From owner-freebsd-net  Sun Jul 14 11:58:53 2002
Delivered-To: freebsd-net@freebsd.org
Received: from mx1.FreeBSD.org (mx1.FreeBSD.org [216.136.204.125])
	by hub.freebsd.org (Postfix) with ESMTP
	id EF01B37B400; Sun, 14 Jul 2002 11:58:46 -0700 (PDT)
Received: from iguana.icir.org (iguana.icir.org [192.150.187.36])
	by mx1.FreeBSD.org (Postfix) with ESMTP
	id 9D02D43E5E; Sun, 14 Jul 2002 11:58:46 -0700 (PDT)
	(envelope-from rizzo@iguana.icir.org)
Received: (from rizzo@localhost)
	by iguana.icir.org (8.11.6/8.11.3) id g6EIwka78526;
	Sun, 14 Jul 2002 11:58:46 -0700 (PDT)
	(envelope-from rizzo)
Date: Sun, 14 Jul 2002 11:58:46 -0700
From: Luigi Rizzo <rizzo@icir.org>
To: arch@freebsd.org
Subject: reintroduce handling of NULL route in ip_output()
Message-ID: <20020714115846.C78146@iguana.icir.org>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
User-Agent: Mutt/1.2.5.1i
Sender: owner-freebsd-net@FreeBSD.ORG
Precedence: bulk
List-ID: <freebsd-net.FreeBSD.ORG>
List-Archive: <http://docs.freebsd.org/mail/> (Web Archive)
List-Help: <mailto:majordomo@FreeBSD.ORG?subject=help> (List Instructions)
List-Subscribe: <mailto:majordomo@FreeBSD.ORG?subject=subscribe%20freebsd-net>
List-Unsubscribe: <mailto:majordomo@FreeBSD.ORG?subject=unsubscribe%20freebsd-net>
X-Loop: FreeBSD.org

[discussion on -arch, Bcc to -net in case someone is interested there]

Hi,
I would like to reintroduce the ability to pass a NULL
route to p_output(), which has been possible up to rev.1.34 of
ip_output().

Making ip_output() unable to handle NULL routes only saved a single
"if (ro == NULL)" check, but caused the work of allocating/deallocating
the route to be replicated in multiple places in the clients of
ip_output(), and the way this is done is very inconsistent, sometimes
involving the use of static variables which (apart from potential
errors due to misuse) are a no-no if we ever want to go to fine-grained
locking in the network stack.  See for example how the route to
ip_output is passed in 

        netinet/igmp.c
        netinet/ip_mroute.c
        netinet/ip_input.c
        netinet/tcp_subr.c

at least in one place there is a "bug" (like resetting some fields
of an automatic variable right before exiting from the block).
Other usages look suspicious too.

My reason to re-allow NULL routes in ip_output() are:

  * simplify the code in the callers by centralizing a common
    function;
  * make another step towards the removal of global variables in the 
    protocol stack;

Does any of you have strong and _motivated_ objections ?
 
        thanks
        luigi


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