[i945gm] GPU lockup (PGTBL_ER: 0x00000102)

[Yura Tolstik]

Binary package hint: xserver-xorg-video-intel

xserver crashed

Workaround: Append i915.powersave=0 to the kernel command line

ProblemType: Crash
Architecture: i386
Chipset: i945gm
Date: Thu Mar 4 20:43:57 2010
DistroRelease: Ubuntu 10.04
DkmsStatus: Error: [Errno 2] No such file or directory
DumpSignature: 585b0b66f331dfb836dfc05f1eb2e8a8
ExecutablePath: /usr/share/apport/apport-gpu-error-intel.py
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release i386 (20091028.5)
InterpreterPath: /usr/bin/python2.6
MachineType: Apple Inc. MacBook2,1
Package: xserver-xorg-video-intel 2:2.9.1-1ubuntu10
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-15-generic root=UUID=0955063e-ecb8-4baa-926a-cad14d3ef918 ro crashkernel=384M-2G:64M,2G-:128M quiet splash
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gpu-error-intel.py
ProcEnviron:

ProcVersionSignature: Ubuntu 2.6.32-15.22-generic
SourcePackage: nvidia-graphics-drivers
Title: [i945gm] GPU lockup 585b0b66f331dfb836dfc05f1eb2e8a8
Uname: Linux 2.6.32-15-generic i686
UserGroups:

dmi.bios.date: 06/27/07
dmi.bios.vendor: Apple Inc.
dmi.bios.version: MB21.88Z.00A5.B07.0706270922
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: Mac-F4208CAA
dmi.board.vendor: Apple Inc.
dmi.board.version: PVT
dmi.chassis.asset.tag: Asset Tag
dmi.chassis.type: 10
dmi.chassis.vendor: Apple Inc.
dmi.chassis.version: Mac-F4208CAA
dmi.modalias: dmi:bvnAppleInc.:bvrMB21.88Z.00A5.B07.0706270922:bd06/27/07:svnAppleInc.:pnMacBook2,1:pvr1.0:rvnAppleInc.:rnMac-F4208CAA:rvrPVT:cvnAppleInc.:ct10:cvrMac-F4208CAA:
dmi.product.name: MacBook2,1
dmi.product.version: 1.0
dmi.sys.vendor: Apple Inc.
system:
 distro: Ubuntu
 codename: lucidarchitecture: i686kernel: 2.6.32-15-generic

[lspci]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
     Subsystem: Intel Corporation Device [8086:7270]

[Geir Ove Myhr (gomyhr)]

Thank you for sending the first ever apport-generated GPU hung bug report. Even though a lot of potentially useful information has been added automatically, we need some more information from you. Could you please update the bug description to describe
- what happened (frozen screen, black screen, nothing, Xorg restart, etc. ?)
- what did you do before it happened (which programs, any suspend/resume, etc.)
- how long had the system been up when it happened (seconds, minutes, hours, days?)
- was it a one-time incident or does it happen regularly? Can you make it happen?

Since this is the first time apport has made an attempt at automatically collecting the information we need, we may also have to do some manual troubleshooting to check that the automatic procedures are doing the right thing. Look at https://wiki.ubuntu.com/X/Troubleshooting/Freeze#How%20to%20Get%20a%20Batchbuffer%20Dump%20%28-intel%20only%29 for how to use a newer intel-drm-next kernel and capture i915_error_state and some other files after the GPU hang.

[Yura Tolstik (yltsrc)]

it looks like nothing happened.
it reproduced every time, when ubuntu boot, when gdm starts and i am login, i have an error report.

can i help you with generate more informational bug report(maybe need run any debug tool)?

[Geir Ove Myhr (gomyhr)]

Thank you for helping us troubleshoot this issue. Could you do the following:

- Install the latest drm-intel-next mainline kernel from [1] (instructions for how to do it on [2]).
- Before rebooting to the new kernel, remove anything in /var/crash, just so we don't get confused: `sudo rm /var/crash/*`.
- 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`.
- Log in and check if there is anything in /var/crash: `ls -l /var/crash`
- 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 ~`
- Upload the files dmesg.txt, Xorg.0.log and i915_error_state from your home directory to here.

