drm/amdkfd: fix circular locking on get_wave_state

get_wave_state acquires the mmap_lock on copy_to_user but so do
mmu_notifiers.  mmu_notifiers allows dqm locking so do get_wave_state
outside the dqm_lock to prevent circular locking.

v2: squash in unused variable removal.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>

drivers/gpu/drm/amd/amdkfd/kfd_device_queue_manager.c

@@ -1674,29 +1674,27 @@ static int get_wave_state(struct device_queue_manager *dqm,
 			  u32 *save_area_used_size)
 {
 	struct mqd_manager *mqd_mgr;
-	int r;
 
 	dqm_lock(dqm);
 
-	if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE ||
-	    q->properties.is_active || !q->device->cwsr_enabled) {
-		r = -EINVAL;
-		goto dqm_unlock;
-	}
-
 	mqd_mgr = dqm->mqd_mgrs[KFD_MQD_TYPE_CP];
 
-	if (!mqd_mgr->get_wave_state) {
-		r = -EINVAL;
-		goto dqm_unlock;
+	if (q->properties.type != KFD_QUEUE_TYPE_COMPUTE ||
+	    q->properties.is_active || !q->device->cwsr_enabled ||
+	    !mqd_mgr->get_wave_state) {
+		dqm_unlock(dqm);
+		return -EINVAL;
 	}
 
-	r = mqd_mgr->get_wave_state(mqd_mgr, q->mqd, ctl_stack,
-			ctl_stack_used_size, save_area_used_size);
-
-dqm_unlock:
 	dqm_unlock(dqm);
-	return r;
+
+	/*
+	 * get_wave_state is outside the dqm lock to prevent circular locking
+	 * and the queue should be protected against destruction by the process
+	 * lock.
+	 */
+	return mqd_mgr->get_wave_state(mqd_mgr, q->mqd, ctl_stack,
+			ctl_stack_used_size, save_area_used_size);
 }
 
 static int process_termination_cpsch(struct device_queue_manager *dqm,