[Mesa-dev] Mesa (master): mesa/st: fix number of ubos being declared in a shader

[Roland Scheidegger]

Am 15.05.2014 14:25, schrieb Jose Fonseca:
> ----- Original Message -----
>> Module: Mesa Branch: master Commit:
>> 3e817e7e56806d8adb8f16c35136045c29908944 URL: 
>> https://urldefense.proofpoint.com/v1/url?u=http://cgit.freedesktop.org/mesa/mesa/commit/?id%3D3e817e7e56806d8adb8f16c35136045c29908944&k=oIvRg1%2BdGAgOoM1BIlLLqw%3D%3D%0A&r=NMr9uy2iTjWVixC0wOcYCWEIYhfo80qKwRgdodpoDzA%3D%0A&m=d14NNjVRGFoM489O%2F3uI7MOVAXXndiX%2ByA49yEXpMHE%3D%0A&s=a6c9fa5849e41346ca1773311daa18b8ca2df2d6eb245de6396cd03f94e7c4dd
>>
>>
>> 
Author: Roland Scheidegger <sroland at vmware.com>
>> Date:   Wed May 14 21:06:23 2014 +0200
>> 
>> mesa/st: fix number of ubos being declared in a shader
>> 
>> Previously the code used the total number of ubos being declared in
>> the linked program (so the ubos of all shaders combined), use the
>> number from the particular shader instead. This fixes an assertion
>> failure with piglit arb_uniform_buffer_object-maxblocks seen in
>> llvmpipe since 8a9f5ecdb116d0449d63f7b94efbfa8b205d826f as it now 
>> emits code for each declared buffer, not just the ones actually
>> used.
> 
> Conform mustpass and a few others conform tests regressed, I believe
> with this change.  Piglit didn't blink though.
> 
>> CC: "10.1 10.2" freedesktop.org>
> 
> This highlights a weakness in the current process: it's hard to know
> before committing what's good candidate for stable, as only after
> commiting will a change get full wide-spread testing.  I think the
> bar should a bit higher though: like potentially fixing issues with
> real apps instead of just making tests happy.
> 

You are quite right. I only did it because the change was quite trivial
(nothing can go wrong right?). It does fix a real issue in llvmpipe -
but only on master