background: HP Pavillion dv9000, running openSuSE 10.2, dual processor, etc; kernel 2.6.18-34

problem: most of the time works like a charm (with the exception of the webcam and external sound, but those are minor quibbles and, I am sure, fixable), but--occasionally--crashes on shutdown when setting hardware clock, or on the startup while doing the same. An attempt to set date works fine, but an attempt to use "hwclock --systohc" crashes the machine also. Does anyone know if this is some hardware defect I need to have taken to be fixed at the shop or a byproduct of running Linux on what is, essentially, a box designed specifically for Windoze and fixable by tweaking?

Oh, it boots into a failsafe mode just fine, and into Xen without a glitch, which is what leads me to believe it has something to do with the kernel itself...

Thanks for all suggestions...

It might be helpful if you answer a few questions:

1. Do you have an AMD or Intel processor? Dual Core?

2. How much ram do you have installed?

3. Do you have a default or custom bios?

4. Do you have an Intel, Nvidia, or ATI chipset?

5. What graphics hardware do you have?

I have a DV6604nr with an AMD x2 processor that needs to have both ACPI and APIC turned off when using the default kernel shipped in Suse 10.2, this has something to do with the AMD processor and Nvidia chipset. I do not have any issues when using newer kernels and I do not have to turn off ACPI or APIC.

Sometimes a custom bios will solve many problems, but installing one can be risky. If you have an incompatible or bad flash, you will have many problems reverting back to the default bios. I do not recommend using a 3rd party bios on this machine.

If you can, try to update the kernel by typing "you" at the CLI or going to yast.

Wow, thanks.

1.AMDx2
2.2 gigs
3.Default
4.Nvidia...I am virtually certain ;P
5.GeForce Go 6150

I'll try updating the kernel. Like I said, the weird thiing is the unpredictability. Sometimes I can go for weeks without a problem and then spend half an hour trying to boot the damn thing 4-5 times, crashing every time; then suddenly, all goes great again.

[QUOTE=usaf_sp;3038826]It might be helpful if you answer a few questions:

1. Do you have an AMD or Intel processor? Dual Core?

2. How much ram do you have installed?

3. Do you have a default or custom bios?

4. Do you have an Intel, Nvidia, or ATI chipset?

5. What graphics hardware do you have?

You could try setting iommu=off as a boot option in Grub/Lilo.
There have been reports that it fixes some problems with AMD cpu's.

Thanks, I'll try that. However, the problem has mutated by now. I got the latest Nvidia drivers and it seems to have fixed the crashing issue; however, the new driver is greedy and seems to think it owns all displays and not just X, making all of my text-only consoles unusable once gdm starts up. I tried booting into RL3, but for some reason am not able to start X from it (it tells me that it cannot move Xorg.0.log to Xorg.0.log.old, which has never been an issue before).

You may be taking the wrong route to fixing the problem.
for some reason i get the feeling that your problem isn't kernel related, it would be more chronic

download
http://www.memtest.org/#downiso i use V1.7
burn it to a cd and boot the machine, it will test the ram.

here is my logic on your issue.
when you turn on your box, linux loads the kernel into memory. However, i think it allocates memory dynamically so it works most the time.

If we had a print out of the message you recieve when your kernel crashes it would help.
also, does it seem to matter if it is a cold boot or not?

I'll do that too. But as I said, I suspect now that the Nvidia software is to blame. It is greedy (see the post above yours for more details).

When it was crashing, it happened while setting the hardware clock. Freeze right on that line. It also crashes whenever I try "hwclock --systohw". (I just tried it, and indeed, it is still going on).

Usually cold boot, but sometimes it would happen 3-4 times in a row.

[QUOTE=chrisortiz;3050636]You may be taking the wrong route to fixing the problem.
for some reason i get the feeling that your problem isn't kernel related, it would be more chronic

download
http://www.memtest.org/#downiso i use V1.7
burn it to a cd and boot the machine, it will test the ram.

here is my logic on your issue.
when you turn on your box, linux loads the kernel into memory. However, i think it allocates memory dynamically so it works most the time.

If we had a print out of the message you recieve when your kernel crashes it would help.
also, does it seem to matter if it is a cold boot or not?[/QUOTE

I have the same chipset. My computer crashes if I use the wrong Nvidia driver. I had better luck downloading the driver from Nvidia than using the one in the repository. I would try to uninstall the NV driver and install the one from Nvidia themselves.

Also I have found that for some reason the AMD X2 processor crashes on the older kernels when ACPI and APIC are both on. I also had a problem when the kernel was looking for the Intel random number generator, but I can not be sure of why.

As a side note, I have had very little trouble using OpenSuse 10.3 on this laptop. The biggest problem that I experienced was the monitor backlight not being set by kpowersave.

If you have the newer broadcom wireless card, the instructions for using ndiswrapper are pretty clear cut on the Opensuse website. The only issue with ndiswrapper was that the information on opensuse's website specified only 2 files that were necessary in the extracted directory prior to calling the ndiswrapper install. I have found that my card only works when ALL files of the XP driver was extracted and present prior to calling ndiswrapper.

I am not sure why opensuse 10.3 works better with this chipset, but I have not researched it.