[Mesa-dev] radeonsi: first successful piglit run

[Christian König]

On 31.08.2012 06:55, Michel Dänzer wrote:
> On Don, 2012-08-30 at 13:30 -0400, Alex Deucher wrote:
>> On Thu, Aug 30, 2012 at 1:14 PM, Tom Stellard <tom at stellard.net> wrote:
>>> On Thu, Aug 30, 2012 at 07:00:44PM +0200, Christian König wrote:
>>>> Hi everybody,
>>>>
>>>> with the following patchset the radeonsi driver finally
>>>> completes a run of nearly all tests in Piglits "quick-driver"
>>>> profile without an GPU hang or X server crash.
>>>>
>>>> The only exceptions are:
>>>> 1. glx/glx-tfp
>>>> Results in a reproducible X server crash in the GLAMOR code.
>>>>
>>>> 2. glean/makeCurrent
>>>> Also crashes the X server, but this time it isn't reproducible
>>>> and the backtrace doesn't show a GLAMOR relation.
>>>>
>>>> 3. texturing/texsubimage
>>>> I have now idea why, but this test results in a hardlock of
>>>> the whole system, going to investigate that one closer next week.
>>>>
>>>> The patchset mostly disable features or adds gracefully handling
>>>> for not implement features, so I'm not completely sure if I should
>>>> commit them or instead try to fix all the features in question.
>>> I'm in favor of committing these.  It will really help to be able to
>>> run through a full piglit run, and it'd be nice not to have to manually
>>> apply these patches to every feature branch.
>> I agree.  We can re-enable as features are fixed/implemented.
> Agreed.

I do agree for patch 1,2,3 and 5.

For patch 4 what Tom has send me for LLVM already solves halve of the 
problem, so I think I'm just going to solve the other half today or on 
Monday and if that takes longer than expected just commit the workaround.

And I'm REALLY unhappy with patch 6. As I already wrote Marek it's just 
a workaround for GLAMOR and GLX not playing nicely together inside the 
X-Server.
Maybe I get a good idea over the weekend how to solve that in a better 
way, but I don't really think so.

Christian.