Michael Graff <explorer@flame.org> writes:

> I don't know if it's the interrupt code or not for CERTAIN, but going
> back to a pre- and then going forward to a post- change kernel seems
> to point that direction.
> 
> Under heavy disk usage, or heavy network usage, my dual AMD machines
> will eventually lock up tight -- not even the kernel debugger can be
> triggered.  I also believe my single-CPU laptop has locked up twice
> due to the new code, once when writing to disk from an rsync over the
> wi0 card.

I've seen this as well since the new interrupt code went in, on my
dual-athlon-MP desktop box. Wedges hard, nothing on the serial
console, doesn't ping, etc. I've noticed it being possibly correlated
with heavy activity on the other serial port.

        - Nathan