Subject: Re: lseek() extension for spare files
To: Matthew Mondor <mm_lists@pulsar-zone.net>
From: Bill Studenmund <wrstuden@netbsd.org>
List: tech-kern
Date: 09/22/2006 10:10:28
--9amGYk9869ThD9tj
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Fri, Sep 22, 2006 at 08:23:35AM -0400, Matthew Mondor wrote:
> On Thu, 21 Sep 2006 17:07:38 -0700
> Bill Studenmund <wrstuden@netbsd.org> wrote:
>=20
> > I expect that how you'd use it is something like (warning, pseudo-code,=
=20
> > lots of stuff glossed over):
> [...]
>=20
> Interesting...  just out of curiosity, could this by any chance also
> potentially allow to implement the -P option of rm(1) in a way that may
> work on file systems such as LFS (for which -P currently makes no
> sense), if LFS supported the new lseek(2) extensions?  Or would writing
> a block at the start of a data position immediately after the seek still
> cause that block to always be appended?

No, writing is always appended, even overwriting. To make rm -P work, we=20
need to have the file system step in and immediately re-write the segments=
=20
containing data for this file.

Take care,

Bil

--9amGYk9869ThD9tj
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (NetBSD)

iD8DBQFFFBkDWz+3JHUci9cRAtv2AKCXPaRM5oLgTnGlCggJeXw+b44doACgmTLN
tiWahI7f6ezsEBqBaNy17x8=
=oRQp
-----END PGP SIGNATURE-----

--9amGYk9869ThD9tj--