[SunHELP] BAD TRAP occurred in module "ufs" when changing local files over NFS

[Sven.Hartrumpf at FernUni-Hagen.de]

Hi all.

Our Sun panics and reboots after receiving a directory (cp, same can happen
for rm and mv) from a SuSE 8.2 PC.
This is the relevant part of /var/adm/messages:
...
Jul  7 09:27:48 node1 unix: [ID 836849 kern.notice] 
Jul  7 09:27:48 node1 ^Mpanic[cpu0]/thread=30004a1e580: 
Jul  7 09:27:48 node1 unix: [ID 340138 kern.notice] BAD TRAP: type=31 rp=2a10074af00 addr=8 mmu_fsr=0 occurred in module "ufs" due to a NULL pointer dereference
Jul  7 09:27:48 node1 unix: [ID 100000 kern.notice] 
Jul  7 09:27:48 node1 unix: [ID 839527 kern.notice] nfsd: 
Jul  7 09:27:48 node1 unix: [ID 520581 kern.notice] trap type = 0x31
Jul  7 09:27:48 node1 unix: [ID 381800 kern.notice] addr=0x8
Jul  7 09:27:48 node1 unix: [ID 101969 kern.notice] pid=405, pc=0x119bef4, sp=0x2a10074a7a1, tstate=0x4480001601, context=0x1899
...

$ uname -a
SunOS node1 5.9 Generic_112233-06 sun4u sparc SUNW,Ultra-250

This error can be reproduced by copying recursively over NFS any (even small)
directory from a SuSE 8.2 Intel PC to the disk shared by the Sun.
(This error can not be reproduced on SuSE 7.2 (and 8.0?).)
The process on the Linux PC can not be killed easily. As long as this process
exists (it can be there for ever) the Sun will stay in an endless reboot-panic
cycle.

The Linux PC is:
> uname -a
Linux node2 2.4.20-4GB-athlon #1 Wed Jun 18 07:44:55 UTC 2003 i686 unknown unknown GNU/Linux
and contains all security patches (same happens for the shipped non-athlon kernel).

Any ideas would be very welcome.
Greetings
Sven