Bug 84853 - [ILK/SNB/IVB/HSW Regression]igt gem_userptr_blits/create-destroy-sync causes system hang
Summary: [ILK/SNB/IVB/HSW Regression]igt gem_userptr_blits/create-destroy-sync causes ...
Status: CLOSED FIXED
Alias: None
Product: DRI
Classification: Unclassified
Component: DRM/Intel (show other bugs)
Version: XOrg git
Hardware: All Linux (All)
: high critical
Assignee: Intel GFX Bugs mailing list
QA Contact: Intel GFX Bugs mailing list
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-10-10 03:06 UTC by lu hua
Modified: 2017-07-03 12:24 UTC (History)
2 users (show)

See Also:
i915 platform:
i915 features:

Attachments
dmesg (4.51 KB, text/plain)
2014-10-10 03:06 UTC, lu hua 		no flags Details
View All

Description lu hua 2014-10-10 03:06:22 UTC 
Created attachment 107640 [details]
dmesg

==System Environment==
--------------------------
Regression: Yes. 
Good commit on -next-queued: 6e47e3f097cc6c4cb470a805a3fa07a8e8376dab
bad commit: a128efeaa64609a9672b51bb37bb703e1b0f0128

==kernel==
--------------------------
drm-intel-nightly/ea4bec8e96ea8b33b49a7892c1c7f20041a56da6

==Bug detailed description==
It causes system hang on ILK with -fixes and -queued kernel, works well on -nightly kernel.

output:
IGT-Version: 1.8-gc7551bf (i686) (Linux: 3.17.0_drm-intel-fixes_bfe01a_20141009+ i686)
Aperture size is 2048 MiB
Total RAM is 1945 MiB
Testing unsynchronized mappings...
Testing synchronized mappings...

Call trace:
[  168.711399] Call Trace:
[  168.715426]  [<f8573d3d>] ? __i915_mm_struct_free__worker+0x1a/0x3b [i915]
[  168.719533]  [<c103fbdd>] ? process_one_work+0x163/0x2ab
[  168.723661]  [<c103007b>] ? copy_process.part.39+0x80b/0x12ec
[  168.727814]  [<c103ff53>] ? worker_thread+0x211/0x2f4
[  168.731984]  [<c103fd42>] ? process_scheduled_works+0x1d/0x1d
[  168.736166]  [<c1042b1b>] ? kthread+0xa1/0xa6
[  168.740340]  [<c16fa5e1>] ? ret_from_kernel_thread+0x21/0x30
[  168.744516]  [<c1042a7a>] ? kthread_freezable_should_stop+0x3b/0x3b
[  168.748720] Code: 5f 8b 1b 85 db 75 e4 8b 14 24 83 c4 08 b8 20 98 f9 c2 5b 5e 5f 5d e9 df 6c f8 ff 57 56 89 c6 53 8b 42 28 89 d3 85 c0 7f 02 0f 0b <83> 7e 04 00 74 58 b8 20 98 f9 c2 e8 98 6c f8 ff 89 c7 8b 46 08
[  168.758209] EIP: [<c10de514>] mmu_notifier_unregister+0x10/0x97 SS:ESP 0068:c0077f18
[  168.762785] CR2: 000000000000000c
[  168.767349] ---[ end trace 6620fa4b34e3f6ac ]---
[  168.772001] BUG: unable to handle kernel paging request at ffffffec

==Reproduce steps==
---------------------------- 
1. ./gem_userptr_blits --run-subtest create-destroy-sync
Comment 1 lu hua 2014-10-10 03:36:07 UTC 
bug 81832 is about "long time to execute", In fact, system hang on some platforms, report this bug to track system hang.
Comment 2 Chris Wilson 2014-10-10 06:10:49 UTC 
The fix is in drm-intel-fixes-next queued for the next stable release.
Comment 3 lu hua 2014-10-15 06:36:20 UTC 
I think keep it open untill fixed on -fixes branch.
Comment 4 lu hua 2014-10-21 01:27:03 UTC 
It also impacts Sandybridge.
Comment 5 lu hua 2014-10-22 05:42:38 UTC 
Fixed.
Comment 6 Jari Tahvanainen 2017-07-03 12:24:27 UTC 
Closing old verified+fixed.

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.