gdm_slave_xioerror_handler: Fatal X error

Bug #152648 reported by Chris Coulson on 2007-10-14
30
This bug affects 2 people
Affects Status Importance Assigned to Milestone
gdm (Ubuntu)
Low
Unassigned
linux-restricted-modules-2.6.22 (Ubuntu)
Undecided
Unassigned
linux-restricted-modules-2.6.24 (Ubuntu)
Undecided
Unassigned

Bug Description

When using the new fast user switch applet to login a second user on my system, X restarts when left at the login screen for more than a few seconds. The error in my syslog is:

Oct 14 15:26:45 localhost gdm[13647]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

Steps to reproduce:
1 - Log in user 1 on first display
2 - Select a second user from the user switch applet
3 - Leave GDM screen showing for a few seconds without entering password

Chris Coulson (chrisccoulson) wrote :

Actually, it still appears to crash even after logging in the second user. This is up-to-date Gutsy RC with nvidia 100.14.19 drivers

Sebastien Bacher (seb128) wrote :

looks like xorg crashing, reassigning

Bryce Harrington (bryce) wrote :

Sounds like a -nvidia binary driver problem. Reassigning.

bojo42 (bojo42) wrote :

Maybe this bug is related to https://bugs.launchpad.net/ubuntu/+bug/140554 . Sounds similar just without the user switching.

Chris Coulson (chrisccoulson) wrote :

bojo42: I think the message 'WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0' is just a generic message saying that something has caused X to crash, and could be caused by anything. The backtrace in my GDM log indicates that my crash is caused by a different problem to the bug report you refer to.

I still get this quite frequently, but it only ever seems to be when 2 users are logged in. I'm attaching my GDM log, which contains a backtrace from the crash (something I should have done when I opened the bug report). The backtrace always seems to look the same after each crash.

Chris Coulson (chrisccoulson) wrote :

Bryce: Does the backtrace contain enough information to be sure that this is a nvidia binary driver problem? If so, I'll pop over to the nvidia forums

Basilio Kublik (sourcercito) wrote :

Hi there
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 development version of Ubuntu, Hardy Heron?

Thanks in advance.

Changed in gdm:
assignee: nobody → sourcercito
importance: Undecided → Low
status: New → Incomplete
Chris Coulson (chrisccoulson) wrote :

Basilio - I was just about to respond to say that I haven't seen this bug for a while on Gutsy, and I that I hadn't seen it at all on Hardy- but it just happened on Hardy beta for the first time. My girlfriend logged in first, then I switched user using fast-user-switch (which is set to leave screen unlocked), then several minutes later - her Xorg dies on screen :0, causing GDM to restart and a login screen to reappear. I then logged in, and returned to my session where I left off before I was rudely interrupted by the crash.

Some info:

chr1s@chris-desktop:~$ lsb_release -rd
Description: Ubuntu hardy (development branch)
Release: 8.04
chr1s@chris-desktop:~$ apt-cache policy linux-restricted-modules
linux-restricted-modules:
  Installed: (none)
  Candidate: 2.6.24.12.13
  Version table:
     2.6.24.12.13 0
        500 http://archive.ubuntu.com hardy/restricted Packages
chr1s@chris-desktop:~$ apt-cache policy nvidia-glx-new
nvidia-glx-new:
  Installed: 169.12+2.6.24.11-12.31
  Candidate: 169.12+2.6.24.11-12.31
  Version table:
 *** 169.12+2.6.24.11-12.31 0
        500 http://archive.ubuntu.com hardy/restricted Packages
        100 /var/lib/dpkg/status
chr1s@chris-desktop:~$ apt-cache policy xserver-xorg
xserver-xorg:
  Installed: 1:7.3+10ubuntu7
  Candidate: 1:7.3+10ubuntu7
  Version table:
 *** 1:7.3+10ubuntu7 0
        500 http://archive.ubuntu.com hardy/main Packages
        100 /var/lib/dpkg/status
chr1s@chris-desktop:~$

Also attached is the Xorg log

Basilio Kublik (sourcercito) wrote :

Hi Chris
The log, as the previously posted one, suggest that this is indeed a restricted driver issue and doesn't have to do with gdm nor fast-user-switch-applet, so i'm closing this part of the bug report to center the attention on the right package, about the gdm_slave_xioerror_handler is just a generic error which can mean almost anything.

Thanks for your time and feedback.

Changed in gdm:
assignee: sourcercito → nobody
status: Incomplete → Invalid
Chris Coulson (chrisccoulson) wrote :

If it's any help, this crash occurs when the screensaver activates on the inactive screen. I've got the screensaver set to 10 minutes, and it always crashes 10 minutes after switching users. If I lock screen and then switch users, Xorg crashes on the inactive screen as soon as the second user logs in.

