[i915gm] X crashes minutes after resuming from suspend to RAM

[roger64]

roger@roger-laptop:~$ lspci | grep VGA
00:02.0 VGA compatible controller: Intel Corporation Mobile 915GM/GMS/910GML Express Graphics Controller (rev 03)

Laptop: Toshiba A 80 Pentium M 1;73 Ghz 32 bits
Kernel Lucid or 2.6.33.04 or 2.6.34-020634-generic (same results with these three kernels)
I have the 2.9 Intel driver. I got the same result with the 2.11

Systematically, after a ten to forty minutes delay when resuming from suspend to RAM, X is crashing (black screen). Since Lucid release, about ten to fifteen seconds after the crash, a small screen pops up and offers to switch to a "degraded mode" and provides other technical choices. After that, suspend to RAM is working again (I think with VESA) but of course, I lose Compiz and these kind of things.

This bug began to appear with Karmic. At that time, I did not have the help screen above so I could not use at all suspend to RAM..

Error consistently at the time of a crash:
[kern.log.txt]
Jun 3 12:07:06 localhost kernel: [ 4760.404031] [drm:i915_hangcheck_elapsed] *ERROR* Hangcheck timer elapsed... GPU hung
Jun 3 12:07:06 localhost kernel: [ 4760.404048] render error detected, EIR: 0x00000000
Jun 3 12:07:06 localhost kernel: [ 4760.405046] [drm:i915_do_wait_request] *ERROR* i915_do_wait_request returns -5 (awaiting 312731 at 312730)
[/kern.log.txt

---
Architecture: i386
DistroRelease: Ubuntu 10.04
EcryptfsInUse: Yes
MachineType: TOSHIBA Satellite A80
Package: xserver-xorg-video-intel 2:2.9.1-4+really2.8.0-0ubuntu4
PackageArchitecture: i386
PccardctlIdent:
 Socket 0:
   no product info available
PccardctlStatus:
 Socket 0:
   no card
ProcCmdLine: BOOT_IMAGE=/boot/vmlinuz-2.6.32-22-generic root=UUID=966722d7-cdc8-4acd-83dc-ea9040b237fe ro splash vga=791 quiet splash
ProcEnviron:
 LANGUAGE=fr_FR:fr:en_GB:en
 LANG=fr_FR.UTF-8
 SHELL=/bin/bash
ProcVersionSignature: Ubuntu 2.6.32-22.35-generic 2.6.32.11+drm33.2
Tags: lucid lucid
Uname: Linux 2.6.32-22-generic i686
UnreportableReason: Ceci n'est pas un authentique paquet Ubuntu
UserGroups: adm admin audio cdrom dialout dip disk floppy fuse lpadmin plugdev sambashare vboxusers video
dmi.bios.date: 12/13/2005
dmi.bios.vendor: TOSHIBA
dmi.bios.version: V2.40
dmi.board.name: EAT10/EAT20
dmi.board.vendor: TOSHIBA
dmi.board.version: Null
dmi.chassis.asset.tag: *
dmi.chassis.type: 10
dmi.chassis.vendor: TOSHIBA
dmi.chassis.version: N/A
dmi.modalias: dmi:bvnTOSHIBA:bvrV2.40:bd12/13/2005:svnTOSHIBA:pnSatelliteA80:pvrPSA80Q-074004:rvnTOSHIBA:rnEAT10/EAT20:rvrNull:cvnTOSHIBA:ct10:cvrN/A:
dmi.product.name: Satellite A80
dmi.product.version: PSA80Q-074004
dmi.sys.vendor: TOSHIBA
system:
 distro: Ubuntu
 codename: lucid
 architecture: i686
 kernel: 2.6.32-22-generic

[Bryce Harrington (bryce)]

Hi roger64,

Please run the command 'apport-collect BUGNUMBER', which will attach several files we need for debugging.

