Bug 107329

Summary: GPU hang with Linux-4.18-rc5
Product: DRI Reporter: udo
Component: DRM/IntelAssignee: Intel GFX Bugs mailing list <intel-gfx-bugs>
Status: CLOSED FIXED QA Contact: Intel GFX Bugs mailing list <intel-gfx-bugs>
Severity: normal    
Priority: medium CC: intel-gfx-bugs
Version: unspecified   
Hardware: x86-64 (AMD64)   
OS: Linux (All)   
Whiteboard: Triaged
i915 platform: BDW i915 features:
Attachments:
Description Flags
Full dmesg output
none
Crash dump none

Description udo 2018-07-22 10:10:02 UTC 
Created attachment 140764 [details]
Full dmesg output

With Linux-4.18-rc5 and an Intel BDW GPU, there is a GPU crash at kernel boot time as follows:

[    0.349074] [drm] VT-d active for gfx access
[    0.349078] [drm] Replacing VGA console driver
[    0.360104] [drm] Supports vblank timestamp caching Rev 2 (21.10.2013).
[    0.360108] [drm] Driver supports precise vblank timestamp query.
[    0.360132] i915 0000:00:02.0: vgaarb: changed VGA decodes: olddecodes=io+mem,decodes=io+mem:owns=io+mem
[    1.312040] tsc: Refined TSC clocksource calibration: 2593.991 MHz
[    1.312049] clocksource: tsc: mask: 0xffffffffffffffff max_cycles: 0x25640fc6114, max_idle_ns: 440795275253 ns
[    1.312069] clocksource: Switched to clocksource tsc
[    4.833977] [drm] GPU HANG: ecode 8:0:0xfffffffe, reason: hang on rcs0, bcs0, vcs0, vecs0, action: reset
[    4.833983] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
[    4.833985] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
[    4.833988] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
[    4.833991] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
[    4.833993] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[    4.834008] i915 0000:00:02.0: Resetting rcs0 for hang on rcs0, bcs0, vcs0, vecs0
[    4.834035] i915 0000:00:02.0: Resetting bcs0 for hang on rcs0, bcs0, vcs0, vecs0
[    4.834053] i915 0000:00:02.0: Resetting vcs0 for hang on rcs0, bcs0, vcs0, vecs0
[    4.834073] i915 0000:00:02.0: Resetting vecs0 for hang on rcs0, bcs0, vcs0, vecs0
[   10.848340] i915 0000:00:02.0: Resetting chip for hang on rcs0, bcs0, vcs0, vecs0
[   10.848362] i915 0000:00:02.0: GPU recovery failed
[   10.848679] [drm] Initialized i915 1.6.0 20180514 for 0000:00:02.0 on minor 0
Comment 1 udo 2018-07-22 10:12:12 UTC 
Created attachment 140765 [details]
Crash dump
Comment 2 Chris Wilson 2018-07-22 12:01:27 UTC 
intel_iommu=igfx_off
Comment 3 Radosław Szwichtenberg 2018-07-23 13:06:37 UTC 
Could you confirm if answer from Chris fixed the problem for you?
Comment 4 Lakshmi 2018-09-11 07:15:00 UTC 
Udo, ping?
Comment 5 Lakshmi 2018-11-13 13:13:23 UTC 
No feedback for more than 2 months. Closing this bug.
Reporter, apply the WA mentioned in the bug. If the issue still appears, reopen the bug with dmesg/error attached. Remember to verify the issue with latest drm-tip.
(https://cgit.freedesktop.org/drm-tip)

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.