diff --git a/include/linux/bpf.h b/include/linux/bpf.h
index a98c833461347..d808253f2e945 100644
--- a/include/linux/bpf.h
+++ b/include/linux/bpf.h
@@ -2499,6 +2499,8 @@ int bpf_map_alloc_pages(const struct bpf_map *map, int nid,
 #ifdef CONFIG_MEMCG
 void *bpf_map_kmalloc_node(const struct bpf_map *map, size_t size, gfp_t flags,
 			   int node);
+void *bpf_map_kmalloc_nolock(const struct bpf_map *map, size_t size, gfp_t flags,
+			     int node);
 void *bpf_map_kzalloc(const struct bpf_map *map, size_t size, gfp_t flags);
 void *bpf_map_kvcalloc(struct bpf_map *map, size_t n, size_t size,
 		       gfp_t flags);
@@ -2511,6 +2513,8 @@ void __percpu *bpf_map_alloc_percpu(const struct bpf_map *map, size_t size,
  */
 #define bpf_map_kmalloc_node(_map, _size, _flags, _node)	\
 		kmalloc_node(_size, _flags, _node)
+#define bpf_map_kmalloc_nolock(_map, _size, _flags, _node)	\
+		kmalloc_nolock(_size, _flags, _node)
 #define bpf_map_kzalloc(_map, _size, _flags)			\
 		kzalloc(_size, _flags)
 #define bpf_map_kvcalloc(_map, _n, _size, _flags)		\
diff --git a/kernel/bpf/helpers.c b/kernel/bpf/helpers.c
index c9fab9a356dfc..c5f63f2685e99 100644
--- a/kernel/bpf/helpers.c
+++ b/kernel/bpf/helpers.c
@@ -1215,13 +1215,20 @@ static void bpf_wq_work(struct work_struct *work)
 	rcu_read_unlock_trace();
 }
 
+static void bpf_async_cb_rcu_free(struct rcu_head *rcu)
+{
+	struct bpf_async_cb *cb = container_of(rcu, struct bpf_async_cb, rcu);
+
+	kfree_nolock(cb);
+}
+
 static void bpf_wq_delete_work(struct work_struct *work)
 {
 	struct bpf_work *w = container_of(work, struct bpf_work, delete_work);
 
 	cancel_work_sync(&w->work);
 
-	kfree_rcu(w, cb.rcu);
+	call_rcu(&w->cb.rcu, bpf_async_cb_rcu_free);
 }
 
 static void bpf_timer_delete_work(struct work_struct *work)
@@ -1230,13 +1237,13 @@ static void bpf_timer_delete_work(struct work_struct *work)
 
 	/* Cancel the timer and wait for callback to complete if it was running.
 	 * If hrtimer_cancel() can be safely called it's safe to call
-	 * kfree_rcu(t) right after for both preallocated and non-preallocated
+	 * call_rcu() right after for both preallocated and non-preallocated
 	 * maps.  The async->cb = NULL was already done and no code path can see
 	 * address 't' anymore. Timer if armed for existing bpf_hrtimer before
 	 * bpf_timer_cancel_and_free will have been cancelled.
 	 */
 	hrtimer_cancel(&t->timer);
-	kfree_rcu(t, cb.rcu);
+	call_rcu(&t->cb.rcu, bpf_async_cb_rcu_free);
 }
 
 static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u64 flags,
@@ -1270,11 +1277,7 @@ static int __bpf_async_init(struct bpf_async_kern *async, struct bpf_map *map, u
 		goto out;
 	}
 
-	/* Allocate via bpf_map_kmalloc_node() for memcg accounting. Until
-	 * kmalloc_nolock() is available, avoid locking issues by using
-	 * __GFP_HIGH (GFP_ATOMIC & ~__GFP_RECLAIM).
-	 */
-	cb = bpf_map_kmalloc_node(map, size, __GFP_HIGH, map->numa_node);
+	cb = bpf_map_kmalloc_nolock(map, size, 0, map->numa_node);
 	if (!cb) {
 		ret = -ENOMEM;
 		goto out;
@@ -1607,7 +1610,7 @@ void bpf_timer_cancel_and_free(void *val)
 		 * completion.
 		 */
 		if (hrtimer_try_to_cancel(&t->timer) >= 0)
-			kfree_rcu(t, cb.rcu);
+			call_rcu(&t->cb.rcu, bpf_async_cb_rcu_free);
 		else
 			queue_work(system_dfl_wq, &t->cb.delete_work);
 	} else {
diff --git a/kernel/bpf/syscall.c b/kernel/bpf/syscall.c
index 2a9456a3e7304..8a129746bd6cc 100644
--- a/kernel/bpf/syscall.c
+++ b/kernel/bpf/syscall.c
@@ -520,6 +520,21 @@ void *bpf_map_kmalloc_node(const struct bpf_map *map, size_t size, gfp_t flags,
 	return ptr;
 }
 
+void *bpf_map_kmalloc_nolock(const struct bpf_map *map, size_t size, gfp_t flags,
+			     int node)
+{
+	struct mem_cgroup *memcg, *old_memcg;
+	void *ptr;
+
+	memcg = bpf_map_get_memcg(map);
+	old_memcg = set_active_memcg(memcg);
+	ptr = kmalloc_nolock(size, flags | __GFP_ACCOUNT, node);
+	set_active_memcg(old_memcg);
+	mem_cgroup_put(memcg);
+
+	return ptr;
+}
+
 void *bpf_map_kzalloc(const struct bpf_map *map, size_t size, gfp_t flags)
 {
 	struct mem_cgroup *memcg, *old_memcg;