mm: memcg: update the correct soft limit tree during migration
authorJohannes Weiner <hannes@cmpxchg.org>
Fri, 20 Jan 2012 22:34:12 +0000 (14:34 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Mon, 23 Jan 2012 16:38:48 +0000 (08:38 -0800)
commitfea4b9558cdace84b56cea3cbc29cb7a90559d12
tree5aef7ff51027d8986113025b3e9b0255476772dd
parent0a2b6f54b75d585ddd31e01603d8cb9bc939d2ce
mm: memcg: update the correct soft limit tree during migration

end_migration() passes the old page instead of the new page to commit
the charge.  This page descriptor is not used for committing itself,
though, since we also pass the (correct) page_cgroup descriptor.  But
it's used to find the soft limit tree through the page's zone, so the
soft limit tree of the old page's zone is updated instead of that of the
new page's, which might get slightly out of date until the next charge
reaches the ratelimit point.

This glitch has been present since 74f9bae ("memcg: condense
page_cgroup-to-page lookup points").

This fixes a bug that I introduced in 2.6.38.  It's benign enough (to my
knowledge) that we probably don't want this for stable.

Reported-by: Hugh Dickins <hughd@google.com>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Acked-by: Michal Hocko <mhocko@suse.cz>
Acked-by: Kirill A. Shutemov <kirill@shutemov.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memcontrol.c