media: uvcvideo: Fix allocation for small frame sizes

[PATCH 3/3] media: uvcvideo: use min() for npacket calculation

Posted by Ricardo Ribalda 3 weeks, 5 days ago

Make the code slightly more appealing by making use of min(). There
shall not be any functional change from this patch.

Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
---
 drivers/media/usb/uvc/uvc_video.c | 4 +---
 1 file changed, 1 insertion(+), 3 deletions(-)

diff --git a/drivers/media/usb/uvc/uvc_video.c b/drivers/media/usb/uvc/uvc_video.c
index 59eb95a4b70c05b1a12986e908b7e9979b064fd0..db02080f15772e0bc1d5cfcadd32463f4e6ea045 100644
--- a/drivers/media/usb/uvc/uvc_video.c
+++ b/drivers/media/usb/uvc/uvc_video.c
@@ -1808,9 +1808,7 @@ static int uvc_alloc_urb_buffers(struct uvc_streaming *stream,
 	 * Compute the number of packets. Bulk endpoints might transfer UVC
 	 * payloads across multiple URBs.
 	 */
-	npackets = DIV_ROUND_UP(size, psize);
-	if (npackets > UVC_MAX_PACKETS)
-		npackets = UVC_MAX_PACKETS;
+	npackets = min(UVC_MAX_PACKETS, DIV_ROUND_UP(size, psize));
 
 	/* Retry allocations until one succeed. */
 	for (; npackets > 0; npackets /= 2) {

-- 
2.52.0.457.g6b5491de43-goog

Re: [PATCH 3/3] media: uvcvideo: use min() for npacket calculation

Posted by Laurent Pinchart 2 weeks, 4 days ago

Hi Ricardo,

Thank you for the patch.

On Wed, Jan 14, 2026 at 10:32:15AM +0000, Ricardo Ribalda wrote:
> Make the code slightly more appealing by making use of min(). There
> shall not be any functional change from this patch.
> 
> Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> ---
>  drivers/media/usb/uvc/uvc_video.c | 4 +---
>  1 file changed, 1 insertion(+), 3 deletions(-)
> 
> diff --git a/drivers/media/usb/uvc/uvc_video.c b/drivers/media/usb/uvc/uvc_video.c
> index 59eb95a4b70c05b1a12986e908b7e9979b064fd0..db02080f15772e0bc1d5cfcadd32463f4e6ea045 100644
> --- a/drivers/media/usb/uvc/uvc_video.c
> +++ b/drivers/media/usb/uvc/uvc_video.c
> @@ -1808,9 +1808,7 @@ static int uvc_alloc_urb_buffers(struct uvc_streaming *stream,
>  	 * Compute the number of packets. Bulk endpoints might transfer UVC
>  	 * payloads across multiple URBs.
>  	 */
> -	npackets = DIV_ROUND_UP(size, psize);
> -	if (npackets > UVC_MAX_PACKETS)
> -		npackets = UVC_MAX_PACKETS;
> +	npackets = min(UVC_MAX_PACKETS, DIV_ROUND_UP(size, psize));

Do you think this improves readability ? I find the existing code easier
to read, its purpose is immediately clear: it computes npackets and
clamps it to a max value. With min() I have to pause and think.

I'll take patches 1/3 and 2/3 in my tree already as 1/3 fixes an issue.

>  
>  	/* Retry allocations until one succeed. */
>  	for (; npackets > 0; npackets /= 2) {

-- 
Regards,

Laurent Pinchart

Re: [PATCH 3/3] media: uvcvideo: use min() for npacket calculation

Posted by Ricardo Ribalda 2 weeks, 4 days ago

Hi Laurent

On Thu, 22 Jan 2026 at 02:49, Laurent Pinchart
<laurent.pinchart@ideasonboard.com> wrote:
>
> Hi Ricardo,
>
> Thank you for the patch.
>
> On Wed, Jan 14, 2026 at 10:32:15AM +0000, Ricardo Ribalda wrote:
> > Make the code slightly more appealing by making use of min(). There
> > shall not be any functional change from this patch.
> >
> > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org>
> > ---
> >  drivers/media/usb/uvc/uvc_video.c | 4 +---
> >  1 file changed, 1 insertion(+), 3 deletions(-)
> >
> > diff --git a/drivers/media/usb/uvc/uvc_video.c b/drivers/media/usb/uvc/uvc_video.c
> > index 59eb95a4b70c05b1a12986e908b7e9979b064fd0..db02080f15772e0bc1d5cfcadd32463f4e6ea045 100644
> > --- a/drivers/media/usb/uvc/uvc_video.c
> > +++ b/drivers/media/usb/uvc/uvc_video.c
> > @@ -1808,9 +1808,7 @@ static int uvc_alloc_urb_buffers(struct uvc_streaming *stream,
> >        * Compute the number of packets. Bulk endpoints might transfer UVC
> >        * payloads across multiple URBs.
> >        */
> > -     npackets = DIV_ROUND_UP(size, psize);
> > -     if (npackets > UVC_MAX_PACKETS)
> > -             npackets = UVC_MAX_PACKETS;
> > +     npackets = min(UVC_MAX_PACKETS, DIV_ROUND_UP(size, psize));
>
> Do you think this improves readability ? I find the existing code easier
> to read, its purpose is immediately clear: it computes npackets and
> clamps it to a max value. With min() I have to pause and think.

I guess it is a matter of taste. I left it as a separate patch for that reason.

>
> I'll take patches 1/3 and 2/3 in my tree already as 1/3 fixes an issue.
Thanks!


Regards!

>
> >
> >       /* Retry allocations until one succeed. */
> >       for (; npackets > 0; npackets /= 2) {
>
> --
> Regards,
>
> Laurent Pinchart



-- 
Ricardo Ribalda