If there is an error report when you boot with the drm-intel-next kernel, file the bug report and mention the bug number here so that we can do some comparisons.

Since the automatic GPU lockup reporting is very new, we don't have a very good feeling form them yet, so feel free to report any information that you think may be useful.

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

[Yura Tolstik (yltsrc)]

display freeze after login with linux 2.6.33-997

[Yura Tolstik (yltsrc)]

apport generates new bug report
https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/532295

[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.

[Geir Ove Myhr (gomyhr)]

Yura, you can ignore the automatic notification from Bryce. you already tested drm-intel-next which has all the bug fixes he mentions.

Thanks a lot for testing. It seems that the automatic and manual logs are fairly consistent, but not 100%. dmesg output is consistent. I think the important part is
[ 21.609431] render error detected, EIR: 0x00000010
[ 21.609496] page table error
[ 21.609548] PGTBL_ER: 0x00000102
[ 21.609605] [drm:i915_handle_error] *ERROR* EIR stuck: 0x00000010, masking
so there are two different error registers that are being triggered, EIR and PGTBL_ER.

The only thing I find a bit confusing is that in IntelGpuDump.txt in bug 532295 it says
PGTBL_ER: 0x00000112
whereas i915_error_state, dmesg, and the other IntelGpuDump.txt says that this has value 0x00000102.

If you find this interesting, the registers are described in http://intellinuxgraphics.org/VOL_1_graphics_core.pdf .

You say that "display freeze after login with linux 2.6.33-997". What happens then? Do you have to use the power button? Does Ctrl-Alt-F1 work? Does the mouse cursor move? Does the apport-generated bug report appear the next time you boot after the freeze? I'm just trying to get a picture of how these bugs manifest themselves with the new kernels.

I can't draw any conclusions yet, but your help definitely helps us understand what is going on.

[Yura Tolstik (yltsrc)]

when display freeze i can't do anything
ctrl+alt+f1 doesn't work
and i can't see cursor
yes, apport generates new bug report every time i boot (maybe that is why you see different errors in logs, in this bug logs from second boot)

[Geir Ove Myhr (gomyhr)]

Originally reported by Yura Tolstik at:
  https://bugs.edge.launchpad.net/bugs/532100

[Problem]
GPU page table error PGTBL_ER: 0x00000102 (but 0x00000112 was also observed once).

[Original report]

Binary package hint: xserver-xorg-video-intel

when display freeze i can't do anything
ctrl+alt+f1 doesn't work
and i can't see cursor

ProblemType: Crash
Architecture: i386
Chipset: i945gm
Date: Thu Mar 4 20:43:57 2010
DistroRelease: Ubuntu 10.04
DkmsStatus: Error: [Errno 2] No such file or directory
DumpSignature: 585b0b66f331dfb836dfc05f1eb2e8a8
ExecutablePath: /usr/share/apport/apport-gpu-error-intel.py
InstallationMedia: Ubuntu 9.10 "Karmic Koala" - Release i386 (20091028.5)
InterpreterPath: /usr/bin/python2.6
MachineType: Apple Inc. MacBook2,1
Package: xserver-xorg-video-intel 2:2.9.1-1ubuntu10
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-15-generic root=UUID=0955063e-ecb8-4baa-926a-cad14d3ef918 ro crashkernel=384M-2G:64M,2G-:128M quiet splash
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gpu-error-intel.py
ProcEnviron:

ProcVersionSignature: Ubuntu 2.6.32-15.22-generic
SourcePackage: nvidia-graphics-drivers
Title: [i945gm] GPU lockup 585b0b66f331dfb836dfc05f1eb2e8a8
Uname: Linux 2.6.32-15-generic i686
UserGroups:

dmi.bios.date: 06/27/07
dmi.bios.vendor: Apple Inc.
dmi.bios.version: MB21.88Z.00A5.B07.0706270922
dmi.board.asset.tag: Base Board Asset Tag
dmi.board.name: Mac-F4208CAA
dmi.board.vendor: Apple Inc.
dmi.board.version: PVT
dmi.chassis.asset.tag: Asset Tag
dmi.chassis.type: 10
dmi.chassis.vendor: Apple Inc.
dmi.chassis.version: Mac-F4208CAA
dmi.modalias: dmi:bvnAppleInc.:bvrMB21.88Z.00A5.B07.0706270922:bd06/27/07:svnAppleInc.:pnMacBook2,1:pvr1.0:rvnAppleInc.:rnMac-F4208CAA:rvrPVT:cvnAppleInc.:ct10:cvrMac-F4208CAA:
dmi.product.name: MacBook2,1
dmi.product.version: 1.0
dmi.sys.vendor: Apple Inc.
system:
 distro: Ubuntu
 codename: lucidarchitecture: i686kernel: 2.6.32-15-generic

