Bugzilla – Bug 152960
Kernel Update results in kernel panic
Last modified: 2006-02-24 11:39:55 UTC
My system hangs with kernel panic and the follwing message: BAD EIP value Kernel panic - not syncing: attempted to kill init! If you "google" for this message you will find many people with the same problem, but each of them has no/different solution. The default 10.0 kernel runs fine, but nighter 2.6.15.7 nor 2.6.15.8 can boot. I tried failsave, but it hangs, too. a) what does bad eip mean? b) how to resolve? Mainboard is a Soltek SL-75DRV4 with VIA KT266A chipset.
Perhaps your custom kernel was not build or installed properly? We don't support non-SuSE kernels, sorry. We suggest asking questions like this on kernel development mailing lists, might I suggest the kernelnewbies mailing list for things like this?
Sorry, wrong version numbers: 2.6.13-15.7 and 2.6.13-15.8 are affected. 2.6.13-15 is OK. That's why you misunderstood it. With other words: The official kernel-updates from Suse/Novell causes the crash. I did not compile a new kernel. Please see the screenshots I will attach now.
Created attachment 69882 [details] kernel panic with 2.6.13-15.8 Same happens with 2.6.13-15.7. Original Suse 10.0 kernel boots fine.
Yes, sorry, I misunderstood your original description.
It dies during the early stages of system setup, when setting up sysfs. I suspect this is a memory problem. Please boot from the installation CD, select memtest and run it overnight.
I did so when it first failed to boot with kernel 2.6.13-15.7. No errors found then. I waited with the bugreport for the next kernel update, and reported it now, as it still exists. Nevertheless I will try another piece of memory and run memtest once more overnight.
Does it boot if you use the Failsafe boot entry or if you try to add additional boot parameters (e.g. 'apic' or 'pci=noacpi' or 'acpi=off') ?
No, it doesn't. Tried almost every boot-option :-(
You won :-) I have two memory modules and after changing the slots the system booted. I run then a memory test with the module, which caused the failure, and after 2 hours of testing it showed the first error. So it's a memory error.
You won, too - now you can stop trying different combinations of kernel command line options :) Thanks!
Silly question: Will the bad memory (two places: 1,6 MB and 114,6 MB) become worse automatically, like a chain reaction?
Every memory error can be worse, depend what data/code is changed. I would not try to use such modules any longer, if the BIOS RAM settings are correct for these modules. Lot memory errors came from wrong timings, sometimes the manufacturer also set wrong values in the eeproms, but in most cases people try overclocking or higher manual speed settings or combine different modules ... If you get more errors or not depend what is the reason behind the error. But anyways this is the wrong place to discuss such things, please go to some HW forum or mailing list.