diff --git a/src/gallium/frontends/lavapipe/lvp_device.c b/src/gallium/frontends/lavapipe/lvp_device.c
index 12e14919898..7bfe36cd0ba 100644
--- a/src/gallium/frontends/lavapipe/lvp_device.c
+++ b/src/gallium/frontends/lavapipe/lvp_device.c
@@ -1709,9 +1709,10 @@ VKAPI_ATTR PFN_vkVoidFunction VKAPI_CALL vk_icdGetInstanceProcAddr(
 static void
 destroy_pipelines(struct lvp_queue *queue)
 {
+   struct lvp_device *device = lvp_queue_device(queue);
    simple_mtx_lock(&queue->lock);
    while (util_dynarray_contains(&queue->pipeline_destroys, struct lvp_pipeline*)) {
-      lvp_pipeline_destroy(queue->device, util_dynarray_pop(&queue->pipeline_destroys, struct lvp_pipeline*), true);
+      lvp_pipeline_destroy(device, util_dynarray_pop(&queue->pipeline_destroys, struct lvp_pipeline*), true);
    }
    simple_mtx_unlock(&queue->lock);
 }
@@ -1721,8 +1722,9 @@ lvp_queue_submit(struct vk_queue *vk_queue,
                  struct vk_queue_submit *submit)
 {
    struct lvp_queue *queue = container_of(vk_queue, struct lvp_queue, vk);
+   struct lvp_device *device = lvp_queue_device(queue);
 
-   VkResult result = vk_sync_wait_many(&queue->device->vk,
+   VkResult result = vk_sync_wait_many(&device->vk,
                                        submit->wait_count, submit->waits,
                                        VK_SYNC_WAIT_COMPLETE, UINT64_MAX);
    if (result != VK_SUCCESS)
@@ -1733,26 +1735,26 @@ lvp_queue_submit(struct vk_queue *vk_queue,
    for (uint32_t i = 0; i < submit->buffer_bind_count; i++) {
       VkSparseBufferMemoryBindInfo *bind = &submit->buffer_binds[i];
 
-      lvp_buffer_bind_sparse(queue->device, queue, bind);
+      lvp_buffer_bind_sparse(device, queue, bind);
    }
 
    for (uint32_t i = 0; i < submit->image_opaque_bind_count; i++) {
       VkSparseImageOpaqueMemoryBindInfo *bind = &submit->image_opaque_binds[i];
 
-      lvp_image_bind_opaque_sparse(queue->device, queue, bind);
+      lvp_image_bind_opaque_sparse(device, queue, bind);
    }
 
    for (uint32_t i = 0; i < submit->image_bind_count; i++) {
       VkSparseImageMemoryBindInfo *bind = &submit->image_binds[i];
 
-      lvp_image_bind_sparse(queue->device, queue, bind);
+      lvp_image_bind_sparse(device, queue, bind);
    }
 
    for (uint32_t i = 0; i < submit->command_buffer_count; i++) {
       struct lvp_cmd_buffer *cmd_buffer =
          container_of(submit->command_buffers[i], struct lvp_cmd_buffer, vk);
 
-      lvp_execute_cmds(queue->device, queue, cmd_buffer);
+      lvp_execute_cmds(device, queue, cmd_buffer);
    }
 
    simple_mtx_unlock(&queue->lock);
@@ -1763,7 +1765,7 @@ lvp_queue_submit(struct vk_queue *vk_queue,
    for (uint32_t i = 0; i < submit->signal_count; i++) {
       struct lvp_pipe_sync *sync =
          vk_sync_as_lvp_pipe_sync(submit->signals[i].sync);
-      lvp_pipe_sync_signal_with_fence(queue->device, sync, queue->last_fence);
+      lvp_pipe_sync_signal_with_fence(device, sync, queue->last_fence);
    }
    destroy_pipelines(queue);
 
@@ -1786,8 +1788,6 @@ lvp_queue_init(struct lvp_device *device, struct lvp_queue *queue,
       return result;
    }
 
-   queue->device = device;
-
    queue->ctx = device->pscreen->context_create(device->pscreen, NULL, PIPE_CONTEXT_ROBUST_BUFFER_ACCESS);
    queue->cso = cso_create_context(queue->ctx, CSO_NO_VBUF);
    queue->uploader = u_upload_create(queue->ctx, 1024 * 1024, PIPE_BIND_CONSTANT_BUFFER, PIPE_USAGE_STREAM, 0);
diff --git a/src/gallium/frontends/lavapipe/lvp_private.h b/src/gallium/frontends/lavapipe/lvp_private.h
index 16aadc9733a..86b47908342 100644
--- a/src/gallium/frontends/lavapipe/lvp_private.h
+++ b/src/gallium/frontends/lavapipe/lvp_private.h
@@ -180,7 +180,6 @@ bool lvp_physical_device_extension_supported(struct lvp_physical_device *dev,
 
 struct lvp_queue {
    struct vk_queue vk;
-   struct lvp_device *device;
    struct pipe_context *ctx;
    struct cso_context *cso;
    struct u_upload_mgr *uploader;
@@ -190,6 +189,12 @@ struct lvp_queue {
    simple_mtx_t lock;
 };
 
+static inline struct lvp_device *
+lvp_queue_device(const struct lvp_queue *queue)
+{
+   return (struct lvp_device *)queue->vk.base.device;
+}
+
 struct lvp_pipeline_cache {
    struct vk_object_base base;
    struct lvp_device *device;