memcg: fix use_hierarchy css_is_ancestor oops regression
authorHugh Dickins <hughd@google.com>
Wed, 20 Jun 2012 19:52:58 +0000 (12:52 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 20 Jun 2012 21:39:35 +0000 (14:39 -0700)
If use_hierarchy is set, reclaim testing soon oopses in css_is_ancestor()
called from __mem_cgroup_same_or_subtree() called from page_referenced():
when processes are exiting, it's easy for mm_match_cgroup() to pass along
a NULL memcg coming from a NULL mm->owner.

Check for that in __mem_cgroup_same_or_subtree().  Return true or false?
False because we cannot know if it was in the hierarchy, but also false
because it's better not to count a reference from an exiting process.

Signed-off-by: Hugh Dickins <hughd@google.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Konstantin Khlebnikov <khlebnikov@openvz.org>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memcontrol.c

index ac35bccadb7b9f53606d445a961e442e891aa94a..f8517cd28feb3f9732276041e8e672897dbb0af7 100644 (file)
@@ -1148,7 +1148,7 @@ bool __mem_cgroup_same_or_subtree(const struct mem_cgroup *root_memcg,
 {
        if (root_memcg == memcg)
                return true;
-       if (!root_memcg->use_hierarchy)
+       if (!root_memcg->use_hierarchy || !memcg)
                return false;
        return css_is_ancestor(&memcg->css, &root_memcg->css);
 }