Created attachment 90491 [details] dmesg System Environment: -------------------------- Arch: x86_64 Platform: ivybridge Kernel: (drm-intel-nightly)ddd0a9e2da6087c7aac74aca34fcdc90645cbdb4 Bug detailed description: --------------------------- It fails on ivybridge with -fixes kernel, has OOM killer Bug 72506 on -queued and -nightly kernel. It's a new case. output: trying buffer count 522904 trying buffer count 522903 trying buffer count 522902 trying buffer count 522901 trying buffer count 522900 trying buffer count 522899 trying buffer count 522898 trying buffer count 522897 trying buffer count 522896 trying buffer count 522895 trying buffer count 522894 trying buffer count 522893 trying buffer count 522892 trying buffer count 522891 trying buffer count 522890 trying buffer count 522889 reduced buffer count to 522889 from 522906 Test assertion failure function big_exec, file gem_ctx_exec.c:160: Failed assertion: drmIoctl(fd, DRM_IOCTL_I915_GEM_EXECBUFFER2, &execbuf) == 0 Subtest eviction: FAIL Reproduce steps: ------------------------- 1. ./gem_ctx_exec --run-subtest eviction
It also happens on Haswell.
This should be fixed with Chris' patch to do a gpu_idle in evict_something. https://patchwork.kernel.org/patch/3295431/ Please test this patch while Chris is hunting down the commit message for it.
(In reply to comment #2) > This should be fixed with Chris' patch to do a gpu_idle in evict_something. > > https://patchwork.kernel.org/patch/3295431/ > > Please test this patch while Chris is hunting down the commit message for it. Fixed by this patch.
commit ad071acb53110c8efd26ff1e5b5d57449b43833b Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Mon Dec 9 10:37:24 2013 +0000 drm/i915: Repeat eviction search after idling the GPU is merged to -fixes.
Verified.Fixed.
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.