[gm45] X crashes and restarts with *ERROR* Failed to pin buffer 4 of 17

[Tommy Brander]

Binary package hint: xserver-xorg-video-intel

A couple of times a day X crashes and restarts on my up-to-date Ubuntu 10.04 Lucid installation. In kern.log and dmesg, I get the following rows after each crash;

[18454.354930] [drm:i915_gem_execbuffer] *ERROR* Failed to pin buffer 4 of 17, total 128389120 bytes: -28
[18454.354935] [drm:i915_gem_execbuffer] *ERROR* 12028 objects [8 pinned], 446992384 object bytes [110735360 pinned], 110735360/201326592 gtt bytes

Fatal server error:
Failed to submit batchbuffer: No space left on device

There is no appearant cause to the crash, sometimes it's when i start an app, other times it's while browsing the net. The last time I did nothing.

It didn't occur when I first installed Lucid (3 of january), but did instead appear around late-january.

If you need any more information, I have no problem extracting them since I'm getting these crashes several times a day :).

ProblemType: Bug
Architecture: amd64
Date: Fri Feb 26 14:18:11 2010
DistroRelease: Ubuntu 10.04
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Alpha amd64 (20091228)
Package: xserver-xorg-video-intel 2:2.9.1-1ubuntu6
ProcEnviron:
 LANGUAGE=sv_SE:sv:en_GB:en
 LANG=sv_SE.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.32-14.20-generic
SourcePackage: xserver-xorg-video-intel
Uname: Linux 2.6.32-14-generic x86_64

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 4 Series Chipset Memory Controller Hub [8086:2a40] (rev 07)
     Subsystem: Hewlett-Packard Company Device [103c:30dd]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 4 Series Chipset Integrated Graphics Controller [8086:2a42] (rev 07)
     Subsystem: Hewlett-Packard Company Device [103c:30dd]

[Bryce Harrington (bryce)]

Hi demom,

Thanks for including the attached files. Could you also include your /var/log/Xorg.0.log (or Xorg.0.log.old) from after reproducing the issue?

Please attach the output of `lspci -vvnn` too.

[This is an automated message. Apologies if it has reached you inappropriately; please just reply to this message indicating so.]

[Bryce Harrington (bryce)]

[This is an automatic notification.]

Thanks for reporting this bug to help making the Intel graphics driver
better. We hear from upstream that a number of bugs (possibly including
this one) have been fixed in the newer DRM code from the 2.6.33 kernel.
I don't know if your bug is one of the ones fixed in this release,
though, but we've prepared a PPA with this DRM update. Would you mind
installing this, rebooting, and testing if the original issue can be
reproduced with it or not?

The DRM PPA is here:

    https://edge.launchpad.net/~apw/+archive/red

Note there could be new bugs... please file these as new reports using
the command 'ubuntu-bug linux' (for kernel or DRM or KMS bugs) or
'ubuntu-bug xorg' if you suspect them to be X.org issues.

[Tommy Brander (demom)]

I installed the PPA and didn't had a single crash - until I tested upgrading to the new Lucid kernel: Linux Pisces 2.6.32-16-generic #24-Ubuntu SMP Sat Mar 6 12:32:35 UTC 2010 x86_64 GNU/Linux.

It took X 35 minutes to crash :).

So I think we safely can assume that my bug is fixed in the newer kernel. But what do I do know? Will there be an official backport to Lucid, or is I forced to compile my own kernel?

[Tommy Brander (demom)]

After re-installing the kernel from the PPA i still get X crashes. Two today.

Should I try installing a 2.6.33 kernel instead?

[Chris Halse Rogers (raof)]

The 2.6.32-16 kernel *is* the kernel from the PPA; if that crashed, then it seems that you were just lucky while testing the PPA packages.

To debug further, it would be useful if you could do the following:

- Install the latest drm-intel-next mainline kernel from [1] (instructions for how to do it on [2]).
- Reboot the computer. If the boot menu shows, choose the 2.6.33-997 kernel. If not, you probably booted it anyway, you can check with `uname -a`.
- Wait until X crashes
- Copy GPU error info: `cp /sys/kernel/debug/dri/0/i915_error_state ~`
- Check if ~/i915_error_state says something like "No error detected" or has lots of information, If no error, STOP.
- Collect some corresponding logs: `dmesg >~/dmesg.txt`, `cp /var/log/Xorg.0.log.old ~`
- Upload the files dmesg.txt, Xorg.0.log and i915_error_state from your home directory to here.

