i915 drm GPU hung errors with kernel 2.6.38.2 on 82865G chipset

Today I have experienced two "GPU hungs" on machine with 82865G chipset working with 2.6.38.2 kernel In the /var/log/syslog I have found the following errors:

Apr 15 22:50:51 wzab kernel: [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Apr 15 22:50:51 wzab kernel: [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -11 (awaiting 1352947 at 1352942, next 1352948)
Apr 15 22:50:51 wzab kernel: [drm:init_ring_common] *ERROR* failed to set render ring head to zero ctl 00000000 head 3280a54c tail 00000000 start 00000000
Apr 15 22:50:51 wzab kernel: [drm:init_ring_common] *ERROR* render ring initialization failed ctl 0001f003 head 3280a54c tail 00000000 start 00000000
Apr 15 22:50:52 wzab kernel: [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Apr 15 22:50:52 wzab kernel: [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -11 (awaiting 1353049 at 1352942, next 1353050)

After reboot:

Apr 15 22:56:41 wzab kernel: [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Apr 15 22:56:41 wzab kernel: [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -11 (awaiting 29134 at 29129, next 29135)

Regards,
Wojtek

(CC'ing Chris Wilson)

Sitsofe | http://sucs.org/~sits/

Hi Wojtek,

Your best chance to see the bugs fixed is to report a bug following guidelines at http://intellinuxgraphics.org/how_to_report_bug.html

For GPU hangs, collecting the error state (while GPU is hung) under debugfs is very helpful.
Report contents of (as well as details on your software stack and if know what kind of action triggered the hung):
/sys/kernel/debug/dri/0/i915_error_state

Also you should CC [email protected] and Chris Wilson <[email protected]> for all issues regarding intel GPUs as thats the place it has better chance of being seen by someone who can help out. (intel-gfx is subscribers-only but at least part of the thread would reach it)

Regards,
Bruno

As Bruno said there is a /sys/kernel/debug/dri/0/i915_error_state file that contains a GPU dump at the time of the error which often contains the vital clue at to what went wrong. If you can also think back to what was happening on the machine at the time of the hang, that can also help identify the trigger and the suspect code.

Thanks,
-Chris

Chris Wilson, Intel Open Source Technology Centre

Ooops, it seems, that my previous message with uncompressed logs was a little too big.
I'm resending it with compressed logs, even though they will be probably discarded in the archive as "unhandled content"

WZab

W dniu 16.04.2011 08:07, Chris Wilson pisze:

Hi,

After I've switched on debugging (booted with drm.debug=0x06 and mounted "sudo mount -t debugfs debugfs /sys/kernel/debug" the probability of the error decreased. However after 2 hours or work it happened again.
There was nothing specific performed on the machine.
OK. This time I had the iceweasel window in the background, and was running an application under wine emulator, but previously
the same problem occured when I had only two gnome-terminals opened and iceweasel.
The problem may be associated with switching of active window in X or with switching of active tab in iceweasel (i.e. with activity related to massive change of displayed contents). I attach the output of the i915_error_state.txt, the log of X server.
Sorry for big uncompressed files, but when I compressed them previously, I saw that they were ignored by archive website as "unhandled content".
Below follows information requested on http://intellinuxgraphics.org/how_to_report_bug.html

output of "uname -m": i686

output of "uname -a":
Linux wzab 2.6.38.2 #1 SMP PREEMPT Fri Apr 8 18:37:23 CEST 2011 i686 GNU/Linux

info about chipset (from lspci):
00:00.0 Host bridge: Intel Corporation 82865G/PE/P DRAM Controller/Host-Hub Interface (rev 02) 00:02.0 VGA compatible controller: Intel Corporation 82865G Integrated Graphics Controller (rev 02) 00:03.0 PCI bridge: Intel Corporation 82865G/PE/P PCI to CSA Bridge (rev 02)
00:06.0 System peripheral: Intel Corporation 82865G/PE/P Processor to I/O Memory Interface (rev 02) 00:1d.0 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #1 (rev 02) 00:1d.1 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #2 (rev 02) 00:1d.2 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #3 (rev 02) 00:1d.3 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB UHCI Controller #4 (rev 02) 00:1d.7 USB Controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) USB2 EHCI Controller (rev 02) 00:1e.0 PCI bridge: Intel Corporation 82801 PCI Bridge (rev c2)
00:1f.0 ISA bridge: Intel Corporation 82801EB/ER (ICH5/ICH5R) LPC Interface Bridge (rev 02) 00:1f.1 IDE interface: Intel Corporation 82801EB/ER (ICH5/ICH5R) IDE Controller (rev 02) 00:1f.2 IDE interface: Intel Corporation 82801EB (ICH5) SATA Controller (rev 02)
00:1f.3 SMBus: Intel Corporation 82801EB/ER (ICH5/ICH5R) SMBus Controller (rev 02)
00:1f.5 Multimedia audio controller: Intel Corporation 82801EB/ER (ICH5/ICH5R) AC'97 Audio Controller (rev 02) 01:01.0 Ethernet controller: Intel Corporation 82547EI Gigabit Ethernet Controller

Version of libdrm2:
Package: libdrm2
Priority: optional
Section: libs
Installed-Size: 500
Maintainer: Debian X Strike Force <[email protected]>
Architecture: i386
Source: libdrm
Version: 2.4.23-3
Depends: libc6 (>= 2.7)
Filename: pool/main/libd/libdrm/libdrm2_2.4.23-3_i386.deb
Size: 421754
MD5sum: e175512785e1db00a09a4ed2063acbeb
SHA1: 1f010300dd200d4a70337f190ecc6848e653bece
SHA256: 2735ec5fbbcad7c48c34308702aa8867c8aa0ae26c5f1be52dadb431f1355c08

I was not able to send the glxinfo, as after login via ssh to the hung machine and running "DISPLAY=:0 glxinfo" the command hung and didn't display anything.

I have also attached the output of "intel_gpu_dump" command.

HTH & Regards,
Wojtek Zabolotny