cgroup: remove redundant cleanup in css_create
authorWenwei Tao <ww.tao0320@gmail.com>
Fri, 13 May 2016 14:59:20 +0000 (22:59 +0800)
committerSasha Levin <sasha.levin@oracle.com>
Mon, 11 Jul 2016 00:19:50 +0000 (20:19 -0400)
commit6ef304587cf41e19312fa9fd81448ea6880ad475
tree9ff90ddf2cfecdcd6d28eff578de3f03590afbc8
parent62f7175fa64005be2e090618da5a20f87752704d
cgroup: remove redundant cleanup in css_create

[ Upstream commit b00c52dae6d9ee8d0f2407118ef6544ae5524781 ]

When create css failed, before call css_free_rcu_fn, we remove the css
id and exit the percpu_ref, but we will do these again in
css_free_work_fn, so they are redundant.  Especially the css id, that
would cause problem if we remove it twice, since it may be assigned to
another css after the first remove.

tj: This was broken by two commits updating the free path without
    synchronizing the creation failure path.  This can be easily
    triggered by trying to create more than 64k memory cgroups.

Signed-off-by: Wenwei Tao <ww.tao0320@gmail.com>
Signed-off-by: Tejun Heo <tj@kernel.org>
Cc: Vladimir Davydov <vdavydov@parallels.com>
Fixes: 9a1049da9bd2 ("percpu-refcount: require percpu_ref to be exited explicitly")
Fixes: 01e586598b22 ("cgroup: release css->id after css_free")
Cc: stable@vger.kernel.org # v3.17+
Signed-off-by: Sasha Levin <sasha.levin@oracle.com>
kernel/cgroup.c