On 9/20/05, Jason Thorpe <thorpej@shagadelic.org> wrote:
>=20
> On Sep 20, 2005, at 2:16 AM, Julio M. Merino Vidal wrote:
>=20
> >> Let's hold off on "feature" fields for now.. I'm working on a
> >> proposal to clean up the advertising of capabilities and file
> >> attributes.
> >
> > Aha.  So should I go with an '#ifdef NFSSERVER' in the main code,
> > at least for now?
>=20
> Hm, read the diff again.  Actually, I like the "vfsfeatureops"
> thing... but I think it's misnamed.  What it really is is a
> notification mechanism for when volumes get mounted and unmounted.
> Maybe you just need to rename it.

Indeed.  The 'features' name comes from the initial changes I did.  At firs=
t,
I had a bit field indicating several features a file system could support, =
and
the file system had to set several flags in it.  One of these features was
"NFS support", but this concept is not applicable any more (the bit field
has gone away, and the idea of a "NFS feature" is no longer applicable).

What about vfshooks, and then make each member match the
corresponding's VFS operation name?  I.e., hook_mount, hook_unmount,
etc.  We might add other hooks in the future for different operations.

Kind regards,

--=20
Julio M. Merino Vidal <jmmv84@gmail.com>
http://www.livejournal.com/users/jmmv/
The NetBSD Project - http://www.NetBSD.org/