This will hopefully catch more information about the error that is occurring, which we will be able to send upstream.

[1]:http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/
[2]: https://wiki.ubuntu.com/KernelTeam/MainlineBuilds

[Robert Hooker (sarvatt)]

Hello, can you please activate this PPA and try the version of xserver-xorg-video-intel that is in it to see if the problems still exist?

https://edge.launchpad.net/~sarvatt/+archive/bugs

Note that the transition from plymouth to X will not be seamless anymore and that is an intentional change.

[Twbecker (twbecker)]

I'm having a similar issue to this where X crashes on login maybe 1 out of 3 times. It seems to be stable if I can get through the login process. Attached is my Xorg log from the crash. Also, I found the following in dmesg output which looks relevant:

[ 85.779437] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 16, total 62062592 bytes: -28
[ 85.779441] [drm:i915_gem_do_execbuffer] *ERROR* 625 objects [7 pinned], 157769728 object bytes [24006656 pinned], 24006656/134217728 gtt bytes

[Dana Olson (adolson)]

My X has crashed and brought me back to the GDM login screen several times in the past 24 hours. This is a brand new PC from Acer (work bought it for me, shipped to my door), with a fresh install using the Ubuntu 10.04.1 LTS ISO that just released. I did update it since the install, but it was only one file, IIRC.

[90236.955921] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 67125248 bytes: -28
[90236.955925] [drm:i915_gem_do_execbuffer] *ERROR* 1691 objects [6 pinned], 280305664 object bytes [47337472 pinned], 47337472/134217728 gtt bytes

I had read that turning off Compiz may help, so I disabled the desktop effects, but this has only been within the past 10 minutes so it's too early to tell if it helps.

[aefaradien (sen-aefaradien-deactivatedaccount)]

I have hit this issue several times over the past week. This is the error in dmesg:

[621913.485057] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 3 of 16, total 163536896 bytes: -28
[621913.485060] [drm:i915_gem_do_execbuffer] *ERROR* 2525 objects [8 pinned], 284733440 object bytes [30347264 pinned], 30347264/234881024 gtt bytes

Lucid 2.6.32-24-generic running on Intel Core i5 661 (which has internal graphics).

I am not using compiz.

[aefaradien (sen-aefaradien-deactivatedaccount)]

Hi again,
This bug is beginning to get really annoying... has happened twice in the last 24 hours. I have started tracking the error messages in dmesg each time it occurs (see bellow).

Done anyone know how I can get this fixed? Is it a kernel bug? Is it a driver bug? Is it a GDM bug? (some googleing around hinted at this). I am desperate for any work-around going!

Thanks in advance for any help you can give.

[111926.196032] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 165175296 bytes: -28
[111926.196041] [drm:i915_gem_do_execbuffer] *ERROR* 2181 objects [7 pinned], 296845312 object bytes [108728320 pinned], 108728320/234881024 gtt bytes

[1691095.768376] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 3 of 16, total 167731200 bytes: -28
[1691095.768384] [drm:i915_gem_do_execbuffer] *ERROR* 2493 objects [8 pinned], 297164800 object bytes [34541568 pinned], 34541568/234881024 gtt bytes

[173841.569203] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 3 of 16, total 137281536 bytes: -28
[173841.569211] [drm:i915_gem_do_execbuffer] *ERROR* 1872 objects [8 pinned], 262864896 object bytes [43978752 pinned], 43978752/234881024 gtt bytes

[223644.030809] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 1 of 2, total 86188032 bytes: -28
[223644.030813] [drm:i915_gem_do_execbuffer] *ERROR* 1722 objects [6 pinned], 330604544 object bytes [42926080 pinned], 42926080/234881024 gtt bytes

[Renaud (renaud)]

Same here. Happens randomly once a day at bare minimum, and it happened twice this morning :

