[intel] xserver lockup, mouse movable, sound in loop, SysRq unusable

Bug #296825 reported by jayprakash
4
Affects Status Importance Assigned to Milestone
mesa (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

Short description:
xserver freeze. Total. Mouse is movable, but anything with keyboard (change to VT1, Ctrl-Alt-Backspace, Ctrl-Alt-Del, Alt-SysRq-B, or anything) is dead. Sound is in a short (0.5-1 sec) loop. CPU is on max (I hear the fan on max).
Absolutely sure way to reproduce any time is to run Steam Counter strike online under wine. Within 3 minutes lockup.

Long desc:
Around a year ago, I had similiar lockups with compiz, but that time Alt-SysRq-B worked always , some workarounds worked. After the move to hardy, no more lockups any kind.
Until this weekend. I instaled (in clean .wine enviroment) Steam Counter strike pack. Local game works (both CS 1.6 and CS:Condition Zero). But after logging onto an online server, freeze every time. (not immediately, but after around 20sec-3minutes). Nonsteam version of CS never had any problem. (I'm running game with compiz turned off).
Here on launchpad are very very similiar bugs (a thought they are the same) "Fixed in intrepid".
So sunday I upgraded.
Now I still have CounterStrike lockups, But now I have also a lower performance of compiz and already had 2 compiz lockups. (plus many other desktop related problems. So "very succesfull" upgrade :( )

I tryed:
* kernels - 2.6.26 (hardy last); 2.6.27-7 generic and server; 2.6.27-8 generic and server (I installed also rt kernel, but I need to try it.)
* xorg.conf - clean (dpkg-reconfigure xserver (no framebuffer)); my old from hardy; this suggestion http://forums.opensuse.org/hardware/388928-using-opengl-google-earth-hardlocks-x-intel-x3100.html#post1836366 . (I'm yet to try XAA, or some other tweaks, but I don't intend to try dri false, I want dri, even I know it will work).

Research:
could be this bug https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-intel/+bug/177518
But that one is marked fixed in intrepid long ago. So I assume I have another bug.
I have no chages in in collors, or pattern on screen (in some other bugs mentioned).
One time I found infinite loop in logs mentioned here https://bugs.launchpad.net/ubuntu/+source/mesa/+bug/279192 . (but in current logs not. That's the reason for filling new bug report.)
Bugs that are filled with 965 seem to have "the little things" different, but maybe its not so much HW related, and it was just another bug.
A have read about similiar bug also on non intel cards. Could by mesa or kernel related.

Hardware:
Fujitsu siemens T4210
Intel graphics 945GM/GMS

Debug:
* "Untrap signals" from https://wiki.ubuntu.com/X/Backtracing
* rebooted in single user mode
* emptyed /var/logs
* normal reboot
* echo 1 | sudo tee /sys/module/drm/parameters/debug
* lockup with countestrike
* boot in single user mode
* copied /var/log

I have no Idea how to strace or gbd a hardlocked pc, so not included.

Included:
whole /var/log (a lot of informations)
lspci -vvnn
xorg.conf

If you need any info, debug, testing, I'm availiable.

I will continue to search for any sollution or workaround.

PS: just because I test crashes with counterstrike, it doesn't mean it schould go to wishlist or low priority. I haven't yet tryed vmware, graphic apps or intensive video play. But two compiz lockups say there will be a problem.

EDIT:
crash happened at Nov 11 12:15:23
propably it's this in kern.log [ 844.717978] and [ 844.726861]
Nov 11 12:15:23 jxt4210-h kernel: [ 844.717978] [drm:drm_lock] 1 has lock
Nov 11 12:15:23 jxt4210-h kernel: [ 844.726861] [drm:drm_lock] 2 has lock

[lspci]
00:00.0 Host bridge [0600]: Intel Corporation Mobile 945GM/PM/GMS, 943/940GML and 945GT Express Memory Controller Hub [8086:27a0] (rev 03)
     Subsystem: Fujitsu Limited. Device [10cf:1378]
00:02.0 VGA compatible controller [0300]: Intel Corporation Mobile 945GM/GMS, 943/940GML Express Integrated Graphics Controller [8086:27a2] (rev 03)
     Subsystem: Fujitsu Limited. Device [10cf:1380]

Revision history for this message
jayprakash (jay-jay-cz) wrote :
Revision history for this message
jayprakash (jay-jay-cz) wrote :
Revision history for this message
jayprakash (jay-jay-cz) wrote :
jayprakash (jay-jay-cz)
description: updated
Revision history for this message
jayprakash (jay-jay-cz) wrote :

Proppably usefull information:

crash is independend of wine graphic settings "vertex shader" and "pixel shader". Still freezes.

