Hi Hans,
On Tue, Dec 20, 2022 at 10:22 AM Hans Verkuil <hverkuil-cisco@xxxxxxxxx> wrote:
> On 20/12/2022 10:09, Geert Uytterhoeven wrote:
> > On Tue, Dec 20, 2022 at 10:01 AM Hans Verkuil <hverkuil-cisco@xxxxxxxxx> wrote:
> >> On 19/12/2022 22:47, Laurent Pinchart wrote:
> >>> (CC'ing Sakari and Hans)
> >>>
> >>> Thank you for the patch.
> >>>
> >>> On Mon, Dec 19, 2022 at 04:01:39PM +0200, Tomi Valkeinen wrote:
> >>>> Add new pixel formats: RGBX1010102, RGBA1010102, ARGB2101010 and Y210.
> >>>>
> >>>> Signed-off-by: Tomi Valkeinen <tomi.valkeinen+renesas@xxxxxxxxxxxxxxxx>
> >>>> ---
> >>>> drivers/gpu/drm/rcar-du/rcar_du_kms.c | 24 +++++++++++++
> >>>> drivers/gpu/drm/rcar-du/rcar_du_vsp.c | 49 +++++++++++++++++++++++++--
> >>>> 2 files changed, 71 insertions(+), 2 deletions(-)
> >>>>
> >>>> diff --git a/drivers/gpu/drm/rcar-du/rcar_du_kms.c b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> index 8c2719efda2a..8ccabf5a30c4 100644
> >>>> --- a/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> +++ b/drivers/gpu/drm/rcar-du/rcar_du_kms.c
> >>>> @@ -259,6 +259,24 @@ static const struct rcar_du_format_info rcar_du_format_infos[] = {
> >>>> .bpp = 32,
> >>>> .planes = 1,
> >>>> .hsub = 1,
> >>>> + }, {
> >>>> + .fourcc = DRM_FORMAT_RGBX1010102,
> >>>
> >>> Ah, here the format makes sense.
> >>>
> >>>> + .v4l2 = V4L2_PIX_FMT_XBGR2101010,
> >>>
> >>> But this is horrible :-( Could we use the same names as DRM for new
> >>> formats, when there is no conflict with existing V4L2 formats ?
> >>>
> >>> Sakari, Hans, what do you think ? Please see patch 1/7 in the series for
> >>> the format definitions.
> >>
> >> V4L2 describes pixel formats based on how they appear in memory from the
> >> lowest to highest memory address.
> >
> > So that means big endian?
>
> Yes.
>
> >> If I am not mistaken, DRM uses the CPU order. So that explains the difference
> >> in naming. I don't think we should hide that difference. And V4L2 has been
> >> quite consistent in following memory ordering in the naming (except possibly
> >> for some of the really old pixelformats).
> >
> > DRM uses little endian.
>
> So not CPU order, but always little endian order? I.e., on a big endian system
> a given DRM_FORMAT_ would have the same memory layout as on a little endian
> system?
Indeed. Big-endian formats must set the DRM_FORMAT_BIG_ENDIAN flag:
#define DRM_FORMAT_BIG_ENDIAN (1U<<31) /* format is big endian
instead of little endian */
unless the big-endian format has a standard (little-endian) equivalent:
Old PPC drivers may violate that, so there is some quirk handling...
/*
* DRM formats are little endian. Define host endian variants for the
* most common formats here, to reduce the #ifdefs needed in drivers.
*
* Note that the DRM_FORMAT_BIG_ENDIAN flag should only be used in
* case the format can't be specified otherwise, so we don't end up
* with two values describing the same format.
*/
#ifdef __BIG_ENDIAN
# define DRM_FORMAT_HOST_XRGB1555 (DRM_FORMAT_XRGB1555 | \
DRM_FORMAT_BIG_ENDIAN)
# define DRM_FORMAT_HOST_RGB565 (DRM_FORMAT_RGB565 | \
DRM_FORMAT_BIG_ENDIAN)
# define DRM_FORMAT_HOST_XRGB8888 DRM_FORMAT_BGRX8888
# define DRM_FORMAT_HOST_ARGB8888 DRM_FORMAT_BGRA8888
#else
# define DRM_FORMAT_HOST_XRGB1555 DRM_FORMAT_XRGB1555
# define DRM_FORMAT_HOST_RGB565 DRM_FORMAT_RGB565
# define DRM_FORMAT_HOST_XRGB8888 DRM_FORMAT_XRGB8888
# define DRM_FORMAT_HOST_ARGB8888 DRM_FORMAT_ARGB8888
#endif
However, given the number of bugs related to big-endian formats,
I doubt DRM has any real use on big-endian hardware (i.e. not
counting hobbyists trying to migrate from fbdev to DRM ;-)
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
