From owner-cvs-all  Tue Aug 15 10:15:17 2000
Delivered-To: cvs-all@freebsd.org
Received: from gatekeeper.whistle.com (gatekeeper.whistle.com [207.76.204.2])
	by hub.freebsd.org (Postfix) with ESMTP
	id 7550237B658; Tue, 15 Aug 2000 10:15:03 -0700 (PDT)
	(envelope-from archie@whistle.com)
Received: from bubba.whistle.com (bubba.whistle.com [207.76.205.7])
	by gatekeeper.whistle.com (8.9.3/8.9.3) with ESMTP id KAA15035;
	Tue, 15 Aug 2000 10:15:03 -0700 (PDT)
	(envelope-from archie@whistle.com)
Received: (from archie@localhost)
	by bubba.whistle.com (8.9.3/8.9.3) id KAA29698;
	Tue, 15 Aug 2000 10:15:02 -0700 (PDT)
	(envelope-from archie)
From: Archie Cobbs <archie@whistle.com>
Message-Id: <200008151715.KAA29698@bubba.whistle.com>
Subject: Re: cvs commit: src/sys/netgraph ng_pppoe.c
In-Reply-To: <200008150917.KAA01236@hak.lan.Awfulhak.org> from Brian Somers at
 "Aug 15, 2000 10:17:36 am"
To: Brian Somers <brian@Awfulhak.org>
Date: Tue, 15 Aug 2000 10:15:02 -0700 (PDT)
Cc: cvs-committers@FreeBSD.ORG, cvs-all@FreeBSD.ORG
X-Mailer: ELM [version 2.4ME+ PL68 (25)]
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Sender: owner-cvs-all@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.ORG

Brian Somers writes:
> > MFC: (rev. 1.26) strip address and control field bytes from outgoing frames
> 
> Is this actually necessary ?  Surely the ppp implementation shouldn't 
> be prepending this stuff ?

Brian-

It depends on what you consider part of the link-layer encapsulation
vs. what is link-layer independent. E.g., a checksum field is clearly
link-dependent, while the protocol field is clearly link independent.
The address/control fields are sortof in between.

RFC 1661 describes address/control fields as part of standard PPP --
hence from that you'd think they are link independent. However, each
link type has something to say about the address and control fields
that is particular to that link type (e.g., in the case of PPPoE,
never send them).

So I decided to include address/control field handling in the ng_ppp(4)
node, since this handling is mandated by RFC 1661 after all, and let
each link type make any necessary "adjustments".

Note that most link types need no "adjustment" -- e.g., async links
(see ng_async(4)), ISDN, frame relay, PPTP, etc. PPPoE is the odd-ball
here, so it is appropriate that it be responsible for making the
adjustment.

>                             I would think that if this functionality 
> is required it should be enabled/disabled via a control message and 
> should work on inbound data too.

It is and it does; please read the ng_ppp(4) man page.

Cheers,
-Archie

___________________________________________________________________________
Archie Cobbs   *   Whistle Communications, Inc.  *   http://www.whistle.com


To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe cvs-all" in the body of the message