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

[roger64 (rogqip-suse)]

Hi Bryce

I did it. I gave all authorizations. Hope it worked?

Cheers

Roger

2010/5/30 Bryce Harrington <email address hidden>

> Hi roger64,
>
>
> Please run the command 'apport-collect BUGNUMBER', which will attach
> several files we need for debugging.
>
> [This is an automated message. Apologies if it has reached you
> inappropriately; please just reply to this message indicating so.]
>
>
> ** Tags added: crash
>
> ** Tags added: needs-xorglog
>
> ** Tags added: needs-lspci-vvnn
>
> ** Changed in: xserver-xorg-video-intel (Ubuntu)
> Status: New => Incomplete
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Stenten (stenten)]

Immediately after you recover from a crash, please open a terminal and run "apport-collect 583902". Please make sure you're booted into the default Lucid versions (2.6.32-22 kernel, 2.9.1-3ubuntu5 xserver-xorg-video-intel, etc).

Can you look more closely on what exactly precedes the crash? If you only suspend/resume once, does the crash still occur? If you don't suspend/resume at all, can you reproduce the crash? What are you doing when it crahes? Any graphics-intensive applications, etc?

[roger64 (rogqip-suse)]

I'll do it and send this report this afternoon.

I start Lucid with Compiz on. I put to sleep. After resuming, and only after
resuming from suspend to RAM, X is crashing. No telltale sign before, no
special activity (using Firefox or OpenOfice or nautilus..). Crashing delay
after resume is between 10 to 40 minutes and it happens systematically. It's
the only predictable thing.

Roger

2010/6/2 Stenten <email address hidden>

> Immediately after you recover from a crash, please open a terminal and
> run "apport-collect 583902". Please make sure you're booted into the
> default Lucid versions (2.6.32-22 kernel, 2.9.1-3ubuntu5 xserver-xorg-
> video-intel, etc).
>
> Can you look more closely on what exactly precedes the crash? If you
> only suspend/resume once, does the crash still occur? If you don't
> suspend/resume at all, can you reproduce the crash? What are you doing
> when it crahes? Any graphics-intensive applications, etc?
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[roger64 (rogqip-suse)]

I sent one report apport-collect some minutes ago. I had to restart the
computer because no help screen appeared. The X freeze occured about ten
minutes after resuming from suspend to RAM. I use the standard Lucid kernel
and Intel driver.

If I can send later a "better" apport-collect report, I'll do it.

Roger

2010/6/2 Roger Jargoyhen <email address hidden>

> I'll do it and send this report this afternoon.
>
> I start Lucid with Compiz on. I put to sleep. After resuming, and only
> after resuming from suspend to RAM, X is crashing. No telltale sign before,
> no special activity (using Firefox or OpenOfice or nautilus..). Crashing
> delay after resume is between 10 to 40 minutes and it happens
> systematically. It's the only predictable thing.
>
> Roger
>
> 2010/6/2 Stenten <email address hidden>
>
> Immediately after you recover from a crash, please open a terminal and
>> run "apport-collect 583902". Please make sure you're booted into the
>> default Lucid versions (2.6.32-22 kernel, 2.9.1-3ubuntu5 xserver-xorg-
>> video-intel, etc).
>>
>> Can you look more closely on what exactly precedes the crash? If you
>> only suspend/resume once, does the crash still occur? If you don't
>> suspend/resume at all, can you reproduce the crash? What are you doing
>> when it crahes? Any graphics-intensive applications, etc?
>>
>> --
>> X crashes minutes after resuming from suspend to RAM
>> https://bugs.launchpad.net/bugs/583902
>> You received this bug notification because you are a direct subscriber
>> of the bug.
>>
>
>

[Stenten (stenten)]

Hmm. For some reason Apport isn't working properly; none of the information was uploaded. Can you describe exactly what happens next time you apport-collect after a crash?

