panic: _sx_xlock (user map): xlock already held _at_ ../../../vm/vm_map.c:2997

I got this while (net)booting an up-to-date 6.0 kernel:

Trying to mount root from nfs:
NFS ROOT: 216.136.204.22:/a/nfs/6.dir1
fxp0: link state changed to UP
panic: _sx_xlock (user map): xlock already held _at_ ../../../vm/vm_map.c:2997
db> wh
Tracing pid 1 tid 100006 td 0xc14e6600
kdb_enter(c06bc443,c077d7a0,c06bcaa0,cbf2268c,100) at kdb_enter+0x30
panic(c06bcaa0,c06a38cb,c06d0ea5,c06d0f25,bb5) at panic+0xd5
_sx_xlock(c14f2044,c06d0f25,bb5,c05726d3,cbf226e0) at _sx_xlock+0x62
_vm_map_lock_read(c14f2000,c06d0f25,bb5,1000101,9cc21000) at _vm_map_lock_read+0x4a
vm_map_lookup(cbf22778,9cc21000,1,cbf2277c,cbf2276c) at vm_map_lookup+0x38
vm_fault(c14f2000,9cc21000,1,0,c14e6600) at vm_fault+0x7e
trap_pfault(cbf22860,0,9cc21603,c13d1bb8,9cc21603) at trap_pfault+0xed
trap(c04d0008,c1040028,28,0,73) at trap+0x35e
calltrap() at calltrap+0x5
--- trap 0xc, eip = 0xc0633eaa, esp = 0xcbf228a0, ebp = 0xcbf228d4 ---
vm_map_pmap_enter(c14f2000,80bc000,3,c1046bdc,73) at vm_map_pmap_enter+0x14a
vm_map_insert(c14f2000,c1046bdc,73000,0,80bc000) at vm_map_insert+0x2f4
elf32_map_insert(c14f2000,c1046bdc,73000,0,80bc000) at elf32_map_insert+0x2d9
elf32_load_section(c14e5c48,c14f2000,c1693990,c1046bdc,736a0) at elf32_load_section+0x117
exec_elf32_imgact(cbf22b9c,0,c06b9a9d,142,12b) at exec_elf32_imgact+0x286
do_execve(c14e6600,cbf22c78,0,3,0) at do_execve+0x24c
kern_execve(c14e6600,cbf22c78,0,bfbfffe4,0) at kern_execve+0xb9
execve(c14e6600,cbf22ce4,0,0,cbf22ce0) at execve+0x4f
start_init(0,cbf22d38,c06b9bd0,30d,366eb4ad) at start_init+0x2de
fork_exit(c04b65e0,0,cbf22d38) at fork_exit+0xc1
fork_trampoline() at fork_trampoline+0x8
--- trap 0x1, eip = 0, esp = 0xcbf22d6c, ebp = 0 ---
db>

Unfortunately since it panicked early in boot dumpdev is not set.  I
will leave the box in DDB in case there is anything else I can get
from it for someone.

Kris