Game did not crash in map without wall textures. There were just single color walls. After switching to a map with full textures. Lockup was inevitable.

So bug is possibly texture related.

Revision history for this message
Pauli (paniemin) wrote :

I'm experienze same X/keyboard freeze when trying to play counter-strike.

It seems like it is wine bug as I'm able to resume X after killing hl.exe using ssh connection to my computer.

I'm using wine 1.1.9 prebuild binaries from winehq.

Error message that is related to freeze in .xsession-errors is
err:ntdll:RtlpWaitForCriticalSection section 0x7e47d880 "x11drv_main.c: X11DRV_CritSection" wait timed out in thread 0024, blocked by 0009, retrying (60 sec)

System:
Athlon XP-M 2800+
Ati mobility radeon 9200 (using radeon driver)
The latest intrepid
XAA+PageFlip is enabled for radeon driver

Lockup sometimes happens very soon and sometimes a lot later.
de_dust2 seems to work about 25 minutes
cs_italy freezes nearly immediately after connect
de_nuke works about 15 minutes

Revision history for this message
Pauli (paniemin) wrote :

This wine bug seems like related to this problem: http://bugs.winehq.org/show_bug.cgi?id=8025

Revision history for this message
jayprakash (jay-jay-cz) wrote :

I was able to ssh and kill wine.
System returned to normal.

So it is the same bug as the wine one from pauli.

But, exactly the same freeze happens on all graphic cards (intel, ati, nvidia), so it must by kernel/mesa based bug.

I'm availiable for testing.
Just write what to do, and I'll freeze my system as necessary :)

Bryce Harrington (bryce)
description: updated
Revision history for this message
Bryce Harrington (bryce) wrote :

Hi jay-jay-cz,

Thank you for taking the time to report this bug and helping to make Ubuntu better. You reported this bug a while ago and there hasn't been any activity in it recently. We were wondering is this still an issue for you? Can you try with the latest development release of Ubuntu? (ISO CD images are available from http://cdimage.ubuntu.com/releases/)

If it remains an issue, could you also attach a new /var/log/Xorg.0.log?
Thanks in advance.

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

tags: added: needs-verification
Changed in mesa (Ubuntu):
status: New → Incomplete
Revision history for this message
jayprakash (jay-jay-cz) wrote :

I'm sorry it took so long, lets explain.

After jaunty upgrade NonSteam CS dropped to 4 FPS and Steam CS didn't even care to start. (with EXA. XAA didn't start X, and UXA was crashing few minutes after desktop start)

Today I found xorg-edgers https://launchpad.net/~xorg-edgers/+archive/ppa
Finally is my desktop fast and usable. Unexpectedly also stable.

But Steam errors, not even interface starts. And NonSteam crashes xorg immediately.

So you can close this bug.
It's impossible to debug it.

Maybe after the problems EXA/UXA/GEM (and so on.. ) I'll see if the problem still persists.

Revision history for this message
jayprakash (jay-jay-cz) wrote :

I'm unable to reproduce this bug anymore.
running up-to-date xorg-edgers
https://launchpad.net/~xorg-edgers/+archive/ppa
mesa 7.6
xserver 1.6.2
xserver-xorg-video-intel 2.8.0
kernel 2.6.30-10 - KMS

performance on i945 is poor ( ~ -40% than I remember on XAA - now around 20fps)
But no crashes or lockups on Steam or NonSteam.

I will continue to test.
But for now closing as fixed.

Changed in mesa (Ubuntu):
status: Incomplete → Fix Released
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.