[lspci]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
     Subsystem: Intel Corporation Device [8086:7270]

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=34224)
Xorg.0.log

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=34225)
dmesg output

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=34226)
i915_error_state

Yes, it's really that short - from drm-intel-next kernel as of 2009-02-27.

[Geir Ove Myhr (gomyhr)]

Created an attachment (id=34227)
Another i915_error_state, this time with empty instead of missing ringbuffer

[Geir Ove Myhr (gomyhr)]

Just remembered that Chris Wilson already commented on the downstream bug report:

This is odd as I can't see the up-to-date output of i915_error_state in
this bug report. Anyway that bug in particular is likely to the fb
misalignment issue.

Assigning to him.

[Geir Ove Myhr (gomyhr)]

I have forwarded this bug to the upstream developers at https://bugs.freedesktop.org/show_bug.cgi?id=27188 . Please register at bugs.freedesktop.org and add yourself to the CC field so that you may receive any update and answer questions from the developers.

[Yura Tolstik (yltsrc)]

I can't boot on 2.6.32-16, 2.6.32-17 and 2.6.32-18 because I have frozen black screen after gdm login.
Unfortunately apport didn't generate any reports about it.
How can I help to fix this bug?

And I want know is this bug will be fixed until 10.04 release?

[Uwe Helm (1forthedoctor)]

appending i915.powersave=0 to the kernel command line solved all of my problems

[Yura Tolstik (yltsrc)]

Uwe Helm thanks for a temporary fix :)

[Geir Ove Myhr (gomyhr)]

From downstream:

appending i915.powersave=0 to the kernel command line stops these freezes.

Yura, please keep an eye on this bug report and add any additional information (like this) that may be useful here.

[Robert Hooker (sarvatt)]

Is i915.powersave=0 still required with the 2.6.32-19 kernel?

[Stefano Rivera (stefanor)]

> Is i915.powersave=0 still required with the 2.6.32-19 kernel?

No, it does not appear to be necessary any more

[Stefano Rivera (stefanor)]

Am I seeing a related issue in #560376 ?

[Geir Ove Myhr (gomyhr)]

The downstream bug has been marked as fixed, due the inclusion of this patch, which I haven't seen applied upstream yet:

  [ Jesse Barnes ]
  * SAUCE: drm/i915: don't change DRM configuration when releasing load
    detect pipe

http://kernel.ubuntu.com/git?p=ubuntu/ubuntu-lucid.git;a=commit;h=0d2907f4bead56cff60f91068b3a3efa7149e702

However, there is a very similar bug at bug #27614, which isn't fixed by that patch or i915.powersave=0.

[Chris Wilson (ickle)]

I believe this should be fixed with

commit ac0c6b5ad3b3b513e1057806d4b7627fcc0ecc27
Author: Chris Wilson <email address hidden>
Date: Thu May 27 13:18:18 2010 +0100

    drm/i915: Rebind bo if currently bound with incorrect alignment.

    Whilst pinning the buffer, check that that its current alignment
    matches the requested alignment. If it does not, rebind.

    This should clear up any final render errors whilst resuming,
    for reference:

      Bug 27070 - [i915] Page table errors with empty ringbuffer
      https://bugs.freedesktop.org/show_bug.cgi?id=27070

      Bug 15502 - render error detected, EIR: 0x00000010
      https://bugzilla.kernel.org/show_bug.cgi?id=15502

      Bug 13844 - i915 error: "render error detected"
      https://bugzilla.kernel.org/show_bug.cgi?id=13844

    Signed-off-by: Chris Wilson <email address hidden>
    Cc: <email address hidden>
    Signed-off-by: Eric Anholt <email address hidden>

in 2.6.35-rc2.