[Bug 101694] New: [BDW] drm crashes on newer kernels

[bugzilla-daemon at freedesktop.org]

https://bugs.freedesktop.org/show_bug.cgi?id=101694

            Bug ID: 101694
           Summary: [BDW] drm crashes on newer kernels
           Product: DRI
           Version: unspecified
          Hardware: x86-64 (AMD64)
                OS: Linux (All)
            Status: NEW
          Severity: critical
          Priority: medium
         Component: DRM/Intel
          Assignee: intel-gfx-bugs at lists.freedesktop.org
          Reporter: bugger at grr.la
        QA Contact: intel-gfx-bugs at lists.freedesktop.org
                CC: intel-gfx-bugs at lists.freedesktop.org

Created attachment 132426
  --> https://bugs.freedesktop.org/attachment.cgi?id=132426&action=edit
/sys/class/drm/card0/error

On newer kernels, Xorg sometimes freezes temporarily. The bug occured on at
least kernels 4.11.8 and 4.12.0, while 4.9.33 did not, after a long while of
intense usage. This indicates that this is a kernel issue, but not necessarily.

chipset: Intel i5-5257U
system architecture: 64-bit
xf86-video-intel: not installed
xserver: 1.19.3
mesa: 17.1.14
libdrm: 2.4.81
kernel: 4.12.0
Linux distribution: Void (with musl libc)
Machine or mobo model: Apple MacBookPro12,1 (2015)
Display connector: internal monitor

Reproduction:
Open chromium and load some intensive web pages. After a while it should freeze
for a few seconds, and later return to normal after some minor graphic
artifacts. This will occur repeatedly indefinitely.

dmesg excerpt:

[  202.561869] DMAR: DRHD: handling fault status reg 3
[  202.561874] DMAR: [DMA Write] Request device [00:02.0] fault addr fd285000
[fault reason 23] Unknown
[  209.700964] [drm] GPU HANG: ecode 8:0:0x85dffffb, in chromium [1064],
reason: Hang on rcs, action: reset
[  209.700965] [drm] GPU hangs can indicate a bug anywhere in the entire gfx
stack, including userspace.
[  209.700966] [drm] Please file a _new_ bug report on bugs.freedesktop.org
against DRI -> DRM/Intel
[  209.700966] [drm] drm/i915 developers can then reassign to the right
component if it's not a kernel issue.
[  209.700966] [drm] The gpu crash dump is required to analyze gpu hangs, so
please always attach it.
[  209.700967] [drm] GPU crash dump saved to /sys/class/drm/card0/error
[  209.701011] drm/i915: Resetting chip after gpu hang
[  219.743979] drm/i915: Resetting chip after gpu hang
[  229.727960] drm/i915: Resetting chip after gpu hang
[  229.728133] DMAR: DRHD: handling fault status reg 2
[  229.728138] DMAR: [DMA Write] Request device [00:02.0] fault addr ffff9000
[fault reason 23] Unknown
[  241.759968] drm/i915: Resetting chip after gpu hang
[  249.760012] drm/i915: Resetting chip after gpu hang
[  249.866657] DMAR: DRHD: handling fault status reg 3
[  249.866664] DMAR: [DMA Write] Request device [00:02.0] fault addr fbb00000
[fault reason 23] Unknown
[  259.744063] drm/i915: Resetting chip after gpu hang
[  269.728105] drm/i915: Resetting chip after gpu hang
[  448.369227] DMAR: DRHD: handling fault status reg 3
...etc...

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.
You are on the CC list for the bug.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.freedesktop.org/archives/intel-gfx-bugs/attachments/20170704/5fad2bbc/attachment-0001.html>