Bug 97423 - Lenovo ThinkPad X1 Carbon GPU hang
Summary: Lenovo ThinkPad X1 Carbon GPU hang
Status: CLOSED WORKSFORME
Alias: None
Product: Mesa
Classification: Unclassified
Component: Drivers/DRI/i965 (show other bugs)
Version: git
Hardware: x86-64 (AMD64) Linux (All)
: medium major
Assignee: Ian Romanick
QA Contact: Intel 3D Bugs Mailing List
URL:
Whiteboard:
Keywords:
Depends on:
Blocks:
 
Reported: 2016-08-20 21:01 UTC by Mark Hedges
Modified: 2016-08-30 16:51 UTC (History)
1 user (show)

See Also:
i915 platform: SKL
i915 features: GPU hang


Attachments
GPU crash dump (40.08 KB, application/x-bzip)
2016-08-20 21:01 UTC, Mark Hedges
Details
dmesg output with debug flags (1000.99 KB, text/plain)
2016-08-20 21:01 UTC, Mark Hedges
Details
attachment-15075-0.html (268 bytes, text/html)
2016-08-30 15:25 UTC, Mark Hedges
Details

Description Mark Hedges 2016-08-20 21:01:00 UTC
Created attachment 125931 [details]
GPU crash dump

GPU hangs when using LibreOffice Writer.  Always happens every once in a while, while typing.  It goes out to lunch for 15 seconds and then comes back when the chip resets.  System does not respond at all during that time.  I lose the keystrokes in the meantime.  It seems to be only on this one document... maybe a weird font rendering issue?

Tried adding intel_iommu=igfx_off to boot params but that did not help, so I removed that option for this test.

`uname -m`      x86_64
`uname -r`      4.6.0-1-rt-amd64
Linux:          Debian Stretch
Machine:        Lenovo X1 Carbon 4th gen
Display:        Built-in 14" WQHD 2560x1440 IPS, 300 nits
Batch buffer:   Attached
dmesg debug:    Attached (using "drm.debug=0x1e log_buf_len=1M")


Syslog excerpt:

===========================================
Aug 20 13:51:10 walnut kernel: [  353.815999] [drm] stuck on render ring
Aug 20 13:51:10 walnut kernel: [  353.816385] [drm] GPU HANG: ecode 9:0:0x86dffffd, in Xorg [2650], reason: Ring hung, action: reset
Aug 20 13:51:10 walnut kernel: [  353.816389] [drm] GPU hangs can indicate a bug anywhere in the entire gfx stack, including userspace.
Aug 20 13:51:10 walnut kernel: [  353.816392] [drm] Please file a _new_ bug report on bugs.freedesktop.org against DRI -> DRM/Intel
Aug 20 13:51:10 walnut kernel: [  353.816394] [drm] drm/i915 developers can then reassign to the right component if it's not a kernel issue.
Aug 20 13:51:10 walnut kernel: [  353.816396] [drm] The gpu crash dump is required to analyze gpu hangs, so please always attach it.
Aug 20 13:51:10 walnut kernel: [  353.816398] [drm] GPU crash dump saved to /sys/class/drm/card0/error
Aug 20 13:51:10 walnut kernel: [  353.818723] drm/i915: Resetting chip after gpu hang
Aug 20 13:51:12 walnut kernel: [  355.816648] [drm] RC6 on
===========================================
Comment 1 Mark Hedges 2016-08-20 21:01:33 UTC
Created attachment 125932 [details]
dmesg output with debug flags
Comment 2 yann 2016-08-30 09:42:09 UTC
Assigning to Mesa product (please let me know if I am mistaken with this GPU Hang).

From this error dump, hung is happening in render ring batch with active head
at 0xf7b64504, with 0x79000002 (3DSTATE_DRAWING_RECTANGLE) as IPEHR.

Batch extract (around 0xf7b64504):

0xf7b644e8:      0x784e0002: 3D UNKNOWN: 3d_965 opcode = 0x784e
0xf7b644ec:      0x00000000: MI_NOOP
0xf7b644f0:      0x00000000: MI_NOOP
0xf7b644f4:      0x00000000: MI_NOOP
0xf7b644f8:      0x780f0000: 3DSTATE_SCISSOR_POINTERS
0xf7b644fc:      0x000073a0:    scissor rect offset
0xf7b64500:      0x79000002: 3DSTATE_DRAWING_RECTANGLE
0xf7b64504:      0x00000000:    top left: 0,0
0xf7b64508:      0x04730570:    bottom right: 1392,1139
0xf7b6450c:      0x00000000:    origin: 0,0
0xf7b64510:      0x784a0000: 3D UNKNOWN: 3d_965 opcode = 0x784a
0xf7b64514:      0x0000c001: MI_NOOP
Comment 3 Mark Hedges 2016-08-30 15:25:53 UTC
Created attachment 126120 [details]
attachment-15075-0.html

May be right.  At the time, I did not have libreoffice-gnome or
libreoffice-gtk installed, and the whole interface was weird, probably
plain mesa.  After I installed the gtk interface, I have not been able to
reproduce the problem.  Thanks.  -Mark
Comment 4 yann 2016-08-30 16:51:17 UTC
Closing bug, please feel free to reopen if it is occurring again.


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.