panic after unloading nfsserver module

I've just been trying to get NFS exports working between a -current box
and 4.7-release computer.   Even when following the intructions in the
handbook but running rpcbind instead of portmap, I always get a message
'RPCPROG_MNT: nfs timeout' on the server.   While trying to fix it and
restart nfs, I unloaded the nfsserver module.   The next command appears
to always crash the computer, with the following:

panic: blockable sleep lock (sleep mutex) Giant _at_
/usr/src/sys/vm/vm_fault.c:206

a backtrace gives the following:

panic
witness_lock
_mtx_lock_flags
vm_fault
trap_pfault
trap
calltrap
--- trap 0xc, eip = 0xc01db4f8, ebp = 0xd7c478a8 ---
strlcpy
kill_kinfo_proc
sysctl_kern_proc
sysctl_root
userland_sysctl
__syscall
syscall
Xint0x80_syscall
--- syscall (202, FreeBSD ELF32, __syscall), eip = 0x804fc67, esp =
0xfbffee7c, ebp = 0xbfbfeea8 ---


I cvsupped and rebuilt the world a couple of weeks ago, and it's a
Athlon-XP system I'm using.

Are there known issues with NFS on -current?

--
Bruce Cran