memcg: fix Bad page state after replace_page_cache
authorHugh Dickins <hughd@google.com>
Thu, 19 Apr 2012 06:34:46 +0000 (23:34 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 19 Apr 2012 06:40:57 +0000 (23:40 -0700)
commit9b7f43afd417a6feb80841d30ced4051c362eb5d
tree695f97a4d028a3463c8e7edaa4c6d410a941d0bb
parent932e9f352b5d685725076f21b237f7c7d804b29c
memcg: fix Bad page state after replace_page_cache

My 9ce70c0240d0 "memcg: fix deadlock by inverting lrucare nesting" put a
nasty little bug into v3.3's version of mem_cgroup_replace_page_cache(),
sometimes used for FUSE.  Replacing __mem_cgroup_commit_charge_lrucare()
by __mem_cgroup_commit_charge(), I used the "pc" pointer set up earlier:
but it's for oldpage, and needs now to be for newpage.  Once oldpage was
freed, its PageCgroupUsed bit (cleared above but set again here) caused
"Bad page state" messages - and perhaps worse, being missed from newpage.
(I didn't find this by using FUSE, but in reusing the function for tmpfs.)

Signed-off-by: Hugh Dickins <hughd@google.com>
Cc: stable@vger.kernel.org [v3.3 only]
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/memcontrol.c