Re: DRM_FORMAT_* byte order (was: Re: [PATCH] drm: virtio: fix virtio_gpu_cursor_formats)

[Gerd Hoffmann <kraxel@redhat.com>]

On Fr, 2017-04-07 at 11:45 +0300, Ville Syrjälä wrote:
> On Fri, Apr 07, 2017 at 10:29:00AM +0200, Gerd Hoffmann wrote:
> >   Hi,
> > 
> > > Hmm. Maybe it's still possible to salvage something by redefining the
> > > BIG_ENDIAN format bit to mean the "the other endianness". Ugly but it
> > > might still result in something usable.
> > 
> > Also at least for the virtual machine use case this doesn't buy us much.
> > The drm drivers (at least the ones used on both big and little endian
> > guests) support only 32 bpp + depth 24 formats.  And for these we don't
> > need a "other endian" flag because we have fourcc codes for all sorts of
> > byte orders (i.e. DRM_FORMAT_XRGB8888 little endian ==
> > DRM_FORMAT_BGRX8888 big endian).
> 
> Yeah, those could be handled without the flag. But when mixed with any
> other format the code would look a bit weird IMO.

Well, there is a reason only the 32bpp formats are supported.  With
those you just adjust your color shifts and you are done.  No need to
actually byte-swap.  In contrast handling 16bpp formats (5:6:5 or 5:5:5)
with the non-native byte order is a PITA.

The other reason of course is that this is the default format these
days.

So, do any "other formats" exist where the byteswapped variant is used
in practice?  Or can we just drop DRM_FORMAT_BIG_ENDIAN?

> So my idea with the
> flag was that if you display is big endian you always have the flag, and
> then you don't have to think so much which way the bytes go for each
> format. Less special casing is good IMO.

Having two valid fourcc (DRM_FORMAT_XRGB8888 + (DRM_FORMAT_BGRX8888 |
DRM_FORMAT_BIG_ENDIAN)) for the same format is confusing IMO.  And I
doubt that it'll be properly implemented everywhere.

cheers,
  Gerd

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel