memcg: add RCU locking around css_for_each_descendant_pre() in memcg_…

…offline_kmem()

memcg_offline_kmem() may be called from memcg_free_kmem() after a css init
failure.  memcg_free_kmem() is a ->css_free callback which is called
without cgroup_mutex and memcg_offline_kmem() ends up using
css_for_each_descendant_pre() without any locking.  Fix it by adding rcu
read locking around it.

 mkdir: cannot create directory `65530': No space left on device
 [  527.241361] ===============================
 [  527.241845] [ INFO: suspicious RCU usage. ]
 [  527.242367] 4.6.0-work+ torvalds#321 Not tainted
 [  527.242730] -------------------------------
 [  527.243220] kernel/cgroup.c:4008 cgroup_mutex or RCU read lock required!
 [  527.243970]
 [  527.243970] other info that might help us debug this:
 [  527.243970]
 [  527.244715]
 [  527.244715] rcu_scheduler_active = 1, debug_locks = 0
 [  527.245463] 2 locks held by kworker/0:5/1664:
 [  527.245939]  #0:  ("cgroup_destroy"){.+.+..}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0
 [  527.246958]  #1:  ((&css->destroy_work)#3){+.+...}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0
 [  527.248098]
 [  527.248098] stack backtrace:
 [  527.249565] CPU: 0 PID: 1664 Comm: kworker/0:5 Not tainted 4.6.0-work+ torvalds#321
 [  527.250429] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.1-1.fc24 04/01/2014
 [  527.250555] Workqueue: cgroup_destroy css_free_work_fn
 [  527.250555]  0000000000000000 ffff880178747c68 ffffffff8128bfc7 ffff880178b8ac40
 [  527.250555]  0000000000000001 ffff880178747c98 ffffffff8108c297 0000000000000000
 [  527.250555]  ffff88010de54138 000000000000fffb ffff88010de537e8 ffff880178747cc0
 [  527.250555] Call Trace:
 [  527.250555]  [<ffffffff8128bfc7>] dump_stack+0x68/0xa1
 [  527.250555]  [<ffffffff8108c297>] lockdep_rcu_suspicious+0xd7/0x110
 [  527.250555]  [<ffffffff810ca03d>] css_next_descendant_pre+0x7d/0xb0
 [  527.250555]  [<ffffffff8114d14a>] memcg_offline_kmem.part.44+0x4a/0xc0
 [  527.250555]  [<ffffffff8114d3ac>] mem_cgroup_css_free+0x1ec/0x200
 [  527.250555]  [<ffffffff810ccdc9>] css_free_work_fn+0x49/0x5e0
 [  527.250555]  [<ffffffff81060b15>] process_one_work+0x1c5/0x4a0
 [  527.250555]  [<ffffffff81060ab5>] ? process_one_work+0x165/0x4a0
 [  527.250555]  [<ffffffff81060e39>] worker_thread+0x49/0x490
 [  527.250555]  [<ffffffff81060df0>] ? process_one_work+0x4a0/0x4a0
 [  527.250555]  [<ffffffff81060df0>] ? process_one_work+0x4a0/0x4a0
 [  527.250555]  [<ffffffff810672ba>] kthread+0xea/0x100
 [  527.250555]  [<ffffffff814cbcff>] ret_from_fork+0x1f/0x40
 [  527.250555]  [<ffffffff810671d0>] ? kthread_create_on_node+0x200/0x200

Link: http://lkml.kernel.org/r/[email protected]
Signed-off-by: Tejun Heo <[email protected]>
Acked-by: Vladimir Davydov <[email protected]>
Acked-by: Johannes Weiner <[email protected]>
Cc: Michal Hocko <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>

mm/memcontrol.c

@@ -2896,13 +2896,16 @@ static void memcg_offline_kmem(struct mem_cgroup *memcg)
 	 * ordering is imposed by list_lru_node->lock taken by
 	 * memcg_drain_all_list_lrus().
 	 */
+	rcu_read_lock(); /* can be called from css_free w/o cgroup_mutex */
 	css_for_each_descendant_pre(css, &memcg->css) {
 		child = mem_cgroup_from_css(css);
 		BUG_ON(child->kmemcg_id != kmemcg_id);
 		child->kmemcg_id = parent->kmemcg_id;
 		if (!memcg->use_hierarchy)
 			break;
 	}
+	rcu_read_unlock();
+
 	memcg_drain_all_list_lrus(kmemcg_id, parent->kmemcg_id);
 
 	memcg_free_cache_id(kmemcg_id);