diff --git a/lib/sbitmap.c b/lib/sbitmap.c
index 4d8e97e..1651ad9d 100644
--- a/lib/sbitmap.c
+++ b/lib/sbitmap.c
@@ -205,11 +205,18 @@
 	if (ret)
 		return ret;
 
+	sbq->alloc_hint = alloc_percpu_gfp(unsigned int, flags);
+	if (!sbq->alloc_hint) {
+		sbitmap_free(&sbq->sb);
+		return -ENOMEM;
+	}
+
 	sbq->wake_batch = sbq_calc_wake_batch(depth);
 	atomic_set(&sbq->wake_index, 0);
 
 	sbq->ws = kzalloc_node(SBQ_WAIT_QUEUES * sizeof(*sbq->ws), flags, node);
 	if (!sbq->ws) {
+		free_percpu(sbq->alloc_hint);
 		sbitmap_free(&sbq->sb);
 		return -ENOMEM;
 	}
@@ -229,6 +236,29 @@
 }
 EXPORT_SYMBOL_GPL(sbitmap_queue_resize);
 
+int __sbitmap_queue_get(struct sbitmap_queue *sbq, bool round_robin)
+{
+	unsigned int hint;
+	int nr;
+
+	hint = this_cpu_read(*sbq->alloc_hint);
+	nr = sbitmap_get(&sbq->sb, hint, round_robin);
+
+	if (nr == -1) {
+		/* If the map is full, a hint won't do us much good. */
+		this_cpu_write(*sbq->alloc_hint, 0);
+	} else if (nr == hint || unlikely(round_robin)) {
+		/* Only update the hint if we used it. */
+		hint = nr + 1;
+		if (hint >= sbq->sb.depth - 1)
+			hint = 0;
+		this_cpu_write(*sbq->alloc_hint, hint);
+	}
+
+	return nr;
+}
+EXPORT_SYMBOL_GPL(__sbitmap_queue_get);
+
 static struct sbq_wait_state *sbq_wake_ptr(struct sbitmap_queue *sbq)
 {
 	int i, wake_index;
@@ -273,10 +303,13 @@
 	}
 }
 
-void sbitmap_queue_clear(struct sbitmap_queue *sbq, unsigned int nr)
+void sbitmap_queue_clear(struct sbitmap_queue *sbq, unsigned int nr,
+			 bool round_robin, unsigned int cpu)
 {
 	sbitmap_clear_bit(&sbq->sb, nr);
 	sbq_wake_up(sbq);
+	if (likely(!round_robin))
+		*per_cpu_ptr(sbq->alloc_hint, cpu) = nr;
 }
 EXPORT_SYMBOL_GPL(sbitmap_queue_clear);