Re: [PATCH] rcar-vin: Limit NV12 availability to supported VIN channels only

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Hi Niklas,

On Thu, Nov 7, 2019 at 8:47 AM Niklas Söderlund
<niklas.soderlund+renesas@xxxxxxxxxxxx> wrote:
> On 2019-11-07 08:41:11 +0100, Geert Uytterhoeven wrote:
> > On Thu, Nov 7, 2019 at 12:25 AM Niklas Söderlund
> > <niklas.soderlund+renesas@xxxxxxxxxxxx> wrote:
> > > When adding support for NV12 it was overlooked that the pixel format is
> > > only supported on some VIN channels. Fix this by adding a check to only
> > > accept NV12 on the supported channels (0, 1, 4, 5, 8, 9, 12 and 13).
> > >
> > > Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
> >
> > Thanks for your patch!
> >
> > > --- a/drivers/media/platform/rcar-vin/rcar-v4l2.c
> > > +++ b/drivers/media/platform/rcar-vin/rcar-v4l2.c
> > > @@ -76,7 +76,12 @@ const struct rvin_video_format *rvin_format_from_pixel(struct rvin_dev *vin,
> > >         if (vin->info->model == RCAR_M1 && pixelformat == V4L2_PIX_FMT_XBGR32)
> > >                 return NULL;
> > >
> > > -       if (pixelformat == V4L2_PIX_FMT_NV12 && !vin->info->nv12)
> > > +       /*
> > > +        * If NV12 is supported it's only supported on some channels (0, 1, 4,
> > > +        * 5, 8, 9, 12 and 13).
> >
> > Is this true for all SoCs, or do you need a vin->info->model == RCAR_GEN3
> > check?
>
> NV12 is only supported by most Gen3 SoCs, but no extra check is needed
> as vin->info->nv12 is only set for the Gen3 SoCs that can support NV12.

Thanks, had missed the meaning of the vin->info->nv12 check.

> > > +        */
> > > +       if (pixelformat == V4L2_PIX_FMT_NV12 &&
> > > +           (!vin->info->nv12 || BIT(vin->id) & 0xcccc))
> > >                 return NULL;
> >
> > So 0xcccc = ~(BIT(0) | BIT(1) | BIT(4) | ...)?
>
> Yes.
>
> > What if you ever have an id larger than 15?
> > Wouldn't it be safer to check for !(BIT(vin->id) & 0x3333)?
>
> There is no SoC with more then 16 VIN instances, today... Maybe your
> suggestion of the inverted check makes more sens. Will respin a v2.

OK.  BTW, the code may look nicer if you start using a
"switch (pixelformat) { ... }" block to handle all special cases.

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@xxxxxxxxxxxxxx

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds



[Index of Archives]     [Linux Input]     [Video for Linux]     [Gstreamer Embedded]     [Mplayer Users]     [Linux USB Devel]     [Linux Audio Users]     [Linux Kernel]     [Linux SCSI]     [Yosemite Backpacking]

  Powered by Linux