memcg: avoid percpu cached charge draining at softlimit
authorKAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Wed, 15 Jun 2011 22:08:46 +0000 (15:08 -0700)
committerLinus Torvalds <torvalds@linux-foundation.org>
Thu, 16 Jun 2011 03:04:01 +0000 (20:04 -0700)
Based on Michal Hocko's comment.

We are not draining per cpu cached charges during soft limit reclaim
because background reclaim doesn't care about charges.  It tries to free
some memory and charges will not give any.

Cached charges might influence only selection of the biggest soft limit
offender but as the call is done only after the selection has been already
done it makes no change.

Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Reviewed-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 c39a177bb641244f9b6b6a6789254012a7564a3c..cf7d027a8844b115bcc6d213264707d220c6c3d9 100644 (file)
@@ -1670,7 +1670,13 @@ static int mem_cgroup_hierarchical_reclaim(struct mem_cgroup *root_mem,
                victim = mem_cgroup_select_victim(root_mem);
                if (victim == root_mem) {
                        loop++;
-                       if (loop >= 1)
+                       /*
+                        * We are not draining per cpu cached charges during
+                        * soft limit reclaim  because global reclaim doesn't
+                        * care about charges. It tries to free some memory and
+                        * charges will not give any.
+                        */
+                       if (!check_soft && loop >= 1)
                                drain_all_stock_async(root_mem);
                        if (loop >= 2) {
                                /*