Chris Coulson (chrisccoulson) wrote :

After doing some debugging, it seems that this crash only occurs when 'BackingStore' is enabled in my xorg.conf. With the option commented out, I can log in multiple users without a crash (so far, so good), so I'm marking it as invalid now.

Changed in linux-restricted-modules-2.6.22:
status: New → Invalid
Andrey Vihrov (andrey.vihrov) wrote :

I'm experiencing this bug also on 64-bit Ubuntu Hardy with latest updates. I'm using Nvidia proprietary driver, indeed (version 169.12). gdm crashes almost every time I press the Switch User button.

Chris Coulson (chrisccoulson) wrote :

Andrey: Do you have the 'BackingStore' option in your /etc/X11/xorg.conf? Do you know if it is the same crash?

Andrey Vihrov (andrey.vihrov) wrote :

No, this is the first time I hear about that option. Nvidia drivers manual also says nothing.

Chris Coulson (chrisccoulson) wrote :

Can you post the backtrace from the end of /var/log/Xorg.0.log.old (or it could be one of the other Xorg log files there).

Andrey Vihrov (andrey.vihrov) wrote :

I found one with such a backtrace, indeed. The log contains usual messages, and then is this:

 Backtrace:
 0: /usr/bin/X(xf86SigHandler+0x6d) [0x48672d]
 1: /lib/libc.so.6 [0x2ba6ac6be7d0]
 2: /usr/lib/xorg/modules//libfb.so(fbSolidFillmmx+0x140) [0x2ba6af1b2c90]
 3: /usr/lib/xorg/modules//libfb.so(fbFillRegionSolid+0xc5) [0x2ba6af1ab185]
 4: /usr/lib/xorg/modules/drivers//nvidia_drv.so(_nv000804X+0x199) [0x2ba6aebb4ec9]

 Fatal server error:
 Caught signal 11. Server aborting

So this is an Nvidia bug indeed.

Chris Coulson (chrisccoulson) wrote :

Could you post your /etc/X11/xorg.conf please? Also, your ~/.nvidia-settings-rc if you have one

Changed in linux-restricted-modules-2.6.24:
assignee: nobody → chrisccoulson
Andrey Vihrov (andrey.vihrov) wrote :

Attached. Once I have enough time, I'll try the default options rather than these.

Andrey Vihrov (andrey.vihrov) wrote :
Chris Coulson (chrisccoulson) wrote :

Yes, please try removing some options. I've posted my xorg.conf, which works at my end - this will give you an idea of which options to try and remove from yours in order to pinpoint the problem.

Chris Coulson (chrisccoulson) wrote :

One other thing, I notice from your .nvidia-settings-rc that you are overriding applications FSAA settings. Do you actually load your nvidia-settings-rc (using nvidia-settings --load-config-only, usually added to session preferences)

Andrey Vihrov (andrey.vihrov) wrote :

hmm, I didn't do that since I used nvidia-settings only a few times and primarily to view information about the card. However, now when you say that, it seems that X server doesn't crash anymore if I execute the command.

Andrey Vihrov (andrey.vihrov) wrote :

Well, today the problem appeared again. Now I have commented out all the custom xorg.conf options, let's see what happens.

This is the backtrace of today's crash:

 Backtrace:
 0: /usr/bin/X(xf86SigHandler+0x6a) [0x48402a]
 1: /lib/libc.so.6 [0x7f704b1b8100]
 2: /usr/lib/xorg/modules/drivers//nvidia_drv.so(_nv000759X+0x68) [0x7f7048c54b78]

Chris Coulson (chrisccoulson) wrote :

When I debugged my crashes, I found that there were 2 problems causing them:

1) 'BackingStore' option in xorg.conf causes crashes when switching user every time.
2) OpenGL applications (glxgears, googleearth) cause Xorg to crash every-time when overriding FSAA settings in nvidia-settings. To fix this, I added the 'UseCompositeWrapper' option to my xorg.conf and I can now use FSAA.

Andrey Vihrov (andrey.vihrov) wrote :

Following days of experiments, here are the results:

1. Default options do not help
2. Disabling/erasing .nvidia-settings-rc doesn't help
3. Option 'UseCompositeWrapper' doesn't help

I've run out of ideas, and thereby am forced to use 'nv' until any progress/change. ):

Maybe the bug isn't in NVIDIA itself, but in it's combination with Xorg 7.3. btw, Xorg states it's a pre-release and is not supported by original maintainers in any way; and we have that in LTS (!).

webservant (doulos) wrote :

I'm having a similar problem. I can reproduce the problem at will with firefox. When I go to my company's external email site (Microsoft OWA), if I click on the link for a large email (2mb), the email partially loads and then gnome crashes.

