Glucose status/instructions request, (and notes on stale branches)

Matthias Hopf mhopf at suse.de
Mon Oct 22 10:58:06 PDT 2007


On Oct 22, 07 20:31:29 +0300, Daniel Stone wrote:
> On Mon, Oct 22, 2007 at 07:24:26PM +0200, Matthias Hopf wrote:
> > On Oct 18, 07 20:28:45 -0700, Keith Packard wrote:
> > > I've spoken often about the desire to reduce our driver complexity by
> > > looking to GL as a shared rendering API, so I'm on the record as
> > > supporting the general plan here, I'd just like to moderate the flux in
> > > master to keep things more stable. With Mesa moving to Gallium, it may
> > > be reasonable for X to look to Gallium as well, instead of using OpenGL.
> > 
> > I don't think Gallium is the right way to do, because OpenGL is a well
> > standardized API, while Gallium is extremely new and will probably not
> > be used by other OpenGL stacks than Mesa. Though this might be not too
> > problematic if both NVidia and AMD won't support glucose anyway, but
> > will continue to use their own acceleration infrastructure.
> 
> Ultimately, the difference is that one requires us to go through another
> layer and deal with more complex mappings, whether the other one lets us
> do more or less what we want.  If the choice between them comes down to
> whether or not to possibly make a worse technical choice in order to
> support people who are developing closed drivers, then I'm pretty sure I
> know what I'd be going for.

With OpenGL 3.0 it shouldn't be that far off. And as I said, NVidia and
AMD probably don't matter that much. OTOH using Gallium will pretty much
fix this in Mesa forever (or we loose our acceleration architecture).
Don't think that this is a good idea.

Matthias

-- 
Matthias Hopf <mhopf at suse.de>      __        __   __
Maxfeldstr. 5 / 90409 Nuernberg   (_   | |  (_   |__          mat at mshopf.de
Phone +49-911-74053-715           __)  |_|  __)  |__  R & D   www.mshopf.de



More information about the xorg mailing list