Hi Michael, Thank you for the patch. On Thu, Mar 23, 2023 at 12:41:14PM +0100, Michael Tretter wrote: > The TRY_FMT call should try to avoid returning EINVAL. If the requested > pixelformat is not supported by the driver, the driver should fall back > to it's own default and only then fail the request. It should actually never fail. You may recall that I didn't like the TRY_FMT implementation, and I still stand by the arguments I provided by then. If you want to handle this properly, revert the commits that added format handling to the driver. > Signed-off-by: Michael Tretter <m.tretter@xxxxxxxxxxxxxx> > --- > drivers/usb/gadget/function/uvc_v4l2.c | 4 ++++ > 1 file changed, 4 insertions(+) > > diff --git a/drivers/usb/gadget/function/uvc_v4l2.c b/drivers/usb/gadget/function/uvc_v4l2.c > index 3f728f451ed5..c5983bb0a8d1 100644 > --- a/drivers/usb/gadget/function/uvc_v4l2.c > +++ b/drivers/usb/gadget/function/uvc_v4l2.c > @@ -317,11 +317,15 @@ uvc_v4l2_try_format(struct file *file, void *fh, struct v4l2_format *fmt) > fmt->fmt.pix.width, fmt->fmt.pix.height); > > uformat = find_format_by_pix(uvc, fmt->fmt.pix.pixelformat); > + if (!uformat) > + uformat = get_default_format(uvc); > if (!uformat) > return -EINVAL; > > uframe = find_closest_frame_by_size(uvc, uformat, > fmt->fmt.pix.width, fmt->fmt.pix.height); > + if (!uframe) > + uframe = get_default_frame(uvc, uformat); > if (!uframe) > return -EINVAL; > -- Regards, Laurent Pinchart
