Subject: Re: newfs: determining file system parameters
To: David Laight <david@l8s.co.uk>
From: Bill Studenmund <wrstuden@netbsd.org>
List: tech-kern
Date: 10/20/2003 15:21:01
--6WlEvdN9Dv0WHSBl
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Oct 20, 2003 at 07:12:54PM +0100, David Laight wrote:
> At the moment newfs defaults to the block/fragment sizes from the disklab=
el
> (in spite of suggestions in the man page that the defaults depend on the
> file system size), and also tries to save the values it used into the lab=
el.
>=20
> Oh, how do you intend to squeeze 64bit offset and size into the 16byte
> disklabel entry?

We don't.

You know all this stuff I've been saying about different partitioning=20
systems and all? Part of it is that our current disklabel has a 2 TB=20
limit. We will need to do something else.

We could go with a scheme like what kre was saying where we do scaling. I=
=20
think though a better system is to move towards the new Intel partitioning=
=20
layout. It's 64-bit clean, and gets around many of the issues we've seen=20
w/ MBR labeling.

So de-coupling the kernel "disklabel" from the on-disk partitioning goop=20
will help us a lot. While you're talking about a different aspect of it=20
now, the ioctl I talked about will help with all of this. :-)

Take care,

Bill

--6WlEvdN9Dv0WHSBl
Content-Type: application/pgp-signature
Content-Disposition: inline

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

iD8DBQE/lF/NWz+3JHUci9cRAngjAJ9yGF09LIwDi7b+wzG4CHhkA2IuzwCdGBpU
ONatlXA2w/UCj0Fqv0TLLE0=
=Bhv5
-----END PGP SIGNATURE-----

--6WlEvdN9Dv0WHSBl--