Bug 100441 - glsl/tests/optimization-test.sh regression
Summary: glsl/tests/optimization-test.sh regression
Status: RESOLVED FIXED
Alias: None
Product: Mesa
Classification: Unclassified
Component: glsl-compiler (show other bugs)
Version: git
Hardware: x86-64 (AMD64) All
: medium normal
Assignee: mesa-dev
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords: regression
Depends on:
Blocks:
 
Reported: 2017-03-29 04:21 UTC by Vinson Lee
Modified: 2017-03-29 09:54 UTC (History)
0 users

See Also:
i915 platform:
i915 features:


Attachments
Test suite log (4.89 KB, text/plain)
2017-03-29 07:46 UTC, Michel Dänzer
Details

Description Vinson Lee 2017-03-29 04:21:04 UTC
mesa: 78ee8b3f849063e3e37db0767212397da522b6fa (master 17.1.0-devel)

====================================================
   Mesa 17.1.0-devel: src/compiler/test-suite.log
====================================================

# TOTAL: 10
# PASS:  9
# SKIP:  0
# XFAIL: 0
# FAIL:  1
# XPASS: 0
# ERROR: 0

.. contents:: :depth: 2

FAIL: glsl/tests/optimization-test.sh
Comment 1 Michel Dänzer 2017-03-29 07:46:58 UTC
Created attachment 130526 [details]
Test suite log

This happens for me since commit 86a937d2644a ("glsl/tests/optimization-test: correctly manage srcdir/builddir"), but AFAICT before that the lower_jumps tests just didn't get generated at all for out-of-tree builds. So one would need to use an in-tree build to find the original regression point.
Comment 2 Michel Dänzer 2017-03-29 07:50:30 UTC
Note that I was seeing different intermittent failures with earlier commits, which make clean resolved.
Comment 3 Michel Dänzer 2017-03-29 08:23:28 UTC
Bisecting with an in-tree build identified the same commit as the culprit. Emil, any ideas?
Comment 4 Timothy Arceri 2017-03-29 08:32:50 UTC
Po(In reply to Michel Dänzer from comment #1)
> Created attachment 130526 [details]
> Test suite log
> 
> This happens for me since commit 86a937d2644a
> ("glsl/tests/optimization-test: correctly manage srcdir/builddir"), but
> AFAICT before that the lower_jumps tests just didn't get generated at all
> for out-of-tree builds. So one would need to use an in-tree build to find
> the original regression point.

Was it getting generated for in-tree builds?

I fixed a bug with lower jumps recently, I wonder if that has change the output the test is expecting.

https://cgit.freedesktop.org/mesa/mesa/commit/?id=c1096b7f1d49f33c85b1042f82ccb063f63e7c20
Comment 5 Michel Dänzer 2017-03-29 08:41:24 UTC
Good catch, yes that commit broke it.
Comment 6 Timothy Arceri 2017-03-29 09:31:14 UTC
This updates the test to reflect the new IR:

https://patchwork.freedesktop.org/patch/147012/
Comment 7 Timothy Arceri 2017-03-29 09:54:21 UTC
Fixed by:

commit e44cba540ee7a07bba643bab4b9e519e90244bd1
Author: Timothy Arceri <tarceri@itsqueeze.com>
Date:   Wed Mar 29 20:30:19 2017 +1100

    mesa: update lower_jumps tests after bug fix
    
    This change updates the tests to reflect the IR after
    the following bug fix.
    
    Fixes: c1096b7f1d49 ("glsl: fix lower jumps for returns when loop is
                          inside an if")
    
    Tested-by: Michel Dänzer <michel.daenzer@amd.com>
    Bugzilla: https://bugs.freedesktop.org/100441


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.