diff --git a/.pick_status.json b/.pick_status.json
index 6f690fdf51a..7147e319653 100644
--- a/.pick_status.json
+++ b/.pick_status.json
@@ -292,7 +292,7 @@
         "description": "lavapipe: fix fencing when submitting multiple cmdbufs",
         "nominated": true,
         "nomination_type": 1,
-        "resolution": 0,
+        "resolution": 1,
         "master_sha": null,
         "because_sha": "b38879f8c5f57b7f1802e433e33181bdf5e72aef"
     },
diff --git a/src/gallium/frontends/lavapipe/lvp_device.c b/src/gallium/frontends/lavapipe/lvp_device.c
index 7f58d1d1d34..f02d4796acd 100644
--- a/src/gallium/frontends/lavapipe/lvp_device.c
+++ b/src/gallium/frontends/lavapipe/lvp_device.c
@@ -1079,9 +1079,18 @@ static int queue_thread(void *data)
       mtx_unlock(&queue->m);
       //execute
       for (unsigned i = 0; i < task->cmd_buffer_count; i++) {
-         lvp_execute_cmds(queue->device, queue, task->fence, task->cmd_buffers[i]);
+         lvp_execute_cmds(queue->device, queue, task->cmd_buffers[i]);
       }
-      if (!task->cmd_buffer_count && task->fence)
+
+      if (task->cmd_buffer_count) {
+         struct pipe_fence_handle *handle = NULL;
+         queue->ctx->flush(queue->ctx, task->fence ? &handle : NULL, 0);
+         if (task->fence) {
+            mtx_lock(&queue->device->fence_lock);
+            task->fence->handle = handle;
+            mtx_unlock(&queue->device->fence_lock);
+         }
+      } else if (task->fence)
          task->fence->signaled = true;
       p_atomic_dec(&queue->count);
       mtx_lock(&queue->m);
diff --git a/src/gallium/frontends/lavapipe/lvp_execute.c b/src/gallium/frontends/lavapipe/lvp_execute.c
index a1b1ab2d5ca..c1e6e396bfc 100644
--- a/src/gallium/frontends/lavapipe/lvp_execute.c
+++ b/src/gallium/frontends/lavapipe/lvp_execute.c
@@ -3142,11 +3142,9 @@ static void lvp_execute_cmd_buffer(struct lvp_cmd_buffer *cmd_buffer,
 
 VkResult lvp_execute_cmds(struct lvp_device *device,
                           struct lvp_queue *queue,
-                          struct lvp_fence *fence,
                           struct lvp_cmd_buffer *cmd_buffer)
 {
    struct rendering_state state;
-   struct pipe_fence_handle *handle = NULL;
    memset(&state, 0, sizeof(state));
    state.pctx = queue->ctx;
    state.blend_dirty = true;
@@ -3156,12 +3154,6 @@ VkResult lvp_execute_cmds(struct lvp_device *device,
    /* create a gallium context */
    lvp_execute_cmd_buffer(cmd_buffer, &state);
 
-   state.pctx->flush(state.pctx, fence ? &handle : NULL, 0);
-   if (fence) {
-      mtx_lock(&device->fence_lock);
-      fence->handle = handle;
-      mtx_unlock(&device->fence_lock);
-   }
    state.start_vb = -1;
    state.num_vb = 0;
    state.pctx->set_vertex_buffers(state.pctx, 0, 0, PIPE_MAX_ATTRIBS, false, NULL);
diff --git a/src/gallium/frontends/lavapipe/lvp_private.h b/src/gallium/frontends/lavapipe/lvp_private.h
index d1e0b27ce31..4567b09db30 100644
--- a/src/gallium/frontends/lavapipe/lvp_private.h
+++ b/src/gallium/frontends/lavapipe/lvp_private.h
@@ -1103,7 +1103,6 @@ struct lvp_cmd_buffer_entry {
 
 VkResult lvp_execute_cmds(struct lvp_device *device,
                           struct lvp_queue *queue,
-                          struct lvp_fence *fence,
                           struct lvp_cmd_buffer *cmd_buffer);
 
 struct lvp_image *lvp_swapchain_get_image(VkSwapchainKHR swapchain,