Move the atpit device model from userspace into vmm.ko for better

precision and lower latency.

Approved by:	grehan (co-mentor)
This commit is contained in:
Tycho Nightingale 2014-03-25 19:20:34 +00:00
parent 717f4815f3
commit e883c9bb40
8 changed files with 395 additions and 306 deletions

View file

@ -191,6 +191,7 @@ struct vmspace *vm_get_vmspace(struct vm *vm);
int vm_assign_pptdev(struct vm *vm, int bus, int slot, int func);
int vm_unassign_pptdev(struct vm *vm, int bus, int slot, int func);
struct vatpic *vm_atpic(struct vm *vm);
struct vatpit *vm_atpit(struct vm *vm);
/*
* Inject exception 'vme' into the guest vcpu. This function returns 0 on

370
sys/amd64/vmm/io/vatpit.c Normal file
View file

@ -0,0 +1,370 @@
/*-
* Copyright (c) 2014 Tycho Nightingale <tycho.nightingale@pluribusnetworks.com>
* Copyright (c) 2011 NetApp, Inc.
* 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 NETAPP, INC ``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 NETAPP, INC 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/types.h>
#include <sys/queue.h>
#include <sys/cpuset.h>
#include <sys/kernel.h>
#include <sys/lock.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/systm.h>
#include <machine/vmm.h>
#include "vmm_ktr.h"
#include "vatpic.h"
#include "vioapic.h"
#include "vatpit.h"
static MALLOC_DEFINE(M_VATPIT, "atpit", "bhyve virtual atpit (8254)");
#define VATPIT_LOCK(vatpit) mtx_lock_spin(&((vatpit)->mtx))
#define VATPIT_UNLOCK(vatpit) mtx_unlock_spin(&((vatpit)->mtx))
#define VATPIT_LOCKED(vatpit) mtx_owned(&((vatpit)->mtx))
#define TIMER_SEL_MASK 0xc0
#define TIMER_RW_MASK 0x30
#define TIMER_MODE_MASK 0x0f
#define TIMER_SEL_READBACK 0xc0
#define PIT_8254_FREQ 1193182
#define TIMER_DIV(freq, hz) (((freq) + (hz) / 2) / (hz))
struct vatpit_callout_arg {
struct vatpit *vatpit;
int channel_num;
};
struct channel {
int mode;
uint16_t initial; /* initial counter value */
sbintime_t now_sbt; /* uptime when counter was loaded */
uint8_t cr[2];
uint8_t ol[2];
int crbyte;
int olbyte;
int frbyte;
struct callout callout;
sbintime_t callout_sbt; /* target time */
struct vatpit_callout_arg callout_arg;
};
struct vatpit {
struct vm *vm;
struct mtx mtx;
sbintime_t freq_sbt;
struct channel channel[3];
};
#define VATPIT_CTR0(vatpit, fmt) \
VM_CTR0((vatpit)->vm, fmt)
#define VATPIT_CTR1(vatpit, fmt, a1) \
VM_CTR1((vatpit)->vm, fmt, a1)
#define VATPIT_CTR2(vatpit, fmt, a1, a2) \
VM_CTR2((vatpit)->vm, fmt, a1, a2)
#define VATPIT_CTR3(vatpit, fmt, a1, a2, a3) \
VM_CTR3((vatpit)->vm, fmt, a1, a2, a3)
#define VATPIT_CTR4(vatpit, fmt, a1, a2, a3, a4) \
VM_CTR4((vatpit)->vm, fmt, a1, a2, a3, a4)
static void pit_timer_start_cntr0(struct vatpit *vatpit);
static void
vatpit_callout_handler(void *a)
{
struct vatpit_callout_arg *arg = a;
struct vatpit *vatpit;
struct callout *callout;
struct channel *c;
vatpit = arg->vatpit;
c = &vatpit->channel[arg->channel_num];
callout = &c->callout;
VATPIT_CTR1(vatpit, "atpit t%d fired", arg->channel_num);
VATPIT_LOCK(vatpit);
if (callout_pending(callout)) /* callout was reset */
goto done;
if (!callout_active(callout)) /* callout was stopped */
goto done;
callout_deactivate(callout);
if (c->mode == TIMER_RATEGEN) {
pit_timer_start_cntr0(vatpit);
}
vatpic_pulse_irq(vatpit->vm, 0);
vioapic_pulse_irq(vatpit->vm, 2);
done:
VATPIT_UNLOCK(vatpit);
return;
}
static void
pit_timer_start_cntr0(struct vatpit *vatpit)
{
struct channel *c;
sbintime_t delta, precision;
c = &vatpit->channel[0];
if (c->initial != 0) {
delta = c->initial * vatpit->freq_sbt;
precision = delta >> tc_precexp;
c->callout_sbt = c->callout_sbt + delta;
callout_reset_sbt(&c->callout, c->callout_sbt,
precision, vatpit_callout_handler, &c->callout_arg,
C_ABSOLUTE);
}
}
static uint16_t
pit_update_counter(struct vatpit *vatpit, struct channel *c, bool latch)
{
uint16_t lval;
sbintime_t delta_ticks;