Oct 11 07:33:52 fractal kernel: [ 42.346051] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 16, total 75886592 bytes: -28
Oct 11 07:33:52 fractal kernel: [ 42.346055] [drm:i915_gem_do_execbuffer] *ERROR* 675 objects [8 pinned], 191365120 object bytes [25247744 pinned], 25247744/134217728 gtt bytes
Oct 11 07:37:59 fractal kernel: [ 288.877544] [drm:i915_gem_object_pin_and_relocate] *ERROR* Relocation beyond target object bounds: obj ef132240 target 1792 delta 537444400 size 4096.
Oct 11 07:37:59 fractal kernel: [ 288.877553] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 21 of 21, total 4808704 bytes: -22
Oct 11 07:37:59 fractal kernel: [ 288.877559] [drm:i915_gem_do_execbuffer] *ERROR* 1938 objects [26 pinned], 344027136 object bytes [30019584 pinned], 134217728/134217728 gtt bytes
Oct 11 07:40:45 fractal kernel: [ 453.838537] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 25 of 34, total 67633152 bytes: -28
Oct 11 07:40:45 fractal kernel: [ 453.838543] [drm:i915_gem_do_execbuffer] *ERROR* 1725 objects [31 pinned], 352296960 object bytes [59015168 pinned], 59015168/134217728 gtt bytes

That's VERY annoying, since i can't rely on my computer...

See lspci output attached.

[Sam Pearson (sagepe)]

Hi

I am also affected by this bug, it's happened twice in the last two days. I'm running 10.04 with all current updates applied.

# uname -a
Linux 4GBL07683 2.6.32-25-generic #44-Ubuntu SMP Fri Sep 17 20:05:27 UTC 2010 x86_64 GNU/Linux

Last relevant messages from syslog:

Oct 14 13:36:424GBL07683 kernel: [18290.015290] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 97714176 bytes: -28
Oct 14 13:36:42 4GBL07683 kernel: [18290.015295] [drm:i915_gem_do_execbuffer] *ERROR* 2091 objects [8 pinned], 466956288 object bytes [75014144 pinned], 75014144/201326592 gtt bytes

In case it's any help to anyone, I've attached lspci -vvnn output.

[aefaradien (sen-aefaradien-deactivatedaccount)]

i guys you will not believe this, but I think I might have managed to make this bug reproducible - an important step to getting it fixed.

So in figuring this out, I have produced the following in dmesg:

[413848.321803] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[413848.321807] [drm:i915_gem_do_execbuffer] *ERROR* 1190 objects [7 pinned], 247103488 object bytes [114286592 pinned], 114286592/234881024 gtt bytes
[413986.513053] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[413986.513062] [drm:i915_gem_do_execbuffer] *ERROR* 1035 objects [7 pinned], 248905728 object bytes [114286592 pinned], 114286592/234881024 gtt bytes
[413995.665969] hda-intel: azx_get_response timeout, switching to polling mode: last cmd=0x302f2d00
[414070.404805] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[414070.404813] [drm:i915_gem_do_execbuffer] *ERROR* 1235 objects [7 pinned], 243036160 object bytes [114286592 pinned], 114286592/234881024 gtt bytes
[414141.474950] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[414141.474959] [drm:i915_gem_do_execbuffer] *ERROR* 1097 objects [7 pinned], 247709696 object bytes [114286592 pinned], 114286592/234881024 gtt bytes
[414225.840459] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[414225.840463] [drm:i915_gem_do_execbuffer] *ERROR* 1239 objects [7 pinned], 273108992 object bytes [114286592 pinned], 114286592/234881024 gtt bytes
[414364.091314] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[414364.091318] [drm:i915_gem_do_execbuffer] *ERROR* 992 objects [7 pinned], 252297216 object bytes [114286592 pinned], 114286592/234881024 gtt bytes

And the test case is to open a specific image in Firefox (does not happen in the gnome image viewer). I must have stumbled across several images that trigger this, and the following is a good example:

http://danbooru.donmai.us/data/5dbfcaeb05aa8a46b0ca2e4ac50a4cde.png
(WARNING: opening this link may crash X if you are affected by this problem - save your work before clicking on it!)

Since I have basically just told the world how to crash my PC, I would now really like to get this defect marked as security threat and escalated via the appropriate channels.

Does anyone know how to escalate a defect?

[Renaud (renaud)]

Hi aefaradien,

The site seems down/in maintenance. Any change to get a copy in order to test it ?

