On Mon, 11 Apr 2011, Aguirre, Sergio wrote:
> Hi Guennadi,
>
> On Mon, Apr 11, 2011 at 3:58 AM, Guennadi Liakhovetski <
> g.liakhovetski@xxxxxx> wrote:
>
> > A recent patch has given individual soc-camera host drivers a possibility
> > to calculate .sizeimage and .bytesperline pixel format fields internally,
> > however, some drivers relied on the core calculating these values for
> > them, following a default algorithm. This patch restores the default
> > calculation for such drivers.
> >
> > Signed-off-by: Guennadi Liakhovetski <g.liakhovetski@xxxxxx>
> > ---
> > diff --git a/drivers/media/video/soc_camera.c
> > b/drivers/media/video/soc_camera.c
> > index 4628448..0918c48 100644
> > --- a/drivers/media/video/soc_camera.c
> > +++ b/drivers/media/video/soc_camera.c
> > @@ -376,6 +376,9 @@ static int soc_camera_set_fmt(struct soc_camera_device
> > *icd,
> > dev_dbg(&icd->dev, "S_FMT(%c%c%c%c, %ux%u)\n",
> > pixfmtstr(pix->pixelformat), pix->width, pix->height);
> >
> > + pix->bytesperline = 0;
> > + pix->sizeimage = 0;
> > +
> > /* We always call try_fmt() before set_fmt() or set_crop() */
> > ret = ici->ops->try_fmt(icd, f);
> > if (ret < 0)
> > @@ -391,6 +394,17 @@ static int soc_camera_set_fmt(struct soc_camera_device
> > *icd,
> > return -EINVAL;
> > }
> >
> > + if (!pix->sizeimage) {
> > + if (!pix->bytesperline) {
> > + ret = soc_mbus_bytes_per_line(pix->width,
> > +
> > icd->current_fmt->host_fmt);
> > + if (ret > 0)
> > + pix->bytesperline = ret;
> > + }
> > + if (pix->bytesperline)
> > + pix->sizeimage = pix->bytesperline * pix->height;
> > + }
> > +
> >
>
> Shouldn't all this be better done in try_fmt?
Not _better_. We might choose to additionally do it for try_fmt too. But
1. we didn't do it before, applications don't seem to care.
2. you cannot store / reuse those .sizeimage & .bytesperline values - this
is just a "try"
3. it would be a bit difficult to realise - we need a struct
soc_mbus_pixelfmt to call soc_mbus_bytes_per_line(), which we don't have,
so, we'd need to obtain it using soc_camera_xlate_by_fourcc().
This all would work, but in any case it would be an enhancement, but not a
regression fix.
Thanks
Guennadi
---
Guennadi Liakhovetski, Ph.D.
Freelance Open-Source Software Developer
http://www.open-technology.de/
--
To unsubscribe from this list: send the line "unsubscribe linux-media" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
