Hey, On Wed, 24 Feb 2021 at 14:58, Daniel Vetter <daniel@xxxxxxxx> wrote: > On Tue, Feb 23, 2021 at 7:50 PM Daniel Stone <daniel@xxxxxxxxxxxxx> wrote: > > The issue is a userspace one though, not a kernel one. Userspace (e.g. > > GNOME Shell, Weston, Xorg) decides ahead of time that it's going to > > use XRGB8888, then use the modifiers available to it for that format. > > There's no logic in those projects to look at the list of 8bpc > > non-alpha formats, examine XRGB vs. XBGR, decide that XBGR is 'better' > > since it has more modifiers available, then go use XBGR. > > That sounds a bit like userspace being too simple. Since if they're ok > with dealing with modifiers accessing the raw buffer is out the window > anyway, so bgr vs rgb shouldn't matter. Sure, you can add colour-channel permutations, if you have concrete 'rough equivalence' mappings, and you check that between KMS + EGL + anything external you want to use like PipeWire. We pretty much punted on that long ago and just use XRGB for everything. Android punted on that from the beginning and just uses XBGR for everything ... so it's not really a matter of dumb vs. smart userspace, just equally dumb userspaces which disagree with each other. ;) Cheers, Daniel _______________________________________________ dri-devel mailing list dri-devel@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/dri-devel
