Re: [PATCH] media: rcar-vin: Return unused buffers correctly

From: Niklas Söderlund

Date: Sat Nov 21 2026 - 03:38:18 EST


Hi Jacopo,

Thanks for your review.

On 2026-11-16 11:21:36 +0100, Jacopo Mondi wrote:
> Hi Niklas
>
> On Fri, Aug 28, 2026 at 12:37:44PM +0200, Niklas Söderlund wrote:
> > Unused VB2 buffers should be returned with state DONE not ERROR when
> > stopping.
>
> Indeed
>
> I think this is worth a fixes tag
>
> I'll add, if you agree:
>
> Fixes: 63a71dd813e2 ("media: rcar-vin: Break out hardware start and stop to new methods")
> Cc: stable@xxxxxxxxxxxxxxx
>
> Not 100% sure what commit to use in the Fixes tag as unused buffers
> were returned in ERROR state before that commit, possibily from the
> very first commit that introduced the driver

If a fixes tag is needed I indeed think it needs be be back to when the
driver was added. But is it worth the churn? I will leave it out for
now.

>
> >
> > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
> > ---
> > drivers/media/platform/renesas/rcar-vin/rcar-dma.c | 2 +-
> > 1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > diff --git a/drivers/media/platform/renesas/rcar-vin/rcar-dma.c b/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > index 73cda0e2d45a..f3003dd61890 100644
> > --- a/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > +++ b/drivers/media/platform/renesas/rcar-vin/rcar-dma.c
> > @@ -1449,7 +1449,7 @@ static void rvin_stop_streaming_vq(struct vb2_queue *vq)
> > dma_free_coherent(vin->dev, vin->format.sizeimage, vin->scratch,
> > vin->scratch_phys);
> >
> > - return_unused_buffers(vin, VB2_BUF_STATE_ERROR);
> > + return_unused_buffers(vin, VB2_BUF_STATE_DONE);
>
> Does the same apply to buffers queued to the hardware and not
> processed released in rvin_stop_streaming() ?
>
> /* Return unprocessed buffers from hardware. */
> for (unsigned int i = 0; i < HW_BUFFER_NUM; i++) {
> if (vin->buf_hw[i].buffer)
> vb2_buffer_done(&vin->buf_hw[i].buffer->vb2_buf,
> VB2_BUF_STATE_ERROR);
> }

Indeed it should.

>
> (nit: there is a stray empty line after this loop, if this hunk needs
> to be patches as well, it might be a good occasion to remove it as
> well)

Nice catch, this in itself is worth a respin ;-)

>
> Thanks
> j
>
>
> > }
> >
> > static const struct vb2_ops rvin_qops = {
> >
> > ---
> > base-commit: 4900cad020c0580dfb1be27776ff10a4ef110cfa
> > change-id: 20260828-vin-fix-vb2-28e1bb24a618
> >
> > Best regards,
> > --
> > Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
> >
> >

--
Kind Regards,
Niklas Söderlund