Ubuntu
xserver-xorg-video-intel package

[i845] GPU lockup da27380f (PGTBL_ER: 0x00000049)

Bug #545201 reported by useResa on 2010-03-23

This bug report is a duplicate of: Bug #541492: MASTER: [i845] GPU lockup. Edit Remove

This bug affects 1 person

Affects		Status	Importance	Assigned to	Milestone
	xserver-xorg-video-intel (Ubuntu)	New	Undecided	Unassigned

Bug Description

Binary package hint: xserver-xorg-video-intel

This bug is most likely a duplicate of bug report #541492

The reason for submitting this report separately is since my initial bug report #529191 was closed due to the fact that updates were provided. This is to confirm that the issue is still present in my situation.

Feel free to contact me if more information is required

ProblemType: Crash
Architecture: i386
Chipset: i845
Date: Tue Mar 23 17:09:49 2010
DistroRelease: Ubuntu 10.04
DkmsStatus: Error: [Errno 2] No such file or directory
DumpSignature: da27380f (PGTBL_ER: 0x00000049)
ExecutablePath: /usr/share/apport/apport-gpu-error-intel.py
InstallationMedia: Ubuntu 10.04 "Lucid Lynx" - Alpha i386 (20100224.1)
InterpreterPath: /usr/bin/python2.6
Lsusb:
Bus 004 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 003 Device 002: ID 04f3:0210 Elan Microelectronics Corp. AM-400 Hama Optical Mouse
Bus 003 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 002 Device 001: ID 1d6b:0001 Linux Foundation 1.1 root hub
Bus 001 Device 001: ID 1d6b:0002 Linux Foundation 2.0 root hub
MachineType: Dell Computer Corporation Dimension 2400
Package: xserver-xorg-video-intel 2:2.9.1-1ubuntu16
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-16-generic root=/dev/sda5
ProcCmdline: /usr/bin/python /usr/share/apport/apport-gpu-error-intel.py
ProcEnviron:

ProcVersionSignature: Ubuntu 2.6.32-16.25-generic
SourcePackage: xserver-xorg-video-intel
Title: [i845] GPU lockup da27380f (PGTBL_ER: 0x00000049)
Uname: Linux 2.6.32-16-generic i686
UserGroups:

XorgLogOld:

dmi.bios.date: 12/02/2003
dmi.bios.vendor: Dell Computer Corporation
dmi.bios.version: A05
dmi.board.name: 0G1548
dmi.board.vendor: Dell Computer Corp.
dmi.board.version: A00
dmi.chassis.type: 15
dmi.chassis.vendor: Dell Computer Corporation
dmi.modalias: dmi:bvnDellComputerCorporation:bvrA05:bd12/02/2003:svnDellComputerCorporation:pnDimension2400:pvr:rvnDellComputerCorp.:rn0G1548:rvrA00:cvnDellComputerCorporation:ct15:cvr:
dmi.product.name: Dimension 2400
dmi.sys.vendor: Dell Computer Corporation
glxinfo: Error: [Errno 2] No such file or directory
system:
distro: Ubuntu
codename: lucid
architecture: i686
kernel: 2.6.32-16-generic

[lspci]
00:02.0 VGA compatible controller [0300]: Intel Corporation 82845G/GL[Brookdale-G]/GE Chipset Integrated Graphics Device [8086:2562] (rev 01)
Subsystem: Dell Device [1028:0160]

See original description

Tags:

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-23:

BootDmesg.txt Edit (31.0 KiB, text/plain; charset="utf-8")
CurrentDmesg.txt Edit (541 bytes, text/plain; charset="utf-8")
Dependencies.txt Edit (3.7 KiB, text/plain; charset="utf-8")
IntelGpuDump.txt Edit (1.2 MiB, text/plain; charset="utf-8")
Lspci.txt Edit (7.9 KiB, text/plain; charset="utf-8")
PciDisplay.txt Edit (783 bytes, text/plain; charset="utf-8")
ProcAttrCurrent.gz Edit (49 bytes, application/x-gzip)
ProcCpuinfo.txt Edit (549 bytes, text/plain; charset="utf-8")
ProcInterrupts.txt Edit (1.2 KiB, text/plain; charset="utf-8")
ProcMaps.txt Edit (5.6 KiB, text/plain; charset="utf-8")
ProcModules.txt Edit (2.0 KiB, text/plain; charset="utf-8")
ProcStatus.txt Edit (726 bytes, text/plain; charset="utf-8")
RelatedPackageVersions.txt Edit (232 bytes, text/plain; charset="utf-8")
UdevDb.txt Edit (87.1 KiB, text/plain; charset="utf-8")
UdevLog.txt Edit (184.5 KiB, text/plain; charset="utf-8")
XorgLog.txt Edit (45.9 KiB, text/plain; charset="utf-8")
Xrandr.txt Edit (8.3 KiB, text/plain; charset="utf-8")
i915_error_state.txt Edit (171 bytes, text/plain; charset="utf-8")
setxkbmap.txt Edit (233 bytes, text/plain; charset="utf-8")
xdpyinfo.txt Edit (19.1 KiB, text/plain; charset="utf-8")
xkbcomp.txt Edit (54.0 KiB, text/plain; charset="utf-8")

Bryce Harrington (bryce) on 2010-03-24

description:

updated

Revision history for this message

Geir Ove Myhr (gomyhr) wrote on 2010-03-24:

You're right, this one is a clear duplicate of bug 541492 (on i845 chipset and page table error 0x49 like many others). With the initial bug report is wasn't possible to see, but now with the right logs attached and the most important information automatically in the title, it's easy to see.

