[PATCH] cgroups: fix 2.6.32 regression causing BUG_ON() in cgroup_diput()

[PATCH] cgroups: fix 2.6.32 regression causing BUG_ON() in cgroup_diput()

[Dave Anderson]

[-- Attachment #1: Type: text/plain, Size: 1591 bytes --]


The LTP cgroup test suite generates a "kernel BUG at kernel/cgroup.c:790!"
here in cgroup_diput():

                 /*
                  * if we're getting rid of the cgroup, refcount should ensure
                  * that there are no pidlists left.
                  */
                 BUG_ON(!list_empty(&cgrp->pidlists));

The cgroup pidlist rework in 2.6.32 generates the BUG_ON, which is caused
when pidlist_array_load() calls cgroup_pidlist_find():

(1) if a matching cgroup_pidlist is found, it down_write's the mutex of the
     pre-existing cgroup_pidlist, and increments its use_count.
(2) if no matching cgroup_pidlist is found, then a new one is allocated, it
     down_write's its mutex, and the use_count is set to 0.
(3) the matching, or new, cgroup_pidlist gets returned back to pidlist_array_load(),
     which increments its use_count -- regardless whether new or pre-existing --
     and up_write's the mutex.

So if a matching list is ever encountered by cgroup_pidlist_find() during
the life of a cgroup directory, it results in an inflated use_count value,
preventing it from ever getting released by cgroup_release_pid_array().
Then if the directory is subsequently removed, cgroup_diput() hits the
BUG_ON() when it finds that the directory's cgroup is still populated
with a pidlist.

The patch simply removes the use_count increment when a matching
pidlist is found by cgroup_pidlist_find(), because it gets bumped by
the calling pidlist_array_load() function while still protected by the
list's mutex.

Signed-off-by: Dave Anderson <anderson@redhat.com>

---



[-- Attachment #2: cgroup.patch --]
[-- Type: text/x-patch, Size: 301 bytes --]

--- linux-2.6-git/kernel/cgroup.c.orig
+++ linux-2.6-git/kernel/cgroup.c
@@ -2468,7 +2468,6 @@ static struct cgroup_pidlist *cgroup_pid
 			/* make sure l doesn't vanish out from under us */
 			down_write(&l->mutex);
 			mutex_unlock(&cgrp->pidlist_mutex);
-			l->use_count++;
 			return l;
 		}
 	}

Re: [PATCH] cgroups: fix 2.6.32 regression causing BUG_ON() in cgroup_diput()

[Li Zefan]

CC: Andrew
CC: Container list

Dave Anderson wrote:
> 
> The LTP cgroup test suite generates a "kernel BUG at kernel/cgroup.c:790!"
> here in cgroup_diput():
> 
>                 /*
>                  * if we're getting rid of the cgroup, refcount should
> ensure
>                  * that there are no pidlists left.
>                  */
>                 BUG_ON(!list_empty(&cgrp->pidlists));
> 

Good catch. Thanks!

This BUG can be easily triggered if 2 threads are reading the same cgroup's
tasks file at the same time, and then the cgroup gets removed.

And this patch needs to be added to 2.6.32.x.

> The cgroup pidlist rework in 2.6.32 generates the BUG_ON, which is caused
> when pidlist_array_load() calls cgroup_pidlist_find():
> 
> (1) if a matching cgroup_pidlist is found, it down_write's the mutex of the
>     pre-existing cgroup_pidlist, and increments its use_count.
> (2) if no matching cgroup_pidlist is found, then a new one is allocated, it
>     down_write's its mutex, and the use_count is set to 0.
> (3) the matching, or new, cgroup_pidlist gets returned back to
> pidlist_array_load(),
>     which increments its use_count -- regardless whether new or
> pre-existing --
>     and up_write's the mutex.
> 
> So if a matching list is ever encountered by cgroup_pidlist_find() during
> the life of a cgroup directory, it results in an inflated use_count value,
> preventing it from ever getting released by cgroup_release_pid_array().
> Then if the directory is subsequently removed, cgroup_diput() hits the
> BUG_ON() when it finds that the directory's cgroup is still populated
> with a pidlist.
> 
> The patch simply removes the use_count increment when a matching
> pidlist is found by cgroup_pidlist_find(), because it gets bumped by
> the calling pidlist_array_load() function while still protected by the
> list's mutex.
> 
> Signed-off-by: Dave Anderson <anderson@redhat.com>
> 

Reviewed-by: Li Zefan <lizf@cn.fujitsu.com>

Re: [PATCH] cgroups: fix 2.6.32 regression causing BUG_ON() in cgroup_diput()

[Ben Blum]

On Wed, Dec 23, 2009 at 01:48:42PM -0500, Dave Anderson wrote:
> 
> The LTP cgroup test suite generates a "kernel BUG at kernel/cgroup.c:790!"
> here in cgroup_diput():
> 
>                 /*
>                  * if we're getting rid of the cgroup, refcount should 
>                  ensure
>                  * that there are no pidlists left.
>                  */
>                 BUG_ON(!list_empty(&cgrp->pidlists));
> 
> The cgroup pidlist rework in 2.6.32 generates the BUG_ON, which is caused
> when pidlist_array_load() calls cgroup_pidlist_find():
> 
> (1) if a matching cgroup_pidlist is found, it down_write's the mutex of the
>     pre-existing cgroup_pidlist, and increments its use_count.
> (2) if no matching cgroup_pidlist is found, then a new one is allocated, it
>     down_write's its mutex, and the use_count is set to 0.
> (3) the matching, or new, cgroup_pidlist gets returned back to 
> pidlist_array_load(),
>     which increments its use_count -- regardless whether new or 
>     pre-existing --
>     and up_write's the mutex.
> 
> So if a matching list is ever encountered by cgroup_pidlist_find() during
> the life of a cgroup directory, it results in an inflated use_count value,
> preventing it from ever getting released by cgroup_release_pid_array().
> Then if the directory is subsequently removed, cgroup_diput() hits the
> BUG_ON() when it finds that the directory's cgroup is still populated
> with a pidlist.
> 
> The patch simply removes the use_count increment when a matching
> pidlist is found by cgroup_pidlist_find(), because it gets bumped by
> the calling pidlist_array_load() function while still protected by the
> list's mutex.
> 
> Signed-off-by: Dave Anderson <anderson@redhat.com>
> 
> ---
> 
> 

Ack! That was probably my fault. Good catch.