/* cannot latch a new value until the old one has been consumed */
if (latch && c->olbyte != 0)
return (0);
if (c->initial == 0) {
/*
* This is possibly an o/s bug - reading the value of
* the timer without having set up the initial value.
*
* The original user-space version of this code set
* the timer to 100hz in this condition; do the same
* here.
*/
c->initial = TIMER_DIV(PIT_8254_FREQ, 100);
c->now_sbt = sbinuptime();
}
delta_ticks = (sbinuptime() - c->now_sbt) / vatpit->freq_sbt;
lval = c->initial - delta_ticks % c->initial;
if (latch) {
c->olbyte = 2;
c->ol[1] = lval; /* LSB */
c->ol[0] = lval >> 8; /* MSB */
}
return (lval);
}
static int
vatpit_update_mode(struct vatpit *vatpit, uint8_t val)
{
struct channel *c;
int sel, rw, mode;
sel = val & TIMER_SEL_MASK;
rw = val & TIMER_RW_MASK;
mode = val & TIMER_MODE_MASK;
if (sel == TIMER_SEL_READBACK)
return (-1);
if (rw != TIMER_LATCH && rw != TIMER_16BIT)
return (-1);
if (rw != TIMER_LATCH) {
/*
* Counter mode is not affected when issuing a
* latch command.
*/
if (mode != TIMER_INTTC &&
mode != TIMER_RATEGEN &&
mode != TIMER_SQWAVE &&
mode != TIMER_SWSTROBE)
return (-1);
}
c = &vatpit->channel[sel >> 6];
if (rw == TIMER_LATCH)
pit_update_counter(vatpit, c, true);
else {
c->mode = mode;
c->olbyte = 0; /* reset latch after reprogramming */
}
return (0);
}
static int
vatpit_get_out(struct vatpit *vatpit, int channel)
{
struct channel *c;
sbintime_t delta_ticks;
int out;
c = &vatpit->channel[channel];
switch (c->mode) {
case TIMER_INTTC:
delta_ticks = (sbinuptime() - c->now_sbt) / vatpit->freq_sbt;
out = ((c->initial - delta_ticks) <= 0);
break;
default:
out = 0;
break;
}
return (out);
}
int
vatpit_handler(void *vm, int vcpuid, struct vm_exit *vmexit)
{
struct vatpit *vatpit;
struct channel *c;
int port;
uint8_t val;
int error;
vatpit = vm_atpit(vm);
if (vmexit->u.inout.bytes != 1)
return (-1);
val = vmexit->u.inout.eax;
port = vmexit->u.inout.port;
if (port == TIMER_MODE) {
if (vmexit->u.inout.in != 0) {
VATPIT_CTR0(vatpit, "vatpit attempt to read mode");
return (-1);
}
VATPIT_LOCK(vatpit);
error = vatpit_update_mode(vatpit, val);
VATPIT_UNLOCK(vatpit);
return (error);
}
/* counter ports */
KASSERT(port >= TIMER_CNTR0 && vmexit->u.inout.port <= TIMER_CNTR2,
("invalid port 0x%x", port));
c = &vatpit->channel[port - TIMER_CNTR0];
VATPIT_LOCK(vatpit);
if (vmexit->u.inout.in) {
/*
* The spec says that once the output latch is completely
* read it should revert to "following" the counter. Use
* the free running counter for this case (i.e. Linux
* TSC calibration). Assuming the access mode is 16-bit,
* toggle the MSB/LSB bit on each read.
*/
if (c->olbyte == 0) {
uint16_t tmp;
tmp = pit_update_counter(vatpit, c, false);
if (c->frbyte)
tmp >>= 8;
tmp &= 0xff;
vmexit->u.inout.eax = tmp;
c->frbyte ^= 1;
} else
vmexit->u.inout.eax = c->ol[--c->olbyte];
} else {
c->cr[c->crbyte++] = vmexit->u.inout.eax;
if (c->crbyte == 2) {
c->frbyte = 0;
c->crbyte = 0;
c->initial = c->cr[0] | (uint16_t)c->cr[1] << 8;
c->now_sbt = sbinuptime();
/* Start an interval timer for channel 0 */
if (port == TIMER_CNTR0) {
c->callout_sbt = c->now_sbt;
pit_timer_start_cntr0(vatpit);
}
if (c->initial == 0)
c->initial = 0xffff;
}
}
VATPIT_UNLOCK(vatpit);
return (0);
}
struct vatpit *
vatpit_init(struct vm *vm)
{
struct vatpit *vatpit;
struct bintime bt;
struct vatpit_callout_arg *arg;
int i;
vatpit = malloc(sizeof(struct vatpit), M_VATPIT, M_WAITOK | M_ZERO);
vatpit->vm = vm;
mtx_init(&vatpit->mtx, "vatpit lock", NULL, MTX_SPIN);
FREQ2BT(PIT_8254_FREQ, &bt);
vatpit->freq_sbt = bttosbt(bt);
for (i = 0; i < 3; i++) {
callout_init(&vatpit->channel[i].callout, true);
arg = &vatpit->channel[i].callout_arg;
arg->vatpit = vatpit;
arg->channel_num = i;
}
return (vatpit);
}
void
vatpit_cleanup(struct vatpit *vatpit)
{
int i;
for (i = 0; i < 3; i++)
callout_drain(&vatpit->channel[i].callout);
free(vatpit, M_VATPIT);
}

