X freezes when booting Thinkpad with DRI enabled

Bug #33368 reported by Carlos Diener / emonkey on 2006-03-02
Affects Status Importance Assigned to Milestone
linux-source-2.6.15 (Ubuntu)
xserver-xorg-video-savage (Ubuntu)

Bug Description

Every time i boot with X, the hole computer freezed after a few seconds. (Dapper Drake 6.04)

I've backuped the hole /var/log an found something in /var/log/kdm.log :
(EE) Failed to load /usr/lib/xorg/modules/extensions/libGLcore.so
(EE) Failed to load module "GLcore" (loader failed, 7)
(EE) Failed to load module "synaptics" (module does not exist, 0)
The hole kdm.log can be downloaded here :

My computer:
IBM Thinkpad T21
kernel 2.6.15-15-686 (386 installed too, same problem)

If you want some other logs or some other information please ask. They're others with the same problem in the german kubuntu forum. We coordinate us in this thread:

I've got gnome installed to. Same problem with gdm/gnome or kdm/gnome.

description: updated

Sorry for the second task, I've never used bevor this bug-tracking system.

Changed the severity and priority because I think this is a breaker and we've different people with this problem on Dapper Drake in our forum. If I'm wrong, feel free to change it again.

My Computer have many different kernels installed because that's an ubuntu installation since Warty, with dist-upgrades to every new Release. I'v tryed to start with 2.6.12-10 kernel and it works to start kdm or gdm without a freeze. The problem exist only with the 2.6.15 kernels. So I assign the bug with the core dev becaus I think it's a problem with the kernel or some modules. If I'm wrong please change, this my first bug report in launchpad and I've got not so much experience with this processes.

Changed in xorg:
assignee: nobody → ubuntu-core-dev
Daniel Holbach (dholbach) wrote :

Please never assign to the core-dev team again, as it sends a mail to every developer. Thanks.

Changed in xorg:
assignee: ubuntu-core-dev → ubuntu-x-swat
Colin Watson (cjwatson) wrote :

Never assign bugs to ubuntu-core-dev; it spams all of us, many of whom have little experience with or clue about the kernel or xorg. In fact, generally in this bug tracking system you shouldn't set the assigned-to field at all unless you're a developer, are managing a developer, or are asked to by a developer, since it indicates that somebody is actually working on the bug at the moment.

I'll send this off to the kernel folks instead.

Changed in xorg:
assignee: ubuntu-x-swat → nobody

I'm really sorry about my mistake, thanks for your correction.

Jeff Stahl (jstahl) wrote :

I can confirm that I've got the exact same problem. X tries to start, the screen flashes some horizontal lines and then I'm left at a black screen w/ an "_" in the upper left hand side. Unable to switch to any other tty and must hard power off or log in via SSH to shut down cleanly.

Using kernel 2.6.15-15-686
IBM Thinkpad T21


I've upgraded to 2.6.15-18-686, the problem still exist. I heard about peoples with other Hardware, I dont' think there's a problem exactly with that Notebook, but maybe with some specific Chipset. I've got a S3 Inc. 86C270-294 Savage/IX-MV (rev 11)

When my computer freezed, after a few seconds the SSH-session freezed too. So can't cleanly shutdown like carpediem91.

Maybe they're some similarities with the following bugs:

wammer (wambam) wrote :

I have the exact same problem, except I can't get X to load at all...ever.. I get that there is a unknown character in libGLCore.so and then all of the modules end up not loading.

I've upgraded to 2.6.15-19-686, the problem still exist.

There was an upgrade of xorg to 7.0.0-0ubuntu22, unfortunately the bug still exist.

Changed in linux-source-2.6.15:
status: Unconfirmed → Rejected

There was another upgrade to 7.0.0-0ubuntu25 and now it WORKS! :)

Unfortunately now it doesn't work again. I've activated ACPI with the klaptop-ui. After the next reboot it freezed. Now I've the same problem as before. :( I'll try to find out how I can deactivate it, deactivating with the kernel option acpi=off doesn't work. I wrote this with lynx, hopefully I can use a X-Browser soon again. :)

bugfriend (mail-scheuler) wrote :

I found out, that this could be a timing thing. After booting recovery mode, i started kdm, no problem. So i got the idea to insert a little sleep to the /etc/init.d/kdm script.............and it worked. After inserting a "sleep 3" to the script, i got no more hangups when booting my T21.

Nice idea but it don't work for me and I think my problem is different. I've removed kdm and use the startx command to start X.

Till some update the computer don't freeze anymore, only the screen becomes black. Remote login with SSH still works and it seems like I still can use the local shell but without any output on the screen.

If I comment out "load GLcore" in Section Module, the computer freeze immediately after the startx command.
Here is the /var/log/Xorg.0.log: http://www.kubuntu.de/forum/forum.php?req=thread&download=15791&inline=1&key=21BED1E4

Knoppix 4.0.2 DVD with pci=noacpi acpi=off and kernel 2.6.12 works fine. Ubuntu with 2.6.12 don't work so I think there's a problem with xorg7. I'm not a specialist, my speculations really could be wrong.

Keep on the good work and thanks for your effort!

Can you attach your Xorg.0.log and xorg.conf (click "Add Attachment" at the left side of this page). The thread of the link you listed here does not exist anymore thanks.

