memcg, slab: RCU protect memcg_params for root caches

We relocate root cache's memcg_params whenever we need to grow the
memcg_caches array to accommodate all kmem-active memory cgroups.
Currently on relocation we free the old version immediately, which can
lead to use-after-free, because the memcg_caches array is accessed
lock-free (see cache_from_memcg_idx()). This patch fixes this by making
memcg_params RCU-protected for root caches.

Signed-off-by: Vladimir Davydov <[e-mail]>
Cc: Michal Hocko <[e-mail]>
Cc: Glauber Costa <[e-mail]>
Cc: Johannes Weiner <[e-mail]>
Cc: Balbir Singh <[e-mail]>
Cc: KAMEZAWA Hiroyuki <[e-mail]>
Cc: Pekka Enberg <[e-mail]>
Cc: Christoph Lameter <[e-mail]>
Signed-off-by: Andrew Morton <[e-mail]>
Signed-off-by: Linus Torvalds <[e-mail]>

include/linux/slab.h

···
  *
  * Both the root cache and the child caches will have it. For the root cache,
  * this will hold a dynamically allocated array large enough to hold
- * information about the currently limited memcgs in the system.
+ * information about the currently limited memcgs in the system. To allow the
+ * array to be accessed without taking any locks, on relocation we free the old
+ * version only after a grace period.
  *
  * Child caches will hold extra metadata needed for its operation. Fields are:
  *
···
 struct memcg_cache_params {
 	bool is_root_cache;
 	union {
-		struct kmem_cache *memcg_caches[0];
+		struct {
+			struct rcu_head rcu_head;
+			struct kmem_cache *memcg_caches[0];
+		};
 		struct {
 			struct mem_cgroup *memcg;
 			struct list_head list;

mm/memcontrol.c

···
 
 	if (num_groups > memcg_limited_groups_array_size) {
 		int i;
+		struct memcg_cache_params *new_params;
 		ssize_t size = memcg_caches_array_size(num_groups);
 
 		size *= sizeof(void *);
 		size += offsetof(struct memcg_cache_params, memcg_caches);
 
-		s->memcg_params = kzalloc(size, GFP_KERNEL);
-		if (!s->memcg_params) {
-			s->memcg_params = cur_params;
+		new_params = kzalloc(size, GFP_KERNEL);
+		if (!new_params)
 			return -ENOMEM;
-		}
 
-		s->memcg_params->is_root_cache = true;
+		new_params->is_root_cache = true;
 
 		/*
 		 * There is the chance it will be bigger than
···
 		for (i = 0; i < memcg_limited_groups_array_size; i++) {
 			if (!cur_params->memcg_caches[i])
 				continue;
-			s->memcg_params->memcg_caches[i] =
+			new_params->memcg_caches[i] =
 						cur_params->memcg_caches[i];
 		}
 
···
 		 * bigger than the others. And all updates will reset this
 		 * anyway.
 		 */
-		kfree(cur_params);
+		rcu_assign_pointer(s->memcg_params, new_params);
+		if (cur_params)
+			kfree_rcu(cur_params, rcu_head);
 	}
 	return 0;
 }

mm/slab.h

···
 	return s->name;
 }
 
+/*
+ * Note, we protect with RCU only the memcg_caches array, not per-memcg caches.
+ * That said the caller must assure the memcg's cache won't go away. Since once
+ * created a memcg's cache is destroyed only along with the root cache, it is
+ * true if we are going to allocate from the cache or hold a reference to the
+ * root cache by other means. Otherwise, we should hold either the slab_mutex
+ * or the memcg's slab_caches_mutex while calling this function and accessing
+ * the returned value.
+ */
 static inline struct kmem_cache *
 cache_from_memcg_idx(struct kmem_cache *s, int idx)
 {
 	struct kmem_cache *cachep;
+	struct memcg_cache_params *params;
 
 	if (!s->memcg_params)
 		return NULL;
-	cachep = s->memcg_params->memcg_caches[idx];
+
+	rcu_read_lock();
+	params = rcu_dereference(s->memcg_params);
+	cachep = params->memcg_caches[idx];
+	rcu_read_unlock();
 
 	/*
 	 * Make sure we will access the up-to-date value. The code updating