From owner-freebsd-current Sun Aug 6 15:56:01 1995 Return-Path: current-owner Received: (from majordom@localhost) by freefall.cdrom.com (8.6.11/8.6.6) id PAA14739 for current-outgoing; Sun, 6 Aug 1995 15:56:01 -0700 Received: from localhost.cdrom.com (localhost.cdrom.com [127.0.0.1]) by freefall.cdrom.com (8.6.11/8.6.6) with SMTP id PAA14730 ; Sun, 6 Aug 1995 15:55:59 -0700 Message-Id: <199508062255.PAA14730@freefall.cdrom.com> X-Authentication-Warning: freefall.cdrom.com: Host localhost.cdrom.com didn't use HELO protocol To: davidg@root.com cc: Mark Murray , freebsd-current@freebsd.org (FreeBSD-current users) Subject: Re: workaround for talk's address problem In-reply-to: Your message of "Sun, 06 Aug 95 15:29:19 PDT." <199508062229.PAA00290@corbin.Root.COM> Date: Sun, 06 Aug 1995 15:55:58 -0700 From: "Justin T. Gibbs" Sender: current-owner@freebsd.org Precedence: bulk >>> > talk(1) has problems with multi-homed hosts. To negotiate the >>> > connection with the remote peer, it uses the first address as returned >>> > by a call to gethostbyname(). This will cause the connection to hang >>> >>> NFS has the exact same problem, FWIW. If there's a more general >>> solution, we should go for it. >> >>Paul Traina fixed a similar problem with kadmin(d?) from eBones a >>couple of days ago. > > What is the problem with specifying INADDR_ANY for the local address? > >-DG For kadmind, that wasn't sufficient. You also had to use getsockname to extract the address of the peer after the connection was accepted since the proper address is required by the protocol. Garrett mentioned something about it only being possible with tcp connections since the binding of addresses is delayed until the accept, but I could have misunderstood him. -- Justin T. Gibbs =========================================== Software Developer - Walnut Creek CDROM FreeBSD: Turning PCs into workstations ===========================================