I'm seeing this on two machines. Initially on 3_STABLE, and now on
 	a 3.99.21 'GENERIC + PF'.

 	Both seem to fail occasionally under heavy IO (One during heavy
 	rsync usage and the other during postgres database backup).

 	Both test clean on memtester, and neither are overclocked.

 	One runs raidframe, the other not.

 	Both run pf.

 	Kernel: http://www.mono.org/abs/uvm_fault/netbsd.bz2
 	Dmesg:  http://www.mono.org/abs/uvm_fault/dmesg

 	Any thoughts?

uvm_fault(0xc0964ea0, 0xd39b7000, 1) -> 0xe
kernel: supervisor trap page fault, code=0
Stopped in pid 26146.1 (find) at        netbsd:cache_lookup+0x5d:       cmpl    0
x20(%ebx),%eax
db> bt
cache_lookup(d0c429ec,cb71ebe8,cb71ebfc,dc035198,d0c429ec) at netbsd:cache_lookup+0x5d
ufs_lookup(cb7aea0c,1000,c072efa0,d0c429ec,cb71ebe8) at netbsd:ufs_lookup+0xde
VOP_LOOKUP(d0c429ec,cb71ebe8,cb71ebfc,200,0) at netbsd:VOP_LOOKUP+0x2b
lookup(cb71ebd8,cdcae800,400,cb71ebf0,286f000) at netbsd:lookup+0x280
namei(cb71ebd8,44ad994c,0,d114d27c,d1150e7c) at netbsd:namei+0xf7
compat_30_sys___lstat13(dc035198,cb71ec48,cb71ec68,206,c1b11580) at netbsd:compat_30_sys___lstat13+0x49
syscall_plain() at netbsd:syscall_plain+0x147
--- syscall (number 280) ---
0xbbbbbd53: