vladimir Grebenschikov <vova@fbsd.ru>
---
 sys/ufs/ffs/ffs_snapshot.c | 8 ++++++++
 sys/ufs/ffs/ffs_softdep.c  | 3 ++-
 2 files changed, 10 insertions(+), 1 deletion(-)

diff --git a/sys/ufs/ffs/ffs_snapshot.c b/sys/ufs/ffs/ffs_snapshot.c
index 11362cfbc75..f6548a385eb 100644
--- a/sys/ufs/ffs/ffs_snapshot.c
+++ b/sys/ufs/ffs/ffs_snapshot.c
@@ -1002,6 +1002,8 @@ expunge_ufs1(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
 	if (lbn < NDADDR) {
 		blkno = VTOI(snapvp)->i_din1->di_db[lbn];
 	} else {
+		if (DOINGSOFTDEP(snapvp))
+			softdep_prealloc(snapvp, MNT_WAIT);
 		td->td_pflags |= TDP_COWINPROGRESS;
 		error = ffs_balloc_ufs1(snapvp, lblktosize(fs, (off_t)lbn),
 		   fs->fs_bsize, KERNCRED, BA_METAONLY, &bp);
@@ -1283,6 +1285,8 @@ expunge_ufs2(snapvp, cancelip, fs, acctfunc, expungetype, clearmode)
 	if (lbn < NDADDR) {
 		blkno = VTOI(snapvp)->i_din2->di_db[lbn];
 	} else {
+		if (DOINGSOFTDEP(snapvp))
+			softdep_prealloc(snapvp, MNT_WAIT);
 		td->td_pflags |= TDP_COWINPROGRESS;
 		error = ffs_balloc_ufs2(snapvp, lblktosize(fs, (off_t)lbn),
 		   fs->fs_bsize, KERNCRED, BA_METAONLY, &bp);
@@ -1746,6 +1750,8 @@ retry:
 		goto retry;
 	TAILQ_FOREACH(ip, &sn->sn_head, i_nextsnap) {
 		vp = ITOV(ip);
+		if (DOINGSOFTDEP(vp))
+			softdep_prealloc(vp, MNT_WAIT);
 		/*
 		 * Lookup block being written.
 		 */
@@ -2268,6 +2274,8 @@ ffs_copyonwrite(devvp, bp)
 	}
 	TAILQ_FOREACH(ip, &sn->sn_head, i_nextsnap) {
 		vp = ITOV(ip);
+		if (DOINGSOFTDEP(vp))
+			softdep_prealloc(vp, MNT_WAIT);
 		/*
 		 * We ensure that everything of our own that needs to be
 		 * copied will be done at the time that ffs_snapshot is
diff --git a/sys/ufs/ffs/ffs_softdep.c b/sys/ufs/ffs/ffs_softdep.c
index 90ed314d3b8..49510a71402 100644
--- a/sys/ufs/ffs/ffs_softdep.c
+++ b/sys/ufs/ffs/ffs_softdep.c
@@ -2475,7 +2475,8 @@ softdep_prealloc(vp, waitok)
 	 * Attempt to sync this vnode once to flush any journal
 	 * work attached to it.
 	 */
-	ffs_syncvnode(vp, waitok);
+	if ((curthread->td_pflags & TDP_COWINPROGRESS) == 0)
+		ffs_syncvnode(vp, waitok);
 	ACQUIRE_LOCK(&lk);
 	process_removes(vp);
 	if (journal_space(ump, 0) == 0) {

From dc510c105f9712a7b36ec428f3700f7e56f63450 Mon Sep 17 00:00:00 2001
From: Edward Tomasz Napierala <trasz@FreeBSD.org>
Date: Fri, 7 May 2010 14:30:21 +0000
Subject: [PATCH 512/532] Enforce RLIMIT_FSIZE in ZFS.

Reviewed by:	pjd@
---
 sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
index 59a58dd7a5a..225beca116d 100644
--- a/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
+++ b/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_vnops.c
@@ -3977,6 +3977,9 @@ zfs_freebsd_write(ap)
 	} */ *ap;
 {
 
+	if (vn_rlimit_fsize(ap->a_vp, ap->a_uio, ap->a_uio->uio_td))
+		return (EFBIG);
+
 	return (zfs_write(ap->a_vp, ap->a_uio, ap->a_ioflag, ap->a_cred, NULL));
 }
 

From 03679e233408833452145f89f49c5929842140d0 Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 7 May 2010 15:49:43 +0000
Subject: [PATCH 513/532] Push down the page queues lock into
 vm_page_activate().

---
 sys/fs/nfsclient/nfs_clbio.c | 17 +++++++++--------
 sys/fs/nwfs/nwfs_io.c        | 17 +++++++++--------
 sys/fs/smbfs/smbfs_io.c      | 17 +++++++++--------
 sys/nfsclient/nfs_bio.c      | 17 +++++++++--------
 sys/vm/vm_fault.c            |  7 +------
 sys/vm/vm_page.c             | 11 +++++++----
 sys/vm/vnode_pager.c         | 15 +++++++++------
 7 files changed, 53 insertions(+), 48 deletions(-)

diff --git a/sys/fs/nfsclient/nfs_clbio.c b/sys/fs/nfsclient/nfs_clbio.c
index 0224e157b6e..7d6fcab6809 100644
--- a/sys/fs/nfsclient/nfs_clbio.c
+++ b/sys/fs/nfsclient/nfs_clbio.c
@@ -196,9 +196,6 @@ ncl_getpages(struct vop_getpages_args *ap)
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
-		vm_page_lock(m);
-		vm_page_lock_queues();
-
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -236,18 +233,22 @@ ncl_getpages(struct vop_getpages_args *ap)
 			 * now tell them that it is ok to use.
 			 */
 			if (!error) {
-				if (m->oflags & VPO_WANTED)
+				if (m->oflags & VPO_WANTED) {
+					vm_page_lock(m);
 					vm_page_activate(m);
-				else
+					vm_page_unlock(m);
+				} else {
+					vm_page_lock(m);
 					vm_page_deactivate(m);
+					vm_page_unlock(m);
+				}
 				vm_page_wakeup(m);
 			} else {
+				vm_page_lock(m);
 				vm_page_free(m);
+				vm_page_unlock(m);
 			}
 		}
-
-		vm_page_unlock_queues();
-		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(object);
 	return (0);
diff --git a/sys/fs/nwfs/nwfs_io.c b/sys/fs/nwfs/nwfs_io.c
index 125201408ad..0bd7dc355ff 100644
--- a/sys/fs/nwfs/nwfs_io.c
+++ b/sys/fs/nwfs/nwfs_io.c
@@ -446,9 +446,6 @@ nwfs_getpages(ap)
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
-		vm_page_lock(m);
-		vm_page_lock_queues();
-
 		if (nextoff <= size) {
 			m->valid = VM_PAGE_BITS_ALL;
 			KASSERT(m->dirty == 0,
@@ -474,18 +471,22 @@ nwfs_getpages(ap)
 			 * now tell them that it is ok to use.
 			 */
 			if (!error) {
-				if (m->oflags & VPO_WANTED)
+				if (m->oflags & VPO_WANTED) {
+					vm_page_lock(m);
 					vm_page_activate(m);
-				else
+					vm_page_unlock(m);
+				} else {
+					vm_page_lock(m);
 					vm_page_deactivate(m);
+					vm_page_unlock(m);
+				}
 				vm_page_wakeup(m);
 			} else {
+				vm_page_lock(m);
 				vm_page_free(m);
+				vm_page_unlock(m);
 			}
 		}
-
-		vm_page_unlock_queues();
-		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(object);
 	return 0;
diff --git a/sys/fs/smbfs/smbfs_io.c b/sys/fs/smbfs/smbfs_io.c
index 1ffcde9fb80..9becfdda13e 100644
--- a/sys/fs/smbfs/smbfs_io.c
+++ b/sys/fs/smbfs/smbfs_io.c
@@ -497,9 +497,6 @@ smbfs_getpages(ap)
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
-		vm_page_lock(m);
-		vm_page_lock_queues();
-
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -538,18 +535,22 @@ smbfs_getpages(ap)
 			 * now tell them that it is ok to use.
 			 */
 			if (!error) {
-				if (m->oflags & VPO_WANTED)
+				if (m->oflags & VPO_WANTED) {
+					vm_page_lock(m);
 					vm_page_activate(m);
-				else
+					vm_page_unlock(m);
+				} else {
+					vm_page_lock(m);
 					vm_page_deactivate(m);
+					vm_page_unlock(m);
+				}
 				vm_page_wakeup(m);
 			} else {
+				vm_page_lock(m);
 				vm_page_free(m);
+				vm_page_unlock(m);
 			}
 		}
-
-		vm_page_unlock_queues();
-		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(object);
 	return 0;
diff --git a/sys/nfsclient/nfs_bio.c b/sys/nfsclient/nfs_bio.c
index 37da34b8cce..e85fab8975d 100644
--- a/sys/nfsclient/nfs_bio.c
+++ b/sys/nfsclient/nfs_bio.c
@@ -194,9 +194,6 @@ nfs_getpages(struct vop_getpages_args *ap)
 		nextoff = toff + PAGE_SIZE;
 		m = pages[i];
 
-		vm_page_lock(m);
-		vm_page_lock_queues();
-
 		if (nextoff <= size) {
 			/*
 			 * Read operation filled an entire page
@@ -234,18 +231,22 @@ nfs_getpages(struct vop_getpages_args *ap)
 			 * now tell them that it is ok to use.
 			 */
 			if (!error) {
-				if (m->oflags & VPO_WANTED)
+				if (m->oflags & VPO_WANTED) {
+					vm_page_lock(m);
 					vm_page_activate(m);
-				else
+					vm_page_unlock(m);
+				} else {
+					vm_page_lock(m);
 					vm_page_deactivate(m);
+					vm_page_unlock(m);
+				}
 				vm_page_wakeup(m);
 			} else {
+				vm_page_lock(m);
 				vm_page_free(m);
+				vm_page_unlock(m);
 			}
 		}
-
-		vm_page_unlock_queues();
-		vm_page_unlock(m);
 	}
 	VM_OBJECT_UNLOCK(object);
 	return (0);
diff --git a/sys/vm/vm_fault.c b/sys/vm/vm_fault.c
index 0088248142d..05c3228643e 100644
--- a/sys/vm/vm_fault.c
+++ b/sys/vm/vm_fault.c
@@ -937,7 +937,6 @@ vnode_locked:
 		vm_fault_prefault(fs.map->pmap, vaddr, fs.entry);
 	VM_OBJECT_LOCK(fs.object);
 	vm_page_lock(fs.m);
-	vm_page_lock_queues();
 
 	/*
 	 * If the page is not wired down, then put it where the pageout daemon
@@ -948,10 +947,8 @@ vnode_locked:
 			vm_page_wire(fs.m);
 		else
 			vm_page_unwire(fs.m, 1);
-	} else {
+	} else
 		vm_page_activate(fs.m);
-	}
-	vm_page_unlock_queues();
 	vm_page_unlock(fs.m);
 	vm_page_wakeup(fs.m);
 
@@ -1267,9 +1264,7 @@ vm_fault_copy_entry(vm_map_t dst_map, vm_map_t src_map,
 			vm_page_unlock(dst_m);
 		} else {
 			vm_page_lock(dst_m);
-			vm_page_lock_queues();
 			vm_page_activate(dst_m);
-			vm_page_unlock_queues();
 			vm_page_unlock(dst_m);
 		}
 		vm_page_wakeup(dst_m);
diff --git a/sys/vm/vm_page.c b/sys/vm/vm_page.c
index afa5cfb7667..455c9902945 100644
--- a/sys/vm/vm_page.c
+++ b/sys/vm/vm_page.c
@@ -1374,22 +1374,25 @@ vm_page_enqueue(int queue, vm_page_t m)
  *	Ensure that act_count is at least ACT_INIT but do not otherwise
  *	mess with it.
  *
- *	The page queues must be locked.
+ *	The page must be locked.
  *	This routine may not block.
  */
 void
 vm_page_activate(vm_page_t m)
 {
 
-	mtx_assert(&vm_page_queue_mtx, MA_OWNED);
 	vm_page_lock_assert(m, MA_OWNED);
 	if (VM_PAGE_GETKNOWNQUEUE2(m) != PQ_ACTIVE) {
-		vm_pageq_remove(m);
 		if (m->wire_count == 0 && (m->flags & PG_UNMANAGED) == 0) {
 			if (m->act_count < ACT_INIT)
 				m->act_count = ACT_INIT;
+			vm_page_lock_queues();
+			vm_pageq_remove(m);
 			vm_page_enqueue(PQ_ACTIVE, m);
-		}
+			vm_page_unlock_queues();
+		} else
+			KASSERT(m->queue == PQ_NONE,
+			    ("vm_page_activate: wired page %p is queued", m));
 	} else {
 		if (m->act_count < ACT_INIT)
 			m->act_count = ACT_INIT;
diff --git a/sys/vm/vnode_pager.c b/sys/vm/vnode_pager.c
index 69fd5d101c7..cfe9f994e0e 100644
--- a/sys/vm/vnode_pager.c
+++ b/sys/vm/vnode_pager.c
@@ -948,8 +948,6 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 		nextoff = tfoff + PAGE_SIZE;
 		mt = m[i];
 
-		vm_page_lock(mt);
-		vm_page_lock_queues();
 		if (nextoff <= object->un_pager.vnp.vnp_size) {
 			/*
 			 * Read filled up entire page.
@@ -992,17 +990,22 @@ vnode_pager_generic_getpages(vp, m, bytecount, reqpage)
 			 * now tell them that it is ok to use
 			 */
 			if (!error) {
-				if (mt->oflags & VPO_WANTED)
+				if (mt->oflags & VPO_WANTED) {
+					vm_page_lock(mt);
 					vm_page_activate(mt);
-				else
+					vm_page_unlock(mt);
+				} else {
+					vm_page_lock(mt);
 					vm_page_deactivate(mt);
+					vm_page_unlock(mt);
+				}
 				vm_page_wakeup(mt);
 			} else {
+				vm_page_lock(mt);
 				vm_page_free(mt);
+				vm_page_unlock(mt);
 			}
 		}
-		vm_page_unlock_queues();
-		vm_page_unlock(mt);
 	}
 	VM_OBJECT_UNLOCK(object);
 	if (error) {

From 97c3834772256c91a828cfe74f03f8366c369dbb Mon Sep 17 00:00:00 2001
From: Alan Cox <alc@FreeBSD.org>
Date: Fri, 7 May 2010 16:22:06 +0000
Subject: [PATCH 514/532] Eliminate unnecessary page queues locking.

---
 sys/vm/swap_pager.c | 4 ----
 1 file changed, 4 deletions(-)

diff --git a/sys/vm/swap_pager.c b/sys/vm/swap_pager.c
index 6017a52d4c7..2f97ee8e122 100644
--- a/sys/vm/swap_pager.c
+++ b/sys/vm/swap_pager.c
@@ -1140,16 +1140,12 @@ swap_pager_getpages(vm_object_t object, vm_page_t *m, int count, int reqpage)
 		
 		for (k = 0; k < i; ++k) {
 			vm_page_lock(m[k]);
-			vm_page_lock_queues();
 			swp_pager_free_nrpage(m[k]);
-			vm_page_unlock_queues();
 			vm_page_unlock(m[k]);
 		}
 		for (k = j; k < count; ++k) {
 			vm_page_lock(m[k]);
-			vm_page_lock_queues();
 			swp_pager_free_nrpage(m[k]);
-			vm_page_unlock_queues();
 			vm_page_unlock(m[k]);
 		}
 	}

From a026a25b1dfe835868324be952a04e6eeba8beac Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 7 May 2010 16:43:00 +0000
Subject: [PATCH 515/532] It seems controller has two types of promiscuous
 control, one for unicast and the other for multicast. To receive multicast
 frames that host didn't join in promiscuous mode, driver have to set
 promiscuous mode for multicast frames as well. The Open Source Software
 Developer Manual for i8255x was not clear how to handle promiscuous mode.

PR:		kern/145905
MFC after:	5 days
---
 sys/dev/fxp/if_fxp.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/dev/fxp/if_fxp.c b/sys/dev/fxp/if_fxp.c
index a480e7cc6a4..4be525d3763 100644
--- a/sys/dev/fxp/if_fxp.c
+++ b/sys/dev/fxp/if_fxp.c
@@ -2345,7 +2345,7 @@ fxp_init_body(struct fxp_softc *sc)
 	cbp->force_fdx =	0;	/* (don't) force full duplex */
 	cbp->fdx_pin_en =	1;	/* (enable) FDX# pin */
 	cbp->multi_ia =		0;	/* (don't) accept multiple IAs */
-	cbp->mc_all =		ifp->if_flags & IFF_ALLMULTI ? 1 : 0;
+	cbp->mc_all =		ifp->if_flags & IFF_ALLMULTI ? 1 : prm;
 	cbp->gamla_rx =		sc->flags & FXP_FLAG_EXT_RFA ? 1 : 0;
 	cbp->vlan_strip_en =	((sc->flags & FXP_FLAG_EXT_RFA) != 0 &&
 	    (ifp->if_capenable & IFCAP_VLAN_HWTAGGING) != 0) ? 1 : 0;

From 8949fcf0dcacfe10e6fdeba6130466d014612866 Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 7 May 2010 17:20:15 +0000
Subject: [PATCH 516/532] raise(3): Note that this sends a signal to the
 current thread, not process.

This is how it works (with threading libraries loaded) and what POSIX
requires.

MFC after:	1 week
---
 lib/libc/gen/raise.3 | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)

diff --git a/lib/libc/gen/raise.3 b/lib/libc/gen/raise.3
index 5f422c5f072..fe75268eb83 100644
--- a/lib/libc/gen/raise.3
+++ b/lib/libc/gen/raise.3
@@ -32,12 +32,12 @@
 .\"     @(#)raise.3	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd June 4, 1993
+.Dd May 7, 2010
 .Dt RAISE 3
 .Os
 .Sh NAME
 .Nm raise
-.Nd send a signal to the current process
+.Nd send a signal to the current thread
 .Sh LIBRARY
 .Lb libc
 .Sh SYNOPSIS
@@ -49,7 +49,7 @@ The
 .Fn raise
 function sends the signal
 .Fa sig
-to the current process.
+to the current thread.
 .Sh RETURN VALUES
 .Rv -std raise
 .Sh ERRORS

From 7251782989820f799d9720f0415d724ea1be71a4 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 7 May 2010 18:03:35 +0000
Subject: [PATCH 517/532] 8255x configure command requires number of bytes of
 configuration table. The default size of the configuration table was 22
 bytes. To use extended feature of 82550/82551 the configuration table size
 was expanded to 32 bytes. The added configuration for 82550/82551 specifies
 VLAN hardware tagging and IPSec configuration as well as TCO. To make
 configuration easier fxp(4) used a configuration template and the template
 was copied to configuration table. After that, some parameters of the
 configuration table was changed depending on controller type and operation
 mode. However the size of template was 22 bytes so some configuration
 parameters were not properly initialized on 82550/82551. Fix this by
 increasing the template size. For 82557, 82558 and 82559 the size of the
 configuration is still 22 bytes.

---
 sys/dev/fxp/if_fxp.c | 17 +++++++++++++----
 1 file changed, 13 insertions(+), 4 deletions(-)

diff --git a/sys/dev/fxp/if_fxp.c b/sys/dev/fxp/if_fxp.c
index 4be525d3763..9c0b44aef75 100644
--- a/sys/dev/fxp/if_fxp.c
+++ b/sys/dev/fxp/if_fxp.c
@@ -106,9 +106,8 @@ static int tx_threshold = 64;
 
 /*
  * The configuration byte map has several undefined fields which
- * must be one or must be zero.  Set up a template for these bits
- * only, (assuming a 82557 chip) leaving the actual configuration
- * to fxp_init.
+ * must be one or must be zero.  Set up a template for these bits.
+ * The actual configuration is performed in fxp_init.
  *
  * See struct fxp_cb_config for the bit definitions.
  */
@@ -137,7 +136,17 @@ static u_char fxp_cb_config_template[] = {
 	0xf0,	/* 18 */
 	0x0,	/* 19 */
 	0x3f,	/* 20 */
-	0x5	/* 21 */
+	0x5,	/* 21 */
+	0x0,	/* 22 */
+	0x0,	/* 23 */
+	0x0,	/* 24 */
+	0x0,	/* 25 */
+	0x0,	/* 26 */
+	0x0,	/* 27 */
+	0x0,	/* 28 */
+	0x0,	/* 29 */
+	0x0,	/* 30 */
+	0x0	/* 31 */
 };
 
 /*

From af4b86b949628f084b659c3b156973b644b9aa2e Mon Sep 17 00:00:00 2001
From: Konstantin Belousov <kib@FreeBSD.org>
Date: Fri, 7 May 2010 18:57:26 +0000
Subject: [PATCH 518/532] One more use for vm_pageout_init_marker().

