--LCbsoPZWBRMqE7TI
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Jun 06, 2005 at 04:01:38PM +0200, Lubomir Sedlacik wrote:
> > That example illustrates exactly where the chroot trick, nice though
> > it is, falls down: what happens when the flash widget is removed and
> > real-/ has to detach?  For that reason, you'd use the md.
>=20
> what happens when you have your / on usb drive and unplug it?  what's
> the difference?

There is no difference, that's exactly what I was talking about. =20

The alternative is to boot and load a kernel from the usb widget, and
have the pre-chroot root in an md image bound into the kernel.  The
usb widget isn't mounted or needed after the kernel has loaded, and
chroot-root is mounted from cgd on the hard disk.

But I'm quite sure you know that, so we must be talking at cross
purposes somehow. What scenario were you envisaging where actually
mounting the usb device as the root filesystem (either before or after
chroot) was desired?

> why isn't the real bug in gettyent() fixed instead of shooting your own
> feet with abandoning the possibility to return back?

I think this is a separate issue, regardless of where things are
mounted from.

--
Dan.

--LCbsoPZWBRMqE7TI
Content-Type: application/pgp-signature
Content-Disposition: inline

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

iD8DBQFCpOVdEAVxvV4N66cRAhwDAJ9swkPEC6Txg7F+adfAdmv+GaCmDgCfQ8NC
95eqJFjsyBO4gcwsnsOBqRY=
=Jpx5
-----END PGP SIGNATURE-----

--LCbsoPZWBRMqE7TI--