netbsd-5 panic in NFS

[Chris Ross <cross+netbsd%distal.com@localhost>]

  I have a dual-UltraSPARC machine (sbus) running a netbsd-5 branch  
kernel and userland as of today/yesterday.  I am 99% sure that I was  
able to run things alright with just the NFS root and swap, and I  
think I even managed to NFS mount NetBSD sources (from the same NFS  
server), and then union-mount a local disk over it.  All of this  
allowed me to do a "df" and see the result.

  However, now twice in a row, I've gotten the following panic:

# df
trap type 0x34: cpu 0, pc=12ab8e4 npc=120ec10 pstate=44820006<PRIV,IE>
kernel trap 34: mem address not aligned
Stopped in pid 572.1 (df) at    netbsd:kauth_cred_ngroups+0x4:   
ld              [%o0 + 0x58], %o0
db{0}> bt
nfs_getattr(bcef418, 8, 164d8e8, 0, 0, 1) at netbsd:nfs_getattr+0x12c
VOP_GETATTR(c9d75b0, bcef5d0, fffffffffffffffe, b7633a4, 0, 1) at  
netbsd:VOP_GETATTR+0x3c
union_allocvp(1, d340000, 0, 0, 20000, c9363a0) at netbsd:union_allocvp 
+0xec
union_lookup(0, 9, 9, 10, 200, 0) at netbsd:union_lookup+0x280
VOP_LOOKUP(c936ac0, bcefbe0, bcefc08, 0, 0, 0) at netbsd:VOP_LOOKUP+0x3c
lookup(bcefbb8, bcef9c8, 20, 0, 0, 0) at netbsd:lookup+0x33c
namei(0, 2650bc0, bcefc08, bcefbb8, 0, 0) at netbsd:namei+0x108
do_sys_pstatvfs(b7633a0, 4080326c, 1, d340970, 0, 0) at  
netbsd:do_sys_pstatvfs+0x20
sys_statvfs1(b7633a0, bcefdc0, bcefe00, 1, 202cc0, 1818c50) at  
netbsd:sys_statvfs1+0x20
syscall_plain(bcefed0, 6, 4053da9c, 4053daa0, 0, bcefdc0) at  
netbsd:syscall_plain+0x134
?(4080326c, 408031a0, 1, 3, fffffffffffffffa, 40803a40) at 0x1008c68

  Anyone have any idea what this is, or how to fix it?  Thanks...   
I'm going to back off to making the mounts one at a time by hand to  
see what I can get working.

                  - Chris