]> git.baikalelectronics.ru Git - kernel.git/commitdiff
mm/vmpressure.c: fix subtree pressure detection
authorVladimir Davydov <vdavydov@virtuozzo.com>
Wed, 3 Feb 2016 00:57:49 +0000 (16:57 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Wed, 3 Feb 2016 16:28:43 +0000 (08:28 -0800)
When vmpressure is called for the entire subtree under pressure we
mistakenly use vmpressure->scanned instead of vmpressure->tree_scanned
when checking if vmpressure work is to be scheduled.  This results in
suppressing all vmpressure events in the legacy cgroup hierarchy.  Fix it.

Fixes: b1d4ae24983c ("mm: memcontrol: hook up vmpressure to socket pressure")
Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Acked-by: Michal Hocko <mhocko@suse.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
mm/vmpressure.c

index 9a6c0704211c856c300f67057fc202e25c9657ab..149fdf6c5c56f927f3613538c61b3c5831c80af9 100644 (file)
@@ -248,9 +248,8 @@ void vmpressure(gfp_t gfp, struct mem_cgroup *memcg, bool tree,
 
        if (tree) {
                spin_lock(&vmpr->sr_lock);
-               vmpr->tree_scanned += scanned;
+               scanned = vmpr->tree_scanned += scanned;
                vmpr->tree_reclaimed += reclaimed;
-               scanned = vmpr->scanned;
                spin_unlock(&vmpr->sr_lock);
 
                if (scanned < vmpressure_win)