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

apport information

[roger64 (rogqip-suse)]

Hi

This morning, some minutes ago, I think the report has been sent. I had to
fully restart the computer because there was no help screen. As soon as I
was online, I sent the report.

Roger

2010/6/2 Stenten <email address hidden>

> Hmm. For some reason Apport isn't working properly; none of the
> information was uploaded. Can you describe exactly what happens next
> time you apport-collect after a crash?
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Stenten (stenten)]

Thank you; Apport worked correctly that time.

When the screen blacks out again, can you:
1) Toggle the Capslock light?
2) Switch to a virtual terminal with Ctrl+Alt+F1?
3) Logout with Ctrl+Alt+Del?
4) REISUB? (Hold down Alt+SysRq and press, in this order, r, e, i, s, u, and b. This should safely reboot your system.)

After you recover your system, please run
cat /var/log/kern.log > kern.log.txt
cat /var/log/Xorg.0.log.old Xorg.0.log.old.txt (If you don't have to reboot your system, use 'Xorg.0.log instead of 'Xorg.0.log.old'.)
That should create two files named kern.log.txt and Xorg.0.log.old.txt (or Xorg.0.log.txt) in your home folder; upload those please.

Also see if you can recreate the problem with Compiz disabled.

[roger64 (rogqip-suse)]

Hi

A quick first reply, another one later, probably this afternoon.
After the X crash following shortly a resume
- The capslock light does not work but toggling it seems to trigger the
appearance of the small help window "degraded mode".
- Ctrl Alt F1: brings me to a terminal login
- Ctrl Alt Del: does nothing
- REISUB: is working, triggers some successive writing in terminal and
finally reboots.

In joint attachment two texts. The second one is a copy of my terminal.

I'll check this afternoon without Compiz.

Roger

2010/6/3 Stenten <email address hidden>

> Thank you; Apport worked correctly that time.
>
> When the screen blacks out again, can you:
> 1) Toggle the Capslock light?
> 2) Switch to a virtual terminal with Ctrl+Alt+F1?
> 3) Logout with Ctrl+Alt+Del?
> 4) REISUB? (Hold down Alt+SysRq and press, in this order, r, e, i, s, u,
> and b. This should safely reboot your system.)
>
> After you recover your system, please run
> cat /var/log/kern.log > kern.log.txt
> cat /var/log/Xorg.0.log.old Xorg.0.log.old.txt (If you don't have to reboot
> your system, use 'Xorg.0.log instead of 'Xorg.0.log.old'.)
> That should create two files named kern.log.txt and Xorg.0.log.old.txt (or
> Xorg.0.log.txt) in your home folder; upload those please.
>
> Also see if you can recreate the problem with Compiz disabled.
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[roger64 (rogqip-suse)]

Hi

To reply to one of your questions:

I changed from compiz to metacity. Put to sleep. Resumed. I had a X crash.
Other thing:
I started the computer with metacity. Put to sleep. Resumed. I had a X crash
after about one hour.
So, it would seem that Compiz is not responsible. I remember using Beryl and
then Compiz on this same graphic card without any suspend to RAM problems.

Also, after a X crash, when I am offered to restart in degraded mode, I have
no more crashes.

Do you need any other technical information?

Roger

2010/6/3 Bryce Harrington <email address hidden>

> ** Tags removed: needs-xorglog
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Stenten (stenten)]

Please try the drm-intel-next mainline kernel here [1]. Download and install the _all headers, i386 headers, and i386 image.

Then boot into that kernel with 'drm.debug=0x02' in the grub kernel line after 'quiet splash'.

Then see if you can reproduce a crash where you can switch to a TTY with Ctrl+Alt+F1. Login and then take a copy of these three files:
cat /sys/kernel/debug/dri/0/i915_error_state > i915_error_state.txt
cat /var/log/Xorg.0.log > Xorg.0.log.txt
dmesg > dmesg.txt
Then try restarting X by running 'startx'.