Reviewed by:	alc
---
 sys/vm/vm_pageout.c | 9 +--------
 1 file changed, 1 insertion(+), 8 deletions(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index a84b6d40965..0524193550e 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -765,14 +765,7 @@ vm_pageout_scan(int pass)
 	 */
 	page_shortage = vm_paging_target() + addl_page_shortage_init;
 
-	/*
-	 * Initialize our marker
-	 */
-	bzero(&marker, sizeof(marker));
-	marker.flags = PG_FICTITIOUS | PG_MARKER;
-	marker.oflags = VPO_BUSY;
-	marker.queue = PQ_INACTIVE;
-	marker.wire_count = 1;
+	vm_pageout_marker_init(&marker, PQ_INACTIVE);
 
 	/*
 	 * Start scanning the inactive queue for pages we can move to the

From 782434abe557eb36ed1da658db4182934e508562 Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Fri, 7 May 2010 19:48:50 +0000
Subject: [PATCH 519/532] Rework the calltree top view by critical callchain.
 The percentage show is the sum of the cost for the codepath.

MFC after: 1 week
---
 usr.sbin/pmcstat/pmcpl_calltree.c | 188 ++++++++++++++++--------------
 1 file changed, 103 insertions(+), 85 deletions(-)

diff --git a/usr.sbin/pmcstat/pmcpl_calltree.c b/usr.sbin/pmcstat/pmcpl_calltree.c
index f8ceece158b..09551334d55 100644
--- a/usr.sbin/pmcstat/pmcpl_calltree.c
+++ b/usr.sbin/pmcstat/pmcpl_calltree.c
@@ -120,8 +120,15 @@ struct pmcpl_ct_node_hash {
 struct pmcpl_ct_sample pmcpl_ct_callid;
 
 #define PMCPL_CT_MAXCOL		PMC_CALLCHAIN_DEPTH_MAX	
-#define PMCPL_CT_MAXLINE	256
-struct pmcpl_ct_node  *pmcpl_ct_topscreen[PMCPL_CT_MAXCOL][PMCPL_CT_MAXLINE];
+#define PMCPL_CT_MAXLINE	1024	/* TODO: dynamic. */
+
+struct pmcpl_ct_line {
+	unsigned	ln_sum;
+	unsigned	ln_index;
+};
+
+struct pmcpl_ct_line	pmcpl_ct_topmax[PMCPL_CT_MAXLINE+1];
+struct pmcpl_ct_node	*pmcpl_ct_topscreen[PMCPL_CT_MAXCOL+1][PMCPL_CT_MAXLINE+1];
 
 /*
  * All nodes indexed by function/image name are placed in a hash table.
@@ -224,28 +231,6 @@ pmcpl_ct_arc_grow(int cursize, int *maxsize, struct pmcpl_ct_arc **items)
 	*maxsize = nmaxsize;
 }
 
-/*
- * Compare two arc by samples value.
- */
-static int
-pmcpl_ct_arc_compare(void *thunk, const void *a, const void *b)
-{
-	const struct pmcpl_ct_arc *ct1, *ct2;
-	int pmcin = *(int *)thunk;
-
-	ct1 = (const struct pmcpl_ct_arc *) a;
-	ct2 = (const struct pmcpl_ct_arc *) b;
-
-	/* Sort in reverse order */
-	if (PMCPL_CT_SAMPLE(pmcin, &ct1->pcta_samples) <
-	    PMCPL_CT_SAMPLE(pmcin, &ct2->pcta_samples))
-		return (1);
-	if (PMCPL_CT_SAMPLE(pmcin, &ct1->pcta_samples) >
-	    PMCPL_CT_SAMPLE(pmcin, &ct2->pcta_samples))
-		return (-1);
-	return (0);
-}
-
 /*
  * Grow the instr table.
  */
@@ -366,9 +351,9 @@ pmcpl_ct_node_cleartag(void)
 
 static int
 pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
-    struct pmcpl_ct_sample *rsamples, int x, int *y, int maxy)
+    struct pmcpl_ct_sample *rsamples, int x, int *y)
 {
-	int i;
+	int i, terminal;
 
 	if (ct->pct_flags & PMCPL_PCT_TAG)
 		return 0;
@@ -382,12 +367,21 @@ pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
 	pmcpl_ct_topscreen[x][*y] = ct;
 
 	/*
-	 * This is a terminal node
+	 * Check if this is a terminal node.
+	 * We need to check that some samples exist
+	 * for at least one arc for that PMC.
 	 */
-	if (ct->pct_narc == 0) {
+	terminal = 1;
+	for (i = 0; i < ct->pct_narc; i++)
+		if (PMCPL_CT_SAMPLE(pmcin,
+		    &ct->pct_arc[i].pcta_samples) != 0) {
+			terminal = 0;
+			break;
+		}
+
+	if (ct->pct_narc == 0 || terminal) {
 		pmcpl_ct_topscreen[x+1][*y] = NULL;
-		if (*y >= PMCPL_CT_MAXLINE ||
-		    *y >= maxy)
+		if (*y >= PMCPL_CT_MAXLINE)
 			return 1;
 		*y = *y + 1;
 		for (i=0; i < x; i++)
@@ -396,14 +390,7 @@ pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
 		return 0;
 	}
 
-	/*
-	 * Quicksort the arcs.
-	 */
-	qsort_r(ct->pct_arc, ct->pct_narc, sizeof(struct pmcpl_ct_arc),
-	    &pmcin, pmcpl_ct_arc_compare);
-
 	for (i = 0; i < ct->pct_narc; i++) {
-		/* Skip this arc if there is no sample at all. */
 		if (PMCPL_CT_SAMPLE(pmcin,
 		    &ct->pct_arc[i].pcta_samples) == 0)
 			continue;
@@ -411,7 +398,7 @@ pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
 		    &ct->pct_arc[i].pcta_samples) > pmcstat_threshold) {
 			if (pmcpl_ct_node_dumptop(pmcin,
 			        ct->pct_arc[i].pcta_child,
-			        rsamples, x+1, y, maxy))
+			        rsamples, x+1, y))
 				return 1;
 		}
 	}
@@ -419,6 +406,25 @@ pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
 	return 0;
 }
 
+/*
+ * Compare two top line by sum.
+ */
+static int
+pmcpl_ct_line_compare(const void *a, const void *b)
+{
+	const struct pmcpl_ct_line *ct1, *ct2;
+
+	ct1 = (const struct pmcpl_ct_line *) a;
+	ct2 = (const struct pmcpl_ct_line *) b;
+
+	/* Sort in reverse order */
+	if (ct1->ln_sum < ct2->ln_sum)
+		return (1);
+	if (ct1->ln_sum > ct2->ln_sum)
+		return (-1);
+	return (0);
+}
+
 /*
  * Format and display given PMC index.
  */
@@ -426,6 +432,11 @@ pmcpl_ct_node_dumptop(int pmcin, struct pmcpl_ct_node *ct,
 static void
 pmcpl_ct_node_printtop(struct pmcpl_ct_sample *rsamples, int pmcin, int maxy)
 {
+#undef	TS
+#undef	TSI
+#define	TS(x, y)	(pmcpl_ct_topscreen[x][y])
+#define	TSI(x, y)	(pmcpl_ct_topscreen[x][pmcpl_ct_topmax[y].ln_index])
+
 	int v_attrs, ns_len, vs_len, is_len, width, indentwidth, x, y;
 	float v;
 	char ns[30], vs[10], is[20];
@@ -433,33 +444,60 @@ pmcpl_ct_node_printtop(struct pmcpl_ct_sample *rsamples, int pmcin, int maxy)
 	struct pmcstat_symbol *sym;
 	const char *space = " ";
 
+	/*
+	 * Sort by line cost.
+	 */
+	for (y = 0; ; y++) {
+		ct = TS(1, y);
+		if (ct == NULL)
+			break;
+
+		pmcpl_ct_topmax[y].ln_sum = 0;
+		pmcpl_ct_topmax[y].ln_index = y;
+		for (x = 1; TS(x, y) != NULL; x++) {
+			pmcpl_ct_topmax[y].ln_sum +=
+			    PMCPL_CT_SAMPLE(pmcin, &TS(x, y)->pct_samples);
+		}
+	}
+	qsort(pmcpl_ct_topmax, y, sizeof(pmcpl_ct_topmax[0]),
+	    pmcpl_ct_line_compare);
+	pmcpl_ct_topmax[y].ln_index = y;
+
 	for (y = 0; y < maxy; y++) {
-		/* Output image. */
-		ct = pmcpl_ct_topscreen[0][y];
-		snprintf(is, sizeof(is), "%-10.10s",
-		    pmcstat_string_unintern(ct->pct_image->pi_name));
-		PMCSTAT_PRINTW("%s ", is);
-		width = indentwidth = 11;
+		ct = TSI(1, y);
+		if (ct == NULL)
+			break;
 
-		for (x = 0; pmcpl_ct_topscreen[x][y] !=NULL; x++) {
+		if (y > 0)
+			PMCSTAT_PRINTW("\n");
 
-			ct = pmcpl_ct_topscreen[x][y];
+		/* Output sum. */
+		v = pmcpl_ct_topmax[y].ln_sum * 100.0 /
+		    rsamples->sb[pmcin];
+		snprintf(vs, sizeof(vs), "%.1f", v);
+		v_attrs = PMCSTAT_ATTRPERCENT(v);
+		PMCSTAT_ATTRON(v_attrs);
+		PMCSTAT_PRINTW("%5.5s ", vs);
+		PMCSTAT_ATTROFF(v_attrs);
+
+		width = indentwidth = 5 + 1;
+
+		for (x = 1; (ct = TSI(x, y)) != NULL; x++) {
 
-			ns[0] = '\0'; ns_len = 0;
 			vs[0] = '\0'; vs_len = 0;
 			is[0] = '\0'; is_len = 0;
 
 			/* Format value. */
 			v = PMCPL_CT_SAMPLEP(pmcin, &ct->pct_samples);
 			if (v > pmcstat_threshold)
-				vs_len  = snprintf(vs, sizeof(vs), "(%.1f%%)", v);
+				vs_len  = snprintf(vs, sizeof(vs),
+				    "(%.1f%%)", v);
 			v_attrs = PMCSTAT_ATTRPERCENT(v);
 
 			if (pmcstat_skiplink && v <= pmcstat_threshold) {
-				PMCSTAT_PRINTW(". ");
-				width += 2;
-				continue;
-			}
+				strlcpy(ns, ".", sizeof(ns));
+				ns_len = 1;
+			} else {
 			sym = pmcstat_symbol_search(ct->pct_image, ct->pct_func);
 			if (sym != NULL) {
 				ns_len = snprintf(ns, sizeof(ns), "%s",
@@ -469,12 +507,14 @@ pmcpl_ct_node_printtop(struct pmcpl_ct_sample *rsamples, int pmcin, int maxy)
 				    (void *)ct->pct_func);
 
 			/* Format image. */
-			if (x > 0 && pmcpl_ct_topscreen[x-1][y]->pct_image != ct->pct_image)
+			if (x == 1 ||
+			    TSI(x-1, y)->pct_image != ct->pct_image)
 				is_len = snprintf(is, sizeof(is), "@%s",
 				    pmcstat_string_unintern(ct->pct_image->pi_name));
 
 			/* Check for line wrap. */
 			width += ns_len + is_len + vs_len + 1;
+			}
 			if (width >= pmcstat_displaywidth) {
 				maxy--;
 				if (y >= maxy)
@@ -487,7 +527,6 @@ pmcpl_ct_node_printtop(struct pmcpl_ct_sample *rsamples, int pmcin, int maxy)
 			PMCSTAT_PRINTW("%s%s%s ", ns, is, vs);
 			PMCSTAT_ATTROFF(v_attrs);
 		}
-		PMCSTAT_PRINTW("\n");
 	}
 }
 
@@ -498,46 +537,25 @@ pmcpl_ct_node_printtop(struct pmcpl_ct_sample *rsamples, int pmcin, int maxy)
 void
 pmcpl_ct_topdisplay(void)
 {
-	int i, x, y, pmcin;
+	int y;
 	struct pmcpl_ct_sample r, *rsamples;
 
 	rsamples = &r;
 	pmcpl_ct_samples_root(rsamples);
 
-	PMCSTAT_PRINTW("%-10.10s %s\n", "IMAGE", "CALLTREE");
+	pmcpl_ct_node_cleartag();
 
-	for (pmcin = 0; pmcin < pmcstat_npmcs; pmcin++) {
-		/* Filter PMCs. */
-		if (pmcstat_pmcinfilter != pmcin)
-			continue;
+	PMCSTAT_PRINTW("%5.5s %s\n", "%SAMP", "CALLTREE");
 
-		pmcpl_ct_node_cleartag();
+	y = 0;
+	if (pmcpl_ct_node_dumptop(pmcstat_pmcinfilter,
+	    pmcpl_ct_root, rsamples, 0, &y))
+		PMCSTAT_PRINTW("...\n");
+	pmcpl_ct_topscreen[1][y] = NULL;
 
-		/* Quicksort the arcs. */
-		qsort_r(pmcpl_ct_root->pct_arc,
-		    pmcpl_ct_root->pct_narc,
-		    sizeof(struct pmcpl_ct_arc),
-		    &pmcin, pmcpl_ct_arc_compare);
+	pmcpl_ct_node_printtop(rsamples,
+	    pmcstat_pmcinfilter, pmcstat_displayheight - 2);
 
-		x = y = 0;
-		for (i = 0; i < pmcpl_ct_root->pct_narc; i++) {
-			/* Skip this arc if there is no sample at all. */
-			if (PMCPL_CT_SAMPLE(pmcin,
-			    &pmcpl_ct_root->pct_arc[i].pcta_samples) == 0)
-				continue;
-			if (PMCPL_CT_SAMPLEP(pmcin,
-			    &pmcpl_ct_root->pct_arc[i].pcta_samples) <=
-			    pmcstat_threshold)
-				continue;
-			if (pmcpl_ct_node_dumptop(pmcin,
-			        pmcpl_ct_root->pct_arc[i].pcta_child,
-			        rsamples, x, &y, pmcstat_displayheight - 2)) {
-				break;
-			}
-		}
-
-		pmcpl_ct_node_printtop(rsamples, pmcin, y);
-	}
 	pmcpl_ct_samples_free(rsamples);
 }
 

From b88c83ff148a17242d52976e95c569222834294b Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Fri, 7 May 2010 20:46:22 +0000
Subject: [PATCH 520/532] sigprocmask(2): pthread_sigmask(3) must be used in
 threaded processes. Although libthr's pthread_sigmask() just calls
 sigprocmask() and this is unlikely to change, mention this POSIX requirement
 on applications.

MFC after:	1 week
---
 lib/libc/sys/sigprocmask.2 | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/lib/libc/sys/sigprocmask.2 b/lib/libc/sys/sigprocmask.2
index ef555a44651..f7c83353588 100644
--- a/lib/libc/sys/sigprocmask.2
+++ b/lib/libc/sys/sigprocmask.2
@@ -28,7 +28,7 @@
 .\"	@(#)sigprocmask.2	8.1 (Berkeley) 6/4/93
 .\" $FreeBSD$
 .\"
-.Dd June 4, 1993
+.Dd May 7, 2010
 .Dt SIGPROCMASK 2
 .Os
 .Sh NAME
@@ -96,6 +96,11 @@ quietly disallows
 or
 .Dv SIGSTOP
 to be blocked.
+.Pp
+In threaded applications,
+.Xr pthread_sigmask 3
+must be used instead of
+.Fn sigprocmask .
 .Sh RETURN VALUES
 .Rv -std sigprocmask
 .Sh ERRORS
@@ -116,6 +121,7 @@ has a value other than those listed here.
 .Xr sigpending 2 ,
 .Xr sigsuspend 2 ,
 .Xr fpsetmask 3 ,
+.Xr pthread_sigmask 3 ,
 .Xr sigsetops 3
 .Sh STANDARDS
 The

From af394cfa3697918cfe0bc4badddee31a637ac941 Mon Sep 17 00:00:00 2001
From: Jung-uk Kim <jkim@FreeBSD.org>
Date: Fri, 7 May 2010 21:06:52 +0000
Subject: [PATCH 521/532] Fix a typo in the previous commit.

---
 sys/vm/vm_pageout.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sys/vm/vm_pageout.c b/sys/vm/vm_pageout.c
index 0524193550e..49b6ac9dba9 100644
--- a/sys/vm/vm_pageout.c
+++ b/sys/vm/vm_pageout.c
@@ -765,7 +765,7 @@ vm_pageout_scan(int pass)
 	 */
 	page_shortage = vm_paging_target() + addl_page_shortage_init;
 
-	vm_pageout_marker_init(&marker, PQ_INACTIVE);
+	vm_pageout_init_marker(&marker, PQ_INACTIVE);
 
 	/*
 	 * Start scanning the inactive queue for pages we can move to the

From b92990ef88c60569cfd1e3ab05e5730da90813ce Mon Sep 17 00:00:00 2001
From: Fabien Thomas <fabient@FreeBSD.org>
Date: Fri, 7 May 2010 22:09:17 +0000
Subject: [PATCH 522/532] Add a fastpath to allocate from packet zone when
 using m_getjcl. This will add support for packet zone for at least igb and
 ixgbe and will avoid to check for that in bce and mxge.

MFC after: 1 week
---
 sys/dev/bce/if_bce.c   | 7 ++-----
 sys/dev/mxge/if_mxge.c | 5 +----
 sys/sys/mbuf.h         | 3 +++
 3 files changed, 6 insertions(+), 9 deletions(-)

diff --git a/sys/dev/bce/if_bce.c b/sys/dev/bce/if_bce.c
index 6608dbde2c1..ad7209e7fd5 100644
--- a/sys/dev/bce/if_bce.c
+++ b/sys/dev/bce/if_bce.c
@@ -5059,11 +5059,8 @@ bce_get_rx_buf(struct bce_softc *sc, struct mbuf *m, u16 *prod,
 #ifdef BCE_JUMBO_HDRSPLIT
 		MGETHDR(m_new, M_DONTWAIT, MT_DATA);
 #else
-		if (sc->rx_bd_mbuf_alloc_size <= MCLBYTES)
-			m_new = m_getcl(M_DONTWAIT, MT_DATA, M_PKTHDR);
-		else
-			m_new = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR,
-			    sc->rx_bd_mbuf_alloc_size);
+		m_new = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR,
+		    sc->rx_bd_mbuf_alloc_size);
 #endif
 
 		if (m_new == NULL) {
diff --git a/sys/dev/mxge/if_mxge.c b/sys/dev/mxge/if_mxge.c
index 956b1aa2dbb..e5a34bc70ed 100644
--- a/sys/dev/mxge/if_mxge.c
+++ b/sys/dev/mxge/if_mxge.c
@@ -2400,10 +2400,7 @@ mxge_get_buf_big(struct mxge_slice_state *ss, bus_dmamap_t map, int idx)
 	mxge_rx_ring_t *rx = &ss->rx_big;
 	int cnt, err, i;
 
-	if (rx->cl_size == MCLBYTES)
-		m = m_getcl(M_DONTWAIT, MT_DATA, M_PKTHDR);
-	else
-		m = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR, rx->cl_size);
+	m = m_getjcl(M_DONTWAIT, MT_DATA, M_PKTHDR, rx->cl_size);
 	if (m == NULL) {
 		rx->alloc_fail++;
 		err = ENOBUFS;
diff --git a/sys/sys/mbuf.h b/sys/sys/mbuf.h
index 158edb4fdc2..95a44a4aecf 100644
--- a/sys/sys/mbuf.h
+++ b/sys/sys/mbuf.h
@@ -523,6 +523,9 @@ m_getjcl(int how, short type, int flags, int size)
 	struct mbuf *m, *n;
 	uma_zone_t zone;
 
+	if (size == MCLBYTES)
+		return m_getcl(how, type, flags);
+
 	args.flags = flags;
 	args.type = type;
 

From 485d41c0a747631060ed7a4b21f9f81ee76aa958 Mon Sep 17 00:00:00 2001
From: Pyun YongHyeon <yongari@FreeBSD.org>
Date: Fri, 7 May 2010 23:05:27 +0000
Subject: [PATCH 523/532] Don't change PCIe maximum read request size to 2048
 on RTL810x controllers. It caused device timeouts.

