upstream/opnsense-src
1
0
Fork 0
mirror of https://github.com/opnsense/src.git synced 2026-06-11 01:30:30 -04:00
Code Issues Projects Releases Packages Wiki Activity Actions

In linux emulation layer try to detect attempt to use linux_clone() to

Browse source 
create kernel threads and call rfork(2) with RFTHREAD flag set in this case,
which puts parent and child into the same threading group. As a result
all threads that belong to the same program end up in the same threading
group.

This is similar to what linuxthreads port does, though in this case we don't
have a luxury of having access to the source code and there is no definite
way to differentiate linux_clone() called for threading purposes from other
uses, so that we have to resort to heuristics.

Allow SIGTHR to be delivered between all processes in the same threading
group previously it has been blocked for s[ug]id processes.

This also should improve locking of the same file descriptor from different
threads in programs running under linux compat layer.

PR:			kern/72922
Reported by:		Andriy Gapon <avg@icyb.net.ua>
Idea suggested by:	rwatson
This commit is contained in:
Maxim Sobolev 2005-03-03 16:57:55 +00:00
parent 24fe1eafe4
commit 4b1783363f
2 changed files with 31 additions and 0 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

19
sys/i386/linux/linux_machdep.c
View file

@ -314,6 +314,9 @@ linux_vfork(struct thread *td, struct linux_vfork_args *args)
#define CLONE_FILES 0x400
#define CLONE_SIGHAND 0x800
#define CLONE_PID 0x1000
#define CLONE_THREAD 0x10000
#define THREADING_FLAGS (CLONE_VM | CLONE_FS | CLONE_FILES | CLONE_SIGHAND)
int
linux_clone(struct thread *td, struct linux_clone_args *args)
@ -349,6 +352,22 @@ linux_clone(struct thread *td, struct linux_clone_args *args)
if (!(args->flags & CLONE_FILES))
ff |= RFFDG;
/*
* Attempt to detect when linux_clone(2) is used for creating
* kernel threads. Unfortunately despite the existence of the
* CLONE_THREAD flag, version of linuxthreads package used in
* most popular distros as of beginning of 2005 doesn't make
* any use of it. Therefore, this detection relay fully on
* empirical observation that linuxthreads sets certain
* combination of flags, so that we can make more or less
* precise detection and notify the FreeBSD kernel that several
* processes are in fact part of the same threading group, so
* that special treatment is necessary for signal delivery
* between those processes and fd locking.
*/
if ((args->flags & 0xffffff00) == THREADING_FLAGS)
ff |= RFTHREAD;
error = fork1(td, ff, 0, &p2);
if (error)
return (error);

12
sys/kern/kern_prot.c
View file

@ -1523,6 +1523,18 @@ p_cansignal(struct thread *td, struct proc *p, int signum)
/* XXX: This will require an additional lock of some sort. */
if (signum == SIGCONT && td->td_proc->p_session == p->p_session)
return (0);
/*
* Some compat layers use SIGTHR for communications between
* different kernel threads of the same process, so that
* they are expecting that it's always possible to deliver
* it, even for suid applications where cr_cansignal() can
* deny such ability for security consideration. It should be
* pretty safe to do since the only way to create two processes
* with the same p_leader is via rfork(2).
*/
if (signum == SIGTHR && td->td_proc->p_leader != NULL &&
td->td_proc->p_leader == p->p_leader)
return (0);
return (cr_cansignal(td->td_ucred, p, signum));
}