A follow up to the message that I posted to the list yesterday:

I don't get any crashes if I mount the NFS disk over TCP, only if I
use UDP.  

It's looking like either:

a) there is a bug in the UDP code in the kernel that has been
introduced recently and is tweaked just the wrong way by being run on
a PCMCIA 3c589 card.

b) there is a bug in the PCMCIA 3c589 drivers that is tweaked in just
the wrong way by some changes that were introduced recently elsewhere
in the kernel (the PCMCIA and 3c589 patches worked just fine under
1.0).

I still haven't received any good ideas on how to go about debugging
this.  I guess what would be really nice is a system for tracking
memory management calls in the kernel.  I don't suppose anyone has
written anything like that?  If not, I guess I'll probably have to do
it myself.

All comments appreciated.

-Todd.