Reported by:	McLone < mclone <> gmail dot com >
Tested by:	McLone < mclone <> gmail dot com >
MFC after:	5 days
---
 sys/dev/re/if_re.c | 8 +++++---
 1 file changed, 5 insertions(+), 3 deletions(-)

diff --git a/sys/dev/re/if_re.c b/sys/dev/re/if_re.c
index 80915061c7d..ef68f145af0 100644
--- a/sys/dev/re/if_re.c
+++ b/sys/dev/re/if_re.c
@@ -1162,9 +1162,11 @@ re_attach(device_t dev)
 	msic = 0;
 	if (pci_find_extcap(dev, PCIY_EXPRESS, &reg) == 0) {
 		sc->rl_flags |= RL_FLAG_PCIE;
-		/* Set PCIe maximum read request size to 2048. */
-		if (pci_get_max_read_req(dev) < 2048)
-			pci_set_max_read_req(dev, 2048);
+		if (devid != RT_DEVICEID_8101E) {
+			/* Set PCIe maximum read request size to 2048. */
+			if (pci_get_max_read_req(dev) < 2048)
+				pci_set_max_read_req(dev, 2048);
+		}
 		msic = pci_msi_count(dev);
 		if (bootverbose)
 			device_printf(dev, "MSI count : %d\n", msic);

From 23d9efa7a81f87c6d31e8d0f7dfc798bed4789d5 Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Sat, 8 May 2010 01:24:18 +0000
Subject: [PATCH 524/532] Patch the experimental NFS client so that it works
 for NFSv2 by adding the necessary mapping from NFSv3 procedure numbers to
 NFSv2 procedure numbers when doing NFSv2 RPCs.

MFC after:	1 week
---
 sys/fs/nfs/nfs_commonkrpc.c | 36 ++++++++++++++++++++++++++++++++++--
 1 file changed, 34 insertions(+), 2 deletions(-)

diff --git a/sys/fs/nfs/nfs_commonkrpc.c b/sys/fs/nfs/nfs_commonkrpc.c
index 384bdb417de..b0756e08d50 100644
--- a/sys/fs/nfs/nfs_commonkrpc.c
+++ b/sys/fs/nfs/nfs_commonkrpc.c
@@ -97,14 +97,37 @@ static void	nfs_up(struct nfsmount *, struct thread *, const char *,
     int, int);
 static int	nfs_msg(struct thread *, const char *, const char *, int);
 
-extern int nfsv2_procid[];
-
 struct nfs_cached_auth {
 	int		ca_refs; /* refcount, including 1 from the cache */
 	uid_t		ca_uid;	 /* uid that corresponds to this auth */
 	AUTH		*ca_auth; /* RPC auth handle */
 };
 
+static int nfsv2_procid[NFS_V3NPROCS] = {
+	NFSV2PROC_NULL,
+	NFSV2PROC_GETATTR,
+	NFSV2PROC_SETATTR,
+	NFSV2PROC_LOOKUP,
+	NFSV2PROC_NOOP,
+	NFSV2PROC_READLINK,
+	NFSV2PROC_READ,
+	NFSV2PROC_WRITE,
+	NFSV2PROC_CREATE,
+	NFSV2PROC_MKDIR,
+	NFSV2PROC_SYMLINK,
+	NFSV2PROC_CREATE,
+	NFSV2PROC_REMOVE,
+	NFSV2PROC_RMDIR,
+	NFSV2PROC_RENAME,
+	NFSV2PROC_LINK,
+	NFSV2PROC_READDIR,
+	NFSV2PROC_NOOP,
+	NFSV2PROC_STATFS,
+	NFSV2PROC_NOOP,
+	NFSV2PROC_NOOP,
+	NFSV2PROC_NOOP,
+};
+
 /*
  * Initialize sockets and congestion for a new NFS connection.
  * We do not free the sockaddr if error.
@@ -533,6 +556,15 @@ newnfs_request(struct nfsrv_descript *nd, struct nfsmount *nmp,
 
 	if (nmp != NULL) {
 		NFSINCRGLOBAL(newnfsstats.rpcrequests);
+
+		/* Map the procnum to the old NFSv2 one, as required. */
+		if ((nd->nd_flag & ND_NFSV2) != 0) {
+			if (nd->nd_procnum < NFS_V3NPROCS)
+				procnum = nfsv2_procid[nd->nd_procnum];
+			else
+				procnum = NFSV2PROC_NOOP;
+		}
+
 		/*
 		 * Now only used for the R_DONTRECOVER case, but until that is
 		 * supported within the krpc code, I need to keep a queue of

From 6dd8e071f09c70e861b73b95db9e155b4dc99d42 Mon Sep 17 00:00:00 2001
From: Ed Maste <emaste@FreeBSD.org>
Date: Sat, 8 May 2010 11:56:00 +0000
Subject: [PATCH 525/532] Add dummy function for ic_update_mcast (a la if_urtw)
 to avoid console spam.

---
 sys/dev/usb/wlan/if_rum.c | 9 +++++++++
 1 file changed, 9 insertions(+)

diff --git a/sys/dev/usb/wlan/if_rum.c b/sys/dev/usb/wlan/if_rum.c
index 155b76787f0..ebb22fe8257 100644
--- a/sys/dev/usb/wlan/if_rum.c
+++ b/sys/dev/usb/wlan/if_rum.c
@@ -197,6 +197,7 @@ static void		rum_enable_tsf(struct rum_softc *);
 static void		rum_update_slot(struct ifnet *);
 static void		rum_set_bssid(struct rum_softc *, const uint8_t *);
 static void		rum_set_macaddr(struct rum_softc *, const uint8_t *);
+static void		rum_update_mcast(struct ifnet *);
 static void		rum_update_promisc(struct ifnet *);
 static void		rum_setpromisc(struct rum_softc *);
 static const char	*rum_get_rf(int);
@@ -514,6 +515,7 @@ rum_attach(device_t self)
 
 	ic->ic_vap_create = rum_vap_create;
 	ic->ic_vap_delete = rum_vap_delete;
+	ic->ic_update_mcast = rum_update_mcast;
 
 	ieee80211_radiotap_attach(ic,
 	    &sc->sc_txtap.wt_ihdr, sizeof(sc->sc_txtap),
@@ -1815,6 +1817,13 @@ rum_update_promisc(struct ifnet *ifp)
 	RUM_UNLOCK(sc);
 }
 
+static void
+rum_update_mcast(struct ifnet *ifp)
+{
+
+	/* XXX do nothing? */
+}
+
 static const char *
 rum_get_rf(int rev)
 {

From 35ad337dcd4bd18eb3129fe4a63600df3dca9c4a Mon Sep 17 00:00:00 2001
From: Jilles Tjoelker <jilles@FreeBSD.org>
Date: Sat, 8 May 2010 14:00:01 +0000
Subject: [PATCH 526/532] sh: Have only one copy of _PATH_STDPATH in the
 binary.

---
 bin/sh/eval.c | 5 ++---
 1 file changed, 2 insertions(+), 3 deletions(-)

diff --git a/bin/sh/eval.c b/bin/sh/eval.c
index c4495d04275..a47d045545a 100644
--- a/bin/sh/eval.c
+++ b/bin/sh/eval.c
@@ -1089,8 +1089,7 @@ breakcmd(int argc, char **argv)
 int
 commandcmd(int argc, char **argv)
 {
-	static char stdpath[] = _PATH_STDPATH;
-	char *path;
+	const char *path;
 	int ch;
 	int cmd = -1;
 
@@ -1101,7 +1100,7 @@ commandcmd(int argc, char **argv)
 	while ((ch = getopt(argc, argv, "pvV")) != -1) {
 		switch (ch) {
 		case 'p':
-			path = stdpath;
+			path = _PATH_STDPATH;
 			break;
 		case 'v':
 			cmd = TYPECMD_SMALLV;

From c19f54267ce601c2c330ce2b9a0ad5c83b10b217 Mon Sep 17 00:00:00 2001
From: Rick Macklem <rmacklem@FreeBSD.org>
Date: Sat, 8 May 2010 14:50:12 +0000
Subject: [PATCH 527/532] Fix typos in macros.

PR:		kern/146375
Submitted by:	simon AT comsys.ntu-kpi.kiev.ua
MFC after:	1 week
---
 sys/fs/nfs/nfskpiport.h | 4 ++--
 sys/fs/nfs/nfsport.h    | 4 ++--
 2 files changed, 4 insertions(+), 4 deletions(-)

diff --git a/sys/fs/nfs/nfskpiport.h b/sys/fs/nfs/nfskpiport.h
index 1e1fb9c7a08..bd26a8e8f06 100644
--- a/sys/fs/nfs/nfskpiport.h
+++ b/sys/fs/nfs/nfskpiport.h
@@ -27,7 +27,7 @@
  */
 
 #ifndef _NFS_NFSKPIPORT_H_
-#define	_NFSKPIPORT_NFS_H_
+#define	_NFS_NFSKPIPORT_H_
 /*
  * These definitions are needed since the generic code is now using Darwin8
  * KPI stuff. (I know, seems a bit silly, but I want the code to build on
@@ -70,4 +70,4 @@ typedef struct mbuf *		mbuf_t;
 #define	uio_iov_len(p)		((p)->uio_iov->iov_len)
 #define	uio_iov_len_add(p, v)	((p)->uio_iov->iov_len += (v))
 
-#endif	/* _NFSKPIPORT_NFS_H */
+#endif	/* _NFS_NFSKPIPORT_H */
diff --git a/sys/fs/nfs/nfsport.h b/sys/fs/nfs/nfsport.h
index 0fa2e14602d..40dfe6cf1dc 100644
--- a/sys/fs/nfs/nfsport.h
+++ b/sys/fs/nfs/nfsport.h
@@ -33,7 +33,7 @@
  */
 
 #ifndef _NFS_NFSPORT_H_
-#define	_NFSPORT_NFS_H_
+#define	_NFS_NFSPORT_H_
 
 /*
  * In general, I'm not fond of #includes in .h files, but this seems
@@ -918,4 +918,4 @@ struct nfsreq {
 
 #endif	/* _KERNEL */
 
-#endif	/* _NFSPORT_NFS_H */
+#endif	/* _NFS_NFSPORT_H */

From 35e68db373802c868554cab5f49ed253cc7aad38 Mon Sep 17 00:00:00 2001
From: Tim Kientzle <kientzle@FreeBSD.org>
Date: Sat, 8 May 2010 15:43:59 +0000
Subject: [PATCH 528/532] Various manpage updates, including many long-option
 synonyms that were previously undocumented.

---
 usr.bin/tar/bsdtar.1 | 318 +++++++++++++++++++++++++++++--------------
 1 file changed, 213 insertions(+), 105 deletions(-)

