Bug 40609 - [llvmpipe] piglit glsl-array-bounds-0[26] failure on 64-bit and fbo
Summary: [llvmpipe] piglit glsl-array-bounds-0[26] failure on 64-bit and fbo
Status: CLOSED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: Other (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium normal
Assignee: Brian Paul
QA Contact:
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-09-03 12:55 UTC by Vinson Lee
Modified: 2011-09-16 10:45 UTC (History)
1 user (show)

See Also:
i915 platform:
i915 features:


Attachments

Description Vinson Lee 2011-09-03 12:55:05 UTC
mesa: 0a00a9a05b357dafae86bf8af879aa601f101eba (master)

The piglit glsl-array-bounds-02 and glsl-array-bounds-06 tests with the -fbo option fail on 64-bit llvmpipe.

The tests pass when run without the -fbo option.
The tests pass when run on 32-bit llvmpipe.
The tests pass on swrast.
The tests pass on sofpipe.

$ ./bin/shader_runner tests/shaders/glsl-array-bounds-02.shader_test -fbo
Probe at (15,15)
  Expected: 0.000000 1.000000 0.000000
  Observed: 1.000000 0.000000 0.000000
PIGLIT: {'result': 'fail' }
Comment 1 Brian Paul 2011-09-12 13:22:03 UTC
I can't reproduce this, Vinson.  I'm on a 64-bit system and I tried Mesa 0a00a9a05b357dafae86bf8af879aa601f101eba and today's head (fb5995cf53000af3bd32ca6effb4b08c4aaffe17) and both glsl-array-bounds-02 and glsl-array-bounds-06 pass with -fbo here.

Maybe rebuild with the latest sources and retest?
Comment 2 Vinson Lee 2011-09-12 14:59:44 UTC
mesa: fb5995cf53000af3bd32ca6effb4b08c4aaffe17 (master)
It's passing again at the above commit.


I looked through past test results.
1465dc26a540e2b6f8a055cf0ab48f774a3775bf - good
9222c497f9e73ae19b1b4dab1d9c9b4aa797b11d - bad


It must have been fixed by one of these commits. Does that make sense?

commit 1465dc26a540e2b6f8a055cf0ab48f774a3775bf
Author: Maarten Lankhorst <m.b.lankhorst@gmail.com>
Date:   Sat Sep 10 09:55:55 2011 +0200

    nouveau: Add support for XvMC acceleration
    
    Try to use the PMPEG where available
    
    Signed-off-by: Maarten Lankhorst <m.b.lankhorst@gmail.com>

commit 10dbd029279dda1689410d8ef2bc5aba64dd5958
Author: Bryan Cain <bryancain3@gmail.com>
Date:   Fri Sep 2 00:10:50 2011 -0500

    glsl_to_tgsi: use UARL instead of I2F and ARL
    
    Since TGSI now has a UARL opcode that takes an integer as the source, it is
    no longer necessary to hack around the lack of an integer ARL opcode using I2F.
    UARL is only emitted when native integers are enabled; ARL is still used
    otherwise.
    
    Reviewed-by: Brian Paul <brianp@vmware.com>

commit 324ac982d8e7c2693035342de2d24baff3042911
Author: Bryan Cain <bryancain3@gmail.com>
Date:   Sat Sep 10 12:31:54 2011 -0500

    gallium: add TGSI opcodes UARL and UCMP
    
    They are needed by glsl_to_tgsi for an efficient implementation using native
    integers.
    
    Reviewed-by: Brian Paul <brianp@vmware.com>
Comment 3 Brian Paul 2011-09-16 06:41:59 UTC
None of those commits seem relevant to the bug.  I don't really feel like spending time investigating so let's just close the bug, if that's OK w/ you.
Comment 4 Vinson Lee 2011-09-16 10:44:50 UTC
mesa: 72e0a9226ddf9282a0351e71c0272882473c836d (master)

The tests pass here. Closing bug.


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.