View file

@ -1,4 +1,5 @@
/*-
* Copyright (c) 2014 Tycho Nightingale <tycho.nightingale@pluribusnetworks.com>
* Copyright (c) 2011 NetApp, Inc.
* All rights reserved.
*
@ -26,20 +27,14 @@
* $FreeBSD$
*/
#ifndef _PIT_8254_H_
#define _PIT_8254_H_
#ifndef _VATPIT_H_
#define _VATPIT_H_
/*
* Borrowed from amd64/include/timerreg.h because in that file it is
* conditionally compiled for #ifdef _KERNEL only.
*/
#include <machine/timerreg.h>
#include <dev/ic/i8253reg.h>
struct vatpit *vatpit_init(struct vm *vm);
void vatpit_cleanup(struct vatpit *vatpit);
#define IO_TIMER1 0x40 /* 8253 Timer #1 */
#define TIMER_CNTR0 (IO_TIMER1 + TIMER_REG_CNTR0)
#define TIMER_CNTR1 (IO_TIMER1 + TIMER_REG_CNTR1)
#define TIMER_CNTR2 (IO_TIMER1 + TIMER_REG_CNTR2)
#define TIMER_MODE (IO_TIMER1 + TIMER_REG_MODE)
int vatpit_handler(void *vm, int vcpuid, struct vm_exit *vmexit);
#endif /* _PIT_8254_H_ */
#endif /* _VATPIT_H_ */

View file

@ -68,6 +68,7 @@ __FBSDID("$FreeBSD$");
#include "vmm_mem.h"
#include "vmm_util.h"
#include "vatpic.h"
#include "vatpit.h"
#include "vhpet.h"
#include "vioapic.h"
#include "vlapic.h"
@ -119,6 +120,7 @@ struct vm {
struct vhpet *vhpet; /* virtual HPET */
struct vioapic *vioapic; /* virtual ioapic */
struct vatpic *vatpic; /* virtual atpic */
struct vatpit *vatpit; /* virtual atpit */
struct vmspace *vmspace; /* guest's address space */
struct vcpu vcpu[VM_MAXCPU];
int num_mem_segs;
@ -349,6 +351,7 @@ vm_create(const char *name, struct vm **retvm)
vm->vioapic = vioapic_init(vm);
vm->vhpet = vhpet_init(vm);
vm->vatpic = vatpic_init(vm);
vm->vatpit = vatpit_init(vm);
for (i = 0; i < VM_MAXCPU; i++) {
vcpu_init(vm, i);
@ -381,6 +384,7 @@ vm_destroy(struct vm *vm)
if (vm->iommu != NULL)
iommu_destroy_domain(vm->iommu);
vatpit_cleanup(vm->vatpit);
vhpet_cleanup(vm->vhpet);
vatpic_cleanup(vm->vatpic);
vioapic_cleanup(vm->vioapic);
@ -1740,3 +1744,9 @@ vm_atpic(struct vm *vm)
{
return (vm->vatpic);
}
struct vatpit *
vm_atpit(struct vm *vm)
{
return (vm->vatpit);
}

