cats hard lockup

Hello!

I have an old CATS ATX-board from Chaltech (is this the same EB110
from Simtec?). It had an uptime of somewhere around 700 days when we
had a power failure. It was running NetBSD 1.4.2 so I deciced to do a
reinstall to 1.6.2 after the crash. The install procedure whent
just fine, but I've run into two problems:

1) The tlp network driver in GENERIC panics whenever I attempt to
access the interface (eg. ifconfig -a). This is probably the MII-bug?
Could anyone point me to a CATS GENERIC kernel binary that has the
old de driver? Or is there a way to make the install kernel mount
root from /dev/wd0a?

2) I got around problem #1 by nicking a 3c905b off a junkheap. I then
ran into a hard lockup problem. I'm not quite sure what causes it.
The lockups seem to happen quite randomly and aren't necessarily
associated with heavy disk activity. They could be associated with
the NIC since on many occasions I've been doing something that causes
network activity when the lockup occurs. The lockup causes the system
to freeze up rock solid. No error messages in any logs, it just stops
responding completely.

As I said, the box was running just fine prior to the power failure.
After that I've installed a second hard disk, changed the NIC and
upgraded the firmware to ABLE 1.97. Does anyone have a clue as to
what causes these lockups and how to fix it?

I still have the original 1.4.2 installation intact on the older
drive. I'm going to detach the newer drive and try running the old
system. If that still works, it's probably not a hardware problem.

-- 
-=[ Count Zero / TBH - Jussi Hämäläinen - email [email protected] ]=-