Xorg.0.log.old has this:

Backtrace:
0: /usr/bin/X(xf86SigHandler+0x7e) [0x80c780e]
1: [0xb7fd8420]

Fatal server error:
Caught signal 11. Server aborting

syslog has this:

May 11 13:14:25 Luther kernel: [ 2045.094801] Xorg[5067]: segfault at fcfefd03 eip b7dbfff9 esp bfa754fc error 4
May 11 13:14:25 Luther console-kit-daemon[5424]: WARNING: Unable to activate console: No such device or address
May 11 13:14:25 Luther kernel: [ 2045.147949] compiz.real[6221]: segfault at 02800001 eip 08055a80 esp bfcb4b20 error 4
May 11 13:14:25 Luther gdm[5063]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
May 11 13:14:36 Luther pulseaudio[6694]: pid.c: Stale PID file, overwriting.
May 11 13:14:37 Luther kernel: [ 2056.316377] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.316511] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.337973] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.338374] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.364073] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.364377] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther hcid[5693]: Default passkey agent (:1.47, /org/bluez/passkey) registered
May 11 13:14:37 Luther hcid[5693]: Default authorization agent (:1.47, /org/bluez/auth) registered
May 11 13:14:37 Luther kernel: [ 2056.500886] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:37 Luther kernel: [ 2056.501198] VMBlock warning: DentryOpRevalidate: invalid args from kernel
May 11 13:14:38 Luther NetworkManager: <info> Updating allowed wireless network lists.
May 11 13:14:38 Luther NetworkManager: <WARN> nm_dbus_get_networks_cb(): error received: org.freedesktop.NetworkManagerInfo.NoNetworks - There are no wireless networks stored..

Here is my xorg.conf

webservant (doulos) wrote :

Here is my version information:

doug@Luther:~/xorg-bug$ lsb_release -rd
Description: Ubuntu 8.04
Release: 8.04
doug@Luther:~/xorg-bug$
doug@Luther:~/xorg-bug$ apt-cache policy linux-restricted-modules
linux-restricted-modules:
  Installed: (none)
  Candidate: 2.6.24.16.18
  Version table:
     2.6.24.16.18 0
        500 http://us.archive.ubuntu.com hardy/restricted Packages

doug@Luther:~/xorg-bug$ apt-cache policy nvidia-glx-new
nvidia-glx-new:
  Installed: 169.12+2.6.24.12-16.34
  Candidate: 169.12+2.6.24.12-16.34
  Version table:
 *** 169.12+2.6.24.12-16.34 0
        500 http://us.archive.ubuntu.com hardy/restricted Packages
        100 /var/lib/dpkg/status

doug@Luther:~/xorg-bug$ apt-cache policy xserver-xorg
xserver-xorg:
  Installed: 1:7.3+10ubuntu10
  Candidate: 1:7.3+10ubuntu10
  Version table:
 *** 1:7.3+10ubuntu10 0
        500 http://us.archive.ubuntu.com hardy/main Packages
        100 /var/lib/dpkg/status

Chris Coulson (chrisccoulson) wrote :

webservant- What NVIDIA card have you got? Are you sure you aren't suffering from bug 212648?

webservant (doulos) wrote :

Actually, I am. Thanks. I followed the diagnostic and reproduced the problem. Then put the symbolic link in place, and all is well now. Simply awesome!! Thanks!

Chris Coulson (chrisccoulson) wrote :

Andrew - What NVIDIA card have you got?

Changed in linux-restricted-modules-2.6.24:
assignee: chrisccoulson → nobody
Andrey Vihrov (andrey.vihrov) wrote :

If you mean me ;), then NVIDIA GeForce 7300 GS (also can be viewed in xorg.conf post up somewhere)

webservant (doulos) wrote :

My NVIDIA is GeForce 8300 GS. It's in a Dell bought with Ubuntu preloaded.

I'm experiencing this bug as well, although it doesn't happen very often. Sometimes when switching desktop X crash. It also happen when using "chvt" from the shell (i.e. when I want to use vnc trough ssh and need to switch user first).

Backtrace:
0: /usr/bin/X(xf86SigHandler+0x6a) [0x48402a]
1: /lib/libc.so.6 [0x7f603e262100]
2: /usr/lib/xorg/modules/drivers//nvidia_drv.so(_nv000759X+0x68) [0x7f603b6d0b78]

Fatal server error:
Caught signal 11. Server aborting

GeForce 6500 rev 161, NVIDIA GLX Module 169.12

I saved the XFree86 log if you want it.

