Hauke Fath wrote:
> At 1:57 Uhr +1100 7.1.2003, matthew green wrote:
>>
>>   How would I proceed in gdb, since I do not get a core file?  Is the kernel
>>   linked [at] an absolute address?
>>
>> give the "call X from Y" numbers below, you can feed X and Y into
>> gdb either via "list *X" or "disassemble X", which will show you
>> source (-g kernels only) or assembly ...
> 
> Nice!
> 
> I am inserting into the ctrace output what I got from a 'list *X' / 'list
> *Y' for the first three stack levels.  Sources are from a cvs update a few
> hours ago.
> 
> Unfortunately, I cannot see any details about where on the bottom level (in
> fdcattach_obio) the cpu was when the watchdog went off. The 'PC: f0004050'
> appears to point into nowhere.

Suggestion: Since it looks like it died looking for the floppy drive, why
not build a new kernel with "fd" support commented out/disabled, and see
how much further you get with the boot using that kernel?

	- Greg (who's looking forward to finally being able to upgrade
		his SPARCserver 20/712 from 1.4.2 once there's a stable
		release with SMP support ... 1.6.2?  1.7?  Anyone?  Bueller?)