Bisected to: bc0af44a5ae9a6671cc6c2f5a380d86fd7babacb Author: Timothy Arceri <tarceri@itsqueeze.com> mesa: add KHR_no_error support for gl{Compressed}TexImage*D() Affects SNB and later: /tmp/build_root/m32/lib/piglit/bin/shader_runner /tmp/build_root/m32/lib/piglit/tests/spec/arb_shader_bit_encoding/execution/and-clamp.shader_test -auto -fbo Probe color at (2,2) Expected: 0.000000 0.250000 0.500000 -1553197254024903250500691171475456.000000 Observed: 0.000000 0.250980 0.501961 0.000000 Test failure on line 43 Mesa CI tests 64bit builds by default. 32bit binaries are checked on a daily basis, so this test failure could not be caught earlier.
(In reply to Mark Janes from comment #0) > Bisected to: > bc0af44a5ae9a6671cc6c2f5a380d86fd7babacb Is the failure intermittent? I suspect the bisect may have come to the wrong conclusion. That commit just adds KHR_no_error paths so there should be no way for it to case any regressions.
The bug is not intermittent for my system.
Ok I was able to reproduce this. It seems like some kind of gcc bug. If I remove the ALWAYS_INLINE from teximage() then the test passes.
(In reply to Timothy Arceri from comment #3) > Ok I was able to reproduce this. > > It seems like some kind of gcc bug. If I remove the ALWAYS_INLINE from > teximage() then the test passes. I should add that I reproduced the issue with gcc 6.3.1 and Mark confirmed it was still present when using gcc 7.1
I've created a gcc bug report and added the url here.
-- GitLab Migration Automatic Message -- This bug has been migrated to freedesktop.org's GitLab instance and has been closed from further activity. You can subscribe and participate further through the new bug through this link to our GitLab instance: https://gitlab.freedesktop.org/mesa/mesa/issues/1608.
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.