Hi Andrew,
Thank you for your -mm1 patch to 2.6.2-rc2, I've now got my system up
and running again. I'm not sure where the problem was, since I didn't
find the problem point after sprinkling a bunch of AAA(); debugging
printks() all over the stuff in arch/i386/kernel/setup.c and some
other files in there as well.
Would it make sense for me to try and work backwards from the -mm1 and
see which patch(es) makes the difference?
One thing I notice is that now my interrupts are all on just one
processor:
> cat /proc/interrupts CPU0 CPU1
0: 59000197 20 IO-APIC-edge timer
1: 1867 1 IO-APIC-edge i8042
2: 0 0 XT-PIC cascade
8: 3 1 IO-APIC-edge rtc
11: 1 0 IO-APIC-edge Cyclom-Y
12: 703 1 IO-APIC-edge i8042
14: 281 0 IO-APIC-edge ide0
16: 42762 3 IO-APIC-level ide2, ide3, ohci_hcd
17: 183325 1 IO-APIC-level ehci_hcd, eth0
18: 119211 1 IO-APIC-level aic7xxx, aic7xxx, Ensoniq AudioPCI
19: 0 0 IO-APIC-level ohci_hcd
NMI: 0 0
LOC: 58999640 58999802
ERR: 0
MIS: 26
Should I think about using that daemon to move interrupts around?
Gotta go find the pointer to that stuff again.
Thanks again Andrew!
John