memcg: avoid oom-killing innocent task in case of use_hierarchy
commitae3559fa2fe48ab923b52885bba2dca0355428f0
authorDaisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Tue, 5 Jan 2010 03:26:33 +0000 (5 12:26 +0900)
committerGreg Kroah-Hartman <gregkh@suse.de>
Wed, 6 Jan 2010 22:26:49 +0000 (6 14:26 -0800)
tree6421e0229dac6f291e41d6bd86047de92a8e6a38
parentf1850a5783908f6528c305e321119e7aa6641151
memcg: avoid oom-killing innocent task in case of use_hierarchy

commit d31f56dbf8bafaacb0c617f9a6f137498d5c7aed upstream

task_in_mem_cgroup(), which is called by select_bad_process() to check whether
a task can be a candidate for being oom-killed from memcg's limit, checks
"curr->use_hierarchy"("curr" is the mem_cgroup the task belongs to).

But this check return true(it's false positive) when:

<some path>/00 use_hierarchy == 0 <- hitting limit
  <some path>/00/aa use_hierarchy == 1 <- "curr"

This leads to killing an innocent task in 00/aa. This patch is a fix for this
bug. And this patch also fixes the arg for mem_cgroup_print_oom_info(). We
should print information of mem_cgroup which the task being killed, not current,
belongs to.

Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp>
Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Balbir Singh <balbir@linux.vnet.ibm.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@suse.de>
mm/memcontrol.c
mm/oom_kill.c