diff --git a/usr.bin/tar/bsdtar.1 b/usr.bin/tar/bsdtar.1
index ac9afbb4692..69e26c781bd 100644
--- a/usr.bin/tar/bsdtar.1
+++ b/usr.bin/tar/bsdtar.1
@@ -51,7 +51,7 @@
 .Nm
 creates and manipulates streaming archive files.
 This implementation can extract from tar, pax, cpio, zip, jar, ar,
-and ISO 9660 cdrom images and can create tar, pax, cpio, ar,
+xar, and ISO 9660 cdrom images and can create tar, pax, cpio, ar, zip,
 and shar archives.
 .Pp
 The first synopsis form shows a
@@ -67,6 +67,8 @@ is a mode indicator from the following list:
 .Bl -tag -compact -width indent
 .It Fl c
 Create a new archive containing the specified items.
+The long option form is
+.Fl Fl create .
 .It Fl r
 Like
 .Fl c ,
@@ -75,8 +77,12 @@ Note that this only works on uncompressed archives stored in regular files.
 The
 .Fl f
 option is required.
+The long option form is
+.Fl Fl append .
 .It Fl t
 List archive contents to stdout.
+The long option form is
+.Fl Fl list .
 .It Fl u
 Like
 .Fl r ,
@@ -86,11 +92,15 @@ Note that this only works on uncompressed archives stored in regular files.
 The
 .Fl f
 option is required.
+The long form is
+.Fl Fl update .
 .It Fl x
 Extract to disk from the archive.
 If a file with the same name appears more than once in the archive,
 each copy will be extracted, with later copies overwriting (replacing)
 earlier copies.
+The long option form is
+.Fl Fl extract .
 .El
 .Pp
 In
@@ -127,14 +137,18 @@ In contrast,
 .Dl Nm Fl c Fl f Pa - Pa newfile Pa original.tar
 creates a new archive with only two entries.
 Similarly,
-.Dl Nm Fl czf Pa - Fl -format Cm pax Cm @ Ns Pa -
+.Dl Nm Fl czf Pa - Fl Fl format Cm pax Cm @ Ns Pa -
 reads an archive from standard input (whose format will be determined
 automatically) and converts it into a gzip-compressed
 pax-format archive on stdout.
 In this way,
 .Nm
 can be used to convert archives from one format to another.
-.It Fl b Ar blocksize
+.It Fl B , Fl Fl read-full-blocks
+Ignored for compatibility with other
+.Xr tar 1
+implementations.
+.It Fl b Ar blocksize , Fl Fl block-size Ar blocksize
 Specify the block size, in 512-byte records, for tape drive I/O.
 As a rule, this argument is only needed when reading from or writing
 to tape drives, and usually not even then as the default block size of
@@ -144,21 +158,22 @@ In c and r mode, this changes the directory before adding
 the following files.
 In x mode, change directories after opening the archive
 but before extracting entries from the archive.
-.It Fl -check-links
-(c and r modes only)
-Issue a warning message unless all links to each file are archived.
-.It Fl -chroot
+.It Fl Fl chroot
 (x mode only)
 .Fn chroot
 to the current directory after processing any
 .Fl C
 options and before extracting any files.
-.It Fl -exclude Ar pattern
+.It Fl Fl disable-copyfile
+Mac OS X specific.
+Disable the use of
+.Xr copyfile 3 .
+.It Fl Fl exclude Ar pattern
 Do not process files or directories that match the
 specified pattern.
 Note that exclusions take precedence over patterns or filenames
 specified on the command line.
-.It Fl -format Ar format
+.It Fl Fl format Ar format
 (c, r, u mode only)
 Use the specified format for the created archive.
 Supported formats include
@@ -172,16 +187,39 @@ Other formats may also be supported; see
 for more information about currently-supported formats.
 In r and u modes, when extending an existing archive, the format specified
 here must be compatible with the format of the existing archive on disk.
-.It Fl f Ar file
+.It Fl f Ar file , Fl Fl file Ar file
 Read the archive from or write the archive to the specified file.
 The filename can be
 .Pa -
 for standard input or standard output.
-If not specified, the default tape device will be used.
-(On
+The default varies by system;
+on
 .Fx ,
-the default tape device is
-.Pa /dev/sa0 . )
+the default is
+.Pa /dev/sa0 ;
+on Linux, the default is
+.Pa /dev/st0 .
+.It Fl Fl gid Ar id
+Use the provided group id number.
+On extract, this overrides the group id in the archive;
+the group name in the archive will be ignored.
+On create, this overrides the group id read from disk;
+if
+.Fl Fl gname
+is not also specified, the group name will be set to
+match the group id.
+.It Fl Fl gname Ar name
+Use the provided group name.
+On extract, this overrides the group name in the archive;
+if the provided group name does not exist on the system,
+the group id
+(from the archive or from the
+.Fl Fl gid
+option)
+will be used instead.
+On create, this sets the group name that will be stored
+in the archive;
+the name will not be verified against the system group database.
 .It Fl H
 (c and r mode only)
 Symbolic links named on the command line will be followed; the
@@ -193,25 +231,36 @@ Synonym for
 .It Fl I
 Synonym for
 .Fl T .
-.It Fl -include Ar pattern
+.It Fl Fl help
+Show usage.
+.It Fl Fl include Ar pattern
 Process only files or directories that match the specified pattern.
 Note that exclusions specified with
-.Fl -exclude
+.Fl Fl exclude
 take precedence over inclusions.
 If no inclusions are explicitly specified, all entries are processed by
 default.
 The
-.Fl -include
+.Fl Fl include
 option is especially useful when filtering archives.
 For example, the command
-.Dl Nm Fl c Fl f Pa new.tar Fl -include='*foo*' Cm @ Ns Pa old.tgz
+.Dl Nm Fl c Fl f Pa new.tar Fl Fl include='*foo*' Cm @ Ns Pa old.tgz
 creates a new archive
 .Pa new.tar
 containing only the entries from
 .Pa old.tgz
 containing the string
 .Sq foo .
-.It Fl j
+.It Fl J , Fl Fl xz
+(c mode only)
+Compress the resulting archive with
+.Xr xz 1 .
+In extract or list modes, this option is ignored.
+Note that, unlike other
+.Nm tar
+implementations, this implementation recognizes XZ compression
+automatically when reading archives.
+.It Fl j , Fl Fl bzip , Fl Fl bzip2 , Fl Fl bunzip2
 (c mode only)
 Compress the resulting archive with
 .Xr bzip2 1 .
@@ -220,69 +269,99 @@ Note that, unlike other
 .Nm tar
 implementations, this implementation recognizes bzip2 compression
 automatically when reading archives.
-.It Fl k
+.It Fl k , Fl Fl keep-old-files
 (x mode only)
 Do not overwrite existing files.
 In particular, if a file appears more than once in an archive,
 later copies will not overwrite earlier copies.
-.It Fl -keep-newer-files
+.It Fl Fl keep-newer-files
 (x mode only)
 Do not overwrite existing files that are newer than the
 versions appearing in the archive being extracted.
-.It Fl L
+.It Fl L , Fl Fl dereference
 (c and r mode only)
 All symbolic links will be followed.
 Normally, symbolic links are archived as such.
 With this option, the target of the link will be archived instead.
-.It Fl l
-This is a synonym for the
-.Fl -check-links
-option.
-.It Fl m
+.It Fl l , Fl Fl check-links
+(c and r modes only)
+Issue a warning message unless all links to each file are archived.
+.It Fl Fl lzma
+(c mode only) Compress the resulting archive with the original LZMA algorithm.
+Use of this option is discouraged and new archives should be created with
+.Fl Fl xz
+instead.
+Note that, unlike other
+.Nm tar
+implementations, this implementation recognizes LZMA compression
+automatically when reading archives.
+.It Fl m , Fl Fl modification-time
 (x mode only)
 Do not extract modification time.
 By default, the modification time is set to the time stored in the archive.
-.It Fl n
+.It Fl n , Fl Fl norecurse , Fl Fl no-recursion
 (c, r, u modes only)
 Do not recursively archive the contents of directories.
-.It Fl -newer Ar date
+.It Fl Fl newer Ar date
 (c, r, u modes only)
 Only include files and directories newer than the specified date.
 This compares ctime entries.
-.It Fl -newer-mtime Ar date
+.It Fl Fl newer-mtime Ar date
 (c, r, u modes only)
 Like
-.Fl -newer ,
+.Fl Fl newer ,
 except it compares mtime entries instead of ctime entries.
-.It Fl -newer-than Pa file
+.It Fl Fl newer-than Pa file
 (c, r, u modes only)
 Only include files and directories newer than the specified file.
 This compares ctime entries.
-.It Fl -newer-mtime-than Pa file
+.It Fl Fl newer-mtime-than Pa file
 (c, r, u modes only)
 Like
-.Fl -newer-than ,
+.Fl Fl newer-than ,
 except it compares mtime entries instead of ctime entries.
-.It Fl -nodump
+.It Fl Fl nodump
 (c and r modes only)
 Honor the nodump file flag by skipping this file.
-.It Fl -null
+.It Fl Fl null
 (use with
-.Fl I ,
-.Fl T ,
+.Fl I
 or
-.Fl X )
+.Fl T )
 Filenames or patterns are separated by null characters,
 not by newlines.
 This is often used to read filenames output by the
 .Fl print0
 option to
 .Xr find 1 .
-.It Fl -numeric-owner
+.It Fl Fl no-same-owner
 (x mode only)
-Ignore symbolic user and group names when restoring archives to disk,
-only numeric uid and gid values will be obeyed.
-.It Fl O
+Do not extract owner and group IDs.
+This is the reverse of
+.Fl Fl same-owner
+and the default behavior if
+.Nm
+is run as non-root.
+.It Fl Fl no-same-permissions
+(x mode only)
+Do not extract full permissions (SGID, SUID, sticky bit, ACLs,
+extended attributes or extended file flags).
+This is the reverse of
+.Fl p
+and the default behavior if
+.Nm
+is run as non-root.
+.It Fl Fl numeric-owner
+This is equivalent to
+.Fl Fl uname
+.Qq
+.Fl Fl gname
+.Qq .
+On extract, it causes user and group names in the archive
+to be ignored in favor of the numeric user and group ids.
+On create, it causes user and group names to not be stored
+in the archive.
+.It Fl O , Fl Fl to-stdout
 (x, t modes only)
 In extract (-x) mode, files will be written to standard out rather than
 being extracted to disk.
@@ -301,11 +380,11 @@ the archive will be discarded.
 .It Fl o
 (c, r, u mode)
 A synonym for
-.Fl -format Ar ustar
-.It Fl -one-file-system
+.Fl Fl format Ar ustar
+.It Fl Fl one-file-system
 (c, r, and u modes)
 Do not cross mount points.
-.It Fl -options Ar options
+.It Fl Fl options Ar options
 Select optional behaviors for particular modules.
 The argument is a text string containing comma-separated
 keywords and values.
@@ -376,7 +455,7 @@ Supported values are store (uncompressed) and deflate (gzip algorithm).
 .El
 If a provided option is not supported by any module, that
 is a fatal error.
-.It Fl P
+.It Fl P , Fl Fl absolute-paths
 Preserve pathnames.
 By default, absolute pathnames (those that begin with a /
 character) have the leading slash removed both when creating archives
@@ -387,21 +466,22 @@ will refuse to extract archive entries whose pathnames contain
 .Pa ..
 or whose target directory would be altered by a symlink.
 This option suppresses these behaviors.
-.It Fl p
+.It Fl p , Fl Fl insecure , Fl Fl preserve-permissions
 (x mode only)
 Preserve file permissions.
 Attempt to restore the full permissions, including owner, file modes, file
 flags and ACLs, if available, for each item extracted from the archive.
-By default, newly-created files are owned by the user running
-.Nm ,
-the file mode is restored for newly-created regular files, and
-all other types of entries receive default permissions.
-If
+This is the default, if
 .Nm
-is being run by root, the default is to restore the owner unless the
-.Fl o
-option is also specified.
-.It Fl q ( Fl -fast-read )
+is being run by root and can be overriden by also specifying
+.Fl Fl no-same-owner
+and
+.Fl Fl no-same-permissions .
+.It Fl Fl posix
+(c, r, u mode only)
+Synonym for
+.Fl Fl format Ar pax
+.It Fl q , Fl Fl fast-read
 (x and t mode only)
 Extract or list only the first archive entry that matches each pattern
 or filename operand.
@@ -416,7 +496,15 @@ Extract files as sparse files.
 For every block on disk, check first if it contains only NULL bytes and seek
 over it otherwise.
 This works similiar to the conv=sparse option of dd.
-.It Fl -strip-components Ar count
+.It Fl Fl same-owner
+(x mode only)
+Extract owner and group IDs.
+This is the reverse of
+.Fl Fl no-same-owner
+and the default behavior if
+.Nm
+is run as root.
+.It Fl Fl strip-components Ar count
 (x mode only)
 Remove the specified number of leading path elements.
 Pathnames with fewer elements will be silently skipped.
@@ -439,7 +527,7 @@ If
 is not matched, the pattern is skipped.
 Within
 .Ar new ,
-~ is substituted with the match, \1 to \9 with the content of
+~ is substituted with the match, \e1 to \e9 with the content of
 the corresponding captured group.
 The optional trailing g specifies that matching should continue
 after the matched part and stopped on the first unmatched pattern.
@@ -448,7 +536,7 @@ of symbolic links.
 The optional trailing p specifies that after a successful substitution
 the original path name and the new path name should be printed to
 standard error.
-.It Fl T Ar filename
+.It Fl T Ar filename , Fl Fl files-from Ar filename
 In x or t mode,
 .Nm
 will read the list of names to be extracted from
@@ -462,25 +550,50 @@ The special name
 on a line by itself will cause the current directory to be changed to
 the directory specified on the following line.
 Names are terminated by newlines unless
-.Fl -null
+.Fl Fl null
 is specified.
 Note that
-.Fl -null
+.Fl Fl null
 also disables the special handling of lines containing
 .Dq -C .
-.It Fl U
+.It Fl Fl totals
+(c, r, u mode only)
+After archiving all files, print a summary to stderr.
+.It Fl U , Fl Fl unlink , Fl Fl unlink-first
 (x mode only)
 Unlink files before creating them.
-Without this option,
+This can be a minor performance optimization if most files
+already exist, but can make things slower if most files
+do not already exist.
+This flag also causes
 .Nm
-overwrites existing files, which preserves existing hardlinks.
-With this option, existing hardlinks will be broken, as will any
-symlink that would affect the location of an extracted file.
-.It Fl -use-compress-program Ar program
+to remove intervening directory symlinks instead of
+reporting an error.
+See the SECURITY section below for more details.
+.It Fl Fl uid Ar id
+Use the provided user id number and ignore the user
+name from the archive.
+On create, if
+.Fl Fl uname
+is not also specified, the user name will be set to
+match the user id.
+.It Fl Fl uname Ar name
+Use the provided user name.
+On extract, this overrides the user name in the archive;
+if the provided user name does not exist on the system,
+it will be ignored and the user id
+(from the archive or from the
+.Fl Fl uid
+option)
+will be used instead.
+On create, this sets the user name that will be stored
+in the archive;
+the name is not verified against the system user database.
+.It Fl Fl use-compress-program Ar program
 Pipe the input (in x or t mode) or the output (in c mode) through
 .Pa program
 instead of using the builtin compression support.
-.It Fl v
+.It Fl v , Fl Fl verbose
 Produce verbose output.
 In create and extract modes,
 .Nm
@@ -493,18 +606,18 @@ will produce output similar to that of
 Additional
 .Fl v
 options will provide additional detail.
-.It Fl -version
+.It Fl Fl version
 Print version of
 .Nm
 and
 .Nm libarchive ,
 and exit.
-.It Fl w
+.It Fl w , Fl Fl confirmation , Fl Fl interactive
 Ask for confirmation for every action.
-.It Fl X Ar filename
+.It Fl X Ar filename , Fl Fl exclude-from Ar filename
 Read a list of exclusion patterns from the specified file.
 See
-.Fl -exclude
+.Fl Fl exclude
 for more information about the handling of exclusions.
 .It Fl y
 (c mode only)
@@ -515,16 +628,7 @@ Note that, unlike other
 .Nm tar
 implementations, this implementation recognizes bzip2 compression
 automatically when reading archives.
-.It Fl z
-(c mode only)
-Compress the resulting archive with
-.Xr gzip 1 .
-In extract or list modes, this option is ignored.
-Note that, unlike other
-.Nm tar
-implementations, this implementation recognizes gzip compression
-automatically when reading archives.
-.It Fl Z
+.It Fl Z , Fl Fl compress , Fl Fl uncompress
 (c mode only)
 Compress the resulting archive with
 .Xr compress 1 .
@@ -533,7 +637,18 @@ Note that, unlike other
 .Nm tar
 implementations, this implementation recognizes compress compression
 automatically when reading archives.
+.It Fl z , Fl Fl gunzip , Fl Fl gzip
+(c mode only)
+Compress the resulting archive with
+.Xr gzip 1 .
+In extract or list modes, this option is ignored.
+Note that, unlike other
+.Nm tar
+implementations, this implementation recognizes gzip compression
+automatically when reading archives.
 .El
+.Sh EXIT STATUS
+.Ex -std
 .Sh ENVIRONMENT
 The following environment variables affect the execution of
 .Nm :
@@ -544,27 +659,19 @@ See
 .Xr environ 7
 for more information.
 .It Ev TAPE
-The default tape device.
+The default device.
 The
 .Fl f
 option overrides this.
+Please see the description of the
+.Fl f
+option above for more details.
 .It Ev TZ
 The timezone to use when displaying dates.
 See
 .Xr environ 7
 for more information.
 .El
-.Sh FILES
-.Bl -tag -width ".Ev BLOCKSIZE"
-.It Pa /dev/sa0
-The default tape device, if not overridden by the
-.Ev TAPE
-environment variable or the
-.Fl f
-option.
-.El
-.Sh EXIT STATUS
-.Ex -std
 .Sh EXAMPLES
 The following creates a new archive
 called
@@ -627,9 +734,9 @@ permissions, or names that differ from existing data on disk:
 .Dl $ tar -cvf output.tar @input.mtree
 .Pp
 The
-.Fl -newer
+.Fl Fl newer
 and
-.Fl -newer-mtime
+.Fl Fl newer-mtime
 switches accept a variety of common date and time specifications, including
 .Dq 12 Mar 2005 7:14:29pm ,
 .Dq 2005-03-12 19:14 ,
@@ -638,7 +745,7 @@ and
 .Dq 19:14 PST May 1 .
 .Pp
 The
-.Fl -options
+.Fl Fl options
 argument can be used to control various details of archive generation
 or reading.
 For example, you can generate mtree output which only contains
@@ -646,9 +753,9 @@ For example, you can generate mtree output which only contains
 and
 .Cm uid
 keywords:
-.Dl Nm Fl cf Pa file.tar Fl -format=mtree Fl -options='!all,type,time,uid' Pa dir
+.Dl Nm Fl cf Pa file.tar Fl Fl format=mtree Fl Fl options='!all,type,time,uid' Pa dir
 or you can set the compression level used by gzip or xz compression:
-.Dl Nm Fl czf Pa file.tar Fl -options='compression-level=9' .
+.Dl Nm Fl czf Pa file.tar Fl Fl options='compression-level=9' .
 For more details, see the explanation of the
 .Fn archive_read_set_options
 and
@@ -794,6 +901,7 @@ components, or symlinks to other directories.
 .Xr mt 1 ,
 .Xr pax 1 ,
 .Xr shar 1 ,
+.Xr xz 1 ,
 .Xr libarchive 3 ,
 .Xr libarchive-formats 5 ,
 .Xr tar 5
@@ -803,7 +911,7 @@ in
 .St -p1003.1-96
 but was dropped from
 .St -p1003.1-2001 .
-The options used by this implementation were developed by surveying a
+The options supported by this implementation were developed by surveying a
 number of existing tar implementations as well as the old POSIX specification
 for tar and the current POSIX specification for pax.
 .Pp
@@ -829,6 +937,9 @@ beginning with
 This is a complete re-implementation based on the
 .Xr libarchive 3
 library.
+It was first released with
+.Fx 5.4
+in May, 2005.
 .Sh BUGS
 This program follows
 .St -p1003.1-96
@@ -838,7 +949,7 @@ option.
 Note that GNU tar prior to version 1.15 treated
 .Fl l
 as a synonym for the
-.Fl -one-file-system
+.Fl Fl one-file-system
 option.
 .Pp
 The
@@ -916,6 +1027,3 @@ Converting between dissimilar archive formats (such as tar and cpio) using the
 convention can cause hard link information to be lost.
 (This is a consequence of the incompatible ways that different archive
 formats store hardlink information.)
-.Pp
-There are alternative long options for many of the short options that
-are deliberately not documented.

From 42a9ad66974eb08d7f5114e5cd5a505f790fcb4b Mon Sep 17 00:00:00 2001
From: Ulf Lilleengen <lulf@FreeBSD.org>
Date: Sat, 8 May 2010 16:19:17 +0000
Subject: [PATCH 529/532] - Remove obsolete flags.

MFC after:	1 week
---
 sys/geom/vinum/geom_vinum_var.h | 6 ------
 1 file changed, 6 deletions(-)

diff --git a/sys/geom/vinum/geom_vinum_var.h b/sys/geom/vinum/geom_vinum_var.h
index c57fef71a8e..eac3c2c60e9 100644
--- a/sys/geom/vinum/geom_vinum_var.h
+++ b/sys/geom/vinum/geom_vinum_var.h
@@ -335,9 +335,6 @@ struct gv_plex {
 	int	flags;
 #define	GV_PLEX_ADDED		0x01	/* Added to an existing volume. */
 #define	GV_PLEX_SYNCING		0x02	/* Plex is syncing from another plex. */
-#define	GV_PLEX_THREAD_ACTIVE	0x04	/* Plex has an active RAID5 thread. */
-#define	GV_PLEX_THREAD_DIE	0x08	/* Signal the RAID5 thread to die. */
-#define	GV_PLEX_THREAD_DEAD	0x10	/* The RAID5 thread has died. */
 #define	GV_PLEX_NEWBORN		0x20	/* The plex was just created. */
 #define GV_PLEX_REBUILDING	0x40	/* The plex is rebuilding. */
 #define GV_PLEX_GROWING		0x80	/* The plex is growing. */
@@ -371,9 +368,6 @@ struct gv_volume {
 #define	GV_VOL_UP	1
 
 	int	flags;
-#define	GV_VOL_THREAD_ACTIVE	0x01	/* Volume has an active thread. */
-#define	GV_VOL_THREAD_DIE	0x02	/* Signal the thread to die. */
-#define	GV_VOL_THREAD_DEAD	0x04	/* The thread has died. */
 #define GV_VOL_NEWBORN		0x08	/* The volume was just created. */
 
 	LIST_HEAD(,gv_plex)	plexes;		/* List of attached plexes. */

From 9b9439b486d4c65b85932df7a656c9c0ec21fc19 Mon Sep 17 00:00:00 2001
From: Tim Kientzle <kientzle@FreeBSD.org>
Date: Sat, 8 May 2010 16:28:22 +0000
Subject: [PATCH 530/532] Config updates.

---
 usr.bin/tar/config_freebsd.h | 26 +++++++++++++++++++-------
 1 file changed, 19 insertions(+), 7 deletions(-)

diff --git a/usr.bin/tar/config_freebsd.h b/usr.bin/tar/config_freebsd.h
index 965fe857ba7..82097587f0e 100644
--- a/usr.bin/tar/config_freebsd.h
+++ b/usr.bin/tar/config_freebsd.h
@@ -30,6 +30,7 @@
 
 #undef	HAVE_ATTR_XATTR_H
 #define	HAVE_CHROOT 1
+#undef	HAVE_DIRECT_H
 #define	HAVE_DIRENT_D_NAMLEN 1
 #define	HAVE_DIRENT_H 1
 #define	HAVE_D_MD_ORDER 1
@@ -38,32 +39,43 @@
 #define	HAVE_FCHDIR 1
 #define	HAVE_FCNTL_H 1
 #define	HAVE_GRP_H 1
+#undef	HAVE_IO_H
 #define	HAVE_LANGINFO_H 1
+#undef	HAVE_LIBACL
 #define	HAVE_LIBARCHIVE 1
 #define	HAVE_LIMITS_H 1
 #undef	HAVE_LINUX_EXT2_FS_H
 #undef	HAVE_LINUX_FS_H
 #define	HAVE_LOCALE_H 1
-#if __FreeBSD_version >= 450002 /* nl_langinfo introduced */
+#define	HAVE_MBTOWC 1
+#undef	HAVE_NDIR_H
 #define	HAVE_NL_LANGINFO 1
-#endif
 #define	HAVE_PATHS_H 1
 #define	HAVE_PWD_H 1
 #define	HAVE_REGEX_H 1
 #define	HAVE_SETLOCALE 1
+#define	HAVE_SIGACTION 1
+#define	HAVE_SIGNAL_H 1
 #define	HAVE_STDARG_H 1
+#define	HAVE_STDINT_H 1
 #define	HAVE_STDLIB_H 1
 #define	HAVE_STRING_H 1
+#undef	HAVE_STRNCPY_S
 #define	HAVE_STRUCT_STAT_ST_FLAGS 1
-#undef	HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC
 #define	HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC 1
+#undef	HAVE_STRUCT_STAT_ST_MTIME_N
+#undef	HAVE_STRUCT_STAT_ST_MTIME_USEC
+#undef	HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC
+#undef	HAVE_STRUCT_STAT_ST_UMTIME
+#define	HAVE_SYS_CDEFS_H 1
+#define	HAVE_SYS_DIR_H 1
 #define	HAVE_SYS_IOCTL_H 1
+#undef	HAVE_SYS_NDIR_H
 #define	HAVE_SYS_PARAM_H 1
 #define	HAVE_SYS_STAT_H 1
-#define	HAVE_TIME_H 1
 #define	HAVE_SYS_TYPES_H 1
-#define	HAVE_UINTMAX_T 1
+#define	HAVE_TIME_H 1
 #define	HAVE_UNISTD_H 1
-#define	HAVE_UNSIGNED_LONG_LONG
 #define	HAVE_WCTYPE_H 1
-#undef	MAJOR_IN_MKDEV
+#undef	HAVE_WINDOWS_H
+

From 9122b590b2b45d5e3a04e85588db3456a3acfd88 Mon Sep 17 00:00:00 2001
From: Tim Kientzle <kientzle@FreeBSD.org>
Date: Sat, 8 May 2010 16:29:02 +0000
Subject: [PATCH 531/532] FreeBSD is now using bsdtar 2.8.3.

---
 usr.bin/tar/Makefile | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/usr.bin/tar/Makefile b/usr.bin/tar/Makefile
index 18289520209..3d782c5cd34 100644
--- a/usr.bin/tar/Makefile
+++ b/usr.bin/tar/Makefile
@@ -2,7 +2,7 @@
 .include <bsd.own.mk>
 
 PROG=	bsdtar
-BSDTAR_VERSION_STRING=2.7.0
+BSDTAR_VERSION_STRING=2.8.3
 SRCS=	bsdtar.c	\
 	cmdline.c	\
 	err.c		\

From b6cb6076444c1b3006d0064be4b6098b8fa92b6c Mon Sep 17 00:00:00 2001
From: Tim Kientzle <kientzle@FreeBSD.org>
Date: Sat, 8 May 2010 16:47:33 +0000
Subject: [PATCH 532/532] bsdcpio 2.8.3

---
 usr.bin/cpio/Makefile         |   4 +-
 usr.bin/cpio/bsdcpio.1        |  31 +-
 usr.bin/cpio/cmdline.c        |  42 +--
 usr.bin/cpio/config_freebsd.h |  67 +---
 usr.bin/cpio/cpio.c           | 576 +++++++++++++---------------------
 usr.bin/cpio/cpio.h           |  23 +-
 usr.bin/cpio/cpio_platform.h  |  43 +--
 usr.bin/cpio/err.c            |  17 +-
 usr.bin/cpio/err.h            |  43 +++
 usr.bin/cpio/line_reader.c    | 171 ++++++++++
 usr.bin/cpio/line_reader.h    |  37 +++
 usr.bin/cpio/matching.c       | 163 ++++++----
 usr.bin/cpio/matching.h       |  18 +-
 usr.bin/cpio/pathmatch.c      |  14 +-
 usr.bin/cpio/pathmatch.h      |   6 +-
 15 files changed, 684 insertions(+), 571 deletions(-)
 create mode 100644 usr.bin/cpio/err.h
 create mode 100644 usr.bin/cpio/line_reader.c
 create mode 100644 usr.bin/cpio/line_reader.h

diff --git a/usr.bin/cpio/Makefile b/usr.bin/cpio/Makefile
index 499b173f64d..80cf1fa55ef 100644
--- a/usr.bin/cpio/Makefile
+++ b/usr.bin/cpio/Makefile
@@ -3,8 +3,8 @@
 .include <bsd.own.mk>
 
