--8NvZYKFJsRX2Djef
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Ok, let me try this again.

We have had a very interesting thread, but it is VERY far from what I was=
=20
asking about, and I'd like input on the original question.

The original question is what to do about the fact that in sddetach(), we
call bufq_free() before all activity on the disk has ended. Thus we can=20
enter code paths that expect to be able to use the bufq routines, when=20
they have already gone away.

So somehow or another we have to not do that.

My thought is to do it later in the detach routine. Any objections?

Further, what's the best way to make sure there are no pending operations=
=20
still on the device? While there is a call to scsipi_kill_pending(), it=20
doesn't actually do anything...

Take care,

Bill

--8NvZYKFJsRX2Djef
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (NetBSD)

iD8DBQFC5stNWz+3JHUci9cRAgZyAJ4lJa0uTJgomNmT22lMNPFILTh0sQCfayQr
P0jhjoQ36O3/05C2soPfS+o=
=9aVT
-----END PGP SIGNATURE-----

--8NvZYKFJsRX2Djef--