segfaults, kernel panic using forcedeth nvidia 4

[emf <i@mindlace.net>]

Hello!

I'm using debian's 2.6.17-2-amd64 (after having these issues with 2.6.16).

I'm running software raid 5 and LVM on the partition in question; the 
machine is connected to my server via gigabit ethernet over the onboard 
nforce4 ethernet- my motherboard is  Gigabyte GA-M51GM-S2G Socket AM2 
NVIDIA GeForce 6100, and the CPU is a AMD Sempron 64 2800+.

When I try to rsync from my old server, I get user-land segfaults (bash 
and sh, mostly) and the machine hangs. This happens with straight scp as 
well, and it is 100% replicable. The error messages look like:

sh[5325]: segfault at 0000000081cc893a rip 00002ae181c2bf02 rsp 
00007fffff927c58 error 4

I've now successfully tried copying a large number of files via a 
machine connected through a 100mbit connection, without issue.

This leads me to think that it's forcedeth having the problem, but I had 
also thought that perhaps I'm asking too much of the md/lvm subsystem.

I've ordered another gigabit ethernet card and a new AMD64x2 cpu, so I 
should be able to test these issues more cleanly.

In the meantime, is there anything I can do to diagnose these issues?

Thanks (and please cc me; I'm not a kernel-type by default.)

~ethan fremen