View file

@ -36,11 +36,16 @@ __FBSDID("$FreeBSD$");
#include <machine/vmm.h>
#include "vatpic.h"
#include "vatpit.h"
#include "vmm_ioport.h"
#define MAX_IOPORTS 1280
ioport_handler_func_t ioport_handler[MAX_IOPORTS] = {
[TIMER_MODE] = vatpit_handler,
[TIMER_CNTR0] = vatpit_handler,
[TIMER_CNTR1] = vatpit_handler,
[TIMER_CNTR2] = vatpit_handler,
[IO_ICU1] = vatpic_master_handler,
[IO_ICU1 + ICU_IMR_OFFSET] = vatpic_master_handler,
[IO_ICU2] = vatpic_slave_handler,

View file

@ -29,6 +29,7 @@ SRCS+= vmm.c \
SRCS+= iommu.c \
ppt.c \
vatpic.c \
vatpit.c \
vhpet.c \
vioapic.c \
vlapic.c

View file

@ -27,7 +27,6 @@ SRCS= \
pci_virtio_block.c \
pci_virtio_net.c \
pci_uart.c \
pit_8254.c \
pm.c \
pmtmr.c \
post.c \

View file

@ -1,292 +0,0 @@
/*-
* Copyright (c) 2011 NetApp, Inc.
* 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 NETAPP, INC ``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 NETAPP, INC 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/types.h>
#include <sys/time.h>
#include <machine/vmm.h>
#include <machine/clock.h>
#include <assert.h>
#include <stdio.h>
#include <string.h>
#include <unistd.h>
#include <vmmapi.h>
#include "acpi.h"
#include "bhyverun.h"
#include "inout.h"
#include "mevent.h"
#include "pci_lpc.h"
#include "pit_8254.h"
#define TIMER_SEL_MASK 0xc0
#define TIMER_RW_MASK 0x30
#define TIMER_MODE_MASK 0x0f
#define TIMER_SEL_READBACK 0xc0
#define TIMER_DIV(freq, hz) (((freq) + (hz) / 2) / (hz))
#define PIT_8254_FREQ 1193182
static const int nsecs_per_tick = 1000000000 / PIT_8254_FREQ;
struct counter {
struct vmctx *ctx;
struct mevent *tevp;
struct timeval tv; /* uptime when counter was loaded */
int mode;
uint16_t initial; /* initial counter value */
uint8_t cr[2];
uint8_t ol[2];
int crbyte;
int olbyte;
int frbyte;
};
static void
timevalfix(struct timeval *t1)
{
if (t1->tv_usec < 0) {
t1->tv_sec--;
t1->tv_usec += 1000000;
}
if (t1->tv_usec >= 1000000) {
t1->tv_sec++;
t1->tv_usec -= 1000000;
}
}
static void
timevalsub(struct timeval *t1, const struct timeval *t2)
{
t1->tv_sec -= t2->tv_sec;
t1->tv_usec -= t2->tv_usec;
timevalfix(t1);
}
static uint64_t pit_mev_count;
static void
pit_mevent_cb(int fd, enum ev_type type, void *param)
{
struct counter *c;
c = param;
pit_mev_count++;
vm_isa_pulse_irq(c->ctx, 0, 2);
/*
* Delete the timer for one-shots
*/
if (c->mode != TIMER_RATEGEN) {
mevent_delete(c->tevp);
c->tevp = NULL;
}
}
static void
pit_timer_start(struct vmctx *ctx, struct counter *c)
{
int msecs;
if (c->initial != 0) {
msecs = c->initial * nsecs_per_tick / 1000000;
if (msecs == 0)
msecs = 1;
if (c->tevp == NULL)
c->tevp = mevent_add(msecs, EVF_TIMER, pit_mevent_cb,
c);
}
}
static uint16_t
pit_update_counter(struct counter *c, int latch)
{
struct timeval tv2;
uint16_t lval;
uint64_t delta_nsecs, delta_ticks;
/* cannot latch a new value until the old one has been consumed */
if (latch && c->olbyte != 0)
return (0);
if (c->initial == 0 || c->initial == 1) {
/*
* XXX the program that runs the VM can be stopped and