I'll mark this as a duplicate and will post any updates to this issue at bug 541492.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-25:

Have installed (as suggested in bug #541492) xserver-xorg-video intel package version: 2:2.10.903-4-g0c47195.1
Will provide feedback as soon as possible on results

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-26:

The trial version (2:2.10.903-4-g0c47195.1) gave me more issues than the current official version (2:2.9.1-1ubuntu17).
Although the current official version crashes every once in a while, with the trial version I have experienced KernelOops errors and furthermore I was unable to use the (commercial) NX client (version: 3.3.0-6).

With the NX client I got logged out each time it tried to establish the display connection.
The KernelOops errors I have not reported separately since they only appeared after installation of the trial version and have not been appearing since I downgraded again to the current official version.

Revision history for this message

Geir Ove Myhr (gomyhr) wrote on 2010-03-26:

Thank you for testing!

Revision history for this message

Geir Ove Myhr (gomyhr) wrote on 2010-03-26:

useResa, the other problems you saw are probably because there have been a lot of other changes between the Lucid driver and the one in Brian Roger's PPA. I have uploaded a driver to my own standard PPA which is only the standard Lucid driver with the backported patch (well, I hope this is the right way to backport it... looks pretty straightforward...) https://launchpad.net/~gomyhr/+archive/standard

Could you test this? If it doesn't crash because I made a mistake in backporting, I will update the master bug description to point to my PPA.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-26:

Hi Geir,
I have installed the following package (partial output of aptitude show xserver-xorg-video-intel):
Package: xserver-xorg-video-intel
State: installed
Automatically installed: no
Version: 2:2.9.1-1ubuntu18~gomyhr2~clipsolids

I will test this version in the course of this weekend. Hope that is okay.
If this is the incorrect version, please let me know.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-26:

What I can already can confirm is that the NX client works with the package from your repository.
The display connection works and I do not get logged out.

Revision history for this message

Geir Ove Myhr (gomyhr) wrote on 2010-03-27:

Good, that probably means that I did't make a mistake in backporting or building. Since this version of the driver is identical to the Lucid one except that it corrects some instructions that will hang the GPU if the show up, everything that works with the Lucid driver should also work with this driver.

I will update the master bug description to point to my PPA.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-27:

#10

Today my system is currently up and running for more than seven and a half hours (not always manned).
But have not encountered one single crash! .... it is looking very hopeful!
Thanks Geir for the effort put in this.

One other thing I noticed (don't know if this has to do with the fix), but I can now use the numeric keypad on my keyboard to enter numbers when using the NX client. I was not able to do this before in Lucid.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-27:

#11

Sorry Geir ... almost wish I did not post the previous comment.
Shortly after my post the system crashed twice :'(

I will try and run an apport-collect to see if I can provide information to help you determine the cause.

Revision history for this message

useResa (rdrijsen) wrote on 2010-03-27:

#12

Sorry ... apport collect did not work
"You are not the reporter or subscriber of this problem report, or the report is a duplicate or already closed.
Please create a new report using "apport-bug"."

Revision history for this message

useResa (rdrijsen) wrote on 2010-04-02:

#13

In the Master bug #541492 I read the following:
However, I think it's pretty well known at this point that upstream's
code is fairly buggered with i845, so if you have that video card, if
you have a freeze it's probably not worth putting in more bug reports
about it at this point. You can track the upstream bug report as it
progresses, and help them with testing things. Be aware you'll probably
be expected to patch and build new kernels in order to do this.

Beyond that, assuming a patch doesn't come to light, the one other knob
we have to fiddle with at the distro level is to blacklist KMS from this
hardware. See http://wiki.ubuntu.com/X/KernelModeSetting for more info
about this and how to configure it off (basically boot i915.modeset=0)
so it uses UMS instead.

Do I understand correctly that it is being advised to turn *off* KMS and check whether the freeze issues are over then?
In the Wiki I read it is possible to turn off KMS and thus reverting to UMS. But also the option of using framebuffering is mentioned. Which option is best to test?

Please advise

Revision history for this message

Geir Ove Myhr (gomyhr) wrote on 2010-04-04:

#14

I didn't know that limiting who could run apport-collect on a bug report had been turned on, but in general it is a good thing, since we often get "apport-collect-bombed" with information that we do not need. Hardly ever would we want someone who is not the original reporter of a bug to run apport-collect to that bug. For a duplicate bug, I'm not so sure, but it is easy to work around by unduplicating the bug temporarily, running apport-collect and then duplicating the bug. The advantage of this is that the other people subscribed to the bug master bug report don't get 20+ mails. Of course, the real problem here is that apport-collect triggers one comment per attachment.

Anyway, in this particular case, I don't think apport-collect will upload the particular dumps that are useful, so you don't have to run it.

In the long term, we want to use KMS and framebuffer compression. But since the Lucid release date is getting close we want to identify workarounds that may make the systems usable until the real bug is fixed. This is why we would like feedback on whether turning KMS off helps, and how much it helps. From the comments upstream it is clear that the bug is not is in the KMS specific code, but that some parts of the driver is much faster with KMS and that the bug therefore may be triggered much more easily with KMS that UMS. In order to decide what our best option is, we need feedback from affected users to know if there is a real improvement.

It looks like framebuffer compression is not supported anyway on 855GM, so disabling it should have no effect. It has recently been disabled on 915GM and 945GM since there were problems with it there. This kind of features can be found in i915_drv.c in the kernel driver [1]. Look for the string "has_fbc".

[1]: http://git.kernel.org/?p=linux/kernel/git/torvalds/linux-2.6.git;a=tree;f=drivers/gpu/drm/i915;h=7de084e6751446b405dda1b2a2e0406b366222cb;hb=5e11611a5d22252f3f9c169a3c9377eac0c32033