Intel Graphics: SNA testing report with xorg-x11-drv-intel >= 2.20

[Sandro Mani]

On 09/03/2012 07:56 PM, Pedro Francisco wrote:
> On Sat, Aug 25, 2012 at 9:49 PM, Sandro Mani <manisandro at gmail.com> wrote:
>> On 08/25/2012 10:08 PM, Pedro Francisco wrote:
>>> On Fri, Aug 24, 2012 at 2:48 PM, Sandro Mani <manisandro at gmail.com> wrote:
>>>> On 08/24/2012 10:40 AM, Pedro Francisco wrote:
>>>>> On Fri, Aug 24, 2012 at 12:00 AM, Pedro Francisco
>>>>> <pedrogfrancisco at gmail.com> wrote:
>>>>>> On Thu, Aug 23, 2012 at 6:53 PM, Sandro Mani <manisandro at gmail.com>
>>>>>> wrote:
>>>>>>> It happens both with and without compositing (actually issues are much
>>>>>>> more
>>>>>>> visible without compositing, possibly due to the fact that various
>>>>>>> compositing effects trigger more screen redraws).
>>>>> Have you ever hit the bug after a system start? Yesterday I couldn't
>>>>> trigger it, today after suspend I can...
>>>> In my case it is hard to reproduce. The issue starts occuring without any
>>>> particular reason after some time, though sometimes I have the feeling
>>>> that
>>>> it is more likely to happen after some heavy openGL load. At the moment I
>>>> am
>>>> testing with xorg-x11-drv-intel-2.20.0-2.20120718.fc19.x86_64 (rebuilt
>>>> for
>>>> new libxcb), and have so far not encountered the issue (uptime 15h).
>>> Can't reproduce it since my last email... grunf...
>> More or less safe reproducer for me: run a graphics-intensive 3D application
>> in windowed mode, and work normally with 2D applications for 10-20 mins in
>> the meantime. Though still not 100% reproducible....
> I've given up trying to reproduce this. Anyway if the bug still is
> present, it's a bug common to UXA and SNA so if it happens again, I
> think it's better to start a new thread with a new title (and add me
> in CC if you do, please).
>
> P.S.: do you have any kind of powersave enabled? I had
> «pcie_aspm=force i915.i915_enable_fbc=1 i915.lvds_downclock=1
> i915.powersave=1» on GRUB and I removed them for further testing,
> unfortunately that was around the time I started to being unable to
> trigger it. I'm readding them just in case it's related.
No, stock options everywhere. I still manage to hit the bug 
occasionally, always apparently caused by heavy 3D load, though never 
reliably ...