 PROG=	bsdcpio
-BSDCPIO_VERSION_STRING=2.7.0
-SRCS=	cpio.c cmdline.c err.c matching.c pathmatch.c
+BSDCPIO_VERSION_STRING=2.8.3
+SRCS=	cpio.c cmdline.c err.c line_reader.c matching.c pathmatch.c
 CFLAGS+= -DBSDCPIO_VERSION_STRING=\"${BSDCPIO_VERSION_STRING}\"
 CFLAGS+= -DPLATFORM_CONFIG_H=\"config_freebsd.h\"
 .ifdef RELEASE_CRUNCH
diff --git a/usr.bin/cpio/bsdcpio.1 b/usr.bin/cpio/bsdcpio.1
index 6017bc5b465..79b6997ef4a 100644
--- a/usr.bin/cpio/bsdcpio.1
+++ b/usr.bin/cpio/bsdcpio.1
@@ -80,6 +80,9 @@ specified directory.
 Unless specifically stated otherwise, options are applicable in
 all operating modes.
 .Bl -tag -width indent
+.It Fl 0
+Read filenames separated by NUL characters instead of newlines.
+This is necessary if any of the filenames being read might contain newlines.
 .It Fl A
 (o mode only)
 Append to the specified archive.
@@ -142,6 +145,11 @@ for more complete information about the
 formats currently supported by the underlying
 .Xr libarchive 3
 library.
+.It Fl H Ar format
+Synonym for
+.Fl -format .
+.It Fl h , Fl -help
+Print usage information.
 .It Fl I Ar file
 Read archive from
 .Ar file .
@@ -154,6 +162,14 @@ Disable security checks during extraction or copying.
 This allows extraction via symbolic links and path names containing
 .Sq ..
 in the name.
+.It Fl J
+(o mode only)
+Compress the file with xz-compatible compression before writing it.
+In input mode, this option is ignored; xz compression is recognized
+automatically on input.
+.It Fl j
+Synonym for
+.Fl y .
 .It Fl L
 (o and p modes)
 All symbolic links will be followed.
@@ -163,6 +179,11 @@ With this option, the target of the link will be archived or copied instead.
 (p mode only)
 Create links from the target directory to the original files,
 instead of copying.
+.It Fl lzma
+(o mode only)
+Compress the file with lzma-compatible compression before writing it.
+In input mode, this option is ignored; lzma compression is recognized
+automatically on input.
 .It Fl m
 (i and p modes)
 Set file modification time on created files to match
@@ -176,6 +197,10 @@ By default,
 displays the user and group names when they are provided in the
 archive, or looks up the user and group names in the system
 password database.
+.It Fl no-preserve-owner
+(i mode only)
+Do not attempt to restore file ownership.
+This is the default when run by non-root users.
 .It Fl O Ar file
 Write archive to
 .Ar file .
@@ -185,6 +210,10 @@ See above for description.
 .It Fl p
 Pass-through mode.
 See above for description.
+.It Fl preserve-owner
+(i mode only)
+Restore file ownership.
+This is the default when run by the root user.
 .It Fl -quiet
 Suppress unnecessary messages.
 .It Fl R Oo user Oc Ns Oo : Oc Ns Oo group Oc
@@ -266,7 +295,7 @@ for more information.
 .Sh EXAMPLES
 The
 .Nm
-command is traditionally used to copy file hierarchies in conjunction
+command is traditionally used to copy file heirarchies in conjunction
 with the
 .Xr find 1
 command.
diff --git a/usr.bin/cpio/cmdline.c b/usr.bin/cpio/cmdline.c
index b6a56bed485..ca50ed2c759 100644
--- a/usr.bin/cpio/cmdline.c
+++ b/usr.bin/cpio/cmdline.c
@@ -46,11 +46,12 @@ __FBSDID("$FreeBSD$");
 #endif
 
 #include "cpio.h"
+#include "err.h"
 
 /*
  * Short options for cpio.  Please keep this sorted.
  */
-static const char *short_options = "0AaBC:F:O:cdE:f:H:hijLlmnopR:rtuvVW:yZz";
+static const char *short_options = "0AaBC:cdE:F:f:H:hI:iJjLlmnO:opR:rtuvW:yZz";
 
 /*
  * Long options for cpio.  Please keep this sorted.
@@ -61,7 +62,6 @@ static const struct option {
 	int equivalent;	/* Equivalent short option. */
 } cpio_longopts[] = {
 	{ "create",			0, 'o' },
-	{ "dot",			0, 'V' },
 	{ "extract",			0, 'i' },
 	{ "file",			1, 'F' },
 	{ "format",             	1, 'H' },
@@ -69,6 +69,7 @@ static const struct option {
 	{ "insecure",			0, OPTION_INSECURE },
 	{ "link",			0, 'l' },
 	{ "list",			0, 't' },
+	{ "lzma",			0, OPTION_LZMA },
 	{ "make-directories",		0, 'd' },
 	{ "no-preserve-owner",		0, OPTION_NO_PRESERVE_OWNER },
 	{ "null",			0, '0' },
@@ -76,10 +77,12 @@ static const struct option {
 	{ "owner",			1, 'R' },
 	{ "pass-through",		0, 'p' },
 	{ "preserve-modification-time", 0, 'm' },
+	{ "preserve-owner",		0, OPTION_PRESERVE_OWNER },
 	{ "quiet",			0, OPTION_QUIET },
 	{ "unconditional",		0, 'u' },
 	{ "verbose",			0, 'v' },
 	{ "version",			0, OPTION_VERSION },
+	{ "xz",				0, 'J' },
 	{ NULL, 0, 0 }
 };
 
@@ -172,7 +175,7 @@ cpio_getopt(struct cpio *cpio)
 				/* Otherwise, pick up the next word. */
 				opt_word = *cpio->argv;
 				if (opt_word == NULL) {
-					cpio_warnc(0,
+					warnc(0,
 					    "Option -%c requires an argument",
 					    opt);
 					return ('?');
@@ -223,13 +226,13 @@ cpio_getopt(struct cpio *cpio)
 
 		/* Fail if there wasn't a unique match. */
 		if (match == NULL) {
-			cpio_warnc(0,
+			warnc(0,
 			    "Option %s%s is not supported",
 			    long_prefix, opt_word);
 			return ('?');
 		}
 		if (match2 != NULL) {
-			cpio_warnc(0,
+			warnc(0,
 			    "Ambiguous option %s%s (matches --%s and --%s)",
 			    long_prefix, opt_word, match->name, match2->name);
 			return ('?');
@@ -241,7 +244,7 @@ cpio_getopt(struct cpio *cpio)
 			if (cpio->optarg == NULL) {
 				cpio->optarg = *cpio->argv;
 				if (cpio->optarg == NULL) {
-					cpio_warnc(0,
+					warnc(0,
 					    "Option %s%s requires an argument",
 					    long_prefix, match->name);
 					return ('?');
@@ -252,7 +255,7 @@ cpio_getopt(struct cpio *cpio)
 		} else {
 			/* Argument forbidden: fail if there is one. */
 			if (cpio->optarg != NULL) {
-				cpio_warnc(0,
+				warnc(0,
 				    "Option %s%s does not allow an argument",
 				    long_prefix, match->name);
 				return ('?');
@@ -283,17 +286,20 @@ cpio_getopt(struct cpio *cpio)
  *
  * Sets uid/gid return as appropriate, -1 indicates uid/gid not specified.
  *
+ * Returns NULL if no error, otherwise returns error string for display.
+ *
  */
-int
+const char *
 owner_parse(const char *spec, int *uid, int *gid)
 {
+	static char errbuff[128];
 	const char *u, *ue, *g;
 
 	*uid = -1;
 	*gid = -1;
 
 	if (spec[0] == '\0')
-		return (1);
+		return ("Invalid empty user/group spec");
 
 	/*
 	 * Split spec into [user][:.][group]
@@ -321,10 +327,8 @@ owner_parse(const char *spec, int *uid, int *gid)
 		struct passwd *pwent;
 
 		user = (char *)malloc(ue - u + 1);
-		if (user == NULL) {
-			cpio_warnc(errno, "Couldn't allocate memory");
-			return (1);
-		}
+		if (user == NULL)
+			return ("Couldn't allocate memory");
 		memcpy(user, u, ue - u);
 		user[ue - u] = '\0';
 		if ((pwent = getpwnam(user)) != NULL) {
@@ -336,9 +340,10 @@ owner_parse(const char *spec, int *uid, int *gid)
 			errno = 0;
 			*uid = strtoul(user, &end, 10);
 			if (errno || *end != '\0') {
-				cpio_warnc(errno,
+				snprintf(errbuff, sizeof(errbuff),
 				    "Couldn't lookup user ``%s''", user);
-				return (1);
+				errbuff[sizeof(errbuff) - 1] = '\0';
+				return (errbuff);
 			}
 		}
 		free(user);
@@ -353,11 +358,12 @@ owner_parse(const char *spec, int *uid, int *gid)
 			errno = 0;
 			*gid = strtoul(g, &end, 10);
 			if (errno || *end != '\0') {
-				cpio_warnc(errno,
+				snprintf(errbuff, sizeof(errbuff),
 				    "Couldn't lookup group ``%s''", g);
-				return (1);
+				errbuff[sizeof(errbuff) - 1] = '\0';
+				return (errbuff);
 			}
 		}
 	}
-	return (0);
+	return (NULL);
 }
diff --git a/usr.bin/cpio/config_freebsd.h b/usr.bin/cpio/config_freebsd.h
index df376ff2d30..ec4e4416e37 100644
--- a/usr.bin/cpio/config_freebsd.h
+++ b/usr.bin/cpio/config_freebsd.h
@@ -25,83 +25,32 @@
  * $FreeBSD$
  */
 
-/* A default configuration for FreeBSD, used if there is no config.h. */
+/* A hand-tooled configuration for FreeBSD. */
 
 #include <sys/param.h>  /* __FreeBSD_version */
 
-#if __FreeBSD__ > 4
-#define	HAVE_ACL_GET_PERM 0
-#define	HAVE_ACL_GET_PERM_NP 1
-#define	HAVE_ACL_PERMSET_T 1
-#define	HAVE_ACL_USER 1
-#endif
-#undef	HAVE_ATTR_XATTR_H
-#define	HAVE_BZLIB_H 1
-#define	HAVE_CHFLAGS 1
-#define	HAVE_DECL_OPTARG 1
-#define	HAVE_DECL_OPTIND 1
-#define	HAVE_DIRENT_D_NAMLEN 1
 #define	HAVE_DIRENT_H 1
-#define	HAVE_D_MD_ORDER 1
 #define	HAVE_ERRNO_H 1
-#undef	HAVE_EXT2FS_EXT2_FS_H
-#define	HAVE_FCHDIR 1
 #define	HAVE_FCNTL_H 1
-#define	HAVE_FNMATCH 1
-#define	HAVE_FNMATCH_H 1
-#define	HAVE_FNM_LEADING_DIR 1
-#define	HAVE_FTRUNCATE 1
 #define	HAVE_FUTIMES 1
-#undef	HAVE_GETXATTR
 #define	HAVE_GRP_H 1
-#define	HAVE_INTTYPES_H 1
-#define	HAVE_LANGINFO_H 1
-#undef	HAVE_LGETXATTR
-#undef	HAVE_LIBACL
 #define	HAVE_LIBARCHIVE 1
-#define	HAVE_LIBBZ2 1
-#define	HAVE_LIBZ 1
-#define	HAVE_LIMITS_H 1
-#undef	HAVE_LINUX_EXT2_FS_H
-#undef	HAVE_LINUX_FS_H
-#undef	HAVE_LISTXATTR
-#undef	HAVE_LLISTXATTR
-#define	HAVE_LOCALE_H 1
+#define	HAVE_LINK 1
+#define	HAVE_LSTAT 1
 #define	HAVE_LUTIMES 1
-#define	HAVE_MALLOC 1
-#define	HAVE_MEMMOVE 1
-#define	HAVE_MEMORY_H 1
-#define	HAVE_MEMSET 1
-#if __FreeBSD_version >= 450002 /* nl_langinfo introduced */
-#define	HAVE_NL_LANGINFO 1
-#endif
-#define	HAVE_PATHS_H 1
 #define	HAVE_PWD_H 1
-#define	HAVE_SETLOCALE 1
+#define	HAVE_READLINK 1
 #define	HAVE_STDARG_H 1
-#define	HAVE_STDINT_H 1
 #define	HAVE_STDLIB_H 1
-#define	HAVE_STRCHR 1
-#define	HAVE_STRDUP 1
-#define	HAVE_STRERROR 1
-#define	HAVE_STRFTIME 1
-#define	HAVE_STRINGS_H 1
 #define	HAVE_STRING_H 1
-#define	HAVE_STRRCHR 1
-#undef	HAVE_STRUCT_STAT_ST_MTIM_TV_NSEC
-#define	HAVE_STRUCT_STAT_ST_MTIMESPEC_TV_NSEC 1
-#define	HAVE_SYS_ACL_H 1
-#define	HAVE_SYS_IOCTL_H 1
-#define	HAVE_SYS_PARAM_H 1
+#define	HAVE_SYMLINK 1
+#define	HAVE_SYS_CDEFS_H 1
 #define	HAVE_SYS_STAT_H 1
+#define	HAVE_SYS_TIME_H 1
 #define	HAVE_TIME_H 1
-#define	HAVE_SYS_TYPES_H 1
 #define	HAVE_UINTMAX_T 1
 #define	HAVE_UNISTD_H 1
 #define	HAVE_UNSIGNED_LONG_LONG 1
+#define	HAVE_UTIME_H 1
 #define	HAVE_UTIMES 1
-#define	HAVE_VPRINTF 1
-#define	HAVE_ZLIB_H 1
-#undef	MAJOR_IN_MKDEV
-#define	STDC_HEADERS 1
 
diff --git a/usr.bin/cpio/cpio.c b/usr.bin/cpio/cpio.c
index 53195aa95af..9b76fe8c9fa 100644
--- a/usr.bin/cpio/cpio.c
+++ b/usr.bin/cpio/cpio.c
@@ -56,6 +56,9 @@ __FBSDID("$FreeBSD$");
 #ifdef HAVE_STDARG_H
 #include <stdarg.h>
 #endif
+#ifdef HAVE_STDINT_H
+#include <stdint.h>
+#endif
 #include <stdio.h>
 #ifdef HAVE_STDLIB_H
 #include <stdlib.h>
@@ -74,11 +77,17 @@ __FBSDID("$FreeBSD$");
 #endif
 
 #include "cpio.h"
+#include "err.h"
+#include "line_reader.h"
 #include "matching.h"
 
 /* Fixed size of uname/gname caches. */
 #define	name_cache_size 101
 
+#ifndef O_BINARY
+#define O_BINARY 0
+#endif
+
 struct name_cache {
 	int	probes;
 	int	hits;
@@ -89,7 +98,8 @@ struct name_cache {
 	} cache[name_cache_size];
 };
 
-static int	copy_data(struct archive *, struct archive *);
+static int	extract_data(struct archive *, struct archive *);
+const char *	cpio_i64toa(int64_t);
 static const char *cpio_rename(const char *name);
 static int	entry_to_archive(struct cpio *, struct archive_entry *);
 static int	file_to_archive(struct cpio *, const char *);
@@ -117,6 +127,7 @@ main(int argc, char *argv[])
 	static char buff[16384];
 	struct cpio _cpio; /* Allocated on stack. */
 	struct cpio *cpio;
+	const char *errmsg;
 	int uid, gid;
 	int opt;
 
@@ -124,33 +135,26 @@ main(int argc, char *argv[])
 	memset(cpio, 0, sizeof(*cpio));
 	cpio->buff = buff;
 	cpio->buff_size = sizeof(buff);
-#if defined(_WIN32) && !defined(__CYGWIN__)
-	/* Make sure open() function will be used with a binary mode. */
-	/* on cygwin, we need something similar, but instead link against */
-	/* a special startup object, binmode.o */
-	_set_fmode(_O_BINARY);
-#endif
 
-	/* Need cpio_progname before calling cpio_warnc. */
+	/* Need progname before calling warnc. */
 	if (*argv == NULL)
-		cpio_progname = "bsdcpio";
+		progname = "bsdcpio";
 	else {
 #if defined(_WIN32) && !defined(__CYGWIN__)
-		cpio_progname = strrchr(*argv, '\\');
+		progname = strrchr(*argv, '\\');
 #else
-		cpio_progname = strrchr(*argv, '/');
+		progname = strrchr(*argv, '/');
 #endif
-		if (cpio_progname != NULL)
-			cpio_progname++;
+		if (progname != NULL)
+			progname++;
 		else
-			cpio_progname = *argv;
+			progname = *argv;
 	}
 
 	cpio->uid_override = -1;
 	cpio->gid_override = -1;
 	cpio->argv = argv;
 	cpio->argc = argc;
-	cpio->line_separator = '\n';
 	cpio->mode = '\0';
 	cpio->verbose = 0;
 	cpio->compress = '\0';
@@ -161,19 +165,17 @@ main(int argc, char *argv[])
 	cpio->extract_flags |= ARCHIVE_EXTRACT_PERM;
 	cpio->extract_flags |= ARCHIVE_EXTRACT_FFLAGS;
 	cpio->extract_flags |= ARCHIVE_EXTRACT_ACL;
-#if defined(_WIN32) || defined(__CYGWIN__)
-	if (bsdcpio_is_privileged())
-#else
+#if !defined(_WIN32) && !defined(__CYGWIN__)
 	if (geteuid() == 0)
-#endif
 		cpio->extract_flags |= ARCHIVE_EXTRACT_OWNER;
+#endif
 	cpio->bytes_per_block = 512;
 	cpio->filename = NULL;
 
 	while ((opt = cpio_getopt(cpio)) != -1) {
 		switch (opt) {
 		case '0': /* GNU convention: --null, -0 */
-			cpio->line_separator = '\0';
+			cpio->option_null = 1;
 			break;
 		case 'A': /* NetBSD/OpenBSD */
 			cpio->option_append = 1;
@@ -187,7 +189,7 @@ main(int argc, char *argv[])
 		case 'C': /* NetBSD/OpenBSD */
 			cpio->bytes_per_block = atoi(cpio->optarg);
 			if (cpio->bytes_per_block <= 0)
-				cpio_errc(1, 0, "Invalid blocksize %s", cpio->optarg);
+				errc(1, 0, "Invalid blocksize %s", cpio->optarg);
 			break;
 		case 'c': /* POSIX 1997 */
 			cpio->format = "odc";
@@ -196,13 +198,14 @@ main(int argc, char *argv[])
 			cpio->extract_flags &= ~ARCHIVE_EXTRACT_NO_AUTODIR;
 			break;
 		case 'E': /* NetBSD/OpenBSD */
-			include_from_file(cpio, cpio->optarg);
+			include_from_file(&cpio->matching,
+			    cpio->optarg, cpio->option_null);
 			break;
 		case 'F': /* NetBSD/OpenBSD/GNU cpio */
 			cpio->filename = cpio->optarg;
 			break;
 		case 'f': /* POSIX 1997 */
-			exclude(cpio, cpio->optarg);
+			exclude(&cpio->matching, cpio->optarg);
 			break;
 		case 'H': /* GNU cpio (also --format) */
 			cpio->format = cpio->optarg;
@@ -215,10 +218,16 @@ main(int argc, char *argv[])
 			break;
 		case 'i': /* POSIX 1997 */
 			if (cpio->mode != '\0')
-				cpio_errc(1, 0,
+				errc(1, 0,
 				    "Cannot use both -i and -%c", cpio->mode);
 			cpio->mode = opt;
 			break;
+		case 'J': /* GNU tar, others */
+			cpio->compress = opt;
+			break;
+		case 'j': /* GNU tar, others */
+			cpio->compress = opt;
+			break;
 		case OPTION_INSECURE:
 			cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_SYMLINKS;
 			cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NODOTDOT;
@@ -229,6 +238,9 @@ main(int argc, char *argv[])
 		case 'l': /* POSIX 1997 */
 			cpio->option_link = 1;
 			break;
+		case OPTION_LZMA: /* GNU tar, others */
+			cpio->compress = opt;
+			break;
 		case 'm': /* POSIX 1997 */
 			cpio->extract_flags |= ARCHIVE_EXTRACT_TIME;
 			break;
@@ -243,23 +255,29 @@ main(int argc, char *argv[])
 			break;
 		case 'o': /* POSIX 1997 */
 			if (cpio->mode != '\0')
-				cpio_errc(1, 0,
+				errc(1, 0,
 				    "Cannot use both -o and -%c", cpio->mode);
 			cpio->mode = opt;
 			break;
 		case 'p': /* POSIX 1997 */
 			if (cpio->mode != '\0')
-				cpio_errc(1, 0,
+				errc(1, 0,
 				    "Cannot use both -p and -%c", cpio->mode);
 			cpio->mode = opt;
 			cpio->extract_flags &= ~ARCHIVE_EXTRACT_SECURE_NODOTDOT;
 			break;
+		case OPTION_PRESERVE_OWNER:
+			cpio->extract_flags |= ARCHIVE_EXTRACT_OWNER;
+			break;
 		case OPTION_QUIET: /* GNU cpio */
 			cpio->quiet = 1;
 			break;
 		case 'R': /* GNU cpio, also --owner */
-			if (owner_parse(cpio->optarg, &uid, &gid))
+			errmsg = owner_parse(cpio->optarg, &uid, &gid);
+			if (errmsg) {
+				warnc(-1, "%s", errmsg);
 				usage();
+			}
 			if (uid != -1)
 				cpio->uid_override = uid;
 			if (gid != -1)
@@ -278,9 +296,6 @@ main(int argc, char *argv[])
 		case 'v': /* POSIX 1997 */
 			cpio->verbose++;
 			break;
-		case 'V': /* GNU cpio */
-			cpio->dot++;
-			break;
 		case OPTION_VERSION: /* GNU convention */
 			version();
 			break;
@@ -293,23 +308,13 @@ main(int argc, char *argv[])
 			break;
 #endif
 		case 'y': /* tar convention */
-#if HAVE_LIBBZ2
 			cpio->compress = opt;
-#else
-			cpio_warnc(0, "bzip2 compression not supported by "
-			    "this version of bsdcpio");
-#endif
 			break;
 		case 'Z': /* tar convention */
 			cpio->compress = opt;
 			break;
 		case 'z': /* tar convention */
-#if HAVE_LIBZ
 			cpio->compress = opt;
-#else
-			cpio_warnc(0, "gzip compression not supported by "
-			    "this version of bsdcpio");
-#endif
 			break;
 		default:
 			usage();
@@ -324,19 +329,16 @@ main(int argc, char *argv[])
 		cpio->mode = 'i';
 	/* -t requires -i */
 	if (cpio->option_list && cpio->mode != 'i')
-		cpio_errc(1, 0, "Option -t requires -i", cpio->mode);
+		errc(1, 0, "Option -t requires -i");
 	/* -n requires -it */
 	if (cpio->option_numeric_uid_gid && !cpio->option_list)
-		cpio_errc(1, 0, "Option -n requires -it");
+		errc(1, 0, "Option -n requires -it");
 	/* Can only specify format when writing */
 	if (cpio->format != NULL && cpio->mode != 'o')
-		cpio_errc(1, 0, "Option --format requires -o");
+		errc(1, 0, "Option --format requires -o");
 	/* -l requires -p */
 	if (cpio->option_link && cpio->mode != 'p')
-		cpio_errc(1, 0, "Option -l requires -p");
-	/* -v overrides -V */
-	if (cpio->dot && cpio->verbose)
-		cpio->dot = 0;
+		errc(1, 0, "Option -l requires -p");
 	/* TODO: Flag other nonsensical combinations. */
 
 	switch (cpio->mode) {
@@ -350,7 +352,7 @@ main(int argc, char *argv[])
 		break;
 	case 'i':
 		while (*cpio->argv != NULL) {
-			include(cpio, *cpio->argv);
+			include(&cpio->matching, *cpio->argv);
 			--cpio->argc;
 			++cpio->argv;
 		}
@@ -361,26 +363,26 @@ main(int argc, char *argv[])
 		break;
 	case 'p':
 		if (*cpio->argv == NULL || **cpio->argv == '\0')
-			cpio_errc(1, 0,
+			errc(1, 0,
 			    "-p mode requires a target directory");
 		mode_pass(cpio, *cpio->argv);
 		break;
 	default:
-		cpio_errc(1, 0,
+		errc(1, 0,
 		    "Must specify at least one of -i, -o, or -p");
 	}
 
 	free_cache(cpio->gname_cache);
 	free_cache(cpio->uname_cache);
-	return (0);
+	return (cpio->return_value);
 }
 
