Bug 93325 - [HSW,BDW,SKL]ES31-CTS.explicit_uniform_location.uniform-loc-* 2 tests fail
Summary: [HSW,BDW,SKL]ES31-CTS.explicit_uniform_location.uniform-loc-* 2 tests fail
Status: RESOLVED DUPLICATE of bug 92589
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: unspecified
Hardware: Other All
: medium normal
Assignee: Tapani Pälli
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks: 92778
  Show dependency treegraph
 
Reported: 2015-12-10 12:47 UTC by Marta Löfstedt
Modified: 2016-01-18 12:29 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments

Description Marta Löfstedt 2015-12-10 12:47:22 UTC
Software versions:
    4.4.0-040400rc4-generic
    OpenGL version string: 3.0 Mesa 11.2.0-devel (git-83e8e07)

GPU hardware:
    OpenGL renderer string: Mesa DRI Intel(R) HD Graphics 530 (Skylake GT2) 
    00:02.0 VGA compatible controller [0300]: Intel Corporation Sky Lake Integrated Graphics [8086:1912] (rev 06)

CPU hardware:
    x86_64
    Intel(R) Core(TM) i5-6600K CPU @ 3.50GHz
-----------------
Software versions:
    4.3.0-rc3+
    OpenGL version string: 3.0 Mesa 11.2.0-devel (git-83e8e07)

GPU hardware:
    OpenGL renderer string: Mesa DRI Intel(R) HD Graphics 5300 (Broadwell GT2) 
    00:02.0 VGA compatible controller [0300]: Intel Corporation Broadwell-U Integrated Graphics [8086:161e] (rev 06)

CPU hardware:
    x86_64
    Genuine Intel(R) CPU 0000 @ 0.60GHz
------------------
Software versions:
    4.2.0-1-generic
    OpenGL version string: 3.0 Mesa 11.2.0-devel (git-83e8e07)

GPU hardware:
    OpenGL renderer string: Mesa DRI Intel(R) Haswell Desktop 
    00:02.0 VGA compatible controller [0300]: Intel Corporation Xeon E3-1200 v3/4th Gen Core Processor Integrated Graphics Controller [8086:0412] (rev 06)

CPU hardware:
    x86_64
    Intel(R) Core(TM) i7-4770 CPU @ 3.40GHz
-----------------
CTS version:
git@67ae88f31295

command: 
./glcts --deqp-case=ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers
./glcts --deqp-case=ES31-CTS.explicit_uniform_location.uniform-loc-implicit-in-some-stages3

Environment:
Mesa built with: --enable-debug
export MESA_GLES_VERSION_OVERRIDE=3.1
export MESA_EXTENSION_OVERRIDE=GL_ARB_compute_shader

Reason for fail: Wrong color. Expected green, got (1, 0, 0, 1).
Comment 1 Tapani Pälli 2015-12-11 07:53:47 UTC
Just making a note here so I don't forget this. I had brief look at 'uniform-loc-implicit-in-some-stages3' some time ago and my impression is that in this test linker overwrites already set correct location because only some of the stages specify location explicitly, I did not start to dig further yet.
Comment 2 Tapani Pälli 2015-12-14 14:35:04 UTC
I've had a better look and it seems explicit locations are broken since:
dcd9cd03837545055ce2a315e7e8840cc3254d1a

We've been lucky to not hit this because I guess explicit uniform locations are still rare to be utilized. Will make a Piglit test that exposes this and do some fixes.
Comment 3 Marta Löfstedt 2015-12-15 13:56:28 UTC
With mesa@10b7a7c344 and Iago/Samuels patches from BUG 93321

ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers

pass on SKL, but not on BDW and HSW.
Comment 4 Tapani Pälli 2015-12-15 14:26:22 UTC
(In reply to Marta Löfstedt from comment #3)
> With mesa@10b7a7c344 and Iago/Samuels patches from BUG 93321
> 
> ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers
> 
> pass on SKL, but not on BDW and HSW.

