| Summary: | [HSW/BDW Regression]igt/kms_flip/vblank-vs-hang causes system hang | ||
|---|---|---|---|
| Product: | DRI | Reporter: | Guo Jinxian <jinxianx.guo> |
| Component: | DRM/Intel | Assignee: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
| Status: | CLOSED FIXED | QA Contact: | Intel GFX Bugs mailing list <intel-gfx-bugs> |
| Severity: | critical | ||
| Priority: | high | CC: | intel-gfx-bugs |
| Version: | DRI git | Keywords: | bisect_pending |
| Hardware: | Other | ||
| OS: | All | ||
| Whiteboard: | |||
| i915 platform: | i915 features: | ||
Cases below also able to reproduce system hang on BDW. kms_flip/flip-vs-modeset-vs-hang kms_flip/flip-vs-modeset-vs-hang-interruptible commit f4cbb3a5f707ae4155beaf103adf50351f6509a0 Author: John Harrison <John.C.Harrison@Intel.com> Date: Fri Dec 5 13:49:34 2014 +0000 drm/i915: Zero fill the request structure There is a general theory that kzmalloc is better/safer than kmalloc, especially for interesting data structures. This change updates the request structure allocation to be zero filled. This also fixes crashes in the reset code. Quoting Mika's patch: "Clean the request structure on alloc. Otherwise we might end up referencing uninitialized fields. This is apparent when we try to cleanup the preallocated request on ring reset, before any request has been submitted to the ring. The request->ctx is foobar and we end up freeing the foobarness." Note that this fixes a regression introduced in commit 9eba5d4a1d79d5094321469479b4dbe418f60110 Author: John Harrison <John.C.Harrison@Intel.com> Date: Mon Nov 24 18:49:23 2014 +0000 drm/i915: Ensure OLS & PLR are always in sync References: https://bugs.freedesktop.org/show_bug.cgi?id=86959 References: https://bugs.freedesktop.org/show_bug.cgi?id=86962 References: https://bugs.freedesktop.org/show_bug.cgi?id=86992 Change-Id: I68715ef758025fab8db763941ef63bf60d7031e2 For: VIZ-4377 Signed-off-by: John Harrison <John.C.Harrison@Intel.com> Reviewed-by: Thomas Daniel <Thomas.Daniel@intel.com> Cc: Mika Kuoppala <mika.kuoppala@intel.com> Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Verified on latest -nightly(bfdd01aa1825aa0068f9236b21362b550f6d630f) [root@x-hsw27 tests]# ./kms_flip --run-subtest vblank-vs-hang IGT-Version: 1.8-g819e68f (x86_64) (Linux: 3.18.0-rc7_drm-intel-nightly_bfdd01_20141208+ x86_64 ) Using monotonic timestamps Beginning vblank-vs-hang on crtc 8, connector 18 1680x1050 60 1680 1784 1960 2240 1050 1053 1059 1089 0x6 0x48 146250 ............................................................ vblank-vs-hang on crtc 8, connector 18: PASSED Beginning vblank-vs-hang on crtc 12, connector 18 1680x1050 60 1680 1784 1960 2240 1050 1053 1059 1089 0x6 0x48 146250 ............................................................ vblank-vs-hang on crtc 12, connector 18: PASSED Beginning vblank-vs-hang on crtc 16, connector 18 1680x1050 60 1680 1784 1960 2240 1050 1053 1059 1089 0x6 0x48 146250 ............................................................ vblank-vs-hang on crtc 16, connector 18: PASSED Subtest vblank-vs-hang: SUCCESS (30.369s) 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.
==System Environment== -------------------------- Regression: Yes. Good commit on -next-queued: 65fa5f773196ccc88c4aae43ece9119f092c20a6(2014_11_26) Non-working platforms: HSW ==kernel== -------------------------- origin/drm-intel-nightly: 2014_12_03(fails) origin/drm-intel-next-queued:d06cced8fa81d56e56ab34e93a83236b151c2a76(fails) drm/i915: Move init_unused_rings to gem_init_hw origin/drm-intel-fixes: b0616c5306b342ceca07044dbc4f917d95c4f825(works) drm/i915: Unlock panel even when LVDS is disabled ==Bug detailed description== ----------------------------- igt/kms_flip/vblank-vs-hang causes system hang. Because system hang, unable to catch dmesg Output: [root@x-hsw27 tests]# ./kms_flip --run-subtest vblank-vs-hang IGT-Version: 1.8-g4e5c16c (x86_64) (Linux: 3.18.0-rc7_drm-intel-nightly_691817_20141202+ x86_64) Using monotonic timestamps Beginning vblank-vs-hang on crtc 8, connector 18 1024x768 60 1024 1048 1184 1344 768 771 777 806 0xa 0x40 65000 ............................................................. vblank-vs-hang on crtc 8, connector 18: PASSED Beginning vblank-vs-hang on crtc 12, connector 18 1024x768 60 1024 1048 1184 1344 768 771 777 806 0xa 0x40 65000 ....................................................Test assertion failure function wait_for_events, file kms_flip.c:1171: Failed assertion: ret > 0 select timed out or error (ret 0) Subtest vblank-vs-hang: FAIL (21.814s) Warning on condition flags != 0 in fucntion check_stop_rings, file drmtest.c:112 i915_ring_stop flags on exit 0x400000ff, can't quiescent gpu cleanly Write failed: Broken pipe ==Reproduce steps== ---------------------------- 1. ./kms_flip --run-subtest vblank-vs-hang