-void
+static void
 usage(void)
 {
 	const char	*p;
 
-	p = cpio_progname;
+	p = progname;
 
 	fprintf(stderr, "Brief Usage:\n");
 	fprintf(stderr, "  List:    %s -it < archive\n", p);
@@ -394,14 +396,9 @@ static const char *long_help_msg =
 	"First option must be a mode specifier:\n"
 	"  -i Input  -o Output  -p Pass\n"
 	"Common Options:\n"
-	"  -v  Verbose filenames    -V  one dot per file\n"
+	"  -v    Verbose\n"
 	"Create: %p -o [options]  < [list of files] > [archive]\n"
-#ifdef HAVE_BZLIB_H
-	"  -y  Compress archive with bzip2\n"
-#endif
-#ifdef HAVE_ZLIB_H
-	"  -z  Compress archive with gzip\n"
-#endif
+	"  -J,-y,-z,--lzma  Compress archive with xz/bzip2/gzip/lzma\n"
 	"  --format {odc|newc|ustar}  Select archive format\n"
 	"List: %p -it < [archive]\n"
 	"Extract: %p -i [options] < [archive]\n";
@@ -423,7 +420,7 @@ long_help(void)
 	const char	*prog;
 	const char	*p;
 
-	prog = cpio_progname;
+	prog = progname;
 
 	fflush(stderr);
 
@@ -455,19 +452,33 @@ version(void)
 static void
 mode_out(struct cpio *cpio)
 {
-	unsigned long blocks;
 	struct archive_entry *entry, *spare;
 	struct line_reader *lr;
 	const char *p;
 	int r;
 
 	if (cpio->option_append)
-		cpio_errc(1, 0, "Append mode not yet supported.");
+		errc(1, 0, "Append mode not yet supported.");
+
+	cpio->archive_read_disk = archive_read_disk_new();
+	if (cpio->archive_read_disk == NULL)
+		errc(1, 0, "Failed to allocate archive object");
+	if (cpio->option_follow_links)
+		archive_read_disk_set_symlink_logical(cpio->archive_read_disk);
+	else
+		archive_read_disk_set_symlink_physical(cpio->archive_read_disk);
+	archive_read_disk_set_standard_lookup(cpio->archive_read_disk);
+
 	cpio->archive = archive_write_new();
 	if (cpio->archive == NULL)
-		cpio_errc(1, 0, "Failed to allocate archive object");
+		errc(1, 0, "Failed to allocate archive object");
 	switch (cpio->compress) {
-#ifndef SMALLER
+	case 'J':
+		r = archive_write_set_compression_xz(cpio->archive);
+		break;
+	case OPTION_LZMA:
+		r = archive_write_set_compression_lzma(cpio->archive);
+		break;
 	case 'j': case 'y':
 		r = archive_write_set_compression_bzip2(cpio->archive);
 		break;
@@ -477,41 +488,30 @@ mode_out(struct cpio *cpio)
 	case 'Z':
 		r = archive_write_set_compression_compress(cpio->archive);
 		break;
-#endif
-	case '\0':
+	default:
 		r = archive_write_set_compression_none(cpio->archive);
 		break;
-	default:
-		cpio_errc(1, 0, "Unrecognized compression option");
 	}
-	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, "Unsupported compression format");
-#ifdef SMALLER
-	if (strcmp(cpio->format, "cpio"))
-		r = archive_write_set_format_cpio(cpio->archive);
-	else if (strcmp(cpio->format, "odc"))
-		r = archive_write_set_format_cpio(cpio->archive);
-	else if (strcmp(cpio->format, "newc"))
-		r = archive_write_set_format_cpio(cpio->archive);
-	else if (strcmp(cpio->format, "ustar"))
-		r = archive_write_set_format_cpio(cpio->archive);
-#else
+	if (r < ARCHIVE_WARN)
+		errc(1, 0, "Requested compression not available");
 	r = archive_write_set_format_by_name(cpio->archive, cpio->format);
-#endif
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(cpio->archive));
+		errc(1, 0, "%s", archive_error_string(cpio->archive));
 	archive_write_set_bytes_per_block(cpio->archive, cpio->bytes_per_block);
 	cpio->linkresolver = archive_entry_linkresolver_new();
 	archive_entry_linkresolver_set_strategy(cpio->linkresolver,
 	    archive_format(cpio->archive));
 
+	/*
+	 * The main loop:  Copy each file into the output archive.
+	 */
 	r = archive_write_open_file(cpio->archive, cpio->filename);
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(cpio->archive));
-	lr = process_lines_init("-", cpio->line_separator);
-	while ((p = process_lines_next(lr)) != NULL)
+		errc(1, 0, "%s", archive_error_string(cpio->archive));
+	lr = line_reader("-", cpio->option_null);
+	while ((p = line_reader_next(lr)) != NULL)
 		file_to_archive(cpio, p);
-	process_lines_free(lr);
+	line_reader_free(lr);
 
 	/*
 	 * The hardlink detection may have queued up a couple of entries
@@ -527,15 +527,14 @@ mode_out(struct cpio *cpio)
 	}
 
 	r = archive_write_close(cpio->archive);
-	if (cpio->dot)
-		fprintf(stderr, "\n");
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(cpio->archive));
+		errc(1, 0, "%s", archive_error_string(cpio->archive));
 
 	if (!cpio->quiet) {
-		blocks = (archive_position_uncompressed(cpio->archive) + 511)
-			      / 512;
-		fprintf(stderr, "%lu %s\n", blocks,
+		int64_t blocks =
+			(archive_position_uncompressed(cpio->archive) + 511)
+			/ 512;
+		fprintf(stderr, "%lu %s\n", (unsigned long)blocks,
 		    blocks == 1 ? "block" : "blocks");
 	}
 	archive_write_finish(cpio->archive);
@@ -549,57 +548,37 @@ mode_out(struct cpio *cpio)
 static int
 file_to_archive(struct cpio *cpio, const char *srcpath)
 {
-	struct stat st;
 	const char *destpath;
 	struct archive_entry *entry, *spare;
 	size_t len;
 	const char *p;
-#if !defined(_WIN32) || defined(__CYGWIN__)
-	int lnklen;
-#endif
 	int r;
 
 	/*
 	 * Create an archive_entry describing the source file.
 	 *
-	 * XXX TODO: rework to use archive_read_disk_entry_from_file()
 	 */
 	entry = archive_entry_new();
 	if (entry == NULL)
-		cpio_errc(1, 0, "Couldn't allocate entry");
+		errc(1, 0, "Couldn't allocate entry");
 	archive_entry_copy_sourcepath(entry, srcpath);
-
-	/* Get stat information. */
-	if (cpio->option_follow_links)
-		r = stat(srcpath, &st);
-	else
-		r = lstat(srcpath, &st);
-	if (r != 0) {
-		cpio_warnc(errno, "Couldn't stat \"%s\"", srcpath);
-		archive_entry_free(entry);
-		return (0);
+	r = archive_read_disk_entry_from_file(cpio->archive_read_disk,
+	    entry, -1, NULL);
+	if (r < ARCHIVE_FAILED)
+		errc(1, 0, "%s",
+		    archive_error_string(cpio->archive_read_disk));
+	if (r < ARCHIVE_OK)
+		warnc(0, "%s",
+		    archive_error_string(cpio->archive_read_disk));
+	if (r <= ARCHIVE_FAILED) {
+		cpio->return_value = 1;
+		return (r);
 	}
 
 	if (cpio->uid_override >= 0)
-		st.st_uid = cpio->uid_override;
+		archive_entry_set_uid(entry, cpio->uid_override);
 	if (cpio->gid_override >= 0)
-		st.st_gid = cpio->gid_override;
-	archive_entry_copy_stat(entry, &st);
-
-#if !defined(_WIN32) || defined(__CYGWIN__)
-	/* If its a symlink, pull the target. */
-	if (S_ISLNK(st.st_mode)) {
-		lnklen = readlink(srcpath, cpio->buff, cpio->buff_size);
-		if (lnklen < 0) {
-			cpio_warnc(errno,
-			    "%s: Couldn't read symbolic link", srcpath);
-			archive_entry_free(entry);
-			return (0);
-		}
-		cpio->buff[lnklen] = 0;
-		archive_entry_set_symlink(entry, cpio->buff);
-	}
-#endif
+		archive_entry_set_gid(entry, cpio->gid_override);
 
 	/*
 	 * Generate a destination path for this entry.
@@ -618,7 +597,7 @@ file_to_archive(struct cpio *cpio, const char *srcpath)
 			free(cpio->pass_destpath);
 			cpio->pass_destpath = malloc(cpio->pass_destpath_alloc);
 			if (cpio->pass_destpath == NULL)
-				cpio_errc(1, ENOMEM,
+				errc(1, ENOMEM,
 				    "Can't allocate path buffer");
 		}
 		strcpy(cpio->pass_destpath, cpio->destdir);
@@ -639,18 +618,18 @@ file_to_archive(struct cpio *cpio, const char *srcpath)
 	 */
 	spare = NULL;
 	if (cpio->linkresolver != NULL
-	    && !S_ISDIR(st.st_mode)) {
+	    && archive_entry_filetype(entry) != AE_IFDIR) {
 		archive_entry_linkify(cpio->linkresolver, &entry, &spare);
 	}
 
 	if (entry != NULL) {
 		r = entry_to_archive(cpio, entry);
 		archive_entry_free(entry);
-	}
-	if (spare != NULL) {
-		if (r == 0)
-			r = entry_to_archive(cpio, spare);
-		archive_entry_free(spare);
+		if (spare != NULL) {
+			if (r == 0)
+				r = entry_to_archive(cpio, spare);
+			archive_entry_free(spare);
+		}
 	}
 	return (r);
 }
@@ -667,8 +646,6 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 	/* Print out the destination name to the user. */
 	if (cpio->verbose)
 		fprintf(stderr,"%s", destpath);
-	if (cpio->dot)
-		fprintf(stderr, ".");
 
 	/*
 	 * Option_link only makes sense in pass mode and for
@@ -686,7 +663,7 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 		/* Save the original entry in case we need it later. */
 		t = archive_entry_clone(entry);
 		if (t == NULL)
-			cpio_errc(1, ENOMEM, "Can't create link");
+			errc(1, ENOMEM, "Can't create link");
 		/* Note: link(2) doesn't create parent directories,
 		 * so we use archive_write_header() instead as a
 		 * convenience. */
@@ -696,15 +673,15 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 		r = archive_write_header(cpio->archive, t);
 		archive_entry_free(t);
 		if (r != ARCHIVE_OK)
-			cpio_warnc(archive_errno(cpio->archive),
-			    archive_error_string(cpio->archive));
+			warnc(archive_errno(cpio->archive),
+			    "%s", archive_error_string(cpio->archive));
 		if (r == ARCHIVE_FATAL)
 			exit(1);
 #ifdef EXDEV
 		if (r != ARCHIVE_OK && archive_errno(cpio->archive) == EXDEV) {
 			/* Cross-device link:  Just fall through and use
 			 * the original entry to copy the file over. */
-			cpio_warnc(0, "Copying file instead");
+			warnc(0, "Copying file instead");
 		} else
 #endif
 		return (0);
@@ -716,9 +693,9 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 	 */
 	if (archive_entry_filetype(entry) == AE_IFREG) {
 		if (archive_entry_size(entry) > 0) {
-			fd = open(srcpath, O_RDONLY);
+			fd = open(srcpath, O_RDONLY | O_BINARY);
 			if (fd < 0) {
-				cpio_warnc(errno,
+				warnc(errno,
 				    "%s: could not open file", srcpath);
 				goto cleanup;
 			}
@@ -730,7 +707,7 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 	r = archive_write_header(cpio->archive, entry);
 
 	if (r != ARCHIVE_OK)
-		cpio_warnc(archive_errno(cpio->archive),
+		warnc(archive_errno(cpio->archive),
 		    "%s: %s",
 		    srcpath,
 		    archive_error_string(cpio->archive));
@@ -744,10 +721,10 @@ entry_to_archive(struct cpio *cpio, struct archive_entry *entry)
 			r = archive_write_data(cpio->archive,
 			    cpio->buff, bytes_read);
 			if (r < 0)
-				cpio_errc(1, archive_errno(cpio->archive),
-				    archive_error_string(cpio->archive));
+				errc(1, archive_errno(cpio->archive),
+				    "%s", archive_error_string(cpio->archive));
 			if (r < bytes_read) {
-				cpio_warnc(0,
+				warnc(0,
 				    "Truncated write; file may have grown while being archived.");
 			}
 			bytes_read = read(fd, cpio->buff, cpio->buff_size);
@@ -776,7 +753,7 @@ restore_time(struct cpio *cpio, struct archive_entry *entry,
 	(void)name; /* UNUSED */
 
 	if (!warned)
-		cpio_warnc(0, "Can't restore access times on this platform");
+		warnc(0, "Can't restore access times on this platform");
 	warned = 1;
 	return (fd);
 #else
@@ -795,7 +772,7 @@ restore_time(struct cpio *cpio, struct archive_entry *entry,
         times[0].tv_sec = archive_entry_atime(entry);
         times[0].tv_usec = archive_entry_atime_nsec(entry) / 1000;
 
-#ifdef HAVE_FUTIMES
+#if defined(HAVE_FUTIMES) && !defined(__CYGWIN__)
         if (fd >= 0 && futimes(fd, times) == 0)
 		return (fd);
 #endif
@@ -811,9 +788,10 @@ restore_time(struct cpio *cpio, struct archive_entry *entry,
 #ifdef HAVE_LUTIMES
         if (lutimes(name, times) != 0)
 #else
-        if (!S_ISLNK(archive_entry_mode(entry)) && utimes(name, times) != 0)
+        if ((AE_IFLNK != archive_entry_filetype(entry))
+			&& utimes(name, times) != 0)
 #endif
-                cpio_warnc(errno, "Can't update time for %s", name);
+                warnc(errno, "Can't update time for %s", name);
 #endif
 	return (fd);
 }
@@ -826,38 +804,32 @@ mode_in(struct cpio *cpio)
 	struct archive_entry *entry;
 	struct archive *ext;
 	const char *destpath;
-	unsigned long blocks;
 	int r;
 
 	ext = archive_write_disk_new();
 	if (ext == NULL)
-		cpio_errc(1, 0, "Couldn't allocate restore object");
+		errc(1, 0, "Couldn't allocate restore object");
 	r = archive_write_disk_set_options(ext, cpio->extract_flags);
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(ext));
+		errc(1, 0, "%s", archive_error_string(ext));
 	a = archive_read_new();
 	if (a == NULL)
-		cpio_errc(1, 0, "Couldn't allocate archive object");
-#ifdef SMALLER
-	archive_read_support_format_cpio(a);
-	archive_read_support_format_tar(a);
-#else
+		errc(1, 0, "Couldn't allocate archive object");
 	archive_read_support_compression_all(a);
 	archive_read_support_format_all(a);
-#endif
 
 	if (archive_read_open_file(a, cpio->filename, cpio->bytes_per_block))
-		cpio_errc(1, archive_errno(a),
-		    archive_error_string(a));
+		errc(1, archive_errno(a),
+		    "%s", archive_error_string(a));
 	for (;;) {
 		r = archive_read_next_header(a, &entry);
 		if (r == ARCHIVE_EOF)
 			break;
 		if (r != ARCHIVE_OK) {
-			cpio_errc(1, archive_errno(a),
-			    archive_error_string(a));
+			errc(1, archive_errno(a),
+			    "%s", archive_error_string(a));
 		}
-		if (excluded(cpio, archive_entry_pathname(entry)))
+		if (excluded(cpio->matching, archive_entry_pathname(entry)))
 			continue;
 		if (cpio->option_rename) {
 			destpath = cpio_rename(archive_entry_pathname(entry));
@@ -867,9 +839,7 @@ mode_in(struct cpio *cpio)
 		if (destpath == NULL)
 			continue;
 		if (cpio->verbose)
-			fprintf(stderr, "%s\n", destpath);
-		if (cpio->dot)
-			fprintf(stderr, ".");
+			fprintf(stdout, "%s\n", destpath);
 		if (cpio->uid_override >= 0)
 			archive_entry_set_uid(entry, cpio->uid_override);
 		if (cpio->gid_override >= 0)
@@ -880,30 +850,34 @@ mode_in(struct cpio *cpio)
 			    archive_entry_pathname(entry),
 			    archive_error_string(ext));
 		} else if (archive_entry_size(entry) > 0) {
-			r = copy_data(a, ext);
+			r = extract_data(a, ext);
+			if (r != ARCHIVE_OK)
+				cpio->return_value = 1;
 		}
 	}
 	r = archive_read_close(a);
-	if (cpio->dot)
-		fprintf(stderr, "\n");
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(a));
+		errc(1, 0, "%s", archive_error_string(a));
 	r = archive_write_close(ext);
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(ext));
+		errc(1, 0, "%s", archive_error_string(ext));
 	if (!cpio->quiet) {
-		blocks = (archive_position_uncompressed(a) + 511)
+		int64_t blocks = (archive_position_uncompressed(a) + 511)
 			      / 512;
-		fprintf(stderr, "%lu %s\n", blocks,
+		fprintf(stderr, "%lu %s\n", (unsigned long)blocks,
 		    blocks == 1 ? "block" : "blocks");
 	}
 	archive_read_finish(a);
 	archive_write_finish(ext);
-	exit(0);
+	exit(cpio->return_value);
 }
 