Alexander Frolushkin (hokum) wrote :

AMD64 Hardy.
Also have this bug, appears while switching to from one X session to another one via fast user switch or by ctrl+alt+Fn, or to any tty console.

Backtrace:
0: /usr/bin/X(xf86SigHandler+0x6a) [0x48402a]
1: /lib/libc.so.6 [0x7f8210878100]
2: /usr/lib/xorg/modules/drivers//nvidia_drv.so(_nv000746X+0x93) [0x7f820e3158f3]

I'm using Nvidia 8800GT and NVIDIA 169.12 from ubuntu repo.

Andrey Vihrov (andrey.vihrov) wrote :

It's still present in the new 173.14.05 driver pack.

I get this problem with only 1 user. The computer crashes at random - no matter if I'm at the computer or not. My graphics card is an ATI 9800XT (with the proprietary drivers enabled).

The error started to appear after I installed a new SATA2-drive (samsung, 500Gb) to my machine. Before this there were no problem. Until yesterday I hadn't found any clues in the logs but yesterday my computer crashed twice and on both occasions the following log entries were found in daemon.log

Jun 19 18:24:42 Lysander gdm[5744]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting: 0
Jun 19 23:42:45 Lysander gdm[5774]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting: 0

Attached is the Xorg.0.log from the time of the later crash.

Chris Coulson (chrisccoulson) wrote :

Matthias:

Your log doesn't contain any details of the crash (the log must be from the session directly after the crash, otherwise a good log gets rotated in to it's place and the important information is lost), and your problem is not related to this one. This is a bug in the NVIDIA driver, and you appear to be using an ATI card. The error in the log (gdm_slave_xioerror_handler) is a very generic message for many different crashes. Please open up a new bug report for your problem (referring to https://wiki.ubuntu.com/X/Reporting)

bsh (bsh) wrote :

i have this bug too. just started to happen today.
i am on 7.04 i386 desktop (running on a file server). the server is headless, so i log in via vnc on screen :1.
since today, when i log in to the desktop using vnc, then close the vnc session, the next time i log in, i am not getting back to the desktop where i left it, but to the login screen again. when i close the vnc session, it crashes X and restarts it, with the above error message.
therefore i doubt it's a nvidia driver issue.
yesterday i installed webmin, but it's unlikely to have anything to do with this.

the other thing i did yesteday was, i found xubuntu-system-tools in the synaptic package manager (i'm using xfce) and installed it, which caused the removal of gnome-system-tools and ubuntu-desktop packages. and in system logs, the x crash error message always appear before the line "gconf is not in use, shutting down".
so i believe, the crash is caused by yesterdays auto-removal of gnome-system-tools.
i check this out now, will reinstall gnome-system-tools

Chris Coulson (chrisccoulson) wrote :

bsh - Remember that the gdm_slave_xioerror_handler message is a very generic message and could be caused by many different things. That message doesn't contain any useful information and doesn't mean that you're seeing the same crash as other people with that message. The only useful information will be contained in a backtrace of the crashing Xorg process. If the backtrace indicates that you're experiencing a different crash to other people in this bug report, then you need to open a new bug report.

And removing gnome-system-tools will have nothing to do with Xorg crashing.

Thanks

Andrey Vihrov (andrey.vihrov) wrote :

Just dropped here to say 173.14.09 is affected also. I forgot about this problem for some time, since I rolled back to 100.xx, but yesterday's update to the .25 kernel forced me install the new drivers. Unfortunately, I found the issue is still present. :-( I'll try to install 177.70 within the next three days.

ikus060 (ikus060) wrote :

Hi,
I buy a NVidia card a while ago and notice this problem very often.

Do any people here have some workaround ? because it's become very annoying this week. EVERY time my computer going in screen saver mode for 20 min of more, the X server crash. So I really need some advice to solve the problem.

Thanks

    Patrik Dufresne

I got fed up last week and disabled the compiz-fusion window manager
and noticed a remarkable drop in the number of crashes. This isn't a
workaround - it's just amputating the problem.

//Matthias

If you don't need any 3D features you can also disable the nvidia
accelerated driver. It will use the open-source driver instead which
doesn't suffer from this problem.

--
Thomas

Andrey Vihrov (andrey.vihrov) wrote :

By the way, 173.14.12 does not seem to be affected so far.

Frodon (frodon) wrote :

Let me join the group.

