Created attachment 105405 [details] dmesg ==System Environment== -------------------------- Regression: No. Only failed on -niggtly branch Non-working platforms: BSW ==kernel== -------------------------- origin/drm-intel-nightly: 257d90d13794c2eb545ab0d6c708f21e2a0378b6(fails) drm-intel-nightly: 2014y-08m-21d-10h-03m-09s integration manifest origin/drm-intel-next-queued: 3a5f87c286515c54ff5c52c3e64d0c522b7570c0(works) drm: fix plane rotation when restoring fbdev configuration origin/drm-intel-fixes: 1a125d8a2c22b11741fc47d4ffcf7a5ffa044dd3(works) drm/i915: don't try to retrain a DP link on an inactive CRTC origin/drm-fixes: c39502c5508fb0106167f80a037c444701faa65f(works) Merge branch 'drm-fixes-3.17' of git://people.freedesktop.org/~agd5f/linux origin/drm-next: 7963e9db1b1f842fdc53309baa8714d38e9f5681(works) Revert "drm: drop redundant drm_file->is_master" ==Bug detailed description== igt/gem_evict_everything some subcases skip Case list: igt/gem_evict_everything/swapping-interruptible igt/gem_evict_everything/swapping-normal Output: [root@x-bsw01 tests]# time ./gem_evict_everything --run-subtest swapping-normal IGT-Version: 1.7-g50166d2 (x86_64) (Linux: 3.17.0-rc1_drm-intel-nightly_257d90_20140825+ x86_64) Estimated that we need 1611399168 bytes for the test, but only have 41943040 bytes available (RAM) Test requirement not met in function swapping_evictions, file eviction_common.c:144: Test requirement: intel_check_memory(working_surfaces, surface_size, CHECK_RAM) Subtest swapping-normal: SKIP ==Reproduce steps== ---------------------------- 1. ./gem_evict_everything --run-subtest swapping-normal
Only 4M after purging caches, that is very suspect (i.e. there is a kernel memory leak such as the userptr mm-deadlock). Does this test skip when run on a clean boot? Can you use: IGT_LOG_LEVEL=debug ./gem_evict_everything --run-subtest swapping-normal
Created attachment 105417 [details] dmesg (In reply to comment #1) > Only 4M after purging caches, that is very suspect (i.e. there is a kernel > memory leak such as the userptr mm-deadlock). Does this test skip when run > on a clean boot? > > Can you use: IGT_LOG_LEVEL=debug ./gem_evict_everything --run-subtest > swapping-normal [root@x-bsw01 tests]# IGT_LOG_LEVEL=debug ./gem_evict_everything --run-subtest swapping-normal IGT-Version: 1.7-g75365eb (x86_64) (Linux: 3.17.0-rc1_drm-intel-nightly_257d90_20140825+ x86_64) Checking 2118 surfaces of size 1048576 bytes (total 2221969408) against RAM + swap Checking 1536 surfaces of size 1048576 bytes (total 1611399168) against RAM Checking 2118 surfaces of size 1048576 bytes (total 2221969408) against RAM + swap Subtest swapping-normal: SUCCESS
commit 77520c8fccc1a599ff7a7b786dacdfdad38b96e4 Author: Chris Wilson <chris@chris-wilson.co.uk> Date: Thu Aug 7 14:20:40 2014 +0100 drm/i915: Prevent recursive deadlock on releasing a busy userptr
Verified on latest -nightly(7101d84020f63f1da7e0c5d021cdd6be4d515de5) [root@x-bsw01 tests]# ./gem_evict_everything --run-subtest swapping-normal IGT-Version: 1.8-g32a0308 (x86_64) (Linux: 3.17.0-rc6_drm-intel-nightly_7101d8_20140929+ x86_64) Subtest swapping-normal: SUCCESS (142.798s) [root@x-bsw01 tests]# ./gem_evict_everything --run-subtest swapping-interruptible IGT-Version: 1.8-g32a0308 (x86_64) (Linux: 3.17.0-rc6_drm-intel-nightly_7101d8_20140929+ x86_64) Subtest swapping-interruptible: SUCCESS (143.629s)
Moving old bug from Verified to Closed.
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.