diff --git a/drivers/gpu/drm/i915/i915_debugfs.c b/drivers/gpu/drm/i915/i915_debugfs.c
index d689e51..4bd1467 100644
--- a/drivers/gpu/drm/i915/i915_debugfs.c
+++ b/drivers/gpu/drm/i915/i915_debugfs.c
@@ -292,6 +292,8 @@
 	struct file_stats *stats = data;
 	struct i915_vma *vma;
 
+	lockdep_assert_held(&obj->base.dev->struct_mutex);
+
 	stats->count++;
 	stats->total += obj->base.size;
 	if (!obj->bind_count)
@@ -476,6 +478,8 @@
 		struct drm_i915_gem_request *request;
 		struct task_struct *task;
 
+		mutex_lock(&dev->struct_mutex);
+
 		memset(&stats, 0, sizeof(stats));
 		stats.file_priv = file->driver_priv;
 		spin_lock(&file->table_lock);
@@ -487,7 +491,6 @@
 		 * still alive (e.g. get_pid(current) => fork() => exit()).
 		 * Therefore, we need to protect this ->comm access using RCU.
 		 */
-		mutex_lock(&dev->struct_mutex);
 		request = list_first_entry_or_null(&file_priv->mm.request_list,
 						   struct drm_i915_gem_request,
 						   client_link);
@@ -497,6 +500,7 @@
 				PIDTYPE_PID);
 		print_file_stats(m, task ? task->comm : "<unknown>", stats);
 		rcu_read_unlock();
+
 		mutex_unlock(&dev->struct_mutex);
 	}
 	mutex_unlock(&dev->filelist_mutex);
diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index a3e59c8..9ad13ee 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -546,11 +546,12 @@
 }
 
 static int
-i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
+i915_gem_execbuffer_relocate_entry(struct i915_vma *vma,
 				   struct eb_vmas *eb,
 				   struct drm_i915_gem_relocation_entry *reloc,
 				   struct reloc_cache *cache)
 {
+	struct drm_i915_gem_object *obj = vma->obj;
 	struct drm_i915_private *dev_priv = to_i915(obj->base.dev);
 	struct drm_gem_object *target_obj;
 	struct drm_i915_gem_object *target_i915_obj;
@@ -628,6 +629,16 @@
 		return -EINVAL;
 	}
 
+	/*
+	 * If we write into the object, we need to force the synchronisation
+	 * barrier, either with an asynchronous clflush or if we executed the
+	 * patching using the GPU (though that should be serialised by the
+	 * timeline). To be completely sure, and since we are required to
+	 * do relocations we are already stalling, disable the user's opt
+	 * of our synchronisation.
+	 */
+	vma->exec_entry->flags &= ~EXEC_OBJECT_ASYNC;
+
 	ret = relocate_entry(obj, reloc, cache, target_offset);
 	if (ret)
 		return ret;
@@ -678,7 +689,7 @@
 		do {
 			u64 offset = r->presumed_offset;
 
-			ret = i915_gem_execbuffer_relocate_entry(vma->obj, eb, r, &cache);
+			ret = i915_gem_execbuffer_relocate_entry(vma, eb, r, &cache);
 			if (ret)
 				goto out;
 
@@ -726,7 +737,7 @@
 
 	reloc_cache_init(&cache, eb->i915);
 	for (i = 0; i < entry->relocation_count; i++) {
-		ret = i915_gem_execbuffer_relocate_entry(vma->obj, eb, &relocs[i], &cache);
+		ret = i915_gem_execbuffer_relocate_entry(vma, eb, &relocs[i], &cache);
 		if (ret)
 			break;
 	}
diff --git a/drivers/gpu/drm/i915/i915_vma.c b/drivers/gpu/drm/i915/i915_vma.c
index 1aba470..f066e2d 100644
--- a/drivers/gpu/drm/i915/i915_vma.c
+++ b/drivers/gpu/drm/i915/i915_vma.c
@@ -650,6 +650,11 @@
 				break;
 		}
 
+		if (!ret) {
+			ret = i915_gem_active_retire(&vma->last_fence,
+						     &vma->vm->i915->drm.struct_mutex);
+		}
+
 		__i915_vma_unpin(vma);
 		if (ret)
 			return ret;