xorg intel driver in rawhide

[Dave Airlie]

On Tue, 2008-03-25 at 21:06 +0000, Peter Robinson wrote:
> >  >  > > > Of course that doesn't mean the driver isn't buggy.  I'm considering
> >  >  > > > this change for upstream, if you get a chance maybe you could try it
> >  >  > > > out to see if it fixes your problems?
> >  >  > >
> >  >  > > mplayer seems to behave with that patch, thanks!
> >  >  >
> >  >  > Great, thanks for testing.  Assuming I don't get any "hey this broke my
> >  >  > machine" mails, I'll push it upstream for the 2.3 release.
> >  >
> >  >  Btw, for the morbidly curious, I posted a summary of the backlight issues we
> >  >  have on the Intel side to xorg at lists.freedesktop.org today (subject "Intel
> >  >  backlight info"), feel free to check it out.
> >  >
> >  just like to add that I am also seeing the backlight issue on similar
> >  hardware to the poster.
> 
> I just grabbed the src rpm and the latest git version of the driver. I
> noticed in the process that there's two versions of the intel driver
> plus the old i810 driver (I thought the later was going to go away in
> F9). In the intel ones there one that's called master and one called
> batchbuffer. What's the difference between the two and how does it
> choose which one to load?
> 
> Also hat's the difference between the two different ones and the git
> master repo? Or even the 2.2.99.910?

F9 contains a couple of pieces of work from Red Hat in the Intel driver,
so we ship the normal intel driver as the master one, and a second copy
from intel-kernelmode branch from upstream. 

The intel-kernelmode branch contains DRI2 and kernel modesetting support
for the Intel driver. It gets enabled if you a) boot with modesetting
enabled (i915.modeset=1) on the kernel command line. or b) Add option
"intel-batchbuffer" "true" to the serverflags section.

Batchbuffer branch contains a redesigned Intel driver that uses the
kernel GPU memory manager and supports DRI2.

Dave.