Message ID | 8002e999-feb5-8715-9e1a-ed67165ade7e@xs4all.nl (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | sun4i-csi: call _vb2_fop_release instead of v4l2_fh_release | expand |
Hi Hans, On Tue, Jul 07, 2020 at 03:36:47PM +0200, Hans Verkuil wrote: > This driver didn't call _vb2_fop_release in the sun4i_csi_release > function. Without that call the vb2 queue is not properly canceled and > buffers may not be freed. > > Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> Acked-by: Maxime Ripard <mripard@kernel.org> Thanks! Maxime
diff --git a/drivers/media/platform/sunxi/sun4i-csi/sun4i_v4l2.c b/drivers/media/platform/sunxi/sun4i-csi/sun4i_v4l2.c index 1721e5aee9c6..8f4e254b6a41 100644 --- a/drivers/media/platform/sunxi/sun4i-csi/sun4i_v4l2.c +++ b/drivers/media/platform/sunxi/sun4i-csi/sun4i_v4l2.c @@ -242,7 +242,8 @@ static int sun4i_csi_release(struct file *file) mutex_lock(&csi->lock); - v4l2_fh_release(file); + _vb2_fop_release(file, NULL); + v4l2_pipeline_pm_put(&csi->vdev.entity); pm_runtime_put(csi->dev);
This driver didn't call _vb2_fop_release in the sun4i_csi_release function. Without that call the vb2 queue is not properly canceled and buffers may not be freed. Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl> --- Compile tested only due to lack of a suitable sensor. I believe I do have an ov7670, but it's in the wrong country :-) ---