Changed in xorg:
status: Unconfirmed → Needs Info

emonkeys Xorg.0.log from the 7th of April 2006

This and the log above are from an Dapper which is dist-upgraded to every Release since Warty. For testing I've installed a new Dapper to a second partition. The log and the config below are from the new "fresh" installation.

I've seen just that there's something other really wrong with the new installation on the second partition, so don't know how much useful these two files are for you. If it's useful for you I can do a new installation on another partition, these laptop is an testing machine only for Dapper. Let me know what I can do to help.

Sorry for the late reply. I just looked at the log files. I'd like to confirm this, but we're already in Beta. Is there a chance you can test this in beta and confirm it still persists?



I can confirm that the problem still exist after a dist-upgrade. I'll try it with a new Installation with an Ubuntu-Install-CD and tell you the results.

I've downloaded the Live CD instead the install CD. But this "mistake" showe me some interessant fact. The first Kubuntu 6.06 Beta freezed shortly after X has started, too. So I've tryed to start with the "save graphic mode" boot option and that works! Unfortunately this freeze with the "normal" boot option I've no logs 'cause it's a live CD...

I'm going this weekend to the mountains so I can try it with the install CD and the Beta 2 soonest on Monday the 1st of May.

I really hope someone can fix this problem before the 1st of June. I'll do my best to give you all the information which are useful to solve this problem. Thank you anyway for your commitment.

x.org from the Live CD (K Beta 1) "save graphic mode"


I have this error but the computer does not "freeze", it just goes into tty and X will not start

Tormod Volden (tormodvolden) wrote :
Andy Brook (javahollic) wrote :

just tried breezy (dual display nvidia driver ok) -> Flight7 upgrade -> kernel 22.34 upgrade, still X will not start due to the above.

I get the following
dlopen: /usr/lib/xorg/modules/extensions/libGLcore.so undefineded symbol: __glXLastContext
Failed to load module "GLcore" (loader failed, 7)
Failed to load module "nvidia" (module does not exist, 0)
No drivers available

Changed in xserver-xorg-driver-savage:
assignee: nobody → ubuntu-x-swat
inos (guehring+ubuntu) wrote :

I got the same on a IBM ThinkPad T20 (Dapper, happend with all kernel-versions, Breezy was fine). As the T21 has also a S3 Savage chip, it seems to be the same Bug as in 33617 (sorry, dont know how to insert a link).
so it is definitely related to DRI. (disabling DRI resolve the problem)
I found some further information on thinkwiki.org (www.thinkwiki.org/wiki/S3_Savage_IX8)

   Option "BusType" "PCI"
   Option "DmaMode" "None"
in xorg.conf also resolve the issue but DRI is still disabled

hopefully you may further investigate the bug with this information

I can confirm that the problem is solved if I disable the DRI Module. Not a Solution but better than to work with vesa.

Please ask if you need any further information, logs or configs. Thanks.

Changed in xserver-xorg-driver-savage:
status: Needs Info → Confirmed
Swen (kernelpanic-1-0) wrote :

Confirmed with Xubuntu on a IBM Thinkpad T21. Adding

Option "ShadowStatus"

to Xorg.conf Device Section did not solve it, but

Option "BusType" "PCI"
Option "DmaMode" "None".

Brian Geppert (briangeppert) wrote :

With a fresh install of 6.06, GDM freezes within a second or two of opening. Once I succeeded in typing my username and password and logging in, only to have X freeze whilst loading the desktop. Adding a 'sleep 5' to the GDM script does not help. This is with the 'nv' driver. I have an Nvidia 5700 OC, an Athlon X2 3800+ and an Elitegroup K8T890-A. I experience this with both -386 and -686 kernels.

By booting in "recovery mode", I can successfully launch GDM and/or startx from the console without freezing. However, if I enable the nvidia drivers, my screen immediately goes black and soon times out/goes into power-saving mode.

I have attached my xorg.conf

Brian Geppert (briangeppert) wrote :

s/Elitegroup/ECS Elitegroup/

AdamPrice (adamkprice) wrote :

Mine was working finwe untill yesterday when i got a xserver update. today im getting the same as above. any ideas whats happened? Im on a samsung x50

AdamPrice (adamkprice) wrote :

Additional to this in my Xorg.0.log i have the following statements:

EE No devices detected.

Fatal server error:
no screens found

Brian Geppert (briangeppert) wrote :

I have upgraded to Edgy Eft, and I am still experiencing the same issues with both -386 and -686 kernels.

Still the same problem with Edgy too. Works only if I deactivate DRI.

This Bug looks like a Duplicate from this one:

Tormod Volden (tormodvolden) wrote :

Carlos, does the DmaMode and DmaType options help in your case? Can you please run "sudo cat /proc/kmsg" in a ssh session and see if there's any messages before it crashes?

Andy, Brian and Adam, you don't have a savage card, so please find/file another bug than this one.

Brian Geppert (briangeppert) wrote :


Tormod Volden (tormodvolden) wrote :

I am not totally sure Carlos' problem is a duplicate, because that other bug sounds more like carpediem's. Please reopen if you like, but please tell us if DmaMode and BusType settings helped (see also man savage).

BusType PCI in Device Section resolved the problem.

I'm not sure it's a duplicate, but my problem is solved. Needless to say that it isn't ideal because it should work out of the box. So if you want to close the bug, for me it's ok.

I wrote an answer on the other bug too:
(But maybe that's another problem from Brian Heung, in my case the hole system freezed.)

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers