For the past week or so, I've been experiencing random lockups in X under SuSE 9.2 Pro. In the last week, there have been kernel, x11-xorg libs, kde-libs and kdebase updates but I think it is the kernel that's responsible since the lockups began shortly after updating the kernel.

Tyan Thunder K7 dual AMD motherboard
AMD Athlon 2000+ (single CPU with second CPU socket left unoccupied)
ASUS GeForce Ti 4200 128MB card (NVIDIA 6629 driver)
Samsung 172T LCD monitor (DVI <---> DVI connection)
on board ATI Rage XL VGA unused

I downgraded kdelibs3 and kdebase3 to the versions before last week
kdebase3-3.3.0-29
kdelibs3-3.3.0-34

and this has reduced the number of lockups a bit. I rarely get a lockup under icewm. Scanning these forums, nvidia's forums etc. hasn't revealed anything. Anyone else have this problem?

Anand

I am running SUSE 9.2 pro using KDE 3.3 and I also get frequent lockups. I didn't get lockups when I was using GNOME. Ctrl-esc do and ctrl-alt-backspace do nothing. I have an nvidia card also. Hmmmmmm

While I'm not positive about this, I think that one of the reasons for my lockups was that I had followed these instructions for speeding up firefox. When I returned firefox's configuration to its default state, I didn't see anywhere near as many lockups. Now, I get a lockup pretty rarely and I think all the current lockups are GLX related.

Anand

I am running firefox, but I am not running the speedup program. I noticed that I don't get these lockups when I am using GNOME. It must be something KDE related. Do you know how I can check my previous logs? When my computer locks up I have to restart so I can check the current one.

Hello,

can you please take a look on your /boot/grub/menu.lst
(You have to be root to open this file)
Is there a boot option called "desktop" ?
e.g.:

kernel (hd1,8)/vmlinuz root=/dev/hdb10 vga=0x31a selinux=0 splash=silent resume=/dev/hdb11 desktop elevator=as showopts

I have had several lockups with KDE after installing the ATI drivers with 3D acceleration enabled.

The option desktop influces the kernel process schedular. After removing the desktop option,
I didn't have any lockups anymore.

Maybe this can help!

As root, you can check the number of occasions that X has crashed by

[root@yourmachine] cd /var/log
[root@yourmachine] grep NVRM messages

However, on SuSE 9.2, there may be more than one messages file indexed by date. if you repeat the above procedure for the different messages files (after gunzipping them), you'll find out the number of times X has crashed and how.

You can also run nvidia-bug-report.sh from the command line and submit it to the nvnews.net linux driver forum. nvidia-bug-report.sh should be in your path. Open up a xterm, su root and run it. It'll generate a log file in the same directory. Submit the log file to the nvidia forum and you should get some responses and clues to what's happening.

Anand