diff --git a/drivers/gpu/drm/i915/i915_sw_fence.c b/drivers/gpu/drm/i915/i915_sw_fence.c
index 8185002d7ec8..95f2f12e0917 100644
--- a/drivers/gpu/drm/i915/i915_sw_fence.c
+++ b/drivers/gpu/drm/i915/i915_sw_fence.c
@@ -13,6 +13,8 @@
 
 #include "i915_sw_fence.h"
 
+#define I915_SW_FENCE_FLAG_ALLOC BIT(3) /* after WQ_FLAG_* for safety */
+
 static DEFINE_SPINLOCK(i915_sw_fence_lock);
 
 static int __i915_sw_fence_notify(struct i915_sw_fence *fence,
@@ -135,6 +137,8 @@ static int i915_sw_fence_wake(wait_queue_t *wq, unsigned mode, int flags, void *
 	list_del(&wq->task_list);
 	__i915_sw_fence_complete(wq->private, key);
 	i915_sw_fence_put(wq->private);
+	if (wq->flags & I915_SW_FENCE_FLAG_ALLOC)
+		kfree(wq);
 	return 0;
 }
 
@@ -192,9 +196,9 @@ static bool i915_sw_fence_check_if_after(struct i915_sw_fence *fence,
 	return err;
 }
 
-int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
-				 struct i915_sw_fence *signaler,
-				 wait_queue_t *wq)
+static int __i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
+					  struct i915_sw_fence *signaler,
+					  wait_queue_t *wq, gfp_t gfp)
 {
 	unsigned long flags;
 	int pending;
@@ -206,8 +210,22 @@ int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
 	if (unlikely(i915_sw_fence_check_if_after(fence, signaler)))
 		return -EINVAL;
 
+	pending = 0;
+	if (!wq) {
+		wq = kmalloc(sizeof(*wq), gfp);
+		if (!wq) {
+			if (!gfpflags_allow_blocking(gfp))
+				return -ENOMEM;
+
+			i915_sw_fence_wait(signaler);
+			return 0;
+		}
+
+		pending |= I915_SW_FENCE_FLAG_ALLOC;
+	}
+
 	INIT_LIST_HEAD(&wq->task_list);
-	wq->flags = 0;
+	wq->flags = pending;
 	wq->func = i915_sw_fence_wake;
 	wq->private = i915_sw_fence_get(fence);
 
@@ -226,6 +244,20 @@ int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
 	return pending;
 }
 
+int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
+				 struct i915_sw_fence *signaler,
+				 wait_queue_t *wq)
+{
+	return __i915_sw_fence_await_sw_fence(fence, signaler, wq, 0);
+}
+
+int i915_sw_fence_await_sw_fence_gfp(struct i915_sw_fence *fence,
+				     struct i915_sw_fence *signaler,
+				     gfp_t gfp)
+{
+	return __i915_sw_fence_await_sw_fence(fence, signaler, NULL, gfp);
+}
+
 struct i915_sw_dma_fence_cb {
 	struct dma_fence_cb base;
 	struct i915_sw_fence *fence;
diff --git a/drivers/gpu/drm/i915/i915_sw_fence.h b/drivers/gpu/drm/i915/i915_sw_fence.h
index cd239e92f67f..707dfc4f0da5 100644
--- a/drivers/gpu/drm/i915/i915_sw_fence.h
+++ b/drivers/gpu/drm/i915/i915_sw_fence.h
@@ -46,6 +46,9 @@ void i915_sw_fence_commit(struct i915_sw_fence *fence);
 int i915_sw_fence_await_sw_fence(struct i915_sw_fence *fence,
 				 struct i915_sw_fence *after,
 				 wait_queue_t *wq);
+int i915_sw_fence_await_sw_fence_gfp(struct i915_sw_fence *fence,
+				     struct i915_sw_fence *after,
+				     gfp_t gfp);
 int i915_sw_fence_await_dma_fence(struct i915_sw_fence *fence,
 				  struct dma_fence *dma,
 				  unsigned long timeout,
@@ -62,4 +65,9 @@ static inline bool i915_sw_fence_done(const struct i915_sw_fence *fence)
 	return atomic_read(&fence->pending) < 0;
 }
 
+static inline void i915_sw_fence_wait(struct i915_sw_fence *fence)
+{
+	wait_event(fence->wait, i915_sw_fence_done(fence));
+}
+
 #endif /* _I915_SW_FENCE_H_ */