[Mesa-dev] [PATCH] i965: Fix Vertex URB Read Length calculation in 3DSTATE_SF on Gen6.

[Kenneth Graunke]

On 02/02/2013 04:50 AM, Martin Steigerwald wrote:
> Am Samstag, 2. Februar 2013 schrieb Kenneth Graunke:
>> The old calculation was off by one in certain cases.  The documentation
>> contains a precise formula for how to calculate it, so just use that.
>>
>> Fixes random corruption in Steam's Big Picture Mode, random corruption
>> in PlaneShift (since the varying reordering code landed), and Piglit
>> test ________.
>>
>> NOTE: This is a candidate for all stable branches.
>>
>> (The code needs to get cleaned up - the final result is quite ugly - but
>>   I wanted to put it on the list for the people working on these bugs.)
>>
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56920
>> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=60172
>
> Tested-by: Martin Steigerwald <martin at lichtvoll.de>
>
> Works just fine on Sandybridge HD 3000 here. Glitches are gone.
>
> Thanks a lot. Great turn around time from bug report to fix!
>
> Will add info to bug report in a moment.
>
>
> For the fun of it I have another gfx glitch for you:
>
> Bug 60185 - Planeshift: Another gfx glitch with some black area around plants
> https://bugs.freedesktop.org/60185
>
>
> This one has been there for longer. Maybe from beginning of playing the
> game.
>
>
> Aside from this I have some short hang of gfx output from time to time.
> I have no idea on how to provide a meaning ful bug report about it tough,
> cause I have no idea what triggers the bug. Its about 5 or 10 seconds and
> then everything fine again.

Crud.  When that happens, do you see any messages about 'GPU hung' in 
dmesg?  If I made it GPU hang, that's really bad, and means I probably 
didn't get this quite right yet...