DRI2 committed

[Maxim Levitsky]

On Friday, 15 February 2008 21:26:19 Kristian Høgsberg wrote:
> On Fri, Feb 15, 2008 at 1:58 PM, Carl Worth <cworth at cworth.org> wrote:
> > On Fri, 15 Feb 2008 09:57:32 -0500, "=?UTF-8?Q?Kristian_H=C3=B8gsberg?=" wrote:
> >  > Last night I committed the DRI2 work that I've been sitting on for
> >  > quite a while now.
> >
> >  Congratulations!
> >
> >
> >  >  - only intel hardware, limited testing on i965 and up.  Gotta
> >  > reinstall that box...
> >
> >  Is that "you've only tested i965 and that's been limited so far"? Or
> >  "you've tested other intel hardware thoroughly but the i965 test has
> >  only been limited"?
> >
> >  I can perhaps help with some i965 testing.
> >
> 
> Oops, you're right, that was unclear: I've tested on my i830 laptop
> and a i945 sony vaio, but not much on i965.  I'm trying to bring the
> stack up on a i965 system right now, but any help is appreciated.
> 
> >
> >  >  - everything committed on master, except the xf86-video-intel parts,
> >  > which are committed on the intel-batchbuffer branch.
> >
> >  OK, and you mentioned in IRC that you'll look into merging that with
> >  the branches that Eric and I have been working on. That sounds great.
> 
> Yes, I did my work on top of the upstrem intel-batchbuffer branch and
> also wasn't sure if Erics branch introduced performance regressions as
> compared to the upstream branch, so I decided to just merge it there
> as a first step.
> 
> cheers,
> Kristian

I tested this on G965 (Intel desktop motherboard DG965RY)
Doesn't work, I first see screen corruption, and then the system crashes.
Tried to start without DRI2 and the system didn't crash, but the corruption was there too.

I used latest graphical stack : (2.6.24 + dri git + mesa git + xserver git + xf86-video-intel#intel-batchbuffer)

I am in process of studying of intel documentation, and I will try to help fix bugs in the driver.

Best regards,
	Maxim Levitsky

PS:

The most important thing that intel driver misses, and I will do my best to fix this,
is that is doesn't initialize the hardware fully, thus in case of error, only a full reboot
brings graphics back.