EXA support for nv driver

Radek Doulík rodo at novell.com
Thu Aug 25 02:37:13 PDT 2005

On Wed, 2005-08-24 at 17:53 -0400, Adam Jackson wrote:
> On Wednesday 24 August 2005 09:00, Christian Parpart wrote:
> > On Wednesday 24 August 2005 10:18, Lars Knoll wrote:
> > > Hi,
> > >
> > > I've hacked the nvidia driver lately to add EXA support to it, and at the
> > > same time cleaned up some (but far from all) of the number magic in
> > > nv_hw.c. I also added some documentation on how to initialize the HW to
> > > be able to do DMA (at least as far as I understood it ;-)
> >
> > [...lots.of.cool.stuff...]
> >
> > If I understand you correctly, this is the "nv" driver you've got patched
> > right? not the "nvidia" (partly-binary) one.
> Correct.  nv is the open driver, nvidia is the closed one.
> > If so, will there be any chance to get them into the "nvidia" one as well?
> > I mean, I wouldn't like to loos 3D support just because of this :)
> This makes no sense.  You can't load both the nv and nvidia drivers for the 
> same card at once.
> exa support in the nvidia driver is totally up to nVidia; but since their 
> internal acceleration architecture is both comparable to and similar 
> performance to exa, I don't see why they'd bother.
> Problems with Render acceleration in the nvidia driver should obviously be 
> addressed to nVidia.

I think the Render acceleration in nvidia driver is only experimental
and disabled by default. (at least the last time I checked)

BTW, how is EXA related to Render extension? Does Render extension use
EXA or are they rather 2 independent things?


More information about the xorg mailing list