+/*
+ * Exits if there's a fatal error.  Returns ARCHIVE_OK
+ * if everything is kosher.
+ */
 static int
-copy_data(struct archive *ar, struct archive *aw)
+extract_data(struct archive *ar, struct archive *aw)
 {
 	int r;
 	size_t size;
@@ -915,14 +889,14 @@ copy_data(struct archive *ar, struct archive *aw)
 		if (r == ARCHIVE_EOF)
 			return (ARCHIVE_OK);
 		if (r != ARCHIVE_OK) {
-			cpio_warnc(archive_errno(ar),
+			warnc(archive_errno(ar),
 			    "%s", archive_error_string(ar));
-			return (r);
+			exit(1);
 		}
 		r = archive_write_data_block(aw, block, size, offset);
 		if (r != ARCHIVE_OK) {
-			cpio_warnc(archive_errno(aw),
-			    archive_error_string(aw));
+			warnc(archive_errno(aw),
+			    "%s", archive_error_string(aw));
 			return (r);
 		}
 	}
@@ -933,32 +907,26 @@ mode_list(struct cpio *cpio)
 {
 	struct archive *a;
 	struct archive_entry *entry;
-	unsigned long blocks;
 	int r;
 
 	a = archive_read_new();
 	if (a == NULL)
-		cpio_errc(1, 0, "Couldn't allocate archive object");
-#ifdef SMALLER
-	archive_read_support_format_cpio(a);
-	archive_read_support_format_tar(a);
-#else
+		errc(1, 0, "Couldn't allocate archive object");
 	archive_read_support_compression_all(a);
 	archive_read_support_format_all(a);
-#endif
 
 	if (archive_read_open_file(a, cpio->filename, cpio->bytes_per_block))
-		cpio_errc(1, archive_errno(a),
-		    archive_error_string(a));
+		errc(1, archive_errno(a),
+		    "%s", archive_error_string(a));
 	for (;;) {
 		r = archive_read_next_header(a, &entry);
 		if (r == ARCHIVE_EOF)
 			break;
 		if (r != ARCHIVE_OK) {
-			cpio_errc(1, archive_errno(a),
-			    archive_error_string(a));
+			errc(1, archive_errno(a),
+			    "%s", archive_error_string(a));
 		}
-		if (excluded(cpio, archive_entry_pathname(entry)))
+		if (excluded(cpio->matching, archive_entry_pathname(entry)))
 			continue;
 		if (cpio->verbose)
 			list_item_verbose(cpio, entry);
@@ -967,11 +935,11 @@ mode_list(struct cpio *cpio)
 	}
 	r = archive_read_close(a);
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(a));
+		errc(1, 0, "%s", archive_error_string(a));
 	if (!cpio->quiet) {
-		blocks = (archive_position_uncompressed(a) + 511)
+		int64_t blocks = (archive_position_uncompressed(a) + 511)
 			      / 512;
-		fprintf(stderr, "%lu %s\n", blocks,
+		fprintf(stderr, "%lu %s\n", (unsigned long)blocks,
 		    blocks == 1 ? "block" : "blocks");
 	}
 	archive_read_finish(a);
@@ -994,23 +962,18 @@ list_item_verbose(struct cpio *cpio, struct archive_entry *entry)
 	char			 uids[16], gids[16];
 	const char 		*uname, *gname;
 	FILE			*out = stdout;
-	const struct stat	*st;
 	const char		*fmt;
-	time_t			 tim;
+	time_t			 mtime;
 	static time_t		 now;
 
-	st = archive_entry_stat(entry);
-
 	if (!now)
 		time(&now);
 
 	if (cpio->option_numeric_uid_gid) {
 		/* Format numeric uid/gid for display. */
-		snprintf(uids, sizeof(uids), "%jd",
-		    (intmax_t)archive_entry_uid(entry));
+		strcpy(uids, cpio_i64toa(archive_entry_uid(entry)));
 		uname = uids;
-		snprintf(gids, sizeof(gids), "%jd",
-		    (intmax_t)archive_entry_gid(entry));
+		strcpy(gids, cpio_i64toa(archive_entry_gid(entry)));
 		gname = gids;
 	} else {
 		/* Use uname if it's present, else lookup name from uid. */
@@ -1024,30 +987,31 @@ list_item_verbose(struct cpio *cpio, struct archive_entry *entry)
 	}
 
 	/* Print device number or file size. */
-	if (S_ISCHR(st->st_mode) || S_ISBLK(st->st_mode)) {
+	if (archive_entry_filetype(entry) == AE_IFCHR
+	    || archive_entry_filetype(entry) == AE_IFBLK) {
 		snprintf(size, sizeof(size), "%lu,%lu",
-		    (unsigned long)major(st->st_rdev),
-		    (unsigned long)minor(st->st_rdev)); /* ls(1) also casts here. */
+		    (unsigned long)archive_entry_rdevmajor(entry),
+		    (unsigned long)archive_entry_rdevminor(entry));
 	} else {
-		snprintf(size, sizeof(size), CPIO_FILESIZE_PRINTF,
-		    (CPIO_FILESIZE_TYPE)st->st_size);
+		strcpy(size, cpio_i64toa(archive_entry_size(entry)));
 	}
 
 	/* Format the time using 'ls -l' conventions. */
-	tim = (time_t)st->st_mtime;
+	mtime = archive_entry_mtime(entry);
 #if defined(_WIN32) && !defined(__CYGWIN__)
 	/* Windows' strftime function does not support %e format. */
-	if (abs(tim - now) > (365/2)*86400)
+	if (mtime - now > 365*86400/2
+		|| mtime - now < -365*86400/2)
 		fmt = cpio->day_first ? "%d %b  %Y" : "%b %d  %Y";
 	else
 		fmt = cpio->day_first ? "%d %b %H:%M" : "%b %d %H:%M";
 #else
-	if (abs(tim - now) > (365/2)*86400)
+	if (abs(mtime - now) > (365/2)*86400)
 		fmt = cpio->day_first ? "%e %b  %Y" : "%b %e  %Y";
 	else
 		fmt = cpio->day_first ? "%e %b %H:%M" : "%b %e %H:%M";
 #endif
-	strftime(date, sizeof(date), fmt, localtime(&tim));
+	strftime(date, sizeof(date), fmt, localtime(&mtime));
 
 	fprintf(out, "%s%3d %-8s %-8s %8s %12s %s",
 	    archive_entry_strmode(entry),
@@ -1066,7 +1030,6 @@ list_item_verbose(struct cpio *cpio, struct archive_entry *entry)
 static void
 mode_pass(struct cpio *cpio, const char *destdir)
 {
-	unsigned long blocks;
 	struct line_reader *lr;
 	const char *p;
 	int r;
@@ -1079,28 +1042,37 @@ mode_pass(struct cpio *cpio, const char *destdir)
 
 	cpio->archive = archive_write_disk_new();
 	if (cpio->archive == NULL)
-		cpio_errc(1, 0, "Failed to allocate archive object");
+		errc(1, 0, "Failed to allocate archive object");
 	r = archive_write_disk_set_options(cpio->archive, cpio->extract_flags);
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(cpio->archive));
+		errc(1, 0, "%s", archive_error_string(cpio->archive));
 	cpio->linkresolver = archive_entry_linkresolver_new();
 	archive_write_disk_set_standard_lookup(cpio->archive);
-	lr = process_lines_init("-", cpio->line_separator);
-	while ((p = process_lines_next(lr)) != NULL)
+
+	cpio->archive_read_disk = archive_read_disk_new();
+	if (cpio->archive_read_disk == NULL)
+		errc(1, 0, "Failed to allocate archive object");
+	if (cpio->option_follow_links)
+		archive_read_disk_set_symlink_logical(cpio->archive_read_disk);
+	else
+		archive_read_disk_set_symlink_physical(cpio->archive_read_disk);
+	archive_read_disk_set_standard_lookup(cpio->archive_read_disk);
+
+	lr = line_reader("-", cpio->option_null);
+	while ((p = line_reader_next(lr)) != NULL)
 		file_to_archive(cpio, p);
-	process_lines_free(lr);
+	line_reader_free(lr);
 
 	archive_entry_linkresolver_free(cpio->linkresolver);
 	r = archive_write_close(cpio->archive);
-	if (cpio->dot)
-		fprintf(stderr, "\n");
 	if (r != ARCHIVE_OK)
-		cpio_errc(1, 0, archive_error_string(cpio->archive));
+		errc(1, 0, "%s", archive_error_string(cpio->archive));
 
 	if (!cpio->quiet) {
-		blocks = (archive_position_uncompressed(cpio->archive) + 511)
-			      / 512;
-		fprintf(stderr, "%lu %s\n", blocks,
+		int64_t blocks =
+			(archive_position_uncompressed(cpio->archive) + 511)
+			/ 512;
+		fprintf(stderr, "%lu %s\n", (unsigned long)blocks,
 		    blocks == 1 ? "block" : "blocks");
 	}
 
@@ -1150,130 +1122,6 @@ cpio_rename(const char *name)
 	return (ret);
 }
 
-
-/*
- * Read lines from file and do something with each one.  If option_null
- * is set, lines are terminated with zero bytes; otherwise, they're
- * terminated with newlines.
- *
- * This uses a self-sizing buffer to handle arbitrarily-long lines.
- */
-struct line_reader {
-	FILE *f;
-	char *buff, *buff_end, *line_start, *line_end, *p;
-	char *pathname;
-	size_t buff_length;
-	int separator;
-	int ret;
-};
-
-struct line_reader *
-process_lines_init(const char *pathname, char separator)
-{
-	struct line_reader *lr;
-
-	lr = calloc(1, sizeof(*lr));
-	if (lr == NULL)
-		cpio_errc(1, ENOMEM, "Can't open %s", pathname);
-
-	lr->separator = separator;
-	lr->pathname = strdup(pathname);
-
-	if (strcmp(pathname, "-") == 0)
-		lr->f = stdin;
-	else
-		lr->f = fopen(pathname, "r");
-	if (lr->f == NULL)
-		cpio_errc(1, errno, "Couldn't open %s", pathname);
-	lr->buff_length = 8192;
-	lr->buff = malloc(lr->buff_length);
-	if (lr->buff == NULL)
-		cpio_errc(1, ENOMEM, "Can't read %s", pathname);
-	lr->line_start = lr->line_end = lr->buff_end = lr->buff;
-
-	return (lr);
-}
-
-const char *
-process_lines_next(struct line_reader *lr)
-{
-	size_t bytes_wanted, bytes_read, new_buff_size;
-	char *line_start, *p;
-
-	for (;;) {
-		/* If there's a line in the buffer, return it immediately. */
-		while (lr->line_end < lr->buff_end) {
-			if (*lr->line_end == lr->separator) {
-				*lr->line_end = '\0';
-				line_start = lr->line_start;
-				lr->line_start = lr->line_end + 1;
-				lr->line_end = lr->line_start;
-				return (line_start);
-			} else
-				lr->line_end++;
-		}
-
-		/* If we're at end-of-file, process the final data. */
-		if (lr->f == NULL) {
-			/* If there's more text, return one last line. */
-			if (lr->line_end > lr->line_start) {
-				*lr->line_end = '\0';
-				line_start = lr->line_start;
-				lr->line_start = lr->line_end + 1;
-				lr->line_end = lr->line_start;
-				return (line_start);
-			}
-			/* Otherwise, we're done. */
-			return (NULL);
-		}
-
-		/* Buffer only has part of a line. */
-		if (lr->line_start > lr->buff) {
-			/* Move a leftover fractional line to the beginning. */
-			memmove(lr->buff, lr->line_start,
-			    lr->buff_end - lr->line_start);
-			lr->buff_end -= lr->line_start - lr->buff;
-			lr->line_end -= lr->line_start - lr->buff;
-			lr->line_start = lr->buff;
-		} else {
-			/* Line is too big; enlarge the buffer. */
-			new_buff_size = lr->buff_length * 2;
-			if (new_buff_size <= lr->buff_length)
-				cpio_errc(1, ENOMEM,
-				    "Line too long in %s", lr->pathname);
-			lr->buff_length = new_buff_size;
-			p = realloc(lr->buff, new_buff_size);
-			if (p == NULL)
-				cpio_errc(1, ENOMEM,
-				    "Line too long in %s", lr->pathname);
-			lr->buff_end = p + (lr->buff_end - lr->buff);
-			lr->line_end = p + (lr->line_end - lr->buff);
-			lr->line_start = lr->buff = p;
-		}
-
-		/* Get some more data into the buffer. */
-		bytes_wanted = lr->buff + lr->buff_length - lr->buff_end;
-		bytes_read = fread(lr->buff_end, 1, bytes_wanted, lr->f);
-		lr->buff_end += bytes_read;
-
-		if (ferror(lr->f))
-			cpio_errc(1, errno, "Can't read %s", lr->pathname);
-		if (feof(lr->f)) {
-			if (lr->f != stdin)
-				fclose(lr->f);
-			lr->f = NULL;
-		}
-	}
-}
-
-void
-process_lines_free(struct line_reader *lr)
-{
-	free(lr->buff);
-	free(lr->pathname);
-	free(lr);
-}
-
 static void
 free_cache(struct name_cache *cache)
 {
@@ -1302,7 +1150,7 @@ lookup_name(struct cpio *cpio, struct name_cache **name_cache_variable,
 	if (*name_cache_variable == NULL) {
 		*name_cache_variable = malloc(sizeof(struct name_cache));
 		if (*name_cache_variable == NULL)
-			cpio_errc(1, ENOMEM, "No more memory");
+			errc(1, ENOMEM, "No more memory");
 		memset(*name_cache_variable, 0, sizeof(struct name_cache));
 		(*name_cache_variable)->size = name_cache_size;
 	}
@@ -1358,8 +1206,8 @@ lookup_uname_helper(struct cpio *cpio, const char **name, id_t id)
 	pwent = getpwuid((uid_t)id);
 	if (pwent == NULL) {
 		*name = NULL;
-		if (errno != 0)
-			cpio_warnc(errno, "getpwuid(%d) failed", id);
+		if (errno != 0 && errno != ENOENT)
+			warnc(errno, "getpwuid(%d) failed", id);
 		return (errno);
 	}
 
@@ -1386,10 +1234,34 @@ lookup_gname_helper(struct cpio *cpio, const char **name, id_t id)
 	if (grent == NULL) {
 		*name = NULL;
 		if (errno != 0)
-			cpio_warnc(errno, "getgrgid(%d) failed", id);
+			warnc(errno, "getgrgid(%d) failed", id);
 		return (errno);
 	}
 
 	*name = grent->gr_name;
 	return (0);
 }
+
+/*
+ * It would be nice to just use printf() for formatting large numbers,
+ * but the compatibility problems are a big headache.  Hence the
+ * following simple utility function.
+ */
+const char *
+cpio_i64toa(int64_t n0)
+{
+	// 2^64 =~ 1.8 * 10^19, so 20 decimal digits suffice.
+	// We also need 1 byte for '-' and 1 for '\0'.
+	static char buff[22];
+	int64_t n = n0 < 0 ? -n0 : n0;
+	char *p = buff + sizeof(buff);
+
+	*--p = '\0';
+	do {
+		*--p = '0' + (int)(n % 10);
+		n /= 10;
+	} while (n > 0);
+	if (n0 < 0)
+		*--p = '-';
+	return p;
+}
diff --git a/usr.bin/cpio/cpio.h b/usr.bin/cpio/cpio.h
index 8c187372513..36dab55bc26 100644
--- a/usr.bin/cpio/cpio.h
+++ b/usr.bin/cpio/cpio.h
@@ -31,7 +31,7 @@
 #include "cpio_platform.h"
 #include <stdio.h>
 
-#define	DEFAULT_BYTES_PER_BLOCK	(20*512)
+#include "matching.h"
 
 /*
  * The internal state for the "cpio" program.
@@ -52,17 +52,16 @@ struct cpio {
 	const char	 *format; /* -H format */
 	int		  bytes_per_block; /* -b block_size */
 	int		  verbose;   /* -v */
-	int		  dot;  /* -V */
 	int		  quiet;   /* --quiet */
 	int		  extract_flags; /* Flags for extract operation */
 	char		  symlink_mode; /* H or L, per BSD conventions */
 	const char	 *compress_program;
-	char		  line_separator; /* --null ? '\0' : '\n' */
 	int		  option_append; /* -A, only relevant for -o */
 	int		  option_atime_restore; /* -a */
 	int		  option_follow_links; /* -L */
 	int		  option_link; /* -l */
 	int		  option_list; /* -t */
+	char		  option_null; /* --null */
 	int		  option_numeric_uid_gid; /* -n */
 	int		  option_rename; /* -r */
 	char		 *destdir;
@@ -77,6 +76,7 @@ struct cpio {
 
 	/* Miscellaneous state information */
 	struct archive	 *archive;
+	struct archive	 *archive_read_disk;
 	int		  argc;
 	char		**argv;
 	int		  return_value; /* Value returned by main() */
@@ -91,30 +91,19 @@ struct cpio {
 	size_t		  buff_size;
 };
 
-/* Name of this program; used in error reporting, initialized in main(). */
-const char *cpio_progname;
-
-void	cpio_errc(int _eval, int _code, const char *fmt, ...) __LA_DEAD;
-void	cpio_warnc(int _code, const char *fmt, ...);
-
-int	owner_parse(const char *, int *, int *);
+const char *owner_parse(const char *, int *, int *);
 
 
 /* Fake short equivalents for long options that otherwise lack them. */
 enum {
 	OPTION_INSECURE = 1,
+	OPTION_LZMA,
 	OPTION_NO_PRESERVE_OWNER,
+	OPTION_PRESERVE_OWNER,
 	OPTION_QUIET,
 	OPTION_VERSION
 };
 
-struct line_reader;
-
-struct line_reader *process_lines_init(const char *, char separator);
-const char *process_lines_next(struct line_reader *);
-void	process_lines_free(struct line_reader *);
-
 int	cpio_getopt(struct cpio *cpio);
-int	include_from_file(struct cpio *, const char *);
 
 #endif
diff --git a/usr.bin/cpio/cpio_platform.h b/usr.bin/cpio/cpio_platform.h
index 037376a8c6d..3043828db2a 100644
--- a/usr.bin/cpio/cpio_platform.h
+++ b/usr.bin/cpio/cpio_platform.h
@@ -37,19 +37,18 @@
 #if defined(PLATFORM_CONFIG_H)
 /* Use hand-built config.h in environments that need it. */
 #include PLATFORM_CONFIG_H
-#elif defined(HAVE_CONFIG_H)
-/* Most POSIX platforms use the 'configure' script to build config.h */
-#include "config.h"
 #else
-/* Warn if cpio hasn't been (automatically or manually) configured. */
-#error Oops: No config.h and no built-in configuration in cpio_platform.h.
-#endif /* !HAVE_CONFIG_H */
+/* Read config.h or die trying. */
+#include "config.h"
+#endif
 
-/* No non-FreeBSD platform will have __FBSDID, so just define it here. */
-#ifdef __FreeBSD__
-#include <sys/cdefs.h>  /* For __FBSDID */
-#elif !defined(__FBSDID)
-/* Just leaving this macro replacement empty leads to a dangling semicolon. */
+/* Get a real definition for __FBSDID if we can */
+#if HAVE_SYS_CDEFS_H
+#include <sys/cdefs.h>
+#endif
+
+/* If not, define it so as to avoid dangling semicolons. */
+#ifndef __FBSDID
 #define	__FBSDID(a)     struct _undefined_hack
 #endif
 
@@ -63,24 +62,6 @@
 #include "archive_entry.h"
 #endif
 
-/*
- * We need to be able to display a filesize using printf().  The type
- * and format string here must be compatible with one another and
- * large enough for any file.
- */
-#if HAVE_UINTMAX_T
-#define	CPIO_FILESIZE_TYPE	uintmax_t
-#define	CPIO_FILESIZE_PRINTF	"%ju"
-#else
-#if HAVE_UNSIGNED_LONG_LONG
-#define	CPIO_FILESIZE_TYPE	unsigned long long
-#define	CPIO_FILESIZE_PRINTF	"%llu"
-#else
-#define	CPIO_FILESIZE_TYPE	unsigned long
-#define	CPIO_FILESIZE_PRINTF	"%lu"
-#endif
-#endif
-
 /* How to mark functions that don't return. */
 #if defined(__GNUC__) && (__GNUC__ > 2 || \
                           (__GNUC__ == 2 && __GNUC_MINOR__ >= 5))
@@ -89,9 +70,7 @@
 #define __LA_DEAD
 #endif
 