I experience xsession crashes with intrepid ibex with nvidia driver 173 and 177 present in the repos. For me this issue is not random and i have been able to identify each time what to do to make session crash :
1- right click any Gdesklet
2- play with system-monitir

The strange part in my case is that if i enable metacity composite manager then i can't reproduce the crash, it goes the same if i replace metacity by xfwm4 the xfce WM.

Anyone tried the 180 driver and can confirm if it solved the issue ?

Glad to see that at least i'm not alone.

Martin Tuzinsky (mato-tuzo) wrote :

i experience this crash while switching from external to notebook screen using xrandr. there is just
gdm[6844]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0
in syslog and i didn't find any error or backtrace in the Xorg.0.log.

interpid ibex,
xserver-xorg-video-intel_2.4.1-1ubuntu10_i386,
xorg_7.4~5ubuntu3_i386,
00:02.0 VGA compatible controller: Intel Corporation Mobile GM965/GL960 Integrated Graphics Controller (rev 03)

ash (jonash99) wrote :

I noticed a similar problem using 8.04, when in a terminal and pushing a down arrow when at the bottom, or left when at the left edge of the characters.
I had recently installed a wacom tablet, and having had trouble with "pad" before decided to take this out of my xorg.conf
And it solved the issue, I don't know wheat the pad input does, am sure I can live without it

---------------------------------<cut
# Section "InputDevice"
# Driver "wacom"
# Identifier "pad"
# Option "Device" "/dev/input/wacom"
# Option "Type" "pad"
# Option "USB" "on"
# EndSection

------------------------------<cut

Section "ServerLayout"
 Identifier "Default Layout"
  screen 0 "Default Screen" 0 0
 Inputdevice "Generic Keyboard"
 Inputdevice "Configured Mouse"
 Inputdevice "stylus" "SendCoreEvents"
 Inputdevice "cursor" "SendCoreEvents"
 Inputdevice "eraser" "SendCoreEvents"
   # Inputdevice "pad"
EndSection

Craig Hewetson (craighewetson) wrote :

Just experienced this on Jaunty (Using Nvidia Graphics, driver:180.44)

snippet from log file:
Oct 7 08:55:06 Baegon kernel: [ 2034.200431] NVRM: Xid (0001:00): 6, PE0001
Oct 7 08:55:06 Baegon kernel: [ 2034.299568] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.399035] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.497329] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.580576] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.664573] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.750847] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:06 Baegon kernel: [ 2034.855478] NVRM: Xid (0001:00): 13, 0001 00000000 00005097 0000194c 00000000 00000100
Oct 7 08:55:08 Baegon gdm[3163]: WARNING: gdm_slave_xioerror_handler: Fatal X error - Restarting :0

....

Oct 7 08:55:12 Baegon gdmgreeter[6964]: GdkPixbuf-CRITICAL: gdk_pixbuf_get_width: assertion `GDK_IS_PIXBUF (pixbuf)' failed
Oct 7 08:55:12 Baegon gdmgreeter[6964]: GdkPixbuf-CRITICAL: gdk_pixbuf_get_height: assertion `GDK_IS_PIXBUF (pixbuf)' failed
Oct 7 08:55:12 Baegon gdmgreeter[6964]: GLib-GObject-CRITICAL: g_object_unref: assertion `G_IS_OBJECT (object)' failed

mc (matic-cankar) wrote :

similar on Ubuntu 9.04, Nvidia driver version 180, GeForce 8800 GT

Oct 13 10:29:27 xxxxxxxx kernel: [ 965.665719] NVRM: Xid (0001:00): 6, PE0001
Oct 13 10:29:27 xxxxxxxx kernel: [ 966.044872] Xorg[3142]: segfault at b6918000 ip b6825de6 sp bfefaa50 error 6 in nvidia_drv.so[b64f0000+3b4000]

Bryce Harrington (bryce) wrote :

Hi all,

Unfortunately, while the title of this bug looks technical and specific, in truth all it means is "X crashed for some reason". You'll see this message in gdm logs for any variety of different crashes, but I'm afraid people are finding this bug report and since the error message is the same are assuming they have the same bug. So I'm going to close this bug report so it doesn't continue to mislead people.

By default the information displayed in the log files when a crash occurs is insufficient to debug the issue. Instead, you'll need to collect a full backtrace - see http://wiki.ubuntu.com/X/Backtracing for guidance on this. (Although with -nvidia since it's a binary proprietary driver, even with a backtrace we likely won't be able to fix it at the distro level anyway. Still, backtraces are worth collecting just in case.)

Changed in linux-restricted-modules-2.6.24 (Ubuntu):
status: New → Invalid
To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers