--tX7/vwywq3jrcw+j
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline

On Tue, Jun 21, 2005 at 10:25:31PM -0400, Allen Briggs wrote:
> I have, however, been toying with adding a 'hw.alldisknames' and
> keeping the components out of 'hw.disknames'.  (Toying with as in
> I have it working here, but I haven't decided if I really like the
> idea of changing the semantics of hw.disknames like that).

I don't think that's a great idea; we don't do it for RF-encapsulated
disks either.  And, as you started out by saying, you do still want to
access the individual disks for SMART and cache control and other such
things.

If a new sysctl node is to be invented, I'd rather that be the
filtered one: hw.availabledisknames or whatever.

Some interaction for locking (ie, returning EBUSY to avoid conflicting
mounts, etc of the same disk sectors) between ld0x and the component
wdNx devices seems like it might be more useful, though I can see that
getting hairy fast too.  Perhaps the wd's should be prevented from
having an in-core label at all, while the ld is open, allowing only
raw access?  (that would prevent the partial-RF usage I described,
though).

Is there a saner way to do this as 'wedges'?

--
Dan.
--tX7/vwywq3jrcw+j
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.1 (NetBSD)

iD8DBQFCuO0BEAVxvV4N66cRAm1hAKDD1s03q2/kMd69uUMSRHq4YlKGgACfRAGM
R6z/TnLrrohVA3wmnCuqDHM=
=JDVU
-----END PGP SIGNATURE-----

--tX7/vwywq3jrcw+j--