From: Vladimir Davydov Date: Tue, 21 Jan 2014 23:49:41 +0000 (-0800) Subject: memcg: fix kmem_account_flags check in memcg_can_account_kmem() X-Git-Tag: v4.9.8~7095^2~58 X-Git-Url: http://review.tizen.org/git/?a=commitdiff_plain;h=1c98dd905ddb7552f13a3f06aa0bd9ef6affeeb7;p=platform%2Fkernel%2Flinux-rpi3.git memcg: fix kmem_account_flags check in memcg_can_account_kmem() We should start kmem accounting for a memory cgroup only after both its kmem limit is set (KMEM_ACCOUNTED_ACTIVE) and related call sites are patched (KMEM_ACCOUNTED_ACTIVATED). Currently memcg_can_account_kmem() allows kmem accounting even if only one of the conditions is true. Fix it. This means that a page might get charged by memcg_kmem_newpage_charge which would see its static key patched already but memcg_kmem_commit_charge would still see it unpatched and so the charge won't be committed. The result would be charge inconsistency (page_cgroup not marked as PageCgroupUsed) and the charge would leak because __memcg_kmem_uncharge_pages would ignore it. [mhocko@suse.cz: augment changelog] Signed-off-by: Vladimir Davydov Cc: Johannes Weiner Acked-by: Michal Hocko Cc: Balbir Singh Cc: KAMEZAWA Hiroyuki Cc: Glauber Costa Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds --- diff --git a/mm/memcontrol.c b/mm/memcontrol.c index 7f1a356..3065fa8 100644 --- a/mm/memcontrol.c +++ b/mm/memcontrol.c @@ -2959,7 +2959,8 @@ static DEFINE_MUTEX(set_limit_mutex); static inline bool memcg_can_account_kmem(struct mem_cgroup *memcg) { return !mem_cgroup_disabled() && !mem_cgroup_is_root(memcg) && - (memcg->kmem_account_flags & KMEM_ACCOUNTED_MASK); + (memcg->kmem_account_flags & KMEM_ACCOUNTED_MASK) == + KMEM_ACCOUNTED_MASK; } /*