[PATCH] Fix race between attach_task and cpuset_exit

[Srivatsa Vaddagiri <vatsa@in.ibm.com>]

On Wed, Apr 04, 2007 at 10:55:01PM -0700, Paul Menage wrote:
> Shouldn't we just put a task_lock()/task_unlock() around these lines
> and leave everything else as-is?
> 
> 	task_lock(tsk);
> 	cs = tsk->cpuset;
> 	tsk->cpuset = &top_cpuset;	/* the_top_cpuset_hack - see above */
> 	task_unlock(tsk)

Andrew,
	Can you drop fix-race-between-attach_task-and-cpuset_exit.patch
and take this fix instead, which addresses some points raised by Paul
Menage?





Currently cpuset_exit() changes the exiting task's ->cpuset pointer w/o
taking task_lock().  This can lead to ugly races between attach_task and
cpuset_exit.  Details of the races are described at
http://lkml.org/lkml/2007/3/24/132.

Patch below closes those races.  It is against 2.6.21-rc6-mm1 and has
undergone a simple compile/boot test on a x86_64 box.

Signed-off-by : Srivatsa Vaddagiri <vatsa@in.ibm.com>


---


diff -puN kernel/cpuset.c~cpuset_race_fix kernel/cpuset.c
--- linux-2.6.21-rc6/kernel/cpuset.c~cpuset_race_fix	2007-04-10 20:53:57.000000000 +0530
+++ linux-2.6.21-rc6-vatsa/kernel/cpuset.c	2007-04-10 22:08:46.000000000 +0530
@@ -2119,10 +2119,6 @@ void cpuset_fork(struct task_struct *chi
  * it is holding that mutex while calling check_for_release(),
  * which calls kmalloc(), so can't be called holding callback_mutex().
  *
- * We don't need to task_lock() this reference to tsk->cpuset,
- * because tsk is already marked PF_EXITING, so attach_task() won't
- * mess with it, or task is a failed fork, never visible to attach_task.
- *
  * the_top_cpuset_hack:
  *
  *    Set the exiting tasks cpuset to the root cpuset (top_cpuset).
@@ -2161,8 +2157,10 @@ void cpuset_exit(struct task_struct *tsk
 {
 	struct cpuset *cs;
 
+	task_lock(current);
 	cs = tsk->cpuset;
 	tsk->cpuset = &top_cpuset;	/* the_top_cpuset_hack - see above */
+	task_unlock(current);
 
 	if (notify_on_release(cs)) {
 		char *pathbuf = NULL;
_

-- 
Regards,
vatsa