[Intel-gfx] [ANNOUNCE] xf86-video-intel 2.7.0

[Stefano Avallone]

On Friday 17 April 2009 17:32:32 Carl Worth wrote:
> On Thu, 2009-04-16 at 20:24 -0700, John Ettedgui wrote:
> > On Thu, Apr 16, 2009 at 11:03 AM, Frederik Himpe <fhimpe at telenet.be>
> > wrote:
> >
> >         On Wed, 15 Apr 2009 18:29:33 -0700, Carl Worth wrote:
> >         > We are pleased to announce a major 2.7.0 release of
> >
> >         xf86-video-intel
> >
> >         > (after three previous 2.6.99.90x release candidates).
> >
> >         This new version is not working for me: the image is
> >         completely corrupted
> >         (except for the mouse cursor), while 2.6.99.903 was working
> >         without any
> >         problem.
> >
> > I have the exact same behavior here.
> > 2.6.29 with KMS enabled and a x3000.
>
> Apparently, Jesse's last-minute fix that made it into xf86-video-intel
> 2.7.0 requires a corresponding fix in the kernel, (which as of yesterday
> was in Eric's tree but not quite into Linus' tree yet---but it might be
> there by now, I haven't checked yet.)

I can confirm that there is no corruption with 2.6.30-rc2 with jbarnes' last-
minute fix manually applied, intel driver 2.7.0 and KMS on GM965. No 
corruption even after a resume from suspend to disk. Great!

However, I am not able to redirect output on VGA connected monitors (it works 
without KMS). I've opened bug #21250 on fd.o with more details.

Also, you might be interested to know that with the current git master of 
xorg, mesa and intel driver there is an additional issue: KDE4 compositing 
manager cannot be enabled when using KMS. 

thanks

>
> So there are some tricky dependencies here that can be hard to get
> right. Hopefully everyone will have a nice, robust driver when all the
> dust settles. And for now, people on older kernels might need to revert
> something in the driver.
>
> I'm sorry we released things in such a delicate state.
>
> -Carl