[BUG] drm/i915 Screen corruption introduced by a00b10c360b35d6431a94cb

[Indan Zupancic]

On Tue, February 22, 2011 04:25, Indan Zupancic wrote:
> Hi,
>
> On Sun, February 20, 2011 14:21, Daniel Vetter wrote:
>> Well, don't start jumping around, yet. These patches are just to rule out
>> some theories. Now: Is it fixed with just the 2nd patch alone or do you need both
>> patches? This is very important, so please test extensively whether there are really
>> no corruptions with just the 2nd patch.
>
> I managed to create some corruption with an xterm above xpdf. It looks  different
> than the original corruption, so I think it's safe to say it's a different bug:
>
> http://img593.imageshack.us/i/ss1298340823.png
> http://img203.imageshack.us/i/ss1298340776.png
>
> This was without xcompmgr running, I don't think I would have seen it otherwise.
> Actually, it turns out it's really easy to trigger as well, so I can more easily
> test patches now. Unfortunately I wasn't smart enough to store the bisecting kernels,
> gah!
>
> I tried with both your patches, as well as the HIC poking patch, but I still get it.
>
> I'll try to pinpoint more exactly when this started to happen.

Okay, I tried older kernels and I get it even with 2.6.34, so I think it's a bug
in userspace, like intel-dri, or perhaps the intel X driver (though I get the same
with an older version).

Greetings,

Indan