Re: i915 GPU Hang on First Login with Display Manager

[Jonathan Nieder <jrnieder@gmail.com>]

reassign 637511 src:linux linux-2.6/3.0.0-1
affects 637511 + xserver-xorg-video-intel
quit

Hi P. J.,

P. J. McDermott wrote:
> On 09/14/2011 02:19 PM, Jonathan Nieder wrote:

>> "Lost in the mists of time" means "Could be discovered by finding out
>> how it regressed and staring at the old code".  Because I don't
>> remember, I'll just ask: does the 845G use KMS in squeeze?
[...]
> I would assume it does, but I don't currently have access to the machine
> to tell for sure.
[...]
> Next time I have access to the machine, I can check for KMS.  And if I
> have time and it might help, I can build a new Linux image with
> additional debugging options (let me know which, if any) enabled in the
> configuration.

Thanks for the offer.  Since then, upstream has found a fix.

| commit c7f7dd61fd07dbf938fc6ba711de07986d35ce1f
| Author: Chris Wilson <chris@chris-wilson.co.uk>
| Date:   Wed Dec 12 19:43:19 2012 +0000
|
|     sna: Pin some batches to avoid CS incoherence on 830/845
|
|     Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=26345
|     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
|
| I can't promise the incoherence won't show up elsewhere as render
| corruption, but my 845g is finally surviving wtf stress tests.

I'm not sure what tree has the patch or what its dependencies are,
though.  If you want to test, you'll need to enable sna by setting
AccelMethod in xorg.conf (see "man intel").

Thanks again for your help tracking it down.
Jonathan