==System Environment== -------------------------- Regression: no, new case Non-working platforms: all ==kernel== -------------------------- drm-intel-fixes/e3e6efcd638d3caa18d9e89f235c7d6fc32c1f0b ==Bug detailed description== ----------------------------- More than 1000+ new cases with (bcs) and (rcs) fail, due to unexpectd token ‘（’。 output: -bash: syntax error near unexpected token `(' ==Reproduce steps== ---------------------------- 1. ./gem_concurrent_blit --run-subtest cpu-bcs-early-read-forked-hang(bcs)
I'm guessing it's caused by IGT commit 16bafdf5bf0248c02ea9824aca003b2a23d464be Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Thu Sep 4 09:26:24 2014 +0100 igt/gem_concurrent_blit: Inject hangs before verifying contents Please roll back to an earlier commit of IGT as a quick workaround.
Pardon? The bug is in the test runner...
Sure, but isn't this the change that screwed up the test runner? Or is there a change that I missed in the test runner that regressed?
QA's test runner appears to be a poorly written bash script.
Or more generally, uses bash to execute a test in a very loose manner.
Did it ever worked? It is not listed on subtests and here it doesn't work even before this commit 16bafdf5b. If this is really a different test case why don't we use: cpu-bcs-early-read-forked-hang_bcs instead? This would keep igt standard and list it as a sub test.
Yes. The bug is in how the test is executed. It is an utterly trivial bug.
(In reply to Rodrigo Vivi from comment #6) > Did it ever worked? > > It is not listed on subtests and here it doesn't work even before this > commit 16bafdf5b. > > If this is really a different test case why don't we use: > cpu-bcs-early-read-forked-hang_bcs instead? This would keep igt standard and > list it as a sub test. It always has this issue. I also test commit before 16bafdf5b.
*** Bug 88349 has been marked as a duplicate of this bug. ***
gem_concurrent_blit has 1575 subcase, 1103 has '('. We add '\', the case is executed, ./gem_concurrent_blit --run-subtest cpu-bcs-early-read-forked-hang\(rcs\) 1575 subcase, one sub case takes 3min on average. On this speed, we need 80+ hours for whole case gem_concurrent_blit.
commit fbcc7ba20f5a45a6e2fc9e74a68608ebc804ec29 Author: Daniel Vetter <daniel.vetter@ffwll.ch> Date: Thu Jan 22 09:43:10 2015 +0100 tests/gem_concurrent_blt: Adjust subtest naming I guess the amount of tests should be tracked somewhere else ... No idea yet what to do there.
Verified.Fixed. [root@x-ivb9 tests]# ./gem_concurrent_blit --run-subtest prw-rcs-overwrite-source-forked-hang-blt IGT-Version: 1.9-gfa94e01 (x86_64) (Linux: 3.19.0-rc5_drm-intel-nightly_67e9eb_20150125_debug+ x86_64) using 2x512 buffers, each 1MiB Subtest prw-rcs-overwrite-source-forked-hang-blt: SUCCESS (65.689s)
Closing old verified.
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.