Hi Hans,
Thank you for the patch.
On Tue, Oct 10, 2023 at 12:24:58PM +0200, Hans de Goede wrote:
> Since the stream API is still experimental it is currently locked away
> behind the internal, default disabled, v4l2_subdev_enable_streams_api flag.
>
> Advertising V4L2_SUBDEV_CAP_STREAMS when the streams API is disabled
> confuses userspace. E.g. it causes the following libcamera error:
>
> ERROR SimplePipeline simple.cpp:1497 Failed to reset routes for
> /dev/v4l-subdev1: Inappropriate ioctl for device
>
> Don't report V4L2_SUBDEV_CAP_STREAMS when the streams API is disabled
> to avoid problems like this.
>
> Reported-by: Dennis Bonke <admin@xxxxxxxxxxxxxxx>
> Signed-off-by: Hans de Goede <hdegoede@xxxxxxxxxx>
Reviewed-by: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> ---
> -Clearing the V4L2_SUBDEV_FL_STREAMS flag from sd.flags might seem
> appealing as an alternative fix. But this causes various v4l2-core bits
> to enter different code paths which confuses drivers which set
> V4L2_SUBDEV_FL_STREAMS, so this is a bad idea.
> -No Closes: for the Reported-by since this was reported by private email
> ---
> drivers/media/v4l2-core/v4l2-subdev.c | 7 +++++++
> 1 file changed, 7 insertions(+)
>
> diff --git a/drivers/media/v4l2-core/v4l2-subdev.c b/drivers/media/v4l2-core/v4l2-subdev.c
> index b92348ad61f6..31752c06d1f0 100644
> --- a/drivers/media/v4l2-core/v4l2-subdev.c
> +++ b/drivers/media/v4l2-core/v4l2-subdev.c
> @@ -502,6 +502,13 @@ static long subdev_do_ioctl(struct file *file, unsigned int cmd, void *arg,
> V4L2_SUBDEV_CLIENT_CAP_STREAMS;
> int rval;
>
> + /*
> + * If the streams API is not enabled, remove V4L2_SUBDEV_CAP_STREAMS.
> + * Remove this when the API is no longer experimental.
> + */
> + if (!v4l2_subdev_enable_streams_api)
> + streams_subdev = false;
> +
> switch (cmd) {
> case VIDIOC_SUBDEV_QUERYCAP: {
> struct v4l2_subdev_capability *cap = arg;
--
Regards,
Laurent Pinchart