If you can't switch to a TTY no matter how many times you reproduce, upload your kern.log file after you recover, noting the exact date/time the crash occurred.

[1]: http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-next/current/

[Stenten (stenten)]

Could you also please attach the output of 'lspci -vvnn' as a .txt?

[roger64 (rogqip-suse)]

Hi

Here it is.

Roger

2010/6/4 Stenten <email address hidden>

> lspci -vvnn

[roger64 (rogqip-suse)]

Hi

I tried this proposed kernel but it does not work properly on my computer
for now.
roger@roger-laptop:~$ uname -r
2.6.34-997-generic

Once started with the option drm.debug=0x02 after quiet splash, I have 68
instances of udevd-daemon running with a plane noise. Wether I kill udevd or
not, suspend to RAM is not working (It suspends and two seconds later, it's
back).

So, I stopped it. Sorry for that. I have another kernel 2.6.34 which is
working properly (with the same bug), do you want me to try with it?

Roger

2010/6/4 Stenten <email address hidden>

> Please try the drm-intel-next mainline kernel here [1]. Download and
> install the _all headers, i386 headers, and i386 image.
>
> Then boot into that kernel with 'drm.debug=0x02' in the grub kernel line
> after 'quiet splash'.
>
> Then see if you can reproduce a crash where you can switch to a TTY with
> Ctrl+Alt+F1. Login and then take a copy of these three files:
> cat /sys/kernel/debug/dri/0/i915_error_state > i915_error_state.txt
> cat /var/log/Xorg.0.log > Xorg.0.log.txt
> dmesg > dmesg.txt
> Then try restarting X by running 'startx'.
>
> If you can't switch to a TTY no matter how many times you reproduce,
> upload your kern.log file after you recover, noting the exact date/time
> the crash occurred.
>
> [1]: http://kernel.ubuntu.com/~kernel-ppa/mainline/drm-intel-<http://kernel.ubuntu.com/%7Ekernel-ppa/mainline/drm-intel->
> next/current/
>
> --
> X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[BavarianEagle (ubuntu-zogi)]

Hello,

I do have exactly the same problem: Minutes after resuming from a suspend to RAM the X-server crashes and I have to restart it in low graphics performance mode. In this mode however it does not crash after a suspend to RAM.

I am using lucid (32 bit) on a Toshiba M45 with a Intel 915GM graphics chip and a Itel Pentium M1.6 GHz.
I did not change anything from the default installation.

Thanks!

[FR. Loïc (hackurx)]

Could you attach your /var/log/Xorg.0.log after getting the bug? that might be an xorg crash.

[roger64 (rogqip-suse)]

Hi

Here is a nice crash...I did /var/log... immediately after recovering from
crash in low resolution mode.

Roger

2010/7/7 HacKurx <email address hidden>

> Could you attach your /var/log/Xorg.0.log after getting the bug? that
> might be an xorg crash.
>
> ** Changed in: xserver-xorg-video-intel (Ubuntu)
> Status: Incomplete => Confirmed
>
> --
> [i915gm] X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Stenten (stenten)]

Sorry, yes, can you install the 2.6.35-rc4 kernel [1] (instructions here [2]) and also the xorg-edgers PPA packages [3] and then reproduce the freeze with drm.debug=0x02 in the kernel parameter line and then either switch to a TTY or SSH in and get a copy of Xorg.0.log, dmesg, and the i915_error_state file, all as .txt's? Thank you.

[1]: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35-rc4-maverick/
[2]: https://wiki.ubuntu.com/KernelTeam/MainlineBuilds
[3]: https://launchpad.net/~xorg-edgers/+archive/ppa

[roger64 (rogqip-suse)]

Hi

To sum it up: I tried and failed.

Installing new kernel was OK (using sudo dpkg -i *.deb)
Installing new PPA (add-apt-repository), updating with added new Lucid
packages : was OK

When booting with new kernel, it stopped with message:
(process : 292) : Glib-Warning ** :get pwuid_r ( ) : failed due to unknown
user_id (0)
Recovery menu with new kernel did not work (everything blocked).

