Yes, its happened in different places; I wish I'd written down the
backtrace, as then I could give you more information.

Don't worry, I'm sure I can make it lock again. ;)

A surefire way to trigger a problem is to set vfs.lfs.pagetrip to less
than the logical segment size.  There's also a wholly different issue
which is exposed if vfs.lfs.ignore_lazy_sync=0 is set:  sometimes
(often it was when synching disks at shutdown), I'd get an error about
a vnode being unexpectedly IN_PAGING.  I'm doing a bunch of stuf f at
once right now, but I promise I will provide a more coherent account
for you soon.  I really appreciate any help I can get on this, as most
people don't want to deal with LFS. ;)

Cheers,

--Blair

"De megamaid nil nisi bonum"