-#if defined(__CYGWIN__)
-#include "cpio_cygwin.h"
-#elif defined(_WIN32)	/* && !__CYGWIN__ */
+#if defined(_WIN32) && !defined(__CYGWIN__)
 #include "cpio_windows.h"
 #endif
 
diff --git a/usr.bin/cpio/err.c b/usr.bin/cpio/err.c
index ad9c0e193e1..c74e097d443 100644
--- a/usr.bin/cpio/err.c
+++ b/usr.bin/cpio/err.c
@@ -24,7 +24,6 @@
  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
  */
 
-
 #include "cpio_platform.h"
 __FBSDID("$FreeBSD$");
 
@@ -39,12 +38,14 @@ __FBSDID("$FreeBSD$");
 #include <string.h>
 #endif
 
-#include "cpio.h"
+#include "err.h"
+
+const char *progname;
 
 static void
-cpio_vwarnc(int code, const char *fmt, va_list ap)
+vwarnc(int code, const char *fmt, va_list ap)
 {
-	fprintf(stderr, "%s: ", cpio_progname);
+	fprintf(stderr, "%s: ", progname);
 	vfprintf(stderr, fmt, ap);
 	if (code != 0)
 		fprintf(stderr, ": %s", strerror(code));
@@ -52,22 +53,22 @@ cpio_vwarnc(int code, const char *fmt, va_list ap)
 }
 
 void
-cpio_warnc(int code, const char *fmt, ...)
+warnc(int code, const char *fmt, ...)
 {
 	va_list ap;
 
 	va_start(ap, fmt);
-	cpio_vwarnc(code, fmt, ap);
+	vwarnc(code, fmt, ap);
 	va_end(ap);
 }
 
 void
-cpio_errc(int eval, int code, const char *fmt, ...)
+errc(int eval, int code, const char *fmt, ...)
 {
 	va_list ap;
 
 	va_start(ap, fmt);
-	cpio_vwarnc(code, fmt, ap);
+	vwarnc(code, fmt, ap);
 	va_end(ap);
 	exit(eval);
 }
diff --git a/usr.bin/cpio/err.h b/usr.bin/cpio/err.h
new file mode 100644
index 00000000000..fe96c701244
--- /dev/null
+++ b/usr.bin/cpio/err.h
@@ -0,0 +1,43 @@
+/*-
+ * Copyright (c) 2009 Joerg Sonnenberger
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef LAFE_ERR_H
+#define LAFE_ERR_H
+
+#if defined(__GNUC__) && (__GNUC__ > 2 || \
+                          (__GNUC__ == 2 && __GNUC_MINOR__ >= 5))
+#define __LA_DEAD       __attribute__((__noreturn__))
+#else
+#define __LA_DEAD
+#endif
+
+extern const char *progname;
+
+void	warnc(int code, const char *fmt, ...);
+void	errc(int eval, int code, const char *fmt, ...) __LA_DEAD;
+
+#endif
diff --git a/usr.bin/cpio/line_reader.c b/usr.bin/cpio/line_reader.c
new file mode 100644
index 00000000000..846194131ec
--- /dev/null
+++ b/usr.bin/cpio/line_reader.c
@@ -0,0 +1,171 @@
+/*-
+ * Copyright (c) 2008 Tim Kientzle
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer
+ *    in this position and unchanged.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ */
+
+#include "cpio_platform.h"
+__FBSDID("$FreeBSD$");
+
+#include <errno.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+#include "err.h"
+#include "line_reader.h"
+
+#if defined(_WIN32) && !defined(__CYGWIN__) && !defined(__BORLANDC__)
+#define strdup _strdup
+#endif
+
+/*
+ * Read lines from file and do something with each one.  If option_null
+ * is set, lines are terminated with zero bytes; otherwise, they're
+ * terminated with newlines.
+ *
+ * This uses a self-sizing buffer to handle arbitrarily-long lines.
+ */
+struct line_reader {
+	FILE *f;
+	char *buff, *buff_end, *line_start, *line_end, *p;
+	char *pathname;
+	size_t buff_length;
+	int nullSeparator; /* Lines separated by null, not CR/CRLF/etc. */
+	int ret;
+};
+
+struct line_reader *
+line_reader(const char *pathname, int nullSeparator)
+{
+	struct line_reader *lr;
+
+	lr = calloc(1, sizeof(*lr));
+	if (lr == NULL)
+		errc(1, ENOMEM, "Can't open %s", pathname);
+
+	lr->nullSeparator = nullSeparator;
+	lr->pathname = strdup(pathname);
+
+	if (strcmp(pathname, "-") == 0)
+		lr->f = stdin;
+	else
+		lr->f = fopen(pathname, "r");
+	if (lr->f == NULL)
+		errc(1, errno, "Couldn't open %s", pathname);
+	lr->buff_length = 8192;
+	lr->buff = malloc(lr->buff_length);
+	if (lr->buff == NULL)
+		errc(1, ENOMEM, "Can't read %s", pathname);
+	lr->line_start = lr->line_end = lr->buff_end = lr->buff;
+
+	return (lr);
+}
+
+const char *
+line_reader_next(struct line_reader *lr)
+{
+	size_t bytes_wanted, bytes_read, new_buff_size;
+	char *line_start, *p;
+
+	for (;;) {
+		/* If there's a line in the buffer, return it immediately. */
+		while (lr->line_end < lr->buff_end) {
+			if (lr->nullSeparator) {
+				if (*lr->line_end == '\0') {
+					line_start = lr->line_start;
+					lr->line_start = lr->line_end + 1;
+					lr->line_end = lr->line_start;
+					return (line_start);
+				}
+			} else if (*lr->line_end == '\x0a' || *lr->line_end == '\x0d') {
+				*lr->line_end = '\0';
+				line_start = lr->line_start;
+				lr->line_start = lr->line_end + 1;
+				lr->line_end = lr->line_start;
+				if (line_start[0] != '\0')
+					return (line_start);
+			}
+			lr->line_end++;
+		}
+
+		/* If we're at end-of-file, process the final data. */
+		if (lr->f == NULL) {
+			/* If there's more text, return one last line. */
+			if (lr->line_end > lr->line_start) {
+				*lr->line_end = '\0';
+				line_start = lr->line_start;
+				lr->line_start = lr->line_end + 1;
+				lr->line_end = lr->line_start;
+				return (line_start);
+			}
+			/* Otherwise, we're done. */
+			return (NULL);
+		}
+
+		/* Buffer only has part of a line. */
+		if (lr->line_start > lr->buff) {
+			/* Move a leftover fractional line to the beginning. */
+			memmove(lr->buff, lr->line_start,
+			    lr->buff_end - lr->line_start);
+			lr->buff_end -= lr->line_start - lr->buff;
+			lr->line_end -= lr->line_start - lr->buff;
+			lr->line_start = lr->buff;
+		} else {
+			/* Line is too big; enlarge the buffer. */
+			new_buff_size = lr->buff_length * 2;
+			if (new_buff_size <= lr->buff_length)
+				errc(1, ENOMEM,
+				    "Line too long in %s", lr->pathname);
+			lr->buff_length = new_buff_size;
+			p = realloc(lr->buff, new_buff_size);
+			if (p == NULL)
+				errc(1, ENOMEM,
+				    "Line too long in %s", lr->pathname);
+			lr->buff_end = p + (lr->buff_end - lr->buff);
+			lr->line_end = p + (lr->line_end - lr->buff);
+			lr->line_start = lr->buff = p;
+		}
+
+		/* Get some more data into the buffer. */
+		bytes_wanted = lr->buff + lr->buff_length - lr->buff_end;
+		bytes_read = fread(lr->buff_end, 1, bytes_wanted, lr->f);
+		lr->buff_end += bytes_read;
+
+		if (ferror(lr->f))
+			errc(1, errno, "Can't read %s", lr->pathname);
+		if (feof(lr->f)) {
+			if (lr->f != stdin)
+				fclose(lr->f);
+			lr->f = NULL;
+		}
+	}
+}
+
+void
+line_reader_free(struct line_reader *lr)
+{
+	free(lr->buff);
+	free(lr->pathname);
+	free(lr);
+}
diff --git a/usr.bin/cpio/line_reader.h b/usr.bin/cpio/line_reader.h
new file mode 100644
index 00000000000..4f5dd6afaf0
--- /dev/null
+++ b/usr.bin/cpio/line_reader.h
@@ -0,0 +1,37 @@
+/*-
+ * Copyright (c) 2009 Joerg Sonnenberger
+ * All rights reserved.
+ *
+ * Redistribution and use in source and binary forms, with or without
+ * modification, are permitted provided that the following conditions
+ * are met:
+ * 1. Redistributions of source code must retain the above copyright
+ *    notice, this list of conditions and the following disclaimer.
+ * 2. Redistributions in binary form must reproduce the above copyright
+ *    notice, this list of conditions and the following disclaimer in the
+ *    documentation and/or other materials provided with the distribution.
+ *
+ * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
+ * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
+ * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
+ * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
+ * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
+ * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
+ * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
+ * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
+ * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
+ * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+ *
+ * $FreeBSD$
+ */
+
+#ifndef LAFE_LINE_READER_H
+#define LAFE_LINE_READER_H
+
+struct line_reader;
+
+struct line_reader *line_reader(const char *, int nullSeparator);
+const char *line_reader_next(struct line_reader *);
+void	line_reader_free(struct line_reader *);
+
+#endif
diff --git a/usr.bin/cpio/matching.c b/usr.bin/cpio/matching.c
index 8d1798fd50f..0fd2b6a90d1 100644
--- a/usr.bin/cpio/matching.c
+++ b/usr.bin/cpio/matching.c
@@ -36,6 +36,8 @@ __FBSDID("$FreeBSD$");
 #include <string.h>
 #endif
 
+#include "err.h"
+#include "line_reader.h"
 #include "matching.h"
 #include "pathmatch.h"
 
@@ -54,7 +56,7 @@ struct matching {
 };
 
 static void	add_pattern(struct match **list, const char *pattern);
-static void	initialize_matching(struct cpio *);
+static void	initialize_matching(struct matching **);
 static int	match_exclusion(struct match *, const char *pathname);
 static int	match_inclusion(struct match *, const char *pathname);
 
@@ -70,52 +72,58 @@ static int	match_inclusion(struct match *, const char *pathname);
  */
 
 int
-exclude(struct cpio *cpio, const char *pattern)
+exclude(struct matching **matching, const char *pattern)
 {
-	struct matching *matching;
 
-	if (cpio->matching == NULL)
-		initialize_matching(cpio);
-	matching = cpio->matching;
-	add_pattern(&(matching->exclusions), pattern);
-	matching->exclusions_count++;
-	return (0);
-}
-
-#if 0
-int
-exclude_from_file(struct cpio *cpio, const char *pathname)
-{
-	return (process_lines(cpio, pathname, &exclude));
-}
-#endif
-
-int
-include(struct cpio *cpio, const char *pattern)
-{
-	struct matching *matching;
-
-	if (cpio->matching == NULL)
-		initialize_matching(cpio);
-	matching = cpio->matching;
-	add_pattern(&(matching->inclusions), pattern);
-	matching->inclusions_count++;
-	matching->inclusions_unmatched_count++;
+	if (*matching == NULL)
+		initialize_matching(matching);
+	add_pattern(&((*matching)->exclusions), pattern);
+	(*matching)->exclusions_count++;
 	return (0);
 }
 
 int
-include_from_file(struct cpio *cpio, const char *pathname)
+exclude_from_file(struct matching **matching, const char *pathname)
 {
 	struct line_reader *lr;
 	const char *p;
 	int ret = 0;
 
-	lr = process_lines_init(pathname, '\n');
-	while ((p = process_lines_next(lr)) != NULL)
-		if (include(cpio, p) != 0)
+	lr = line_reader(pathname, 0);
+	while ((p = line_reader_next(lr)) != NULL) {
+		if (exclude(matching, p) != 0)
 			ret = -1;
-	process_lines_free(lr);
+	}
+	line_reader_free(lr);
+	return (ret);
+}
+
+int
+include(struct matching **matching, const char *pattern)
+{
+
+	if (*matching == NULL)
+		initialize_matching(matching);
+	add_pattern(&((*matching)->inclusions), pattern);
+	(*matching)->inclusions_count++;
+	(*matching)->inclusions_unmatched_count++;
+	return (0);
+}
+
+int
+include_from_file(struct matching **matching, const char *pathname,
+    int nullSeparator)
+{
+	struct line_reader *lr;
+	const char *p;
+	int ret = 0;
+
+	lr = line_reader(pathname, nullSeparator);
+	while ((p = line_reader_next(lr)) != NULL) {
+		if (include(matching, p) != 0)
+			ret = -1;
+	}
+	line_reader_free(lr);
 	return (ret);
 }
 
@@ -123,15 +131,15 @@ static void
 add_pattern(struct match **list, const char *pattern)
 {
 	struct match *match;
+	size_t len;
 
-	match = malloc(sizeof(*match) + strlen(pattern) + 1);
+	len = strlen(pattern);
+	match = malloc(sizeof(*match) + len + 1);
 	if (match == NULL)
-		cpio_errc(1, errno, "Out of memory");
-	if (pattern[0] == '/')
-		pattern++;
+		errc(1, errno, "Out of memory");
 	strcpy(match->pattern, pattern);
 	/* Both "foo/" and "foo" should match "foo/bar". */
-	if (match->pattern[strlen(match->pattern)-1] == '/')
+	if (len && match->pattern[len - 1] == '/')
 		match->pattern[strlen(match->pattern)-1] = '\0';
 	match->next = *list;
 	*list = match;
@@ -140,13 +148,11 @@ add_pattern(struct match **list, const char *pattern)
 
 
 int
-excluded(struct cpio *cpio, const char *pathname)
+excluded(struct matching *matching, const char *pathname)
 {
-	struct matching *matching;
 	struct match *match;
 	struct match *matched;
 
-	matching = cpio->matching;
 	if (matching == NULL)
 		return (0);
 
@@ -166,7 +172,7 @@ excluded(struct cpio *cpio, const char *pathname)
 			 */
 			if (match->matches == 0) {
 				match->matches++;
-				matching->inclusions_unmatched_count++;
+				matching->inclusions_unmatched_count--;
 				return (0);
 			}
 			/*
@@ -198,7 +204,7 @@ excluded(struct cpio *cpio, const char *pathname)
  * gtar.  In particular, 'a*b' will match 'foo/a1111/222b/bar'
  *
  */
-int
+static int
 match_exclusion(struct match *match, const char *pathname)
 {
 	return (pathmatch(match->pattern,
@@ -210,50 +216,69 @@ match_exclusion(struct match *match, const char *pathname)
  * Again, mimic gtar:  inclusions are always anchored (have to match
  * the beginning of the path) even though exclusions are not anchored.
  */
-int
+static int
 match_inclusion(struct match *match, const char *pathname)
 {
+#if 0
 	return (pathmatch(match->pattern, pathname, 0));
+#else
+	return (pathmatch(match->pattern, pathname, PATHMATCH_NO_ANCHOR_END));
+#endif	
 }
 
 void
-cleanup_exclusions(struct cpio *cpio)
+cleanup_exclusions(struct matching **matching)
 {
 	struct match *p, *q;
 
-	if (cpio->matching) {
-		p = cpio->matching->inclusions;
-		while (p != NULL) {
-			q = p;
-			p = p->next;
-			free(q);
-		}
-		p = cpio->matching->exclusions;
-		while (p != NULL) {
-			q = p;
-			p = p->next;
-			free(q);
-		}
-		free(cpio->matching);
+	if (*matching == NULL)
+		return;
+
+	for (p = (*matching)->inclusions; p != NULL; ) {
+		q = p;
+		p = p->next;
+		free(q);
 	}
+
+	for (p = (*matching)->exclusions; p != NULL; ) {
+		q = p;
+		p = p->next;
+		free(q);
+	}
+
+	free(*matching);
+	*matching = NULL;
 }
 
 static void
-initialize_matching(struct cpio *cpio)
+initialize_matching(struct matching **matching)
 {
-	cpio->matching = malloc(sizeof(*cpio->matching));
-	if (cpio->matching == NULL)
-		cpio_errc(1, errno, "No memory");
-	memset(cpio->matching, 0, sizeof(*cpio->matching));
+	*matching = calloc(sizeof(**matching), 1);
+	if (*matching == NULL)
+		errc(1, errno, "No memory");
 }
 
 int
-unmatched_inclusions(struct cpio *cpio)
+unmatched_inclusions(struct matching *matching)
 {
-	struct matching *matching;
 
-	matching = cpio->matching;
 	if (matching == NULL)
 		return (0);
 	return (matching->inclusions_unmatched_count);
 }
+
+int
+unmatched_inclusions_warn(struct matching *matching, const char *msg)
+{
+	struct match *p;
+
+	if (matching == NULL)
+		return (0);
+
+	for (p = matching->inclusions; p != NULL; p = p->next) {
+		if (p->matches == 0)
+			warnc(0, "%s: %s", p->pattern, msg);
+	}
+
+	return (matching->inclusions_unmatched_count);
+}
diff --git a/usr.bin/cpio/matching.h b/usr.bin/cpio/matching.h
index a66bc86bc73..e98e82e9c02 100644
--- a/usr.bin/cpio/matching.h
+++ b/usr.bin/cpio/matching.h
@@ -29,12 +29,18 @@
 #ifndef MATCHING_H
 #define MATCHING_H
 
-#include "cpio.h"
+struct matching;
 
-int	exclude(struct cpio *, const char *pattern);
-int	include(struct cpio *, const char *pattern);
-int	excluded(struct cpio *cpio, const char *pathname);
-void	cleanup_exclusions(struct cpio *cpio);
-int	unmatched_inclusions(struct cpio *cpio);
+int	exclude(struct matching **matching, const char *pattern);
+int	exclude_from_file(struct matching **matching,
+			       const char *pathname);
+int	include(struct matching **matching, const char *pattern);
+int	include_from_file(struct matching **matching,
+			       const char *pathname, int nullSeparator);
+
+int	excluded(struct matching *, const char *pathname);
+void	cleanup_exclusions(struct matching **);
+int	unmatched_inclusions(struct matching *);
+int	unmatched_inclusions_warn(struct matching *, const char *msg);
 
 #endif
diff --git a/usr.bin/cpio/pathmatch.c b/usr.bin/cpio/pathmatch.c
index 40fa836041b..75f9095af01 100644
--- a/usr.bin/cpio/pathmatch.c
+++ b/usr.bin/cpio/pathmatch.c
@@ -131,7 +131,6 @@ pm(const char *p, const char *s, int flags)
 				s = pm_slashskip(s);
 			}
 			return (*s == '\0');
-			break;
 		case '?':
 			/* ? always succeds, unless we hit end of 's' */
 			if (*s == '\0')
@@ -150,7 +149,6 @@ pm(const char *p, const char *s, int flags)
 				++s;
 			}
 			return (0);
-			break;
 		case '[':
 			/*
 			 * Find the end of the [...] character class,
@@ -229,9 +227,17 @@ pathmatch(const char *p, const char *s, int flags)
 		flags &= ~PATHMATCH_NO_ANCHOR_START;
 	}
 
-	/* Certain patterns anchor implicitly. */
-	if (*p == '*' || *p == '/')
+	if (*p == '/' && *s != '/')
+		return (0);
+
+	/* Certain patterns and file names anchor implicitly. */
+	if (*p == '*' || *p == '/' || *p == '/') {
+		while (*p == '/')
+			++p;
+		while (*s == '/')
+			++s;
 		return (pm(p, s, flags));
+	}
 
 	/* If start is unanchored, try to match start of each path element. */
 	if (flags & PATHMATCH_NO_ANCHOR_START) {
diff --git a/usr.bin/cpio/pathmatch.h b/usr.bin/cpio/pathmatch.h
index fd2c2575cc0..3e82830b456 100644
--- a/usr.bin/cpio/pathmatch.h
+++ b/usr.bin/cpio/pathmatch.h
@@ -26,13 +26,13 @@
  * $FreeBSD$
  */
 
-#ifndef PATHMATCH_H
-#define PATHMATCH_H
+#ifndef LAFE_PATHMATCH_H
+#define LAFE_PATHMATCH_H
 
 /* Don't anchor at beginning unless the pattern starts with "^" */
 #define PATHMATCH_NO_ANCHOR_START	1
 /* Don't anchor at end unless the pattern ends with "$" */
-#define PATHMATCH_NO_ANCHOR_END	2
+#define PATHMATCH_NO_ANCHOR_END 	2
 
 /* Note that "^" and "$" are not special unless you set the corresponding
  * flag above. */