* restarted at any time. This means that state that was
* created by the guest is destroyed between invocations
* of the program.
*
* If the counter's initial value is not programmed we
* assume a value that would be set to generate 100
* interrupts per second.
*/
c->initial = TIMER_DIV(PIT_8254_FREQ, 100);
gettimeofday(&c->tv, NULL);
}
(void)gettimeofday(&tv2, NULL);
timevalsub(&tv2, &c->tv);
delta_nsecs = tv2.tv_sec * 1000000000 + tv2.tv_usec * 1000;
delta_ticks = delta_nsecs / nsecs_per_tick;
lval = c->initial - delta_ticks % c->initial;
if (latch) {
c->olbyte = 2;
c->ol[1] = lval; /* LSB */
c->ol[0] = lval >> 8; /* MSB */
}
return (lval);
}
static int
pit_8254_handler(struct vmctx *ctx, int vcpu, int in, int port, int bytes,
uint32_t *eax, void *arg)
{
int sel, rw, mode;
uint8_t val;
struct counter *c;
static struct counter counter[3];
if (bytes != 1)
return (-1);
val = *eax;
if (port == TIMER_MODE) {
assert(in == 0);
sel = val & TIMER_SEL_MASK;
rw = val & TIMER_RW_MASK;
mode = val & TIMER_MODE_MASK;
if (sel == TIMER_SEL_READBACK)
return (-1);
if (rw != TIMER_LATCH && rw != TIMER_16BIT)
return (-1);
if (rw != TIMER_LATCH) {
/*
* Counter mode is not affected when issuing a
* latch command.
*/
if (mode != TIMER_INTTC &&
mode != TIMER_RATEGEN &&
mode != TIMER_SQWAVE &&
mode != TIMER_SWSTROBE)
return (-1);
}
c = &counter[sel >> 6];
c->ctx = ctx;
if (rw == TIMER_LATCH)
pit_update_counter(c, 1);
else {
c->mode = mode;
c->olbyte = 0; /* reset latch after reprogramming */
}
return (0);
}
/* counter ports */
assert(port >= TIMER_CNTR0 && port <= TIMER_CNTR2);
c = &counter[port - TIMER_CNTR0];
if (in) {
/*
* The spec says that once the output latch is completely
* read it should revert to "following" the counter. Use
* the free running counter for this case (i.e. Linux
* TSC calibration). Assuming the access mode is 16-bit,
* toggle the MSB/LSB bit on each read.
*/
if (c->olbyte == 0) {
uint16_t tmp;
tmp = pit_update_counter(c, 0);
if (c->frbyte)
tmp >>= 8;
tmp &= 0xff;
*eax = tmp;
c->frbyte ^= 1;
} else
*eax = c->ol[--c->olbyte];
} else {
c->cr[c->crbyte++] = *eax;
if (c->crbyte == 2) {
c->frbyte = 0;
c->crbyte = 0;
c->initial = c->cr[0] | (uint16_t)c->cr[1] << 8;
/* Start an interval timer for counter 0 */
if (port == 0x40)
pit_timer_start(ctx, c);
if (c->initial == 0)
c->initial = 0xffff;
gettimeofday(&c->tv, NULL);
}
}
return (0);
}
INOUT_PORT(8254, TIMER_MODE, IOPORT_F_OUT, pit_8254_handler);
INOUT_PORT(8254, TIMER_CNTR0, IOPORT_F_INOUT, pit_8254_handler);
INOUT_PORT(8254, TIMER_CNTR1, IOPORT_F_INOUT, pit_8254_handler);
INOUT_PORT(8254, TIMER_CNTR2, IOPORT_F_INOUT, pit_8254_handler);
static void
pit_dsdt(void)
{
dsdt_line("");
dsdt_line("Device (TIMR)");
dsdt_line("{");
dsdt_line(" Name (_HID, EisaId (\"PNP0100\"))");
dsdt_line(" Name (_CRS, ResourceTemplate ()");
dsdt_line(" {");
dsdt_indent(2);
dsdt_fixed_ioport(IO_TIMER1, 4);
dsdt_fixed_irq(0);
dsdt_unindent(2);
dsdt_line(" })");
dsdt_line("}");
}
LPC_DSDT(pit_dsdt);