Trying to start with old kernel, same error message.
Recovery menu with old kernel: got some long and complicated error messages
about xorg versions, failed to startx

I decided to call it a day and I reinstalled a two days old Clonezilla Live
image and here I am.

Sorry for this.

Roger

2010/7/11 Stenten <email address hidden>

> Sorry, yes, can you install the 2.6.35-rc4 kernel [1] (instructions here
> [2]) and also the xorg-edgers PPA packages [3] and then reproduce the
> freeze with drm.debug=0x02 in the kernel parameter line and then either
> switch to a TTY or SSH in and get a copy of Xorg.0.log, dmesg, and the
> i915_error_state file, all as .txt's? Thank you.
>
> [1]: http://kernel.ubuntu.com/~kernel-ppa/mainline/v2.6.35-rc4-maverick/<http://kernel.ubuntu.com/%7Ekernel-ppa/mainline/v2.6.35-rc4-maverick/>
> [2]: https://wiki.ubuntu.com/KernelTeam/MainlineBuilds
> [3]: https://launchpad.net/~xorg-edgers/+archive/ppa<https://launchpad.net/%7Exorg-edgers/+archive/ppa>
>
> --
> [i915gm] X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Stenten (stenten)]

That's a problem with with xorg-edgers PPA, most likely because it was in the process of updating or one of the daily git snapshots has a bug in it. From the xorg-edgers page:
"To revert to official packages, you can install the ppa-purge package and run "sudo ppa-purge xorg-edgers". If you have driconf it is recommended to remove ~/.drirc when changing versions!"

[roger64 (rogqip-suse)]

Hi

I am ready to resume trials to solve this pesky bug. I know the 2.6.35
kernel just run fine on my laptop but I do not know exactly which
xorg-edgers package I should install to complement it.
I currently have Intel driver 2.9.1 and standard Lucid kernel 2.6.32 and
something.

Cheers

2010/7/14 Stenten <email address hidden>

> That's a problem with with xorg-edgers PPA, most likely because it was in
> the process of updating or one of the daily git snapshots has a bug in it.
> From the xorg-edgers page:
> "To revert to official packages, you can install the ppa-purge package and
> run "sudo ppa-purge xorg-edgers". If you have driconf it is recommended to
> remove ~/.drirc when changing versions!"
>
> --
> [i915gm] X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[roger64 (rogqip-suse)]

Hi,

I downloaded yesterday the Maverick beta iso. I installed it on a USB key and ran it on my computer. The same bug happened: I experimented an X freeze some minutes after suspend.

So, if this can be of some help, this bug appeared in Karmic, was present on Lucid and is now present on the standard iso beta edition of Maverick...

[roger64 (rogqip-suse)]

 I downloaded yesterday and installed Maverick. After waiting several hours
using suspend and resuming, t's my pleasure to inform you that the bug seems
finally to be GONE!!

GONE!

Thank you for all your supporting work. With this, I do not need anymore to
use X in a degraded mode and it seems I have now a new and bright computer.
The planned birthday celebration of the bug has been hastily cancelled.

Cheers

2010/9/12 roger64 <email address hidden>

> Hi,
>
> I downloaded yesterday the Maverick beta iso. I installed it on a USB
> key and ran it on my computer. The same bug happened: I experimented an
> X freeze some minutes after suspend.
>
> So, if this can be of some help, this bug appeared in Karmic, was
> present on Lucid and is now present on the standard iso beta edition of
> Maverick...
>
> --
> [i915gm] X crashes minutes after resuming from suspend to RAM
> https://bugs.launchpad.net/bugs/583902
> You received this bug notification because you are a direct subscriber
> of the bug.
>

[Jouni Mettala (jouni-mettala)]

Thanks. I am marking it fix released then.