A dual Xeon wedged today on me, with sources from May 14 and a kernel
config that is pretty much GENERIC.MPACPI+LOCKDEBUG+DEBUG+DIAGNOSITIC.
The traces from both CPUs are

ffs_balloc(cdc22040,0,1,0,10200) at netbsd:ffs_balloc
ffs_truncate(cdc22040,0,0,0,ffffffff) at netbsd:ffs_truncate+0x162b
ufs_inactive(cc347e14,2e6c6961,676e70,286,c06d25c0) at netbsd:ufs_inactive+0x1fd  
VOP_INACTIVE(cdc22040,cc0561c8,4e5,286,9) at netbsd:VOP_INACTIVE+0x28
vput(cdc22040,cc347eb4,cc0561c8,cc0561c8,8082948) at netbsd:vput+0x116
sys___lstat30(cc0561c8,cc347f64,cc347f5c,c08bfbdc,c03ea1cb) at netbsd:sys___lstat30+0x8a
syscall_plain() at netbsd:syscall_plain+0x1a5

and

_simple_lock_try(c094ba08,c07ff29f,5f4,d,44a2cc1c) at netbsd:_simple_lock_try+0x74
_kernel_lock_acquire_count(1,c07fef77,c080030d,3fd,cb32c4a4) at netbsd:_kernel_lock_acquire_count+0
x6d
mi_switch(cb32c4a4,0,c080030d,22c,0) at netbsd:mi_switch+0x1c5
ltsleep(c092b548,28,c07fe4f0,64,0) at netbsd:ltsleep+0x4cf
sched_sync(cb32c4a4,a2c000,a35000,0,c01002e3) at netbsd:sched_sync+0x17b

I couldn't get a crash dump though (dump device not ready), so is it
worth filing a PR?  I couldn't see anything similar in gnats either.

Simon.