[Sam Pearson (sagepe)]

Hi,

That image makes X crash for me using either Metacity or Compiz. I had switched to Metacity since my last post and had not experienced a crash, clearly this was just good luck.

dmesg output:

[ 1674.365828] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[ 1674.365833] [drm:i915_gem_do_execbuffer] *ERROR* 1181 objects [8 pinned], 317698048 object bytes [97525760 pinned], 97525760/201326592 gtt bytes
[ 1756.663140] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[ 1756.663145] [drm:i915_gem_do_execbuffer] *ERROR* 1040 objects [8 pinned], 255336448 object bytes [97525760 pinned], 97525760/201326592 gtt bytes
[ 1813.581319] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[ 1813.581326] [drm:i915_gem_do_execbuffer] *ERROR* 1071 objects [8 pinned], 245129216 object bytes [97525760 pinned], 97525760/201326592 gtt bytes
[ 1881.021524] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 142737408 bytes: -28
[ 1881.021529] [drm:i915_gem_do_execbuffer] *ERROR* 1079 objects [8 pinned], 258760704 object bytes [97525760 pinned], 97525760/201326592 gtt bytes

Xorg.0.log:
Fatal server error:
Failed to submit batchbuffer: No space left on device

[Renaud (renaud)]

Hi there,

That image makes X crash for me too:

Oct 28 07:54:52 fractal kernel: [ 1398.219342] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 1 of 2, total 71376896 bytes: -28
Oct 28 07:54:52 fractal kernel: [ 1398.219346] [drm:i915_gem_do_execbuffer] *ERROR* 1936 objects [6 pinned], 484548608 object bytes [25227264 pinned], 25227264/134217728 gtt bytes
Oct 28 07:55:09 fractal kernel: [ 1414.707426] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 16, total 75886592 bytes: -28
Oct 28 07:55:09 fractal kernel: [ 1414.707430] [drm:i915_gem_do_execbuffer] *ERROR* 1343 objects [8 pinned], 326594560 object bytes [25247744 pinned], 25247744/134217728 gtt bytes
Oct 28 07:55:24 fractal kernel: [ 1429.906538] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 16, total 75886592 bytes: -28
Oct 28 07:55:24 fractal kernel: [ 1429.906544] [drm:i915_gem_do_execbuffer] *ERROR* 1004 objects [8 pinned], 291229696 object bytes [25247744 pinned], 25247744/134217728 gtt bytes

renaud@fractal:~$ sudo cat /var/log/Xorg.* | grep space
Failed to submit batchbuffer: No space left on device

How to raise the bug priority ?

[Jörg Herzinger (joerg-herzinger)]

This bug also happens in my small company. Our IBM ThinkCentre Workstations keep crashing when accessing certain pages in firefox. Two pages we already found that trigger this crash are:
http://webmail.aon.at and
http://ec.europa.eu/sanco_pesticides/public/index.cfm
Those crashes happen in about 9 out of 10 tries and they do not happen when opening those pages in Google Chrome (installed via googles apt repo).
Error logs are pretty much the same as the others have:

Nov 2 14:09:00 donauauen20 kernel: [ 393.091751] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 67112960 bytes: -28
Nov 2 14:09:00 donauauen20 kernel: [ 393.091765] [drm:i915_gem_do_execbuffer] *ERROR* 263 objects [5 pinned], 158416896 object bytes [47333376 pinned], 47333376/100794368 gtt bytes

[Marek Bohuncak (bohuncak)]

Hi the same crash on my PC .... kernel ver. 2.6.32-25-generic-pae, Ubuntu 10.04

2 weeks ago it crashed really often then it worked fine and finaly another today.

the crash shutdowns all the running apps and swithch me to the login screen.

Nov 5 15:21:34 bohuncak-work kernel: [715458.577901] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 87359488 bytes: -28
Nov 5 15:21:34 bohuncak-work kernel: [715458.577905] [drm:i915_gem_do_execbuffer] *ERROR* 4717 objects [7 pinned], 419315712 object bytes [60510208 pinned], 60510208/134217728 gtt bytes

If you need any other attachment let me know.

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

Maybe it's a driver bug - should be reported to Intel ?

