[BUGFIX][PATCH] memcg: fix oom killer kills a task in other cgroup

[KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>]

Please take this patch in different context with recent discussion.
This is a quick-fix for a terrible bug.

This patch itself is against mmotm but can be easily applied to mainline or
stable tree, I think. (But I don't CC stable tree until I get ack.)

==
Now, oom-killer kills process's chidlren at first. But this means
a child in other cgroup can be killed. But it's not checked now.

This patch fixes that.

CC: Balbir Singh <balbir@linux.vnet.ibm.com>
CC: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
---
 mm/oom_kill.c |    3 +++
 1 file changed, 3 insertions(+)

Index: mmotm-2.6.33-Feb03/mm/oom_kill.c
===================================================================
--- mmotm-2.6.33-Feb03.orig/mm/oom_kill.c
+++ mmotm-2.6.33-Feb03/mm/oom_kill.c
@@ -459,6 +459,9 @@ static int oom_kill_process(struct task_
 	list_for_each_entry(c, &p->children, sibling) {
 		if (c->mm == p->mm)
 			continue;
+		/* Children may be in other cgroup */
+		if (mem && !task_in_mem_cgroup(c, mem))
+			continue;
 		if (!oom_kill_task(c))
 			return 0;
 	}

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>