Sweet! Today I found that commit dcd9cd03837545055ce2a315e7e8840cc3254d1a is OK after all, it just changes the way how location is read and propagated forward. I'll try to investigate why HSW and BDW fails. It could be still related to the way this commit changes things but not sure.
Comment 5 Marta Löfstedt 2015-12-15 14:40:48 UTC
You will also need:
http://patchwork.freedesktop.org/patch/68192/

I thought I had included it in BUG 93321, but I hadn't since it was not needed for those test-cases.
Comment 6 Tapani Pälli 2016-01-14 11:17:34 UTC
I've sent a fix to:
ES31-CTS.explicit_uniform_location.uniform-loc-implicit-in-some-stages3

http://lists.freedesktop.org/archives/mesa-dev/2016-January/104929.html
Comment 7 Tapani Pälli 2016-01-15 05:36:49 UTC
these 2 should be fixed now by separate patches:

ES31-CTS.explicit_uniform_location.uniform-loc-implicit-in-some-stages3
was fixed by

commit cf96bce0ca8b2d6b5d4641efea1f2feedc024957
Author: Tapani Pälli <tapani.palli@intel.com>
Date:   Thu Jan 14 14:10:59 2016 +0200

    glsl: mark explicit uniforms as explicit in other stages too
    
    If shader declares uniform explicit location in one stage but
    implicit in another, explicit location should be used. Patch marks
    implicit uniforms as explicit if they were explicit in previous stage.
    This makes sure that we don't treat them implicit later when assigning
    locations.
    
    Fixes following CTS test:
       ES31-CTS.explicit_uniform_location.uniform-loc-implicit-in-some-stages3
    
    v2: move check to cross_validate_globals (Timothy)
    
    Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
    Reviewed-by: Timothy Arceri <timothy.arceri@collabora.com>

--- 8< -----------------------------------------------

ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers
was fixed by

commit 18c76551ee425b981efefc61f663a7781df17882
Author: Francisco Jerez <currojerez@riseup.net>
Date:   Sat Jan 2 19:02:09 2016 -0800

    i965/gen6-7: Implement stall and flushes required prior to switching pipelines.
    
    Switching the current pipeline while it's not completely idle or the
    read and write caches aren't flushed can lead to corruption.  Fixes
    misrendering of at least the following Khronos CTS test:
    
     ES31-CTS.shader_image_load_store.basic-allTargets-store-fs
    
    The stall and flushes are no longer required on Gen8+.
    
    v2: Emit PIPE_CONTROL with non-zero post-sync op before the write
        cache flush on SNB due to hardware bug. (Ken)
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=93323
    Reviewed-by: Matt Turner <mattst88@gmail.com>
    Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Comment 8 Marta Löfstedt 2016-01-15 11:21:45 UTC
Sorry, Tapani!

With: Mesa: git@781d2787bc1

./glcts --deqp-case=ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers

fails on both BDW and SKL.
Comment 9 Marta Löfstedt 2016-01-15 11:24:19 UTC
see: BUG 93725
Comment 10 Marta Löfstedt 2016-01-15 11:34:23 UTC
./glcts --deqp-case=ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers

cause GPU_HANG on SKL, with Mesa: git@781d2787bc1

Above, test passed on SKL and BDW with Mesa: git@f4ab7340c
Comment 11 Marta Löfstedt 2016-01-15 12:12:35 UTC
I bisected the SKL regression to:

commit 044acb9256046bebec890cac7e42043754459fc2
Author: Francisco Jerez <currojerez@riseup.net>
Date:   Sat Jan 2 18:35:28 2016 -0800

    i965/gen8+: Invalidate color calc state when switching to the GPGPU pipeline.
Comment 12 Francisco Jerez 2016-01-17 00:13:31 UTC
The regression seems to be another symptom of missing sampler state setup for the compute pipeline, the patch I attached to bug 92589 apparently fixes this too.
Comment 13 Marta Löfstedt 2016-01-18 12:29:31 UTC
I conform that Curros patch from 92589 fix: ES31-CTS.explicit_uniform_location.uniform-loc-types-samplers

*** This bug has been marked as a duplicate of bug 92589 ***


Use of freedesktop.org services, including Bugzilla, is subject to our Code of Conduct. How we collect and use information is described in our Privacy Policy.