On Mon, Aug 31, 2020 at 01:46:00PM -0700, Sowjanya Komatineni wrote: > > On 8/31/20 1:23 PM, Sakari Ailus wrote: > > > @@ -1968,19 +2087,45 @@ static int imx274_remove(struct i2c_client *client) > > > struct v4l2_subdev *sd = i2c_get_clientdata(client); > > > struct stimx274 *imx274 = to_imx274(sd); > > > + pm_runtime_get_sync(&imx274->client->dev); > > > + > > > /* stop stream */ > > This really shouldn't happen and the driver isn't expected to handle it > > either. > > Do you mean to remove stop stream during remove()? > > Stop stream is not part of this change and as writes to sensor can't happen > when power off, added pm_runtime_get_sync Indeed. But there certainly isn't a need to power the sensor on to stream off, is there? > > > > imx274_write_table(imx274, imx274_stop); > > > v4l2_async_unregister_subdev(sd); > > > v4l2_ctrl_handler_free(&imx274->ctrls.handler); > > > + > > > + pm_runtime_put(&client->dev); > > > + pm_runtime_disable(&client->dev); > > > + pm_runtime_set_suspended(&client->dev); > > > + > > > media_entity_cleanup(&sd->entity); > > > mutex_destroy(&imx274->lock); > > > return 0; > > > } -- Sakari Ailus
