Created attachment 114143 [details] dmesg ==System Environment== -------------------------- Non-working platforms: ILK Regression :No, I didn't find find a good commit. ==kernel== -------------------------- origin/drm-intel-nightly: c2f6e584b215dd0d7e5a8a02716a89e985366ec0(2015-03-08) ==Bug detailed description== ----------------------------- [root@x-pk5 tests]# ./gem_wait --run-subtest render_timeout IGT-Version: 1.9-g07be8fe (x86_64) (Linux: 4.0.0-rc2_drm-intel-nightly_fd3701_20150305+ x86_64) ^[[A2048 iters is enough work Finished with 545667483 time remaining Test assertion failure function render_timeout, file gem_wait.c:217: Failed assertion: gem_bo_wait_timeout(fd, dst2->handle, &negative_timeout) == 0 Last errno: 62, Timer expired error: -62 != 0 Subtest render_timeout failed. **** DEBUG **** Test requirement passed: !igt_run_in_simulation() Test requirement passed: !(gem_bo_wait_timeout(fd, dst->handle, &timeout) == -EINVAL) 2048 iters is enough work Finished with 545667483 time remaining Test assertion failure function render_timeout, file gem_wait.c:217: Failed assertion: gem_bo_wait_timeout(fd, dst2->handle, &negative_timeout) == 0 Last errno: 62, Timer expired error: -62 != 0 **** END **** Subtest render_timeout: FAIL (12.280s) Received signal SIGTERM. This case passed when it was run by piglit. But when I switch to the same kernel and run the same case manually, it fails. ==Reproduce steps== ---------------------------- 1. ./gem_wait --run-subtest render_timeout
http://patchwork.freedesktop.org/patch/43961/
Fixed by commit 762e45836a047323defe9bdbbac534f0675ff027 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Wed Mar 4 18:09:26 2015 +0000 drm/i915: Make WAIT_IOCTL negative timeouts be indefinite again in drm-intel-fixes, thanks for the report.
(In reply to Chris Wilson from comment #1) > http://patchwork.freedesktop.org/patch/43961/ This patch has been seen in nightly latest branch, but the issue still exist.
(In reply to Jani Nikula from comment #2) > Fixed by > > commit 762e45836a047323defe9bdbbac534f0675ff027 > Author: Chris Wilson <chris@chris-wilson.co.uk> > Date: Wed Mar 4 18:09:26 2015 +0000 > > drm/i915: Make WAIT_IOCTL negative timeouts be indefinite again > > in drm-intel-fixes, thanks for the report. This issue still exist on this commit. Reopen.
Well then, update the log outputs...
Created attachment 114236 [details] dmesg on commit 762e45
Created attachment 114237 [details] dmesg on nightly-2015-03-12
It impacts i965 platforms.
And the actual test output.
It fails on all platforms. output on PNV: [root@x-pnv2 tests]# ./gem_wait --run-subtest render_timeout IGT-Version: 1.9-gf59935c (i686) (Linux: 4.0.0-rc4_drm-intel-nightly_3753ea_20150318+ i686) 2048 iters is enough work Finished with 220664141 time remaining Test assertion failure function render_timeout, file gem_wait.c:218: Failed assertion: negative_timeout == 0 Last errno: 62, Timer expired error: -1 != 0 Subtest render_timeout failed. **** DEBUG **** Test requirement passed: !igt_run_in_simulation() Test requirement passed: !(gem_bo_wait_timeout(fd, dst->handle, &timeout) == -EINVAL) 2048 iters is enough work Finished with 220664141 time remaining Test assertion failure function render_timeout, file gem_wait.c:218: Failed assertion: negative_timeout == 0 Last errno: 62, Timer expired error: -1 != 0 **** END **** Subtest render_timeout: FAIL (28.735s) Received signal SIGTERM.
commit ea4aa70f3c83a942ee34868ade1eba126b15d879 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Thu Mar 19 08:36:25 2015 +0000 igt/gem_wait: When waiting for infinity, an infinity remains Incorrect testing of out parameters leads to bug noise... Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89494#c4 Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
(In reply to Chris Wilson from comment #11) > commit ea4aa70f3c83a942ee34868ade1eba126b15d879 > Author: Chris Wilson <chris@chris-wilson.co.uk> > Date: Thu Mar 19 08:36:25 2015 +0000 > > igt/gem_wait: When waiting for infinity, an infinity remains > > Incorrect testing of out parameters leads to bug noise... > > Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=89494#c4 > Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk> Case pass on this igt commit. Change state to verified.
It also fails on 3.19.2 kernel.
It will take some time for the patch to percolate back through stable kernels.
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.