On Fri, 24 Jan 2003, Arador wrote:
> machine: p3 2x800, 128 mb of ram, via VT82C686 chipset
> The problem is that in 2.5.59-mm kernels;
> /proc/interrupts shows something like this:
>
> diego@estel:~$ cat /proc/interrupts
> CPU0 CPU1
> 0: 49316 0 IO-APIC-edge timer
> 1: 95 0 IO-APIC-edge i8042
> 2: 0 0 XT-PIC cascade
> 3: 1859 0 IO-APIC-edge serial
> 4: 912 0 IO-APIC-edge serial
> 8: 2 0 IO-APIC-edge rtc
> 10: 1291 0 IO-APIC-level VIA686A
> 14: 2407 0 IO-APIC-edge ide0
> 15: 11 0 IO-APIC-edge ide1
> NMI: 101404 101288
> LOC: 101258 101296
> ERR: 0
> MIS: 0
>
> plain 2.5.59 works well. reverting kirq and kirq-up-fix
> patches (as suggested by Andrew Morton) from the -mm patchset
> fixes the problem.
That's the thing I would expect to see if you used 'noapic' and watchdog.
I posted over the weekend that I have been seeing some inobvious results
to IPC benchmarking with scombinations of noapic and watchdog, but I
didn't snap the interrupts. I'll be happy to add that to my list of stuff
to try next weekend, but the box is not a toy during the week, and I would
have a five hour round trip drive if a reboot failed, so I'll pass on
trying it until I'll in the same room.
What I posted may or may not be related, I was running 2.5.59 as stock as
I could and still compile.
-- bill davidsen <davidsen@tmr.com> CTO, TMR Associates, Inc Doing interesting things with little computers since 1979.
This archive was generated by hypermail 2b29 : Fri Jan 31 2003 - 22:00:17 EST