Skip site navigation (1)Skip section navigation (2)
Date:      Tue, 28 Mar 2006 08:24:31 -0800
From:      Brooks Davis <brooks@one-eyed-alien.net>
To:        Max Laier <max@love2party.net>
Cc:        freebsd-net@freebsd.org, Brooks Davis <brooks@freebsd.org>
Subject:   Re: Interface groups (from OpenBSD)
Message-ID:  <20060328162431.GA9637@odin.ac.hmc.edu>
In-Reply-To: <200603281131.28240.max@love2party.net>
References:  <200603281131.28240.max@love2party.net>

next in thread | previous in thread | raw e-mail | index | archive | help

--YiEDa0DAkWCtVeE4
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Tue, Mar 28, 2006 at 11:31:22AM +0200, Max Laier wrote:
> Hi,
>=20
> while porting OpenBSD 3.9 (soon to be released) pf I stumbled on interfac=
e=20
> groups.  This is a mechanism to group arbitrary interfaces into logical=
=20
> groups.  It is just naming (not functional change), but it helps to conve=
y=20
> semantic information (e.g. group "LAN", "DMZ" ...) about your interface t=
o=20
> supporting applications.  This way you can write a policies for interface=
=20
> group "LAN" and have it applied to all the VLAN interfaces that come and =
go. =20
> Administration is done via ifconfig.  We currently have "ifconfig name" w=
hich=20
> does part of the job.
>=20
> My question: Does that sound like something interesting for us and should=
 I go=20
> for importing it into FreeBSD proper, or is it not at all interesting and=
 we=20
> don't want it (in which case I'd hack something up for pf).

Sounds like a reasonable feature.  I think it's orthogional to renaming.

> Technical reasoning:  A proper import would add an additional TAILQ link =
into=20
> struct ifnet (which is a great deal of ABI change and causes the usual=20
> headaches).  The hack would use a single void *, but we'd have to pay for=
 the=20
> additional indirection.  Also yet another config tool would be required t=
o=20
> administer the interface <-> group binding.

Adding a TAILQ to the end of struct ifnet would not be an ABI change in 6
because drivers don't know or care how big struct ifnet is anymore and I
can't think of an implementation where the drive code would need to care.

-- Brooks

--=20
Any statement of the form "X is the one, true Y" is FALSE.
PGP fingerprint 655D 519C 26A7 82E7 2529  9BF0 5D8E 8BE9 F238 1AD4

--YiEDa0DAkWCtVeE4
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)

iD8DBQFEKWM+XY6L6fI4GtQRAhoQAJ41pqwaheC1iAd5jcmXk6nPTUQCpACfcYly
wiFAsgULr9EQ9ldqSuWYXwg=
=X5zs
-----END PGP SIGNATURE-----

--YiEDa0DAkWCtVeE4--



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20060328162431.GA9637>