diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu.h b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
index d050f503d9ff..4addac5f6763 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu.h
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu.h
@@ -440,9 +440,12 @@ int amdgpu_fence_wait_next(struct amdgpu_ring *ring);
 int amdgpu_fence_wait_empty(struct amdgpu_ring *ring);
 unsigned amdgpu_fence_count_emitted(struct amdgpu_ring *ring);
 
-signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
-			  struct amdgpu_fence **fences,
-			  bool intr, long t);
+signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev,
+				       struct amdgpu_fence **array,
+				       uint32_t count,
+				       bool wait_all,
+				       bool intr,
+				       signed long t);
 struct amdgpu_fence *amdgpu_fence_ref(struct amdgpu_fence *fence);
 void amdgpu_fence_unref(struct amdgpu_fence **fence);
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
index 98500f1756f7..ae014fcf524e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_fence.c
@@ -836,13 +836,12 @@ static inline bool amdgpu_test_signaled(struct amdgpu_fence *fence)
 	return test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags);
 }
 
-static inline bool amdgpu_test_signaled_any(struct amdgpu_fence **fences)
+static bool amdgpu_test_signaled_any(struct amdgpu_fence **fences, uint32_t count)
 {
 	int idx;
 	struct amdgpu_fence *fence;
 
-	idx = 0;
-	for (idx = 0; idx < AMDGPU_MAX_RINGS; ++idx) {
+	for (idx = 0; idx < count; ++idx) {
 		fence = fences[idx];
 		if (fence) {
 			if (test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
@@ -852,6 +851,22 @@ static inline bool amdgpu_test_signaled_any(struct amdgpu_fence **fences)
 	return false;
 }
 
+static bool amdgpu_test_signaled_all(struct amdgpu_fence **fences, uint32_t count)
+{
+	int idx;
+	struct amdgpu_fence *fence;
+
+	for (idx = 0; idx < count; ++idx) {
+		fence = fences[idx];
+		if (fence) {
+			if (!test_bit(FENCE_FLAG_SIGNALED_BIT, &fence->base.flags))
+				return false;
+		}
+	}
+
+	return true;
+}
+
 struct amdgpu_wait_cb {
 	struct fence_cb base;
 	struct task_struct *task;
@@ -867,33 +882,56 @@ static void amdgpu_fence_wait_cb(struct fence *fence, struct fence_cb *cb)
 static signed long amdgpu_fence_default_wait(struct fence *f, bool intr,
 					     signed long t)
 {
-	struct amdgpu_fence *array[AMDGPU_MAX_RINGS];
 	struct amdgpu_fence *fence = to_amdgpu_fence(f);
 	struct amdgpu_device *adev = fence->ring->adev;
 
-	memset(&array[0], 0, sizeof(array));
-	array[0] = fence;
-
-	return amdgpu_fence_wait_any(adev, array, intr, t);
+	return amdgpu_fence_wait_multiple(adev, &fence, 1, false, intr, t);
 }
 
-/* wait until any fence in array signaled */
-signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
-				struct amdgpu_fence **array, bool intr, signed long t)
+/**
+ * Wait the fence array with timeout
+ *
+ * @adev:     amdgpu device
+ * @array:    the fence array with amdgpu fence pointer
+ * @count:    the number of the fence array
+ * @wait_all: the flag of wait all(true) or wait any(false)
+ * @intr:     when sleep, set the current task interruptable or not
+ * @t:        timeout to wait
+ *
+ * If wait_all is true, it will return when all fences are signaled or timeout.
+ * If wait_all is false, it will return when any fence is signaled or timeout.
+ */
+signed long amdgpu_fence_wait_multiple(struct amdgpu_device *adev,
+				       struct amdgpu_fence **array,
+				       uint32_t count,
+				       bool wait_all,
+				       bool intr,
+				       signed long t)
 {
 	long idx = 0;
-	struct amdgpu_wait_cb cb[AMDGPU_MAX_RINGS];
+	struct amdgpu_wait_cb *cb;
 	struct amdgpu_fence *fence;
 
 	BUG_ON(!array);
 
-	for (idx = 0; idx < AMDGPU_MAX_RINGS; ++idx) {
+	cb = kcalloc(count, sizeof(struct amdgpu_wait_cb), GFP_KERNEL);
+	if (cb == NULL) {
+		t = -ENOMEM;
+		goto err_free_cb;
+	}
+
+	for (idx = 0; idx < count; ++idx) {
 		fence = array[idx];
 		if (fence) {
 			cb[idx].task = current;
 			if (fence_add_callback(&fence->base,
-					&cb[idx].base, amdgpu_fence_wait_cb))
-				return t; /* return if fence is already signaled */
+					&cb[idx].base, amdgpu_fence_wait_cb)) {
+				/* The fence is already signaled */
+				if (wait_all)
+					continue;
+				else
+					goto fence_rm_cb;
+			}
 		}
 	}
 
@@ -907,7 +945,9 @@ signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
 		 * amdgpu_test_signaled_any must be called after
 		 * set_current_state to prevent a race with wake_up_process
 		 */
-		if (amdgpu_test_signaled_any(array))
+		if (!wait_all && amdgpu_test_signaled_any(array, count))
+			break;
+		if (wait_all && amdgpu_test_signaled_all(array, count))
 			break;
 
 		if (adev->needs_reset) {
@@ -923,13 +963,16 @@ signed long amdgpu_fence_wait_any(struct amdgpu_device *adev,
 
 	__set_current_state(TASK_RUNNING);
 
-	idx = 0;
-	for (idx = 0; idx < AMDGPU_MAX_RINGS; ++idx) {
+fence_rm_cb:
+	for (idx = 0; idx < count; ++idx) {
 		fence = array[idx];
 		if (fence)
 			fence_remove_callback(&fence->base, &cb[idx].base);
 	}
 
+err_free_cb:
+	kfree(cb);
+
 	return t;
 }
 
diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
index d6398cf45f24..4597899e9758 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_sa.c
@@ -352,7 +352,8 @@ int amdgpu_sa_bo_new(struct amdgpu_device *adev,
 		} while (amdgpu_sa_bo_next_hole(sa_manager, fences, tries));
 
 		spin_unlock(&sa_manager->wq.lock);
-		t = amdgpu_fence_wait_any(adev, fences, false, MAX_SCHEDULE_TIMEOUT);
+		t = amdgpu_fence_wait_multiple(adev, fences, AMDGPU_MAX_RINGS, false, false,
+						MAX_SCHEDULE_TIMEOUT);
 		r = (t > 0) ? 0 : t;
 		spin_lock(&sa_manager->wq.lock);
 		/* if we have nothing to wait for block */