Chipset is E7505 with dual Xeons.
Under X, I can provoke a lock just by waggling the mouse. I've had the
machine connected up to a serial console with nmi_watchdog=1 and, when
the machine dies, nothing is printed on the console (I guess that makes
it *very* bad :( ).
Booting with "noapic" helps stability considerably (though I have had
one lockup with even that this morning). I've also tried disabling AGP
support in-kernel and that makes no difference.
Strangely, 2.4 is considerably more stable (without "noapic") - anybody
have any idea what's going on or what to look at next (I'm happy to roll
my sleeves up and stare at code - just need some pointers)?
If you'd like more details, let me know and I'll post them.
Many thanks,
Dan.
P.S. There appear to have been other recent discussions on this topic
with complaints of IDE (I have an MPT/Fusion SCSI setup) and ethernet
load being enough to cause the hard lock.
On Sat, 15 Nov 2003, Dan Creswell wrote:
> Chipset is E7505 with dual Xeons.
>
> Under X, I can provoke a lock just by waggling the mouse. I've had the
> machine connected up to a serial console with nmi_watchdog=1 and, when
> the machine dies, nothing is printed on the console (I guess that makes
> it *very* bad :( ).
Is NMI really enabled?
$ cat /proc/interrupts
- Davide