From owner-freebsd-bugs Sun Jan 4 02:14:17 1998 Return-Path: Received: (from root@localhost) by hub.freebsd.org (8.8.7/8.8.7) id CAA16167 for bugs-outgoing; Sun, 4 Jan 1998 02:14:17 -0800 (PST) (envelope-from owner-freebsd-bugs) Received: from sax.sax.de (sax.sax.de [193.175.26.33]) by hub.freebsd.org (8.8.7/8.8.7) with ESMTP id CAA16162; Sun, 4 Jan 1998 02:14:14 -0800 (PST) (envelope-from j@uriah.heep.sax.de) Received: (from uucp@localhost) by sax.sax.de (8.8.8/8.8.8) with UUCP id LAA09862; Sun, 4 Jan 1998 11:14:02 +0100 (CET) (envelope-from j@uriah.heep.sax.de) Received: (from j@localhost) by uriah.heep.sax.de (8.8.8/8.8.5) id LAA17951; Sun, 4 Jan 1998 11:05:21 +0100 (MET) Message-ID: <19980104110521.14399@uriah.heep.sax.de> Date: Sun, 4 Jan 1998 11:05:21 +0100 From: J Wunsch To: Greg Lehey Cc: Brian Somers , John-Mark Gurney , freebsd-bugs@hub.freebsd.org, freebsd-hackers@FreeBSD.ORG Subject: Re: kern/5404: slXX slip (tun & ppp) interfaces always point to point Reply-To: Joerg Wunsch References: <199801010130.RAA10049@hub.freebsd.org> <199801011325.NAA17803@awfulhak.demon.co.uk> <19980102105504.61189@lemis.com> <19980102102027.41384@uriah.heep.sax.de> <19980104174838.41538@lemis.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.88 In-Reply-To: <19980104174838.41538@lemis.com>; from Greg Lehey on Sun, Jan 04, 1998 at 05:48:38PM +1030 X-Phone: +49-351-2012 669 X-PGP-Fingerprint: DC 47 E6 E4 FF A6 E9 8F 93 21 E0 7D F9 12 D6 4E Sender: owner-freebsd-bugs@FreeBSD.ORG X-Loop: FreeBSD.org Precedence: bulk As Greg Lehey wrote: > I had a discussion about this with a bloke here in Adelaide a couple > of months ago. He runs a large ISP, and he came up with some > plausible reason, but unfortunately I've forgotten the details. It > had to do with Microslop: they use broadcasts a lot, and this would abuse :) > seem to indicate that they expected broadcasts on a /26 subnet, or at > least were prepared to respond to them. So this netmask crap might apply for M$ then -- they still have a long way to be internet-ready. For us, an interface is either IFF_BROADCAST, or IFF_POINTOPOINT, but not both. Thus, the netmask and broadcast addresses of a p2p interface are irrelevant. (If you think about it, both, broadcasting and p2p are mutually exclusive by logic. p2p is always unicasting to just one peer.) > It sounds like you're saying that PPP shouldn't be allowed to set the > default route automatically when the link comes up. No, i was really telling that there should be _only_ the implied host route automatically installed when upping a p2p interface, but no other route. ISTR Brian suggested a second route should be derived from the (bogus) netmask, and installed {too or instead}. So per Brian's suggestion: ifconfig foobar0 1.2.3.4 1.4.5.6 netmask 0xffffff00 would have implied route add 1.4.5.6 -iface foobar0 route add -net 1.4.5.0 -netmask 0xffffff00 1.4.5.6 while the existing behaviour (and what i think is the Right Thing) is only the first of both routes. Installing a default route is always done separately anyway. -- cheers, J"org joerg_wunsch@uriah.heep.sax.de -- http://www.sax.de/~joerg/ -- NIC: JW11-RIPE Never trust an operating system you don't have sources for. ;-)