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

(maybe related to https://bugs.launchpad.net/ubuntu/+source/xorg-server/+bug/611619 ? )

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

Bug opened @freedesktop (didn't find any opened bug related to this error) : https://bugs.freedesktop.org/show_bug.cgi?id=31520

[Jörg Herzinger (joerg-herzinger)]

I can provide you with any Information or testing you want, just name what you need. I really need this to be fixed, cause otherwise thunderbird and firefox will keep crashing my 45 workstations regularly.

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

errm.... according to what Chris Wilson answered me on freedesktop, this bug was corrected about one year ago :
"""
This bug has been fixed for over a year. Please do try testing upstream
packages before filing a bug against upstream.
"""
(source: https://bugs.freedesktop.org/show_bug.cgi?id=31520#c1)

What's the status of this driver in ubuntu ? are you sync with upstream ?

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

Any news in here ?

[Renaud (renaud)]

Anybody tested the upstream version ?

[aefaradien (sen-aefaradien-deactivatedaccount)]

@Renaud Do you have a link to the upstream package to test? I do not know where to get it from / how to install it.

[Jörg Herzinger (joerg-herzinger)]

To me this is currently unacceptable. This bug has been reported even before the release of lucid and it has been confirmed now for just one month. According to freedesktop.org this bug has even been fixed over one year ago.
I don't know if you are seeing the severity of this. Every IBM ThinkCentre (and there do exist a lot of them around the globe) is using this graphics card/driver and they all crash randomly when browsing with firefox, reading emails with thunderbird or editing some files in OpenOffice. I think this bug really has a very high importance and I personally will start tesing around with packages from maverick and debian squeeze to get my 35 users a stable system.

[Cédric Jeanneret deactivated (cjeanneret-c2c-deactivated)]

another workaround (really ugly) is to deactivate images in firefox... but as you can imagine, it's not a solution.

[Renaud (renaud)]

Latest (fixed ?) drivers are located at http://intellinuxgraphics.org/2010Q3.html
But my skills are not enough to test by myself.

[Bryce Harrington (bryce)]

The xorg maintainer gave some directions for diagnosing the issue in comment #8. There's been two dozen comments since then but no one gave the info he requested...

In any case, judging from the comments this is a kernel bug (not X) and that if someone did a git bisect on the kernel they should be able to locate the patch that fixes it, and propose for inclusion in a stable kernel update, if you feel strongly about getting this fix into lucid.

[Bryce Harrington (bryce)]

Closing the natty task since this is sorted for the development version; opened a Lucid task for the backport.

[Bryce Harrington (bryce)]

Moving to linux since the kernel fix needs the backport.

[Evgeny Burzak (buzzilo)]

Affected me too after upgrading my notebook up to 4Gb RAM (maybe that have matter, because of Intel using chunk of RAM as video memory?)

Jun 23 21:14:18 bo kernel: [98940.623214] [drm:i915_gem_do_execbuffer] *ERROR* Failed to pin buffer 2 of 3, total 147300352 bytes: -28
Jun 23 21:14:18 bo kernel: [98940.623220] [drm:i915_gem_do_execbuffer] *ERROR* 1925 objects [7 pinned], 459808768 object bytes [86282240 pinned], 86282240/234881024 gtt bytes

Linux bo 2.6.32-30-generic-pae-tuxonice #59~ppa1-Ubuntu SMP Fri Mar 18 15:49:05 UTC 2011 i686 GNU/Linux

[Roland Hieber (rohieb)]

I was getting this error in dmesg too on my Lenovo ThinkPad SL510, with linux-image-2.6.32-38-generic_2.6.32-38.83. However, i915_error_state and Xorg.0.log were both free of error messages, although X crashed & restarted. When I tried to debug the X server, gdb said that the program exited with code 1.

In my case, it was the picture http://stratum0.org/images/8/8e/20120204-Pony.jpg that caused Xorg to crash when opened in Firefox, but it seems that the error is gone in linux-image-3.2.0-997-generic_3.2.0-997.201201050433 from the drm-intel-next mainline builds. When I open that link in Firefox, all I get now is:

[drm:i915_gem_mmap_gtt] *ERROR* Attempting to mmap a purgeable buffer

I tried to do a git bisect of v2.6.32.52.21 to 3.2.0, but have not yet been successful compiling the kernel from Git... :-| I will keep on trying though.

[dino99 (9d9)]

Lucid is no more a supported version now